Analytical Separation Science 9783527333745

Citation preview

Edited by Jared L. Anderson Alain Berthod Verónica Pino Estévez Apryll M. Stalcup

Analytical Separation Science

Editors Prof. Jared L. Anderson

The University of Toledo Department of Chemistry & Biochemistry 2801 W. Bancroft St., MS 602 OH United States

All books published by Wiley-VCH are carefully produced. Nevertheless, authors, editors, and publisher do not warrant the information contained in these books, including this book, to be free of errors. Readers are advised to keep in mind that statements, data, illustrations, procedural details or other items may inadvertently be inaccurate. Library of Congress Card No.: applied for

Prof. Alain Berthod

Uni. Claude-Bernard, Lyon 1 Bat. CPE-Lyon 308-D 69622 Villeurbanne Cedex France Prof. Verónica Pino Estévez

University of La Laguna C/Molinos de Agua 1 38207 San Cristobal la Laguna Spain Apryll M. Stalcup

Dublin City University Irish Separation Science Cluster Glasnevin 9 Dublin Ireland

British Library Cataloguing-in-Publication Data A catalogue record for this book is available from the British Library. Bibliographic information published by the Deutsche Nationalbibliothek The Deutsche Nationalbibliothek lists this publication in the Deutsche Nationalbibliografie; detailed bibliographic data are available on the Internet at http://dnb.d-nb.de.  2015 Wiley-VCH Verlag GmbH & Co. KGaA, Boschstr. 12, 69469 Weinheim, Germany All rights reserved (including those of translation into other languages). No part of this book may be reproduced in any form – by photoprinting, microfilm, or any other means – nor transmitted or translated into a machine language without written permission from the publishers. Registered names, trademarks, etc. used in this book, even when not specifically marked as such, are not to be considered unprotected by law. Print ISBN: 978-3-527-33374-5 Cover Design Formgeber, Mannheim, Germany Typesetting Thomson Digital, Noida, India Printed on acid-free paper

V

Contents About the Editors XVII Preface XIX List of Contributors XXI

Volume 1 1

Basic HPLC Theory and Definitions: Retention, Thermodynamics, Selectivity, Zone Spreading, Kinetics, and Resolution 1 Torgny Fornstedt, Patrik Forssén, and Douglas Westerlund

1.1 1.1.1 1.1.2 1.1.3 1.1.3.1 1.1.3.2 1.1.3.3 1.1.4 1.2 1.3 1.3.1 1.3.2 1.3.3 1.4

Basic Definitions 2 Basic Retention Models and Kinetics 6 Band Broadening and the Plate Height Concept 7 Sources of Zone Broadening 9 Eddy Diffusion 10 Molecular Diffusion 10 Slow Equilibration 10 Dependence of Zone Broadening on Flow Rate 11 Resolution 12 Modern Trends in Liquid Chromatography 14 Efficiency Trend 15 Permeability Trend 17 Selectivity and New Material Trend 19 Conclusions 21 References 22

2

Basic LC Method Development and Optimization Victoria F. Samanidou

2.1 2.2 2.2.1 2.2.2 2.2.3

Introduction 25 Theoretical Aspects 26 Retention Factor k 27 Selectivity α 27 Peak Asymmetry 27

25

VI

Contents

2.2.4 2.2.5 2.2.6 2.3 2.3.1 2.3.1.1 2.3.2 2.3.3 2.3.3.1 2.3.3.2 2.3.3.3 2.3.3.4 2.3.3.5 2.4 2.4.1 2.4.2 2.5 2.5.1 2.6

Efficiency of Chromatographic Column and Theoretical Plates 27 Resolution Rs 28 The Fundamental vanDeemter Equation 29 Controlling Resolution 30 How to Improve N 32 Physical Characteristics of Packing Material 32 Increase of k 33 Factors Influencing Selectivity or How to Improve α? 33 Optimization of Mobile-Phase Composition 34 pH Control, Ion-Pair Reagents, and Other Additives 35 Temperature 35 Stationary Phase and Column Selection 35 Stationary Phase and Packing Material Composition 36 Method Development Strategy 37 Gradient Elution versus Isocratic 38 Other Parameters in LC Method Development 38 Current and Future Trends 39 Two-Dimensional Chromatography 39 Conclusions 40 References 40

3

Recent Advances in Column Technology Ross Andrew Shalliker and Danijela Kocic

43

Introduction 43 Column Packing: Downward Slurry Packing 45 Column Bed Heterogeneity 46 Axial Heterogeneity 46 Radial Heterogeneity and the Wall Effect 49 Active Flow Technology: A New Design Concept in Chromatography Columns 51 3.4.1 AFT Columns: Parallel Segmented Flow 51 3.4.2 AFT Columns: Curtain Flow 52 3.4.3 Performance of AFT Columns 53 3.4.3.1 Sensitivity 53 3.4.3.2 Efficiency 54 3.4.3.3 Speed 58 3.5 Summary 60 References 61 3.1 3.2 3.3 3.3.1 3.3.2 3.4

4

4.1 4.2 4.3 4.3.1

Hydrophilic Interaction Liquid Chromatography Xinmiao Liang, Aijin Shen, and Zhimou Guo

63

Introduction 63 Separation Mechanism in HILIC 64 Stationary Phases for HILIC 67 Conventional NPLC Stationary Phases for HILIC 67

Contents

4.3.2 4.3.2.1 4.3.2.2 4.3.2.3 4.3.2.4 4.4 4.4.1 4.4.2 4.4.3 4.4.4 4.5

Stationary Phases Developed for HILIC 75 Polyaspartamide-Based Stationary Phases 75 Amide-Based Stationary Phases 75 Saccharides-Based Stationary Phases 76 Zwitterionic Stationary Phases 76 Application of HILIC 77 Application in the Pharmaceutical Field 77 Application in the Separation of Carbohydrates 78 Application in Proteome, Glycoproteome, and Phosphoproteome 78 Application in Metabolomics/Metabonomics 80 Conclusions and Outlook 81 References 81

5

LC–MS Interfaces 87 Pierangela Palma, Elisabetta Pierini, and Achille Cappiello

5.1 5.2 5.2.1 5.2.1.1 5.2.1.2 5.2.1.3 5.2.2 5.2.2.1 5.2.3 5.2.3.1 5.2.4 5.2.4.1 5.3 5.3.1 5.3.2 5.3.3 5.3.4

6

6.1 6.2 6.2.1 6.2.2 6.2.3 6.3 6.3.1 6.3.2

Introduction 87 API Sources 88 Electrospray Interface (ESI) 89 Principles of Operation and Ion Formation 90 Factors Influencing ESI Response 92 Modes of Operation 92 Atmospheric Pressure Chemical Ionization 93 Principles of Operation and Ion Formation 94 Atmospheric Pressure Photoionization 95 Principle of Operation 96 Atmospheric Pressure Laser Ionization 98 Principle of Operation and Ion Formation 98 Non-API Sources 99 Direct-EI 100 EI of Cold Molecules in Supersonic Molecular Beam (SMB) 103 Combined Single-Photon Low-Pressure Photoionization and EI Ionization 104 LC/DESI–MS Interface 106 References 107 LC–MS Applications in Environmental and Food Analysis 111 Alessandra Gentili, Fulvia Caretti, and Virginia Pérez Fernández

Introduction 111 Environmental Applications 112 Last Trends in Sample Preparation for LC–MS Analysis 112 Advances and Trends in Liquid Chromatography 113 Advances and Trends in Mass Spectrometry 113 Food Toxicant Applications 117 Recent Trends in Sample Preparation for LC–MS Analysis 117 Recent Trends in LC–MS Screening Analysis 118

VII

VIII

Contents

6.3.3 6.4 6.4.1 6.4.2 6.4.3 6.5

7

7.1 7.2 7.2.1 7.2.2 7.2.3 7.2.4 7.2.5 7.3 7.3.1 7.3.2 7.3.3 7.4 7.4.1 7.4.2 7.4.3 7.5

8

8.1 8.2 8.2.1 8.2.2 8.2.3 8.2.4 8.3 8.3.1 8.3.2 8.3.3 8.4

Recent Trends in LC–MS Confirmatory Analysis 120 Foodomics as a Recent Approach Embracing Metabolomics, Proteomics, and Lipidomics 121 Food Proteomics 121 Food Metabolomics 124 Food Lipidomics 125 Trends and Future Developments 127 References 128 Solvents in Chromatography and Electrophoresis Alain Berthod and Karine Faure

135

Introduction 135 Physicochemical Properties of Solvents 135 Melting and Boiling Points, and Vapor Pressure 135 Molecular Weight, Density, and Molar Volume 136 Viscosity, Surface Tension, UV Cutoff, and Refractive Index 136 Solvent Polarity Scales 137 New Solvents 142 Physicochemical Properties of Mixtures of Solvents 143 Fully Miscible Solvents 143 Nonfully Miscible Solvents and Phase Diagrams 144 Solvent Mixtures and Chromatographic Retention Times: Elution Strength 146 Mobile-Phase pH and Buffers 147 pH Definition 147 pH in Hydro-organic Mobile Phases 147 pKa Shifts in Hydro-organic Mobile Phases 148 Conclusions 151 Acknowledgments 157 References 157

Reversed Phase Liquid Chromatography 159 Maria C. García-Alvarez-Coque, Juan J. Baeza-Baeza, and Guillermo Ramis-Ramos

Introduction 159 The Stationary Phase 160 Silica Support and Chemical Bonding 161 Types of Phases 163 Silanol Effects 164 Silanol Deactivation 166 The Mobile Phase 167 Mobile Phase Components 167 Snyder’s Solvent Selectivity Triangle 168 Control of the Mobile-Phase pH 170 Temperature as Chromatographic Factor 172

Contents

8.5 8.5.1 8.5.2 8.5.3 8.5.4 8.5.5 8.5.6 8.6 8.6.1 8.6.2 8.6.3 8.6.4 8.6.5 8.6.5.1 8.6.5.2 8.6.5.3 8.6.5.4 8.7

9

9.1 9.2 9.2.1 9.2.2 9.2.3 9.3 9.3.1 9.3.2 9.4 9.4.1 9.4.2 9.5 9.5.1 9.5.2 9.6 9.6.1 9.6.2 9.7

Gradient versus Isocratic Elution 174 Solute Retention and Peak Width 174 Isocratic Elution 175 Gradients of Modifier: The Usual Solution for the General Elution Problem 175 Development of Gradients of Modifier 176 Strengths and Weaknesses of Gradients of Modifier 179 Other Types of Gradients 181 Attempts to Explain the Retention Mechanisms in RPLC 181 Solvent Adsorption and Partitioning in RPLC 181 The Solvophobic Theory 182 Solute Adsorption or Partitioning? 183 Investigating How RPLC Really Works 184 Going Down to the Molecular Detail 186 Chain Conformation 186 Adsorption and Partitioning of Common Solvents 186 Adsorption and Partitioning of Solutes 188 Anomalous Behavior with Highly Aqueous Mobile Phases 189 Development and Trends in RPLC 190 References 192 Modeling of Retention in Reversed Phase Liquid Chromatography Maria C. García-Alvarez-Coque, Guillermo Ramis-Ramos, José R. Torres-Lapasió, and C. Ortiz-Bolsico

Introduction 199 Isocratic Elution 199 Polynomial Models to Describe Retention Using Modifier Content as a Factor 199 Polarity Models 201 pH as an Experimental Factor 202 Dead Time Estimation 206 Static Methods 207 Dynamic Methods 207 Effect of Temperature 209 Van’t Hoff Equation 209 Combined Effect of Modifier Content, pH, and Temperature 210 Effect of Pressure 211 Deviations of Retention Factors 211 Correction of Pressure Effects 212 Enhancing the Prediction of Retention 214 Practical Considerations 214 Influence of the Model Regression Process on the Quality of Predictions 215 Gradient Elution 216

199

IX

X

Contents

9.7.1 9.7.2 9.8 9.9 9.9.1 9.9.2

Integration of the Fundamental Equation for Gradient Elution 216 Nonintegrable Retention Models 217 Computer-Assisted Interpretive Optimization 218 Stationary-Phase Characterization 220 Linear Solvation Energy Relationships 220 Local Models for Characterizing RPLC Columns 221 References 223

10

Normal-Phase and Polar Organic Solvents Chromatography 227 Ahmed A. Younes, Charlene Galea, Debby Mangelings, and Y. Vander Heyden

10.1 10.2 10.2.1 10.2.2 10.2.3 10.2.4 10.3 10.3.1 10.3.2 10.3.2.1 10.3.2.2 10.3.2.3 10.3.3 10.3.3.1 10.3.3.2 10.3.3.3 10.4

11

11.1 11.2 11.2.1 11.2.2 11.2.3 11.2.4 11.2.5 11.3 11.3.1 11.3.2 11.4 11.5 11.6

Introduction 227 HPLC Retention and Separation Mechanisms 228 Polarity-Based Separations 228 Charge-Based Separations 232 Size-Based Separations 232 Other Separation Mechanisms 232 Normal-Phase and Polar Organic Solvents Chromatography 233 Retention Mechanism 234 Stationary Phases 234 Nonbonded Phases 234 Bonded Phases 235 Stationary Phases and Selectivity 236 Mobile Phases 238 Mobile-Phase Selection 238 Solvent Strength and Selectivity 239 Isocratic and Gradient Elution 241 Conclusions 242 References 243

Inline Detectors 245 Ramisetti Nageswara Rao and Pothuraju Nageswara Rao

Introduction 245 Detector Characteristics 246 Sensitivity 246 Selectivity 246 Linearity 247 Dynamic Range 247 Detector Cell Volume 247 UV-Visible Absorbance Detector 247 Fixed Wavelength Detector 249 Variable Wavelength Detector 250 Photodiode Array Detector (PDA) 251 Fluorescence Detector 252 Refractive Index Detector (RID) 255

Contents

11.7 11.8 11.9 11.10 11.11 11.12

Evaporative Light-Scattering Detector 256 Electrochemical Detector 257 Charged Aerosol Detection 258 Conductivity Detector 259 Coupling Detectors 260 Comparison of HPLC Detectors 260 References 261

12

pH Effects on Chromatographic Retention Modes Paweł Wiczling, Łukasz Kubik, and Roman Kaliszan

263

12.1 12.2 12.3 12.4 12.5 12.6 12.7 12.8

Introduction 263 pH Measurements of Mobile Phase 264 Effect of pH on Isocratic Retention 266 pH Effect on Organic Modifier Gradients 268 pH Gradient 269 Determination of pKa, log kw (Hydrophobicity), and S 274 Effect of pH in Normal-Phase Mode 275 Summary 277 References 277

13

Chemometrics in Data Analysis and Liquid Chromatographic Method Development 279 Biljana Jančic ́-Stojanovic ́ and Tijana Rakic ́

Introduction 279 Chemometrics in Data Analysis 280 Data Preprocessing 280 Data Analysis 284 Chemometrics in LC Method Development 285 Analytical Target Profile and Critical Quality Attributes (Definition of the Objectives of the Method) 286 13.3.2 Quality Risk Assessment and Critical Process Parameters (Definition of Investigated Factors and Their Levels) 287 13.3.3 Investigation of the Knowledge Space (Selection of an Appropriate Experimental Design) 288 13.3.3.1 Screening Designs 289 13.3.3.2 Optimization Designs 291 13.3.4 Critical Quality Attributes Modeling (Creation of Mathematical Models) 293 13.3.5 Design Space 294 13.3.6 Selection of the Working Points 295 13.3.7 Robustness Testing 295 13.4 Conclusions 296 References 296 13.1 13.2 13.2.1 13.2.2 13.3 13.3.1

Index to Volume 1 I1-I18

XI

XII

Contents

Volume 2

Part One Special Liquid Chromatography Modes 299 1

Chiral Liquid Chromatography: Recent Applications with Special Emphasis on the Enantioseparation of Amino Compounds 301 István Ilisz

2

Chiral Separation of Some Classes of Pesticides by HPLC Method 321 Imran Ali, Iqbal Hussain, Mohd Marsin Sanagi, and Hassan Y. Aboul-Enein

3

Micellar Liquid Chromatography: Fundamentals 371 Maria C. García-Alvarez-Coque, Maria J. Ruiz-Angel, and Samuel Carda-Broch

4

Micellar Liquid Chromatography: Method Development and Applications 407 Maria C. García-Alvarez-Coque, Maria J. Ruiz-Angel, and Samuel Carda-Broch

5

Affinity Chromatography 461 Erika L. Pfaunmiller, Jesbaniris Bas, Marissa Brooks, Mitchell Milanuk, Elliott Rodriguez, John Vargas, Ryan Matsuda, and David S. Hage

6

Immunoaffinity Chromatography: Advantages and Limitations Nancy E. Thompson and Richard R. Burgess

Part Two

483

Capillary Electromigration Techniques 503

7

Capillary Electromigration Techniques: Capillary Electrophoresis Václav Kašička

8

Modern Injection Modes (Stacking) for CE Joselito P. Quirino

9

Capillary Gel Electrophoresis 555 Márta Kerékgyártó and András Guttman

10

Nonaqueous Capillary Electrophoresis Julie Schappler and Serge Rudaz

581

11

Detectors in Capillary Electrophoresis Petr Tůma and František Opekar

607

12

Trends in CE-MS and Applications Anna Tycova and Frantisek Foret

13

Capillary Electrochromatography Kai Zhang and Ruyu Gao

629 653

531

505

Contents

14

Micellar Electrokinetic Chromatography 675 Paolo Iadarola, Marco Fumagalli, and Simona Viglio

15

Chip-Based Capillary Electrophoresis 707 Yuanhong Xu, Jizhen Zhang and Jingquan Liu

16

Chiral Separations by Capillary Electrophoresis 731 E. Sánchez-López, M. Castro-Puyana, M.L. Marina, and A.L. Crego

Index to Volume 2 I1-I24

Volume 3 1

Gas Chromatography: Theory and Definitions, Retention and Thermodynamics, and Selectivity 775 Glenn E. Spangler

2

Basic Overview on Gas Chromatography Injectors 807 Md. Musfiqur Rahman, A.M. Abd El-Aty, and Jae-Han Shim

3

Basic Overview on Gas Chromatography Columns 823 Md. Musfiqur Rahman, A.M. Abd El-Aty, Jeong-Heui Choi, Ho-Chul Shin, Sung Chul Shin, and Jae-Han Shim

4

Overview of Detectors in Gas Chromatography 835 Md. Musfiqur Rahman, A.M. Abd El-Aty, and Jae-Han Shim

5

Current Use of Gas Chromatography and Applications 849 Walter Vetter

6

Gas Chromatography with Mass Spectrometry (GC-MS) Walter Vetter

7

Chiral GC 927 Volker Schurig

8

New Essential Events in Modern Applications of Inverse Gas Chromatography 979 Adam Voelkel, Henryk Grajek, Beata Strzemiecka, and Katarzyna Adamska

9

Chip-Based Gas Chromatography Hamza Shakeel and Masoud Agah

999

883

XIII

XIV

Contents

1021

10

Portable Gas Chromatography Philip A. Smith

11

Packed Column Sub- and Supercritical Fluid Chromatography Caroline West, Syame Khater, and Eric Lesellier

12

Instrumentation for Sub- and Supercritical Fluid Chromatography Taghi Khayamian, Ali Daneshfar, and Hassan Ghaziaskar

1051

Index to Volume 3 I1-I18 Volume 4 1

High-Performance Thin-Layer Chromatography 1093 Vicente L. Cebolla, Luis Membrado, Carmen Jarne, and Rosa Garriga

2

Field Flow Fractionation 1143 Gaëtane Lespes, Julien Gigault, and Serge Battu

3

Separations with a Liquid Stationary Phase: Countercurrent Chromatography or Centrifugal Partition Chromatography 1177 Alain Berthod and Karine Faure

4

Preparative Chromatography: Batch and Continuous 1207 José P.S. Aniceto and Carlos M. Silva

5

Fast and Miniaturized Chromatography 1315 Bárbara Socas-Rodríguez, Antonio V. Herrera-Herrera, Miguel Ángel González-Curbelo, Javier González-Sálamo, and Javier Hernández-Borges

6

Two-Dimensional Liquid Chromatography 1357 Morgan Sarrut, Nicola Marchetti, and Sabine Heinisch

Index to Volume 4 I1-I14

Volume 5 1

Sampling Strategies: Statistics of Sampling 1385 Małgorzata Bodnar, Piotr Konieczka, and Jacek Namieśnik

2

Targeted and Non-Targeted Analysis 1401 Luis E. Rodriguez-Saona, Marçal Plans Pujolras, and M. Monica Giusti

1075

Contents

3

Conventional Extraction Techniques: Soxhlet and Liquid–Liquid Extractions and Evaporation 1437 Adegbenro Peter Daso and Okechukwu Jonathan Okonkwo

4

Main uses of Microwaves and Ultrasounds in Analytical Extraction Schemes: an Overview 1469 Idaira Pacheco-Fernández, Providencia González-Hernández, Priscilla RocíoBautista, María José Trujillo-Rodríguez, and Verónica Pino

5

Membrane-assisted Separations Jan Åke Jönsson

6

Dispersive Solid-Phase Extraction 1525 Bárbara Socas-Rodríguez, Antonio V. Herrera-Herrera, María Asensio-Ramos, and Javier Hernández-Borges

7

Solid-Phase Extraction 1571 Nil Ozbek, Asli Baysal, Suleyman Akman, and Mehmet Dogan

8

Solid-Phase Microextraction 1595 Ali Mehdinia and Mohammad Ovais Aziz-Zanjani

9

Liquid-Phase Microextraction 1625 Mohammad Reza Ganjali, Morteza Rezapour, Parviz Norouzi, and Farnoush Faridbod

10

Analytical Supercritical Fluid Extraction 1659 Julian Martínez and Ana Carolina de Aguiar

11

Extraction Methods Facilitated by the use of Magnetic Nanoparticles 1681 Priscilla Rocío-Bautista and Verónica Pino

12

Sample Derivatization in Separation Science 1725 Pascal Cardinael, Hervé Casabianca, Valerie Peulon-Agasse, and Alain Berthod

13

Validation of Analytical Methods Based on Chromatographic Techniques: An Overview 1757 Juan Peris-Vicente, Josep Esteve-Romero, and Samuel Carda-Broch

14

“Omics” and Biomedical Applications 1809 Pasquale Ferranti, Chiara Nitride, and Monica Gallo

15

Food Applications: Using Novel Sample Preparation Modes Mónica González and Venerando González

1503

1859

XV

XVI

Contents

16

Forensic Applications 1877 Matías Calcerrada Guerreiro, María López-López, Ma Ángeles Fernández de la Ossa, and Carmen García-Ruiz

17

Environmental Applications of Solid Phase Microextraction Techniques 1897 Sarah Montesdeoca-Esponda, M Esther Torres-Padrón, Zoraida Sosa-Ferrera, and José Juan Santana-Rodríguez

Index to Volume 5 I1-I20 Index 1929

Editors Prof. Jared L. Anderson

The University of Toledo Department of Chemistry & Biochemistry 2801 W. Bancroft St., MS 602 OH United States

All books published by Wiley-VCH are carefully produced. Nevertheless, authors, editors, and publisher do not warrant the information contained in these books, including this book, to be free of errors. Readers are advised to keep in mind that statements, data, illustrations, procedural details or other items may inadvertently be inaccurate. Library of Congress Card No.: applied for

Prof. Alain Berthod

Uni. Claude-Bernard, Lyon 1 Bat. CPE-Lyon 308-D 69622 Villeurbanne Cedex France Prof. Verónica Pino Estévez

University of La Laguna C/Molinos de Agua 1 38207 San Cristobal la Laguna Spain Apryll M. Stalcup

Dublin City University Irish Separation Science Cluster Glasnevin 9 Dublin Ireland

British Library Cataloguing-in-Publication Data A catalogue record for this book is available from the British Library. Bibliographic information published by the Deutsche Nationalbibliothek The Deutsche Nationalbibliothek lists this publication in the Deutsche Nationalbibliografie; detailed bibliographic data are available on the Internet at http://dnb.d-nb.de.  2015 Wiley-VCH Verlag GmbH & Co. KGaA, Boschstr. 12, 69469 Weinheim, Germany All rights reserved (including those of translation into other languages). No part of this book may be reproduced in any form – by photoprinting, microfilm, or any other means – nor transmitted or translated into a machine language without written permission from the publishers. Registered names, trademarks, etc. used in this book, even when not specifically marked as such, are not to be considered unprotected by law. Print ISBN: 978-3-527-33374-5 Cover Design Formgeber, Mannheim, Germany Typesetting Thomson Digital, Noida, India Printed on acid-free paper

1

1 Basic HPLC Theory and Definitions: Retention, Thermodynamics, Selectivity, Zone Spreading, Kinetics, and Resolution Torgny Fornstedt, Patrik Forssén, and Douglas Westerlund

Liquid chromatography is a very important separation method used in practically all chemistry fields. For many decades, it has played a key role in academic and industrial laboratories where it is used to analyze or purify components from complex mixtures. For example, it is used to separate proteins/drugs from impurities and to analyze drugs and endogenous components in biological materials. Most breakthroughs in biochemical and pharmaceutical sciences would probably not have been possible without chromatography. Chromatography is generally considered to have been developed in the early twentieth century by the Russian botanist Tswett. He found that he could separate components from plant extracts by flushing a sample with organic solvents through a glass tube packed with an inorganic adsorbent. Distinct bands of various colors evolved and migrated at different rates down the column. The bands corresponding to the different plant pigments could be collected at the outlet at the bottom of the tube. Tswett chose to call his technique “chromatography,” which means “color writing” in Greek. The name has been kept for historical reasons, although it is not very descriptive of the method in general. His publications had, however, little impact, and the technique fell into oblivion for several decades. Chromatography is based on the partitioning of solutes between two phases and is, therefore, related to simple liquid–liquid extraction. In chromatography, however, one phase (the mobile phase) is in constant movement relative to the other one (the stationary phase). The sample molecules are partitioned between the phases; those in the stationary phase are retained, whereas those in the mobile phase move. The interaction between the solutes and the stationary phase is most often based on adsorption. During a chromatographic separation, a solute normally partitions between the phases many thousand times. The basis of separation is that different kinds of molecules on average spend different amount of time in the stationary phase. Due to the large number of partitioning steps, chromatography has enormous resolving power and can separate mixtures of components with very similar physical properties. In the most common format, called column chromatography, the stationary phase is a highly porous solid material packed inside a cylindrical column (steel or glass), whereas the mobile phase is a liquid, a gas, or a supercritical fluid. If a successful separation has been Analytical Separation Science, First Edition. Edited by Jared L. Anderson, Alain Berthod, Verónica Pino Estévez, and Apryll M. Stalcup.  2015 Wiley-VCH Verlag GmbH & Co. KGaA. Published 2015 by Wiley-VCH Verlag GmbH & Co. KGaA.

2

1 Basic HPLC Theory and Definitions: Retention, Thermodynamics, Selectivity, Zone Spreading, Kinetics

Figure 1.1 Schematic representation of an ideal analytical chromatogram for a binary sample mixture with an unretained component (t0). The retention times of the peaks tR are determined at the peak maxima and the peak widths W at the baseline.

made of a binary sample, this will in the ideal case result in the elution of two Gaussian-shaped concentration peaks (see Figure 1.1). Mathematical models for chromatography were formulated in the 1940s [1] and in 1952 Martin and Synge were awarded the Nobel Prize in chemistry for their work on partition chromatography [2]. The work by Giddings [3] in the 1960s is also considered a milestone in the history of chromatography, for its stringent description of the causes of zone spreading (also called band broadening). The theory describes how the column should be designed and treated to result in efficient separations. The stationary-phase particles should have a small particle diameter, uniform geometry and small-size distribution, and should be homogeneously packed, and any extra column volume should be minimized. Academic scientists, and later manufacturers, followed these directions that resulted in improved liquid chromatography; for example, high-performance liquid chromatography (HPLC) in the 1970s illustrates the dramatic improvements achieved. Chromatography is categorized after the type of mobile phase used, liquid, gas, or supercritical chromatography, which will be described in more detail in later chapters. In this chapter, we will focus on the basic theory necessary for a deeper understanding of the separation process but will also cover new trends in liquid chromatography today.

1.1 Basic Definitions

In analytical chromatography, we want to obtain quantitative and/or qualitative information about one or several components in a sample mixture, whereas in preparative chromatography, the aim is to purify the individual components. The qualitative information is obtained from the retention times in an analytical

1.1 Basic Definitions

Figure 1.2 Schematic representation of a peak with a perfect (ideal) Gaussian normal distribution. Note that the base width Wb is defined as the distance between the points

where the front and rear peak tangents cross the baseline, Wb, which corresponds to four standard deviations (σ).

chromatogram and the quantitative information from the areas, or height, of the peaks (see Figure 1.1). In the ideal case, the analytical peaks are Gaussian (see Figure 1.2); however, in reality, the peaks are often slightly distorted with a small tail. The adsorption isotherms relate the mobile-phase concentration with the stationary-phase concentration (see Figure 1.3a). This relation is linear in analytical chromatography because it is performed at low concentrations corresponding to the initial, practically linear, section of the adsorption isotherm. Because of this, analytical chromatography is sometimes also called linear chromatography. In preparative separations, we instead want to purify as much as possible and the sample concentrations are normally very high, corresponding to regions where the adsorption isotherms exhibit strong curvature. In that region, a further increase of the mobile-phase concentration of the component does not lead to a proportional increase of the stationary-phase concentration. These conditions, which prevail under most preparative separations, are called nonlinear chromatography and we have severe peak deformations. If the adsorption isotherm of the component is convex upward, the resulting elution profile will have a sharp front and a diffused rear (see Figure 1.3b). But if the adsorption isotherm has the opposite shape, that is, concave upward (see Figure 1.4a), the resulting elution profile will have a diffused front and a sharp rear (see Figure 1.4b). These peak shapes are not uncommon in chiral preparative chromatography, especially

3

1 Basic HPLC Theory and Definitions: Retention, Thermodynamics, Selectivity, Zone Spreading, Kinetics

(b)

C

CS

(a)

CM

VR

Figure 1.3 In (a) a schematic representation of a convex upwards (“Langmurian”) adsorption isotherm with the initial linear part indicated by the dotted tangent, CM and CS are the concentration in the mobile and

stationary phase, respectively. In (b) the shape of the resulting overloaded elution profile with sharp front and diffusive rear, VR is the eluted volume, and C is the concentration in the eluted mobile phase at the column outlet.

when there exist an adsorbing additive in the mobile phase [4,5]. Recent research has revealed that adsorption is surprisingly complex and that advanced models often apply [6–9]. Analytical (linear) chromatography can be described by relatively simple models: injection of an n component mixture will give n Gaussianshaped peaks, which are more or less separated in the chromatogram. Here, we focus on analytical (linear) models assuming Gaussian peaks (see Figure 1.2). The most common deviation from this in analytical chromatography is peak tailing. Figure 1.5 shows peak tailing in a schematic way and how it is measured.

(a)

(b)

C

CS

4

CM Figure 1.4 In (a) a schematic representation of a concave upward (“anti-Langmurian”) adsorption isotherm, CM and CS are the concentration in the mobile and stationary phase, respectively. In (b) the shape of the

VR resulting overloaded elution profile with diffused front and sharp rear, VR is the eluted volume, and C is the concentration in the eluted mobile phase at the column outlet.

1.1 Basic Definitions

a+b Tf 5% = 2a C

h

a

b VR

Figure 1.5 Illustration of how the tailing factor at 5%, Tf5%, is calculated from a chromatogram where VR is the eluted volume and C is the concentration in the eluted

mobile phase at the column outlet. A line parallel to the baseline at 5% of the peak height is drawn and the distances a (front part) and b (rear part) are determined.

The pharmaceutical industry prefers to use the term tailing factor (Tfx%), which is defined in Figure 1.5, whereas in the academic community, the asymmetry factor, Asfx% (b/a) is commonly used. In both cases, x stands for at which peak height, relative to the baseline, the asymmetry is calculated. If Asf ∼ 1, we have a close to “Gaussian peak,” whereas Asf > 1 indicates peak tailing and Asf < 1 indicates peak leading, alternatively called “peak fronting.” In this context, it must be mentioned that nonlinear conditions resulting in peak tailing (see Figure 1.3b and Figure 1.5) are very common, not only in preparative chromatography but unfortunately also in analytical chromatography, since the solid phase may contain two or more different adsorption sites. If one has a few numbers of the so called “strong sites,” the curved, nonlinear, section of the adsorption isotherm is reached very early for these sites (see Figures 1.3 and 1.4) and peak tailing will occur. This is especially the case in analytical chiral chromatography [10,11] and when separating basic amines at low-to-moderate pH; here, the amines are charged leading to strong polar interactions besides the hydrophobic interactions [12,13]. Since the traditional reversed phase (= nonpolar stationary phase combined with polar mobile phase, opposite to Tswett’s original straight-phase mode) columns could not stand pH > 7, where the basic amines become more uncharged, material research was during the 1990s focused on eliminating the influence of these sites. Manufacturers investigated different ways to eliminate the strong sites during the production process, whereas academic researchers worked more on operating the existing columns in a different way, for example, to reduce the impact of the polar sample interactions by adding different competitive transparent amines in the mobile phase [14,15]. Today, hybrid materials have been introduced solving the problem in a more consistent way by allowing highly efficient silica-based stationary phases to be combined with strongly basic mobile phases [10,11,16,17]; see more below.

5

6

1 Basic HPLC Theory and Definitions: Retention, Thermodynamics, Selectivity, Zone Spreading, Kinetics

1.1.1 Basic Retention Models and Kinetics

A sample mixture applied on the top of a chromatographic column will be transported through the column by the mobile phase and in a properly designed column, the solutes will be eluted at the column outlet as separate zones. Now, we will look at this process in more detail and develop expressions that describe how quickly the sample zone migrates and how this can be related to the solute distribution between the mobile phase and the stationary phase. The solutes travel can be described by the well-established relation: speed = road/time. Even if the solute zone migrates at a constant rate through the column bed, at the molecular level, the individual molecules do a “random walk.” More specifically, a single molecule transported by the mobile phase is adsorbed for a certain time and in the next moment desorbed, adsorbed again, and so on in thousands of steps during its travel along the column. We can say that the molecules behave like a bunch of rabbits on the way home through a salad field. They stop, on an individual basis, for a moment here and there to eat. On average, the individual molecule will be in the mobile phase during the time tm and in the stationary phase during the time ts. Its total residence time in the column becomes tm + ts. A molecule is adsorbed for an average time of ts, while it is desorbed for an average time of tm, and it then migrates with the velocity of the mobile phase, ux. This means that the molecule will spend a fraction tm/(tm + ts) of its time moving with this speed. The velocity of the molecule, us, is then us ˆ ux


tm : tm ‡ ts

(1.1)

Shortly after the sample starts to travel, the distribution equilibrium of the solute between the stationary phase and the mobile phase is established. If we consider a limited part of the column and assume that it represents the solute molecules’ average behavior, we will have a certain amount of solute in the stationary phase, Qs, and a certain amount in the mobile phase, Qm. The total amount in the zone will be (Qm + Qs) and the percentage of the solute that is in the mobile phase is Qm/(Qm + Qs). There must be a direct correlation between the amount in the zones in the mobile phase and the corresponding time fraction for the solute in the mobile phase. The larger fraction of time the solute spends in the mobile phase, a correspondingly larger quantity is present in the mobile phase. This gives us ˆ ux


Qm 1 1 ˆ ux
ˆ ux
: Qm ‡ Qs 1 ‡ Qs =Qm 1‡k

(1.2)

1.1 Basic Definitions

Here, k is the retention factor. The definition of k, assuming the conditions are the same throughout the chromatographic bed, is given by the following simple relation: kˆ

total amount of solute in the stationary phase Cs
V s : ˆ Cm
V m total amount of solute in the mobile phase

(1.3)

Thus, since the amount (moles) is equal to the concentration (C) multiplied by volume (V), the retention factor is simply defined as a ratio between the solute amounts in the stationary phase and those in the mobile phase. Equation 1.3 is the IUPAC definition of k [18], whereas Equation 1.4 provides an easy way to calculate k from a chromatogram: kˆ

tR

t0 t0

;

(1.4)

where tR is the component solute retention time and t0 is the travel time for the mobile phase or the retention time of an unretained component (Figure 1.1). 1.1.2 Band Broadening and the Plate Height Concept

To perform chromatographic separation of a binary sample mixture, the two substances must have different retention times, which in turn means that they must have different distribution ratios. However, this is not the only necessary condition. Looking at a chromatogram, it is obvious that the peak widths will also affect the ability to separate one component from the other. The initially introduced solute zone must have a very narrow width, but the spreading phenomenon in the column will increase the final zone width significantly. This means that the sample molecules come out in a zone that contains a larger volume than the sample injection and this dilution increases with increasing retention factor. There are some models available that describe the effects that lead to band broadening. The easiest to understand is probably the so-called random walk model [3], which looks at band broadening as a result of a random process that exposes every molecule of movements forward or backward relative to the zone center. With a sufficiently large number of steps and molecules, the result is a normal distribution of the molecules. The variance can be estimated from the chromatogram according to the rules for a Gaussian distribution, which states that the standard deviation in time units is σ t = Wt/4, where Wt is the base width of the peak defined by the tangents to the inflection points (see Figure 1.2). The relationship (expressed in time units) between the width of the peak, Wt, and the variance σ 2t is then
σ 2t ˆ

Wt 4

2 :

(1.5)

7

8

1 Basic HPLC Theory and Definitions: Retention, Thermodynamics, Selectivity, Zone Spreading, Kinetics

Since the variance of the peak increases with the number of steps, which is proportional to the distance migrated, L, a quantity σ 2/L (σ expressed in length units) is used as a general measure of band broadening (i.e., the efficiency of the column). It is usually called plate height (or “height equivalent to a theoretical plate”), H: Hˆ

σ2 : L

(1.6)

Do not get confused by the expression “plate height” that is used for historical reasons, consider it simply as a quantity that is proportional to the variance. If we have several different types of random processes that occur independently and with different step lengths and number of steps, the end result is a composite profile that also is Gaussian, with a total variance that is the sum of the variances of each individual process (random walk). So, in accordance with the law on summation of variances σ 2 ˆ σ 21 ‡ σ 22 ‡ σ 23 ‡ σ 24 ‡ . . .

(1.7)

The corresponding expression for the summation of the individual contributions to the plate height will then be Hˆ

σ 2 σ 21 σ 22 σ 23 ˆ ‡ ‡ ‡ . . . ˆ H1 ‡ H2 ‡ H3 ‡ . . . L L L L

(1.8)

The variance also increases with the travel time of the zones, so normalization has to be done if you measure variance in time units based on the simple expression time = path/speed. The relationship between the base width of the peak in time units, Wt, and H is then obtained from σ 2 ˆ …σ t
us †2 :

(1.9)

So, how do we go from the definition of the plate height to an expression that can be used to calculate the plate height directly from a chromatogram? Let us first take a look again at the schematic representation of a Gaussian peak in Figure 1.2. This schematic figure shows that if you calculate the width of the peak at the base, as defined by the tangents of the slope of the rear and the front of the peak crossing the baseline, the peak width contains four standard deviations, that is, 4σ. By combining the definition of H in Equation 1.6 with Equation 1.9, with the assumption of ideal Gaussian shape, where Wt = 4σ, we obtain an equation to calculate H directly from the chromatogram: Hˆ

W t2
L : 16
t 2R

(1.10)

1.1 Basic Definitions

H can thus be calculated from the base width, retention time, and column length. H will be in length unit and can apparently be attributed to the width of the peak relative to its retention time. If the ratio is constant, then H is constant, assuming the same column length, L. A measure of the efficiency of the column as a whole is given by the number of plates, N: Nˆ

L 16
t 2R : ˆ H W 2t

(1.11)

N gives the total number of theoretical plates for a certain column length. Since N = L/H, the number N says something about the potential performance of a certain column where H is a more general property. For columns with the same length, the higher the number of plates, the lower the peak width expressed as a fraction of the retention time. For a column with 1 600 plates, the peak width is 10% of the retention time, while 6 400 plates are obtained for a peak that has a width that is 5% of the retention time. 1.1.3 Sources of Zone Broadening

There are mainly three different physicochemical phenomena that cause zone broadening, graphically illustrated in Figure 1.6. The three different physicochemical phenomena are as follows: 1) Eddy diffusion (multiple paths) 2) General molecular diffusion 3) Slow equilibration.

Figure 1.6 Illustration of the three major contributions to band broadening in chromatography: (a) eddy diffusion, (b) molecular diffusion, and (c) slow equilibration.

9

10

1 Basic HPLC Theory and Definitions: Retention, Thermodynamics, Selectivity, Zone Spreading, Kinetics

1.1.3.1

Eddy Diffusion

This type of zone broadening is due to the uneven path lengths and velocities in the packed column (see Figure 1.6). It is due to the fact that as the mobile phase is pumped through the column, the flow will pass between the particles using different paths with different local flow rates and molecules trapped in these different paths will travel at different speeds. In the random walk of a solute molecule, the length of the step, l, is assumed to be equivalent to the diameter of a stationary phase (support) particle, dp. The number of steps, n, is equal to L/dp, that is, the length of the migration zone divided by the length of each step. This gives σ 2E ˆ l2
n ˆ 2
λ
dp
L:

(1.12)

The factor, λ, depends on the structure of the packing; a more homogeneous packing has a lower number. The derivations applied for the other mechanisms of band broadening by the random walk model is not given here. The readers can consult [3] for details. 1.1.3.2

Molecular Diffusion

This type of diffusion occurs in all directions, but since the separation occurs in the flow (axial) direction, the longitudinal diffusion has generally the largest impact on band broadening (see Figure 1.6). Diffusion also happens in the pores of the stationary phase, but the diffusion in the bulk of the mobile phase dominates. The variance, σ 2D , is σ 2D ˆ

2
Dm
L ; ux

(1.13)

where Dm is the diffusion coefficient of the solute in the mobile phase. We see that the variance for molecular diffusion contributes mainly to the band broadening at low flow rates, and increases with increasing diffusion coefficient (larger for smaller molecules) in the mobile phase. 1.1.3.3

Slow Equilibration

During the sample transport through the chromatographic bed, local distribution equilibrium is established in which molecules are constantly adsorbed and desorbed by the stationary phase. This process is almost always fast enough to establish equilibrium, except in those areas of the sample zone where the sample concentration changes more drastically, that is, at the front and at the back of the sample zone. At the front of the zone, higher concentration of sample molecules is pushed in all the time and the local equilibrium at the front does not have time to be established (Figure 1.6). Therefore, Cm becomes high and Cs low, and thus the ratio Cs/Cm (in Equation 1.3) is lower compared to an established equilibrium. This means that the retention factor k locally becomes small (see Equation 1.3) and thus the speed of the front zone becomes high. In the back of the zone, the opposite happens: Cs becomes high and thus k in this zone is high. In this way, the zone will broaden in both directions. Equation 1.14 gives

1.1 Basic Definitions

an expression for the variance due to this source of band broadening assuming a partitioning of solutes into the pores with an average depth of df and a packing factor, qs, which is smaller with increasing homogeneity of the packing, and k is the retention factor: σ 2c ˆ

k
qs
d 2f
ux
L : Ds … 1 ‡ k †

(1.14)

An important conclusion from the variance due to slow equilibration is that it increases with the mobile-phase velocity (ux) and decreases with the higher diffusion/adsorption constant in the stationary phase (Ds). 1.1.4 Dependence of Zone Broadening on Flow Rate

In the total zone broadening, the variances of the individual sources for band broadening, as described above, are added. The relation between the total zone broadening and the linear flow rate is important and has been described with different degrees of complexity in the literature. The simple expression for packed gas chromatography (GC) by van Deemter [19] is still used as a didactic introduction to the area and also for liquid chromatography (LC). The relationship is illustrated in Figure 1.7 (solid line); it shows that there exists an optimal

B A+ u + Cux

Plate height

x

B Longitudinal ux diffusion

Cux Equilibration time Multiple A paths

Flow rate Figure 1.7 Illustration of the three different sources for band broadening with increasing flow rate (dotted lines). The solid line shows the sum of the three individual contributions in accordance with Equation 1.15.

11

12

1 Basic HPLC Theory and Definitions: Retention, Thermodynamics, Selectivity, Zone Spreading, Kinetics

flow rate which gives a minimum plate height, H. This flow rate should be used to obtain the maximal number of theoretical plates (N) for the column. Let us take a look at Equation 1.15 and the expression derived by van Deemter for the relationship between H and the linear flow rate ux: H ˆA‡

B ‡ C
ux : ux

(1.15)

Here, A represents eddy diffusion, B/ux molecular diffusion, and Cux slow equilibria. This relationship is illustrated in Figure 1.7: the solid line at the top shows the total band broadening that is the sum of the three dashed lines below that shows the three different sources of band broadening. From Figure 1.7 and Equation 1.15, we can see that at low flow rates, the molecular diffusion is the dominating source for the band broadening, whereas at high flow rates it is the slow equilibrium that is the main reason. At the optimum flow rate, the band broadening is mainly caused by the eddy diffusion that can be decreased only by decreasing the particle sizes. Figure 1.7 is a didactic example, based on the situation in packed GC columns. In LC, the minimum is often at very low flow rates since the influence of diffusion in the mobile phase (B/ux) is very small as the diffusion coefficients in a liquid phase are more than 100 times lower than in a gas phase.

1.2 Resolution

The selectivity α between two components is the ratio of the retention factors between the more retained neighboring peak (k2) and the less retained one (k1): αˆ

k2 : k1

(1.16)

The selectivity is a good measure of a stationary phase ability to discriminate between two components in GC and is, for example, used by organic chemists to find the best stationary phase for separating a certain component or a class of target components. However, in LC the composition of the mobile phase also influences the selectivity, and furthermore, the selectivity does not say all about how good the separation is between two components. Here, we must find an expression that also accounts for the width of the peaks, that is, the column efficiency expressed by the number of theoretical plates. This is illustrated in Figure 1.8a–c. Figure 1.8a and c shows a chromatogram for two components with the same selectivity but different column efficiency, whereas we only have complete resolution in Figure 1.8c. Figure 1.8b has the same low column efficiency as in Figure 1.8a but much higher selectivity, α, and this is another way to achieve resolution starting out from the nonresolved situation in Figure 1.8a.

1.2 Resolution

Figure 1.8 Illustration of different ways to improve the separation of two components where we have both low efficiency and selectivity (a), in (b) the selectivity is increased, and in (c) the efficiency is increased.

Thus, the ultimate measure of a successful separation between the closely eluted peaks is the resolution (Rs) accounting both for the selectivity and for the column efficiency. Here, the degree of chromatographic separation, or resolution, of two adjacent solute zones is defined as the distance between the zone centers divided by the average zone width (Wt): Rs ˆ

  2 t R;2 t R;1 : W t;1 ‡ W t;2

(1.17)

For closely adjacent bands, the average zone width is approximately equal. Since Wt = 4σ t, Equation 1.17 can be written: Rs ˆ

t R;2 t R;1 : 4σ t

(1.18)

When Rs = 1, there are four σ units between the zone centers. This corresponds to a 2% contamination of each band by the other. Rs = 1.5 corresponds to a distance of six σ units between the zone centers. The cross-contamination of the zones is then not more than 0.2% and the separation is almost complete, provided that the amount of solute in the two zones is approximately equal. Now, we have a quantitative measure of the degree of resolution between the two substances and by the definition of the resolution, we understand that it depends on both the sample retention and the band broadening. Through theoretical analysis, one can show the following: Rs ˆ

pffiffiffiffi

 N k2 α 1

; 4 1 ‡ k2 α

(1.19)

13

1 Basic HPLC Theory and Definitions: Retention, Thermodynamics, Selectivity, Zone Spreading, Kinetics

(a)

(b)

5000 N

10000

1 (α−1)/α

50

0 0

(c)

1 k2/(1+k2)

100 √N

14

0.5 0 0

10

20 k2

0.5 0

5

10

15

20

α

Figure 1.9 The influence on resolution when increasing (a) the efficiency, (b) the retention factor for second component, and (c) the separation factor (see Equation 1.19).

where we know that the resolution depends on the column efficiency N, the retention factor for the more retained neighboring peak, k2, and the selectivity, α. A closer study on the effects of these different parameters on resolution will help us understand, control, and improve the resolution. As we can see, Equation 1.19 consists of three multiplied terms. Figure 1.9a–c illustrates, for each of these three terms, how a variation of the corresponding parameter affects each of these terms individually. From Figure 1.9a, we see that increasing the efficiency will continuously increase the resolution since it is proportional to the square root of N. Increasing the retention factor has a large effect when going from low values but reaches a limiting value of 1 when k2 is high; the effect on resolution is very small when k2 > 10 (see Figure 1.9b). When the separation factor is increased from 1, a similar dramatic improvement of the resolution is obtained, although this effect is greatly reduced at higher α-values and reaches the limit of 1 (see Figure 1.9c). However, Equation 1.19 is valid only for low values of α, since when deriving it was assumed that the peak widths of the two separated peaks were approximately equal, and this is true only when the separation factor is small (< 2). The simplest way to improve the resolution is to increase the retention factor, since this often can be done by a simple modification of the mobile phase. This is, however, effective only when the initial retention factor is smaller than 5. An increase of N should be the next choice; the simplest way to do this is to use a longer column. But a drawback with longer columns is decreasing sensitivity, since the dilution of the peaks will increase. The best way to increase N is, therefore, to decrease the particle diameter. The most difficult way to increase resolution is to try to increase the selectivity; this generally requires a radical change of the mobile or stationary phase that in many cases gives unpredictable results.

1.3 Modern Trends in Liquid Chromatography

Separation scientists in both academy and industry always try to continuously improve their separations, starting out from the actual knowledge, and there are

1.3 Modern Trends in Liquid Chromatography

several simultaneous trends in liquid chromatography today. Many of the trends are logical consequences of the basic theory we have just covered. For example, we learned from the basic theory that the resolution (Rs) must be  1.5 for a complete separation between two neighboring peaks. If the resolution is 1.5, we have 6 σ units between the peaks’ zone centers corresponding to a 0.2% coelution of the two components. From Equation 1.18, we could see that the resolution depends on the efficiency (N), the retention factor of the second eluting component, k2, and the selectivity, α. Thus, Equation 1.18 tells us that for a pair of solutes with reasonable retention factors, we could increase the resolution by improving either the selectivity or the efficiency. One important trend is faster analysis, with higher sample throughput, as this saves time. One way to achieve this is to develop a system with higher resolution than you need and thereafter increase the flow rate. For example, by using smaller particles (sub-2 μm particles), the resolution can be much larger (Rs >> 1.5) than for larger particles and one can then increase the flow rate until Rs is the same as for the larger particles. This increased flow rate will result in a significantly higher pressure; therefore, this technique was named ultrahighpressure liquid chromatography (UHPLC). Another way is to use a more permeable matrix material in the column packing giving a conventional pressure, which is one of the underlying reasons for the trend in monolithic materials. Yet, another way is to use core–shell particles. Next follow further details on these materials.

1.3.1 Efficiency Trend

Already in the early days of LC [2], it was realized that in order to achieve high efficiency, the use of small particles was necessary. However, it took several decades (until the 1970s) before scientists and manufacturers learned how to make small particles  10 μm. The new technique using small particles was given the name high-performance liquid chromatography (HPLC). A step further was to decrease the number of irregular particles in favor of spherical homogeneous particles. A challenge with the smaller particles is the need to develop instruments that can be operated under higher pressures. If we look at the parameters that determine the pressure drop over the column in Equation 1.20, we can see that the pressure is inversely related to the squared particle diameter: ΔP ˆ ϕ

ux ηL : d2p

(1.20)

Here, ΔP is the pressure drop, dp is the mean particle diameter, η is the viscosity and ϕ the column resistance factor. ϕ depends on the method for packing the column and on the porosity of the packing material. This means that going from 100 μm particles in the 1950s to 10 μm particles in the 1970s (a 10-fold

15

1 Basic HPLC Theory and Definitions: Retention, Thermodynamics, Selectivity, Zone Spreading, Kinetics

60 Plate height [μm]

16

10 μm

40 20

5 μm 3 μm

0

2 4 6 Flow rate [mL/min]

8

Figure 1.10 The plate height versus the flow rate for different particle sizes: 10, 5, and 3 μm. (Adapted from Ref. [38].)

decrease in particle size) increased the pressure drop over the column 100 times, given that all other conditions are identical. The trend has continued with the modern sub-2 μm porous spherical particles of today (UHPLC). The optimum H is around two-particle diameters; but as discussed above, the pressure increases inversely proportional to the square of the particle diameter (Equation 1.20). Note that increasing the mobile-phase linear velocity, ux, also increases the pressure and the column efficiency; see Equations 1.15 and 1.20. Figure 1.10 shows the van Deemter curves (see Equation 1.15) for the three different particle sizes: 10, 5, and 3 μm. More specifically, Figure 1.10 illustrate that with decreasing particle sizes, the A term and C term in Equation 1.15 will decrease due to shorter irregular longitudinal path length and reduced diffusion distance. This leads to higher optimal mobile-phase linear velocity, that is, the linear velocity where H has a minimum. For example, if the particle size is decreased in an already established method, the method can be operated under larger linear velocity achieving the same efficiency (and thus maintained RS) using shorter columns. Reducing the particle size will not only increase the separation speed but also increase the sensitivity as the sample zone(s) get less diluted in the shorter column, provided the same injected volume is used. Because of the limited operational pressure for HPLC (400 bar), the full potential of reduced particle size cannot be utilized. As a consequence, the column length has also been decreased with decreasing particle size to operate the separation system within the pressure limit, leading to nonsubstantial increases in efficiency. Today, up to 1000 bars can be delivered by new commercial HPLC systems, called UHPLC. They first became commercially available in 2004 (Waters Acquity UPLC) [20] and shortly afterward most manufacturers offered similar equipment. UHPLC provides faster separations with lower solvent consumptions compared to HPLC with preserved column efficiency [21,22]. Its popularity has grown steadily and UHPLC is today well-established in the pharmaceutical

1.3 Modern Trends in Liquid Chromatography

industry. From an instrumental perspective, the main difference between HPLC and UHPLC is that smaller particles are used in UHPLC (< 2 μm particles). As a consequence, pumps able to manage pressures up to 1000 bar are required, compared to around 100–250 bar for conventional HPLC ( 3 μm particles). UHPLC is nevertheless a success story because the manufacturing companies have succeeded in producing small-particle stationary phases with very similar properties as the ones with HPLC size particles. However, at high flow rates, there is a serious risk that the use of the reduced particle size in UHPLC leads to temperature gradients in the column that can seriously affect the chromatographic performance, that is, retention, efficiency, and resolution. More specifically, when these columns are operated at high flow rates, and thus with high inlet pressures, heat is generated in the column due to viscous friction that causes both axial and radial temperature gradients. Consequently, these columns become heterogeneous and several physicochemical parameters, including the retention factors and the mass transfer kinetics parameters of the analytes, are no longer constant along and across the column. This has been demonstrated theoretically with advanced modeling, which combines the heat and mass balance of the column, and verified experimentally by Kaczmarski et al. [23,24]. In this context, it should be mentioned that a good way to decrease the pressure drop over the columns is to increase the column temperature. In accordance with the classic Equation 1.21 that follows, we can easily see that the viscosity decreases significantly with increased temperature, T: η ˆ A…Ev=kT † ;

(1.21)

where A, Ev, and k (Boltzmann’s constant) are constants. From Equation 1.20, we see that the pressure drop over the column is proportional to the viscosity. To summarize, higher temperature leads to a smaller viscosity (Equation 1.21) that in turn leads to lower pressure drop over the column (Equation 1.20). Therefore, in practice UHPLC systems are operated at somewhat elevated temperatures; often 40 °C instead of 25 °C used in HPLC. 1.3.2 Permeability Trend

The use of the so-called semiporous material (also called fused core particle technology, superficially porous material, or core–shells) is a strong trend today. This is another way to achieve high-efficiency separations, with higher resolution than necessary, which means one can increase the flow rate to obtain speedier analysis. From a technical perspective, the “pellicular material” introduced in 1967 [25] consisted of a solid core covered with a porous shell. The “pellicular” material did not become popular immediately, because in the 1970s the more competitive 10 μm porous silica particles were introduced. Figure 1.11a–c illustrates the principle for semiporous packing materials. Figure 1.11a shows a modern standard HPLC particle, that is, a fully porous 3 μm diameter particle with a

17

1 Basic HPLC Theory and Definitions: Retention, Thermodynamics, Selectivity, Zone Spreading, Kinetics

Figure 1.11 An illustration of how fused core paths (0.5 μm) compared to HPLC (1.5 μm) and particles (c) as related to modern standard UHPLC (0.9 μm). (Adapted from an illustration HPLC (a) and UHPLC (b) particles. Note that provided by HALO Ltd.) the fused core particles have shorter diffusion

radius of 1.5 μm. Figure 1.11b shows a modern UHPLC particle, that is, a totally porous 1.8 μm diameter particle with a radius of 0.9 μm. The radius of these particles gives the maximum diffusion length that a molecule has to travel to interact fully with the particle surface. Finally, Figure 1.11c shows the so-called semiporous (or superficially porous) packing material. This particle has a porous 0.5 μm layer on top of a 1.7 μm fused core that gives a 2.7 μm particle size. This means that the use of semiporous particles results in relatively moderate pressure drops, almost comparable to standard HPLC particles since the total particle size is almost the same (2.7 μm compared to 3 μm). But at the same time, the efficiency is very high because of the short diffusion length for semiporous particles compared to UHPLC particles (0.5 μm versus 0.9 μm). The shorter diffusion length increases the mass transfer kinetics (smaller C term in Equation 1.15) and thus decreases the band broadening. Figure 1.12 shows the relation between the pressure drop over the column (in bar) and the mobile-phase velocity (in millimeter per second) for three different particle sizes 1.7 μm UHPLC packing material, 2.7 μm fused core material, and 3.6 μm standard HPLC packing material particles. 600 Pressure [bar]

18

1.7 µm porous

400 200 0

2.7 µm fused core 3.6 µm porous 2 4 6 Mobile Phase Velocity [mm/s]

Figure 1.12 The pressure drop across a column when different particle sizes and particle types. (Adapted from an illustration provided by HALO Ltd.)

1.3 Modern Trends in Liquid Chromatography

To summarize, semiporous materials can be used with standard HPLC instruments since they have almost the same size as standard particles; this is in contrast to sub-2 μm porous particles where the low-permeability forces the user to invest in UPLC systems. At the same time, the increased mass transfer for semiporous materials (due to shorter diffusion distances for semiporous particles, see Figure 1.11) will give very high efficiency. Monoliths provide another solution to increase the permeability by increasing the external porosity [26,27]. Monoliths consist of single continuous porous materials, polymerized in situ and covalently bound to the column wall, with large through-pores that allow the flow to percolate with a convective flow through the pores and with diffusive smaller pores to increase the surface area. One disadvantage is that it is difficult to polymerize very wide and very narrow columns in a reproducible way. However, the increased permeability allows the user to employ longer columns to obtain higher column effectiveness or increased flow rates for shorter analysis times. For monolithic columns, the C term in Equation 1.15 is much smaller than for standard HPLC packing due to shorter diffusion distances. However, the A term is larger probably because of the large-size distribution of the through-pores. Chromolith (Merk) was the first commercially available silica rod monolith column with efficiency comparable to that provided by 5 μm particles and with a high permeability similar to a column packed with 10–15 μm particles. Monolithic column materials are especially well suited for peptide separations. 1.3.3 Selectivity and New Material Trend

One big problem since the birth of HPLC has been the peak tailing of basic amines that adsorbs on low-capacity, strongly polar (charged interactions) adsorption sites on otherwise reversed phase material. The pKa of an aliphatic amine is around 9 and thus the sample components are charged at the mobilephase pH usually used in reversed phase LC, that is, between 2 and 7. If the pH is increased above 7, the silica matrices will be destroyed. However, from a theoretical standpoint, it makes sense to separate amines at pH around 11 since the amines will be uncharged and no charged interaction with the stationary phase will occur; therefore, many attempts have been made to strengthen the silica matrix in different ways. Today, newly developed silica stationary phases, the so-called hybrid phases, are available that are stable at a pH between 1 and 12, under reversed phase LC (RPLC) conditions. The pHstable silica was mentioned already in the 1970s and is actually a hybrid material that consists of a combination of silica and organic polymers [28]. The hybrid materials available on the market today use methyl or ethyl groups distributed throughout the particle (Xterra and XBridge from Waters) [29] or surface-grafted ethyl bridged (Gemini from Phenomenex). The stability is increased because the Si-C bond can withstand hydrolysis much better than the Si-O bond. The Xbridge particles are prepared from tetraethoxysilane and

19

1 Basic HPLC Theory and Definitions: Retention, Thermodynamics, Selectivity, Zone Spreading, Kinetics

(a) (1.65)

150

(1.52) (1.66)

(2.10)

(1.59)

(2.30)

50 pH 3

pH 8

pH 7

pH 6

pH 5

pH 4

(1.27)

(3.85)

100

(1.24) pH 10

pH 9

pH 11

Metoprolol

0 150 response [mAU]

20

(b)

100 (1.14)

50

(1.13)

pH 3

(1.08)

pH 4

(1.12)

(1.13)

pH 5

(1.11)

pH 7

pH 6

(1.14)

pH 8

(1.16) (1.10) pH 10

pH 9

pH 11

3−Phenyl−1−Propanol

0 (c) 200

(1.32)

(1.36)

(1.34)

(1.35)

(1.30)

(1.29)

150 (1.10)

100 50 0 0

(1.20)

(1.78) pH 3

10

20

30

pH 6 pH 4

pH 7

pH 8

pH 9

pH 10

2−Phenylbutyric acid

pH 5

40 50 time [min]

Figure 1.13 Analytical chromatograms for (a) metoprolol, (b) 3-phenyl-1-propanol, and (c) 2-phenylbyturic acid for pH 3–11 using an XBridge column. The asymmetry values at

pH 11

60

70

80

90

10% of the peak height are shown in parenthesis. (Reprinted from Ref. [10], with permission from Elsevier.)

bis(triethoxysilyl)ethane. The column lifetime is reduced at low pH due to acid hydrolysis of the bound alkylsilane. Bulky side groups on the alkylsilane have shown to sterically shield the bound siloxane from acid hydrolysis [30,31]. Hybrid materials and bidentate-bound stationary phases also seem to increase the stability [32]. Figure 1.13a–c shows the resulting eluted analytical peaks at a wide range of pH used in the mobile phase all the way from pH 3 to pH 11 using an early generation of pH-stable hybrid column (XBridge). Figure 1.13a shows the resulting peaks for a basic amine (metoprolol), Figure 1.13b for the neutral component 3-phenyl-1-propranolol (PP) and Figure 1.13c for the acidic component 2-phenylbutyric (PB) acid. The asymmetry factors (Asf10 values) are shown in parenthesis. In the case of the neutral PP, it can be noted that the asymmetry is constant over the whole studied pH range (3–11). In the case of the acid PB, the asymmetry factor is approximately unity for all pH values, because at higher pH values, the retention time is very short. For the base ME, the asymmetry factor is maximal at pH 8; at lower pH, the

1.4 Conclusions

asymmetry is near unity and at pH 9–11, it decreases to below 2. At pH > 9, it is remarkable that we have a low degree of tailing of ME, in spite of the adsorption, and thus the retention of the uncharged component, increases tremendously (see Figure 1.13c). The silica hybrid material has improved the chemical stability, but only for the short-term use. For the long-term use under chemically aggressive conditions and elevated temperature, only polymeric and metal oxide materials work. Hightemperature separations are sometimes called green chromatography because less organic modifier or even pure water can be used as eluent. The reason for this is that the solvent strength in the eluent increases with increasing temperature, for example, a 4–5 °C increase of temperature approximately equals a reduction of 1% methanol or acetonitrile [33,34]. Also, the mass transfer and molecular diffusion increase with temperature, resulting in a lower C term but a higher B term in Equation 1.15 [35]. Another trend is the use of polar stationary phases combined with aqueous mobile phases, similar to those used in reversed phase mode, called hydrophilic interaction liquid chromatography (HILIC) [36]. The technique is especially useful for polar – neutral and ionic – components that are difficult to retain by reversed phase system. In HILIC, the retention increases with increasing content of organic solvent in the mobile phase. HILIC, therefore, has a special interest [37] due to its nearly orthogonal selectivity compared to RPLC; for example, at 30/70 water/acetonitrile using a ZIC -HILIC column, uracil is retained more than naphthalene [36]. HILIC is also an alternative for separating organic bases, sugar, and other polar components that have low solubility in the nonpolar eluents used in normal-phase liquid chromatography (NPLC). As eluent, 5–40% aqueous buffer, mixed with polar organic solvents, primarily acetonitrile, are used. Examples of polar stationary phases are pure silica, diols, and amides, mixed modes with ion-exchangers, zwitterions, or sugars [36]. The retention mechanism is not completely understood but is believed to be a complex blend of liquid–liquid partitioning between the bulk eluent and an adsorbed aqueous layer on the stationary phase involving hydrogen bonding, electrostatic interactions, and weaker interactions (see Figure 1.14). The relative importance of the different interactions certainly depends on the chemical characters of both the solutes and the stationary phase.

1.4 Conclusions

The main objective of this chapter is to provide a comprehensive description of the basic theory of liquid chromatography. This is necessary in order to understand the following chapters in the volume, where different chromatographic techniques, modes, and methods will be presented and discussed. An adequate theoretical knowledge of separation science is further important in order to understand the background to many of today’s trends in chromatography that

21

22

1 Basic HPLC Theory and Definitions: Retention, Thermodynamics, Selectivity, Zone Spreading, Kinetics

Figure 1.14 Schematic illustration of the tentative retention mechanism in HILIC. (Adapted from an illustration provided by Phenomenex.)

are described briefly here, such as UHPLC (using < 2 μm porous particles) and utilization of core–shell particles or monolithic materials in the applications of separation science. Another trend that is also presented briefly here is the development of modern pH-stable silica-based C18 hybrid packing materials and hydrophilic interaction chromatography.

References 1 Gluckauf, E. (1946) Contributions to the

2

3

4

5

theory of chromatography. Proc. Roy. Soc. London A., 186 (1004), 35–57. Martin, A.J. and Synge, R.L. (1941) A new form of chromatogram employing two liquid phases: a theory of chromatography. 2. Application to the micro-determination of the higher monoamino-acids in proteins. Biochem. J., 35 (12), 1358–1368. Giddings, J.C. (1965) Dynamics of Chromatography: Principles and Theory, Marcel Dekker, New York. Arnell, R., Forssén, P., and Fornstedt, T. (2007) Tuneable peak deformations in chiral liquid chromatography. Anal. Chem., 79 (15), 5838–5847. Forssén, P., Arnell, R., and Fornstedt, T. (2009) A quest for the optimal additive in chiral preparative chromatography. J. Chromatogr. A., 1216 (23), 4719–4727.

6 Guiochon, G., Shirazi, D.G., Felinger, A.,

and Katti, A.M. (2006) Fundamentals of Preparative and Nonlinear Chromatography, Academic Press, Boston, MA. 7 Felinger, A., Cavazzini, A., and Dondi, F. (2004) Equivalence of the microscopic and macroscopic models of chromatography: stochastic–dispersive versus lumped kinetic model. J. Chromatogr. A., 1043 (2), 149–157. 8 Fornstedt, T. (2010) Characterization of adsorption processes in analytical liquid– solid chromatography. J. Chromatogr. A., 1217 (6), 792–812. 9 Samuelsson, J., Arnell, R., and Fornstedt, T. (2009) Potential of adsorption isotherm measurements for closer elucidating of binding in chiral liquid chromatographic phase systems. J. Separ. Sci., 32 (10), 1491–1506.

References 10 Samuelsson, J., Franz, A., Stanley, B.J., and

11

12

13

14

15

16

17

18

Fornstedt, T. (2007) Thermodynamic characterization of separations on alkaline-stable silica-based C18 columns: why basic solutes may have better capacity and peak performance at higher pH. J. Chromatogr. A., 1163 (1–2), 177–189. Undin, T., Samuelsson, J., Törncrona, A., and Fornstedt, T. (2013) Evaluation of a combined linear–nonlinear approach for column characterization using modern alkaline-stable columns as model. J. Separ. Sci., 36 (11), 1753–1761. McCalley, D.V. (2000) Effect of temperature and flow-rate on analysis of basic compounds in high-performance liquid chromatography using a reversedphase column. J. Chromatogr. A., 902 (2), 311–321. Ståhlberg, J. (1999) Retention models for ions in chromatography. J. Chromatogr. A., 855 (1), 3–55. Sokolowski, A. and Wahlund, K.-G. (1980) Peak tailing and retention behaviour of tricyclic antidepressant amines and related hydrophobic ammonium compounds in reversed-phase ion-pair liquid chromatography on alkyl-bonded phases. J. Chromatogr. A., 189 (3), 299–316. Tilly Melin, A., Liungcrantz, M., and Schill, G. (1979) Reversed-phase ion-pair chromatography with an adsorbing stationary phase and a hydrophobic quaternary ammonium ion in the mobile phase. J. Chromatogr. A., 185, 225–239. Neue, U.D., Wheat, T.E., Mazzeo, J.R., Mazza, C.B., Cavanaugh, J.Y., Xia, F., and Diehl, D.M. (2004) Differences in preparative loadability between the charged and uncharged forms of ionizable compounds. J. Chromatogr., 1030 (1–2), 123–134. Davies, N.H., Euerby, M.R., and McCalley, D.V. (2006) Study of overload for basic compounds in reversed-phase high performance liquid chromatography as a function of mobile phase pH. J. Chromatogr. A., 1119 (1–2), 11–19. Nič, M., Jirát, J., Košata, B., Jenkins, A., and McNaught, A. (eds) (2009) IUPAC Compendium of Chemical Terminology:

19

20

21

22

23

24

25

26

27

28

Gold Book, IUPAC, Research Triagle Park, NC. Van Deemter, J.J., Zuiderweg, F.J., and Klinkenberg, A. (1956) Longitudinal diffusion and resistance to mass transfer as causes of nonideality in chromatography. Chem. Eng. Sci., 5 (6), 271–289. Swartz, M.E. (2005) UPLCTM: an introduction and review. J. Liq. Chromatogr. R. T., 28 (7–8), 1253–1263. Petersson, P. and Euerby, M.R. (2007) Characterisation of RPLC columns packed with porous sub-2μm particles. J. Separ. Sci., 30 (13), 2012–2024. Fanali, S., Haddad, P.R., Poole, C., Schoenmakers, P., and Lloyd, D.K. (eds) (2013) Liquid Chromatography: Fundamentals and Instrumentation, Elsevier, Waltham. Kaczmarski, K., Kostka, J., Zapała, W., and Guiochon, G. (2009) Modeling of thermal processes in high pressure liquid chromatography. I. Low pressure onset of thermal heterogeneity. J. Chromatogr. A., 1216 (38), 6560–6574. Kaczmarski, K., Gritti, F., Kostka, J., and Guiochon, G. (2009) Modeling of thermal processes in high pressure liquid chromatography. II. Thermal heterogeneity at very high pressures. J. Chromatogr. A., 1216 (38), 6575–6586. Horvath, C.G., Preiss, B.A., and Lipsky, S.R. (1967) Fast liquid chromatography: investigation of operating parameters and the separation of nucleotides on pellicular ion exchangers. Anal. Chem., 39 (12), 1422–1428. Guiochon, G. (2007) Monolithic columns in high-performance liquid chromatography. J. Chromatogr. A., 1168 (1–2), 101–168. Vervoort, N., Gzil, P., Baron, G.V., and Desmet, G. (2004) Model column structure for the analysis of the flow and band-broadening characteristics of silica monoliths. J. Chromatogr. A., 1030 (1–2), 177–186. Unger, K.K., Becker, N., and Roumeliotis, P. (1976) Recent developments in the evaluation of chemically bonded silica packings for liquid chromatography. J. Chromatogr. A., 125 (1), 115–127.

23

24

1 Basic HPLC Theory and Definitions: Retention, Thermodynamics, Selectivity, Zone Spreading, Kinetics 29 Wyndham, K.D., O’Gara, J.E., Walter,

T.H., Glose, K.H., Lawrence, N.L., Alden, B.A., Izzo, G.S., Hudalla, C.J., and Iraneta, P.C. (2003) Characterization and evaluation of C18 HPLC stationary phases based on ethyl-bridged hybrid organic/ inorganic particles. Anal. Chem., 75 (24), 6781–6788. 30 Scholten, A.B., de Haan, J.W., Claessens, H.A., van de Ven, L.J.M., and Cramers, C.A. (1994) 29-Silicon NMR evidence for the improved chromatographic siloxane bond stability of bulky alkylsilane ligands on a silica surface. J. Chromatogr. A., 688 (1–2), 25–29. 31 Kirkland, J.J., Adams, J.B., van Straten, M.A., and Claessens, H.A. (1998) Bidentate silane stationary phases for reversed-phase high-performance liquid chromatography. Anal. Chem., 70 (20), 4344–4352. 32 Kirkland, J.J. (2004) Development of some stationary phases for reversed-phase HPLC. J. Chromatogr. A., 1060 (1–2), 9–21.

33 Bowermaster, J. and McNair, H.M. (1984)

34

35

36

37

38

Temperature programmed microbore HPLC. Part I. J. Chromatogr. Sci., 22 (4), 165–170. Chen, M.H. and Horváth, C. (1997) Temperature programming and gradient elution in reversed-phase chromatography with packed capillary columns. J. Chromatogr. A., 788 (1–2), 51–61. Yang, Y. (2006) A model for temperature effect on column efficiency in hightemperature liquid chromatography. Anal. Chim. Acta, 558 (1–2), 7–10. Hemström, P. and Irgum, K. (2006) Hydrophilic interaction chromatography. J. Separ. Sci., 29 (12), 1784–1821. Alpert, A.J. (1990) Hydrophilic-interaction chromatography for the separation of peptides, nucleic acids and other polar compounds. J. Chromatogr. A., 499, 177–196. Harris, D.C. (2007) Quantitative Chemical Analysis, W.H. Freeman & Co., New York, NY.

25

2 Basic LC Method Development and Optimization Victoria F. Samanidou

2.1 Introduction

Chromatography is considered to be the main part of separation science that has been established over a century ago, and since then it has been catering to the needs of many scientific areas, including biological, pharmaceutical, food, forensic, environmental, and so on. High-pressure liquid chromatography (HPLC) is one of the most reliable analytical technologies used in any analytical laboratory. It is, beyond any doubt, the most powerful weapon in an analyst’s arsenal and the most useful tool toward any analytical challenge, no matter what the sample matrix is or how complicated it appears to be. In all chromatographic techniques, analytes are separated as they are distributed between two phases, the mobile phase and the stationary phase. It may sound an easy task, but an analyst knows that this is not true at all. The final chromatographic method is a jigsaw that has to be constructed by the analyst or the so-called “chromatographer.” He or she has to choose the right column, the effective mobile phase, the proper detection technique, and all necessary operational parameters, summarized under the term “chromatographic conditions,” which have to be used in order to achieve the desired target, that is, wellresolved peaks. The whole procedure is called HPLC method development, and undoubtedly, it is a complicated process. The physicochemical properties of the analytes to be separated are the driving force of the final goal, that is, resolution, and then the exciting journey to an optimum method starts. In the beginning, the chromatographer has to decide where to start from. These are the “initial conditions,” which are further optimized and the final method has to be validated before its application to real samples. The KISS principle is the preferred approach during method development. This means “Keep It Short and Simple,” in order to make a “chromatographer’s

Analytical Separation Science, First Edition. Edited by Jared L. Anderson, Alain Berthod, Verónica Pino Estévez, and Apryll M. Stalcup.  2015 Wiley-VCH Verlag GmbH & Co. KGaA. Published 2015 by Wiley-VCH Verlag GmbH & Co. KGaA.

26

2 Basic LC Method Development and Optimization

life” easier. Simplicity should be a key goal while designing the method development strategy and unnecessary complexity should be avoided. For example, a gradient is not necessary, when analytes can be easily separated under isocratic conditions. Similarly, a ternary gradient is not the best choice, when a binary seems to be efficient. Regardless of the final scope of the method application, whether it is routine analysis or a single-sample analysis, some steps in LC method development are common. In this chapter, the principal factors that have to be investigated during LC method development and optimization are discussed.

2.2 Theoretical Aspects

The separation of target analytes in HPLC mixtures relies on the following properties: charge, hydrophobicity, affinity, solubility, and molecular weight. According to these properties, chromatographic separations are classified into five major separation modes: 1) Adsorption chromatography, based on adsorption/desorption procedures. 2) Partition chromatography, based on partitioning of analytes between two liquid phases. 3) Ion-exchange chromatography, based on exchange of ions between surface ionic groups and ions in mobile phase. Ionic interactions depend on charge, ion size, and polarization. The pH also affects the separation. 4) Size exclusion chromatography, also known as gel permeation or gel filtration chromatography, based on molecular size. According to their size, some molecules are included in the pores of the stationary phase and thus they are retained, while some are excluded, so they pass through the column unretained. In this case, the pore size is critical and affects the retention and as a result the retention time. 5) Affinity chromatography in which analytes (enzymes, antibodies, etc.) bind to a ligand bound on a substrate and are subsequently eluted by using a chaotropic agent, by changing pH, or by using a specific eluent. As far as interactions that take place during chromatographic separation are concerned, electrostatic forces are stronger, while the polar interactions including hydrogen bonding (permanent dipoles), dipole-induced dipole, London dispersion forces, and van der Waals forces are the most universal interactions between molecules, although relatively weak [1–6]. Before delving deeper into the various aspects of the method development, it is necessary to describe briefly some fundamental terms [7].

2.2 Theoretical Aspects

2.2.1 Retention Factor k

The retention factor, k, describes the ability of the stationary phase to retain analytes and is given by Equation 2.1 k ˆ …t R

t 0 †=t 0 ;

(2.1)

where tR is the analyte’s retention time and t0 the column dead time, that is, the time that an unretained compound needs to pass through the column and reach the detector. Usually, uracil or nitrite and nitrate salts can be used in reversed phase HPLC for the determination of t0. 2.2.2 Selectivity α

Selectivity α, also called the separation factor, is an indication of the separation degree of two adjacent peaks and is given by Equation 2.2. α ˆ k 2 =k 1 ˆ …t R2

t 0 †=…t R1

t 0 †:

(2.2)

2.2.3 Peak Asymmetry

Any chromatographer wishes to obtain perfectly symmetric peaks, which can be more accurately quantified, but in the real world of chromatography, most peaks are asymmetric to some degree. Peak tailing factor (Tf) and peak asymmetry factor (As) are two terms that are used to describe the same phenomenon, and they result in slightly different numeric values due to the different calculation approach. Peak asymmetry factor is given by the equation As = b/a, where a is the width of the front half of the peak and b is the width of the back half of the peak at 5 or 10% of the peak height from baseline to a line dropped perpendicularly from the peak apex. However, according to US Pharmacopoeia, Tf = (a + b)/2a is used to describe peak asymmetry as shown in Figure 2.1. In general, a 10% peak height is considered for the peak asymmetry factor and 5% peak height for the peak tailing factor [5,6]. 2.2.4 Efficiency of Chromatographic Column and Theoretical Plates

The quality and subsequently the efficiency of a column are characterized by the separation ability and the separation efficiency, expressed by the height equivalent of theoretical plates (H = HETP) as given by Equation 2.3: L Hˆ ; (2.3) N where H and N are the height and the number of theoretical plates, respectively, and L is the length of the analytical column in centimeters.

27

28

2 Basic LC Method Development and Optimization

Figure 2.1 Peak asymmetry.

Actually, the efficiency of the stationary phase is described by the number of theoretical plates, N, and is calculated by Equation 2.4 or 2.5: N ˆ 16…t R =t b †2 N ˆ 5:54…t R =t h †

(2.4) 2

(2.5)

where tb is the peak width at baseline and th is the half-height peak width, both expressed in time units. 2.2.5 Resolution Rs

Resolution describes the separation degree between a pair of adjacent peaks and is given by Equations 2.6 and 2.7 Rs ˆ 2…t R2

t R1 †=…t b2 ‡ t b1 †

(2.6)

Rs ˆ 2…t R2

t R1 †=1:70…t h2 ‡ t h1 †

(2.7)

or

depending on the method of peak width measurement used. In Equation 2.6, peak width is measured at the baseline, whereas in (2.7) it is measured at the half height of the peak.

2.2 Theoretical Aspects

According to the Purnell equation, the factors that regulate the resolution of a column are as follows: pffiffiffiffi!   N α 1 k2 ; (2.8) Rs ˆ 4 k2 ‡ 1 α where N is the number of the theoretical plates, α is the selectivity, and k2 is the retention factor for the compound, which is eluted last. 2.2.6 The Fundamental vanDeemter Equation

According to the fundamental van Deemter equation (2.9), the efficiency of a column is related to the linear velocity of mobile phase u: H ˆA‡

B ‡ Cu u

(2.9)

where A is the eddy diffusion parameter, B is the diffusion coefficient that results in dispersion, and C is the resistance to mass transfer coefficient. The comparison of the performance of different chromatographic columns is made by the H versus u plots, known as the van Deemter plots. The optimum linear velocity is achieved when the slope of the plot H versus u is zero (dH/du = 0): rffiffiffiffi Dm B uopt ˆ (2.10) dp C where dp is the average particle size and Dm is the diffusion coefficient of the analyte in mobile phase. The value of H at the optimum linear velocity is given by Equation 2.11: pffiffiffiffiffiffiffiffiffiffi H min ˆ dP …A ‡ B
C † (2.11) At this point of the curve, the performance of the column is the highest. A theoretical van Deemter plot showing the relationship between the theoretical plate height and the mobile-phase velocity is presented in Figure 2.2. The A, B, and C terms of the van Deemter equation are given by (2.12)–(2.14) relationships. A ˆ 2λd p

(2.12)

B ˆ 2γ…Dm †

(2.13)

Cˆ

wd 2p Dm

(2.14)

where dp is the average particle diameter and λ is a constant almost close to 1, Dm is the analyte diffusion coefficient in the mobile phase, γ is a factor related to the diffusion restriction by column packing, and w is a coefficient determined by the pore size distribution, pore shape, and particle size distribution [1,4,8,9].

29

30

2 Basic LC Method Development and Optimization

Figure 2.2 A theoretical van Deemter plot showing the relationship between theoretical plate height and mobile-phase velocity ( http://www.restek.com/Technical-Resources/TechnicalLibrary/Pharmaceutical/pharm_A016).

2.3 Controlling Resolution

Resolution is governed by different physicochemical phenomena. Physics (band spreading) and chemistry in terms of selectivity and retention lead to wellresolved or not so well-resolved peaks. It is up to the chromatographer to obtain the desired resolution. Resolution can be influenced by changing one of the three parameters, selectivity, efficiency, and retention, as described by the Purnell equation (Equation 2.8), which determines the method development strategy. As shown in Figure 2.3, the factors that control resolution are N, k, and α. So, in order to reach the desired separation, the chromatographer has to

 increase N (either by increasing column length or by using more efficient column),

 increase k-values (by increasing the retention of analytes), and  increase α (by using a more selective column or mobile phase). As expected by the square root sign, efficiency (N) has a relatively smaller effect on resolution, but it can significantly influence run time. The number of theoretical plates can be increased by increasing the length of the analytical column, according to the van Deemter equation H = L/N. Resolution is improved as the peaks become narrower. The retention term k/(1 + k) can never exceed unity and, of course, an increase in k-values increases the total run time, while generally contributing to resolution to a lesser extent.

2.3 Controlling Resolution

Figure 2.3 Resolution versus selectivity, retention, and efficiency.

Obviously, selectivity (α) is the most important parameter as far as resolution is concerned, while it is hard to change. By changing selectivity, the peaks are resolved and k´ is often also changed as shown in Figure 2.4 [1,10].

Figure 2.4 The factors that control resolution.

31

32

2 Basic LC Method Development and Optimization

2.3.1 How to Improve N

While choosing the right packing material of the analytical column, the most common choice is the almost universal solution of octadecylsilane (C18 or ODS), which is very nonpolar. In this case, retention is based on London (dispersion) interactions with hydrophobic compounds. Octyl C8-bonded phases are also common, and are less hydrophobic than C18; therefore, retention times for hydrophobic compounds are typically shorter, with to some extent different selectivity. Phenyl-bonded phases are also nonpolar with retention being based on a mixed mechanism of hydrophobic and p p interactions. Exceptional selectivity results in p p interaction of the bonded phase with electron-deficient functional moieties of analytes. Cyanopropyl phases are of intermediate polarity, where retention is a mixed mechanism, resulting from both hydrophobic interactions and dipole interactions of the bonded phase CN group with solute amino groups or p p interactions with unsaturated sites. This phase is the best for polar organic compounds and is flexible enough to be used in both normal- and reversed phase modes. 2.3.1.1

Physical Characteristics of Packing Material

When considering the influence of the physical properties of the packing material, it is also important to take into account the influence of the column size, the particle size and shape, the surface area, the pore size, the carbon load, the bonding type, and the base material. Column dimensions: Refer to the length of column and the internal diameter of the packing bed within the column. The use of short columns (from 30 to 50 mm) results in short run times, faster equilibration, and low back pressure. On the other hand, the use of long columns (250–300 mm) results in higher resolution, higher back pressure, longer analysis times, and higher consumption of solvents used. Narrow columns produce narrower and taller peaks and a lower limit of detection. Particle shape and size: Chromatographic packing materials may be spherical particles or irregular in shape. Spherical particles offer reduced back pressures. Particle size: Refers to the average diameter of the particles. Although the manufacturers give a nominal size, this is typically the average size and some bigger or smaller particles are also included. Standard particle sizes range from 3 μm, which exhibit high efficiency, to 10 μm, although the latter are not so common in recent applications. Smaller particles allow less diffusion of analytes and thus they result in narrower and sharper peaks. However, smaller particles cause higher back pressures. Therefore, particles of 5 μm prevail in conventional systems and offer a good compromise between efficiency and back pressure.

2.3 Controlling Resolution

Recently, smaller porous particles, with a diameter of less than 2 μm, have gained significant attention. These particles, which are named sub-2 μm particles, offer both faster separations, without any loss of column efficiency, and lower limits of detection. The main drawback is that columns packed with such particles require special chromatographic systems, the so-called ultrahigh-pressure liquid chromatography (UHPLC) in order to overcome the high column pressure drop [4,9,11–14]. Surface area: The total surface area of a particle is the sum of the outer particle surface and the internal pore surface (expressed in m2/g). Packing materials with high surface area (300 m2/g) result in longer retention, greater capacity, and higher resolution than those with low surface area (200 m2/g). Pore size: it refers to the average size of the pores in porous packing materials. Larger pores allow larger solute molecules to be retained longer through maximum exposure to the surface area of the particles. A pore size of 150 Å or less is chosen for a sample with molecular weight (MW)  2000 and a pore size of 300 Å or greater is usually chosen for a sample with MW>2000. Bonding type: It refers to how the bonded phase is attached to the substrate. Monomeric bonding offers increased mass transfer rates, higher column efficiency, and faster column equilibration, while polymeric bonding results in increased column stability, especially with highly aqueous mobile phases. Carbon load: It is a good indicator of hydrophobic retention and refers to the amount of bonded phase attached to the base material. Endcapping: Important in reversed phase chromatography, endcapping is the process of bonding short hydrocarbon chains to free silanols remaining after the primary bonded phase has been added to the silica base. Endcapping reduces peak tailing of polar analytes that interact to a great extent with the most acidic silanols. 2.3.2 Increase of k

The retention factor k is increased by decreasing the percentage of organic modifier in the mobile phase. The pH of mobile phase also affects k, in the case of weak organic acids and amines, depending on their pKa value. In the range of pKa ± 1.5, a linear relationship between k and pH is observed. 2.3.3 Factors Influencing Selectivity or How to Improve α?

Selectivity α can be influenced by changing the mobile-phase type and composition, the pH, the concentration of buffer, the column temperature, and the packing material of analytical column.

33

34

2 Basic LC Method Development and Optimization

Figure 2.5 Snyder triangle.

2.3.3.1

Optimization of Mobile-Phase Composition

The solvent type (e.g., acetonitrile, methanol, and THF) is beyond any doubt the best approach to control selectivity. Kirkland and Snyder introduced “solvent triangle” (Figure 2.5), which describes the effect of organic modifier on the separation of analytes. A change in the percentage of organic modifier, the application of gradient elution, the type of gradient (e.g., binary and ternary), and the gradient profile (e.g., linear and multistep) are the main factors for altering selectivity. The Snyder triangle helps in finding the optimum mobile phase. The idea behind the triangle is that solvents in the same group will provide a comparable chromatographic selectivity. Therefore, switching from one solvent to another within the same group would not yield a spectacular change in selectivity as expected by switching to a solvent in a group with different characteristics, the same way as it happens by switching from group I to group VII, for example. Snyder’s triangle helps the chromatographer to choose wisely from among a selection of various solvents. For example, if methanol in the mobile phase is ineffective, a similar result is expected from the use of another alcohol such as ethanol or propanol. Choosing a solvent from an entirely different part of the triangle is more likely to yield the desired separation. On the other hand, larger

2.3 Controlling Resolution

alcohols, such as propanol, tend to be less denaturing to biomolecules than methanol; so, it may be the solvent of choice in some cases. Another useful application of the solvent triangle is to identify alternative mobile-phase solvents in terms of cost or availability [15,16]. 2.3.3.2

pH Control, Ion-Pair Reagents, and Other Additives

Controlling the pH can also play a significant role in changing the selectivity, when the analytes are weak acids or bases. Ionized forms are strongly retained in ion chromatography but are retained less in reversed phase, while nonionized forms exhibit the opposite behavior. Buffer type and buffer concentration are among the factors that need to be taken into consideration during method development. Ion-pair reagents, for example, hexane sulfonate, and other additives like tailing suppression reagents, such as triethylamine, can also affect resolution directly or indirectly. Optimization process can also be performed by the prediction of elution times, tR, of the analytes under examination using mathematical models in an attempt to minimize both the solvent consumption and the trial time [17]. 2.3.3.3

Temperature

Temperature is a significant factor that is often neglected or underestimated in liquid chromatography because it is erroneously related to gas chromatography (GC). However, it can also be powerful in resolution control and can be a critical parameter in HPLC, but in a narrower range and to a lesser extent than in GC. Resistance in mass transfer (the C term in van Deemter equation) is significantly reduced in elevated temperatures. In nearly all separations, an increase in temperature will also cause a decrease in retention. Moreover, a decreased solvent viscosity at elevated temperatures leads to lower back pressure. This allows the use of higher flow rates using standard equipment. Since high temperature leads to a flatter van Deemter curve, it enables the use of higher flow rates without sacrificing efficiency and thus optimizing the resolution (Figure 2.6). The temperature-programmed liquid chromatography (TPLC) is a favorable technique in many applications. However, disadvantages of elevated temperatures in HPLC should also be mentioned. For example, there are inconveniences related to instrumentation or type of columns. Undoubtedly, more research in this field is required [18]. 2.3.3.4

Stationary Phase and Column Selection

The stationary phase is another significant factor, which also plays an important role in selectivity control. A radical change of the sorbent type is more effective with regard to its polarity; however, a smaller effect is also expected among similar sorbents from different companies.

35

36

2 Basic LC Method Development and Optimization

Figure 2.6 Effect of temperature on column efficiency. (Reproduced from http://www.richrom .com/application/v2/public/upload/0/default/157.pdf.)

Obviously, all parameters exerting an influence on selectivity also have a significant influence on the retention factor. The particle size is an important parameter in separation control as is already mentioned by van Deemter equation and the A, B, and C terms as shown in Equations 2.12–2.14. By taking a closer look, one can see that A term (eddy diffusion) is independent of linear velocity of the mobile phase. Concerning the effect on efficiency, A is smaller for small particle size and for spherical particles or even better for no particles at all. However, a small particle size leads to a high back pressure in conventional HPLC, although the recent development in technology has solved this problem, either by using suitable instrumentation or by using modern fused core silica particles, with low pressure drop. The B term due to molecular diffusion (longitudinal) is caused by random motions of molecules and is large for smaller molecules. Dispersion increases when the analyte spends more time in mobile phase like it happens at slow flow rates. The C term due to mass transfer to and within the stationary phase yields a higher dispersion for high flow rates and less dispersion for small particles. Clearly, there are contradictory conditions, so the optimum one has to be chosen. 2.3.3.5

Stationary Phase and Packing Material Composition

Silica-based packing material is the most commonly used material available in a wide variety of bonded substrates. The main disadvantage is the pH range stability, which is limited from pH 2 to pH 8. Polymeric packing materials are more stable in the total pH range but have a poorer performance. These are the most often used materials in ion chromatography since anion- or cation-exchange sites are embedded in the styrene– divinylbenzene backbone. Other packing materials include zirconia-based columns as an alternative to the silica-based ones, which offer a unique interaction mechanism based on Lewis acid–base theory. In addition to the main hydrophobic interactions with the modified surface of the zirconia substrate, ion-exchange and ligand-exchange

2.4 Method Development Strategy

interactions also take place. Porous graphite carbon is a highly stable packing material mostly used for the resolution of geometrical isomers, and hybrids of silica core and polymers between silica and bonded phase form a combination that improves silica stability [19]. Recently, the use of superficially porous or solid core particles has been shown to have resulted in a better resolution. The terms superficially porous, core–shell, fused core, or solid core particles refer to particles that consist of a solid core and a porous outer shell. These particles provide faster separations compared to the large porous particles [20–23]. Monolithic columns where the stationary phase is a continuous porous material lead to low back pressures, and thus higher flow rates can be applied minimizing the analysis time without sacrificing resolution. These columns have several advantages over particulate columns [24], some of which are as follows:

 The porous polymeric rod improves both mass transfer and separation efficiency.

 They allow higher mobile phase flow rates with lower back pressure.  They exhibit stability over a wide pH range. 2.4 Method Development Strategy

During method development, the chromatographer has to take into consideration the following main parameters as shown in Figure 2.7: 1) Which column is the most suitable one? 2) Should a buffer be used in a mobile phase? 3) Which is the best organic modifier? For example, should acetonitrile or methanol be preferred? Obviously, each one has its own benefits and drawbacks. 4) Should the use of a mixture of organic solvents be considered, since sometimes the combination leads to a better resolution? 5) Which mixtures should be chosen: binary or ternary? 6) Is isocratic elution sufficient or should a gradient system be followed? 7) In case gradient elution is applied, is a binary sufficient or should a ternary one be followed? 8) In the case of gradient, what is the expected gradient time? How many steps are required? 9) What is the optimum temperature? 10) What is the optimum flow rate? Reasonably, changes start with those variables that are most possible to alter separation. The easier one to manipulate should be preferred and then one should turn to more complicated approaches.

37

38

2 Basic LC Method Development and Optimization

Figure 2.7 HPLC method development strategy.

2.4.1 Gradient Elution versus Isocratic

When two adjacent peaks cannot be well resolved by using isocratic elution, then gradient elution may help that also leads to a faster analysis. But still optimization is needed with regard to total gradient time, type of gradient, steepness, and so on. However, gradient elution has its own set of problems, such as ghost peaks, peak asymmetry, baseline drift, and so on. 2.4.2 Other Parameters in LC Method Development

Needless to say that additional parameters with regard to sample preparation, detection, and quantification also require optimization. External standard calibration is simpler, but injection errors may lead to false results. Internal standards are essential especially in the case of significant matrix effects. In this case, the difficulty of finding the most suitable internal standards is the main concern that arises. This compound has to fulfill some requirements, such as being absent from real samples, having similar properties but being differentiated and separated from analytes. Matrix effects are often a limited factor in HPLC. Therefore, the optimum sample preparation technique should be chosen. Finally, once desired resolution has been reached and the optimum LC method has been developed, this has to be validated before it is applied to the routine analysis. Method validation is performed in terms of precision, accuracy,

2.5 Current and Future Trends

limits of detection and quantitation, specificity, selectivity, linearity, range, robustness, and system suitability [25]. 2.5 Current and Future Trends

The “three S” model describes the current status in HPLC analysis requirements: separation (resolution), sample capacity, and speed are the principal demands. Advances in column and instrumentation technology are the key factors in present and future liquid chromatography. The combination of analytical columns with smaller particle size and the use of more sophisticated instrumentation with reduced dead volume, low sample carryover, and better pump and detector specifications ensure high speed and optimum separations. The van Deemter plot shows that smaller particles provide not only increased efficiency but also the ability to take advantage of this efficiency over an extended flow range. However, small particles cause increased back pressures. Moreover in conventional HPLC, faster chromatography reduces resolution. By increasing the flow rate, compression is observed. In ultra performance LC, efficiency is maintained over large areas of linear velocity (flow), which means that the flow can be increased without losing efficiency and this is the key for faster chromatography. Sub-2 μm particles offer the required improvement in column performance [11]. Numerous significant developments in materials science have been given by core–shell particle technology and monolithic columns, both of which have led to improved separation efficiency using relatively low pressures. Columns with core–shell particles provide excellent mass transfer kinetics and a lower C term, compared to totally porous particles, in the separation of both large and small molecules, due to the lower A and B terms of the van Deemter equation. The most common size of shell particles used nowadays is 2.5–2.7 μm. These particles compete with the sub-2 μm particles, as they provide faster separations with the same efficiency but with only half the back pressure. Therefore, the most important advantage of these particles is that they do not require a special LC system [12–14,26]. Monolithic columns consist of a single rod of a highly porous material and are prepared as a single-piece block into a tube. Silica-based or polymer-based monolithic columns offer many advantages compared to particle columns; for instance, since they have a higher column permeability, they yield faster separations and low column back pressure at high flow rates of the mobile phase [27,28]. 2.5.1 Two-Dimensional Chromatography

In some cases where a higher degree of selectivity is required, one-dimensional separation alone is not enough and multidimensional chromatography is required. In this case, parts of the separated sample components can be subjected to additional separation procedures, and the procedure is called two-dimensional

39

40

2 Basic LC Method Development and Optimization

HPLC according to IUPAC definition, 1997. A fraction eluting from the column can pass into another column with different separation characteristics. Instrumentation is based on a normal LC instrument equipped with an extra pump and a switching valve. The two columns are connected together via a multiport switching valve. The effluent from the first column can be directed, by switching the valve, to waste, to detector, or to the second column. The two-dimensional systems can be divided into two categories, namely, heart-cut and comprehensive techniques. In the comprehensive two-dimensional liquid chromatography (LC/LC) mode, the entire sample is subjected to two different separations, representative of the entire sample. It is used when information is needed from all sample components. In heart-cut approach, only one or a few fractions of the first separation are collected and then transferred to the second column for further separation (second dimension). These techniques are applied when only some components are analyzed from a complex matrix. A typical example of heart-cut LC (LC–LC) is the analysis of drugs in a biological sample (e.g., urine) and the first column is used mainly for selective cleanup and concentration [29]. 2.6 Conclusions

Separation science is an analytical tool of utmost importance. High-pressure liquid chromatography is one of the most reliable analytical technologies in any analytical laboratory. One of the most difficult tasks is to develop selective and fast separations in manual and/or automated approaches. No matter how sophisticated the instrumental design developments are, and no matter what improved sorbent materials are synthesized, the HPLC method development will always need the expertise of the chromatographer to arrive at an optimum final analytical method. There are various factors affecting resolution and method developments, each one to a different extent. Practically, a combination of solvent strength (% B or gradient type) and temperature provides the most powerful control in peak resolution. In the immediate future, new approaches will need to be explored. The ability to separate more and more species in complex matrices will remain a critical challenge. Advances in instrumentation and column technology, as well as in multidimensional separation approaches, will be the cornerstone of the chromatographic separation science. Although HPLC method development will continue to be based on chromatographer’s experience, software and mathematical models in method prediction may save a lot of the laboratory budget for organic solvents, not to mention the greener chemistry that will be achieved. References 1 Snyder, L.R., Kirkland, J.J., and Glajch, J.L.

(1997) Practical HPLC Method Development, 2nd edn, John Wiley & Sons, Inc., New York.

2 Kromidas, S. (2006) HPLC Made to

Measure: A Practical Handbook for Optimization, Wiley-VCH Verlag GmbH, Weinheim.

References 3 Samanidou, V.F. and Papadoyannis, I.N.

4

5

6

7

8

9

10

11

12

13

(2009) HPLC: the dominant separation technique with a wide range of applications, in Chromatography: Types, Techniques and Methods (ed. Frank Columbus), Nova Science Publishers, Hauppauge, NY. Unger, K.K. and Liapis, A.I. (2012) Adsorbents and columns in analytical high-performance liquid chromatography: a perspective with regard to development and understanding. J. Sep. Sci., 35, 1201–1212. Snyder, L.R., Kirkland, J.J., and Dolan, J.W. (2010) Introduction to Modern Liquid Chromatography, John Wiley & Sons, Inc., New Jersey, USA. Samanidou, V. and Nazyropoulou, C. (2014) Chapter 2, in High-Performance Liquid Chromatography (HPLC): Principles, Practices and Procedures (ed. Y. Zuo), Nova Science Publishers, Hauppauge, NY. Gritti, F., Sanchez, C.A., Farkas, T., and Guiochon, G. (2010) Achieving full performance of columns by optimizing HPLC instruments. J. Chromatogr. A, 1217, 3000–3012. Oláh, E., Fekete, S., Fekete, J., and Ganzler, K. (2010) Comparative study of new shelltype, sub-2 μm fully porous and monolith stationary phases, focusing on masstransfer resistance. J. Chromatogr. A, 1217, 3642–3653. Unger, K.K., Skudas, R., and Schulte, M.M. (2002) Particle packed columns and monolithic columns in high-performance liquid chromatography: comparison and critical appraisal. J. Chromatogr. A, 1184, 393–415. Dolan, J.W. (2010) Selectivity in reversedphase LC separations, part I: solvent-type selectivity. LCGC North Am., 28 (12) 1022–1027. Majors, R.E. (2012) Developments in HPLC/UHPLC column technology. LCGC Eur., Suppl. 14, 7–14. Fekete, S., Fekete, J., and Ganzler, K. (2009) Shell and small particles: evaluation of new technology. J. Pharm. Biomed. Anal., 49, 64–71. Fekete, S., Ganzler, K., and Fekete, J. (2011) Efficiency of the new sub-2μm

14

15

16

17

18

19

20

21

22

23

24

core–shell (KinetexTM) column in practice, applied for small and large molecule separation. J. Pharm. Biomed. Anal., 54, 482–490. Fekete, S., Oláh, E., and Fekete, J. (2012) Fast liquid chromatography: the domination of core–shell and very fine particles. J. Chromatogr. A, 1228, 57–71. Johnson, A.R. and Vitha, M.F. (2011) Chromatographic selectivity triangles. J. Chromatogr. A, 1218, 556–586. Vanbel, P.F. and Schoenmakers, P.J. (2009) Selection of adequate optimization criteria in chromatographic separations. Anal. Bioanal. Chem., 394, 1283–1289. Nikitas, P. and Pappa-Louisi, A. (2009) Retention models for isocratic and gradient elution in reversed phase liquid chromatography. J. Chromatogr. A, 1216 (10), 1737–1755. Vanhoenacker, G. and Sandra, P. (2006) Elevated temperature and temperature programming in conventional liquid chromatography: fundamentals and applications. J. Sep. Sci., 29 (12), 1822–1835. Janečková, L., Kalíková, K., Bosáková, Z., and Tesařová, E. (2010) Study of interaction mechanisms on zirconia-based polystyrene HPLC column. J. Sep. Sci., 33 (19), 3043–3051. Cunliffe, J. and Maloney, T. (2007) Fused core particle technology as an alternative to sub-2 μm particles to achieve high separation efficiency with low backpressure. J. Sep. Sci., 30, 3104–3109. DeStefano, J.J., Schuster, S.A., Lawhorn, J.M., and Kirkland, J.J. (2012) Performance characteristics of new superficially porous particles. J. Chromatogr. A, 1258, 76–83. Kirkland, J., Langlois, T., and DeStefano, J. (2007) Fused core particles for HPLC columns. Am. Lab., 39, 18–21. Omamogho, J.O., Nesterenko, E., Connolly, D., and Glennon, J.D. (2012) Next-generation stationary phases: properties and performance of core–shell columns. LCGC Eur., 4, 31–34. Samanidou, V. and Karageorgou, E. (2011) On the use of KinetexTM-C18 core–shell 2.6 μm stationary phase to the multi-class

41

42

2 Basic LC Method Development and Optimization

25

26

27

28

29

determination of antibiotics. Drug Test Anal., 3 (4), 234–244. Papadoyannis, I.N. and Samanidou, V.F. (2005) Validation of HPLC instrumentation, in Encyclopedia of Chromatography, 2nd edn (ed. J. Cazes), Marcel Dekker, New York, pp. 1743–1758. Gritti, F. and Guiochon, G. (2012) The current revolution in column technology: how it began, where is it going? J. Chromatogr. A, 1228, 2–19. Guiochon, G. (2007) Monolithic columns in high-performance liquid chromatography. J. Chromatogr. A, 1168, 101–168. Jandera, P. (2011) Stationary and mobile phases in hydrophilic interaction chromatography: a review. Anal. Chim. Acta, 692, 1–25. Tanaka, N., Kimura, H., Tokuda, D., Hosoya, K., Ikegami, T., Ishizuka, N., Minakuchi, H., Nakanishi, K., Shintani, Y., Furuno, M., and Cabrera, K. (2004) Simple and comprehensive two-dimensional reversed-phase HPLC using monolithic silica columns. Anal. Chem., 76 (5), 1273–1281.

Related Web Sites http://www.chem.agilent.com/Library/ eseminars/Public/Microsoft%20PowerPoint %20-%20rapid%20HPLC%20Method% 20Development.pdf (accessed September15, 2013). www.hplc.eu/halo.htm (accessed March 28, 2013).

http://www.interscience.nl/promotiesites/ hypersil/topics/promotiesites/hypersil/ nieuws/hypercarb_technical.pdf (accessed March 29, 2013). http://www.knauer.net/fileadmin/user_ upload/produkte/files/Dokumente/ columns/lc_columns/brochures/b_e_co_ blueshell_columns.pdf (accessed March 28, 2013). http://www.merckmillipore.si/.../201207.144 (accessed March 28, 2013). http://www.pharmtech.com/pharmtech/ Analytical/HPLC-Method-Developmentand-Validation-for-Pharmac/ArticleLong/ Article/detail/89002 (accessed September 15, 2013). http://www.phenomenex.com/Kinetex/ CoreShellTechnology (accessed March 28, 2013). http://www.restek.com/Technical-Resources/ Technical-Library/Pharmaceutical/ pharm_A016 (accessed March 28, 2013). http://www.richrom.com/application/v2/ public/upload/0/default/157.pdf (accessed September 15, 2013). http://www.separationsnow.com/coi/cda/ detail.cda?id=17076&type=Education Feature&chId=4&page=1 (accessed September 15, 2013). www.sequant.com/default.asp?ml=11625 (accessed March 21, 2013). http://www.sigmaaldrich.com/catalog/ product/supelco/814005?lang=en& region=GR (accessed March 29, 2013). http://www.sigmaaldrich.com/etc/medialib/ docs/Supelco/Posters/11887.Par.0001.File .tmp/11887.pdf (accessed March 29, 2013). http://www.shimadzu.com/an/hplc/support/ lib/lctalk/2dlc.html.

43

3 Recent Advances in Column Technology Ross Andrew Shalliker and Danijela Kocic

3.1 Introduction

Although it may not appear so, the HPLC column has undergone substantial changes in design since the first high-pressure packed columns were made. These early columns were manufactured using 10 μm ion-exchange particles and they were packed in a downward slurry process [1]. Since then, columns are typically packed with sub-3 μm particles, and we seem to be heading toward sub-2 μm particles, with even the 5 μm particles becoming less commonplace. With this downscale in particle size has come an increase in the number of theoretical plates per meter of column, which in the earlier days may have been generously estimated to be around 30 000–40 000 plates per meter to now typically on the order of 130 000 plates per meter for a 3 μm particle packed column, and nearing 180 000 plates per meter for 1.9 μm particle packed columns. This decrease in particle size has driven the need to increase the operational pressure of these columns, which in turn has led to the development of the UHPLC system. One added benefit of the advancements made in column technology was the design of chromatographic systems improved with greater attention paid to the extra column dead volume features of the instrument – but that is a different story. Columns are now typically 5 cm long rather than 25–30 cm; however, we still use just the same number, or even less, of plates in a given separation, some 40 years on. In the past decade or so, the concept of pellicular particles was revisited, but in particle sizes less than 3 μm, most often 2.6 μm. The first of these reborn pellicular particles, or core–shell particles as they are termed, was the halo phase, but since then numerous types of core–shell particles have been commercialized, such as the Kinetex and Accucore columns. These types of particles have provided a higher separation performance, in part, because they pack more uniformly into a tube yielding higher numbers of theoretical plates [2]; typically, 250 000 plates per meter offer far greater efficiency than the fully porous particles. Once silica was the primary stationary phase support, but now other types of ceramics are employed, for example, zirconia [3,4], titania [4], and alumina [4] which offer several significant benefits: pH stability, mechanical strength, and Analytical Separation Science, First Edition. Edited by Jared L. Anderson, Alain Berthod, Verónica Pino Estévez, and Apryll M. Stalcup.  2015 Wiley-VCH Verlag GmbH & Co. KGaA. Published 2015 by Wiley-VCH Verlag GmbH & Co. KGaA.

44

3 Recent Advances in Column Technology

differing selectivity. Silica has also been developed as a hybrid phase incorporating organics into the support matrix [5]. Likewise, polymeric particles have found widespread use in ion chromatography and in gel permeation chromatography. Selectivity is important in the development of separations and is, indeed, essential when dealing with complex samples. One needs only to glance through a supply catalog to testify the diversity of phases on offer. Such choice is particularly useful when developing two-dimensional separations, since for optimal performance each dimension should provide selectivity for the target species, ideally orthogonal retention behavior. Monolithic columns have been developed in silica and polymeric forms. Their advantage is the increased bed permeability [6], which allows for higher speed in separation performance. Their disadvantage is that the separation performance is poorer than that of the particle packed columns. Further development work is required here. While monoliths are an important aspect of column technology, they will not be discussed further here since an entire chapter could be devoted to monoliths. Indeed, Guiochon [6] comprehensively reviewed monoliths in 2007, and the reader can refer to his works for more information, albeit, an update is soon warranted. Perhaps, the most stagnant region of this development has been in the design of the container that houses the stationary phase. Largely, this container is a tube whose purpose is to house the particles in a tightly packed order and allow the solvent to traverse efficiently through the bed. Some novel design strategies were constructed in an effort to improve the solvent transport, namely, radial compression columns [7], which have soft walls that could be pressurized and bent around the particles to overcome wall effects (this will be discussed in more detail later). However, the radial pack column could not gain market acceptance and its use faded. Some specialized column formats have been tested, for example, an injection technique known as central point injection [8], where a needle is inserted into the bed at the inlet and the sample is introduced and restricted during migration in the radial central region of the column – the so-called “infinite diameter column” [8]. Its use in real-world applications was, however, limited. Similarly, in an attempt to overcome the wall effect, end-point detection processes were investigated [9–12], where instead of the bulk eluent passing through a detection source being used to collect chromatographic data, a localized detector is employed which is located usually in the central region of the outlet end of the column, such that the detector sees just the solute species that migrate through the radial central region of the column. This process greatly improves the separation performance, but it is difficult to implement in an automated and general run-of-the-mill laboratory. Hence, it has not found acceptance in commercial practice. More recently, a new design concept in HPLC columns, referred to as active flow technology (AFT) [13–16], is undergoing development, and this is the focus of this chapter. In order to understand this design feature, we first need to understand how columns are packed. Hence, we will begin our discussion by summarizing the column packing techniques and then we will evaluate factors that prove to be detrimental to column performance.

3.2 Column Packing: Downward Slurry Packing

3.2 Column Packing: Downward Slurry Packing

Downward slurry packing methods dominate commercial manufacturing processes, for at least analytical scale columns. However, manufacturers treat their packing protocols as closely guarded secrets. Furthermore, there are no generic methods for column packing since different types of materials demand different packing conditions, even to the point whereby not all types of C18 columns can be successfully packed using the same packing method. All this indicates that column packing processes remain incompletely understood. The particle size and the size distribution further complicate the process. The general principle of downward slurry packing is, however, relatively straightforward. In the first instance, a suitable solvent must be identified that allows the particles to be dispersed, avoiding the formation of agglomerates. The particles are stirred in this solvent until completely dispersed: ultrasonication should be employed to remove trapped air from within the particle pore structure. Next, a suitable packing or pushing solvent needs to be identified. The purpose of this solvent is to carry the particles into the column at high velocity and subsequently consolidate the particles tightly in the column tube. As a guide, acetone generally serves as a useful dispersive solvent for C18 particles and methanol as a packing solvent; although this combination does not work for all C18s, it is at least a useful starting point. These solvents are likely not suitable for all types of bonded phases. A typical example of column packing equipment used for downward slurry packed columns is shown in Figure 3.1. An air-driven fluid pump (a), capable of delivering solvent at high pressure and high volumetric flow rates, is usually employed. The piston volume of this pump should be small so as to avoid large-pressure fluctuations during piston chamber refill, or the piston volume

(b) (a) (c)

(d) (e) (f)

Figure 3.1 Schematic diagram of a typical downward slurry packing system: (a) high-pressure gas-driven fluid pump, (b) compressed gas, (c) gate valve, (d) slurry reservoir, (e) column blank, and (f) sacrificial column blank.

45

46

3 Recent Advances in Column Technology

should exceed that of the volume required to pack the column entirely so as to avoid any refill at all. The latter is technically more difficult to achieve due to the rigors associated with high-pressure and high-volumetric flow rates, but this depends on the column’s internal diameter. The pump is driven by compressed air (b). Pressure from the pump is built up against a gate valve (c), typically at around 7000 psi. To initiate the packing process, this gate valve is opened rapidly. Below the gate valve, there is a slurry reservoir (d), the column (e), and a sacrificial outlet section (f). Prior to loading the slurry reservoir, the column blank and the sacrificial outlet section are filled with a displacement solvent. Ideally, this solvent should have a similar density to that of the slurry, although obtaining a solvent to match is difficult, especially one that is safe and environment-friendly to use. Although not perfectly matched to the typical slurries used in reversed phase columns, dichloromethane serves as a suitable choice. The purpose of using this solvent is to prevent the premature settling of particles within the column blank prior to the initiation of high-pressure and high-velocity fluid flow. The slurry will remain suspended above the displacement solvent through the duration of the packing process [17]. The particle slurry is added to the slurry reservoir (d) carefully, but quickly, capped, and then the packing process begins. Upon the application of pressure to the slurry reservoir, the slurry travels into the column blank more or less as a discrete plug, which gradually compresses like a spring as it moves through the tubing [17]. Once the column cylinder is filled with particles, the packing solvent should be passed through the column bed until such a time that the pulsations on the packing pump reach a constant time interval.

3.3 Column Bed Heterogeneity

It is now well understood that packed chromatographic beds are heterogeneous, both in the axial and in the radial directions. While heterogeneity in the axial direction leads to a general broadening of an elution profile and a loss in efficiency, it is radial heterogeneity that is a far more serious contributor to the loss in column performance, since radial heterogeneity leads to a distortion of the elution profile, from the ideal cylindrical plug flow to a parabolic, cup-like elution profile. In the following discussion, the terms “radial” and “axial” heterogeneity have been evaluated separately, although there are commonalities in the cause of each aspect of heterogeneity. 3.3.1 Axial Heterogeneity

Even in modern HPLC columns, with the exception of core–shell particles, reduced plate heights lower than 2 are rarely reported. There are a few works that have reported h values near unity [18–21], but these were especially

3.3 Column Bed Heterogeneity

prepared columns and operated under strictly controlled conditions, certainly not suitable for routine applications. Perhaps, we have reached our limits with respect to how well spheres, whether monodisperse or not, can be physically packed inside a column, since inside the column there are significant variations in the packing density and hence local differences in the void space. One important aspect of the column packing process that leads to column packing irregularities is friction, which can affect both axial and radial homogeneity of the column bed. First, however, we will consider how friction is a contributing factor to axial heterogeneity. One analogy that describes very well this effect was articulated by Jaeger, Nagel, and Behringer [22]. The authors related the close packing of particles to the “close” packing of cars in parking spaces inside a parking lot. In this parking lot, there are no assigned spaces and the aim is to completely fill the lot with uniformly parked cars. If, however, one car in the lot is poorly parked, occupying more than an equivalent single space, then in order for another car to park in this lot, the poorly parked vehicle must be moved and parked in a more orderly fashion. To do this, many cars must be moved at the same time, since other cars have effectively parked in a fashion dictated by the poorly parked car. Particles undergoing packing in a cylindrical tube behave in the same manner; once a particle is poorly packed and other particles are consolidated around this particle, they are more or less fixed in place. A uniform bed may build up beyond their space, but their poorly packed presence has created a void space within the tube. In order for this void space to be filled and additional particles added to the column, many particles must be reordered. While on an individual basis the frictional forces between particles may be insignificant, when taken on a whole across the entire column, there is an enormous resistance to particle movement. Hence, how these particles enter the column and are then consolidated is critically important to the quality of the packed bed. For interest, however, the reader can refer to reference [23], which shows that these small individual friction forces between particles can be overcome using ultrasonic radiation applied to columns running very low-viscosity mobile phases at low flow rates. Understanding the significance of the parking lot effect is an important consideration when understanding how columns are physically prepared. In Section 3.2, the downward slurry packing process was discussed, and this technique is largely used to pack analytical scale columns. When doing so, pressure is built up against a gate valve and then rapidly released to drive the particles from the slurry reservoir into the column tubing. The particles enter the column tube at very high velocity, and the bed progressively builds up. If packing is undertaken at constant pressure, the flow velocity must, therefore, decrease as the bed length increases. Hence, particles on the outlet section of the bed have traveled at a higher velocity than particles at the inlet section of the bed. Systematically, therefore, there is a variation in the bed density along the axial direction of the column [24]. In a 30 cm bed, the bed density may increase by as much as 7% from the inlet to the outlet [24]; even on a 10 cm bed, the average bed density for the outlet 5 cm section of the column was observed to be 4% higher than the inlet 5 cm section [24].

47

48

3 Recent Advances in Column Technology

Figure 3.2 Plot of HETP curves for six, 5 cm sections of 30 cm column. (Reproduced with permission from Ref. [24].)

More serious than this variation in bed density is, however, the uniformity of the packing material. At the column inlet, where the bed is least dense and the packing material is not protected from the high-velocity fluid flow stream (since there is no frit), the column bed is poorly packed. Channeling as a consequence of the unprotected bed surface being exposed to the flow stream may be a factor in the poor performance. The HETP curves shown in Figure 3.2 illustrate this phenomenon. The column labeled as C61 is the inlet 5 cm section of a 30 cm bed (isolated from the 30 cm section using techniques described in reference [24]). The HETP curve reflects the poorest separation performance of any of the sections within the bed. At the outlet section of the column, that is, a 5 cm section extracted from a 30 cm bed [24], the packing density is at its highest, although the HETP curve (Figure 3.2) does not reflect that of the best performing section of the column. Hence, a high bed density does not necessarily reflect a better packed bed. To understand why this section of the column is poorly packed, we need to appreciate the packing process itself. Packing commences following the rapid opening of the gate valve. Studies on glass column have shown that during the transportation process of slurry to the column, the slurry moves as a plug and compresses like a spring [17]. The packing pump used to pack the column in reference [24], the HETP curves for which are shown in Figure 3.2, had a stroke volume of 11 ml. Approximately, three pump strokes were required before the bed was fully built within the column. During refill of the pump piston chamber, a pressure pulse was established–around 2000–3000 psi. After these initial pump pulsations, the fluctuations decrease by a factor of 10.

3.3 Column Bed Heterogeneity

As the bed builds within the column housing, the velocity of fluid decreases to maintain constant pressure. This means that the particles that impinge upon the outlet frit of the column blank do so with the greatest momentum of any other particle impinging upon the bed. This results in a decreasing bed density from the outlet to the inlet of the column, and leads to the question why the mass of the packing material was greatest in the outlet section and least in the inlet section. This was observed in the columns packed in reference [24], the HETP curves for which are shown in Figure 3.2. One might consider that the greater the packing density, the less voids within the bed, and hence the better the column efficiency. Yet, the HETP curves do not support this assumption. Rather, the efficiency of this outlet section of the column was very poor. The poor performance of this section is likely to arise from the initial two–three pumping strokes in which the pressure change in the column is on the order of 2000–3000 psi. During these pulsations, the bed at this point in time has not yet been consolidated and during these first pump strokes, the bed undergoes a rapid succession of large expansions and compressions. This causes the particles in the outlet section to move upward from the frit in a manner similar to the fluffing of a finely divided solid in a vacuum that is rapidly released to atmospheric pressure. Even a minor disturbance here is catastrophic to the bed quality, since these particles cannot return to their original position within the bed with the same degree of momentum and as such voids are established [24]. Once these voids are established, they remain fixed in place since the bed builds up quickly on top of these particles. To remove these voids, the parking lot effect must come into play. Once the outlet and inlet sections of the column have been removed, the remaining internal sections of the column are relatively homogeneous, with respect to the axial direction [24] (Figure 3.2). 3.3.2 Radial Heterogeneity and the Wall Effect

The effects of radial bed heterogeneity on column efficiency are far more serious. The details of radial bed heterogeneity were recently reviewed [16], and as such only a brief account of the bed heterogeneity is considered here. Knox, Laird, and Raven [8,25], Golay [26], and Eon [7] were the pioneers who investigated the column bed heterogeneity. These early works were undertaken in dry packed beds and independently they identified that the flow velocity of the mobile phase in the radial column center was different from the flow velocity of the mobile phase near the wall. In the dry packed beds of reference [8], the flow velocity increased as the wall was approached, and at the same time, the reduced plate height increased from the radial center (h = 1.7) toward the wall (h = 4.7). It was these early studies of Knox, Laird, and Raven [8,25] that paved the way for the notion of a “wall effect.” Eon [7] then studied in more detail the “wall effect” and in particular focused attention on minimizing wall effects by using soft-walled columns or radial compression columns as they are more commonly known.

49

50

3 Recent Advances in Column Technology

Eon’s study [7] was one of the earliest that described the concept of these radial compression columns. Later, Baur and Wightman [9] studied the column bed heterogeneity in slurry packed columns, the particle size of which was 3 μm. They used an end-column microelectrode that could be precisely positioned at various radial locations. They were able to show that the reduced plate height increased significantly as the wall was approached; in the radial column center, the reduced plate height was 1.9, but near the wall, the reduced plate height was 4.2. This finding was similar to that of Knox and Parcher [8], but there was one significant difference: the flow velocity in the slurry packed columns was ∼5% higher in the radial column center than near the wall. Later, studies by Farkas et al. [10–12] investigated the radial column bed heterogeneity. Like Baur and Wrightman [9], they used microelectrodes [10] at the column outlet to observe how the flow velocity and column efficiency varied as a function of the radial location. They also used fluorescence detection to improve spatial resolution [11,12]. They found that the velocity was systematically lower near the wall than in the radial center of the column, irrespective of the particle diameter, and the efficiency was highest in the radial column central region of the bed. The magnitude in the difference between the flow velocities through the center compared to the wall region decreased as the flow rate decreased, until it became effectively negligible when the column was operated at its optimal flow velocity. Tallarek and coworkers used magnetic resonance imaging to evaluate the column bed heterogeneity [27–30], and like Eon their studies focused on radial compression columns. In columns packed with large irregular particles, localized variations in radial packing density (as much as 30%) were observed. When the column was radially compressed, the radial packing density varied even more. Columns packed with smaller spherical particles (6 μm) were in contrast effectively homogeneous, with variation across the column being less than 1%. The on-column matched refractive index detection process developed by Shalliker, Broyles, and Guiochon [31] was also used to study a variety of flow phenomena that detailed heterogeneity in the radial flow of the mobile phase through slurry packed beds. The most important outcome of these studies was the physical proof that detailed the wall effect [32]. They showed for the first time that there were two wall effects: the first wall effect, being a consequence of the column and particle geometry, neither the rigid particles nor the rigid column wall could bend to accommodate the other. As a consequence, the void space next to the wall was the greatest within the column. Hence, in this region of the packed bed, the permeability was at its highest. So, the flow velocity here was the highest; in fact, they observed a 37% increase in the flow velocity at the wall compared to the column radial center. The second wall effect was a result of the packing density increasing gradually as the wall was approached (but not at the wall). Near the wall, the flow velocity was the slowest, since the packing density was the greatest. In the radial central section of the bed, which extended from the radial center to around two-thirds of the column internal diameter, the

3.4 Active Flow Technology: A New Design Concept in Chromatography Columns

flow was relatively homogeneous. The solute migration efficiency was at its highest in this central region, decreasing as the wall region was approached. Subsequently, these studies verified the general findings of Baur and Wightman [9] and Farkas et al. [10–12] and provided a conclusive proof of bed heterogeneity with a detailed description of the wall effect.

3.4 Active Flow Technology: A New Design Concept in Chromatography Columns

In an effort to overcome column bed heterogeneity, at least with respect to radial heterogeneity and wall effects, a new design concept in column technology was developed. This technology has been referred to as AFT [13–16] and comprises a suite of columns that offer various performance attributes beyond the capabilities of conventional HPLC columns. Two of the most important columns within the AFT suite are (i) the parallel segmented flow column (PSF) [14,16] and (ii) the curtain flow column (CF) [13,16]. Both of these columns provide a distinct advantage over conventional columns and their design features and mode of operation will now be discussed in detail. 3.4.1 AFT Columns: Parallel Segmented Flow

Figure 3.3 is an illustration of a parallel segmented flow chromatography column, which has been designed to separate the flow streams that migrate in the wall region from the radial central region of the column [16]. In that way,

Figure 3.3 Illustration of the AFT end fitting, that is, the parallel segmented flow fitting, showing a sketch of the outlet frit and the outlet end fitting. (Reproduced with permission from Ref. [14].)

51

52

3 Recent Advances in Column Technology

inefficiency associated with wall effects and bed heterogeneity is overcome. The design of the column consists of an annular frit whereby the central portion of the frit is separated from its outer portion by a solid PEEK ring. This frit prevents cross dispersion of solute between the radial central region of the migration zone and the wall or peripheral zone of the column. The frit is housed in an outlet fitting that has multiple exit ports: a central port that directs the flow from the radial central region of the column and a peripheral port or ports that capture the flow eluting from the outer annulus or wall region of the frit. The purpose of this outlet fitting is to segment the flow into two portions: (1) the central portion and (2) the peripheral or wall region portion. The flow from either of these ports can be processed further, that is, passed through a detector or the sample collected in a fraction collector. The segmentation ratio (central flow : wall flow) is easily adjusted by regulating the pressure in any of the respective outlet ports. In fact, differential flow rates can be established through any of the exit ports, whether it is the two-port end-fitting design or the four-port endfitting design, or any other multi-port design. Another benefit of the multiport end fitting is that it enables multiplexed detection processes to be developed that enhance the analysis of complex samples [33–35]. A key to understanding the benefits of active flow technology is the realization that segmentation of the outlet flow from the chromatography column creates what is in effect a “virtual” column inside the physical column [15]. Solutes that migrate through the radial central region of the column exit the column via the radial central exit port, that is, they traverse the bed through the virtual column and migrate without any interaction with the column wall. The virtual column is in effect a wall-less column. The virtual diameter of this wall-less column can be tuned, perhaps, either to suit the needs of the detection mode [15,36] or to influence the level of column efficiency [14,15] or sensitivity [14]. For example, if the physical internal diameter of the parallel segmented flow column is 4.6. mm, and 21% of the flow eluted from the radial central outlet port, then effectively a virtual column is established, which has a virtual internal diameter equivalent to 2.1 mm. Likewise, if 43% of the flow were allowed to elute through the radial central exit port, then a virtual 3.0 mm i.d. column would be established. It is a very simple task to tune this segmentation ratio, and column efficiency and sensitivity both depend on the segmentation ratio. More details regarding this will follow as we discuss the separation performance. 3.4.2 AFT Columns: Curtain Flow

Curtain flow chromatography columns [13] utilize the same outlet fittings as detailed in Figure 3.3 for the PSF columns. But, curtain flow columns differ from the PSF columns because the multiple port end fitting and frit are located at both the column inlet and the column outlet. The sample is introduced into the radial central region of the column through the radial central inlet port. Mobile phase is introduced into the column through both the radial central inlet port and the

3.4 Active Flow Technology: A New Design Concept in Chromatography Columns

peripheral port(s). This can be achieved by either splitting the flow prior to the injector or by using two pumping devices, one for the central flow and the other for the peripheral flow [37]. On a 4.6 mm i.d. column, the ratio of the central flow to the peripheral should be somewhere in the vicinity of 35 : 65 [13] for best performance. Since the sample enters the column via the radial central inlet port, it then enters into the central porous region of the annular frit. The impermeable ring located inside the inlet frit prevents the sample from dispersing through the frit to the wall. Once the sample load enters the column, the peripheral flow of mobile phase then further inhibits the diffusion of solute to the wall, effectively establishing a curtain flow environment. Solutes then migrate through the column with no interaction with the wall. The principles of the infinite diameter column are thus applied. The outlet segmentation ratio can, however, be adjusted to tune the internal diameter of the virtual column so as to optimize efficiency and sensitivity, with respect to the detection mode being utilized. 3.4.3 Performance of AFT Columns 3.4.3.1

Sensitivity

Because packed column beds are not perfectly homogeneous, and that the sample is not uniformly distributed across the inlet section on the column, the AFT end fitting and frit assembly does not function in the same manner as the postcolumn flow stream splitting. The heterogeneity of the packed bed generates

Figure 3.4 Photograph illustrating a regular injection profile through a glass packed column. (Reproduced with permission from Ref. [38].)

53

54

3 Recent Advances in Column Technology

flow profiles that resemble partially filled bowls [38], similar to that shown in Figure 3.4 (a photograph of a band profile inside a glass column operating in a matched refractive index system according to reference [38]). When viewed from the trailing edge of the band profile, it is apparent that the band is partially hollow. The post-column flow stream splitting samples the entire band profile equally, that is, the concentration of the solute in either portion of the split that emanates from a post-column splitting system will be in proportion to the split ratio. However, the segmented outlet fitting on an AFT column samples the inner core region of the profile separately from the outer core region of the profile, effectively heart cutting the radial central region of the band, resembling a doughnut. The concentration of the radial central portion is not equal to the concentration of the outer core region because (1) frits do not distribute the sample uniformly across the head of the column [39–41] and (2) the column is less efficient near the wall [32], so band broadening is greater near the wall and thus the sample is diluted. Hence, even when small portions of the flow are sampled from the radial central outlet section of the column, the concentration response on a detector can be as high as, or even higher than, the larger volume peripheral portion of the flow, or for that matter, the entire portion of the flow eluting from the exact same format conventional column carrying the same solute loading. This is clearly shown in Figure 3.5, where it can be seen that at a segmentation ratio of 15% through the column central exit port, there is the same sensitivity in UV detection as 100% of the sample eluting from a conventional column [14], such is the significance of the radial heterogeneity and the bowl-shaped elution profile. Another important aspect from this data is that the band volume is decreased in the exact proportion to the outlet segmentation ratio, and this has major benefits when utilizing flow sensitive detectors, such as the mass spectrometer (to be discussed later). Sensitivity in detection can be boosted in CF column mode, because the sample is restricted to the radial central region of the bed. In CF, sensitivity with UV detection is almost three times higher than conventional columns [37] as shown by the chromatographic trace in Figure 3.5. Even greater gains in sensitivity are achieved when CF columns are used with MS detection [36]. In some instances, almost 70-fold gains in S/N responses have been observed as a consequence of a substantial reduction in the baseline noise when AFT columns are used [36]. The noise response has been shown to be solute dependent, with the largest variations arising from separations undertaken on conventional columns [36], rather than the AFT columns. For example, an illustration of the noise response is shown in Figure 3.6a for extracted ion chromatograms of phenylalanine and Figure 3.6b for extracted ion chromatograms of methionine. Note the noise response for these two amino acids on the AFT columns remains almost constant, but is highly variable for the data collected on the conventional columns [36]. 3.4.3.2

Efficiency

When comparing the efficiency of AFT columns and conventional columns, two sets of reference points need to be established. The first is a comparison between

3.4 Active Flow Technology: A New Design Concept in Chromatography Columns

14

Butyl benzene

12

Intensity (mV)

10 8 6 4 2 100

75 63 45 30 15

0 -2 -0.5

-0.4

-0.3

-0.2

-0.1 0.0 0.1 Peak volume (ml)

Figure 3.5 Band profiles of butyl benzene eluting from the column operating in a normal mode (100) and segmented modes 75–15% solvent exiting via the central port as marked. Bandwidth expressed in units of volume.

0.2

0.3

0.4

0.5

Mobile phase 30/70 water/methanol, flow rate 1.1 ml/min, injection volume 2 μl, and detection 250 nm. (Reproduced with permission from Ref. [14].)

AFT columns and conventional columns whereby they both have the same physical internal diameters; of course, the particle size and length also remain constant. The second reference point is the comparison between a conventional column that has the same physical internal diameter as the internal diameter of the virtual column that is established by the AFT column. The reason this second reference point is required is that there is a comparison made at constant linear velocity through the flow cell of a detector between the two column sets. When the flow is segmented at the column outlet on an AFT column, the flow rate through the detector is reduced and depends on the segmentation ratio; hence, the residence time in the detector is increased. Thus, when making comparisons between the virtual column and the conventional column with a matching internal diameter, flow rates and detector residence times are the same. This provides for an even comparison, with respect to the flow rate and to the volume of an elution profile. The efficiency of the AFT column is generally underestimated by 7–10% because of the increase in residence time within the detection flow cell. The efficiency of an AFT column depends on the segmentation ratio [14,15]. At very low segmentation ratios, post-column dead volume contributions become more significant and this decays the efficiency gains obtained, relative to a conventional column with the same physical internal diameter as the AFT column. Albeit, the efficiency of the AFT column remains very much higher than the comparative conventional column that has the same internal diameter as the

55

3 Recent Advances in Column Technology

(a)

Void time –vacancy peak

3.0x105

Signal intensity

2.5x105

2.0x105

1.5x105

1.0x105

5.0x104

0.0 0.000

(b)

0.025

0.050

0.075 0.100 0.125 Normalized time

0.150

0.175

0.200

0.025

0.050

0.075 0.100 0.125 Normalized time

0.150

0.175

0.200

3.0x105 2.5x105

Signal intensity

56

2.0x105 1.5x105 1.0x105 5.0x10

4

0.0 0.000

Figure 3.6 (a) Plots showing the noise contributions for each of the three columns for extracted ion chromatograms of methionine. Note the time axis is expressed as a fraction of the total run time, that is, normalized time between 0 and 1. Each plot represents the first 20% of the total time. (i) Curtain flow column,

(ii) 2.1 mm i.d. conventional column (bold), and (iii) 4.6 mm i.d. conventional column. (Reproduced with permission from Ref. [36].) (b) Plot of the baseline noise response for the EIC for m/z 150 (methionine) on the same scale as that for phenylalanine in Figure 3.6a. (Reproduced with permission from Ref. [36].)

3.4 Active Flow Technology: A New Design Concept in Chromatography Columns

12000 Parallel Segmented Flow Column

11000 10000

N

9000

4.6 mm conventional

3.0 mm conventional

8000 7000

2.1 mm conventional

6000 5000

0

10

20 30 40 50 60 70 80 % volumetric flow through column center

Figure 3.7 Comparison in N-values obtained on the 4.6 mm i.d. parallel segmented outlet flow column, the 4.6 mm i.d. conventional column, the 3.0 mm i.d. conventional column, and the 2.1 mm i.d. conventional column. Test solutes: toluene (squares), propylbenzene (circles), and butylbenzene (diamonds). All columns were packed with 5 μm Hypersil Gold particles in formats of 2.1, 3.0, and 4.6 mm i.d.,

90

100

and each column was 100 mm in length. (Reproduced with permission from Ref. [15].) Note the data obtained on the 2.1 mm and 3.0 mm i.d. columns are centered on the 21% and 43% volumetric flow positions to correspond to the equivalent flow through the 4.6 mm i.d. column at that specific segmentation ratio.

virtual column established by the AFT column [15]. Figure 3.7, for example, shows the magnitude of typical gains in efficiency: AFT versus a variety of conventional columns [15]. Gains in N were as great as 70%. The efficiency gain of an AFT column is also length dependent. Smaller gains are apparent using longer columns, presumably because the solute species has greater opportunity to migrate as part of its elution time in the wall region [42]. For columns packed with 5 μm particles, the optimal length was on the order of 15 cm [42], where gains in efficiency were by as much as a unit value in the reduced plate height term. While for columns packed with 3 μm particles, the gain in efficiency decreased as the column length increased for columns between 3 and 10 cm in length. However, the gain in efficiency for the 3 μm particle-packed columns was far greater than for the 5 μm particle-packed columns. On the 3 cm column, for example, gains in efficiency were by as much as 4 reduced plate height units and about 1.8 reduced plate units on the 10 cm column. A very significant benefit of the AFT columns is that the efficiency gain when compared to conventional columns increases as the flow rate increases, and this favors high-throughput applications. The HETP curves in Figure 3.8, for

57

3 Recent Advances in Column Technology

20

3 μm - 2.1 mm Conventional 5 μm - 2.1 mm PSF

18

H (μm)

58

20% efficiency gain 50% decrease in back pressure

16

14

12 10 0.0

0.5

1.0

1.5 2.0 υ (mm/s)

2.5

3.0

3.5

Figure 3.8 HETP curves on 50 mm × 2.1 mm format columns: the conventional column (squares) was packed with 3 μm Hypersil Gold particles and the PSF column was packed with 5 μm Hypersil Gold particles. (Reproduced with permission from Ref. [43].)

example, show that a 5 μm particle-packed AFT column, with a virtual internal diameter of 1 mm outperforms a 2.1 mm internal diameter conventional column packed with 3 μm particles when the flow velocity exceeds ∼1.6 mm/s [43]. When the 3 μm particle-packed column has reached its operational pressure limit, the 5 μm particle-packed AFT column has 20% more theoretical plates and functions at half the back pressure [43]. 3.4.3.3

Speed

Perhaps, the most impressive benefit of AFT columns can be seen in ultrahighspeed separations, especially when these are coupled to flow-rate-limited detectors, such as the mass spectrometer. In general, the HPLC process limits the throughput of HPLC–MS analyses because of the time required to undertake a separation in the LC step. The MS process limits the throughput of the HPLC– MS assay because of the limited volume of the solvent that can be processed by the MS. This is essentially the reason why HPLC–MS is undertaken using narrow bore columns – higher throughput in the LC, but without overloading the volume capacity of the MS. Narrow bore AFT columns (2.1 mm i.d.) provide for very high efficiency at very high linear velocities compared to the conventional columns. The HETP curves in Figure 3.9 show the efficiencies of conventional and AFT columns over a very wide range of linear velocities. The upper velocity tested corresponded to a flow rate of 2 ml/min. On the conventional column, 1000 theoretical plates were obtained, but on the AFT column operating with a 21% segmentation ratio, 2000 theoretical plates were obtained; effectively, the 5 μm PSF column functions from an efficiency perspective as if packed with 2.5 μm particles, but with the pressure of a 5 μm particle-packed column.

3.4 Active Flow Technology: A New Design Concept in Chromatography Columns

5.5x10-5

2.1 mm i.d. conventional 2.1 mm i.d. PSF

5.0x10-5

1000 plates

4.5x10-5 4.0x10-5 3.5x10-5 H (m)

3.0x10-5 2.5x10-5

2000 plates

2.0x10-5 1.5x10-5 1.0x10-5 -6

5.0x10

0.0 0

2

4

6

8 v (mm/s)

10

12

14

16

Figure 3.9 HETP curves on 50 mm × 2.1 mm format columns both packed with 5 μm particles (Hypersil Gold). Test solute: propylbenzene, 40/60 water/methanol mobile phase. Injection volume: 1 μl. UV detection at 254 nm.

Given these gains in efficiency, and the fact that the solvent load to the MS detector is greatly reduced, there is an exceptional potential for very highthroughput operations. We have, in fact, named this technique “pulsed direct injection” HPLC–MS. This refers to a separation technique with HPLC and MS, whereby the process acts in a manner similar to direct injection into the MS (in the absence of a column); however, this direct injection is pulsed, by virtue of the fact that an AFT column is incorporated into the scheme, but it operates in a manner such that it has virtually no impact on the processing speed. It does, however, establish a “pulse” into the assay, the frequency of which is dictated by the injection cycle time. The extracted ion chromatograms in Figure 3.10 illustrate nicely this technique. In this separation, a 50 mm × 2.1 mm i.d. PSF column, packed with 5 μm particles, was employed, and the flow rate was 3 ml/min, resulting in a cycle time (injection to injection) of 18 s, 6 s of which was the time required to inject the sample; the separation was completed in 12 s. In Figure 3.10, there are 10 replicate injections of a three-component sample: acetaminophen, caffeine, and piroxicam. We deliberately chose conditions such that acetaminophen and caffeine coeluted in the HPLC separation but were separated by the triple-quadrupole MS. The precision in area quantification was excellent: 2.4% for acetaminophen, 1.2% for caffeine, and 1.4% for piroxicam. This type of analytical technique shows a great promise in processes that involve reaction monitoring, where assay speed and selectivity is essential.

59

60

3 Recent Advances in Column Technology Acetaminophen Caffeine Piroxicam

5

7x10

5

6x10

5

Response

5x10

5

4x10

5

3x10

5

2x10

5

1x10

0 0.10

0.15

0.20

0.25

0.30

0.35

0.40

Time (min)

Acetaminophen Caffeine Piroxicam

5

7x10

5

6x10

5

Response

5x10

5

4x10

5

3x10

5

2x10

5

1x10

0 0.0

0.5

1.0

1.5

2.0

2.5

Time (min) Figure 3.10 The analysis of three pharmaceuticals using pulsed direct injection HPLC– MS. Flow rate, 3.0 ml/min, 40/60 water/ methanol mobile phase (0.1% FA). Column

50 mm × 2.1 mm i.d. Hypersil Gold in PSF mode. Outlet segmentation ratio 15% through the column center to MS detector.

3.5 Summary

There have to date been no advances in column technology that are able to provide the benefits of increased sensitivity, efficiency, and speed in a single design

References

format. Usually, gains in sensitivity come at the cost of speed, since separations need to be run at a flow rate near the minimum in a HETP curve. Curtain flow chromatography, and even PSF chromatography, overcomes this limitation. Likewise, the cost of column efficiency is usually paid for in the currency of time. But AFT columns can provide higher efficiency than conventional columns, especially as the flow rate increases. Hence, the AFT columns function at their best when separations are run fast. Another factor that usually makes the cost of speed too much to bear is the limitation in detection processing, especially if the detector is a mass spectrometer. With AFT columns, less solvent enters the MS, in fact, in direct proportion to the segmentation ratio. Thus, ultrafast, highly efficient, and very sensitive assays can be undertaken using AFT columns with MS detectors.

References 1 Scott, C.D. and Lee, N.E. (1969) 2

3

4

5

6 7 8 9 10 11 12 13

14

J. Chromatogr., 42, 263–265. Bruns, S., Stoeckel, D., Smarsly, B.M., and Tallarek, U. (2012) J. Chromatogr., 1268, 53–63. Nawrocki, J., Rigney, M., McCormick, A., and Carr, P.W. (1993) J. Chromatogr., 657, 229–282. Nawrocki, J., Dunlap, C.A., McCormick, A., and Carr, P.W. (2004) J. Chromatogr., 1028, 1–30. Yu, H., Jia, C., Wu, H., Song, G., Jin, Y., and Ke, Y., and Liang, X. (2012) J. Chromatogr., 1247, 63–70. Guiochon, G. (2007) J. Chromatogr., 1168, 101–168. Eon, C.H. (1978) J. Chromatogr., 149, 29–42. Knox, J.H. and Parcher, J.F. (1969) Anal. Chem., 41, 1599–1606. Baur, J.E. and Wightman, R.M. (1989) J. Chromatogr., 482, 65–73. Farkas, T., Chambers, J.Q., and Guiochon, G. (1994) J. Chromatogr. A., 679, 231–245. Farkas, T., Sepaniak, M.J., and Guiochon, G. (1996) J. Chromatogr. A., 740, 169–181. Farkas, T. and Guiochon, G. (1997) Anal. Chem., 69, 4592–4600. Camenzuli, M., Ritchie, H.J., and Ladine, J.R., and Shalliker, R.A. (2011) Analyst, 136, 5127–5130. Camenzuli, M., Ritchie, H.J., Ladine, J.R., and Shalliker, R.A. (2012) J. Chromatogr. A, 1232, 47–51.

15 Shalliker, R.A., Camenzuli, M., Pereira, L.,

16 17

18 19 20 21 22

23

24 25 26

27

28

and Ritchie, H.J. (2012) J. Chromatogr. A, 1262, 64–69. Shalliker, R.A. and Ritchie, H.J. (2014) J. Chromatogr. A., 1335, 122–135. Shalliker, R.A., Broyles, B.S., and Guiochon, G. (1998) Am. Lab., 30 (21), 124–129. Giddings, J.C. and Robison, R.A. (1962) Anal. Chem., 34, 885–890. Giddings, J.C. (1963) Anal. Chem., 35, 1338–1341. Halasz, I. and Heine, E. (1962) Nature, 194, 971. Sternberg, J.C. and Poulson, R.E. (1964) Anal. Chem., 36, 1492–1502. Jaeger, H.M., Nagel, S.R., and Behringer, R.P. (1996) Rev. Mod. Phys., 68, 1259–1273. Shalliker, R.A., Broyles, B.S., and Guiochon, G. (2000) J. Chromatogr. A., 878, 153–163. Wong, V., Shalliker, R.A., and Guiochon, G. (2004) Anal. Chem., 76 (9), 2601–2608. Knox, J.H., Laird, G.R., and Raven, P.A. (1976) J. Chromatogr., 122, 129–145. Golay, M.J.E. (1961) Gas Chromatography (eds H.J. Noebels, R.F. Wall, and N. Brenner), Academic Press, New York, p. 11. Tallarek, U., Baumeister, E., Albert, K., Bayer, E., and Guiochon, G. (1995) J. Chromatogr. A., 696, 1–18. Van Dusschoten, D., Tallarek, U., Scheenen, T., Neue, U.D., and Van As, H.

61

62

3 Recent Advances in Column Technology

29 30

31

32

33 34

35

36

(1998) Mag. Res. Imaging, 16 (5/6), 703–706. Tallarek, U., Bayer, E., and Guiochon, G. (1998) J. Am. Chem. Soc., 120, 1494–1505. Tallarek, U., Bayer, E., Van Dusschoten, D., Scheenen, T., Van As, H., Guiochon, G., and Neue, U.D. (1998) AIChE J., 44, 1962–1975. Shalliker, R.A., Broyles, B.S., and Guiochon, G. (1998) J. Chromatogr. A., 826, 1–13. Shalliker, R.A., Broyles, B.S., and Guiochon, G. (2000) J. Chromatogr. A., 888, 1–12. Camenzuli, M., Ritchie, H.J., and Shalliker, R.A. (2013) Microchem. J. 110, 473–479. Camenzuli, M., Ritchie, H.J., Dennis, G.R., and Shalliker, R.A. (2013) Microchem. J., 110, 726–730. Camenzuli, M., Terry, J.M., Shalliker, R.A., Conlan, X.A., Barnett, N.W., and Francis, P.S. (2013) Ana. Chim. Acta, 803, 154–159. Kocic, D., Pereira, L., Foley, D., Edge, T., Mosely, J.A., Ritchie, H., Conlan, X.A., and

37

38

39

40

41

42

43

Shalliker, R.A. (2013) J. Chromatogr. A, 1305, 102–108. Foley, D., Pereira, L., Camenzuli, M., Edge, T., Ritchie, H., and Shalliker, R.A. (2013) Microchem. J., 110, 127–132. Catchpoole, H.J., Shalliker, R.A., Dennis, G.R., and Guiochon, G. (2006) J. Chromatogr. A, 1117, 137–145. Broyles, B.S., Shalliker, R.A., and Guiochon, G. (1999) J. Chromatogr. A., 855, 367–382. Shalliker, R.A., Broyles, B.S., and Guiochon, G. (1999) J. Chromatogr. A., 865, 83–95. Broyles, B.S., Shalliker, R.A., and Guiochon, G. (2001) J. Chromatogr. A., 917, 1–22. Gritti, F., Pynt, J., Soliven, A., Dennis, G.R., Shalliker, R.A., and Guiochon, G. (2014) J. Chromatogr. A., 1333, 32–44. Soliven, A., Foley, D., Pereira, L., Hua, S., Edge, T., Ritchie, H., Dennis, G.R., and Shalliker, R.A. (2014) Microchem. J., 116, 230–234.

63

4 Hydrophilic Interaction Liquid Chromatography Xinmiao Liang, Aijin Shen, and Zhimou Guo

4.1 Introduction

High-performance liquid chromatography (HPLC) has been recognized as a powerful analytical technique in many fields including pharmaceutical, environmental, food industries, clinical diagnosis, bioanalysis, and so on. Although reversed phase liquid chromatography (RPLC) has widely been used because of its high resolution and superior reproducibility, the separation of polar and hydrophilic compounds is challenging owing to the insufficient retention. Normal phase liquid chromatography (NPLC) and ion-exchange chromatography (IEX) can be utilized for resolving polar analytes. However, NPLC suffers from poor reproducibility and low solubility of hydrophilic analytes, whereas IEX is merely applicable for ionic compounds. In contrast, hydrophilic interaction liquid chromatography (HILIC) [1] turns out to be a valuable complement for the resolution of polar compounds and has received increasing popularity during the past two decades [2–12]. The concept of HILIC was coined by Alpert in 1990 [1], although it was originally introduced in the 1970s [13,14] for the analysis of carbohydrates. The main characteristic of HILIC is the combination of polar stationary phase with aqueous-organic mobile phase (most frequently acetonitrile). Generally, a water content of higher than 2% is essential for the formation of water-enriched layer semi-immobilized on the surface of the stationary phase. The retention of polar analytes decreases with increasing polar solvent (water) in the mobile phase. Comparing with NPLC, the solubility of hydrophilic analytes is greatly improved, and the high organic component together with volatile buffer is suitable for combining with mass spectrometric (MS) analysis. Meanwhile, low viscosity of the mobile phase enables the efficient separation at lower pressure than RPLC. In addition, HILIC affords alternative separation selectivity to RPLC and IEX for polar compounds, thus enhancing the analytes’ coverage. Publications involving HILIC have increased rapidly since 2005 (as displayed in Figure 4.1), covering stationary phases and chromatographic method development, retention behavior and mechanism, separation efficiency and peak shape, Analytical Separation Science, First Edition. Edited by Jared L. Anderson, Alain Berthod, Verónica Pino Estévez, and Apryll M. Stalcup.  2015 Wiley-VCH Verlag GmbH & Co. KGaA. Published 2015 by Wiley-VCH Verlag GmbH & Co. KGaA.

64

4 Hydrophilic Interaction Liquid Chromatography

Figure 4.1 Number of publications indexed on Web of Science with terms “HILIC” or “hydrophilic interaction liquid chromatography” or “hydrophilic interaction chromatography.”

and numerous applications in diverse fields. Among them, a number of remarkable reviews have been contributed by different authors. In 2006, Hemström and Igrum provided a comprehensive summary of HILIC [7]. In 2008, Tanaka and others presented an excellent review regarding the separation efficiency in HILIC [9]. Heck and coworkers summarized the application of HILIC in proteomics, along with the analysis of protein post-translational modifications [15]. In 2011, Guo and Gaiki reviewed the retention and selectivity of HILIC stationary phases [16]. Based on previous publications and research works, in this chapter, we will provide a general description of the principles and research progresses of HILIC, including the separation mechanism, stationary phases, and practical applications.

4.2 Separation Mechanism in HILIC

The separation mechanism in HILIC is not so well understood as that of RPLC. It is difficult to predict the retention or selectivity by the functional groups on the HILIC stationary phase. As early as 1990, Alpert suggested the partitioning of analytes between the water-enriched layer on the surface of the stationary phase and the hydro-organic mobile phase as the potential mechanism [1]. The existence of a surface water-enriched layer has been studied and validated through chromatographic [17], spectrometric [18], and molecular dynamics simulation methods [19]. Nonetheless, numerous investigations [6,7,20–25] regarding the mechanism characterization or application in HILIC reflect the complexity of interaction mechanism. As illustrated in Figure 4.2, apart from

4.2 Separation Mechanism in HILIC

Figure 4.2 Schematic diagram of the interaction mechanism in HILIC.

hydrophilic partitioning, surface adsorption (such as hydrogen bonding and dipole–dipole interaction), as well as electrostatic interaction, may affect the retention of analytes. Moreover, the interaction mechanism exhibits significant differences among different stationary phases. In order to understand the retention mechanism in HILIC, several retention models have been employed to preliminarily determine the relative contribution of partitioning and surface adsorption [7]. The relationship established for characterizing the partitioning interaction is log k ´ ˆ log k ´w

Sϕ

(4.1)

where k ´w is the retention factor for the weaker eluent (organic solvent in HILIC) only as the mobile phase, ϕ is the volume fraction (concentration) of water in the mobile phase, and S is the slope of the linear regression model. On the other hand, the basic equation for describing the surface adsorptive interaction is log k ´ ˆ log k ´B

As log N B nB

(4.2)

65

66

4 Hydrophilic Interaction Liquid Chromatography

where k ´B is the retention factor with pure B (H2O in HILIC) as an eluent, As and nB are the cross-sectional areas occupied by the solute molecule on the surface and the B molecules, respectively. NB is the mole fraction of water in the mobile phase. The relationship between log k´ and linear or logarithmical function of water content in the mobile phase is supposed to indicate the predominant retention mechanism in HILIC. However, sometimes the linearity of either log-linear plots or log–log plots is not satisfactory, indicating the absence of pure partitioning or adsorption mechanism. Thus, the third model was proposed for characterizing solute–solvent–stationary-phase interactions [26]. ln k ´ ˆ a ‡ b ln φ ‡ cφ

(4.3)

where a is a constant relating to the molecular volume of solutes, the interaction energy between solutes with the stationary phase and the mobile phase; b is the coefficient involving the direct interaction between analyte and stationary phase; and c relates to the interaction energy between solutes and solvents. The retention model based on Equation 4.3 was found to be suitable for describing the retention factors [22,26], revealing the diversity of intermolecular interactions in HILIC. In addition, the existence of multiple-interaction mechanism has been demonstrated using the retention model based on a linear solvation energy relationship (LESR) approach [25,27]. log k ´ ˆ c ‡ eE ‡ sS ‡ aA ‡ bB ‡ vV ‡ d D ‡ d‡ D‡

(4.4)

The capital factors represent the particular interaction properties of the analytes (solute descriptors), whereas lower case letters are the system constants related to the complementary effect of the stationary phase. c is a constant independent of the analytes and is dominated by the phase ratio and specific column parameters, such as porosity, as well as other properties. E is the excess molar refraction and model polarizability contributions from n and π electrons. S represents the solute dipolarity/polarizability. A and B refer to the overall hydrogen bond acidity (H donor) and basicity (H acceptor). V is the McGowan characteristic volume. D reflects the negative charge carried by anionic and zwitterionic species, and D+ reflects the positive charge carried by cationic and zwitterionic species. The differential contribution of partitioning, hydrogen bonding, and electrostatic attractive/repulsive interaction among different HILIC columns represents the predominant mechanistic interaction. In general, the relative contribution of partitioning and surface adsorption mechanisms in HILIC depends on the comprehensive effect of the type of stationary phase, the composition of the mobile phase, and the structures of the analytes (as illustrated by the “LC retention Troika” in Figure 4.3). The contribution of multiple intermolecular interactions is considered to be favorable for regulating the separation selectivity in HILIC.

4.3 Stationary Phases for HILIC

Figure 4.3 LC retention “Troika.” (Reproduced with permission from Ref. [25].)

4.3 Stationary Phases for HILIC

In contrast to RPLC, stationary phases in HILIC are of polar characteristics (Table 4.1). The most frequently utilized chromatographic support in HILIC is silica gel (SiO2), which can easily be modified and is of high mechanical strength. 4.3.1 Conventional NPLC Stationary Phases for HILIC

Stationary phases for conventional NPLC can be employed in HILIC, such as underivatized silica and amino-/cyano-/diol-modified silica. Bare silica possesses superior stability at low pH and is not subject to the bleeding of bonded phases from the column. It is favorable to combine with MS for selective and quantitative analysis. Thus, bare silica has appeared in a wide range of applications, especially in pharmaceutical analysis. However, basic compounds may irreversibly absorb onto silanol groups owing to the strongly electrostatic interaction. Besides, the long-term stability of underivatized silica under neutral or basic conditions is poor. To improve the stability, ethylene-bridged hybrid silica (BEH HILIC) was introduced. However, the hydrophilicity and selectivity of either conventional underivatized silica or BEH HILIC are limited. Amino-modified silica was initially applied for the separation of carbohydrates before the introduction of the term “HILIC.” In 1975, Linden and Palmer reported the

67

68

Representative stationary phases for HILIC.

Types of polar groups

Representative column

Bonded phase

Company/ ref.

Silica

Atlantics HILIC/BEH HILIC

Underivatized silica

Waters

Amino

Luna NH2/TSKgel NH2

Silica

NH2

Phenomenex/ Tosoh

Cyano

YMC-Pack Cyano

Silica

CN

YMC

Diol

Inertsil diol

Silica

O

Silica

Luna HILIC

OH

GL Science

OH O

O OH O OH

Phenomenex

OH

4 Hydrophilic Interaction Liquid Chromatography

Table 4.1

Amide

TSKgel amide-80

H2N

O

H

m

R4 H n R3

R1

Tosoh

R2

Silica

Silica

Urea-based

Silica

BEH amide

O

NH2

linker

O Si O

Waters

O

O O O

Si OH

N H

NH2

[28]

O O

Si OH

N NH HO

Si OH

N H

[29]

O HO

N H

n

Sorbitol-based

Silica

Saccharides

CH O O C C CH3 O

OH [30]

OH

HO HO

69

OH

4.3 Stationary Phases for HILIC

Silica

USP-HILIC

O

(continued)

Bonded phase

Click Glucose

Silica

Click Maltose

Company/ ref.

O

O Si O OH

O

Si O OH

HO

N N N

OH OH

O

[31]

OH

O

HO

N N N

OH OOH

O

OH

O

OH OH

[31,32]

OH

Silica

Click β-CD

O

Si O OH

N N N

H N

(OH)6 [31]

(OH)14

Silica

Galactose-based

O

Si O OH

HO O

N OH N N

O HO

OH OH

[33]

4 Hydrophilic Interaction Liquid Chromatography

Representative column

Silica

Types of polar groups

70

Table 4.1 (Continued)

Silica

Lactose-based

HO

O

Si O OH

O

N OH N N

O

OH HO O O

O HO

HO

OH

NH2

NH

O Si O OH

[34]

ica

O O O

Si OH

N H

O

CF6

O

H N

O

Si O OH

N H

[35]

N H

O

CF6

[35]

O

O

Silica

Sulfonated CF6 stationary phase

O O

Si OH

N H

O

CF6

SO3

[36] (continued)

4.3 Stationary Phases for HILIC

Silica

Dicarbamoxyl-hexyl CF6 stationary phase

Silica

Propyl carbamate cyclofructan 6 (CF6)

NH2

NH

OH N N N n

Sil

OH

OH

O OH O

O O

O O

Click Chitooligosaccharides

HO

[33]

OH

OH

OH

OH

71

Table 4.1 (Continued) Representative column

Bonded phase

Company/ ref.

Poly(succinimide) silica [PolyGlycoplex A]

O

Silica

Polyaspartamide

N

N N H

H N

O Poly LC

O

O

n

O OH

HN O H N

O

Poly(2-hydroxyethyl aspartamide) silica [PolyHydroxyethyl A]

Silica

N H

H N

O N H O n

O

Poly LC

HN O OH SO3-

HN O

Silica

Poly(2-sulfoethyl aspartamide) silica [PolySulphoethyl A]

O H N N H

H N

O N H O n

O HN

O

SO3-

Poly LC

4 Hydrophilic Interaction Liquid Chromatography

O

72

Types of polar groups

HO O H N

O

Silica

poly(aspartic acid) silica [PolyCAT A]

N H

H N

O N H O n

O

Poly LC

OH O

ZIC–HILIC

Silica

Zwitterionic groups

N

Merck

SO3

Silica

3-P,P-diphenylphosphoniumpropylsulfonate

Silica

Click lysine

Silica

O O P O O

O

Si O OH

Merck

N

O O

Si O OH

[37]

SO3

P

N H

N H

N N N

COO

[38]

NH3 (continued)

4.3 Stationary Phases for HILIC

ZIC–cHILIC

73

74

Types of polar groups

Representative column

Bonded phase

Company/ ref.

OOC NH3+

-OOC

S

-OOC

NH3+

NH3+ S

S

[39,40]

Click TE-Cys

Si O

Si

O

O

O

Si O

O

Si O

O

O

Silica Gel HO O HN

O

NH

NH2

S O

OH

Click TE–GSH

O O

Si O

O

Si O

Si O

Silica Gel

O

O

[41]

4 Hydrophilic Interaction Liquid Chromatography

Table 4.1 (Continued)

4.3 Stationary Phases for HILIC

separation and quantitation of saccharides on amino-modified material using acetonitrile and water as eluents [13,14]. Because of their good retention and special selectivity, amino-based materials still play an important role in the resolution of carbohydrates. Meanwhile, the existence of amino on the stationary phase can efficiently eliminate the splitting or broadening of peak shape due to α/β sugar anomerization. But, amino-modified silica suffers from the formation of Schiff base with reducing sugars, affecting the accuracy of quantification and altering the chemical characteristics of the bonded phase. Besides, aminomodified silica is not stable over a long period. Cyano-based silica displays poor retention for hydrophilic analytes owing to its low polarity. Diol-based silica can provide additional hydrogen bonding and possesses higher polarity than cyanobased silica. Nonetheless, the hydrophilicity of both materials is limited. 4.3.2 Stationary Phases Developed for HILIC

After the introduction of HILIC, many stationary phases (academically and commercially) with diverse polar functional groups have been developed for improving the hydrophilicity and separation selectivity in HILIC. Based on the inherent characteristics of bonded functional groups, HILIC stationary phases can be divided into several categories, including polyaspartamide-, amide-, saccharides-, and zwitterionic-based materials. 4.3.2.1

Polyaspartamide-Based Stationary Phases

Polyaspartamide-based stationary phases were the first chemically modified materials designed for HILIC by Alpert [1]. Poly(succinimide) was bonded onto aminopropyl silica to obtain the poly(succinimide) silica. Then, subsequent alkaline hydrolysis or modification with taurine or ethanolamine was carried out to form the corresponding poly(aspartic acid), poly(2-sulfoethyl aspartamide), and poly(2-hydroxyethyl aspartamide) stationary phases [1,42,43]. With the combination of hydrophilicity and unique electrostatic interaction, poly(2-sulfoethyl aspartamide) silica revealed excellent selectivity for the separation of peptides and proteins [9,15]. But, polyaspartamide-based materials present limited longtime stability and low efficiency [9]. 4.3.2.2

Amide-Based Stationary Phases

Different from amino-modified silica, amide-based stationary phase is not susceptible to Schiff-base formation. Besides, the stability of amide-based silica is greatly improved compared to amino-modified silica. Among the commercially available amide-bonded phases, TSK Amide-80 column [4,44,45] proves to be a good choice because of its high hydrophilicity and separation efficiency. In addition, urea or bidentate-bonded urea-modified silica [28,29] has been explored for HILIC.

75

76

4 Hydrophilic Interaction Liquid Chromatography

4.3.2.3

Saccharides-Based Stationary Phases

Saccharides, which are composed of abundant hydroxyl groups, are highly hydrophilic. The polyhydroxyl groups are beneficial for the formation of waterenriched layer and the enhancement of secondary interaction with polar analytes, such as hydrogen bonding. Thus, saccharides represent an interesting alternative as bonded phases for HILIC. In 2007, Liang and others designed and synthesized several saccharides-based HILIC materials through alkyne–azide click chemistry. Monosaccharide (Click Glucose), disaccharide (Click Maltose), and oligosaccharides (Click β-CD, Click Chitooligosaccharides) were bonded onto silica [31,32,34]. Click Maltose material exhibited good selectivity in the enrichment of glycopeptides [46]. The analysis and preparation of oligosaccharides and peptides were successfully realized on the Click Maltose column [47,48]. In 2008, a sorbitol-based material was developed by graft polymerization from the surface of silica [30]. The sorbitol-based silica exhibited better hydrophilicity than underivatized silica. Lactose- and galactose-modified silica were synthesized and applied in the resolution of sugar anomers [33]. Several cyclofructan 6 (CF6)-based stationary phases were also developed by Armstrong and others for the separation of small polar analytes [35,36]. 4.3.2.4

Zwitterionic Stationary Phases

Zwitterionic material is characterized by the existence of both positively and negatively charged groups on the bonded phase. The incorporation of high hydrophilicity and unique ion-exchange interaction greatly improve the separation selectivity in HILIC. Consequently, the application of zwitterionic material in HILIC has received increasing popularity, especially in proteomics and metabonomics. The typical zwitterionic stationary phases include sulfobetaine (ZIC–HILIC)- and phosphocholine (ZIC–cHILIC)-modified silica, of which the oppositely charged groups are distributed perpendicular to the silica surface. In 2011, Liang and others designed a novel type of zwitterionic stationary phase with a uniform distribution of both positive and negative charges that are parallel to the surface of the silica gel [39,40]. The zwitterionic material was facilely synthesized by the modification of vinyl silica with cysteine through thiol-ene click chemistry (Click TE-Cys). Click TE-Cys displayed high separation efficiency and better hydrophilicity than many commercial HILIC stationary phases (Figure 4.4). In addition, glutathione was immobilized onto vinyl silica to obtain a hydrophilic interaction/cation-exchange (HILIC/CEX) mixed mode zwitterionic material [41]. In 2011, Armstrong and others developed a zwitterionic stationary phase based on 3-P,P-diphenylphosphonium-propylsulfonate [37]. The resulting material presented a higher efficiency and greater retention than ZIC–HILIC and a bare silica column. Besides, lysine-based zwitterionic material [38] was also prepared and applied in the separation of cephalosporins and carbapenems. With expanding application of HILIC, the development of chromatographic columns with higher efficiency, better hydrophilicity and selectivity is of great significance. One way of increasing the separation efficiency is to minimize the particle size and inner diameter of the column (capillary column). Stationary

4.4 Application of HILIC

10

Click maltose Click CD ZIC-HILIC Atlantis HILIC silica TSKgel Amide-80 Tigerkin Diol Venusil HILIC Click TE-Cys

9 8 7

k

6 5 4 3 2 1 0

Uracil

Uridine

Cytosine

Cytidine

Orotic acid

Figure 4.4 Comparison of capacity factors on Click TE-Cys and seven different HILIC columns. Mobile phase, 15 mM HCOONH4 in ACN/H2O (85 : 15), pH 3.28. (Reproduced with permission from Ref. [39].)

phases with particle size less than 3 μm (Kinetex HILIC core–shell column) or even 1.7 μm (BEH HILIC) have been successfully developed. Hydrophilic monolithic columns represent a novel trend among HILIC column techniques because of their good permeability, low resistance to mass transfer, and easy preparation within capillaries [9,49–52]. On the other hand, silica has been replaced by metal-oxide such as titania or zirconia as packing material to improve the pH stability and introduce alternative selectivity [11,12]. Nonetheless, the selection of appropriate column for separating particular compounds becomes more and more complicated with the increasing diversity of HILIC stationary phases.

4.4 Application of HILIC 4.4.1 Application in the Pharmaceutical Field

As the classical chromatographic technique, RPLC is the predominant approach for qualitative and/or quantitative analysis of pharmaceutical components. However, RPLC cannot provide sufficient retention and resolution with regard to polar drugs and their metabolites. HILIC has been proven to be a valuable alternative to biological and nonbiological pharmaceutical assays [53–58], such as the determination of active ingredients in pharmaceutical formulations, the intermediates or impurities involving in drug development, and the quantification of drug compounds and/or their metabolites. The application of HILIC for the

77

78

4 Hydrophilic Interaction Liquid Chromatography

analysis of polar pharmaceuticals in complex biological samples (e.g., plasma, serum, or urine) can greatly increase the retention factors and separation selectivity, which is conducive to minimize matrix interference and ion suppression. Meanwhile, the high organic content in the mobile phase shows better compatibility with LC–MS interface, thus improving the detection sensitivity. On the other hand, the commonly utilized sample cleanup procedures, such as protein precipitation (PP), liquid–liquid extraction (LLE), or solid-phase extraction, always generate extracts of high organic content. The extracts can be directly injected into HILIC columns without evaporation and reconstitution steps. Bare silica (Atlantis HILIC, YMC silica, etc.) is the most widely used stationary phase in pharmaceutical applications [55,57]. In contrast, a limited number of pharmaceutical analysis was performed using chemically modified stationary phases that are subject to the bleeding of bonded ligands. 4.4.2 Application in the Separation of Carbohydrates

As a large group of organic polar compounds, carbohydrates possess diverse biological functions and play a significant role in the biochemical system. The separation and identification of carbohydrates are gaining more and more attention. Derivatization of carbohydrates with hydrophobic chromophore is always required before the separation by conventional RPLC owing to the insufficient retention. High-performance anion-exchange chromatography (HPAEC), in combination with pulsed amperometric detection (PAD), has widely been used for analyzing carbohydrates. However, the mobile phase used in HPAEC is incompatible with mass spectrometry (MS). In comparison, HILIC provides adequate retention for carbohydrates and has become an efficient technique for the separation of complex carbohydrates, quantification, or structure identification [2,10,43,47,59]. The resolution of small carbohydrates, native glycan cleaved from native bovine fetuin through hydrazinolysis, and derivatized complex carbohydrates was successfully realized on the PolyGLYCOPLEX column with comparable selectivity to HPAEC and RPLC [43]. In 2009, Wuhrer et al. comprehensively reviewed the application of HILIC–MS in structural glycomics at both glycan and glycopeptides levels (Figure 4.5), indicating that HILIC combined with MS is an attractive and powerful approach for the analysis of glycans and glycoconjugates [10]. Furthermore, the resolving ability of HILIC for oligosaccharides with a high degree of polymerization has been greatly improved with increasing hydrophilicity and selectivity of HILIC stationary phase [41]. 4.4.3 Application in Proteome, Glycoproteome, and Phosphoproteome

RPLC is an indispensible technique for separating and purifying peptide mixtures in proteomics. However, the separation selectivity in the analysis of complex proteomic samples based on one-dimensional RPLC separation is limited.

4.4 Application of HILIC

Figure 4.5 Scheme of HILIC–MS approaches in structural glycomics. (Reproduced with permission from Ref. [10].)

Therefore, the incorporation of online/offline multidimensional separation approaches has received increasing popularity and widespread application. Strong cation exchange chromatography (SCX), which displays different retention mechanism from RPLC, is commonly utilized for peptide separation. But peptides with the same charge always elute in narrow windows and the mobile phase is incompatible with MS. In contrast, HILIC is an attractive choice because of its distinct selectivity and superior orthogonality to RPLC. What is more, HILIC presents higher resolution capacity than SCX [60]. In 2007, Mohammed and others explored an alternative two-dimensional liquid chromatography (2D-LC), that is, ZIC–HILIC hyphenated offline with RPLC for analyzing the cellular nuclear lysate, revealing its great potential as multidimensional protein identification technology (MudPIT) in proteomics [61]. In 2011, 2D-LC systems based on ZIC–HILIC or ZIC–cHILIC in combination with RPLC were further applied for peptide identification from 1.5 μg of Hela lysate digestion (Figure 4.6). Approximately 20 000 unique peptides corresponding to over 3500 proteins were successfully identified, demonstrating that HILIC can represent a powerful foundation for a sensitive multidimensional strategy [62]. In addition, HILIC/CEX mixed mode chromatography has been proved to be an excellent complement to RPLC. The great potential of HILIC/CEX in the separation of peptides (e.g., α-helical peptides and modified or deletion products of synthetic peptides) and modified histone proteins was demonstrated by Hodges and Lindner [63]. Furthermore, the HILIC technique plays a significant role in the enrichment and analysis of protein post-translational modifications (e.g., glycosylation and phosphorylation) [8,10,45,64–66] that have significant effects on protein

79

80

4 Hydrophilic Interaction Liquid Chromatography

Figure 4.6 (a) Schematic design of the nano-online HILIC–MS systems (1) and firstdimensional separation configurations (2). After fractionation, the eluent was collected as 1-min fractions in a 96-well plate containing 10% formic acid and subsequently analyzed

by nano-RP–LC–MS. (b) Schematic workflow of different settings employed with the above offline configuration shown in part (a2). (Reproduced with permission from Ref. [61]. Copyright 2011 American Chemical Society.)

functions. In 2008, Heck and coworkers reviewed the versatility of HILIC both in the enrichment of phosphorylated, N-terminally blocked, and glycosylated peptides and in the separation of differentially modified histones [15]. 4.4.4 Application in Metabolomics/Metabonomics

Metabolomics/metabonomics, aiming at the unbiased analysis and quantification of metabolites in a biological system, can provide a comprehensive signature of the physiological state of an organism, as well as insights into specific biochemical processes [67]. NMR spectrometry and MS are the major analytical techniques for metabolite profiling, and MS exhibits better sensitivity. However, MS-based metabolomics is susceptible to matrix effects owing to the complexity of metabolites in biological samples. Consequently, the coupling of chromatographic separation to MS for reducing the interference is of great importance. RPLC–MS-based metabolomics is the most common approach for biofluids analysis. Nonetheless, RPLC is not applicable to polar components that are abundant in predominantly aqueous biofluids. The introduction of HILIC in MS metabolomics provides an alternative choice for resolving highly polar metabolites [67–73]. Besides, the MS detection selectivity is greatly improved because of the high organic content in the mobile phase. HILIC–MS-based metabolomics

References

was first applied in plant metabolite profiling in 2002 [4]. With the combination of HILIC (TSKgel Amide-80 columns) and electrospray quadrupole iontrap mass spectrometry, the identification and quantification of polar metabolites of plant origin were successfully realized. Both novel and general components, including oligosaccharides, glycosides, amino sugars, and sugar nucleotides, were detected from Cucurbita maxima phloem exudates. HILIC coupled to MS has been widely utilized for the analysis of cellular metabolites [70] (bacterial or microbial samples) such as water-soluble cellular metabolites extracted from Escherichia coli, sulfur endogenous metabolites related to glutathione biosynthesis from Saccharomyces cerevisiae cells, and so on. In addition, the metabolite profiling of urine samples in mammalian systems based on the HILIC–MS approach can provide complementary information involving polar metabolites to RPLC–MS-based metabolomics, thus increasing the coverage of metabolome [71].

4.5 Conclusions and Outlook

Since the introduction of HILIC in 1990, the versatility and capability of HILIC in the analysis of polar and hydrophilic compounds have widely been accepted. Applications concerning small compounds (nucleosides, nucleotides, etc.), pharmaceuticals, natural products, peptide mixtures, metabolites, and so on were successfully achieved. The basic principles and mechanisms of HILIC have been gradually understood. Numerous stationary phases with distinct selectivities have been developed, and particular materials, including TSK Amide-80, ZIC– HILIC, Click Maltose, and Click TE-Cys, exhibit excellent chromatographic properties. Nonetheless, the increasing complexity of samples such as in proteomics and metabolomics presents further challenges with regard to the separation selectivity and efficiency. Besides, the high organic content in the mobile phase is a disadvantage for the dissolution of polar analytes. Therefore, the development of novel HILIC stationary phase with superior hydrophilicity and resolution ability is significant. On the other hand, the elaborate illustration of retention mechanism is crucial for predicting the retention behaviors in HILIC.

References 1 Alpert, A.J. (1990) Hydrophilic-interaction

chromatography for the separation of peptides, nucleic-acids and other polar compounds. J. Chromatogr., 499, 177–196. 2 Churms, S.C. (1996) Recent progress in carbohydrate separation by highperformance liquid chromatography based

on hydrophilic interaction. J. Chromatogr. A, 720, 75–91. 3 Strege, M.A. (1998) Hydrophilic interaction chromatography electrospray mass spectrometry analysis of polar compounds for natural product drug discovery. Anal. Chem., 70, 2439–2445.

81

82

4 Hydrophilic Interaction Liquid Chromatography 4 Tolstikov, V.V. and Fiehn, O. (2002)

5

6

7

8

9

10

11

12

13

14

15

Analysis of highly polar compounds of plant origin: combination of hydrophilic interaction chromatography and electrospray ion trap mass spectrometry. Anal. Biochem., 301, 298–307. Yoshida, T. (2004) Peptide separation by hydrophilic-interaction chromatography: a review. J. Biochem. Biophys. Methods, 60, 265–280. Guo, Y. and Gaiki, S. (2005) Retention behavior of small polar compounds on polar stationary phases in hydrophilic interaction chromatography. J. Chromatogr. A, 1074, 71–80. Hemström, P. and Irgum, K. (2006) Hydrophilic interaction chromatography. J. Sep. Sci., 29, 1784–1821. Alpert, A.J. (2008) Electrostatic repulsion hydrophilic interaction chromatography for isocratic separation of charged solutes and selective isolation of phosphopeptides. Anal. Chem., 80, 62–76. Ikegami, T., Tomomatsu, K., Takubo, H., Horie, K., and Tanaka, N. (2008) Separation efficiencies in hydrophilic interaction chromatography. J. Chromatogr. A, 1184, 474–503. Wuhrer, M., de Boer, A.R., and Deelder, A.M. (2009) Structural glycomics using hydrophilic interaction chromatography (HILIC) with mass spectrometry. Mass Spectrom. Rev., 28, 192–206. Gama, M.R., Silva, R.G.D., Collins, C.H., and Bottoli, C.B.G. (2012) Hydrophilic interaction chromatography. Trend. Anal. Chem., 37, 48–60. Garcia-Gomez, D., Rodriguez-Gonzalo, E., and Carabias-Martinez, R. (2013) Stationary phases for separation of nucleosides and nucleotides by hydrophilic interaction liquid chromatography. Trend. Anal. Chem., 47, 111–128. Linden, J.C. and Lawhead, C.L. (1975) Liquid-chromatography of saccharides. J. Chromatogr., 105, 125–133. Palmer, J.K. (1975) Versatile system for sugar analysis via liquid-chromatography. Anal. Lett., 8, 215–224. Boersema, P.J., Mohammed, S., and Heck, A.J.R. (2008) Hydrophilic interaction liquid chromatography (HILIC) in

16

17

18

19

20

21

22

23

proteomics. Anal. Bioanal. Chem., 391, 151–159. Guo, Y. and Gaiki, S. (2011) Retention and selectivity of stationary phases for hydrophilic interaction chromatography. J. Chromatogr. A, 1218, 5920–5938. McCalley, D.V. and Neue, U.D. (2008) Estimation of the extent of the water-rich layer associated with the silica surface in hydrophilic interaction chromatography. J. Chromatogr. A, 1192, 225–229. Wikberg, E., Sparrman, T., Viklund, C., Jonsson, T., and Irgum, K. (2011) A (2)H nuclear magnetic resonance study of the state of water in neat silica and zwitterionic stationary phases and its influence on the chromatographic retention characteristics in hydrophilic interaction high-performance liquid chromatography. J. Chromatogr. A, 1218, 6630–6638. Melnikov, S.M., Holtzel, A., SeidelMorgenstern, A., and Tallarek, U. (2012) A molecular dynamics study on the partitioning mechanism in hydrophilic interaction chromatography. Angew. Chem. Int. Ed., 51, 6251–6254. McCalley, D.V. (2010) Study of the selectivity, retention mechanisms and performance of alternative silica-based stationary phases for separation of ionised solutes in hydrophilic interaction chromatography. J. Chromatogr. A, 1217, 3408–3417. Dinh, N.P., Jonsson, T., and Irgum, K. (2011) Probing the interaction mode in hydrophilic interaction chromatography. J. Chromatogr. A, 1218, 5880–5891. Karatapanis, A.E., Fiamegos, Y.C., and Stalikas, C.D. (2011) A revisit to the retention mechanism of hydrophilic interaction liquid chromatography using model organic compounds. J. Chromatogr. A, 1218, 2871–2879. Kawachi, Y., Ikegami, T., Takubo, H., Ikegami, Y., Miyamoto, M., and Tanaka, N. (2011) Chromatographic characterization of hydrophilic interaction liquid chromatography stationary phases: hydrophilicity, charge effects, structural selectivity, and separation efficiency. J. Chromatogr. A, 1218, 5903–5919.

References 24 Kumar, A., Heaton, J.C., and McCalley,

25

26

27

28

29

30

31

32

D.V. (2013) Practical investigation of the factors that affect the selectivity in hydrophilic interaction chromatography. J. Chromatogr. A, 1276, 33–46. Schuster, G. and Lindner, W. (2013) Comparative characterization of hydrophilic interaction liquid chromatography columns by linear solvation energy relationships. J. Chromatogr. A, 1273, 73–94. Jin, G., Guo, Z., Zhang, F., Xue, X., Jin, Y., and Liang, X. (2008) Study on the retention equation in hydrophilic interaction liquid chromatography. Talanta, 76, 522–527. Chirita, R.I., West, C., Zubrzycki, S., Finaru, A.L., and Elfakir, C. (2011) Investigations on the chromatographic behaviour of zwitterionic stationary phases used in hydrophilic interaction chromatography. J. Chromatogr. A, 1218, 5939–5963. Bicker, W., Wu, J., Yeman, H., Albert, K., and Lindner, W. (2011) Retention and selectivity effects caused by bonding of a polar urea-type ligand to silica: a study on mixed-mode retention mechanisms and the pivotal role of solute–silanol interactions in the hydrophilic interaction chromatography elution mode. J. Chromatogr. A, 1218, 882–895. Kotoni, D., D’Acquarica, I., Ciogli, A., Villani, C., Capitani, D., and Gasparrini, F. (2012) Design and evaluation of hydrolytically stable bidentate urea-type stationary phases for hydrophilic interaction chromatography. J. Chromatogr. A, 1232, 196–211. Persson, J., Hemström, P., and Irgum, K. (2008) Preparation of a sorbitol methacrylate grafted silica as stationary phase for hydrophilic interaction chromatography. J. Sep. Sci., 31, 1504–1510. Guo, Z., Lei, A., Zhang, Y., Xu, Q., Xue, X., Zhang, F., and Liang, X. (2007) “Click saccharides”: Novel separation materials for hydrophilic interaction liquid chromatography. Chem. Commun., 2491–2493. Fu, Q., Guo, Z.M., Liang, T., Zhang, X.L., Xu, Q., and Liang, X.M. (2010) Chemically

33

34

35

36

37

38

39

40

bonded maltose via click chemistry as stationary phase for HILIC. Anal. Methods, 2, 217–224. Moni, L., Ciogli, A., D’Acquarica, I., Dondoni, A., Gasparrini, F., and Marra, A. (2010) Synthesis of sugar-based silica gels by copper-catalysed azide-alkyne cycloaddition via a single-step azidoactivated silica intermediate and the use of the gels in hydrophilic interaction chromatography. Chem. Eur. J., 16, 5712–5722. Huang, H., Jin, Y., Xue, M., Yu, L., Fu, Q., Ke, Y., Chu, C., and Liang, X. (2009) A novel click chitooligosaccharide for hydrophilic interaction liquid chromatography. Chem. Commun., 6973–6975. Qiu, H., Loukotkova, L., Sun, P., Tesarova, E., Bosakova, Z., and Armstrong, D.W. (2011) Cyclofructan 6 based stationary phases for hydrophilic interaction liquid chromatography. J. Chromatogr. A, 1218, 270–279. Padivitage, N.L.T. and Armstrong, D.W. (2011) Sulfonated cyclofructan 6 based stationary phase for hydrophilic interaction chromatography. J. Sep. Sci., 34, 1636–1647. Qiu, H., Wanigasekara, E., Zhang, Y., Tran, T., and Armstrong, D.W. (2011) Development and evaluation of new zwitterionic hydrophilic interaction liquid chromatography stationary phases based on 3-P,P-diphenylphosphoniumpropylsulfonate. J. Chromatogr. A, 1218, 8075–8082. Guo, H.Y., Liu, R.H., Yang, J.J., Yang, B.C., Liang, X.M., and Chu, C.H. (2012) A novel click lysine zwitterionic stationary phase for hydrophilic interaction liquid chromatography. J. Chromatogr. A, 1223, 47–52. Shen, A.J., Guo, Z.M., Yu, L., Cao, L.W., and Liang, X.M. (2011) A novel zwitterionic HILIC stationary phase based on “thiol-ene” click chemistry between cysteine and vinyl silica. Chem. Commun., 47, 4550–4552. Shen, A., Guo, Z., Cai, X., Xue, X., and Liang, X. (2012) Preparation and chromatographic evaluation of a cysteinebonded zwitterionic hydrophilic

83

84

4 Hydrophilic Interaction Liquid Chromatography

41

42

43

44

45

46

47

48

49

interaction liquid chromatography stationary phase. J. Chromatogr. A, 1228, 175–182. Shen, A.J., Li, X.L., Dong, X.F., Wei, J., Guo, Z.M., and Liang, X.M. (2013) Glutathione-based zwitterionic stationary phase for hydrophilic interaction/cationexchange mixed-mode chromatography. J. Chromatogr. A, 1314, 63–69. Zhu, B.Y., Mant, C.T., and Hodges, R.S. (1991) Hydrophilic-interaction chromatography of peptides on hydrophilic and strong cation-exchange columns. J. Chromatogr., 548, 13–24. Alpert, A.J., Shukla, M., Shukla, A.K., Zieske, L.R., Yuen, S.W., Ferguson, M.A.J., Mehlert, A., Pauly, M., and Orlando, R. (1994) Hydrophilic-interaction chromatography of complex carbohydrates. J. Chromatogr. A, 676, 191–202. Yoshida, T. (1997) Peptide separation in normal phase liquid chromatography. Anal. Chem., 69, 3038–3043. McNulty, D.E. and Annan, R.S. (2008) Hydrophilic interaction chromatography reduces the complexity of the phosphoproteome and improves global phosphopeptide isolation and detection. Mol. Cell. Proteomics, 7, 971–980. Yu, L., Li, X., Guo, Z., Zhang, X., and Liang, X. (2009) Hydrophilic interaction chromatography based enrichment of glycopeptides by using click maltose: a matrix with high selectivity and glycosylation heterogeneity coverage. Chem. Eur. J., 15, 12618–12626. Fu, Q., Liang, T., Zhang, X., Du, Y., Guo, Z., and Liang, X. (2010) Carbohydrate separation by hydrophilic interaction liquid chromatography on a ‘click’ maltose column. Carbohydr. Res., 345, 2690–2697. Xu, J., Zhang, X., Guo, Z., Yan, J., Yu, L., Li, X., Xue, X., and Liang, X. (2012) Short-chain peptides identification of scorpion Buthus martensi Karsch venom by employing high orthogonal 2D-HPLC system and tandem mass spectrometry. Proteomics, 12, 3076–3084. Horie, K., Ikegami, T., Hosoya, K., Saad, N., Fiehn, O., and Tanaka, N. (2007) Highly efficient monolithic silica capillary

50

51

52

53

54

55

56

57

58

59

columns modified with poly(acrylic acid) for hydrophilic interaction chromatography. J. Chromatogr. A, 1164, 198–205. Jiang, Z.J., Smith, N.W., Ferguson, P.D., and Taylor, M.R. (2007) Hydrophilic interaction chromatography using methacrylate-based monolithic capillary column for the separation of polar analytes. Anal. Chem., 79, 1243–1250. Nunez, O., Nakanishi, K., and Tanaka, N. (2008) Preparation of monolithic silica columns for high-performance liquid chromatography. J. Chromatogr. A, 1191, 231–252. Jiang, Z., Smith, N.W., and Liu, Z. (2011) Preparation and application of hydrophilic monolithic columns. J. Chromatogr. A, 1218, 2350–2361. Strege, M.A., Stevenson, S., and Lawrence, S.M. (2000) Mixed mode anion–cation exchange/hydrophilic interaction liquid chromatography–electrospray mass spectrometry as an alternative to reversed phase for small molecule drug discovery. Anal. Chem., 72, 4629–4633. Olsen, B.A. (2001) Hydrophilic interaction chromatography using amino and silica columns for the determination of polar pharmaceuticals and impurities. J. Chromatogr. A, 913, 113–122. Weng, N.D. (2003) Bioanalytical liquid chromatography tandem mass spectrometry methods on underivatized silica columns with aqueous/organic mobile phases. J. Chromatogr. B, 796, 209–224. Hsieh, Y. (2008) Potential of HILIC–MS in quantitative bioanalysis of drugs and drug metabolites. J. Sep. Sci., 31, 1481–1491. Dejaegher, B. and Heyden, Y.V. (2010) HILIC methods in pharmaceutical analysis. J. Sep. Sci., 33, 698–715. Jian, W., Edom, R.W., Xu, Y., and Weng, N. (2010) Recent advances in application of hydrophilic interaction chromatography for quantitative bioanalysis. J. Sep. Sci., 33, 681–697. Takegawa, Y., Deguchi, K., Keira, T., Ito, H., Nakagawa, H., and Nishimura, S. (2006) Separation of isomeric 2aminopyridine derivatized N-glycans and N-glycopeptides of human serum

References

60

61

62

63

64

65

66

immunoglobulin G by using a zwitterionic type of hydrophilic-interaction chromatography. J. Chromatogr. A, 1113, 177–181. Gilar, M., Olivova, P., Daly, A.E., and Gebler, J.C. (2005) Orthogonality of separation in two-dimensional liquid chromatography. Anal. Chem., 77, 6426–6434. Boersema, P.J., Divecha, N., Heck, A.J.R., and Mohammed, S. (2007) Evaluation and optimization of ZIC–HILIC–RP as an alternative MudPIT strategy. J. Proteome. Res., 6, 937–946. Di Palma, S., Boersema, P.J., Heck, A.J.R., and Mohammed, S. (2011) Zwitterionic hydrophilic interaction liquid chromatography (ZIC–HILIC and ZIC– cHILIC) provide high resolution separation and increase sensitivity in proteome analysis. Anal. Chem., 83, 3440–3447. Mant, C.T. and Hodges, R.S. (2008) Mixed-mode hydrophilic interaction/ cation-exchange chromatography (HILIC/ CEX) of peptides and proteins. J. Sep. Sci., 31, 2754–2773. Hagglund, P., Bunkenborg, J., Elortza, F., Jensen, O.N., and Roepstorff, P. (2004) A new strategy for identification of Nglycosylated proteins and unambiguous assignment of their glycosylation sites using HILIC enrichment and partial deglycosylation. J. Proteome Res., 3, 556–566. Wuhrer, M., Deelder, A.M., and Hokke, C.H. (2005) Protein glycosylation analysis by liquid chromatography–mass spectrometry. J. Chromatogr. B, 825, 124–133. Wuhrer, M., Koeleman, C.A.M., Hokke, C.H., and Deelder, A.M. (2005) Protein glycosylation analyzed by normal-phase nano-liquid chromatography–mass

67

68

69

70

71

72

73

spectrometry of glycopeptides. Anal. Chem., 77, 886–894. Idborg, H., Zamani, L., Edlund, P.O., Schuppe-Koistinen, I., and Jacobsson, S.P. (2005) Metabolic fingerprinting of rat urine by LC/MS Part 1. Analysis by hydrophilic interaction liquid chromatography–electrospray ionization mass spectrometry. J. Chromatogr. B, 828, 9–13. Bajad, S.U., Lu, W., Kimball, E.H., Yuan, J., Peterson, C., and Rabinowitz, J.D. (2006) Separation and quantitation of water soluble cellular metabolites by hydrophilic interaction chromatography–tandem mass spectrometry. J. Chromatogr. A, 1125, 76–88. Cubbon, S., Bradbury, T., Wilson, J., and Thomas-Oates, J. (2007) Hydrophilic interaction chromatography for mass spectrometric metabonomic studies of urine. Anal. Chem., 79, 8911–8918. Cubbon, S., Antonio, C., Wilson, J., and Thomas-Oates, J. (2010) Metabolomic applications of HILIC–LC–MS. Mass Spectrom. Rev., 29, 671–684. Spagou, K., Tsoukali, H., Raikos, N., Gika, H., Wilson, I.D., and Theodoridis, G. (2010) Hydrophilic interaction chromatography coupled to MS for metabonomic/metabolomic studies. J. Sep. Sci., 33, 716–727. Jian, W., Xu, Y., Edom, R.W., and Weng, N. (2011) Analysis of polar metabolites by hydrophilic interaction chromatography– MS/MS. Bioanalysis, 3, 899–912. Zhang, T., Creek, D.J., Barrett, M.P., Blackburn, G., and Watson, D.G. (2012) Evaluation of coupling reversed phase, aqueous normal phase, and hydrophilic interaction liquid chromatography with orbitrap mass spectrometry for metabolomic studies of human urine. Anal. Chem., 84, 1994–2001.

85

87

5 LC–MS Interfaces Pierangela Palma, Elisabetta Pierini, and Achille Cappiello

5.1 Introduction

Liquid chromatography (LC) coupled to mass spectrometry (MS) is today a wellestablished analytical technique (LC–MS) that, in the last few decades, has opened the door to many challenging applications. MS is undoubtedly the most powerful detector that can exploit the separation capability of an LC column, and although coupling these two techniques was a demanding operation that required many efforts, now it can depend on reliable and rugged instrumentation that allows the determination of a large number of thermally labile molecules with quite different chemical properties. From a historical point of view, the thermospray interface developed by Vestal and coworkers can be considered one of the first attempts of LC–MS techniques. It was based on a spray formation followed by solvent elimination using differential pumping stages [1]. More recently, the development of “soft” ionization techniques and the availability of efficient LC columns that operate at lower flow rates opened the way to a skyrocketing success of LC–MS. As a matter of fact, if we look carefully at the literature and the instruments available on the market, we can observe that most of the new benefits of LC–MS come from developments in each of the two coupled techniques, LC and MS, rather than from the interface itself. If we take into consideration the improvements in the chromatographic separations offered by ultra high-pressure liquid chromatography (UHPLC), solid-core packing materials, and nano-flow, chip technologies coupled to faster, more accurate, and sensitive MS analyzers, such as orbitrap, linear trap, and time-of-flight, we have an idea of the arsenal of today’s LC–MS. As opposed to GC–MS that relies on electron ionization (EI), LC–MS mostly utilizes atmospheric pressure ionization (API) techniques, a totally different approach that leads to radically different MS results. EI is a high-energy, gasphase ionization technique that generates odd electron ions (M+• ) and provides a reproducible fragmentation of molecules that can be recorded in electronic

Analytical Separation Science, First Edition. Edited by Jared L. Anderson, Alain Berthod, Verónica Pino Estévez, and Apryll M. Stalcup.  2015 Wiley-VCH Verlag GmbH & Co. KGaA. Published 2015 by Wiley-VCH Verlag GmbH & Co. KGaA.

88

5 LC–MS Interfaces

libraries for unparalleled identification purposes. Soft ionization techniques provide ions with low-energy transfer that typically results in a single-molecular ion (M + H)+ or (M H) , and several adduct ions. In-source fragmentation can be achieved through collision-induced dissociation (CID), but the quality and the reproducibility of the resulting mass spectra are not sufficient to create suitable electronic libraries. High-resolution or tandem (MS/MS) instruments are needed to compensate for the lack of mass spectral information. The major issues in coupling LC with MS include both mobile phase and sample restrictions. The mobile phase can have a variable composition and carries the sample to the MS influencing, in many cases, the ionization of analytes. The physicochemical properties of the analytes, on the other hand, may be very different, imposing a huge demand in terms of system requirements. All these difficulties have stimulated the development of different approaches to satisfy the demand arising from many application fields that can benefit from LC–MS [2]. In this chapter, attention will be focused on the ionization process. The most commonly used interfaces and ion sources, as well as several new approaches, will be discussed [2–4]. Developments and improvements in the widely used API sources technology, including atmospheric pressure photoionization (APPI) and atmospheric pressure laser ionization (APLI), will be discussed. In addition, non-API sources, such as direct EI and supersonic molecular beams (SMB) LC– MS, will be taken into account, due to the growing attention gained in the analysis of low molecular weight compounds, including nonpolar molecules that are difficult to ionize with the two mostly used API sources, namely, electrospray interface (ESI) and atmospheric pressure chemical ionization (APCI).

5.2 API Sources

API ionization techniques are the driving force in LC–MS and are at the basis of the enormous success of LC–MS. To date, no other approach can compete with API interfaces for the number of instruments, applications, and presentation at conferences. They are all “soft” ionization techniques, which are widely used for the analysis of a large number of compounds in a wide range of molecular weights and polarities. Among them, ESI ionization is the most widespread source, present in the majority of LC–MS systems, in many cases associated with APCI for the analysis of organic molecules in a wide range of polarities [4,5]. To a lesser extent, APPI and APLI sources are employed. The operational principle of API interfaces is the nebulization of a liquid phase into an atmospheric pressure ion source region, followed by the separation of ions from neutral molecules. Nebulization can be achieved by an electric field (ESI), which can be assisted pneumatically (ion spray), ultrasonically, or by heat (such as in APCI). Whatever ions are generated in the gas phase, they are sampled through an orifice that acts as a fixed restriction between the pressure region and the first sampling stage before

5.2 API Sources

ASMS 2013 ESI

APCI

APPI 2% 1%

APLI

OTHERS

0%

10%

87%

Figure 5.1 Orientative distribution of the different LC–MS interfaces in the abstract presented at the annual conference of the ASMS in 2013 (nonofficial data).

reaching the analyzer. Modern API instruments have become very much userfriendly, and this has contributed to their widespread diffusion. API platforms have reached a solid level of maturity in terms of reliability and robustness so much so that most of the research is addressed to new applications (proteomics, cancer research, glycomics, metabolomics, and many others) rather than to technological improvements. In particular, the advent of ESI has revolutionized biological and biomedical research, opening the door to a great number of new research lines. In the abstracts presented at the annual conference of the American Society for Mass Spectrometry (ASMS) in 2013, approximately 87% of the presentations contained the word “ESI” or “electrospray” in their title, nearly 13% contained either the word “APCI” or “APPI” or “APLI,” whereas only less than 1% of the presentations had the use of non-API sources (Figure 5.1, nonofficial data). 5.2.1 Electrospray Interface (ESI)

ESI was developed in the 1980s by John Fenn, who was awarded the Noble Prize for his studies on “soft desorption ionization methods for mass spectrometric analyses of biological macromolecules.” ESI rapidly became the LC–MS interface everybody was waiting for, opening the door to a wide variety of applications for high- to medium-polarity compounds in an extended range of molecular weights, due to its high sensitivity and versatility. ESI technology has a dominant role in the LC–MS market; in fact, nearly all the LC–MS instruments are equipped with an ESI interface, which can be coupled to an APCI interface for the analysis of less polar compounds. As it can be seen in Figure 5.2, modern instrument configurations, due to improvements in both LC and MS, allow very high detection specificity and sensitivity [6].

89

90

5 LC–MS Interfaces

Figure 5.2 High-throughput UHPLC–ESI–MS/ MS analysis of 125 pesticide residues. Column: Pinnacle DB Aqueous C18; 50 mm × 2.1 mm ID; particle size: 1.9 μm; pore size: 140 Å; temperature: 35 °C; mobile phase A: 10 mM

5.2.1.1

NH4OAc in water; B: 10 mM NH4OAc in methanol. Gradient: from 10 to 90% B in 10 min. Flow rate: 0.60 ml/min. (Reproduced with permission).

Principles of Operation and Ion Formation

As illustrated in Figure 5.3, ESI ions and molecules, already present in bulk, are carried by a mobile phase inside the interface and converted into ions in the gas phase at atmospheric pressure by the vaporization of the charged droplets of the solution [7–9]. Heat may be applied to compensate for the heat of vaporization of the solvent. When the mobile phase passes through metallic capillary tubing, a strong electrical field, of the order of 106 V/m, causes the formation of charged droplets. Under the effect of the voltage applied, the ions tend to move toward the surface of the liquid, and at a proper voltage, the solution forms a typical Taylor cone, from which a spray plume of charged droplets breaks out. A coaxial low gas flow is also used to keep the droplets in a limited space. As the droplets’ size decreases, the ratio of surface charge to surface area increases. When the charge repulsion overcomes the surface tension, the Rayleigh stability limit is reached and, at that point, a cascade fission process begins [10,11]. A heated curtain gas or a heated capillary can support the evaporation of the solvent. Several theories have been proposed to describe the ion formation in ESI [12–14] that depends on many different mechanisms that may occur in the bulk solution, or during the formation of the charged droplets, or in the gas phase by protonation or deprotonation or salt adduct formation, or by an electrochemical redox reaction. Molecules with more than one protonation or deprotonation site can form multiple charged ions. It is a peculiarity of ESI to form ions such as [M + nH]n+ and [M nH]n . In this case, depending on the molecular mass, a

5.2 API Sources

Figure 5.3 Schematic of the ESI nebulization process (reproduced with permission).

number of molecular ions are present in the spectrum at different m/z values. With appropriate software, it is possible to resolve all these signals in a deconvoluted mass spectrum to calculate the molecular mass of the analyte [5,7]. The possibility of forming multiple charged ions allows the detection of high molecular weight molecules with mass analyzers having a limited m/z scan range. The behavior of high molecular weight compounds, in particular protein and peptides, as well as many others, to form multiple charged ions has made this ionization method essential in biomedical investigations. Adducts with various ions (Na+, K+, NH4+, Cl , and acetate) can be formed at different stages of the ESI process [15]. The adduct formation can be stimulated to promote the ionization of weakly basic or polar analytes, adding proper salts bearing the desired cations. Other adducts, such as noncovalently bound ones, can involve enzyme and substrate, protein and ligand, protein and protein, and antigen and antibody, can be studied using ESI–MS with the advantage of requiring a low amount of sample and a short time to provide structural information [14,17]. ESI is a soft ionization technique, which means that the ion fragmentation is limited; therefore, the mass spectrum provides information primarily on the molecular ion. As a consequence, tandem mass spectrometry (MS/MS) is mandatory to obtain the identification and structural elucidation of the analytes. ESI–MSn is a basic technique in the identification and sequencing of high molecular weight molecules, such as proteins, by using bottom-up and/or topdown strategies.

91

92

5 LC–MS Interfaces

5.2.1.2

Factors Influencing ESI Response

The response in ESI–MS strongly depends on the physico-chemical properties of the analytes; however, the signal can also be influenced by the presence of coeluted compounds or by changes in the mobile phase composition [11,18–21]. This phenomenon, peculiar to API, is commonly described as matrix effects (ME), and it is responsible for either the suppression or the enhancement of the signal. The many factors that contribute to ME in API have been described elsewhere, and they can occur either in the solution or in the gas phase, or both [22]. The ESI interface is more prone to ME among API interfaces because ionization reactions occur mainly in the liquid phase, wherein the concentration of interfering compounds is higher. ME greatly affects not only the ionization but also the vaporization process. The consequence of ME is an unreliable quantitation of analytes, which can lead to erroneous results. While making the method setup, it is wise to prevent ME, for example, by using sample cleanup procedures to get rid of the matrix components. Enhancing the chromatography is also a good way to separate coeluted compounds from those of interest. Another approach to compensating for ME is dilution or the use of costly isotope-labeled internal standards (ISs), whenever available [23–26]. 5.2.1.3

Modes of Operation

Depending on the nature of analytes, the ESI analysis can be performed either in a positive or in a negative ion mode. Although modern instruments allow switching modes during a single run, this operation can end up in spray instability; therefore, if possible, it is always preferable to run separate analyses of the same sample. The typical mobile phase in ESI is primarily composed of polar solvents, such as water, methanol, and, to a lesser extent, acetonitrile with or without the addition of volatile buffers. Pure water as a mobile phase might decrease sensitivity because of its low vapor pressure. When adding a modifier, for example, for pH adjustment or ion pairing, it is mandatory to avoid the introduction of nonvolatile salts or compounds that can induce signal alteration [24]. The ESI technology can work at nano-flow and high flow rates, with the purpose of optimizing ionization efficiency, minimizing ME, and improving sensitivity, in both cases. In Table 5.1, ion sources working at different flow rates are Table 5.1 Different ESI sources. Source

Flow rate (μl/min)

Aerosol generation

Nano-ESI

0.01–0.1

Electrostatic

Micro-ESI

0.1–4

Electrostatic

ESI

1–10

Electrostatic

Ionspray

10–500

Pneumatic/electrostatic

Turboionspray

500–1000

Pneumatic/electrostatic

5.2 API Sources

summarized. The capability of ESI to operate in a very wide range of liquid intake permits easy hyphenation with LC and capillary electrophoresis (CE). Low flow rates (1–1000 nl/min) have the advantage to achieve higher sensitivity because the ionization efficiency is optimal. Nano-ESI is particularly useful in biochemistry and protein analysis where very small amounts of sample are available. Smaller inner diameter tips give a more efficient generation of gas phase ions and allow placement of the tip closer to the sampling orifice resulting in optimal sampling of ions [5–27]. However, nowadays the sensitivity has approached its theoretical limit with nano-ESI systems. As a consequence, future efforts are mainly aimed at enhancing the high-flow ESI technique, which has an ionization efficiency lower than the theoretical limit considering that most commercial LC instruments operate at high flow rates (1–1000 μl/min). At high flow rates, it is possible to inject large volumes, compensating for the minor absolute sensitivity, thus lowering the detection limits. At high flow rates, nebulization is often assisted pneumatically (ion spray), by heat (turboionspray), or ultrasonically with the aim of enhancing the efficiency and tolerance to mobile phase composition, reducing ME. In the routine analysis of biological samples, source contamination from salts and nonvolatile compounds can occur. In this case, off-axis ESI geometries (with respect to the heated capillary) perform better than on-axis geometries. Some splitting devices have been developed with the aim of reducing the MS input flow rates. Concentric nanosplitter can be considered an interesting example, which adopts a second tube, with a smaller inner diameter, coaxial to that exiting from the HPLC column developed by Vouros and coworkers. Part of the LC effluent enters into this smaller tube, whereas the rest of the mobile phase is split into a waste or directed to a second detector or to a fraction collector. This splitter device reduces the flow rates entering into the analyzer from 200–400 μl/ min to 200 nl/min, avoiding turbulences and peak broadening that occur with splitting devices with T or Y configurations [28,29]. As a result, a significant increase in sensitivity is achieved, regardless of the removal of more than 99% of the sample. 5.2.2 Atmospheric Pressure Chemical Ionization

Chemical ionization in MS was first used in the 1960s by Munson and Field in a typical electron ionization source [30]. A controlled amount of reagent gas introduced into the ion source is ionized in the classical EI mode. The gas ions transfer their charge to the analytes’ molecules via an acid–base reaction, generating an intense molecular ion and a limited fragmentation. This approach is helpful in the determination of compounds that tend to give a very weak molecular ion in EI. The first attempt of an atmospheric pressure chemical ionization source used a 63 Ni foil or, alternatively, a corona discharge needle as a primary source of electrons [31,32]. However, in those days, a suitable LC–MS instrumentation capable of accommodating an APCI source was not fully developed yet. It was

93

94

5 LC–MS Interfaces

not before the end of the 1980 that this valuable approach could become commercially available, and today the APCI interface is essential in the analysis of low-polarity compounds for its sensitivity and robustness [5,33–35]. 5.2.2.1

Principles of Operation and Ion Formation

APCI is a soft ionization technique and an LC–MS interface that operates in gas phase. APCI and ESI interfaces are hosted normally in the same instrumentation. A liquid effluent coming from the LC at flow rates between 0.2 and 2 ml/min is nebulized into a heated (350–600 °C) vaporization tube. The high temperature compensates for the latent heat of vaporization and does not spoil the analytes that remain at temperatures slightly higher than the ambient. A buffer gas (air, nitrogen) is also used at this point to favor the ionization. A schematic of an APCI source is reported in Figure 5.4. The gas–vapor mixture enters the atmospheric pressure region in which a corona discharge needle, kept at a high voltage (5–6 kV), is positioned. A beam of electrons is accelerated in this electric field accomplishing the task of ionizing the buffer gas, in most cases nitrogen, generating the so-called reagent or primary ions In positive ion mode mainly N2 ‡ • , O2 ‡ • , H2 O‡ • , and NO‡ • ions are generated, whereas in negative ion mode O2 • , O • , NO2 • , NO3 • , O3 • , and CO3 • ions are formed as shown in Figure 5.5 [36,37]. A series of reactions take place, involving the charge transfer from the primary ions to solvent molecules for the production of “reactant ions.” These reactant ions transfer the charge to the analyte molecules that, depending on their proton affinity (PA), form either protonated or deprotonated molecular ions. The ionization process leads to [M + H]+ ions’ formation in the positive ion mode and [M H] ions’ formation in the negative ion mode. The formation of adduct ions is less prevalent in APCI than in ESI; however, [M-HCOO] or [M-CH3COO] ions may be observed. During the APCI process, parallel reactions can occur, some of which can cause detrimental effects on the overall process. External components, such as impurities or additives of the mobile phase, can provoke ion suppression or may generate interfering adduct ions. These matrix effects are observed to a lesser

Figure 5.4 Schematic of the APCI interface (reproduced with permission).

5.2 API Sources

Figure 5.5 Ion formation in (a) positive and (b) negative ion mode in APCI (reproduced with permission).

extent in APCI than in ESI. Air components can enter the plasma as well, producing unwanted species that have an effect on the signal response. The APCI interface is successfully used in the analysis of volatile and semivolatile molecules of low polarity that are neither strong acids nor strong bases, with a molecular weight up to 1500 u. Polar and nonpolar solvents can be used as mobile phase, and the pH does not have a significant influence on the ionization of the analytes. The ionization process produces a few fragments that do not provide structural information; therefore, multiple-stage MS is required for compound characterization. In contrast to ESI, APCI does not produce multiple charged ions. MS/MS can also be helpful to enhance the sensitivity and the selectivity, especially when dealing with trace determinations in complex matrices. The APCI interface is used in a large number of applications due to its robustness and high sensitivity. In particular, it is widely used in pharmaceutical chemistry, pharmacology, biotechnology, biochemistry, food, and environmental analyses. 5.2.3 Atmospheric Pressure Photoionization

The principle of photoionization (PI) was already used a few decades ago in gas chromatography (GC) and LC, although only recently it has found a widespread use as an ionization method for MS. Atmospheric pressure photoionization (APPI) is a relatively new LC–MS interface, presented in 2000 for the analysis of nonpolar molecules that do not ionize with APCI. Like the other API interfaces, it is a soft ionization technique [38–45]. Ions are generated from the interaction of a molecule with a photon in an ion source with a design very similar to that of

95

96

5 LC–MS Interfaces

an APCI source. Combined APCI/APPI sources are available on the market and allow one to operate in either one or both modes. In APPI, a vacuum UV (VUV) lamp is used for the ionization of the molecules in the gas phase. 5.2.3.1

Principle of Operation

In Figure 5.6, a schematic of the interface is shown in comparison with an APCI source [45]. A kripton VUV lamp emits photons at 123.9 nm and 116.5 nm with energies of 10.03 and 10.64 eV, respectively. This energy is not sufficient to ionize the major components of air and most solvents. The radiation beam can be placed orthogonally or in-line with respect to the MS ion path, depending on the manufacturer [40–42,45]. The molecules (M) absorb the UV photons leading to electronically excited species; if the photons’ energy (E = hν) exceeds the ionization energy (IE) of the analyte, it is ionized forming a radical molecular ion (primary ionization). M ‡ hν ! M*

(5.1)

M* ! M• ‡ ‡ e

(5.2)

However, solvent molecules, present at higher percentage, can also undergo the same primary process depleting the emitted photons [43]. Therefore, direct ionization is not so efficient to be competitive with other API sources. Moreover, at atmospheric pressure, the ion’s free pathway being very short, it is very likely

Figure 5.6 (a,b) Schematic of APPI source in comparison with an APCI one (reproduced with permission).

5.2 API Sources

that radical molecular ions undergo collisional reactions. As a consequence, secondary reactions can occur. The main mechanisms involved in a secondary ionization are charge exchange, electron capture (EC), and proton transfer. Other reactions are also possible; when the IE > hν, M∗ may undergo a deexcitation process, such as photodissociation, photon emission, or collisional quenching with a nonexcited molecule. All these reactions are undesirable. Essentially, it was demonstrated that the ionization efficiency can greatly be improved (two to three orders of magnitude) when a photoionizable compound (dopant) is present in the LC effluent that undergoes primary ionization and promotes the analyte ionization exploiting secondary ionization pathways [44]. The dopant acts as an intermediate between photons and analytes. The dopant forms radical molecular ions with high recombination energy and/or a low PA. The use of the dopant (D) enhancing gas-phase reactions allows to obtain a higher yield of certain species and is necessary to achieve higher sensitivities. The ionization mechanism depends on the PA values of the molecules involved (dopant, solvent, and analyte) and on their capacity to capture an electron in the gas phase, called electron affinity (EA) [40–42]. Charge exchange is the most important mechanism for apolar and low-polarity compounds. Proton transfer is common for acidic or basic compounds. D• ‡ ‡ M ! D ‡ M• ‡ D• ‡ ‡ M ! ‰D

if EAD > EAM

HŠ• ‰M ‡ HŠ‡

…charge exchange†

if PAM > PA‰D-HŠ•

…proton transfer†

(5.3) (5.4)

In the negative ion mode, the ionization process is initiated by the photoionization of the dopant according to reaction (5.2). If electron affinitive species are present in the source, such as oxygen, for example, they can capture the lowenergy electrons, as in reaction (5.5). O2 ‡ e ! O2

•

electron capture

(5.5)

The superoxide O2 • ion can promote charge exchange (reaction (5.6)), proton transfer (reaction (5.7)), or substitution reactions (reaction (5.8)) with analytes or solvent molecules. M ‡ O2

•

!M

M ‡ O2

•

! ‰M

HŠ ‡ HO2 •

(5.7)

M ‡ O2

•

! ‰M

H ‡ OŠ ‡ OH•

(5.8)

•

‡ O2

if EA…M† > EA…O2 † ˆ 0:451 eV

(5.6)

In any case, the reactions must be addressed to obtain a high yield of a single species choosing the proper solvent and dopant. Commonly used dopants are benzene (it leads mainly to protonated analytes), acetone, toluene, anisole, tetrahydrofuran, chlorobenzene, bromobenzene, and hexafluorobenzene. However, it is necessary to have a good knowledge of the chemistry involved in the gas-phase reactions to control the ionization process and to obtain the expected results. The sensitivity strongly depends upon the method conditions and the operator

97

98

5 LC–MS Interfaces

ability. An expert operator can achieve detection limits in the low nmol/l range [41,43–50]. For reversed phase (RP) analysis, the combination water/methanol yields a higher sensitivity than water/acetonitrile, due to the higher PA of acetonitrile [47]. Typical normal-phase (NP) mobile phases can be added to nonpolar solvents with low PA. Toluene, as well as acetone, is often used as a dopant [48,49]. The addition of weak acids may have a negative influence on the ionization process in negative ion mode. Flow rate has the main role in the ionization efficiency by proton transfer. Generally, high flow rates are reported to reduce ionization efficiency [50]. Recently, Robb and Blades proposed an improved PI source, an atmospheric pressure electron capture dissociation (APECD) source, for the analysis of polar, nonvolatile, and thermally labile compounds, such as peptide and protein at subpicomolar concentrations, obtaining performances comparable to ESI [51]. This source greatly extends the range of compounds analyzable with APPI, which can become an a alternative technique to ESI. The apparatus consists of a commercial APPI source equipped with an electropneumatic heated nebulizer obtained by electrifying the sprayer, using an external high-voltage power supply. The results are interesting; in fact, the response increased by two-order magnitude [40,51]. A future development of the technique could be the study of a nano-spray model, which might give better performances. 5.2.4 Atmospheric Pressure Laser Ionization

APLI has recently been introduced in the realm of API sources as an alternative or complementary ionization for APCI and APPI with a high specificity in the analysis of small, nonpolar molecules, in particular aromatic compounds. This method employs laser beams to promote the ionization of the analytes [52–54]. It is a soft ionization method, and modulating the laser power density can induce fragmentation. 5.2.4.1

Principle of Operation and Ion Formation

In Figure 5.7, a schematic of the APLI interface is illustrated [55]. Ionization is promoted at atmospheric pressure by a pulsed fixed frequency laser beam placed in front of the MS sampling orifice. The position of the laser beam does not seem to be crucial. The corona discharge needle is disconnected, and a field gradient of 50–200 V/cm is used to improve ion transmission. Resonantly enhanced multiphoton ionization (REMPI) at atmospheric pressure is the process that governs the ion formation through very well-described reactions [56–60]. The adsorption of the radiation leads to the formation of the excited molecule and to loss of an electron, with the formation of a radical molecular ion, M+• . This reaction is unlikely to have any efficiency to be useful in analytical

5.3 Non-API Sources

Figure 5.7 Schematic of APLI source (reproduced with permission).

applications. Due to laser pulses, further absorption of photons can lead to the fragmentation of the precursor ion. Photodissociation can occur; however, it does not provide ions in sufficient numbers. When an additional light source is applied, an intersystem crossing (ISC) phenomenon can deactivate resonantly excited states and promote ionization. Mobile phase and solvent molecules do not interfere in the ionization process because they do not absorb at the UV wavelengths used. With APLI, instead of APPI, even without dopants, an efficient formation of radical cations is observed in most cases. In addition, the sensitivity for nonpolar to very low-polarity aromatic or highly conjugated compounds is up to three orders of magnitude higher compared to APPI [52]. At present, APLI outperforms all other API techniques for the analysis of polycyclic aromatic hydrocarbons (PAH) in terms of sensitivity and detection limits. However, its use is limited to some specific applications, mainly concerning PAH.

5.3 Non-API Sources

All API sources are soft ionization-based interfaces and therefore are not very informative, because they produce poor fragmentation or none at all. To get more structural information and to increase the identification capability, two strategies are generally adopted. The first consists in causing the fragmentation by collisional-induced dissociation (CID) with argon or helium molecules with adjustable kinetic energy after the ionization process. To this aim, two analyzers (MS/MS) and a collision cell are

99

100

5 LC–MS Interfaces

required, increasing the analysis cost. Furthermore, CID is not very reproducible and the mass spectra cannot be collected into reliable libraries. The second highly expensive strategy is to use high-resolution analyzers such as Orbitrap or TOF for exact mass measurements. In this case, a number of possible molecular formulae, depending on mass accuracy, are obtained for the analyte identification. In addition, at atmospheric pressure, the ion’s mean free pathway is just 65 nm. As a consequence, the probability of collisions with molecules present in the gas phase, solvent, for example, is extremely high and many reactions can occur at the same time. An expert operator must govern these reactions accurately to get the desired ionization yield. In ESI, the matrix could also affect vaporization and ion evaporation leading to worse performances in quantitative analysis [7,8,54]. The setting up of the analytical method with an API source and the interpretation of multicharged spectra, when present (in ESI or ESI–APPI interface), could be difficult and time-consuming. For all these reasons, a few groups are working to realize a suitable non-API interface aimed at overcoming all the listed drawbacks, with challenging results. Predecessors of these interfaces could be considered the particle beam and the thermospray interfaces, based on different physical properties, used to remove the solvent from the LC effluent before entering the MS [1,61]. 5.3.1 Direct-EI

Electron ionization is a well-known ionization technique, widely used in hyphenated GC–MS systems. It offers plenty of fragmentation for an unparalleled identification capacity. In the past, many authors made considerable efforts to generate EI spectra from LC effluent. However, with the advent of ESI, researchers’ interest was completely drawn to exploit the potential of the new interface, which opened the way to new applications in many fields, particularly of biological interest. However, soft ionization techniques are less informative than EI and do not completely fill the gap toward the detection of small nonpolar molecules. For these reasons, further attempts at LC–EI–MS coupling were done by Cappiello et al. who realized a microparticle beam interface. This interface was still an external interface between a high-vacuum source and the atmospheric pressure liquid phase, necessary to evaporate and eliminate the solvent before entering the ion source, without hampering the analyte access. The interface could cause contamination, analyte losses, poor linearity, and reproducibility, especially at a high concentration of water in the mobile phase. Thus, the particle beam interface resulted in being less attractive than the GC–EI–MS interface and not competitive with the new API sources for LC. In recent years, due to the advent of nano-LC technology, a new, more challenging interface was realized by Cappiello et al., which introduced all the column effluents directly into the EI source, avoiding sample loss, contaminations, and irreproducibility. Because of its straightforward design, the interface was called direct Direct-EI.

5.3 Non-API Sources

Figure 5.8 Schematic of the Direct-EI interface and mechanism of ion formation.

Nano-LC columns work at submicroliter/min flow rates. The LC effluent enters directly into the EI source through the new interface, and solvent evaporation occurs inside the source in which the solvent excess is eliminated by diffusive or turbomolecular pumps, which are similar to those used in a commercial GC–MS system. In fact, a submicroliter/min flow rate produces a submilliter/ min vapor rate, which is within the pumping capacity of most MS systems. The lowest flow rate able to generate a fine aerosol is 100 nl/min. The system works at flow rates ranging from 100 to 900 nl/min. At higher flow rates (more than 1 μl/min), chemical ionization activated by solvent vapors can occur, and the protonated molecule abundance increases. It is worth noting that the interface is concentration sensitive so that higher flow rates imply lower signal intensity. Therefore, flow rates higher than 900 nl/min imply not only chemical ionization but also a reduced response. Commercial GC–MS can easily be adapted to work as a nano-LC–MS system. The new interfacing apparatus, schematized in Figure 5.8, is entirely held into the source. This internal interface insulates the LC effluent, which must enter the source as a liquid phase to prevent premature in-tube solvent evaporation and clogging. Analytes enter the MS source through a capillary tubing (25-μm internal diameter (ID)) that protrudes a few millimeters inside the source. The ion source temperature must range from 250 to 350 °C. EI differs from other ionization techniques in that it produces a hard ionization, it operates under high-vacuum and high-temperature conditions, and it is based on a physical ionization process. In all API sources, ionization is due, to some extent, to a chemical reaction so that coeluted species, such as a matrix or a solvent, can strongly interfere with analyte ionization changing the reaction yield. EI ionization is, on the opposite, a physical process; the interaction with 70-eV energy electrons released from a tungsten filament produces a radical molecular ion that undergoes abundant fragmentation inside the source (caused by a surplus of vibrational energy that exceeds bondage energy), without interfering with the coeluted substances. In fact, under high-vacuum conditions, ion–molecule collisions (and reactions) are not likely to occur. As a consequence, signal intensity is

101

102

5 LC–MS Interfaces

related only to the concentration of each analyte and ionization is unaffected by chemical reactions with the matrix (or to a very low extent). The coeluted compounds are ionized independent of one another. Fragmentation is highly reproducible and generates high-quality and informative spectra, which can be compared with those present in reliable and commercially available libraries (National Institute of Standard and Technology (NIST), Wiley). NIST and Spectral Deconvolution and Identification System (AMDIS) also developed algorithms, which extract the single-analyte mass spectrum in the case of unresolved chromatographic peaks [62,63]. Solvent ions are present in the low mass range of the spectrum recorded; however, they can be eliminated by background subtraction, an operation commonly used in the GC–MS analysis also to get a better quality spectrum [64,65]. Indeed, the new interface can work like a GC–MS interface enlarging the range of compounds that can be analyzed to thermolabile, polar, and nonpolar molecules with low molecular weight (approximately up to 500 u), as shown in Figure 5.9, in which 11 compounds (9–19) are not suitable for GC–MS analysis, whereas the first group of substances is not amenable to ESI detection [66]. Therefore, with a low-cost LC–MS instrumentation, without the use of a highresolution MS, a highly informative or highly selective analysis can be performed for a very wide range of small-molecule applications. The direct EI interface is compatible with volatile buffers, acids, and modifiers and shows an increased tolerance to nonvolatile buffers. This characteristic widens the choice of suitable chromatographic methods, improving de facto the LC–MS application potential.

Abundance

5 TIC: MIX_002.D\data.ms

1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19.

18000 16000 14000 2,3

12000 10000

4

7 6

8000 6000

13 12 11

9 1

4000

10

18

Dicamba (12.5 mg/L) Bromoxynil (12.5 mg/L) 2,4 D (12.5 mg/L) Mecoprop (12.5 mg/L) MCPB (12.5 mg/L) 2,4 DB (12.5 mg/L) Silvex (12.5 mg/L) Lindane (6.7 mg/L) Methoxychlor(6.7 mg/l) β -Endosulfan (11.0 mg/L) α-Endosulfan (22.0 mg/L) o,p’-DDD (6.7 mg/L) p,p’-DDD (6.7 mg/L) Heptachlor (6.7 mg/L) o,p’-DDT (6.7 mg/L) p,p’-DDT (6.7 mg/L) HCB (6.7 mg/L) p,p’-DDE (6.7 mg/L) Aldrin (6.7 mg/L)

16,17 15 14 19

2000

8

16.00 18.00 20.00 22.00 24.00 26.00 28.00 30.00 32.00 34.00 36.00 38.00

Time

All the pesticides were dissolved in THF/water (50/50, v/v) acidified with 1% formic acid Figure 5.9 HPLC–EI–MS separation of a mixture of pesticides. Compounds from 9 to 19 are not suitable for GC–MS analysis.

5.3 Non-API Sources

Moreover, the technique is easy to use and easy to automatize with respect to API techniques and does not require an expert operator. Direct-EI allows an accurate quantitative detection giving up to four to five orders of magnitude of linear response. An improvement in LODs, hitherto in the order of picograms in SIM mode, is expected by interfacing the direct EI to a triple quadrupole operating in selected reaction monitoring (SRM) mode. 5.3.2 EI of Cold Molecules in Supersonic Molecular Beam (SMB)

The SMB approach proposed by Amirav and coworkers can be considered an improvement upon the particle beam technology, focused on achieving an EI spectrum with an enhanced molecular ion so that the identification of the analytes is improved. The radical molecular ion intensity is enhanced preserving the typical EI library searchable fragmentation, which is the main advantage of EI. The schematic of the apparatus is reported in Figure 5.10. The vaporization is realized in two steps. In the first step, the liquid effluent is heated by a thermospray system, as in an APCI interface. In the second step, the sample expands from the supersonic nozzle. The LC effluent is vaporized at about 1–2 bar inside a glass tube, connected to a supersonic nozzle by a short fused silica capillary transfer line. The capillary impedes the flow, leading to an effective high-pressure sample vaporization combined with 0.1 bar low pressure behind a 300-μm supersonic nozzle to suppress cluster formation during the supersonic expansion, yet obtaining an efficient vibrational cooling. Sample vaporization is obtained by the pneumatically assisted spray formation followed by a fast thermal vaporization of the analytes before their expansion from the

Figure 5.10 Schematic of supersonic EI– LC–MS apparatus (reproduced with permission).

103

104

5 LC–MS Interfaces

supersonic nozzle. This new approach of EI–LC–MS, called capillary separated vaporization chamber and nozzle (CSVCN) system, gives performances comparable to APCI in terms of robustness and usability. Because the sample flow rate is more than 100-fold greater in the capillary transfer line than in the vaporization chamber, thermal degradation can be considered negligible in the transfer line, and it could eventually occur in the chamber. However, this thermal degradation does not occur because the molecules are supercooled, avoiding any further dissociation, as soon as they expand from the supersonic nozzle. The CSTNV vaporization chamber is heated at an external temperature of about 500–800 °C, as well as in APCI, although with respect to APCI, it is fully thermally assisted and not thermally/gas assisted. No nitrogen gas generator is required (in contrast to APCI and ESI) because the vaporized solvent acts as a carrier gas. The SMB of vibrationally cold indissociated molecules is collimated by a skimmer into a second vacuum chamber, equipped with two diffusion pumps. The analytes are ionized by an EI source and ions are deflected at 90 °C through an ion mirror and sent to the analyzer. Unlike APCI, all the analytes can be ionized regardless of their polarity and no gas-phase ion–molecule reactions occur, so that no ME is observed. The spectra are the same way obtainable from an EI ionization source; thus, they can be compared with NIST spectra, although they have an enhanced molecular ion. In Figure 5.11, a supersonic EI spectrum compared with the EI spectrum of the same compound in the NIST library is shown. Amirav also developed a software that converts experimental MS data into a chemical formula, confirming or rejecting automatically the NIST library identification [67]. The response is rather uniform due to constant ionization efficiency so that an estimation of the concentration of unknown samples can be done without knowing the single compound identity. Higher EI sensitivities are obtained compared to particle beam MS. LODs are in the low picogram range in SIM mode, comparable to those obtained with the Direct-EI approach. Furthermore, like Direct-EI, the MS system is low cost and can be used with both in GC–MS and in LC–MS. 5.3.3 Combined Single-Photon Low-Pressure Photoionization and EI Ionization

An interesting attempt was recently made by Zimmerman and Cappiello who realized an interface that uses photoionization under vacuum conditions in order to avoid the main drawbacks of working at atmospheric pressure: ion–molecule reactions and matrix effect. In Figure 5.12, the ion source asset is shown. The LC effluent enters the ion source through a Direct-EI interface. Inside the source, an electron beam (12 keV) is directed through a 300-nm-thin SiNx foil into a reservoir filled with argon (2000 hPa pressure). As a result, these atoms are excited and form excimers that, upon their decay, emit VUV radiation at 126 nm focused by an optical system and used for single-photon ionization (SPI).

5.3 Non-API Sources

Figure 5.11 A comparison of cold EI mass spectrum of corticosterone obtained with the supersonic LC–EI–MS system and the NIST spectrum, with NIST library matching factors and probability of identification (reproduced with permission).

Figure 5.12 Schematic of the single photon low-pressure photoionization–electron ionization interface.

105

106

5 LC–MS Interfaces

The quadrupole analyzer is set orthogonally to the capillary carrying the LC effluent [68]. With this combined interface, both reproducible, library-matchable EI–MS spectra and soft PI spectra can be obtained. As a consequence, the advantages of the two ionization techniques are summed. In fact, good sensitivities can be obtained working in PI and, at the same time, collateral reactions and ME are avoided because ionization is realized under high-vacuum conditions. Furthermore, the technique has the potential of EI in getting structural information on analytes. 5.3.4 LC/DESI–MS Interface

An interesting approach has been proposed recently by Cai et al. [69]. An LC effluent passes through a PEEK capillary tube with a microdrilled orifice on its wall, as shown in the schematic reported in Figure 5.13. A small droplet of liquid emerges out of the orifice, whereas the LC effluent runs toward the outlet. The droplet is ionized by desorption electrospray ionization (DESI), whereas the great amount of analytes remains into the peak capillary tube and can be collected for preparative purposes or analyzed by other methods. This new technique allows to avoid dead volume formation, which could determine chromatographic peak band broadening. Reactive DESI can also be applied for online derivatization. The interface works efficiently also with higher mobile phase flow rates, such as the flow rates employed working with monolithic columns.

Figure 5.13 Schematic of the new LC/DESI–MS interface (reproduced with permission).

References

References 1 Blakley, C.R. and Vestal, M.L. (1983)

2

3

4

5

6

7

8

9

10

11

Thermospray interface for liquid chromatography/mass spectrometry. Anal. Chem., 55, 750–754. Himmelsbach, M. (2012) 10 years of MS instrumental developments: impact on LC–MS/MS in clinical chemistry. J. Chromatogr. B, 883, 3–17. Sosa-Ferrera, Z., Mahungo-Santana, C., and Santana-Rodriguez, J.J. (2012) New developments in liquid chromatography mass spectrometry for the determination of micropollutants. Chromatogr. Res. Int., 2012, 1–15. Holčapek, M., Jirásko, R.J., and Lísa, M. (2012) Recent developments in liquid chromatography-mass spectrometry and related techniques. J. Chromatogr. A, 1259, 3–15. Covey, T.R., Bruce, T., and Bradley, B.S. (2009) Atmospheric pressure ion sources. Mass Spectrom. Rev., 28 (6), 870–897. Wittrig, B. (2013) Comprehensive pesticide residue analysis by LC–MS/MS using an ultra aqueous C18 column. Available at: http://www.restek.com/ Technical-Resources/Technical-Library/ Foods-Flavors-Fragrances/fff_A020. Cech, N.B. and Enke, C.G. (2001) Electrospray ionization mass spectrometry: history, theory, and instrumentation. Mass Spectrom. Rev., 20 (6), 362–387. Cody, R.B. (2002) Electrospray ionization mass spectrometry: history, theory, and instrumentation, in Applied Electrospray Mass Spectrometry, vol. 8 (eds B.N. Pramanik, A.K. Ganguly, and M.L. Gross), Marcel Dekker, Inc., New York, pp. 1–104. Cech, N.B., Enke, C.G. (2006) Electrospray ionization mass spectrometry: how and when it works, in The Encyclopedia of Mass Spectrometry, vol. 8 (eds M.L. Gross and R.M. Caprioli), Elsevier, The Netherlands, pp. 171–180. Kebarle, P. and Yeunghaw, H. (1997) On the mechanism of electrospray mass spectrometry, in Electrospray Ionization Mass Spectrometry (ed. R.B. Cole), Wiley, New York, p. 14. Tang, L. and Kebarle, P. (1993) From ions in solution to ions in the gas phase: The

12

13

14

15

16

17

18

19

20

21

mechanism of electrospray mass spectrometry. Anal. Chem., 65, 972A–986A. Cole, R.B. (2000) Some tenets pertaining to electrospray ionization mass spectrometry. J. Mass Spectrom., 35 (7), 763–772. Kebarle, P. and Peschke, M. (2000) On the mechanisms by which the charged droplets produced by electrospray lead to gas phase ions. Anal. Chim. Acta, 406 (1), 11–35. Thomson, B.A. and Iribarne, J.V. (1979) Field-induced ion evaporation from liquid surfaces at atmospheric pressure. J. Chem. Phys., 71, 4451–4463. Leitner, A., Emmert, J., Boerner, K., and Lindner, W. (2007) Influence of solvent additive composition on chromatographic separation and sodium adduct formation of peptides in HPLC–ESI MS. Chromatographia, 65 (11/12), 649–653. Saf, R., Mirtl, C., and Hummel, K. (1994) Electrospray mass spectrometry using potassium iodide in aprotic organic solvents for the ion formation by cation attachment. Tetrahedron Lett., 35, 6653–6656. Ackloo, S.Z., Smith, R.W., Terlouw, J.K., and McCarry, B.E. (2000) Characterization of ginseng saponins using electrospray mass spectrometry and collision induced dissociation experiments of metal attachment ions. Analyst, 125, 591–597. Giorgianni, F., Cappiello, A., BeranovaGiorgianni, S., Palma, P., Trufelli, H., and Desiderio, D.M. (2004) LC–MS/MS analysis of peptides with methanol as organic modifier: improved limits of detection. Anal. Chem., 76 (23), 7028–7038. Cech, N.B. and Enke, C.G. (2000) Relating electrospray ionization response to nonpolar character of small peptides. Anal. Chem., 72, 2717–2723. Zhou, S. and Cook, K.D. (2001) A mechanistic study of ESI–MS: charge gradients within ESI droplets and their influence on ion response. J. Am. Soc. Mass Spectrom., 12, 206–214. Cech, N.B., Krone, J.R., and Enke, C.G. (2001) Predicting electrospray response

107

108

5 LC–MS Interfaces

22

23

24

25

26

27

28

29

30

31

from chromatographic retention time. Anal. Chem., 73, 208–213. King, R., Bonfiglio, R., Fernandez-Metzler, C., Miller-Stein, C., and Olah, T. (2000) Mechanistic investigation of ionization suppression in electrospray ionization. J. Am. Soc. Mass Spectrom., 11 (11), 942–950. Trufelli, H., Palma, P., Famiglini, G., and Cappiello, A. (2011) An overview of matrix effects in liquid chromatographymass spectrometry. Mass Spectrom. Rev., 30 (3), 491–509. Voyksner, R.D. (1997) Combining liquid chromatography with electrospray mass spectrometry, in Electrospray Ionization Mass Spectrometry (ed. R.B. Cole), John Wiley & Sons, Inc., New York, pp. 323–341. Antignac, J.P., Wash, F., Monteau, H.D., Brabander, F.A., and Bizet, B.L. (2005) The ion suppression phenomenon in liquid chromatography–mass spectrometry and its consequences in the field of residue analysis. Anal. Chim. Acta, 529, 129–136. Tonidandel, L. and Seraglia, R. (2007) Matrix effect, signal suppression and enhancement in LC–ESI–MS, in Advances in LC–MS Instrumentation, vol. 72 (ed. A. Cappiello), Elsevier, Amsterdam, The Netherlands, pp. 193–210. Kebarle, P. and Verkerk, U.H. (2009) Electrospray: from ions in solution to ions in the gas phase, what we know. Mass Spectrom. Rev., 28, 898–917. Gangl, E.T., Annan, M., Spooner, N., and Vouros, P. (2001) Reduction of signal suppression effects in ESI–MS using a nanosplitting device. Anal. Chem., 73, 5635–5644. Andrewsa, C.L., Yub, C.P., Yangb, E., and Vourosa, P. (2004) Improved liquid chromatography–mass spectrometry performance in quantitative analysis using a nanosplitter interface. J. Chromatogr. A, 1053, 151–159. Munson, M.S.B. and Field, F.H. (1966) Chemical ionization mass spectrometry: I. General introduction. J. Am. Chem. Soc., 88 (12), 2621–2630. Horning, E.C., Horning, M.G., Carroll, D.I., Dzidic, I., and Stillwell, R.N. (1973)

32

33

34

35

36

37

38

39

40

41

New picogram detection system based on a mass spectrometer with an external ionization source at atmospheric pressure. Anal. Chem., 45, 936–943. Carroll, D.I., Dzidic, I., Stillwell, R.N., Horning, M.G., and Horning, E.C. (1974) Subpicogram detection system for gas phase analysis based upon atmospheric pressure ionization. Anal. Chem., 46, 706–710. Bruins, A.P. (1994) Atmospheric pressureionization mass spectrometry: II. Applications in pharmacy, biochemistry and general chemistry. Trend. Anal. Chem., 13 (2), 81–90. Bruins, A.P. (1991) Mass spectrometry with ion sources operating at atmospheric pressure. Mass Spectrom. Rev., 10 (1), 53–77. Andrade, F.J., Shelley, J.T., Wetzel, W.C., Webb, M.R., Gamez, G., Ray, S.J., and Hieftje, G.M. (2008) Atmospheric pressure chemical ionization source. 2. Desorption-ionization for the direct analysis of solid compounds. Anal. Chem., 80, 2654–2663. Raffaelli, A. (2007) Atmospheric pressure chemical ionization (APCI): new avenues for an old friend, in Advances in LC–MS Instrumentation, vol. 72 (ed. A. Cappiello), Elsevier, Amsterdam, The Netherlands, pp. 11–25. University of Bristol, School of Chemistry, Mass spectrometry facility. Atmospheric pressure chemical ionisation (APCI) (2013). Available at: http://www.chm.bris. ac.uk/ms/theory/apci-ionisation.html. Robb, D.B., Covey, T.R., and Bruins, A.P. (2000) Atmospheric pressure photoionization: an ionization method for liquid chromatography-mass spectrometry. Anal. Chem., 72, 3653–3659. Syage, J.A., Evans, M.D., and Hanold, K.A. (2000) Photoionization mass spectrometry. Am. Lab., 32, 24–29. Marchi, I., Rudaz, S., and Veuthey, J.L. (2009) Atmospheric pressure photoionization for coupling liquidchromatography to mass spectrometry: a review. Talanta, 78, 1–18. Robb, D.B. and Blades, W.B. (2008) State of the art in atmospheric pressure

References

42

43

44

45

46

47

48

49

50

photoionization for LC/MS. Anal. Chim. Acta., 627, 34–49. Syage, J.A., Hanold, K.A., Lynn, T.C., Horner, J.A., and Thakur, R.A. (2004) Atmospheric pressure photoionization II. Dual sources ionization. J. Chromatogr. A, 1050, 137–149. Raffaelli, A. and Saba, A. (2003) Atmospheric pressure photoionization mass spectrometry. Mass Spectrom. Rev., 22 (5), 318–331. Kauppila, T.J., Kotiaho, T., Kostiainen, R., and Bruins, A.P. (2004) Negative ionatmospheric negative ion-atmospheric pressure photoionization-mass spectrometry. Rapid Commun. Mass Spectrom., 18 (7), 808–815. Robb, D.B. and Blades, W.B. (2005) Factors affecting primary ionization in dopant assisted atmospheric pressure photoionization (DA-APPI) for LC/MS. Anal. Chim. Acta., 17, 130–138. Robb, D.B., Rogalski, J.C., Kast, J., and Blades, W.B. (2005) Effect of solvent flow, dopant flow and lamp current on dopantassisted atmospheric pressure photoionization (DA-APPI) for LC–MS. Ionization via proton transfer. J. Am. Soc. Mass Spectrom., 16, 1275–1290. Cai, S.S. and Syage, J.A. (2006) Atmospheric pressure photoionization mass spectrometry for analysis of fatty acid and acylglycerol lipids. J. Chromatogr. A, 1110, 15–26. Kauppila, T.J., Kuuranne, T., Meurer, E.C., Eberlin, M.N., Kotiaho, T., and Kostiainen, R. (2002) Atmospheric pressure photoionization mass spectrometry. Ionization mechanism and the effect of solvent on the ionization of naphthalenes. Anal. Chem., 74, 5470–5479. Mol, R., de Jong, G.J., and Somsen, G.W. (2005) On-line capillary electrophoresis mass spectrometry using dopant-assisted atmospheric pressure photoionization: setup and system performance. Electrophoresis, 26, 146–154. Robb, D.B. and Blades, MW. (2005) Effects of solvent flow, dopant flow, and lamp current on dopant-assisted atmospheric pressure photoionization (DAAPPI) for LC–MS. Ionization via proton transfer.

51

52

53

54

55

56

57

58

59

60

J. Am. Soc. Mass Spectrom., 16, 1275–1290. Robb, D.B. and Blades, W.B. (2008) Atmospheric pressure-electron capture dissociation of peptides using a modified photospray ion source. Rapid Commun. Mass Spectrom., 24, 3303–3308. Constanpel, M., Schellenträger, M., Schmitz, O.J., Gäb, S., Brockmann, K.J., Giese, R., and Benter, Th. (2005) Atmospheric-pressure laser ionization: a novel ionization method for liquid chromatography/mass spectrometry. Rapid Commun. Mass Spectrom., 19, 326–336. Schmitz, O.J. and Benter, T. (2007) Atmospheric pressure laser ionization (APLI), in Advances in LC–MS Instrumentation, vol. 72 (ed. A. Cappiello), Elsevier, Amsterdam, The Netherlands, pp. 89–113. Balogh, M.P. (2010) When electrospray just isn’t enough. LCGC North America, 28 (6), 440–453. Kersten, H., Lorenz, M., Brockmann, K.J., and Benter, T. (2011) Evaluation of the performance of small diode pumped UV solid state (DPSS) Nd: YAG lasers as new radiation source for atmospheric pressure laser ionization mass spectrometry (APLI–MS). J. Am. Soc. Mass Spectr., 22, 1063–1069. Robin, M.B. (1980) Multiphoton fragmentation and ionization. Appl. Opt., 19, 3941–3947. Goodman, L. and Philis, J. (1992) Multiphoton absorption spectroscopy, in Applied Laser Spectroscopy: Techniques, Instrumentation, and Applications (ed. D.L. Andrews), Wiley-VCH Verlag GmbH, Weinheim. Ashfold, M.N.R. and Howe, J.D. (1994) Multiphoton spectroscopy of molecular species. Annu. Rev. Phys. Chem., 45, 57–82. Pfab, J. (1995) Laser-induced fluorescence and ionization spectroscopy of gas phase species, in Spectroscopy in Environmental Science (eds R.J.H. Clark and R.E. Hester), John Wiley & Sons, Inc., New York. Boesl, U. (1999) Multiphoton excitation in mass spectrometry, in Encyclopedia of

109

110

5 LC–MS Interfaces

61

62

63

64

Spectroscopy and Spectrometry (ed. C. McNeil), Academic Press, New York. Willoughby, R.C. and Browner, R.F. (1984) Monodisperse aerosol generation interface for combining liquid chromatography with mass spectroscopy. Anal. Chem., 56, 2626–2631. Palma, P., Famiglini, G., Trufelli, H., Pierini, E., Termopoli, V., and Cappiello, A. (2011) Electron ionization in LC–MS: recent developments and applications of the direct-EI LC–MS interface. Anal. Bioanal. Chem., 399, 2683–2693. Cappiello, A., Famiglini, G., Palma, P., Pierini, E., Termopoli, V., and Trufelli, H. (2011) Direct-EI in LC–MS: towards a universal detector for small-molecules applications. Mass Spectrom. Rev., 30, 1242–1255. Cappiello, A., Famiglini, G., Termopoli, V., Trufelli, H., Zazzeroni, R., Jacquoilleot, S., Radici, L., and Saib, O. (2011) Application of liquid chromatography-Direct-Electron ionization–MS in an in vitro dermal absorption study: quantitative determination of trans-cinnamaldehyde. Anal. Chem., 83 (22), 8537–8542.

65 Famiglini, G., Palma, P., Pierini, E.,

66

67

68

69

Trufelli, H., and Cappiello, A. (2008) Organochlorine pesticides by LC–MS. Anal. Chem., 80 (9), 3445–3449. Famiglini, G., Palma, P., Termopoli, V., Trufelli, H., and Cappiello, A. (2009) Single-step LC/MS method for the simultaneous determination of GCamenable organochlorine and LCamenable phenoxy acidic pesticides. Anal. Chem., 81 (17), 7373–7378. Granot, O. and Amirav, A. (2005) LC–MS with electron ionization of cold molecules in supersonic molecular beams. Int. J. Mass Spectrom., 244, 15–28. Schepler, C., Sklorz, M., Passig, J., Famiglini, G., Cappiello, A., and Zimmermann, R. (2013) Flow injection of liquid samples to a mass spectrometer with ionization under vacuum conditions: a combined ion source for single-photon and electron impact ionization. Anal. Bioanal. Chem., 405, 6953–6957. Cai, Y., Adams, D., and Chen, H. (2014) A new splitting method for both analytical and preparative LC/MS. JASMS, 25, 286–292.

111

6 LC–MS Applications in Environmental and Food Analysis Alessandra Gentili, Fulvia Caretti, and Virginia Pérez Fernández

6.1 Introduction

Over the past few years, the number of applications of LC–MS has increased considerably in many areas of chemistry, pharmaceutical sciences, and biochemistry because of the rapid advances in LC–MS technology. The volume and quality of knowledge acquired in these fields depend directly on such developments. Triple-quadrupole (QqQ) mass spectrometers are still preserving their market leadership due to their competitive price and unsurpassed performances in quantitative analysis. Hybrid instruments such as quadrupole time-of-flight (QqTOF) and Q-linear ion traps (QqQLIT) have already been affirmed for their peculiar advantages: QqQLIT for its capability to combine the selective scanning modes of QqQ with LIT experiments providing improved performance and enhanced sensitivity such as in enhanced full scan and product ion modes, and QqTOF for its high scanning speed, accurate mass measurement, and MS2 operations. First, Fourier transform ion cyclotron resonance (FT ICR) and, more recently, orbitrap detectors have gained position in several application areas due to their superior resolving power. At the same time, newly introduced improvements in LC [1], involving monolith technology, fused core columns, hightemperature LC (HTLC), and ultrahigh-performance LC (UHPLC), make the LC–MS technology more attractive and powerful. The improved separation and detection capabilities of LC–MS instruments, as well as sophisticated software tools for data acquisition and processing, have played a central role in the development of new analytical strategies to perform high-throughput screening of contaminants, fingerprint analysis of natural products, quantitative analysis with enhanced identification power, and characterization of biomolecules such as peptides, proteins, oligosaccharides, lipids, and oligonucleotides. Environmental and food fields are two of the most important application areas of LC–MS technology. In this chapter, the practical aspects, advantages, and weaknesses of the

Analytical Separation Science, First Edition. Edited by Jared L. Anderson, Alain Berthod, Verónica Pino Estévez, and Apryll M. Stalcup.  2015 Wiley-VCH Verlag GmbH & Co. KGaA. Published 2015 by Wiley-VCH Verlag GmbH & Co. KGaA.

112

6 LC–MS Applications in Environmental and Food Analysis

several analytical approaches will be discussed along with trends and future potential developments within each of these application areas.

6.2 Environmental Applications

Environmental analysis is a very important application area of LC–MS mainly related to the study about the occurrence and fate of organic micropollutants in wastewater, sludge, natural waters, drinking water, sediments, soil, and aquatic biota. The term “organic micropollutants” is meant to include any organic contaminant – pesticides, pharmaceuticals, personal care products, industrial chemicals, hormones, flame-retardants, plasticizers, and others – which enters the environment during its production, consumption, and disposal at ppm or lower level. Among all the known organic contaminants, pesticides are the most relevant and investigated since approximately 900 approved active ingredients, belonging to more than 100 different classes, are still being used worldwide. The same applies to legal and illegal drugs: at present, about 4000 pharmaceuticals are in use in the European Union (EU) [2]. Growing concern and awareness about the potential toxicity to living organisms or/and eco-toxicity of these contaminants is the driving force for developing fast and sensitive multicomponent methods in order to expand monitoring strategies and to investigate the micropollutant fate in the treatment processes and environment. Nevertheless, the wide variety of chemicals occurring in the several environmental compartments exhibits significant differences in physicochemical properties (pKa, log P, etc.) causing serious problems in the development of a “universal” residue analytical method. Recently, much effort has been made in order to reduce time, complexity, and cost of analysis by developing multiclass, multiresidue methods both for screening and for confirmation purposes. 6.2.1 Last Trends in Sample Preparation for LC–MS Analysis

Most of the analytical methods for the determination of micropollutants in water samples include at least one step for the simultaneous sample enrichment and sample cleanup with solid-phase extraction (SPE) cartridges, SPE extraction disks, or LC–LC column switching [3,4]. These steps, allowing typical enrichment factors varying between 20 and 1000, constitute the most time- and laborconsuming parts of the analytical process. In the past few years, the most recent tendencies are toward the development of online sample preparation units: SPE cartridges, turbulent flow chromatography (TFC), and direct sample injection into an LC–MS system without any sample treatment. The latter approach, uniquely allowed by the high sensitivity and selectivity of LC–MS instrumentations, can be applied to isolate many different families of contaminants simultaneously and quickly. Nevertheless, the coextraction of interferences is

6.2 Environmental Applications

responsible for low sensitivity, poor selectivity, and stress of analytical systems (chromatographic column and/or MS) [5]. 6.2.2 Advances and Trends in Liquid Chromatography

Over the past decade, advances on the LC side have also contributed to the development of high-throughput methods. UHPLC columns with sub-2 μm diameter porous particles have been widely and successfully employed to speed up the analysis of drugs [6], UV filters [7], pesticides [8], and perfluorinated compounds [9] maintaining similar or even better efficiency than the classical HPLC columns. The extremely narrow peaks produced by such columns have required the use of mass spectrometers with elevated acquisition rates in order to collect enough data points across the LC peak for accurate and reproducible data [10]. Other solutions to reduce the chromatographic run times are based on the use of monolithic and HTLC columns. Monolithic supports consist of a unique piece of porous material that offers a lower solvent resistance and enables ultrafast separations down to only a few seconds. Stationary phases of HTLC columns are based on materials with high temperature stability such as graphitized carbon, zirconium oxide, and polystyrene/divinylbenzene; application of temperatures between 90 and 200 °C is able to reduce column back pressure, to improve selectivity, to allow the use of high percentage of water (superheated water at 200 °C has a similar eluting power as methanol at ambient temperature), and to increase the separation speed by a factor of 3–20. However, both approaches present clear disadvantages that make their use in environmental analysis very limited [11]: stationary phases of monolithic columns have compositions and efficiencies unsuitable for the separation of a significant number of compounds of environmental concern. On the other hand, the high temperatures applied during HTLC separations can affect the stability of both the analytes and the packing materials [12]. Finally, the separation on fused core columns (superficially porous particles also called core–shell or porous shell particles) is gaining attention in this research field. Such columns, characterized by a smaller van Deemter A term, provide more than twice the speed and efficiency of columns with sub-2 μm totally porous particles at half the back pressure. This performance enhancement is also applicable to all HPLC instruments by increasing the column temperature between 30 and 40 °C in order to further decrease the back pressure. Wode et al. [13] separated 72 micropollutants (industrial chemicals, analgesics, anticonvulsants, antihypertensive, psychoactive substances, flame-retardants, and neutral and acidic pesticides) on a C18 core–shell column (2.6 μm) kept at 30 °C, by using a UHPLC coupled to a single-stage orbitrap. 6.2.3 Advances and Trends in Mass Spectrometry

Up to now, electrospray (ESI) has been the most applied source in this research area, initially together with atmospheric pressure chemical ionization (APCI)

113

114

6 LC–MS Applications in Environmental and Food Analysis

that is nowadays less used [14,15]. More recently, atmospheric pressure photoionization (APPI) has been emerging as an alternative for less polar pollutants such as steroids [16], even if it has not yet found versatile applications, whereas direct electron ionization has been valued for the detection of organochlorinated pesticides, which are not amenable to LC–API–MS [17]. Multimode ionization source is another option offered by the latest generation of instruments, which are able to simultaneously carry out both ESI and APCI experiments. Even if dual sources appeared to be a promising solution to increase screening productivity, they always compromise the sampling rate and sensitivity relative to a single ionization mode. Moreover, due to temperature constraint, the APCI mode is clearly not equal to a dedicated APCI source, and most apolar compounds could not be as efficiently ionized. In the past few years, environmental analysis has been performed by employing both low-resolution (LR) and high-resolution (HR) mass spectrometers. The recent advances in mass analyzer technology have favored the development of fast large-scale methods to determine a huge number of targeted and untargeted micropollutants and their transformation products. QqQ mass analyzers operating under selected reaction monitoring (SRM) have still been the most applied MS detectors because of their unsurpassed selectivity and wide linear dynamic range even if different solutions had to be adopted to overcome the limited number of SRM transitions monitored in a single chromatographic run. Greulich and Alder [18] conducted the determination of 300 pesticides in mineral water by direct sample injection into the LC–QqQ system. Two SRM transitions per analyte did not allow a sensitive detection of all analytes within a cycle time of 2 or 3 s; therefore, data acquisition was performed in two runs using time windows. Huntscha et al. [19] developed an automated multiresidue method consisting of an SPE–HPLC-QqQ system able to determine 88 polar organic micropollutants (pharmaceuticals, pesticides, biocides, corrosion inhibitors, an artificial sweetener, and their transformation products) with a broad range of physicochemical properties in groundwater, surface water, and wastewater. The global analysis took 36 min per sample. Of late, there is a growing use of QqQLIT, an LR hybrid mass spectrometer that combines fully functional QqQ and LIT within the same instrument. In addition to classical QqQ scan modes that make it suitable for sensitive and selective quantitative analysis, QqQLIT can perform MS3 operations useful for the identification of transformation products. Reemtsma et al. [20] succeeded to develop a multimethod for the determination of 150 metabolites of pesticides in groundand surface water by using direct injection LC–QqLIT. The analysis was performed in two analytical runs for positive or negative ESI, monitoring two SRM transitions for the analyte. The cycle time was 1.8 s, resulting in variable dwell times for each transition with a minimum of 30 ms. Gros et al. [21] described the development of a fast and robust analytical method, based on an automated offline SPE followed by UHPLC–QqLIT analysis, for the determination of 53 antibiotics and their metabolites in environmental matrices such as hospital wastewaters, urban wastewaters, and river waters. The use of HR mass spectrometers in environmental field began in 1999 with pesticide analysis [22], but

6.2 Environmental Applications

it has been in the recent years that their use has been increasing noticeably due to two major advantages: the accurate mass full-spectrum acquisition that allows the screening of thousands of compounds both target or unknown [23] and the improved selectivity able to distinguish between target ions and quasi-isobaric interfering ions. Although in comparison with LR mass spectrometers, HR mass analyzers have lower linear dynamic ranges and sensitivity, TOF, QqTOF, and orbitrap are endowed with insuperable identification power and a great versatility in performing a variety of tasks: pre-target analysis, post-target analysis, and nontarget analysis [23]. A pre-target approach requires compound-specific information before measurement, as typically occurs in LC–LRMS-based methods. In post-target analysis, the search for compounds is conducted after accurate mass full-spectrum acquisition and can be done even without using reference standards. The m/z values of the target analytes are extracted from the full-scan total ion current (TIC) chromatogram after its acquisition using a narrow mass window (±10 mDa); the identification is based on the retention time window, the measurement of the exact mass (accuracy < 5 ppm for orbitrap, 5–20 ppm for TOF), and the isotopic pattern. Finally, non-target screening methods evaluate the occurrence of compounds after the acquisition of the full-scan chromatogram by searching in-house spectral libraries, to identify unknown compounds without having any previous information. Martínez-Bueno et al. performed the analysis of 56 organic pollutants in wastewater by applying an efficient LC–MS strategy based on QqQLIT in combination with TOF-MS [24]. Quantification was performed by LC–QqQLIT operating in SRM mode in both positive and negative ion. Unequivocal identification was provided by the acquisition of three SRM transitions per compound and by LC-TOF analysis, which allowed achieving accurate mass measurements of the identified compounds with errors lower than 2 ppm. The flexibility of QqQLIT instruments allowed improving confirmatory information by the application of additional operation modes based on the use of the LIT mode (see Figure 6.1). Furthermore, the acquisition of full spectra at all times permitted a post-target analysis of non-target compounds, metabolites, or degradation products, thus avoiding additional cost and time. Another LC-TOF method providing the advantages of this technique is described in the work of Lara-Martin, González-Mazo, and Brownawell [25]. The comprehensive analysis of the most commonly used anionic and nonionic surfactants and their main degradation intermediates in aqueous and solid environmental matrices was carried out. Apart from the identification of the target compounds, the recorded full-scan spectra permitted to identify non-target metabolites such as alkyl sulfates and alkyl ether sulfates. Pesticides and their degradation products have also been studied by LC–TOF-MS. A multiclass method for the chromatographic separation and accurate mass identification of 101 pesticides and their metabolites using LC–TOF-MS was reported by Ferrer and Thurman [4]. Hybrid tandem instruments, such as QqTOF, provide HR data and structural information by performing product ion scan (PIS) experiments. This instrument can collect full-scan spectra and PIS spectra in a single injection by planning data-dependent acquisition (DDA) or information-dependent

115

116

6 LC–MS Applications in Environmental and Food Analysis

Figure 6.1 SRM total ion chromatogram (a), extracted ion chromatogram (b), and enhanced product ion spectrum (c) of ibuprofen simultaneously obtained by IDA experiments. (Reproduced with permission from Ref. [24]. Copyright 2007 American Chemical Society.)

acquisition (IDA) experiments in which TOF-MS full scan is the survey scan, while PIS scan is the dependent scan [23]. Hernández et al. [26] used QqTOF and a specialized processing data application manager to perform the retrospective analysis of pharmaceutical metabolites in urban wastewater without the need of an additional injection of sample extracts. Around 160 metabolites were investigated in wastewater samples using LC–QqTOF under MSE mode, that is, the simultaneous recording of two acquisition functions, at low and high collision energy, to maximize the number of structural information. Finally, orbitrap mass analyzer is an alternative to QqTOF instruments for identification of transformation products and metabolites or screening over a large mass range. Advantages of the LIT-orbitrap over QqTOF are high ion transmission resulting in higher MSn sensitivity and detection limits and a higher intensity range over that accurate mass data can be acquired. Helbling et al. [27] identified a variety of previously reported and unreported microbial transformation products of organic micropollutants by using LIT-orbitrap MS. Pharmaceuticals and pesticides were spiked into batch reactors seeded with activated sludge, and the

6.3 Food Toxicant Applications

candidate transformation products were preliminarily identified with an innovative postacquisition data processing method based on target and nontarget screenings of the full-scan MS data. Structures were proposed following the interpretation of MS spectra and MS/MS fragments. The results showed that the complementary use of both approaches allowed for a more comprehensive interpretation than either would have provided individually. However, the application of orbitrap in the environmental field is still rare [28] because of the two main drawbacks: a low scanning speed that makes it difficult to couple to UHPLC systems, where very narrow peaks are obtained, and a phenomenon of postinterface signal suppression that affects the sensitivity in detecting small molecular weight compounds.

6.3 Food Toxicant Applications

Food toxicants are mainly small molecules (100–1000 Da) ranging from environmental contaminants, such as pesticides, heavy metals, dyes, and mycotoxins, to registered veterinary drugs and banned substances. Both the parent compound and its metabolites can occur in foodstuffs individually or as multicomponent mixtures with enhanced adverse effects on public health. For these reasons, in the past few years, the most relevant LC–MS applications have especially focused on the development of multiclass residue determinations. The regulatory agencies of many countries have been establishing restrictive food control measures to protect consumer health. In this respect, the EU has pursued a very tough policy and issued several regulations and directives, and maximum residue limits (MRLs) have been set both for the allowed veterinary drugs and for mycotoxins and other contaminants in foodstuffs (Commission Regulation (EC) No 1881/2006), whereas the use of hormones and other performance enhancers for animal fattening have been prohibited (Commission Regulation (EU) No. 37/ 2010). Instead of the standardized methods, criteria and procedures to develop novel analytical methods (Commission Decision 657/2002/EC and its implementation) have been laid down so as to ensure flexibility and ready adaptation to technical developments, useful to face new emerging problems efficiently. 6.3.1 Recent Trends in Sample Preparation for LC–MS Analysis

The latest trends in preparing food samples for LC–MS analysis are geared for generic and nonselective sample preparation procedures in order to maximize the number of analytes belonging to different toxicant families. Recovery efficiency can be low, but these multicomponent protocols have to be time-/costeffective, simple, and with high sample throughput. Basically, procedures with these characteristics can be categorized in three groups: solvent extraction (SE), solid-phase extraction, and the QuEChERS methodology. SE has been applied to

117

118

6 LC–MS Applications in Environmental and Food Analysis

isolate more than 100 substances (up to 350 toxicants), among them antibiotics, food additives, mycotoxins, pesticides, sedatives, and growth promoters, from different kinds of foods with a minimum sample preparation without further purification [29,30]. Conventional SPE is the most used technique, but it is less suitable than SE for comprehensive multiresidue methods (up to 50 veterinary drugs) [31]. The latest trends show a growing use of the online SPE–LC–MS to save time and solvents [32]. Dispersive SPE (dSPE) is a recent variant based on the uniform dispersion of the sorbent in a sample solution/suspension that has been applied for single-class [33] and multiclass determinations of more than 100 veterinary drugs [34]. Molecularly imprinted SPE (MISPE) is a selective technique that can be compared with synthetic antibodies assays [35]. For this reason, these materials are mainly suitable for the single-class determination of a limited number of toxicants [36]. QuEChERS (quick, easy, cheap, effective, rugged, and safe) is a two-stage process, combining SE with dispersive SPE. This extractive methodology allows preparing many samples in few minutes and extracting more than 200 different compounds (above all pesticides and veterinary drugs) with good efficiencies and repeatability [37–39] Matrix solid-phase dispersion (MSPD) is a flexible one-step sample treatment useful for extracting/ purifying contaminants from a variety of solid, semisolid, viscous, and liquid foodstuffs. It has been applied for the LC–MS analysis of pesticides, veterinary drugs, persistent environmental chemicals, naturally occurring toxicants, and surfactants in several food matrices [40]. To achieve a faster and more efficient extraction of target compounds, MSPD with heated and/or pressurized solvents has also been proposed with pressurized liquid extraction (PLE) [41]. TFC is a compromise between size exclusion and chromatographic adsorption that has been devised as online automated system for achieving high sample throughput. Notwithstanding its potentialities, its use in food analysis is rare and so far has been limited to the analysis of veterinary drugs (up to 40) [42]. 6.3.2 Recent Trends in LC–MS Screening Analysis

LC coupled to LR- or HR-MS has been emerging as an alternative technique to microbiological inhibition assay and immunoassay due to its capability to perform large-scale screening of more than 100 residues in a single chromatographic run (up to 300–500 substances in ∼5 min). QqQ instruments operating in neutral loss scan (NLS), precursor ion scan (PrIS), or SRM have been used to carry out pretarget screening methods, employing authentic standards for the preselection of ion currents [34]. Martínez Vidal et al. [43] developed two multiclass screening methods for the rapid identification of 21 veterinary drugs in milk. The method was based on NLS and PrIS and allowed a rapid identification of residue families in real samples but with high cutoff levels. The other method used one SRM transition per compound and was more suitable for screening purposes at low concentration levels. Although a limitation of QqQ mass spectrometers is the number of analytes that can be monitored per injection, their

6.3 Food Toxicant Applications

ability to detect residues not known a priori is the main reason why full-scan HR mass spectrometers are becoming more and more attractive [44]. TOF and orbitrap are especially used to perform post-target and nontarget screening analysis. The highest resolving power of orbitrap (up to 100 000 FWHM) is important to avoid both false positives and false negatives, while the product ion data acquired by QqTOF hybrid instruments are useful to identify compounds belonging to a specific class. Gómez-Pérez et al. [29] have created a database for the simultaneous analysis of more than 350 pesticides, biopesticides, and veterinary drugs using UHPLC–orbitrap MS. The developed database includes exact masses and retention times of the target ions, allowing their automated search, and then, the quantification of the detected compounds within the same run (see Figure 6.2). Generic chromatographic and MS conditions have been used, so that new compounds can be included and the database can be easily upgraded. In the similar way, but using a LC–QqTOF-MS, Turnipseed et al. [45] collected exact mass data for approximately 200 veterinary drugs. Farré, Picó, and Barceló [8] performed the screening and quantification of a large number of pesticides and the characterization of other several contaminants in a number of environmental and food samples by using UHPLC–LIT-orbitrap MS. The full-scan MS data

Figure 6.2 Workflow scheme used to create and apply the database to detect and identify the analytes in the samples. (Reproduced with permission from Ref. [29].

119

120

6 LC–MS Applications in Environmental and Food Analysis

were acquired for quantification, while data on MS2 and MS3 product ion spectra were obtained for identification and/or confirmation. In addition to target analytes, this method enabled the simultaneous detection/identification of nontarget pesticides, pharmaceuticals, and drugs of abuse, mycotoxins, and their metabolites. 6.3.3 Recent Trends in LC–MS Confirmatory Analysis

According to the criteria of the Commission Decision 657/2002/EC, an indispensable requirement to perform confirmation analysis is to provide a clear identification of the target analytes. To date, the QqQ analyzer running in SRM mode is still the most used chromatographic detector to perform reliable quantitative multiresidue analysis due to its high duty cycle in this scan mode, high sensitivity and selectivity, robustness, extended linear dynamic range, and low cost. Since this analyzer does not record a full mass spectrum, at least two SRM transitions have to be selected in order to achieve a number of identification points (IPs), enough to confirm both group B (IPs  3) and group A (IPs  3) substances. Following the recent trends in LC–MS, several methods based on UHPLC and fast scanning mass spectrometers have been developed to separate different kinds of toxicants in short run times (less than 10 min) with high efficiency and sensitivity. A representative example is the UHPLC–QqQ method developed by Gómez Pérez et al. [39] for the determination of 17 veterinary drugs (macrolides, sulfonamides, and anthelmintics) in cheese in less than 9 min. In recent years, some methods have been developed using UHPLC coupled to polarity switching QqQ spectrometers to perform the simultaneous detection of positive and negative ions in a single chromatographic run. Leandro et al. [46] used an UHPLC–QqQ system for the determination of 52 pesticides in cerealbased baby foods, oranges, and potatoes. The UHPLC separated all of the pesticides, including the structural isomers butocarboxim sulfoxide and aldicarb sulfoxide, whereas the dual polarity detection enabled the determination of 44 compounds in the positive ionization mode and 8 compounds in the negative ionization mode in a single run. Whelan et al. [47] analyzed 38 anthelmintic drug residues (benzimidazoles, avermectins, and flukicides) in a 13 min run time by using an UHPLC–QqQ with fast polarity switching. Notwithstanding the high selectivity of the SRM scan mode, the unit resolution of the quadrupole analyzer cannot distinguish between the target analytes and the coeluting isobaric interferences, giving rise to potential false identifications and/or quantitation. This problem can be solved by increasing either the chromatographic or the MS resolution. The latter solution was adopted by Martínez-Villalba et al. [48] who carried out the selective LC–SRM analysis of a coccidiostatic drug and its metabolites in meat by using an enhanced resolution QqQ with 0.1 Da at FWHM. Comparing performances of QqQ, TOF, and orbitrap mass spectrometers, Kaufmann et al. [49] demonstrated that orbitrap can discern isobaric interferences better than TOF (12 000 FWHM but QqTOF can reach 50 000) due to

6.4 Foodomics as a Recent Approach Embracing Metabolomics, Proteomics, and Lipidomics

its superior mass resolution. Notwithstanding the potentialities of such instruments, LC–HR-MS still remains exceptional for target quantitative approaches. Vanhaecke et al. [50] have demonstrated that steroid analysis based on orbitrap, operating at a resolution of 50 000 FWHM, can compete with QqQ in terms of selectivity and linearity. Nevertheless, QqQ-MS has been proved superior in terms of precision (10) and low ( 4, where both silanols and basic compounds are partially or totally ionized. Cationic compounds show broad tailing peaks for packings containing appreciable amounts of free silanols (Figure 8.2) [25,26]. This behavior has been explained by both the wide range of polarities of silanols and the slow sorption–desorption kinetics of the cationic solutes on silanols. The problem is of considerable concern, since a large number of compounds of biomedical, pharmaceutical, and environmental significance have a basic character. The inhibition of silanol activity by using an acidic mobile phase at pH close to 3 is a usual practice. The alternative is to select a modern high-purity silica packing having less acidic silanol groups. Along half a century, the evolution of silica from type A first-generation to type B packings has been boosted by the need of improving the efficiency in the separation of basic compounds [27,28]. Type A packings of the first generation were synthesized with silica of lesser purity and showing less batch-to-batch reproducibility. The presence of trace metal impurities in these packings, such as aluminum, iron, calcium, potassium, and sodium, increased significantly the acidity of silanols (from pKa = 6.2–6.8 to 3.5–4.6), thus enhancing silanophilic interactions with greater peak tailing for basic compounds [22]. In the second generation

8.2 The Stationary Phase

CA

(a) ACE

MET CE

0

10

20

CA ACE

0

1

MET

2

(b)

CE

3

CA

4 (c)

MET ACE CE

0

1

2

3 (d)

CA

MET ACE CE

0

1

2

4

3

4

Time (min) Figure 8.2 Chromatograms for a mixture of basic compounds (β-blockers) eluted with 30% v/v acetonitrile using: (a) Lichrospher, (b) Kromasil, (c) Zorbax Sb, and (d) Zorbax XDB columns. Although all are C18 columns, their

chromatographic performance is significantly different from each other. Compounds: CA, carteolol; ACE, acebutolol; MET, metoprolol; and CE, celiprolol) [25].

165

166

8 Reversed Phase Liquid Chromatography

of type A packings, the silica support was chemically deactivated by washing with acids, covering or encapsulating the silica surface with a polymer or other proprietary methods addressed to shield the surface or remove the metal impurities. The third generation of silica packings (type B), developed in the late 1980s and early 1990s, show a more homogeneous distribution of silanols, higher bonding density, and better batch-to-batch reproducibility than type A silicas. Also, type B silica contains only trace amounts of metal impurities ( 6, since the hydrophobic stationary phase minimizes dissolution of the silica support. Unfortunately, the absence of data on the amount and accessibility of the surface silanol groups does not allow, in many instances, the selection of an RPLC stationary phase to meet the demands of a particular problem, although several tests have been developed to evaluate the silanol activity [26,30]. Manufacturers have invested considerable resources in attempts to reduce further the density of the undesirable accessible silanols. For this purpose, new bonding procedures or alternative end-capping reagents have been described. Thus, a densely covered hydrophobic surface is obtained by the combined use of

8.3 The Mobile Phase

a silanizing reagent with longer R´ groups than methyl (as hexyl), followed by end-capping with C1 groups. In another approach, a polar functional group (an amide, carbamate, or carbamide, among others) is incorporated into the packing, close to the silica surface, to help in repelling the basic species from the surface without hindering their interaction with the upper parts of the bonded phase. The polar group may be either linked to the alkyl ligand (embedded polar group or EPG stationary phases) or added to the end-capping reagent after the main derivatization reaction (polar end-capped stationary phases) [31]. These phases are referred to by trade names that suggest “polar” or “water-compatible” phases. The presence of the polar group close to the surface may effectively suppress the influence of silanols on the retention of basic solutes, also allowing the work with highly aqueous mobile phases (see Section 8.6.5.4). Thus, these phases also exhibit different selectivities from classical C8 and C18 packings, while having comparable run times (as evaluated by the retention time of the same highly retained solutes). Finally, a well-established dynamic approach is to block the silanols with reagents (silanol blockers or masking agents), such as amines (as triethylamine and triethanolamine). These reagents added to the mobile phase, at low concentration, interact electrostatically with ionized silanols [7]. Other additives, such as anionic surfactants, associate with the bonded alkyl chains, effectively masking the silanols by restricting the accessibility [25]. Silanol blockers may have, however, some undesirable effects, such as the difficulty of removal from the stationary phase after use, chemical reaction with some functional groups, and the generation of background noise in evaporative light-scattering and spectral interferences and ion suppression effects in mass spectrometric detection.

8.3 The Mobile Phase 8.3.1 Mobile Phase Components

The second major contributor to the RPLC environment is the mobile phase, which has a considerable influence on the separation of solutes. Usually, binary mixtures of water and an organic solvent are used, but ternary or quaternary mixtures of water and two or three organic solvents, respectively, can be an option to control both elution strength and selectivity. The selection of the mobile-phase composition, including solvents, buffer, and other additives, will determine the degree of interaction between the solutes and the stationary phase. Several approaches can be followed to optimize all these factors in order to succeed in the separation (see Chapter 9 on “Modeling of retention in RPLC”). In principle, there is a wide range of water-miscible organic solvents that can be used as modifiers; however, only a few of them are used in RPLC, namely, acetonitrile, methanol, ethanol (a greener solvent), tetrahydrofuran, and isopropanol [32,33]. Among them, acetonitrile followed by methanol are by far the most frequently used. Tetrahydrofuran is sometimes added in proportions of

167

168

8 Reversed Phase Liquid Chromatography

5–20% v/v along with acetonitrile or methanol to fine-tune the selectivity. It is also useful in larger proportions to elute highly hydrophobic solutes as carotenoids. If required, the separation can be enhanced by addition of buffers, weak acids, neutral salts, surfactants, and ion-pairing reagents (often as ammonium or acetate salts) [7]. The reasons that make acetonitrile the most popular choice as a modifier are its low viscosity that helps in reducing the back pressure, low cutoff wavelength for UV detection, sufficiently large elution strength, reduced reactivity, and ability to dissolve a wide range of solutes. The viscosity of methanol–water mixtures is particularly large, having a maximum at about 50% methanol, which does not occur with acetonitrile. In turn, methanol is less expensive and less toxic, and its higher polarity reduces the risk of buffer precipitation. There may be other reasons to switch to another solvent, including unavailability or legal restrictions. For instance, from late 2008 to early 2009, the production of acetonitrile came down giving rise to a substantial increase in its price [34]. As alternative, solvent consumption was recommended to be reduced by using columns with narrower diameters. Ethanol, which is less toxic than acetonitrile and methanol, was also suggested, but it is subjected to restrictions in some countries to avoid illegal diversion to human consumption. Finally, solvent recycling technologies were also given as a solution to reduce organic solvent consumption and its environmental impact. 8.3.2 Snyder’s Solvent Selectivity Triangle

The selectivity of a particular chromatographic system is a function of many intermolecular interactions of solutes with both mobile and stationary phases [33]. A common approach in method development is maintaining the nature of the stationary-phase constant, thus proceeding to adjust first the elution strength of the mobile phase and then its selectivity. The so-called solvent selectivity triangles (SSTs) can assist in this task [35] (Figure 8.3). These diagrams, which are constructed by following different criteria, have in common a simple idea: illustrating the multiple interactions between solutes and mobile phase, whatever their nature, by means of only three parameters that describe the properties of the solvent mixture. Since SSTs focus on the solvent properties, while ignoring the solute properties, they are rather rough descriptions of the chromatographic behavior. However, they are useful for qualitative predictions. To construct SSTs, it is assumed that the forces that hold the solvent molecules together are of the same nature and similar intensities as those maintaining the solute molecules in solution. These forces are mainly proton donation (acidity) and acceptance (basicity), interactions between permanent dipoles (polarity), induced dipoles (dipolarity), and permanent and induced dipoles (polarity and polarizability), to which electrostatic interactions should be added [33]. The energy required to create the cavity where the solute molecule is located should also be considered. As indicated, to construct SSTs, all these forces and the

8.3 The Mobile Phase

0.2

0.6 isopropanol

xe

ethanol 0.3

methanol

pyridine

0.4

acetic acid

xd

0.4

DMF

water

0.5 0.2

0.5

DMSO

tetrahydrofuran 1,4-dioxane acetone

acetonitrile 0.3

0.5

0.4

DIPOLARITY Figure 8.3 Snyder’s selectivity triangle for water and water-miscible solvents. The arrows starting at the methanol location show how to read the scales. The area highlighted in gray

xn

0.6

0.3

represents the selectivity region covered by the four most common solvents used in RPLC. Acronyms: DMF, dimethylformamide; DMSO, dimethylsulfoxide.

involved energies should be summarized in only three parameters. The absolute values of these forces and associated energies are not represented in the diagrams, as only a picture of their relevance relative to each other is intended. In this way, SSTs will show the selectivity or “character” of the addressed solvent or solvent mixture, rather than its elution strength. Among the proposed approaches to construct SSTs, the original semiempirical method developed by Snyder is still the most useful [36]. According to Snyder, the selectivity of a solvent is characterized by three parameters which mainly represent its acidity, basicity, and dipolarity (which embraces both polarity and polarizability). These three properties are estimated by measuring, in closed vials where a given volume of the addressed solvent has been introduced, the gas– liquid distribution of three probes: ethanol, 1,4-dioxane, and nitromethane. These substances mainly represent the proton donor and acceptor capabilities, and the dipolarity of solutes, respectively. None of the three probes represent these properties uniquely (e.g., ethanol is not only predominantly a proton donor but also a weak acceptor and a moderate dipole), but they are different enough to give rise to a diagram with a remarkable capacity of distinguishing among the “character” of the solvents. Thus, a solvent which strongly retains ethanol (predominantly a proton donor) reveals its strong basic character, and a solvent which strongly retains 1,4-dioxane (predominantly a proton acceptor) has a strong acidic character. Finally, a solvent which strongly retains nitromethane is strongly dipolar. The expected behavior of solutes with known properties can also be deduced from the diagram.

169

170

8 Reversed Phase Liquid Chromatography

The Snyder’s SST diagram is built with three normalized indices for the addressed solvent: xe ‡ xd ‡ xn ˆ 1

(8.1)

where xe (ethanol index), xd (1,4-dioxane index), and xn (nitromethane index) represent the basic, acidic, and dipolar character of the solvent, respectively. For a given solvent, the three indices give rise to a point in the diagram. In Figure 8.3, the properties of a series of common water-miscible solvents, compatible with RPLC, have been plotted. The internal shadowed area results from joining the coordinates for the four most common RPLC solvents (i.e., water, acetonitrile, methanol, and tetrahydrofuran), thus representing the character of all possible mobile phases that can be prepared by mixing them. Water is both a strong donor and a strong acceptor of protons, but it is a weak dipole, and accordingly, it is located at half-height left position in the SST. Methanol is predominantly basic, and then it is in an upper location in the SST, whereas acetonitrile is more acidic and predominantly dipolar. Tetrahydrofuran is as dipolar as acetonitrile, but as water, it has a more balanced acidity–basicity character than methanol and acetonitrile. As already explained, the SST indicates only the character of the solvent mixtures and not its elution strength, which increases when water is partially substituted by an organic solvent. Thus, for example, for mixtures having the same elution strength (equieluotropic), the basicity of the mixture increases and the acidity and dipolarity decrease if acetonitrile is substituted by methanol. This speeds up the elution of acidic solutes and increases the retention of basic solutes. Substitution of acetonitrile by tetrahydrofuran yields a similar but smaller effect on the selectivity. Thus, when a solvent is substituted by another without increasing the elution strength, the expected changes in the relative retention of solutes of well-known properties can be predicted from the character of the solvents, and vice versa, the changes observed in the retention can be used to deduce solute properties. 8.3.3 Control of the Mobile-Phase pH

The mobile phase pH is an important parameter that determines the chromatographic retention of solutes with acid–base properties (ionizable solutes), among which many compounds of pharmaceutical and biological interest are found. As explained, column stability should be considered when manipulating the pH, since the working pH of conventional columns is usually in the 2.5–7.5 range. Outside this range, important damage can be inflicted (i.e., hydrolysis of the siloxane bond below pH = 2 and dissolution of silica above pH = 8). However, innovative silica supports that contain short carbon chains between the Si atoms, as well as protecting polymer layers, for which the range can be extended to pH = 1–12, have become commercially available [8]. Manipulation of the pH to change the selectivity for mixtures containing ionizable solutes is described

8.3 The Mobile Phase

in Chapter 9. We will discuss here only the methodology required to adjust the pH of the mobile phase. Common buffers are based on the phosphoric (with useful pKa values of 2.15 and 7.2 at 25 °C), citric (3.1, 4.7, and 6.4), acetic (4.8), trifluoroacetic acid (0.5), formic (3.8), and ammonium (9.25) acid–base systems. Phosphoric and citric buffers, which provide control over wide pH ranges, are the most popular. Their main disadvantage is that their inorganic salts may precipitate inside the column if the proportion of modifier is too high, particularly with acetonitrile or at low column temperatures. Phosphoric and citric salts are also nonvolatile and, therefore, noncompatible with evaporative light scattering and mass spectrometric detectors. Only acetic, trifluoroacetic, and formic acids and their ammonium salts should be used to prepare buffers compatible with these detectors. However, trifluoroacetic acid reduces the sensitivity of mass spectrometric detection, particularly when working in the negative ion mode. Analysts often do not clarify in their reports how they measure the pH of the mobile phase, without being aware that there are several alternative procedures. The pH is obtained by measuring the electromotive forces, Ex , provided by a potentiometric cell that contains a pH-sensitive electrode. Cell calibration is performed by using the IUPAC’s operational definition of pH, which can be reduced to pHx ˆ A

B Ex

(8.2)

During calibration, the two constants, A and B (which vary with temperature), are established. For this purpose, two reference buffers are measured. Cell calibration is most conveniently performed by adjusting the electrical zero and gain of the millivoltmeter, which will directly provide the reading in the pH scale. This is done by adjusting the zero (A) with a buffer of pH ∼ 7 (such as an equimolar mixture of KH2PO4 and Na2HPO4, with pH = 6.865 at 25 °C), and the slope (B) with a standard buffer in the acidic or basic regions, depending on the pH that is subsequently measured. However, standard reference buffers should be prepared in pure water, whereas mobile phases for RPLC are most frequently aqueous-organic mixtures. The presence of the organic solvent may significantly modify the pH scale. Thus, to maintain the accuracy when measuring the pH of the aqueous-organic mixtures, three alternative procedures can be used [37]: w w pH

scale where cell calibration is performed with reference aqueous buffers, followed by the measurement of pH in the aqueous buffer of the mobile phase before adding the organic modifier; ii) ss pH scale where cell calibration is performed with reference buffers prepared with the aqueous-organic mixture of the mobile phase, followed by measurement of the pH in the mobile phase; and iii) sw pH scale where cell calibration is performed with reference aqueous buffers, followed by the measurement of pH in the aqueous-organic mobile phase. i)

171

172

8 Reversed Phase Liquid Chromatography

The first procedure (ww pH) is the most simple and may be adequate when the pH should not be strictly controlled, although the pH modification the organic modifier may produce in the mixture should be still considered. This method is the most usual for repetitive procedures where a specific buffer and the same mobile-phase composition are routinely used. The second procedure (ss pH) is the most accurate. Note that the ss pH scale is directly related to the thermodynamic acid–base constants of ionizable solutes in the aqueous-organic mixture (which are usually only established in water). Therefore, the preparation of reference buffers in a variety of mobile phases of different composition is problematic. Reference buffers in a variety of aqueous-organic mixtures are not needed by using the third procedure (sw pH). Fortunately, the sw pH scale can be converted to the ss pH scale through a simple correction parameter: δ ˆ sw pH

s s pH

(8.3)

Even if the δ-value is not known, Equation 8.3 indicates that using the sw pH scale, all pH readings will be displaced with the same unknown but constant amount, relative to the ss pH values at a specific mobile-phase composition.

8.4 Temperature as Chromatographic Factor

Many separations in RPLC are carried out at room temperature for convenience. However, temperature is an important factor for accurate method development, and thus, it should be controlled. An accurate control of temperature is necessary to achieve reproducible retention and peak profile [9]. Temperature can also play a role in reducing the run time, controlling the selectivity, improving the efficiency, and consequently, enhancing the detection sensitivity [9,38]. The use of temperatures of up to 100 °C can have noticeable benefits. Thus, retention usually decreases at increasing temperature, which can be used to accelerate the separations [39,40] (see Chapter 9). For solutes experiencing slow mass transfer kinetics, working at a higher temperature can improve the efficiency [41]. Temperature may also affect the separation selectivity in a number of situations, for example, i) when the relative retention of the solutes is sensitive to changes in the conformation of both the solute itself and the stationary phase, as temperature is varied (case of synthetic polymers and biomolecules); ii) when the relative molecular size or shape of the solutes differs with temperature, leading to entropic discrimination; iii) when the solutes have functional groups with different dependence of retention on temperature; and iv) for partially ionized solutes with different dependence of their acid-base constants (pKa values) with temperature.

8.4 Temperature as Chromatographic Factor

High-temperature liquid chromatography (HTLC) often refers to any separation carried out above room temperature, typically within a range from 40 to 200 °C [40]. The development of HTLC has been driven by the need for faster analysis, the reduction of mobile-phase viscosity to decrease column pressure, the search for improved selectivity and efficiency, and the reduction of the modifier content in the mobile phase. HTLC is often combined with ultrafast HPLC (UHPLC) in the separation of complex mixtures, in order to achieve faster separations without losing resolution (although in some cases, there is some loss). Most column compartments are designed for temperature control and programming up to 100 or 150 °C, although commercial equipment prepared to reach 200 °C and higher is also available. For high working temperatures, there is a compelling need of both preheating the mobile phase and cooling it down again before the detector cell. Without preheating, the entrance of the cool mobile phase into the warm column would create a radial temperature gradient that deteriorates the efficiency. The use of temperatures above 100 °C is at present also restricted by the availability of thermally stable stationary phases. The lifetime of conventional stationary phases is largely reduced at 60 °C, particularly if pH is close to the limits of stability of the siloxane bond. Therefore, the reluctance of chromatographers to work at high temperature is mainly based on the possible reduction in the number of runs that can be performed on a given column, without any noticeable damage. An additional unfavorable characteristic is the complexity of the changes involved by the variation of numerous relevant physicochemical parameters with temperature. Finally, the role of water as RPLC eluent at high temperature, where it becomes less polar, should be commented [42]. The elution strength of water increases with temperature, also involving selectivity changes. The decrease of water polarity as temperature increases has an effect similar to the increase of the organic modifier content in the mobile phase at constant temperature. The retention decreases to a greater or lesser extent depending on the solute molecules, due to entropic, steric, conformational, and ionization effects. The approach has allowed the development of water-based more environmentfriendly or “greener” RPLC procedures, although at the cost of the additional energy needed to maintain the oven temperature, and the preheating and cooling systems. The technique has been called either “chromatography with pressurized water”, “superheated water chromatography”, or “subcritical water chromatography”. In addition to the attractive advantage of the very low toxicity, the use of pressurized water as mobile phase offers the possibility of hyphenation to low-wavelength UV, flame ionization, inductively coupled plasma mass spectrometry, and nuclear magnetic resonance (NMR) spectroscopy using deuterium oxide as mobile phase. Unfortunately, the elution strength of water is not sufficiently high below 200 °C, which in most cases hinders total replacement of organic solvents by water. Thus, the polarity of water at 150 °C is approximately the same as for an aqueous mobile phase containing 50% v/v methanol at room temperature, for which nonpolar solutes can be still excessively retained. Further reduction of

173

174

8 Reversed Phase Liquid Chromatography

water polarity can be achieved above 200 °C, but the choice of suitable stationary phases with adequate stabilities is rather short. Another problem is the aggressiveness of very hot pressurized water (i.e., high acidity and dissolving power for metals). The development of more stable stationary phases and advances in instrumentation are necessary before HTLC moves from the domain of academic research into routine analysis.

8.5 Gradient versus Isocratic Elution 8.5.1 Solute Retention and Peak Width

RPLC is the technique of choice for the separation of complex mixtures of solutes within a wide range of polarities. Retention in RPLC is mainly related to solute hydrophobicity. Therefore, the 1-octanol/water partition coefficient (log Po/w) for solutes can be used to assess retention. If the solute is ionized at the mobile-phase pH, log Papp (the apparent partition coefficient, which is smaller than log Po/w) will take into account the ionization degree [43]. Separations can be carried out in either the isocratic or the gradient elution modes. In isocratic elution, the mobile-phase composition is held constant during the separation (e.g., 30% v/v acetonitrile), while in gradient elution, it is varied (e.g., changing from 10 to 70% v/v acetonitrile). Isocratic and gradient elution are the same processes viewed from different perspectives. This allows to transfer some of our intuitive understanding of isocratic separations to gradients or vice versa [44,45]. The peaks occur in the same order in both the isocratic and the gradient elution modes. Also, the absolute peak broadening increases with the retention time: the earliest peaks are the narrowest. In both cases, the retention is controlled by varying the composition of the “weak” solvent A (water) and “strong” solvent B (the organic modifier). In isocratic elution, a decrease in the constant percentage of solvent B (φB) results in longer run time, with resolution improvements. Similar effects will be observed in gradient elution by decreasing either the average φB or the gradient slope. The peaks will be broader, resulting in peak height drops with poorer detection sensitivity. Solute retention within the retention factor range 1 < k < 10 (although k = 20 may be still acceptable) should be procured to achieve resolution within reasonable run times. As explained below, gradient elution helps in avoiding peaks with either too low or too large k-values within the same chromatogram. Meanwhile, a change in flow rate will expand or compress the isocratic or gradient elution chromatogram, having little (if any) effect on the selectivity. However, the flow rate significantly influences the efficiency (and therefore, the resolution). According to the Van Deemter equation, the efficiency will improve by increasing the flow rate at low values, but after it reaches the maximum, the efficiency will decrease for higher flow rates.

8.5 Gradient versus Isocratic Elution

8.5.2 Isocratic Elution

Due to its inherent simplicity, lower cost, and higher robustness, many RPLC separations are carried out in the isocratic mode [46]. This approach is suitable when the chromatogram contains a few peaks for solutes within a small or moderate polarity range. In such cases, all solutes in the sample can be separated over a reasonable time period (i.e., within the optimal k-range). However, isocratic elution is not recommended if the polarity of the solutes in the sample spreads over a wide range. In this case, if the most retained solutes are eluted within the optimal k-range (at high φB), the early eluted peaks will be poorly resolved and even lost in the solvent front. On the contrary, if the less retained solutes are well resolved (at low φB), then the most retained solutes will be eluted with too large k-values, that is, with excessively long run times and broadened peaks; thus, with a reduced detection sensitivity. Some highly retained solutes may even be buried into the baseline noise or overlap the following sample. Therefore, it is not possible to improve both extremes of the chromatogram at the same time by using isocratic elution. This situation has been called the “general elution problem of RPLC”. 8.5.3 Gradients of Modifier: The Usual Solution for the General Elution Problem

As described, a mobile phase with fixed composition is often not suitable to satisfactorily resolve complex samples with ratios over ∼20 between the k-values of the last and first eluting solutes. The usual solution to the general elution problem is the application of gradient elution with programmed changes of organic modifier [47–50]. The main objective of gradient elution is to achieve adequate resolution, and acceptably short times during a single run, by increasing the retention of the poorly retained solutes and speeding up the elution of those strongly retained. For this purpose, the elution strength of the mobile phase should be initially low and become steadily stronger by increasing the concentration of the modifier as the separation progresses. The effect of gradient elution on retention times can be graphically expressed as progressively contracting an elastic rubber hold by the starting extreme to a fixed location. As the concentration of the organic modifier is increased, the polarity of the mobile phase decreases. In mobile phases with low elution strength, the strongly retained solutes are stuck in a narrow band near the head of the column, migrating only very slowly, so that this range of mobile phase compositions does not contribute significantly to their elution. As the elution strength of the mobile phase increases, the rate of partitioning into the mobile phase also increases and the solutes are “accelerated” through the column. At some point within the column, a given solute may partition mostly into the mobile phase, and will move almost with the same linear velocity as the mobile phase. A simple and graphical image has being given for this behavior: “a solute sits at the head of a column

175

176

8 Reversed Phase Liquid Chromatography

until a strong enough solvent comes along to push it through the column leaving the other solutes behind, then it travels to the column outlet fairly quickly” [51]. The point at which this occurs depends on the strength of solute interaction with the mobile and stationary phases. Solutes in gradient RPLC seldom experience the whole range of mobile-phase compositions. The fraction of the solvent composition range that actually affects solute migration has been called “significant solvent concentration range” [52]. Gradient experiments are often used in a scouting technique to obtain information for selecting isocratic separation conditions. From the above discussion, it can be concluded that gradient data can be transformed into accurate isocratic retention factors only in a narrow concentration range, and the concentration dependence of the extrapolated values in a broader concentration range may be subjected to significant errors. Isocratic measurements are more laborious but generally provide more reliable information for selecting optimal isocratic separation conditions. 8.5.4 Development of Gradients of Modifier

The simplest and most usual approach to gradient elution makes use of binary eluents of water and modifier, whose concentrations are varied linearly according to a program. Using a linear single ramp to optimize the resolution within the shortest run times, three variables should be controlled: the initial (φ0) and final (φ1) modifier concentrations, and the total gradient time (tG), or alternatively, the gradient slope (m = (φ1 φ0)/tG). The retention of the first and last eluted solutes is taken into account to establish φ0 and φ1. The gradient is usually stopped right after the elution of the most strongly retained solute [53]. Plots of log k (or ln k) against the modifier content are conventionally used to illustrate the isocratic retention behavior. Similarly, for linear gradients with different slopes, plots of ln kg (kg being the retention factor in gradient elution), as a function of the initial modifier concentration in the gradients, offer a convenient representation of gradient elution (Figure 8.4) [45]. In these plots, the upper line depicts the isocratic elution for which the gradient slope is null (m = 0), whereas the bottom line describes the retention with the steepest gradient. Since these diagrams define a triangular region, they have been called “triangular elution plots”. The triangle sides delimit all possible values of retention factor for a given solute and experimental design, with either isocratic or gradient elution. The effect of the gradient is larger for the most retained solutes, and practically null for those weakly retained (Figure 8.4b). Therefore, the left side of the triangle is wider for the former. Besides simple linear gradients with binary eluents, other formats have been developed to optimize gradient elution: multicomponent gradients (ternary, quaternary, etc.), segmented gradients, which can be multilinear, curved or stepwise, and the so-called “relay gradients”, which is a special type of segmented gradient where the nature of the modifiers is changed between segments. In a ternary gradient, the concentrations of two strong solvents (such as acetonitrile

8.5 Gradient versus Isocratic Elution

5

(a) m=0

ln kg

4

3

m = 0.2 m = 0.4 m = 0.6 m = 0.8 m = 1.0 m = 1.2

2

1 20

30

40

50

60

m = 1.2

5

(b)

m=0

4

ln kg

3 2 1 0 -1 20

30

40

50

60

Initial acetonitrile, % v/v Figure 8.4 Retention behavior, expressed as logarithm of the retention factor in gradient elution, at different initial concentrations of acetonitrile and gradient slopes (change in acetonitrile percentage per min, m):

(a) xipamide, (b) the diuretics trichloromethiazide, furosemide, and ethacrynic acid (from top to bottom). For isocratic elution, m = 0. (Reproduced with permission from Ref. [45].)

and methanol in water) change simultaneously, often following linear functions [47]. Three types of ternary gradients can be generated: i) “elution strength” gradients, where the concentration ratio for the two strong solvents is kept constant to maintain constant selectivity, but the sum of their concentrations increases linearly; ii) “selectivity” gradients, where the sum of the concentrations of the two strong solvents is kept constant, but their concentration ratio changes linearly to improve the separation of closely eluting solutes; and

177

178

8 Reversed Phase Liquid Chromatography

iii) mixed ternary gradients, with simultaneous change of both the ratio and the sum of the concentrations of the two solvents, thus to take advantage of the simultaneous modification of both the elution strength and the selectivity. In multilinear gradients, several gradients of different slopes are sequentially combined [54]. These gradients are particularly useful to slow down the elution of hard to resolve pairs or groups of solutes, while speeding up the elution within those regions in the chromatogram having well-resolved solutes or with empty gaps between solutes. Nonlinear gradients with concave or convex profiles are occasionally applied when dealing with multicomponent samples requiring extra resolution [47]. In practice, curvilinear gradients are usually substituted by gradients with many small linear consecutive segments, where the slope is stepwise increased. The changes in mobile-phase composition may also be carried out in sequences of a few isocratic steps of increasing modifier concentration with equal or different lengths (stepwise gradients). Gradients include often isocratic hold periods, at the beginning and/or the end of the runs, or inserted between linear gradient segments. The rarely used reverse gradients (with decreasing modifier concentration) can be useful in some cases. Real gradients are never perfectly smoothed variations of the mobile-phase composition. In fact, they are actually constituted by small juxtaposed segments of nonperfectly mixed solvents, where the modifier concentration increases by following a slightly oscillating imitation of the programmed ramp. This is because continuous gradients are produced in a gradient mixer that approximates the nominal changes in composition. The accuracy of the actual mixture with respect to the programmed composition depends on the pump module and mixer designs. Owing to the difficulty in mixing pure solvents, smoother gradients are obtained with premixed solvents. For instance, by mixing water containing at least 5% v/v acetonitrile with acetonitrile containing at least 5% v/v water, a much smoother gradient is achieved than by mixing the two pure solvents. The mobile phase leaving the gradient mixer needs some delay time (“dwell time”, tD) to reach the column inlet. This delay is related to the “dwell volume”, which is the system volume that should be refilled for the mobile phase to arrive to the column inlet [55]. The dwell volume is constituted by the inner volume of the gradient mixing chamber, connecting tubing, frits, and usually also the internal volume of the injection device (including the injection loop volume). At the start of the gradient, the dwell volume contains the mobile phase at the initial gradient composition, which is delivered before the gradient front, so that the sample is initially exposed to isocratic elution. In conventional RPLC, the influence of the dwell volume on solute retention is small but not negligible, having greater relevance in capillary- and nanoflow RPLC. This topic is further commented in Section 8.5.5. Finally, other important time parameters to consider in gradient elution, after a sample run, are related with purging and column re-equilibration. Purging is necessary to elute strongly retained components of a previous sample. This is

8.5 Gradient versus Isocratic Elution

usually achieved by finishing the gradient with a high slope segment, thus to reach a high φ1 value. An isocratic hold for a few minutes can also be added after the high slope segment. In modern instruments, returning to the initial gradient composition is programmed to occur very rapidly. However, before injection of the following sample, a time for column re-equilibration with the mobile phase at the gradient initial composition is required. Without re-equilibration, solute retention and peak shape will not be reproduced. Manufacturers recommend between 5 and 10 column void volumes for proper re-equilibration, but smaller or larger mobile-phase volumes are needed depending on the stationary-phase nature and mobile-phase composition. A reproducible and stable back pressure under the initial conditions is an indication that equilibrium between the mobile and the stationary phases has been reached, but the ultimate criterion is attaining reproducible solute retention and peak shape. 8.5.5 Strengths and Weaknesses of Gradients of Modifier

Gradient elution reduces the distance between the peaks of solutes of progressively decreasing polarity, which become more regularly spaced in the chromatogram. The retention times are shorter, the peaks are appreciably narrower, and for peaks of the same height, the widths are also approximately the same. This means taller peaks and lower detection limits. Also, in gradient elution, the “peak capacity” (number of peaks that can be resolved in a certain time range) is considerably larger with regard to isocratic elution. Finally, at the end of the gradient, the mobile phase is usually strong enough to remove all sample impurities, which otherwise may remain retained, thus interfering with the chromatograms of the following samples. For all these reasons, gradient elution is preferred for the separation of many samples, being widely used in RPLC. It is also recommended during the initial stages of method development, when the retention of solutes and even the composition of the sample are not known. However, gradient elution is not always the most convenient approach: an isocratic separation may be a better choice when the range of polarities of solutes is not too wide. It should be noted that the maximal resolution capability of the system is usually obtained with an optimized isocratic mobile phase and that it will be difficult to reduce the run time if the critical peak pair (that one offering the lowest resolution) corresponds to the two most retained solutes. Gradient elution also requires some additional time for column re-equilibration between subsequent runs. Depending on the nature of the mobile and stationary phases, the re-equilibration time can range from several minutes up to half an hour (or longer), which would call into question the advantage of the reduction of the run time by using gradient elution. Another drawback is that gradient elution is not compatible with some detectors, such as the refractive index and most electrochemical detectors, which need stable conditions. Also, with UV detection, high-purity “gradient grade” solvents and mobile-phase

179

180

8 Reversed Phase Liquid Chromatography

additives having negligible concentrations of UV-absorbing impurities are required to reduce baseline drift [56]. Below 220 nm, baseline drift caused by differential solvent absorbance can be sufficient to prevent the practical use of certain solvents, such as methanol or tetrahydrofuran. Sometimes, it is possible to compensate for differences in solvent absorptivity by adding a UV-absorbing compound to the less absorbing solvent component of the mobile phase; however, this will make the optical transmittance of the whole mobile phase to further decrease, thus increasing the baseline noise. Gradient elution needs more sophisticated and expensive instrumentation, as well as somewhat greater care than isocratic elution. The selection of an optimal gradient program is also less straightforward than the development of isocratic separations, as there are more parameters to be fixed. Another big concern about gradient methods is that they are hard to transfer, whether a published method is being reproduced, a method is transferred between laboratories, or the method is moved from one instrument to another in the same laboratory. Often, the problem can be traced to differences in dwell volume between the various HPLC systems, which will yield an offset in retention time and changes in selectivity [55]. The early peaks in the chromatogram, corresponding to the solutes partially eluted under isocratic conditions during the dwell time, will be most affected by dwell time differences between instruments. Later eluting peaks usually present offsets in time, but resolution changes are less. The undesirable effects due to dwell time differences can be masked by adding an isocratic hold at the beginning of the run. This initial isocratic period can be increased or decreased, thus to compensate for the shorter or longer dwell time of the current instrument with respect to the instrument where the method was developed. A problem with this approach is that a long isocratic hold at the beginning of the method is not always convenient or wished. A better solution for method development and transfer is to synchronize the system [57]. Synchronization requires the gradient to be started at the pump module a time after or before the injection, thus to avoid the early or late arrival of the gradient front to the injection valve. If this event is matched with the exit of the sample plug from the valve, then the effects of the dwell time will be fully removed without introducing any unwanted initial isocratic hold. System synchronization also requires switching the injection valve to bypass after exiting the sample plug and immediately before the arrival of the gradient front to the valve. Valve switching at the right moment introduces the gradient front into the column path immediately after the sample plug. For a given instrument and working conditions, simple experiments can be carried out to establish both the valve switching and gradient match times. System synchronization is of interest in conventional RPLC, particularly for method transfer, becoming extremely important in capillary- and nanoflow RPLC, especially to have control on the real elution conditions of early eluted solutes. However, not all the instrument control software systems have the necessary options to independently set the sample injection, valve switching, and gradient starting times.

8.6 Attempts to Explain the Retention Mechanisms in RPLC

8.5.6 Other Types of Gradients

Even though gradients of modifier are the most frequent in RPLC, gradients of pH may be useful for separation of weak acids or bases [58,59], although their use is still limited. Linear gradients of pH are more difficult to create than linear gradients of modifier, but they can be implemented by using universal buffers. A common universal buffer can be constructed by mixing phosphoric, acetic, and boric acids, to which increasing amounts of sodium hydroxide solution are added [60]. Universal buffers that do not interfere with biological solutes have also been described [61]. By mixing two solvents, one containing the acids and the other the strong base, the ionic strength of the mobile phase will be nearly constant all along the pH gradient. The use of a stable RPLC column over a wide range of pH is crucial for these gradients. Flow programming (increasing the flow rate of an isocratic mobile phase during the analysis) is limited by the maximal pressure allowed by the instrumentation and, therefore, the range of variation is small, except in case of using monolithic columns [62–64]. Today this type of gradient is still anecdotic. Temperature programming offers an attractive alternative to gradients of organic modifier, as they do not need complex gradient pumps, and column reequilibration after the end of the experiment requires less time [65,66]. Nevertheless, temperature programming is still rarely used in RPLC. One reason is that the effect of solvent strength on solute retention is usually more significant than the effect of temperature. Another important limiting factor is the relatively slow response of a column located in a thermostated column compartment to a temperature change. However, more sophisticated equipment has become available recently that allows both mobile phase preheating and rapid temperature ramp inside the column. This makes temperature programming more feasible, especially with microbore and capillary columns, where the small column diameter allows rapid radial heat transfer and much faster column temperature equilibration with regard to conventional analytical columns.

8.6 Attempts to Explain the Retention Mechanisms in RPLC 8.6.1 Solvent Adsorption and Partitioning in RPLC

Retention in RPLC is driven by the distribution of solutes between two immiscible phases that move in relation to each other: the mobile phase (an aqueous-organic mixture) and the stationary phase (typically alkyl chains tethered to a silica surface that contains residual silanols). Close to the molecular level, binary mixtures of water and organic solvents are heterogeneous media. Water is a highly associated solvent in which the continuous restructuring of

181

182

8 Reversed Phase Liquid Chromatography

water–water interactions constitutes a dynamic mechanism that assists in the transfer of solutes to the stationary phase, thus contributing to the driving force for retention in RPLC. Aqueous-organic mixtures also contain water–organic solvent aggregates and associated organic solvent molecules [67]. These continuously reorganized clusters are in close contact with the bonded stationary phase, where adsorption and partitioning of both solvent and solute molecules take place. Adsorption is produced at two interfaces, namely, the stationary-phase surface and the accessible sites of the silica surface, whereas partitioning is provided not only by the free spaces between the partially mobile bonded alkyl chains but also by the layers of the adsorbed solvent molecules. The mobile phase fills all volume elements of the column not physically occupied by the column packing (i.e., the solid support plus the stationary phase). This includes portions of the pore volume and interparticle volume close to the points of contact between particles in which the mobile phase is stagnant, which can be considered as part of the stationary phase, and the portion of the interparticle volume containing the streaming mobile phase. When the mobile phase first contacts the stationary phase, solvation (wetting) of the stationary-phase/ mobile-phase interfacial region occurs. An immobilized liquid film is created on the top of this region, with different composition with regard to the bulk mobile phase due to preferential adsorption of the organic component through hydrophobic interactions with the bonded phase. The volume and composition of the immobilized film varies with the composition of the mobile phase, column characteristics, and other experimental parameters. Meanwhile, water and organic solvent molecules adsorb on residual silanols. The stationary phase in RPLC represents thus a complex heterogeneous interfacial region in dynamic equilibrium with the bulk mobile phase. In other types of chromatography, the distinction between mobile and stationary phases to calculate the stationary-phase/mobilephase volume ratio is obvious. In RPLC, the phase volume ratio cannot be easily calculated, since the volumes of the two phases (especially the stationary phase) are not well defined, and this ratio changes with the composition of the mobile phase and other operating factors. 8.6.2 The Solvophobic Theory

The complexity of the environment inside an RPLC column, especially with regard to the location of the mobile-phase components, makes the underlying principles of separation to be extremely complex. The topic has been studied since the birth of the technique, with considerable experimental effort, in order to elucidate the retention mechanism(s), and thereby help in the development of novel RPLC systems. An early attempt to interpret RPLC retention is the solvophobic theory proposed by Horváth and coworkers, based on thermodynamic principles [6,68,69]. According to this theory, the stationary phase plays a passive role, and is regarded as being saturated with molecules of the organic component. Meanwhile, the water molecules in the mobile phase prefer to interact

8.6 Attempts to Explain the Retention Mechanisms in RPLC

with each other rather than with the stationary phase or the nonpolar regions of the solute molecules. In this environment, the solute is excluded from the mobile phase and driven into the layer of the organic solvent, which is adsorbed on the stationary phase. In essence, the most favorable free energy contribution driving the solute retention process is predicted to come from the solvent–solute interaction when the solute is transferred into the retentive phase. In the solvophobic theory, the transfer of a solute into a solvent is viewed as a three-step process: (i) cavity formation of a solute molecule in the solvent system, (ii) insertion into the cavity, and (iii) intermolecular interaction with the surrounding solvent, or “charging” of the solute. First, a cavity of the correct size and shape to accommodate the solute is created in the solvent, which requires “turning off” the intermolecular interactions between the solvent molecules, which is energetically unfavorable. Next, the cavity is filled with the solute, and finally, the solute–solvent interactions are “turned on”. In the process of “charging” the solute, the solvent molecules close to the solute are reorganized to some extent. This is a simple model that assumes the corresponding free energies are independent. Also, a debate has arisen whether the driving forces for the transfer of solutes from one environment to another are primarily solvophobic or lipophilic [6]. Solvophobic refers to the unfavorable interaction that solute molecules experience with the polar mobile phase, and lipophilic to the favorable interaction with the retentive phase. However, RPLC retention cannot be reduced to a single mechanism that holds for all types of solutes. Also, the solvophobic theory offers an incomplete description of the retention process. It is focused on the mobile phase, being not specifically directed to explain the dependence of solute retention on the length and density of the alkyl chains bonded to the silica surface or on the embedded polar groups. However, if a solute has a relatively large hydrophobic surface area, then different retention can be anticipated when a C4 and a C18 ligand are compared, simply because the hydrophobic contact area between the C4 ligand and the solute is smaller than that of the C18 ligand with the same solute. Equally, if the solute is a macromolecule such as a protein, then the ligand density may have a profound effect on retention, due to a difference in the number of ligands that can interact simultaneously with the macromolecule. 8.6.3 Solute Adsorption or Partitioning?

Carr, Martire, and Snyder [70] emphasized the following questions: Is retention in RPLC dominated by an adsorption or a partitioning process? Do the same general principles apply to different solutes and stationary phases? To what extent chromatographic parameters, such as mobile-phase composition and bonding density, affect the retention mechanism? Adsorption occurs when the solute lies at the interface between the bonded chains and the bulk solvent, or near the silica support when this contains residual silanols. In an adsorption mechanism on the bonded chains, the stationary

183

184

8 Reversed Phase Liquid Chromatography

phase is considered as an impenetrable barrier, and solute molecules simply migrate from the mobile phase to the interface (the adsorptive monolayer), displacing the adsorbed molecules of solvent. Partitioning (or absorption) refers to a liquid–liquid-like transfer process where the solute is embedded deep within the bonded phase, occupying a space within it. Therefore, the stationary phase is no longer a passive support [67,71]. The debate on the type of retention mechanism is complicated by the fact that the surface of conventional alkyl-bonded silica materials is heterogeneous [72–74]. This results from the nature of the bonded layer structure, the silica surface heterogeneity, and the randomness of the distribution of the alkyl-bonded chains and endcapping groups, giving rise to multiple sites of interaction. Therefore, solute retention seems to be more a result of a combination of different simultaneous adsorption and partitioning processes that depend on the nature of the solute molecules and stationaryphase environment. The mixed retention mechanism involves hydrophobic, polar, hydrogen bond, ion-exchange, and other interactions both at adsorption and in partitioning sites. However, adsorption and partitioning explain only partially the nature of solute interactions with the stationary phase. The separation of isomers and other compounds are associated with the shape selectivity of the stationary-phase architecture. Also, larger (bulkier) molecules experience greater difficulty in entering the stationary phase, because they are too large to easily fit between the bonded alkyl ligands (steric hindrance), thereby decreasing the solute retention [75]. An illustration of the stationary-phase environment, experienced differently by different solutes, is given by the relationship between isocratic log k-values and the carbon number of the compounds in a homologous series when retained on C18, C8, and C4 modified silica [4]. Log k increases linearly with respect to the number of carbon atoms of the homologs, until this number approximately equals the length of the organic ligand of the stationary phase. This discontinuity indicates a change in retention mechanism from pure partitioning to a mixed process of adsorption and partitioning, after which steric factors prevent solute molecules undergoing partitioning, adsorption being the predominant retention mechanism for very large homologs. 8.6.4 Investigating How RPLC Really Works

Although rotational and translational molecular motion is relatively not impeded in liquids, these motions are restricted for bonded ligands, which are chained to the support surface by one of their extremities in RPLC columns. This results in partial ordering. Spacing between the bonded phase molecules reflects both the irregular physical nature of the silica surface and the chemical heterogeneity of silanols. The layer of bonded alkyl chains, in contact with the liquid mobile phase, adsorbs selectively the components of the mobile phase, swells to some degree, and forms an interface with a complex structure. The conformation of the bonded chains depends on a number of factors,

8.6 Attempts to Explain the Retention Mechanisms in RPLC

including mobile-phase composition, chemical structure of the grafted chains, bonding density, nature of the support surface, pore diameter and shape, pressure, and pH. A long-term goal for RPLC researchers has been providing a definitive description of the structure of the bonded chains and solvent molecules within the stationary phase region and at the interface with the mobile phase, fully explaining the retention mechanisms of different solutes [76]. For this purpose, a variety of chromatographic, spectroscopic, and molecular simulation methods have been employed. The vast majority of the studies corresponds to C18 phases. However, understanding the stationary-phase architecture and its implication in the separation performance is still incomplete. Indirect evidence of the stationary-phase architecture is provided by the chromatographic retention behavior of probe solutes. The retention time is a thermodynamic parameter that indicates solute distribution between mobile and stationary phases. However, being based on macroscopic measurements, thermodynamics cannot directly offer microscopic mechanistic insights to infer a molecular picture for the mixed retention mechanisms. More direct evidence of the conformational structure, motion of alkyl-bonded chains and silica support is achieved through diverse studies with spectroscopic techniques, such as NMR, small-angle neutron scattering, Fourier transform infrared, Raman, sum-frequency generation, and fluorescence spectroscopies, which directly probe the alkyl chains or molecules that interact with the chains [77–81]. These techniques have yielded a wealth of information on structural and dynamic properties of alkyl stationary phases. For example, solid-state NMR spectroscopy has revealed the distribution and types of silanols on silica, and silica bonding details of derivatizing reagents. However, the studies with spectroscopic techniques involve averaging over time and space. Thus, the information gained, involving chain conformations and solute interactions, corresponds to the average behavior of a large number of molecules, and hence, offers only limited insight into the individual grafted chains and their distribution, the penetration of solvent into the different regions of the retentive phase, and the spatial and orientational preferences of solute molecules within and near the retentive phase. Taking into account that the properties of a chromatographic system are described by averages over millions of configurations, it is obvious that the details of the retention mechanisms in RPLC cannot be revealed. The use of molecular simulation to investigate the RPLC behavior goes back to the last decade of the twentieth century [6,82–87]. Virtual computer experiments may provide molecular-level details that are impossible to achieve through experimental chromatographic and spectroscopic techniques. Some studies involve the presence of solvents to mimic the mobile phase, whereas others utilize the bonded material in the absence of liquids. Two main approaches are used: molecular dynamics and Monte Carlo methods. Molecular dynamics provides information on chain dynamics and transport properties, whereas Monte Carlo methods are suited for the investigation of sorption

185

186

8 Reversed Phase Liquid Chromatography

equilibria underlying RPLC retention. Both approaches have been used to elucidate bonded chain conformations and solvent distributions. With molecular simulation, models can be precisely described in terms of well-controlled physical conditions (molecular structure of the constituents, bonding density, concentrations, temperature, pressure, etc.), using a noninvasive approach, where the influence of changes in specific parameters can be examined directly without altering other parameters. With the development of more efficient algorithms and more powerful computers, simulation approaches have become more practical, facilitating the investigation of complex RPLC systems, the design of improved stationary phases, and the optimal use of existing materials. The good agreement between simulations and laboratory experiments is impressive and provides a guarantee of sufficient realism to learn about chromatographic systems. 8.6.5 Going Down to the Molecular Detail

Experimental measurements and molecular simulations have given insights at the molecular level in RPLC systems. Some conclusions regarding the chain conformation, solvent penetration, and mechanisms of retention are next summarized. 8.6.5.1

Chain Conformation

For all solvents, the grafted chains in RPLC stationary phases are rather disordered, with a significant fraction of gauche defects (dihedral angle deviations larger than 60° with respect to the angle of the trans conformer) [81,88–90]. Temperature increases the conformational disorder. Meanwhile, the surface modification (i.e., monomeric, polymeric, etc.) does not directly affect the order. All this information is highly relevant as any change that results in increased stationary-phase order also enhances recognition of shape-constrained solutes. For relatively low bonding density, the alkyl chains may lean over the support but are generally oriented away from the surface (Figure 8.5). The C8 chains prefer to align perpendicular to the support, whereas the distribution for the C18 and C30 chains is much broader. Only the lower parts of the long C18 and C30 chains are preferentially aligned perpendicular to the silica support. Instead, the terminal segments show no significant orientation preference relative to the substrate surface. The mobility of the alkyl chains plays an important role in the retention mechanism, since it influences the accessibility of solutes and mobilephase components to surface silanols. This mobility is a consequence of solvation processes and thermal motion, and depends strongly on the type and length of the bonded chains, being hampered at increasing bonding density. 8.6.5.2

Adsorption and Partitioning of Common Solvents

The stationary phase is a heterogeneous medium with multiple preferred regions for solutes and solvent molecules [91]. Changes are simultaneously produced by

8.6 Attempts to Explain the Retention Mechanisms in RPLC

Figure 8.5 Simulation snapshots for RPLC dimethyl octadecylsilane systems with increasing surface coverage, using methanol/water solvent. The stationary phase is depicted as tubes with CHx groups in gray, silicon in

yellow, hydrogen in white, and oxygen in red. Solvent molecules are shown in the ball and stick representation with CH3 in cyan, oxygen in red, and hydrogen in white. (Reproduced with permission from Ref. [90].)

both adsorption and partitioning mechanisms, and the relative contribution of each mechanism varies with the chain length, bonding density, solute polarity, and mobile-phase composition. Solvent penetration enhances the bonded chain alignment, increasing the average height of the terminal methyl group over the silica surface. With pure water, there is very little penetration of the solvent and the least chain alignment is observed. The extent of solvent penetration into the hydrophobic bonded chain region increases as the polarity of the mobile phase decreases, being also larger for stationary phases having low grafting density [92]. Acetonitrile shows greater penetration relative to methanol at the same solvent to water ratio. In the interfacial region, methanol molecules are preferentially oriented with the hydroxyl group pointing toward the bulk mobile phase, whereas there is little orientation ordering for acetonitrile in this region. In addition to partitioning into the bonded phase, solvents are adsorbed on the silica surface. This is most important for methanol and especially for water, which interact strongly with the free surface silanols through hydrogen bonding. Acetonitrile has much weaker interaction with the silica surface due to its low hydrogen-bond acceptor ability. Therefore, in acetonitrile–water mobile phases, residual silanols are covered mostly by water molecules, being a very large fraction available for interaction with solute molecules. In fact, a main difference between mobile phases prepared with methanol and acetonitrile is the difference in the fraction of free silanols [29]. The differences between these two solvents are larger at high modifier contents, or in the presence of electrolytes, when less water is available for adsorption on residual silanols. However, the differences at the same modifier concentration are much smaller than those observed by changing the modifier content. For a wide range of stationary phases (different ligand types, alkyl chain lengths, and bonding densities), the adsorbed solvent layer between the bonded alkyl chains and the mobile phase is roughly a monolayer for methanol, while multilayers several molecules thick are formed for acetonitrile and tetrahydrofuran. The formation of solvent multilayers at the top of the bonded stationary

187

188

8 Reversed Phase Liquid Chromatography

phase is the reason of the larger column loadability, when acetonitrile is used as modifier instead of methanol. The presence of a wide interfacial region also explains the large re-equilibration times required when the mobile phase contains tetrahydrofuran. 8.6.5.3

Adsorption and Partitioning of Solutes

Solutes compete with solvent molecules for the accessible volume and sites of interaction. A complete description of the retention mechanism requires the knowledge of the location where a solute molecule is retained within the stationary phase and how it is oriented. It is known that hydrophobic solutes are retained in C18 phases according to: (i) adsorption on the chain’s outer surface near the ligand–solution interface, preferentially parallel to the silica surface, and (ii) partitioning both within the layer of adsorbed solvent molecules at the top of the bonded phase and within the chains. In contrast, solutes with hydroxyl groups (such as alcohols) are preferentially located closer to the ligand–solution interface, due to hydrogen bonding with water [91]. At increasing bonding density, the order of the bonded chains and their alignment perpendicular to the surface increase (Figure 8.5). Solute partitioning becomes thus entropically unfavorable, since the inclusion of a solute molecule into the stationary phase further increases the order within the system. Hence, adsorption dominates both for highly dense bonded phases and for solutes having large molecular sizes, up to the point of being excluded from the bonded layer. Accordingly, the retention into C8 phases, and with shorter ligands, is best described as an adsorption process for both nonpolar and polar solutes. Retention also depends on the shape of the solute molecules and their orientation with respect to the bonded chains. These perform as a sort of imprinted material, which preferentially retains certain solutes that have specific geometry to interact with the bonded layer microenvironment. This explains, for instance, the separation of isomers. As indicated, RPLC packings are not homogeneous but include different sites of interaction. The low-energy sites are the most abundant, being located at the interface between the mobile phase and the bonded layer [73]. Other sites are located inside the bonded layer, more or less deep, causing stronger retention. The availability of these high-energy sites depends on the relative sizes of the bonded chains and solutes. For a given bonded phase, there is a size range for which the solute fits snugly, and high-energy adsorption sites become available. However, solutes of the same molecular size interact differently. Accordingly, the overall retention in RPLC is more the result of a complicated convolution of many different interactions happening simultaneously, with the different sites available on the surface of the bonded layer and underlying silica surface, and inside the bonded layer. The intrinsic heterogeneity of the packings is one of the keys of the successful application of RPLC to the separation of an extremely wide variety of solutes.

8.6 Attempts to Explain the Retention Mechanisms in RPLC

8.6.5.4

Anomalous Behavior with Highly Aqueous Mobile Phases

RPLC is often carried out using water-rich mobile phases, sometimes containing less than 5% v/v organic modifier. With high water contents, very polar solutes that would otherwise elute close to the dead time are sometimes satisfactorily retained. However, under these conditions, a drastic loss of retention is observed with conventional columns after stopping and restarting the flow. Also, retention times are not reproducible, efficiency is low with significant tailing, and equilibration times are long [93–95]. These undesirable effects depend on the nature of the bonded phase, bonding density, pore diameter, pressure, and temperature. The problem arises especially with stationary phases having long alkyl chains (especially C18 and longer chains), dense surface coverages, and small pores. Fortunately, the anomalous effect does not damage the column permanently, regeneration being achieved by simply washing with a mobile phase containing 50% v/ v or higher concentration of an organic solvent. This anomalous behavior has been one of the mysteries of RPLC. It was first thought that water would force the chains to aggregate and lie flat on the support surface, giving rise to a nonsolvated collapsed state, inaccessible to solutes. Because most of the bonded phase is inside the pores of the silica support, this collapse would occur within the pores rather than on the outer surface. From this interpretation, the phenomenon was called “phase collapse”, “hydrophobic collapse”, or “chain folding”, names still in use. However, the dramatic loss of retention in water-rich mobile phases seems to be more the result of “pore dewetting” and not a change in chain conformation. In RPLC, the presence of the organic modifier is essential to ensure that the mobile phase wets the hydrophobic chains and enters a substantial fraction into the pores of the silica particles, offering a large surface area for solute partitioning. In water-rich mobile phases, the solute cannot fully access the interior surfaces of the stationary phase, because the pressure is unable to overcome the surface tension forces resisting penetration of the mobile phase into the pores. When pressure decreases, water is excluded from the highly hydrophobic pore environment and the pore becomes dewetted. Consequently, the stationary phase located at the pore walls has no contact with the mobile phase and does not show any retention capability. This effect occurs unless the pressure is high enough to force the mobile phase into the pores. Retention is thus a function of the column inlet pressure with higher retention observed for higher pressure. Furthermore, owing to the pressure gradient, the pore penetration of highly aqueous mobile phases declines along the column with almost the full exclusion of the mobile phase from the stationary phase pores toward the column outlet. Therefore, the solutes are more retained at the column head, where the pores are better wetted. This interpretation was confirmed by molecular simulation [76]. Dewetting is less an issue when the bonding density is low (< 3 μmol/m2) presumably because residual silanols allow the silica surface to be wetted with water molecules. The solutes are, therefore, able to penetrate the bonded layer, where they interact. Dewetting is also rare with very short, alkyl-bonded phases, such

189

190

8 Reversed Phase Liquid Chromatography

as C3, because little shielding of silanols occurs. The problem is also avoided by using especial columns designed for operation in highly aqueous environments: nonencapped, polar-encapped, hydrophilic, polar-enhanced, and polarembedded alkyl stationary phases, where the bonded layer can be totally penetrated by the solutes, independent of the water content of the mobile phase.

8.7 Development and Trends in RPLC

RPLC has evolved with the driving force of improving the batch-to-batch reproducibility, stability, and mechanical strength of the column packings. Stationary phases with an extended working pH range and longer life, and increased efficiency, selectivity, and sensitivity, have become available. Furthermore, the run times have decreased largely, and bonded phases with diversified chemistries have appeared. The technique that we are practicing today is the result of the hard work of hundreds of researchers who have elucidated the possibilities and potentials of RPLC, and highlighted the challenges that are still awaiting further research. Since the birth of silica-based RPLC, chromatographic packing materials have developed from irregularly shaped to spherical particles with narrower size distributions, from large to small, and from impure to highly pure uniform silica. RPLC columns are today available with a great variety of hydrophobic nonpolarizable, moderately polar, and polarizable (mainly with phenyl rings) groups and bonding densities, grafted to silica particles of different diameters, surface areas, and pore sizes. In the early days of RPLC, the market was dominated by columns packed with large irregular silica particles, where residual silanols had not been end-capped. Major developments in column supports were the reduction of the average particle diameter of porous silica from the former 10 to 5 μm spherical end-capped particles, with a narrow particle size range (mid-1990s), and further to 4, 3, and finally, sub-2 μm particles. A smaller particle size allows a substantial increase in the number of theoretical plates per unit of column length, and a reduction of the run time [96]. The sub-2 μm ultrahigh performance HPLC (UHPLC) requires, however, a more advanced instrument technology (introduced in 2004) that allows column pressures up to 150 MPa. Another disadvantage of sub-2 μm packings is the higher risk of irreversible damage if solvents and samples are not properly filtered. It is, thus, not surprising that 5 and 3 μm spherical particles are still the norm in most analytical laboratories. Simultaneously, due to economic and environmental considerations, involving the reduction of solvent consumption, shortening of the run times, and increased sample throughput in routine analysis, the column length was reduced from a typical 10–15 cm to a mere 2–5 cm. With the increasing use of mass spectrometric detection, there is less need for complete resolution of all peaks, so that a shorter column could be enough in many instances. So far as the column diameter (ID) is concerned, still 4.6 and 4.0 mm ID columns dominate the

8.7 Development and Trends in RPLC

market, but narrow-bore 3.5 and 2 mm ID columns are becoming common. A limitation is the lower sensitivity for evaporative detectors, which depends on the absolute mass of the solute. Scaling down below 1 mm ID, using, for instance, fused silica capillary columns from 75 to 500 μm ID, requires special instrumentation and has not reached broad acceptance. A problem with silica columns is the fragility of the column ends when connections are tightened. However, capillary- and nanoflow RPLC are playing an important role in proteomics (and other “omics”), being also widely used in tandem with high-resolution mass spectrometers. Besides fully porous silica microparticles, there are other interesting materials that are gaining popularity, such as core–shell (also called fused core) silica particles and monolithic materials, which allow fast analysis with common HPLC instrumentation. The core–shell approach was developed by Kirkland in 1975; however, commercial columns became available only about 30 years later [97]. A core–shell particle consists of a solid core of silica, typically of 1.7 μm ID, coated with a thin porous silica layer with a thickness of 0.5 μm. In modern commercial core–shell particles, the shell is obtained by sintering or gluing together several layers of silica nanoparticles on the solid silica core. The thin porous shell allows very fast mass transfer kinetics of the solutes, which results in impressive efficiency enhancements. Silica and organic or polymeric monoliths have a double-pore structure of macropores, which provide high permeability permitting high flow rates, and mesopores (similar to those in silica microparticles), which increase the surface area [96,98–100]. Organic monoliths are not widely commercially available yet [99]. Silica-based monoliths are, in contrast, commercialized [100]. However, these are constructed by following a unique chemistry and, therefore, have not reached all the expected success. Developments of new stationary phases will continue, but already much has been done to produce silica-based phases usable over wide pH ranges, with high reproducibility and reduced peak tailing for basic solutes. Working at extreme pH values is also possible by using zirconia- and titania-based stationary phases [17], as well as graphitic carbon and carbon-core phases with a nano-diamond shell. However, the surface chemistry of zirconium and titanium oxides is rather complex, resulting in unexpected retention values for many solutes, and it is not as easily manipulable as that of silica. The true secret of the success of silica seems to lie in its simple and versatile silanization chemistry. Porous graphitic carbon has also been developed as an alternative to silica to eliminate the problems associated with residual silanols [101]. The carbon phase has a rigid structure, is chemically stable and hydrophobic, and displays a different selectivity from that of silica-based stationary phases. Retention is mainly due to the attraction between the solute dipoles and the induced dipoles of the carbon surface. This material is particularly useful for the separation of isomeric mixtures and for highly polar solutes normally not retained on C18-silica. Carbon-clad zirconia, where zirconia microspheres are coated with a thin carbon monolayer, is another type of carbonaceous RPLC support that has high

191

192

8 Reversed Phase Liquid Chromatography

mechanical resistance and excellent chemical stability [4]. The main disadvantage is that the carbon coverage is not complete and the uncovered highly acidic sites (zirconia groups) can cause peak tailing and even irreversible adsorption of basic compounds. Meanwhile, small particles of polymeric resins such as polystyrenedivinylbenzene are employed in combination with either very basic or acidic solvents, allowing the work with basic compounds [4]. However, polymer particles are deformed under pressure. This prevents the use of very small particles, which limits the efficiency. The introduction in 2014 of carbon-core particles covered with a shell of nano-diamonds in a polymeric matrix is promising. Commercially available RPLC columns of different manufacturers exhibit significantly different chromatographic behaviors, even among columns that are nominally similar [25]. Not only columns from different brands, such as those shown in Figure 8.2, but also columns from among batches of the same brand can differ from each other. These differences, instead of being a handicap, offer an interesting variety for method development. However, in spite of the many parameters that can be adjusted, the necessary selectivity may not be found. In such a case, a solution is to serially connect columns of different selectivity, varying also their length and order, which provides enhanced capabilities for the fine-tuning of the selectivity [102,103]. For years, a typical RPLC column in a routine laboratory has been a 15 cm long, 4.6 mm ID stainless steel column packed with 5 μm spherical end-capped C18 porous silica microparticles. This has changed in the past decade, since columns packed with both conventional full-porous and core–shell particles below 3 μm have become increasingly common. However, it is surprising that classical superseded porous materials have been also maintained on the market for many decades, whereas alternative materials, either well proved or promising, are not recognized or only slowly accepted by the users. Considerable difficulties still remain in the analysis of basic compounds, understanding and remediation of overloading, and instability of silica at extreme pH values. The limited pH range prevents exploitation of selectivity effects for ionizable compounds, for instance, increasing the retention of acidic solutes upon protonation. The retention of cations and anions can also be increased by ion-pairing with a counter-ion; however, the low or null retention of poly-ions is still a problem. The development of alternatives to the current stationary phases, including improved silica materials and nonsilica-based (e.g., organic polymers) or composite (inorganic– organic) materials, can provide solutions to many of these problems.

References 1 Melander, W.R. and Horváth, C. (1980)

Reversed-phase chromatography, in High-Performance Liquid Chromatography, vol. 2, Academic Press, New York, pp. 113–319.

2 Lough, W.J. (2006) Reversed

phase liquid chromatography, in Encyclopedia of Analytical Chemistry (on line) (ed R.A. Meyers), John Wiley & Sons, Inc., New York.

References 3 Pesek, J.J. and Matyska, M.T. (2006)

4

5

6

7

8

9

10

11

Reversed-phase chromatography: description and applications, in Encyclopedia of Chromatography (ed J. Cazes), Taylor & Francis, New York, pp. 719–722. Soliven, A., Kayillo, S., and Shalliker, R.A. (2013) Reversed phase liquid chromatography, in Reference Module in Chemistry, Molecular Sciences and Chemical Engineering (ed J. Reedijk), Elsevier, Amsterdam. Snyder, L.R. and Kirkland, J.J. (1979) Introduction to Modern Liquid Chromatography, 2nd edn, John Wiley & Sons, Inc., New York. Rafferty, J.L., Siepmann, J.I., and Schure, M.R. (2010) Understanding the retention mechanism in reversed-phase liquid chromatography: insights from molecular simulation. Adv. Chromatogr., 48, 1–55. García-Alvarez-Coque, M.C. and TorresLapasió, J.R. (2013) Secondary chemical equilibria in reversed-phase liquid chromatography, in Liquid Chromatography: Fundamentals and Instrumentation (eds S. Fanali, P. Haddad, C.F. Poole, P.J. Schoenmakers, and D. Lloyd), Elsevier, Amsterdam, pp. 87–104. Torres-Lapasió, J.R., García-AlvarezCoque, M.C., Bosch, E., and Rosés, M. (2005) Considerations on the modelling and optimisation of resolution of ionisable compounds in extended pHrange columns. J. Chromatogr. A, 1089, 170–186. Wolcott, R.G., Dolan, J.W., Snyder, L.R., Bakalyar, S.R., Arnold, M.A., and Nichols, J.A. (2000) Control of column temperature in reversed-phase liquid chromatography. J. Chromatogr. A, 892, 211–230. Staroverov, S.M. and Fadeev, A.Y. (1991) Apparent simplicity of reversed stationary phases for high-performance liquid chromatography. J. Chromatogr., 544, 77–98. Wheeler, J.F., Beck, T.L., Klatte, S.J., Cole, L.A., and Dorsey, J.G. (1993) Phase transitions of reversed-phase stationary phases: cause and effects in the

12

13

14

15

16

17

18

19

20

21

mechanism of retention. J. Chromatogr. A, 656, 317–333. Kirkland, J.J. (2004) Development of some stationary phases for reversedphase HPLC. J. Chromatogr. A, 1060, 9–21. Pesek, J.J. and Matyska, M.T. (2006) Reversed-phase stationary phases, in Encyclopedia of Chromatography (ed J. Cazes), Taylor & Francis, New York, pp. 723–726. Neue, U.D. (2007) Stationary phase characterization and method development. J. Sep. Sci., 30, 1611–1627. Qiu, H., Liang, X., Sun, M., and Jiang, S. (2011) Development of silica-based stationary phases for high-performance liquid chromatography. Anal. Bioanal. Chem., 399, 3307–3322. Unger, K.K. and Liapis, A.I. (2012) Adsorbents and columns in analytical high-performance liquid chromatography: a perspective with regard to development and understanding. J. Sep. Sci., 35, 1201–1212. Neue, U.D. (2009) Silica gel and its derivatization for liquid chromatography, in Encyclopedia of Analytical Chemistry (on line) (ed R.A. Meyers), JohnWiley & Sons, Inc., New York. Hetem, M.J.J., de Haan, J.W., Claessens, H.A., van de Ven, L.J.M., Cramers, C.A., and Kinkel, J.N. (1990) Influence of alkyl chain length on the stability of n-alkyl-modified reversed phases. 1. Chromatographic and physical analysis. Anal. Chem., 62, 2288–2296. Claessens, H.A. and van Straten, M.A. (2004) Review on the chemical and thermal stability of stationary phases for reversed-phase liquid chromatography. J. Chromatogr. A, 1060, 23–41. Croes, K., Steffens, A., Marchand, D.H., and Snyder, L.R. (2005) Relevance of π–π and dipole–dipole interactions for retention on cyano and phenyl columns in reversed-phase liquid chromatography. J. Chromatogr. A, 1098, 123–130. Nawrocki, J. (1997) The silanol group and its role in liquid chromatography. J. Chromatogr. A, 779, 29–71.

193

194

8 Reversed Phase Liquid Chromatography 22 Bocian, S. and Buszewski, B. (2012)

23

24

25

26

27

28

29

30

31

Residual silanols at reversed-phase silica in HPLC: a contribution for a better understanding. J. Sep. Sci., 35, 1191–1200. Méndez, A., Bosch, E., Rosés, M., and Neue, U.D. (2003) Comparison of the acidity of residual silanol groups in several liquid chromatography columns. J. Chromatogr. A, 986, 33–44. Marchand, D.H., Carr, P.W., McCalley, D.V., Neue, U.D., Dolan, J.W., and Snyder, L.R. (2011) Contributions to reversed-phase column selectivity. II. Cation exchange. J. Chromatogr. A, 1218, 7110–7129. Ruiz-Angel, M.J., Pous-Torres, S., CardaBroch, S., and García-Alvarez-Coque, M.C. (2014) Performance of different C18 columns in reversed-phase liquid chromatography with hydro-organic and micellar-organic mobile phases. J. Chromatogr. A, 1344, 76–82. Baeza-Baeza, J.J., Ruiz-Angel, M.J., CardaBroch, S., and García-Alvarez-Coque, M.C. (2013) Half-width plots, a simple tool to predict peak shape, reveal column kinetics and characterise chromatographic columns in liquid chromatography: state of the art and new results. J. Chromatogr. A, 1314, 142–153. Gilroy, J.J., Dolan, J.W., and Snyder, L.R. (2003) Column selectivity in reversedphase liquid chromatography: IV. Type-B alkyl–silica columns. J. Chromatogr. A, 1000, 757–778. McCalley, D.V. (2010) The challenges of the analysis of basic compounds by high performance liquid chromatography: some possible approaches for improved separations. J. Chromatogr. A, 1217, 858–880. Bocian, S., Vajda, P., Felinger, A., and Buszewski, B. (2010) Effect of endcapping and surface coverage on the mechanism of solvent adsorption. Chromatographia, 71, 5–11. Rogers, S.D. and Dorsey, J.G. (2000) Chromatographic silanol activity test procedures: the quest for a universal test. J. Chromatogr. A, 892, 57–65. Wilson, N.S., Gilroy, J., Dolan, J.W., and Snyder, L.R. (2004) Column selectivity in

32

33

34

35

36

37

38

39

40

41

reversed-phase liquid chromatography: VI. Columns with embedded or endcapping polar groups. J. Chromatogr. A, 1026, 91–100. Barwick, V.J. (1997) Strategies for solvent selection. Trends Anal. Chem., 16, 293–309. Ramis-Ramos, G. and García-AlvarezCoque, M.C. (2013) Solvent selection in liquid chromatography, in Liquid Chromatography: Fundamentals and Instrumentation (eds S. Fanali, P. Haddad, C.F. Poole, P.J. Schoenmakers, and D. Lloyd), Elsevier, Amsterdam, pp. 223–249. Dolan, J.W. (2010) Selectivity in reversedphase LC separations. Part I. Solvent-type selectivity. LCGC North Am., 28, 1022–1027. Snyder, L.R., Carr, P.W., and Rutan, S.C. (1993) Solvatochromically based solventselectivity triangle. J. Chromatogr. A, 656, 537–547. Snyder, LR. (1974) Classification of the solvent properties of common liquids. J. Chromatogr., 92, 223–230. Rosés, M. (2004) Determination of the pH of binary mobile phases for reversedphase liquid chromatography. J. Chromatogr. A, 1037, 283–298. Viseras, C., Cela, R., Barroso, C.G., and Pérez-Bustamante, J.A. (1987) Influence of temperature in reverse-phase highperformance liquid chromatography with gradient elution. Anal. Chim. Acta, 196, 115–122. Pous-Torres, S., Torres-Lapasió, J.R., Baeza-Baeza, J.J., and García-AlvarezCoque, M.C. (2007) Combined effect of solvent content, temperature and pH on the chromatographic behaviour of ionisable compounds. J. Chromatogr. A, 1163, 49–62. Heinisch, S. and Rocca, J.L. (2009) Sense and nonsense of high-temperature liquid chromatography. J. Chromatogr. A, 1216, 642–658. Baeza-Baeza, J.J., Dávila, Y., FernándezNavarro, J.J., and García-AlvarezCoque, M.C. (2012) Measurement of the elution strength and peak shape enhancement at increasing modifier concentration and temperature in

References

42

43

44

45

46

47

48

49

50

51

52

RPLC. Anal. Bioanal. Chem., 404, 2973–2984. Hartonen, K. and Riekkola, M.L. (2008) Liquid chromatography at elevated temperatures with pure water as the mobile phase. Trends Anal. Chem., 27, 1–14. Ruiz-Angel, M.J., Carda-Broch, S., García-Alvarez-Coque, M.C., and Berthod, A. (2005) Effect of ionization and the nature of the mobile phase in quantitative structure–retention relationship studies. J. Chromatogr. A, 1063, 25–34. Dolan, J.W. (2013) Gradient elution, Part I. Intuition. LCGC North Am., 31, 204, 206, 208, 209. Baeza-Baeza, J.J., Ortiz-Bolsico, C., Torres-Lapasió, J.R., and García-AlvarezCoque, M.C. (2013) Approaches to model the retention and peak profile in linear gradient reversed-phase liquid chromatography. J. Chromatogr. A, 1284, 28–35. Dolan, J.W. (2014) LC method scaling, Part I. Isocratic separations. LCGC North Am., 32, 98, 100, 102. Jandera, P. (2006) Can the theory of gradient liquid chromatography be useful in solving practical problems? J. Chromatogr. A, 1126, 195–218. Haky, J.E. and Teifer, D.A. (2006) Gradient elution, in Encyclopedia of Chromatography (ed J. Cazes), Taylor & Francis, New York, pp. 393–396. Snyder, L.R. and Dolan, J.W. (2007) HighPerformance Gradient Elution, John Wiley & Sons, Inc., Hoboken, NJ. Dolan, J.W. and Snyder, L.R. (2012) Gradient elution chromatography, in Encyclopedia of Analytical Chemistry (on line) (ed R.A. Meyers), John Wiley & Sons, Inc., New York. Dolan, J.W. (2002) The hazards of adjusting gradients. LCGC North Am., 20, 940–946. Vivó-Truyols, G., Torres-Lapasió, J.R., and García-Alvarez-Coque, M.C. (2004) Estimation of significant solvent concentration ranges and its application to the enhancement of the accuracy of gradient predictions. J. Chromatogr. A, 1057, 31–39.

53 Dolan, J.W. and Snyder, L.R. (2013)

54

55

56

57

58

59

60

61 62

63

64

Theory and practice of gradient elution, in Liquid Chromatography: Fundamentals and Instrumentation (eds S. Fanali, P. Haddad, C.F. Poole, P.J. Schoenmakers, and D. Lloyd), Elsevier, Amsterdam, pp. 269–282. Concha-Herrera, V., Vivó-Truyols, G., Torres-Lapasió, J.R., and García-AlvarezCoque, M.C. (2005) Limits of multi-linear gradient optimisation in reversed-phase liquid chromatography. J. Chromatogr. A, 1063, 79–88. Dolan, J.W. (2013) Gradient elution, Part IV. Dwell-volume problems. LCGC North Am., 31, 456, 458, 460–463. Dolan, J.W. (2013) Gradient elution, Part V. Baseline drift problems. LCGC North Am., 31, 538, 540, 542, 543. Carrasco-Correa, E.J., Acquaviva, A., Herrero-Martínez, J.M., and RamisRamos, G. (2013) Synchronized gradient elution in capillary liquid chromatography. J. Chromatogr. A, 1318, 142–148. Kaliszan, R., Wiczling, P., and Markuszewski, M.J. (2004) pH gradient reversed-phase HPLC. Anal. Chem., 76, 749–760. Nikitas, P., Pappa-Louisi, A., and Zisi, C. (2012) pH-gradient reversed-phase liquid chromatography of ionogenic analytes revisited. Anal. Chem., 84, 6611–6618. Britton, H.T.K. and Robinson, R.A. (1931) Universal buffer solutions and the dissociation constant of veronal. J. Chem. Soc., 1931, 1456–1462. Ellis, D.A. (1961) A new universal buffer system. Nature, 191, 1099–1100. Lesins, V. and Ruckenstein, E. (1989) Gradient flow programming: a coupling of gradient elution and flow programming. J. Chromatogr., 467, 1–14. Nikitas, P., Pappa-Louisi, A., and Balkatzopoulou, P. (2007) Fundamental equation of dual-mode gradient elution in liquid chromatography involving simultaneous changes in flow rate and mobile-phase composition. Anal. Chem., 79, 3888–3893. Kaminski, L., El Deeb, S., and Wätzig, H. (2008) Repeatability of monolithic HPLC

195

196

8 Reversed Phase Liquid Chromatography

65

66

67

68

69

70

71

72

73

columns while using a flow program. J. Sep. Sci., 31, 1745–1749. Zhu, P.L., Snyder, L.R., Dolan, J.W., Djordjevic, N.M., Hill, D.W., Sander, L.C., and Waeghe, T.J. (1996) Combined use of temperature and solvent strength in reversed-phase gradient elution. Part I. Predicting separation as a function of temperature and gradient conditions. J. Chromatogr. A, 756, 21–39. Nikitas, P., Pappa-Louisi, A., Papachristos, K., and Zisi, C. (2008) Theory and application of the two-mode gradient elution in liquid chromatography involving simultaneous changes in temperature and mobilephase composition. Anal. Chem., 80, 5508–5514. Poole, C.F. and Poole, S.K. (2009) Foundations of retention in partition chromatography. J. Chromatogr. A, 1216, 1530–1550. Horváth, C., Melander, W., and Molnár, I. (1976) Solvophobic interactions in liquid chromatography with nonpolar stationary phases. J. Chromatogr., 125, 129–156. Horváth, C. and Melander, W. (1977) Liquid chromatography with hydrocarbonaceous bonded phases: theory and practice of reversed phase chromatography. J. Chromatogr. Sci., 15, 393–404. Carr, P.W., Martire, D.E., and Snyder, L.R. (1993) Preface to the retention process in reversed-phase liquid chromatography. J. Chromatogr. A, 656, 1–2. Vailaya, A. and Horváth, C. (1998) Retention in reversed-phase chromatography: partition or adsorption? J. Chromatogr. A, 829, 1–27. Dorsey, J.G. (1989) The molecular mechanism of retention in reversedphase liquid chromatography. Chem. Rev., 89, 331–346. Gritti, F. and Guiochon, G. (2005) Critical contribution of nonlinear chromatography to the understanding of retention mechanism in reversed-phase liquid chromatography. J. Chromatogr. A, 1099, 1–42.

74 Gritti, F. and Guiochon, G. (2006)

75

76

77

78

79

80

81

82

83

84

Heterogeneity of the adsorption mechanism of low molecular weight compounds in reversed-phase liquid chromatography. Anal. Chem., 78, 5823–5834. Carr, P.W., Dolan, J.W., Neue, U.D., and Snyder, L.R. (2011) Contributions to reversed-phase column selectivity. Part I. Steric interaction. J. Chromatogr. A, 1218, 1724–1742. Schure, M.R., Rafferty, J.L., Zhang, L., and Siepmann, J.I. (2013) How reversed-phase liquid chromatography works. LCGC North Am., 31, 630–637. Hansen, R.L. and Harris, J.M. (1996) Lateral diffusion of molecules partitioned into silica-bound alkyl chains: influence of chain length and bonding density. Anal. Chem., 68, 2879–2884. Pursch, M., Sander, L.C., and Albert, K. (1999) Understanding reversed-phase liquid chromatography with solid-state NMR. Anal. Chem., 71, 733A–741A. Ducey, M.W., Orendorff, C.J., Pemberton, J.E., and Sander, L.C. (2002) Structure– function relationships in high-density octadecylsilane stationary phases by Raman spectroscopy. Part 2. Effect of common mobile-phase solvents. Anal. Chem., 74, 5585–5592. Wirth, M.J., Swinton, D.J., and Ludes, M.D. (2003) Adsorption and diffusion of single molecules at chromatographic interfaces. J. Phys. Chem. B, 107, 6258–6268. Sander, L.C., Lippa, K.A., and Wise, S.A. (2005) Order and disorder in alkyl stationary phases. Anal. Bioanal. Chem., 382, 646–668. Yarovsky, I., Aguilar, M.I., and Hearn, M.T.W. (1995) Influence of the chain length and surface density on the conformation and mobility of n-alkyl ligands chemically immobilized onto a silica surface. Anal. Chem., 67, 2145–2153. Klatte, S.J. and Beck, T.L. (1996) Microscopic simulation of solute transfer in reversed phase liquid chromatography. J. Phys. Chem., 100, 5931–5934. Slusher, J.T. and Mountain, R.D. (1999) A molecular dynamics study of a

References

85

86

87

88

89

90

91

92

93

reversed-phase liquid chromatography model. J. Phys. Chem. B, 103, 1354–1362. Zhang, L., Rafferty, J.L., Siepmann, J.I., Chen, B., and Schure, M.R. (2006) Chain conformation and solvent partitioning in reversed-phase liquid chromatography: Monte Carlo simulations for various water/methanol concentrations. J. Chromatogr. A, 1126, 219–231. Mansfield, E.R., Mansfield, D.S., Patterson, J.E., and Knotts, T.A. (2012) Effects of chain grafting positions and surface coverage on conformations of model reversed-phase liquid chromatography stationary phases. J. Phys. Chem. C, 116, 8456–8464. Lindsey, R.K., Rafferty, J.L., Eggimann, B.L., Siepmann, J.I., and Schure, M.R. (2013) Molecular simulation studies of reversed-phase liquid chromatography. J. Chromatogr. A, 1287, 60–82. Lippa, K.A., Sander, L.C., and Mountain, R.D. (2005) Molecular dynamics simulations of alkylsilane stationaryphase order and disorder. Part 1. Effects of surface coverage and bonding chemistry. Anal. Chem., 77, 7852–7861. Lippa, K.A., Sander, L.C., and Mountain, R.D. (2005) Molecular dynamics simulations of alkylsilane stationaryphase order and disorder. Part 2. Effects of temperature and chain length. Anal. Chem., 77, 7862–7871. Rafferty, J.L., Siepmann, J.I., and Schure, M.R. (2008) Influence of bonded-phase coverage in reversed-phase liquid chromatography via molecular simulation. Part I. Effects on chain conformation and interfacial properties. J. Chromatogr. A, 1204, 11–19. Rafferty, J.L., Siepmann, J.I., and Schure, M.R. (2011) Mobile phase effects in reversed-phase liquid chromatography: a comparison of acetonitrile/water and ethanol/water solvents as studied by molecular simulation. J. Chromatogr. A, 1218, 2203–2213. Li, Z., Rutan, S.C., and Dong, S. (1996) Wetting of octadecylsilylated silica in methanol–water eluents. Anal. Chem., 68, 124–129. Majors, R.E. and Przybyciel, M. (2002) Columns for reversed-phase liquid

94

95

96

97

98

99

100

101

102

103

chromatographic separations in highly aqueous mobile phases. LCGC North Am., 20, 584, 587–593. Przybyciel, M. and Majors, R.E. (2002) Phase collapse in reversed-phase liquid chromatography. LCGC North Am., 20, 516, 520–523. Walter, T.H., Iraneta, P., and Capparella, M. (2005) Mechanism of retention loss when C8 and C18 HPLC columns are used with highly aqueous mobile phases. J. Chromatogr. A, 1075, 177–183. Unger, K.K., Skudas, R., and Schulte, M.M. (2008) Particle packed columns and monolithic columns in highperformance liquid chromatography: comparison and critical appraisal. J. Chromatogr. A, 1184, 393–415. Fekete, S., Oláh, E., and Fekete, J. (2012) Fast liquid chromatography: the domination of core–shell and very fine particles. J. Chromatogr. A, 1228, 57–71. Svec, F., Tennikova, T.B., and Deyl, Z. (eds) (2003) Monolithic materials: Preparation, Properties and Applications, Journal of Chromatography Library, Elsevier, Amsterdam. Svec, F. (2010) Porous polymer monoliths: amazingly wide variety of techniques enabling their preparation. J. Chromatogr. A, 1217, 902–924. Cabrera, K. (2004) Applications of silicabased monolithic HPLC columns. J. Sep. Sci., 27, 843–852. Pereira, L. (2008) Porous graphitic carbon as a stationary phase in HPLC: theory and applications. J. Liq. Chromatogr. Rel. Technol., 31, 1687–1731. Ortiz-Bolsico, C., Torres-Lapasió, J.R., and García-Alvarez-Coque, M.C. (2013) Simultaneous optimization of mobile phase composition, column nature and length to analyse complex samples using serially coupled columns. J. Chromatogr. A, 1314, 39–48. Ortiz-Bolsico, C., Torres-Lapasió, J.R., and García-Alvarez-Coque, M.C. (2014) Optimization of gradient elution with serially-coupled columns. Part I. Single linear gradients. J. Chromatogr. A, 1350, 51–60.

197

199

9 Modeling of Retention in Reversed Phase Liquid Chromatography Maria C. García-Alvarez-Coque, Guillermo Ramis-Ramos, José R. Torres-Lapasió, and C. Ortiz-Bolsico

9.1 Introduction

Interpretive optimization strategies in liquid chromatography are based on the accurate description (modeling) of the chromatographic behavior [1,2]. The first step in these strategies consists of gathering information about the compounds in the sample, usually attending only to retention and covering reasonably wide regions of the factors involved. For this purpose, either isocratic or gradient experiments can be used. For each solute, a relationship that describes the retention is obtained as a function of the experimental factors. This function (conventionally an empirical or mechanistic algebraic expression) will allow the prediction of isocratic or gradient retention times for a given solute under different conditions. The accuracy of predictions is decisive for the reliability of optimizations [3,4]. For a given solute, the prediction quality depends on the information richness provided by the experimental design (i.e., the number and distribution of the experiments within the factor domain), the equation selected to fit the training data, and the fitting procedure. The ranges of the experimental factors and the elution mode (isocratic or gradient) can also influence the accuracy. An overview of retention modeling in reversed phase liquid chromatography (RPLC) is next presented. The most widely used models for stationary-phase characterization are also outlined.

9.2 Isocratic Elution 9.2.1 Polynomial Models to Describe Retention Using Modifier Content as a Factor

The organic solvent (the modifier) content in the mobile phase is the factor most frequently optimized in RPLC [5]. It has not only a large impact on both elution Analytical Separation Science, First Edition. Edited by Jared L. Anderson, Alain Berthod, Verónica Pino Estévez, and Apryll M. Stalcup.  2015 Wiley-VCH Verlag GmbH & Co. KGaA. Published 2015 by Wiley-VCH Verlag GmbH & Co. KGaA.

200

9 Modeling of Retention in Reversed Phase Liquid Chromatography

strength and relative retention (selectivity) but it can also be easily and accurately modified over wide ranges. Hence, the importance of obtaining reliable retention models involving the modifier content as a factor. In RPLC, the retention for a solute i can be expressed in terms of solubility parameters as follows [6]: ln k i ˆ ln

t Ri

t0 t0

ˆ

νi
…δM RT

δi †2

…δS

 nS δi †2 ‡ ln nM

(9.1)

where ki is the retention factor (or relative retention) of the solute; tRi the corresponding isocratic retention time; t0 the dead time; R the gas constant (1.9865 cal/ (K mole)); T the absolute temperature (K); νi the solute molar volume (cm3/mole); δM, δS, and δi (cal0.5/cm1.5) the solubility parameters of mobile phase, stationary phase, and solute, respectively; and nM and nS the number of moles of mobile phase and stationary phase, respectively, in the column. For a binary mixture of water (w) and organic solvent (o), the mobile-phase polarity can be calculated as follows: δM ˆ …1

φ†δw ‡ φδo

(9.2)

where φ is the modifier content (usually expressed as the volume fraction). If Equation 9.2 is substituted in Equation 9.1, a quadratic relationship is obtained between the logarithm of the retention factor (usually expressed as a decimal logarithm) and the modifier volume fraction [7]: log k ˆ c0 ‡ c1 φ ‡ c2 φ2

(9.3)

The fitting parameters c0 c2 gather all the constants and solute/mobilephase/stationary-phase parameters outlined in Equation 9.1, with particular values for a given solute and chromatographic system (column/modifier). In narrow modifier content ranges, the quadratic term can be dropped. This simplifies the model to the usual linear relationship: log k ˆ c0 ‡ c1 φ ˆ log k w

Sφ

(9.4)

The intercept of the fitted straight-line refers to the extrapolated log k-value if water were used as mobile phase (log kw). The slope indicates the sensitivity of retention to changes in the modifier content, that is, the solvent elution strength (S). In wide ranges of modifier, deviations from linearity are observed, which are especially significant at the highest and lowest modifier contents. Equations 9.3 and 9.4 are also valid in the presence of specific interactions produced by mobile-phase additives at fixed concentration. Retention in ternary solvent systems, composed of water and two organic solvents (usually methanol and acetonitrile), has been described with similar equations: log k ˆ c0 ‡ c1 φ1 ‡ c2 φ2 ‡ c12 φ1 φ2 ‡ c11 φ21 ‡ c22 φ22

(9.5)

9.2 Isocratic Elution

where φ1 and φ2 are the contents for the two organic solvents in the mobile phase [8,9]. The main purpose of using ternary systems in RPLC is the possibility of extending the fine-tuning of the selectivity with regard to binary systems. The search for improved models has not yet finished. The topic was reviewed in 2009 by Nikitas and Pappa-Louisi [10]. The most reported models are logarithmic. However, there are other proposals. In this regard, the excellent accuracy, when applied to RPLC, of a nonlogaritmic model proposed by Jandera for normal-phase liquid chromatography, deserves to be mentioned [11]: 1 ˆ ‰a ‡ b φŠn k

(9.6)

9.2.2 Polarity Models

Good linear relationships have also been found by relating log k to polarity parameters other than the solubility parameter that also depend on φ, such as the Dimroth–Reichardt parameter, ET(30), or its normalized value, E N T (30). Its range of linearity is, however, limited (e.g., 0–80% v/v for acetonitrile and 20–100% v/v for methanol). A derived normalized parameter (P N M ) was proposed by Bosch and Rosés [12,13] to extend the linearity to the full range of mobilephase compositions (0–100%). For acetonitrile–water mixtures, P N M is defined as PN M ˆ 1:00

2:068 φ 1 ‡ 1:341 φ

(9.7)

and for methanol–water as PN M ˆ 1:00

1:33 φ : 1 ‡ 0:47 φ

(9.8)

It is possible to describe the retention with a linear model that isolates the polarity contributions of the three agents involved in the separation process (i.e., solute, stationary phase, and mobile phase): log k ˆ …log k†0 ‡ pS …P N M

PN S†

(9.9)

where pS and PN M are polarity descriptors for the solute and mobile phase, respectively, and (log k)0 and P N S account for the stationary-phase polarity. The term (log k)0 represents the retention in a hypothetical mobile phase with the N same polarity as the stationary phase (PN M ˆ P S ). However, the separation of the polarities of solute, mobile phase, and stationary phase in Equation 9.9 is not

201

202

9 Modeling of Retention in Reversed Phase Liquid Chromatography

perfect: although pS mainly gathers the solute contributions, it is not an absolute value, rather a relative measurement that depends on the column environment (mobile phase and stationary phase). The same holds for (log k)0 and P N S. An interesting advantage of Equation 9.9 is that it includes only one solute N descriptor, ps, and two column descriptors, (log k)0 and P N S (P M is obtained from the mobile-phase composition). This facilitates the transfer of k-values to other columns and modifiers by establishing simple correlations between pS values, for small sets of selected reference compounds. When (log k)0 and P N S are fitted individually for each solute, Equation 9.9 is simplified to N log k ˆ q ‡ ps P N m ˆ c0 ‡ c1 P m

(9.10)

Equations 9.4 and 9.10 are both two-parameter models, but the polarity model (Equation 9.10) is valid for a wider range of modifier content. In fact, the prediction accuracy of this model is comparable to that of the quadratic model (Equation 9.3) (Figure 9.1) [13]. 9.2.3 pH as an Experimental Factor

For ionizable compounds, the pH is an additional experimental factor with a large influence on retention and selectivity. When the analyzed mixture contains one or more compounds with acid–base behavior, pH tuning can offer unique opportunities to improve resolution. However, modeling the changes in retention with pH is particularly cumbersome [14,15]. Not surprisingly, a widely extended practice consists of fixing the pH at a convenient value, which means that the benefits of this experimental factor on selectivity are missed. At fixed modifier content, RPLC retention of ionizable compounds is a weighted mean of the behavior of the basic and acidic species: k ˆ k A δA ‡ k HA δHA ˆ k A

1 Kh k A ‡ k HA K h ‡ k HA ˆ 1‡K h 1‡K h 1‡K h

(9.11)

where k is the measured retention factor, kA and kHA are the retention factors for the basic and acidic species, respectively, δA and δHA are their molar fractions, and h is the molar proton concentration. For convenience, the acid–base equilibrium is here characterized by the apparent protonation constant, which is the reciprocal of the acid–base dissociation constant (K ˆ K a 1 ). This constant is affected by all interactions of both the acidic and the basic species with the components of the mobile and stationary phases. Since the retention for each species (kA and kHA) is different, a sudden change in retention will happen at pH values close to log K (Figure 9.2). Depending on the charge and polarity of the acidic and basic species, the retention may decrease, increase, or remain constant with the pH (note that the charged species

9.2 Isocratic Elution

3 (a)

log kpred

2

1

0

-1 3 (b)

log kpred

2

1

0

-1 3 (c)

log kpred

2

1

0

-1 -1

0

1

log kexp

2

3

Figure 9.1 Comparison of the accuracy of predictions of the retention of 152 compounds, eluted with methanol–water mixtures (n = 745), provided by several retention models: (a) Equation 9.4, (b) Equation 9.3, (c) Equation 9.10. (Reproduced with permission from Ref. [13].)

are less retained). The protonation of a solute is produced along at least two pH units (within pH = log K ± 1). Therefore, for conventional RPLC columns, where the working pH range is relatively narrow (3–7), the change in retention will be fully sampled only for some solutes. As a result of the incomplete information, the retention of the acidic and basic species should be extrapolated.

203

204

9 Modeling of Retention in Reversed Phase Liquid Chromatography

k

9 8 7 6 5 4 (b)

3

(c)

(d)

2 1

(e)

(a)

0 2

k

4

6

8

10

12

6 (f) 5 (i)

4 (g) 3 2 (h) 1 0 2

4

6

8

10

12

pH Figure 9.2 Dependence of the retention with pH for several acidic (a–e) and basic (f–i) compounds, eluted with 40% acetonitrile. Compounds (log K in 40% acetonitrile is given): (a) 2-nitrobenzoic acid (3.59), (b) naphthoic acid (5.09), (c) 2-nitrophenol

(7.91), (d) 3,5-dichlorophenol (9.34), (e) phenol (11.61), (f) 4-chloroaniline (3.11), (g) 4-methylaniline (4.58), (h) 2,4,6-trimethylpyridine (6.59), and (i) N,N-dimethylbenzylamine (8.11). The working pH range for conventional columns is marked. (Data reproduced from Ref. [15].)

9.2 Isocratic Elution

The following transformations of Equation 9.11 allow a more convenient fitting: …k A =k HA † ‡ K h f ‡K h‡f K h k HA ˆ 1 ‡ K h 1‡K h   Kh ˆ f ‡ …1 f † k HA 1‡K h

kˆ

f Kh

k HA

  Kh log k ˆ log k HA ‡ log f ‡ …1 f † 1‡K h   log K 10 h ˆ log k HA ‡ log f ‡ …1 f † 1 ‡ 10log K h

(9.12)

(9.13)

To account for the simultaneous effect of the modifier content and pH, the equations that describe retention with regard to both factors should be combined. Also, the influence of the modifier content on the protonation constant should be considered. The interaction of both modifier content and pH twists and shifts the retention surfaces, complicating the treatment notably. A linear function of log k and log K versus modifier content can be assumed for simplicity:   10…log K w ‡mφ† h …1 f † log k ˆ log k HA ‡ log f ‡ 1 ‡ 10…log K w ‡mφ† h (9.14)   10…log K w ‡mφ† h ˆ …log k w;HA Sφ† ‡ log f ‡ …1 f † 1 ‡ 10…log K w ‡mφ† h where kw,HA and Kw are the retention of the acidic species and protonation constant in pure water. If the dependence of the retention factor and protonation constant on the modifier content is assumed to be quadratic, log k ˆ log k w ‡ S φ ‡ T φ2

(9.15)

log K ˆ log K w ‡ Q1 φ ‡ Q2 φ2

(9.16)

From Equation 9.11, it follows: kˆ

2 2 k w;A 10…SA φ‡T A φ † k w;HA K w 10‰…SHA ‡Q1 †φ‡…T HA ‡Q2 †φ Š h ‡ 2 2 1 ‡ K w 10…Q1 φ‡Q2 φ † h 1 ‡ K w 10…Q1 φ‡Q2 φ † h

(9.17)

where S, T, Q1, and Q2 are model parameters. The fitting is numerically easier if Equation 9.17 is rewritten as follows: 10…RA ‡SA φ‡T A φ † ‡ 10…RHA ‡SHA φ‡T HA φ † h 2 1 ‡ 10…Q0 ‡Q1 φ‡Q2 φ † h 2

kˆ

2

(9.18)

205

206

9 Modeling of Retention in Reversed Phase Liquid Chromatography

The dependence of retention on pH can be established most conveniently by using the polarity model (Equation 9.10). Thus, the combination of Equation 9.10 and the left-hand term in Equation 9.14 gives rise to  log k ˆ c0 ‡ c1 P N ‡ log f ‡ M

10…log K w ‡mφ† h …1 1 ‡ 10…log K w ‡mφ† h

 f†

(9.19)

where both c0 and c1 are coefficients of the acidic species.

9.3 Dead Time Estimation

The void volume in liquid chromatographic systems has been defined as “the volume of mobile phase that fills the space between the injector and the detector cell, which includes the accessible interstitial or interparticle volume (the volume between packed particles) and the intraparticle volume (the volume of the particle pores accessible to the mobile phase), as well as the volume of tubing and any other component in the system (the extra–column volume)” [16,17]. Related with this concept is the dead time, which is the time that an ideal unretained compound (i.e., a compound that does not interact with the stationary phase) needs to cross the distance between the injector and the detector cell when eluted at a constant flow rate. The dead time estimation in chromatographic systems is the basis for the calculation of retention factors (left-hand term in Equation 9.1). The calculation of the retention factor implies moving the origin of retention times to the time for an unretained compound, and division of the resulting net time by the dead time. The retention factor normalizes the retention and allows for comparisons among different column lengths or different flow rates for the same column. An accurate knowledge of retention factors is needed in the prediction of retention and resolution for the optimization of chromatographic separation, in the estimation of partition coefficients, selectivity factors, and other thermodynamic parameters, and in the establishment of correlations with several chemical–physical properties. However, dead time estimation involves two main difficulties, especially in RPLC, as follows: i) The stationary phase adsorbs a certain amount of modifier, forming a layer and thus reducing the accessible volume inside the packing. ii) Solute molecules may be partially or completely excluded from the particle pores. Since this exclusion effect is due to the volume and shape of the solute molecules, different solutes may have different associated dead times. To this size exclusion effect, the electrostatic exclusion of charged solutes should be added.

9.3 Dead Time Estimation

The methods that have been reported for dead time estimation have been classified as static and dynamic. In static methods, there is no flow and the column is kept at atmospheric pressure. In dynamic methods, the mobile phase is flowing, and there is a decreasing linear pressure gradient along the column. The results yielded by different methods differ because different properties are measured, to which the experimental uncertainties should be added. The discussion and controversy on dead time estimation have been kept alive along decades. It can be said that there is no universally accepted method for the accurate estimation of dead time yet. 9.3.1 Static Methods

Static methods are not commonly used owing to several problems that are discussed next. The most usual static method is the pycnometry or weight difference method. This consists in filling the packed column successively with two solvents of sufficiently different density, such as carbon tetrachloride and methanol, and weighting. The dead volume is then obtained from the differences in density and weight. In another method, the column is flushed with water and dried with a nitrogen stream, followed each by weighting. Both methods ignore the solvation of the stationary phase by the components of the mobile phase. Thus, unless a correction is made for the solvation layer, an error in column dead volume arises. In addition, a small difference between two large measurements (i.e., the weight of the column filled with two different solvents, or filled with water and dried) has a large uncertainty. The one solvent method (with water) is also impractical because, once the column is dried, it is often irreversibly damaged. 9.3.2 Dynamic Methods

Dynamic methods are divided into direct and indirect. Direct methods include the injection of unretained compounds, such as mobile-phase solvents or markers [17]. Thus, a baseline disturbance is obtained by injecting water, an organic solvent, or a solution with a composition slightly different from the mobile phase. A refractive index detector is not necessary to measure the disturbance, since UV detectors frequently evidence changes in the refractive index at low wavelengths. The injection of markers, either organic (e.g., acetone, N,N-dimethylformamide, nitrobenzene, picric acid, or uracil) or inorganic (e.g., KI, KNO3, NaCl, NaNO3, and NaNO2), is the most widespread method owing to its simplicity. Usually, the possible retention of the marker is neglected. The ideal marker should be small enough to penetrate the whole accessible volume in the stationary phase, and hydrophilic enough to be unretained. However, in practice, all

207

208

9 Modeling of Retention in Reversed Phase Liquid Chromatography

markers are either partially excluded from the particle pores or slightly retained. Uracil and KBr seem to be the best markers. Indirect methods are also called “mathematical methods”. These methods obviate the difficulties associated with the selection of truly unretained compounds. The most known indirect method is based on the elution of homologous series [16,17]. The approach is based on an assumed linear relationship between the logarithm of retention factor and the homolog carbon number, nC: log k ˆ c0 ‡ c1 nC

(9.20)

which is combined with the left-hand term of Equation 9.1, as follows: t R ˆ t 0 …1 ‡ k† ˆ t 0 …1 ‡ k 0 ec1 nC †

(9.21)

where k0 is the residual retention factor for nC = 0, and c1 the slope in Equation 9.20. The parameters t0, k0, and c1 in Equation 9.21 are fitted by nonlinear regression. The homologous series method is more time-consuming than dynamic methods based on unretained compounds, and the assumed linear relationship is not always valid along the whole series. Deviations from linearity have been observed for both the smaller homologs and the homologs exceeding the alkyl chain length of the bonded phase. The choice of the homologous series is based on their availability, solubility in the mobile phase, retention, and detection. The estimated dead time depends significantly on the number and choice of homologs. Furthermore, since the dead time is an extrapolated value, it is strongly influenced by small errors in the retention times of the homologs (it requires highly precise and accurate data) and by the applied mathematical approach. The retention times of compounds eluted at several mobile-phase compositions can also be used to estimate the dead time [18]. The compounds used for the estimation can be those for which there is an interest to calculate the retention factor. For this purpose, a variety of retention models can be used, but the linear relationship between log k and φ is the most convenient (Equation 9.4). The dead time is calculated from t R ˆ t 0 …1 ‡ k† ˆ t 0 …1 ‡ k 0 ec1φ †

(9.22)

The model parameters can be obtained through nonlinear regression, by fitting the retention times at several mobile-phase compositions within a narrow modifier content range, where the linear retention model fits properly and the column dead time is negligibly affected by the change in modifier content. The approach is strongly affected by the magnitude of the retention times being processed, as is the case for the homologous series approach. In order to get reliable estimations, the retention data of several compounds eluting under the same conditions (rather than the data of a single compound) should be treated

9.4 Effect of Temperature

altogether. It is, thus, implicitly accepted that the measured void volume is the space accessible to the solutes during elution, the accessibility to the pores of the solutes treated simultaneously is similar, and the void volume is not affected by changes in the modifier content. The approach includes an iterative algorithm to compensate for the lack of accuracy or inadequacy of some data.

9.4 Effect of Temperature 9.4.1 Van’t Hoff Equation

The influence of temperature on retention is described by the Vant’Hoff equation [19]: ln k ˆ

ΔS R

ΔH ‡ ln ϕ RT

(9.23)

where ϕ is the system phase ratio, ΔH and ΔS are the standard enthalpy and entropy changes, respectively, R is the universal gas constant, and T is the absolute temperature. For sufficiently narrow temperature ranges (of about 90 °C), where ΔH and ΔS are constant, Equation 9.23 can be transformed into a simple two-parameter equation: ln k ˆ c0 ‡

c1 T

(9.24)

where c0 and c1 are the model parameters. For wider ranges, due to the dependence of ΔH and ΔS on temperature, the plots are curvilinear and a third term is needed: log k ˆ c0 ‡

c1 c2 ‡ T T2

(9.25)

Nonlinear plots are also obtained for mixed retention mechanisms, for changes in the conformation of solute molecules (case of large molecules) or bonded chains in the stationary phase, which influence adsorption processes, or for retention mechanisms affected by secondary chemical equilibria in the mobile phase. The usefulness of temperature as a factor to improve separations is controversial [20]. At increasing temperature, retention decreases, longitudinal diffusion increases (which deteriorates the efficiency at low flow rates), and mass transfer kinetics is stimulated (which improves the efficiency at sufficiently large flow rates) [21]. The slope of the Van’t Hoff equation for several analytes may also differ to significantly affect the selectivity, occasionally giving rise to peak pair

209

210

9 Modeling of Retention in Reversed Phase Liquid Chromatography

reversals with varying temperature. Usually, these changes in selectivity are larger for ionizable or polar compounds. The changes are particularly intense for large molecules, such as proteins and soluble polymers that can exhibit diverse conformations depending on temperature. The effects of temperature and elution strength (modifier content) on the selectivity are approximately orthogonal to each other. Therefore, a pair of solutes, which are difficult to resolve by optimizing the elution strength, may be separated by optimizing the temperature. Although generally not exercising as much influence on the selectivity as the organic modifier nature and content, gradient slope or pH do, temperature optimization can be very useful in method development to improve selectivity while saving analysis time. 9.4.2 Combined Effect of Modifier Content, pH, and Temperature

For ionizable compounds, synergic interactions between the modifier content, the pH, and the temperature can be expected [22]. Modeling of retention including all significant interactions for a group of target compounds is challenging, since a unique design with a necessarily rather large number of experiments is needed to satisfy the information requirements of several compounds simultaneously (often, some compounds may be deficiently sampled, Figure 9.2). The dependence of the protonation constant on temperature through the van’t Hoff equation, log K ˆ c0 ‡

c1 T

(9.26)

should be considered when combining Equations 9.13 and 9.24 to account for the simultaneous effects of pH and temperature on retention: log k ˆ c0 ‡

 c1 10c3 ‡…c4 =T † h …1 ‡ log c2 ‡ T 1 ‡ 10c3 ‡…c4 =T † h

 c2 †

(9.27)

where the model parameters c0 c4 are characteristics for a given solute and chromatographic system. The combined effect of the three factors (modifier content, pH, and temperature) can be expressed by the following eight-parameter (c0 c7) equation built using the polarity model: log k ˆ c0 ‡

 c1 10c4 ‡c5 φ‡…c6 =T †‡c7 …φ=T † h …1 ‡ c2 P N ‡ log c3 ‡ M T 1 ‡ 10c4 ‡c5 φ‡…c6 =T †‡c7 …φ=T † h

 c3 † (9.28)

As can be observed, Equations 9.19 and 9.27 are particular cases of the general description given by Equation 9.28.

9.5 Effect of Pressure

9.5 Effect of Pressure 9.5.1 Deviations of Retention Factors

In an ideal chromatographic system, the mobile-phase properties, the packing and column cartridge geometries, and other system parameters are not affected by pressure. When the system departs from ideality (at increasing pressure), deviations in retention times are observed owing to [23,24]: (i) changes in the mobile-phase volume and viscosity; (ii) deviations in Darcy’s law, which describes the flow of a fluid through a porous medium, relating the local linear velocity of an unretained solute to the specific permeability, external porosity, mobile-phase viscosity, and local pressure gradient; (iii) column expansion and packing compression, which makes the dead volume to increase; (iv) perturbation of the equilibria inside the column when the partial molar volumes of the two species involved in the partition equilibrium are different, which is especially significant for large molecular mass compounds; (v) generation of a certain amount of heat (frictional heating) produced by the high velocity of a liquid through a low-permeability bed, which affects the partition equilibria, decreasing the retention and diminishing the efficiency. Discriminating the contributions of the different sources to the observed deviations from ideality is not easy. Also, the deviations in retention are only evidenced when working at varying flow rates or at varying pressures with constant flow rate (which is unusual in practice). The effects of pressure are particularly intense in fast chromatography. In this context, a particular behavior is observed with silica-based monolithic columns. These columns are made of a silica skeleton rod encapsulated in a polyether ether ketone (PEEK) tube, resulting in a network of macropores interconnected by channels through which the mobile phase percolates [25]. The volume of the channels in a monolithic column is larger than the volume between the particles in microparticle packed columns, the channel structure being also less tortuous. This allows larger flow rates, beyond those feasible for conventional packed columns at relatively low pressures. Flow rate becomes, therefore, an important factor to be considered with monolithic columns, in addition to the mobile-phase composition, in order to achieve good resolution at sufficiently low analysis time. As a result of the bed and tube elasticity, mobile-phase linear velocity for any column may vary with pressure to a certain extent. The initially cylindrical column at atmospheric pressure may be deformed under the influence of the pressure gradient and becomes approximately a truncated cone. The deformation is, however, usually negligible for columns made of stainless steel tube and pressures up to 100 MPa (it amounts only to some tenths percentage). However, the Young modulus of PEEK is 3.6 GPa compared to 210 GPa for stainless steel, meaning that PEEK is 50 times more elastic. Therefore, a column encapsulated

211

212

9 Modeling of Retention in Reversed Phase Liquid Chromatography

in PEEK undergoes a larger stress with pressure, which tends to inflate it [26]. This decreases the linear velocity, and increases the retention at relatively low pressures ( 0 (Figure 9.3b), with more significant differences from zero for the most retained compounds (i.e., c0 increases as the solute polarity decreases). Therefore, in RPLC, the pressuredependent deviation in retention time for dead time markers is smaller than for retained solutes. This makes the retention factors (k) depend on the flow rate. The tR versus 1/Fd lines for all solutes in a sample converge approximately at a common point on the 1/F-axis. This crossing point allows the estimation of an apparent flow rate (Fapp) as follows [28]: 1 1 1 ˆ ‡ F app F d ΔF c

(9.30)

ΔFc being the flow rate deviation. Several authors have reported retention models for RPLC packed columns, such as the following: y ˆ c0 ‡ c1 φ ‡ c2 F ‡ c12 φ F ‡ c11 φ2 ‡ c22 F 2

(9.31)

which includes the modifier content and flow rate as factors, with either k or log k as the response (y). However, independently of the origin of the deviations, the retention factors can be corrected to the ideal behavior as follows: k ci ˆ

t cR;i t c0 …t R;i ˆ t c0

c0;i † …t 0 c0;0 † …t 0 c0;0 †

(9.32)

9.5 Effect of Pressure

(a) 10

k 8

6

4

2

0 (b)

0

1

2

3

4

5

6

0.6

0.8

1.0

1.2

Fd , mL/min

16 2.0

14 12

1.5 1.0 0.5

tR , min

10

0.0 0.00

0.05

0.10

8 6 4 2 0

0.0

0.2

0.4

1/Fd , min/mL Figure 9.3 Dependence of the retention factor (a) and retention time (b) on the delivered flow rate for a set of compounds eluted with 20% acetonitrile from a silica-based monolithic column encapsulated in PEEK.

Compounds (from bottom to top): Nadolol, metoprolol, acebutolol, oxprenolol, labetalol, and propranolol. The inner plot in (b) magnifies the region at short retention times. (Reproduced with permission from Ref. [28].)

where t 0 and t c0 are the experimental and corrected dead time at each flow rate, and c0,i and c0,0 the deviations in retention time for a retained compound and marker, respectively. Once the retention factors are corrected, retention models (e.g., Equations 9.3 or 9.10) do not need to include the flow rate as a factor, since k ci values do not depend on the flow rate anymore [28].

213

214

9 Modeling of Retention in Reversed Phase Liquid Chromatography

9.6 Enhancing the Prediction of Retention 9.6.1 Practical Considerations

In chromatographic practice, replicated measurements of retention times at a given mobile-phase composition always fluctuate within a certain range. The possible reasons of these fluctuations are the uncertainty in injection valve operation, flow rate variations produced by defective pumping, changes in temperature, insufficient power supply stabilization, insufficient stationary-phase reequilibration, stationary-phase degradation (a long-term effect), and deviations from the nominal mobile-phase composition (e.g., due to errors in the preparation of the mobile phase or organic solvent evaporation), among others. These sources of error can deteriorate the accuracy of predictions [29]. Another point to consider in the measurement of retention times is the incorrect peak assignment when mixtures of standard compounds are injected. This can be carried out by the one-at-a-time injection of standards. Multistandard solutions can also be used, provided the differences in retention allow for identifying the peaks correctly. If needed, peak tracking can be applied by varying the concentration of the injected standards or with the aid of a selective detection technique. The reduction of the experimental effort, without deteriorating the quality of the prediction model, is also important. Usually, when only the modifier content or the temperature is taken as a factor, models yielding accurate predictions will be constructed (such as Equations 9.3, 9.4 and 9.24), even using minimal experimental designs with only one degree of freedom. Moreover, since the reliability of the models has been extensively demonstrated, in many practical instances, experimental designs with no degrees of freedom are used. Only when significant deviations are found in a further prediction, more experiments than those strictly required to construct the model are included in the design to reduce the uncertainty of the predictions. In comparison to the modifier content and temperature, the robustness of the predictions is poorer when the pH is used as a factor [22], owing to i) the complexity of modeling the retention of ionizable compounds, due to the sudden solute-dependent logarithmic drops, which requires extensive experimental designs to assure accurate descriptions. ii) the frequent different acid–base properties of the solutes in a mixture, which forces to explore different pH regions; the complexity of the required experimental design increases with the number of solutes. Therefore, designs involving the pH and one or more additional experimental factors should usually consider the data from 12 to 15 or more mobile phases (e.g., three mobile-phase compositions, each at four or five pH levels).

9.6 Enhancing the Prediction of Retention

Frequently, for complex mixtures, the retention will remain unchanged for some solutes all along the sampled pH region, while sudden drops will be observed for others. This can result in multiple peak crossings along the pH range, and unrobust separations at those pH values close to the log K of the ionizable compounds (Figure 9.2). Therefore, the experimental reproduction of a predicted separation will probably require small corrections of the mobile-phase pH. 9.6.2 Influence of the Model Regression Process on the Quality of Predictions

The regression process builds the best possible relationship between response and predictors. This is done by minimizing the sum of squared residuals (i.e., the squared difference between the actual and the predicted responses, extended to the whole set of training experiments) [1,29]. Under the perspective of its mathematical resolution, the retention models can be classified into linear and nonlinear, for which appropriate linear and nonlinear regression procedures, respectively, are required. Some nonlinear models can be transformed in general linear models, by modifying the response variable. Once this transformation is made, linear regression procedures (which are universal, safer, and easier) can be applied. It should be clarified that in this context, “linear” means “linear combinations of predictor variables”, where each term in the summation can be linear, quadratic, and so on, or a product of factors (see Equation 9.5 as an example). It should be noted that all equations describing the retention in RPLC with k as response variable are nonlinear. When the response variable is transformed (e.g., log k) to get a simpler linear model, the distribution of the uncertainties associated with the solution is affected: the set of optimal (regressed) parameters will minimize the residuals for the transformed response (e.g., log k), but not for the original response (k), which is the actual interest. Naturally, this problem does not exist if the regression is carried out nonlinearly with the original response variable. Nevertheless, there is a generalized practice of ignoring these effects. The fitting bias can be fully compensated through the introduction of weights (Figure 9.4), which can be obtained by the application of the error theory:  2 @F 1 2 ˆ wˆ  ˆ …2:303 k† @f @ log k 2

(9.33)

@k where F and f are the nonlinearized and linearized functions, respectively. To reduce the uncertainty of predictions, the use of weights is recommended as a general rule. Weighted regression, however, deteriorates the accuracy of predictions when mobile phases of large elution strength are used.

215

9 Modeling of Retention in Reversed Phase Liquid Chromatography

10

(a)

0

-10 10

(b)

0

kexp – kpred

216

-10 10

(c)

0

-10 10

(d)

0

-10

1

10

100

kexp Figure 9.4 Uncertainty in the prediction of the retention factors for a set of 13 phenols eluted with 9 mobile phases in the 20–100% acetonitrile range, according to Equation 9.4

(a, c) and Equation 9.3 (b, d), after unweighted (a, b) and weighted (c, d) linear regression. (Reproduced with permission from Ref. [3].)

9.7 Gradient Elution 9.7.1 Integration of the Fundamental Equation for Gradient Elution

The effect of a gradient of organic modifier on retention is usually described through an integral equation, which is called the fundamental equation for gradient elution [30,31]. Z t0 ˆ 0

tg t0

dt k…φ…t††

(9.34)

9.7 Gradient Elution

where φ is the modifier content at the solute location along its migration inside the column and tg the target variable (the time the solute needs to reach the column outlet), which can be calculated for any gradient provided the composite function k(φ(t)) is known. This implies two nested equations: the dependence of the retention factor on the modifier content (i.e., the retention model) and the change in the modifier content with time (i.e., the gradient program). Equation 9.34 can be solved analytically only in a few cases. The most simple case is presented when both log k(φ) and φ(t) functions are linear. This is the basis of the linear solvent strength theory [31], which has been widely used to model gradient elution. This theory can be easily extended to the construction of several consecutive linear segments (i.e., multilinear gradients). From Equation 9.4 and considering a linear gradient: φ ˆ φ0 ‡ m t

(9.35)

the following equation is obtained: k ˆ kw e

Sφ

ˆ kw e

S…φ0 ‡mt†

ˆ kw e

Sφ0

e

Smt

ˆ k0 e

Smt

(9.36)

where k0 is the retention factor at the beginning of the gradient. Equation 9.34 assumes that the gradient front reaches instantaneously the column inlet. However, actually, the solute is eluted isocratically until the gradient reaches it (i.e., during the so-called dwell time, td). Taking into account Equations 9.34 and 9.36: Z td Z tg t0  dt dt td 1
Sm…t g t 0 td † t0 ˆ ‡ ˆ ‡ e 1 (9.37) Sm…t t † d k 0 Smk 0 k0 e 0 k0 td From Equation 9.37, the retention time in gradient elution is obtained: tg ˆ

ln‰1 ‡ Sm…k 0 t 0 Sm

t d †Š

‡ t0 ‡ td

(9.38)

This outline can be applied to other integrable equations, such as Equation 9.6. 9.7.2 Nonintegrable Retention Models

As commented, the linear retention model (Equation 9.4) is valid only within narrow ranges of modifier content. For wider ranges, a quadratic term (Equation 9.3) should be added to make a proper description of the retention. A drawback of the quadratic equation, when used in gradient elution, is that the analytical integration of Equation 9.34 is not possible or at least not straightforward. Also, independent of the retention model, the analytical integration cannot be applied with nonlinear gradient programs. However, these can be satisfactorily

217

218

9 Modeling of Retention in Reversed Phase Liquid Chromatography

emulated by approximating them to summations of several consecutive linear gradients. In any case, the numerical integration is a suitable solution, preferred by some analysts independent of the retention model and gradient program. In this case, tg is calculated by splitting the integral into multiple small isocratic steps: Z t0 ˆ 0

tg t0

dt ˆ k…t†

Z

t1 0

dt ‡ k…t†

Z

t2

t1

dt ‡ ... ‡ k…t†

Z

ti ti

dt ‡ k…t† 1

Z ti

t i‡1

dt k…t†

(9.39)

to which the dwell time should be added as in Equation (9.37). Since k(t) can be assumed to be constant inside each step, t0 can be approximated to t0 

t1 t2 t1 t i t i 1 t i‡1 t i ‡ ‡ ... ‡ ‡ k 0;1 k 1;2 k i 1;i k i;i‡1

(9.40)

with k i;i‡1 ˆ

k…t i † ‡ k…t i‡1 † 2

(9.41)

The precision in tg can be increased to any desired level by just reducing the time intervals (ti+1 ti). However, the higher the required precision, the longer the computation time. Also, there is a limitation associated with the pumping system, since gradients are generated by the instruments by approximating changes in composition to small steps. Therefore, investing effort in obtaining a precision level in tg beyond the instrument performance is useless.

9.8 Computer-Assisted Interpretive Optimization

When the chromatographer faces a new problem, some decisions should be first taken concerning nonadjustable factors (e.g., the column nature and length, and the solvent and the buffer nature). These factors are usually selected by trial and error or on the basis of prior knowledge. Next, the effects of readily adjustable properties, such as the modifier content, pH, temperature, or the gradient program, are examined by varying one or more of these factors. When the results are not satisfactory, a major change in the chromatographic system is required. This panorama suggests that finding the best chromatographic separation conditions is not easy. Several strategies for optimizing the experimental conditions have been proposed to assist the resolution of complex elution problems [1,3,4,32]. In spite of being particularly slow and inefficient, trial-and-error strategies are still frequent.

9.8 Computer-Assisted Interpretive Optimization

Many samples are, however, so complex that the protocol can be too long, and often, the best (or at least acceptable) conditions are missed. Fortunately, method development can be notably expedited with more reliable results by applying computer-assisted interpretive optimization strategies. This kind of optimization includes two steps: system modeling based on training data and prediction of the resolution through computer simulation [4]. In order to carry out a rigorous optimization, it is convenient that not only the retention but also the peak profile defined by their left and right half-widths (and eventually, the peak height) is modeled [33,34]. In the first step, the chromatographer develops a number of experiments as reduced and informative as possible, in order to fit equations that will allow the prediction of chromatograms under any new condition within the modeled experimental space. In the second step, the separation quality is prospected for a large number of separation conditions, to find out the set of conditions giving the maximal (or at least an appropriate) resolution. This is done by simulating the chromatograms for a prefixed distribution of the experimental factors being optimized. For this purpose, synthetic chromatograms constituted by the predicted individual signals of the targeted compounds in the analyzed mixture are built. The search of the best conditions implies the reduction of the information contained in the simulated chromatograms to a numerical value depicting the achieved overall separation level, which is monitored throughout the separation space. Ideally, to be meaningful, this value should correlate with the analyst’s appraisal of resolution. The mathematical function that allows the evaluation of the separation quality is called “chromatographic objective function”, which is maximized throughout the optimization process. Usually, the objective function considers only the resolution, but it may achieve additional aims, such as a short analysis time, low cost, or desirable peak profiles (i.e., high efficiencies and low asymmetries) [1,35]. Very often only one factor (e.g., the modifier content in isocratic elution) is enough to succeed in the separation. The optimization of two or more factors (e.g., the modifier content, pH, and/or temperature, or the concentration of two modifiers and pH) is less frequent, due to the increased experimental effort needed to achieve accurate predictions. Gradient elution implies finding the suitable gradient program, which is, numerically speaking, more complex. The factors that should be determined are the gradient time, the gradient shape, the initial value of the experimental factor(s), and the number of nodes and their coordinates in a multilinear gradient. The resolution of a mixture is usually faced in a first trial by looking for a unique experimental condition, able to resolve all compounds in the sample. If this is not possible, the problem can be outlined with less ambitious aims, focusing on only some compounds. In an extreme case, a single analyte can be individually optimized [4]. Optimization oriented to subsequent deconvolution is also possible in case of unavoidable partial coelution. Several software packages have been marketed to facilitate the implementation of interpretive methodologies, especially for RPLC. The development of analytical methods can be

219

220

9 Modeling of Retention in Reversed Phase Liquid Chromatography

greatly benefited with these packages, which in addition can assist untrained users to set up separations. Some examples are DryLab [36], ChromDream [37], PREOPT-W [38], and OSIRIS [39]. The user can also develop his or her own software with the aid of a spreadsheet, or using programming languages, preferably an efficient computing environment such as MATLAB or R. These tools give freedom to the analyst to implement new developments and strategies.

9.9 Stationary-Phase Characterization

The wide variety of stationary phases commercially available (several hundreds) may make the appropriate selection for method development extremely difficult. The decision is often based on previous experience, stationary-phase availability in the laboratory, or advice and availability from customary suppliers. In some cases, as commented above, trial-and-error or screening approaches are applied, which are time-consuming and expensive. More rational and less labor-intensive approaches are based on the prediction of differences in absolute retention and selectivity, considering the relative importance of the intermolecular interactions between analytes and stationary phases [40]. Ideally, the approaches should help in identifying alternative stationary phases that can provide either equivalent selectivity for routine assay procedures or very different selectivity for orthogonal separations. In any case, the analyst should be guided in the selection of the best option for a given separation. Regrettably, this possibility remains still a grand challenge in separation science, although some interesting initiatives have been developed. The developed approaches have helped in understanding the intermolecular interactions governing the separation process (see Chapter 8). The approaches enjoying more acceptability are described next. 9.9.1 Linear Solvation Energy Relationships

Linear solvation energy relationships (LSER) is a specific subset of a broader class of thermodynamic relationships known as linear free energy relationships (LFER), which describe the transfer of a solute between two solvents. The most widely accepted LSER model was developed by Abraham and coworkers [41–43]: SP ˆ c ‡ eE ‡ sS ‡ aA ‡ bB ‡ vV

(9.42)

In Equation 9.42, SP can be any free energy-related property. In chromatography, it is any measure related to solute partitioning (most often log k), and c is a system constant independent of the probe solutes. The capital letters (the solute dependent input parameters E, S, A, B, and V) represent solute descriptors that account for specific intermolecular interactions: the solute’s polarizability in excess with regard to a comparably sized n-alkane (E), dipolarity (with some

9.9 Stationary-Phase Characterization

contribution from polarizability) (S), hydrogen bond acidity (proton-donating ability) (A), hydrogen bond basicity (proton-accepting ability) (B), and molecular size (specifically, the McGowan characteristic molar volume of the solute) (V). The lower case letters (the coefficients) are system constants, which depend on the nature of both mobile and stationary phases. These coefficients are determined by measuring the solute property (SP) for a wide range of solutes with known E, S, A, B, and V values, spanning a reasonably wide range of interaction abilities, and then performing multiparameter linear least squares regression. The interpretation of the coefficients e, s, a, b and v is based on their magnitude, sign, and chemical meaning. The magnitude estimates the difference in the interaction ability of the two phases, and the sign indicates which phase has the greater ability. Regarding the chemical meaning, the coefficients represent a complementary property of the solute (e.g., solute basicity/solvent acidity). Once a column has been characterized, it is possible to predict the retention for other solutes for which the solute descriptors are known. It is important to realize that the origin of the LSER solute descriptors is not chromatographic but much more general. The LSER model treats both mobile and stationary phases as homogenous bulk phases akin to simple pure solvents. Clearly, this is an oversimplification for RPLC, where the bonded alkyl chains, the silanol groups, and clusters of sorbed solvent and water molecules embedded in the stationary phase establish a very heterogeneous environment. Also, not all kinds of interactions inside a column are considered, such as those accounting for differences in the retention for isomers or shape selectivity observed on some RPLC columns, the enhanced retention of highly fluorinated species on highly fluorinated bonded phases, and metal complexation interactions such as those taking place in zirconia-based stationary phases with certain solutes, among others. 9.9.2 Local Models for Characterizing RPLC Columns

The Abraham’s LSER approach yields a general model that contributes to our understanding of the retention process, but the predictions of solute retention are not sufficiently accurate. Snyder, Dolan, and coworkers created a new approach entirely based on solute properties captured from the analysis of RPLC data: the hydrophobic subtractive model [44,45]. This is thus a local model restricted to RPLC that has shown more successful predictions compared to the LSER model. Essentially, the information in the hydrophobic subtractive approach was obtained from RPLC retention data for a large set of solutes (about 80 solutes) on different columns, which were correlated with each other. The model recognizes that hydrophobic retention is the dominant process in RPLC, and in the absence of other retention mechanisms, plots of log k for one column versus another column, for solutes having different polarities, should be a straight line. The presence of other types of interactions gives rise to scatter in the plots.

221

222

9 Modeling of Retention in Reversed Phase Liquid Chromatography

Deviations from the primary correlation are then assigned to specific properties of the solutes and columns. The hydrophobic subtractive model is outlined as follows:   ki log α ˆ log (9.43) ˆ η´ H σ ´ S ‡ β´ A ‡ α´ B ‡ κ ´ C k ethylbenzene Greek letters represent properties of the solute relative to the values for ethylbenzene, which is taken as reference. Capital letters represent column properties relative to a hypothetical average type B C18 column. Any column which behaves identically to the hypothetical reference column will have H = 1, with the value for all other parameters being zero. The η´ H term describes the hydrophobicity contribution to the relative retention; the σ ´ S term indicates the contribution from the steric resistance to the insertion of the analyte into the stationary phase or inability of solute molecules to penetrate freely into the solvated stationaryphase volume because of its size and/or shape (σ ´ represents the solute bulkiness); A is the stationary-phase hydrogen bond acidity, which combines with the solute hydrogen-bond basicity β´ , while the term in B represents the stationaryphase hydrogen-bond basicity and α´ the solute hydrogen-bond acidity. The last term (κ ´ C) reflects the ion-exchange properties of the packing, which are attributed to the surface silanols. The five parameters are of primary practical interest because they determine the selectivity and applicability of most RPLC stationary phases. The H, S, A, B, and C parameters for at least 400 stationary phases have been evaluated, mainly of type-B-based silica C18. For phenylpropyl and cyanopropyl columns, two additional terms are required, π ´ P and μ´ D, referring to p–p and dipole–dipole interactions, respectively. The limitation of the hydrophobic subtractive approach lies on the fact that the stationary-phase properties were initially obtained in a single type of mobile phase (50% v/v acetonitrile). Later on, it was found that for columns that are similar to each other the solvent effect can be ignored. It should be also noted that all model parameters are temperature dependent, and the κ´ C term is pH dependent. Therefore, C-values at both pH 3 and 7 were obtained. Another popular method for stationary-phase characterization was proposed by Tanaka and coworkers [40]. The initial test was modified by Euerby and has been established as an industrial standard test to assess selectivity and performance differences between HPLC columns. In the so-called Tanaka–Euerby test, the probes are pentylbenzene, butylbenzene, triphenylene, o-terphenyl (whose retention is measured with 80% methanol), and benzylamine, phenol, and caffeine (whose retention is measured with 30% methanol). Columns are characterized by six parameters: retention capacity or absolute hydrophobicity, which depends on the carbon load and stationary-phase surface area, measured by the k-value for pentylbenzene; hydrophobic selectivity or methylene group selectivity, indicated as the ratio of k-values for pentylbenzene and butylbenzene; shape selectivity as the ratio of k-values for triphenylene and o-terphenyl; hydrogen bonding as the ratio of k-values for caffeine and phenol measured in unbuffered

References

mobile phase; and total and acidic ion-exchange capacity as the ratio of k-values for benzylamine and phenol at two pH values (2.7 and 7.6). The latter three tests are of particular interest for the analysis of basic solutes.

References 1 García-Alvarez-Coque, M.C., Torres-

2

3

4

5

6

7

Lapasió, J.R., and Baeza-Baeza, J.J. (2006) Models and objective functions for the optimisation of selectivity in reversedphase liquid chromatography. Anal. Chim. Acta, 579, 125–145. Baeza-Baeza, J.J., Ortiz-Bolsico, C., TorresLapasió, J.R., and García-Alvarez-Coque, M.C. (2013) Approaches to model the retention and peak profile in linear gradient liquid chromatography. J. Chromatogr. A, 1284, 28–35. Torres-Lapasió, J.R., Rosés, M., Bosch, E., and García-Alvarez-Coque, M.C. (2000) Interpretive optimisation strategy applied to the isocratic separation of phenols by reversed-phase liquid chromatography with acetonitrile–water and methanol– water mobile phases. J. Chromatogr. A, 886, 31–46. Torres-Lapasió, J.R. and García-AlvarezCoque, M.C. (2006) Levels in the interpretive optimisation of selectivity in high-performance liquid chromatography: a magical mystery tour. J. Chromatogr. A, 1120, 308–321. Ramis-Ramos, G. and García-AlvarezCoque, M.C. (2013) Solvent selection in liquid chromatography, in Liquid Chromatography: Fundamentals and Instrumentation (eds S. Fanali, P. Haddad, C.F. Poole, P.J. Schoenmakers, and D. Lloyd), Elsevier, Amsterdam, pp. 223–249. Schoenmakers, P.J., Billiet, H.A.H., and de Galan, L. (1982) The solubility parameter as a tool in understanding liquid chromatography. Chromatographia, 15, 205–214. Schoenmakers, P.J., Billiet, H.A.H., and de Galan, L. (1983) Description of solute retention over the full range of mobile phase compositions in reversed-phase liquid chromatography. J. Chromatogr., 282, 107–121.

8 Pappa-Louisi, A., Nikitas, P., and

9

10

11

12

13

14

15

Karageorgaki, M. (2005) Retention prediction in ternary solvent reversedphase liquid chromatography systems based on the variation of retention with binary mobile phase composition. J. Chromatogr. A, 1091, 21–31. García-Lavandeira, J., MartinezPontevedra, J.A., and Cela, R. (2012) A binary-like approach for the computer assisted method development of isocratic and programmed ternary solvent elutions in reversed-phase liquid chromatography. J. Chromatogr. Sci., 50, 33–42. Nikitas, P. and Pappa-Louisi, A. (2009) Retention models for isocratic and gradient elution in reversed-phase liquid chromatography. J. Chromatogr. A, 1216, 1737–1755. Jandera, P. (1998) Optimisation of gradient elution in normal-phase highperformance liquid chromatography. J. Chromatogr. A, 797, 11–22. Bosch, E., Bou, P., and Rosés, M. (1994) Linear description of solute retention in reversed-phase liquid chromatography by a new mobile phase polarity parameter. Anal. Chim. Acta, 299, 219–229. Torres-Lapasió, J.R., García-AlvarezCoque, M.C., Rosés, M., and Bosch, E. (2002) Prediction of the retention in reversed-phase liquid chromatography using solute-mobile phase–stationary phase polarity parameters. J. Chromatogr. A, 955, 19–34. Schoenmakers, P.J. and Tijssen, R. (1993) Modelling retention of ionogenic solutes in liquid chromatography as a function of pH for optimization purposes. J. Chromatogr. A, 656, 577–590. Torres-Lapasió, J.R., García-AlvarezCoque, M.C., Bosch, E., and Rosés, M. (2005) Considerations on the modelling and optimisation of resolution of ionisable compounds in extended pH-range

223

224

9 Modeling of Retention in Reversed Phase Liquid Chromatography

16

17

18

19

20

21

22

23

24

columns. J. Chromatogr. A, 1089, 170–186. Rimmer, C.A., Simmons, C.R., and Dorsey, J.G. (2002) The measurement and meaning of void volumes in reversedphase liquid chromatography. J. Chromatogr. A, 965, 219–232. Pous-Torres, S., Torres-Lapasió, J.R., and García-Alvarez-Coque, M.C. (2009) Performance of markers and the homologous series method for dead time estimation in reversed-phase liquid chromatography. J. Liq. Chromatogr. Rel. Technol., 32, 1065–1083. Pous-Torres, S., Torres-Lapasió, J.R., Baeza-Baeza, J.J., and García-AlvarezCoque, M.C. (2009) Alternating iterative regression method for dead time estimation from experimental designs. Anal. Bioanal. Chem., 394, 625–636. Dolan, J.W. (2002) Temperature selectivity in reversed-phase high performance liquid chromatography. J. Chromatogr. A, 965, 195–205. Heinisch, S. and Rocca, J.L. (2009) Sense and nonsense of high-temperature liquid chromatography. J. Chromatogr. A, 1216, 642–658. Baeza-Baeza, J.J., Dávila, Y., FernándezNavarro, J.J., and García-Alvarez-Coque, M.C. (2012) Measurement of the elution strength and peak shape enhancement at increasing modifier concentration and temperature in RPLC. Anal. Bioanal. Chem., 404, 2973–2984. Pous-Torres, S., Torres-Lapasió, J.R., Baeza-Baeza, J.J., and García-AlvarezCoque, M.C. (2007) Combined effect of solvent content, temperature and pH on the chromatographic behaviour of ionisable compounds. J. Chromatogr. A, 1163, 49–62. Martin, M. and Guiochon, G. (2005) Effects of high pressure in liquid chromatography. J. Chromatogr. A, 1090, 16–38. Fekete, S., Veuthey, J.L., McCalley, D.V., and Guillarme, D. (2012) The effect of pressure and mobile phase velocity on the retention properties of small analytes and large biomolecules in ultra-high pressure liquid chromatography. J. Chromatogr. A, 1270, 127–138.

25 Svec, F. and Huber, C.G. (2006)

26

27

28

29

30

31

32

33

34

Monolithic materials: promises, challenges, achievements. Anal. Chem., 78, 2100–2107. Pous-Torres, S., Torres-Lapasió, J.R., RuizAngel, M.J., and García-Alvarez-Coque, M.C. (2010) Origin and correction of the deviations in retention times at increasing flow rate with Chromolith columns. J. Chromatogr. A, 1217, 5440–5443. Fallas, M.M., Hadley, M.R., and McCalley, D.V. (2009) Practical assessment of frictional heating effects and thermostat design on the performance of conventional (3 μm and 5 μm) columns in reversed-phase high-performance liquid chromatography. J. Chromatogr. A, 1216, 3961–3969. Pous-Torres, S., Torres-Lapasió, J.R., RuizÁngel, M.J., and García-Alvarez-Coque, M.C. (2011) Correction of the deviations in the retention times with Chromolith columns associated to the flow rate: implications in the modelling of the retention behaviour. J. Sep. Sci., 34, 931–938. Concha-Herrera, V., Vivó-Truyols, G., Torres-Lapasió, J.R., and García-AlvarezCoque, M.C. (2004) Enhancement of retention predictions in reversed-phase liquid chromatography using reference compounds. Anal. Chim. Acta, 518, 191–197. Snyder, L.R. and Dolan, J.W. (2007) HighPerformance Gradient Elution, John Wiley & Sons. Inc., Hoboken, NJ. Snyder, L.R. and Dolan, J.W. (1998) The linear solvent strength model of gradient elution. Adv. Chromatogr., 38, 115–185. Cela, R., Ordoñez, E.Y., Quintana, J.B., and Rodil, R. (2013) Chemometric-assisted method development in reversed-phase liquid chromatography. J. Chromatogr. A, 1287, 2–22. Torres-Lapasió, J.R., Baeza-Baeza, J.J., and García-Alvarez-Coque, M.C. (1997) A model for the description, simulation and deconvolution of skewed chromatographic peaks. Anal. Chem., 69, 3822–3831. Baeza-Baeza, J.J., Ruiz-Angel, M.J., CardaBroch, S., and García-Alvarez-Coque, M.C. (2013) Half-width plots, a simple tool to predict peak shape, reveal column

References

35

36

37

38

39

kinetics and characterize chromatographic columns in liquid chromatography: state of the art and new results. J. Chromatogr. A, 1314, 142–153. Siouffi, A.M. and Phan-Tan-Luu, R. (2000) Optimization methods in chromatography and capillary electrophoresis. J. Chromatogr. A, 892, 75–106. Molnár, I. (2002) Computerized design of separation strategies by reversed-phase liquid chromatography: development of DryLab software. J. Chromatogr. A, 965, 175–194. Galushko, S.V., Kamenchuk, A.A., and Pit, G.L. (1994) Calculation of retention in reversed-phase liquid chromatography. IV. ChromDream software for the selection of initial conditions and for simulating chromatographic behaviour. J. Chromatogr. A, 660, 47–59. Cela, R. and Lores, M. (1996) PREOPT-W: a simulation program for off-line optimization of binary gradient separations in HPLC. I. Fundamentals and overview. Comput. Chem., 20, 175–191. Heinisch, S., Lesellier, C., Podevin, C., Rocca, J.L., and Tchapla, A. (1997) Computerized optimization of RP-HPLC separation with nonaqueous or partially

40

41

42

43

44

45

aqueous mobile phases. Chromatographia, 44, 529–537. Lesellier, E. and West, C. (2007) Description and comparison of chromatographic tests and chemometric methods for packed column classification. J. Chromatogr. A, 1158, 329–360. Vitha, M. and Carr, P.W. (2006) The chemical interpretation and practice of linear solvation energy relationships in chromatography. J. Chromatogr A, 1126, 143–194. Put, R. and Vander Heyden, Y. (2007) Review on modelling aspects in reversedphase liquid chromatographic quantitative structure-retention relationships. Anal. Chim. Acta, 602, 164–172. Neue, U.D. (2007) Stationary phase characterization and method development. J. Sep. Sci., 30, 1611–1627. Snyder, L.R., Dolan, J.W., and Carr, P.W. (2007) A new look at the selectivity of RPC columns. Anal. Chem., 79, 3254–3262. Zhang, Y. and Carr, P.W. (2009) A visual approach to stationary phase selectivity classification based on the Snyder–Dolan hydrophobic-subtraction model. J. Chromatogr. A, 1216, 6685–6694.

225

227

10 Normal-Phase and Polar Organic Solvents Chromatography Ahmed A. Younes, Charlene Galea, Debby Mangelings, and Y. Vander Heyden

10.1 Introduction

Chromatography is a collective term utilized for a range of physical methods for the separation of (complex) mixtures with analytical and preparative purposes. The mixture is separated via distribution of their components between two phases: a stationary phase, which is either the surface of a solid or a liquid layer fixed on a solid support, and a mobile phase, which moves through the system. The constituents of the mixture exhibit different affinities for the mobile and stationary phases. Components that display stronger interactions with the stationary phase will be retained longer than those with weaker interactions. The individual components can then be detected visually or with a detection system [1]. Chromatography was first described by the Russian botanist Mikhail S. Tswett in the early 1900s [2]. He developed a colorful separation of plant pigments through a column packed with calcium carbonate as a stationary phase. Liquid mobile phase was then continuously added to the column, allowing the sample components to be sequentially eluted based on their relative affinities for the stationary phase and mobile phases. Years after, in the late 1930s and early 1940s, Martin and Synge [3] separated some acetyl amino acids by using a water layer supported on a silica packed bed as a stationary phase. Chromatography is probably the most powerful and versatile separation technique available to the modern analyst. In a single process, a mixture can be separated into its individual components, which is important both for preparative and for analytical scenarios, and it can also provide a quantitative estimate of each constituent (analytical chromatography) [4]. A commonly applied form of chromatography is liquid chromatography (LC), where the stationary phases are either solid surfaces or more frequently liquid-like compounds (ligands) bonded on solid particles and the mobile phases are liquids. The mobile phase can also be a gas (gas chromatography (GC)) or a supercritical fluid (supercritical fluid chromatography (SFC)). The stationary phase might also be a liquid that is nonmiscible with the mobile phase as is the case in countercurrent chromatography (CCC). Analytical Separation Science, First Edition. Edited by Jared L. Anderson, Alain Berthod, Verónica Pino Estévez, and Apryll M. Stalcup.  2015 Wiley-VCH Verlag GmbH & Co. KGaA. Published 2015 by Wiley-VCH Verlag GmbH & Co. KGaA.

228

10 Normal-Phase and Polar Organic Solvents Chromatography

The types of compounds or mixtures that can be separated by chromatography are also very diverse. Gases, volatile compounds, or substances that can be made volatile upon derivatization can be analyzed by GC. In LC, very different types of compounds can also be analyzed ranging from small inorganic ions to small organic molecules (e.g., amino acids, drug molecules, and secondary plant metabolites) to large and very large (polymeric) compounds, which could either have a natural (peptides, proteins, and polynucleotides) or a synthetic (plastics) origin. The analysis of these diverse compounds requires different mobile- and/or stationary-phase properties. Different physicochemical processes, such as adsorption, ion-exchange, partition, exclusion, and affinity, are applied in the separation of compounds. They give rise to different separation modes, which are shortly discussed further. LC employs a liquid mobile phase. Liquid–solid chromatography utilizes a solid stationary phase, and the major mechanisms of retention are adsorption or ionexchange. Liquid–liquid chromatography, on the other hand, uses a stationary phase, usually bonded to a solid support, having the properties of a liquid. In this case, the major mechanism of retention is partition. In earlier days, the liquid mobile phase percolated through a column packed with the stationary phase by means of gravity. Nowadays, high pressures are applied to allow the liquid to flow through columns packed with small stationary-phase particles (usually 3–5 μm I.D.). Alternatively, a monolith, which is a single rod of porous silica (sometimes a porous polymer), can also be used. Owing to the capability of yielding very narrow peaks of the sample components, the technique is named after high-performance liquid chromatography (HPLC). However, the continuous developments in instrumentation and packing materials led to improvements in separation, identification, purification, and quantification of molecules, and the name was changed to high-pressure liquid chromatography (HPLC). At this moment, the application of HPLC, in areas such as pharmaceutical sciences, life sciences, foods, polymers, and forensics, is widespread. To improve separation efficiency, sub-2 μm stationary phases were recently developed [5]. Analyses on such phases require pressures that cannot be handled by the regular HPLC instrumentation. This gave rise to the development of a new type of chromatography, ultrahigh-performance liquid chromatography (UHPLC). 10.2 HPLC Retention and Separation Mechanisms

The classification of the different HPLC separation modes can be based on the principle with which a chemical compound interacts with the chromatographic system. In general, three primary chemical properties, that is, polarity, ionic charge, and molecular size, determine the different HPLC separation modes. 10.2.1 Polarity-Based Separations

The structure of the chromatographed molecule determines whether it is polar or nonpolar. The larger the electronegativity differences between atoms in a

10.2 HPLC Retention and Separation Mechanisms

Table 10.1 Summary of intermolecular interactions that can take place between solute and solvent molecules in LC systems. Interaction

Description

Dispersion

A temporary attractive force that results when the electrons in two adjacent atoms occupy positions that make the atoms form temporary dipoles

Dipole

The attraction between two polar molecules having permanent or induces dipole moments, that is, charge separation within the molecule

Hydrogen bonding

Interactions between a proton donor and a proton acceptor molecule

Dielectric

Interaction of sample ions with liquids of high dielectric constants

bond, the more polar the bond is. Interactions that can take place between the chromatographed molecules and its surrounding phase are summarized in Table 10.1. Molecules with similar polarities attract and interact with each other, while those with dissimilar polarities repel one another, that is, “like attracts like” or “like dissolves like.” Chromatographic separations based on polarity depend upon the extent of interaction of the molecules to be separated with the stationary and mobile phases. Therefore, to develop a successful chromatographic system, the competition of the mixture’s components should be kept in mind when choosing mobile and stationary phases with different polarities. Consequently, compounds in the mixture having their polarity comparable to that of the stationary phase will elute late because they interact more strongly with the stationary phase. On the other hand, compounds with polarities resembling that of the mobile phase will elute earlier. In this way, the differences in the relative attraction or interaction of each compound with each phase will result in different speeds of the analytes through the stationary phase, and separation would be achieved. The rule “like attracts/dissolves like” will determine which analytes are retained longer and which proceed at a faster speed. The different types of polarity-based chromatographic systems are summarized in Table 10.2. Two well-known primary separation modes, normal-phase and reversed phase chromatography, are based on polarity. Normal-phase liquid chromatography (NPLC) uses a polar stationary phase and a nonpolar mobile phase. Sample components are separated according to their relative polarities, that is, the more polar a component, the longer its retention time. The most widely used solvents in this mode are hexane or heptane, usually mixed with polar modifiers, such as methanol, acetonitrile, dichloromethane, isopropanol, or ethanol. In comparison to heptane, isohexane, and cyclohexane, the use of hexane is much unhealthier; also, acetonitrile is not soluble in hydrocarbons, while methanol and ethanol are only partially miscible with hydrocarbons. The addition of more polar solvents to the mobile phase increases its polarity and decreases the retention times of the late-eluting polar analytes. The stationary phases most frequently used for

229

230

10 Normal-Phase and Polar Organic Solvents Chromatography

Table 10.2 Polarities of the stationary and mobile phases for different polarity-based chromatographic systems. Mode/ Phases

Normal phase

Reversed phase

Aqueous normal phase/ HILIC

Hydrophobic Interaction

Ion-pair

Micellar Liquid

Polar Organic solvent

Stationary Polar phase

Nonpolar

Polar

Nonpolar

Nonpolar

Nonpolar

Nonpolar

Mobile phase

Polar and water

Polar and water

Water and salts

Polar with ionpairing reagents

Micellar

Polar without water

Nonpolar

NPLC are untreated porous silica gel (bare silica (SIL)) or silica chemically derivatized with polar functional groups at the surface. Common examples of such groups are the aminopropyl group (amino phase), the 1,2-dihydroxypropyl group (diol phase), or the cyanopropyl group (cyano or CN phase). The term reversed phase liquid chromatography (RPLC) is used to describe the chromatographic mode that uses a polar mobile phase and a nonpolar stationary phase. The mobile phase is generally a mixture of water and a miscible polar organic solvent, such as acetonitrile, methanol, and/or tetrahydrofuran, commonly referred to as organic modifiers. The retention of compounds increases as the percentage of modifier (the organic solvent) decreases. The most popular stationary phases are surface-modified silica, RMe2SiCl, where R most often is a straight chain alkyl group, such as octadecyl C18H37 (ODS or C18), octyl C8H17 (C8) or hexyl C6H13 (C6). The elution order of compounds is the opposite of what is seen in normal phase, that is, components elute in order of decreasing polarity. Both aqueous normal-phase chromatography (ANPC) and hydrophilic interaction liquid chromatography (HILIC) are types of chromatography situated between the normal- and reversed phase modes. As in NPLC, both use polar stationary phases. In HILIC, it is formed by a water-rich layer dynamically retained on either the polar surface of a solid or a layer of a polar compound that is bonded to the solid surface. In ANPC the mobile phase is based on a water-miscible organic solvent (such as acetonitrile or methanol, rather than hexane or heptane). Thus, the mobile phase is both aqueous (since water is present) and normal (because it is less polar than the stationary phase). In HILIC, the mobile phase is considered as being of the reversed phase type. In practice, the mobile phases used are the same as in ANPC and both terms may be considered synonyms. Addition of water to the organic mobile phase enhances the elution of the polar (even ionic) compounds, strongly retained under the classical normal-phase conditions. HILIC conditions are typically used to separate highly polar compounds that are not sufficiently retained under reversed phase conditions and that show a too long retention under NPLC conditions.

10.2 HPLC Retention and Separation Mechanisms

Hydrophobic interaction chromatography (HIC) is a variant of RPLC. It is used to separate large biomolecules, such as proteins. Separation using HIC is based on the reversible interaction between the biomolecules and the hydrophobic groups, for example, octadecyl, phenyl, octyl, or butyl, covalently bonded to the silica backbone of the stationary phase. The bonded stationary phases show weak hydrophobic interactions with biomolecules. However, high salt concentrations, usually ammonium sulfate or sodium chloride in water, enhance the retention of proteins. Separation of the retained proteins is achieved by a stepwise or gradient decrease in salt concentration in the mobile phase. In this way, the biomolecules elute in order of increasing hydrophobicity. Another variant of the reversed phase mode is ion-pair chromatography (IPC). IPC uses ion-pairing reagents as mobile phase additives to allow the retention and separation of ionic and highly polar substances on reversed phase columns. Ion-pairing agents are usually ionic compounds, including ionic surfactants (tensides) and others, which contain a hydrocarbon chain so that the ion-pair, formed between the analyte and the reagent, becomes more nonpolar and can be retained on a reversed phase column. The ion-pairing agent may also interact with the stationary phase through its hydrocarbon chain. Since the ionic or polar part of the ion-pairing agent is then faced toward the mobile phase, the stationary phase becomes more polar compared to the classical RPLC stationary phases. Another variant to RPLC is micellar liquid chromatography (MLC). MLC utilizes a solution of surfactant (tenside), containing either ionic or nonionic groups as in IPC, but now at concentrations exceeding the critical micellar concentration (CMC) in the mobile phase, while the stationary phases are still nonpolar entities as in the reversed phase mode. The surfactant forms micelles in the mobile phase. Monomeric surfactant molecules also interact with the stationary phase becoming part of it as already discussed with regard to IPC. Thus, a solute can interact with the mobile phase, the micelles, and the dynamically modified stationary phase. Three types of partitioning can be considered: (i) between the mobile phase and the modified stationary phase, (ii) between the mobile phase and the micelles, and (iii) between the micelles and the modified stationary phase. The MLC technique is used mainly to enhance retention and selectivity of various solutes that are not resolved or are partially resolved by RPLC [6]. Like IPC, the stationary phase becomes more polar; therefore, the retention of polar compounds is increased, while that of nonpolar is decreased. Consequently, using only one mobile phase under isocratic conditions, compounds having a broader range of polarities can be analyzed compared to similar RPLC conditions. Polar organic solvents chromatography (POSC) is another related mode. POSC, also known as nonaqueous reversed phase mode, usually uses methanol, ethanol, acetonitrile, or their combinations as mobile phase. The stationary phase is nonpolar, as in RPLC. The POSC mode combines the advantages of RPLC, that is, increased solubility and ionic and polar interactions, with those of NPLC, that is, the use of volatile solvents, which is interesting in preparative

231

232

10 Normal-Phase and Polar Organic Solvents Chromatography

chromatography where the mobile phase is evaporated to collect the separated compounds [7]. The mobile phases have large eluent strengths leading to the fast analyses of compounds. 10.2.2 Charge-Based Separations

Charge-based HPLC separation, commonly known as ion-exchange chromatography (IEC), is a chromatographic mode based on ionic interactions between a sample ion and an oppositely charged functional group on the stationary phase. Stationary phases for IEC carry either acidic (cation exchangers) or basic (anion exchangers) functional groups on their surfaces. The sample ion competes for the functional group on the stationary phase against a counterion added to the mobile phase as a salt. Thus, elution is most often accomplished by increasing the salt concentration over time. Here, ion-exchange is an interaction involved in the retention and separation of compounds. IEC is, in fact, a mode of adsorption chromatography. When adsorption chromatography is performed on an organic stationary phase, for example, derivatized cellulose, resins, or modified silica (with bonded ionic groups) and the retention mechanism is at least in part electrostatic attraction, then the technique is called IEC. When inorganic stationary phases are used, such as bare silica or alumina, then one speaks about adsorption chromatography. 10.2.3 Size-Based Separations

Size-exclusion chromatography (SEC) is an HPLC separation mode that separates molecules on the basis of their size. The separation process is based on size discrimination of sample components by a controlled porosity stationary phase. SEC is usually applied to large molecules or macromolecular complexes, such as proteins and industrial polymers. When sample components are passing through a column having the range of pore diameters that is well distributed, large molecules might not be able to penetrate the pores of the stationary phase; so, they spend less time in the column bed and elute fast, while smaller molecules can access all or a large number of pores and thus spend more time in the stationary phase and elute later. Two terms are used in relation to SEC, namely, gel filtration chromatography (GFC) and gel permeation chromatography (GPC). In GFC, an aqueous mobile phase is used to transport the sample through the column, while GPC utilizes an organic solvent as mobile phase. 10.2.4 Other Separation Mechanisms

Affinity chromatography (AC) differs from other chromatographic methods in such a way that the driving force for retention is a highly specific interaction

10.3 Normal-Phase and Polar Organic Solvents Chromatography

between the analyte of interest and a structure attached to the stationary-phase bed. These highly specific interactions can be between an antigen and an antibody, between an enzyme and a substrate, or between a receptor and a ligand. Almost all biomolecules can be purified on the basis of AC. AC can be seen as an entrapment with the target molecule becoming trapped on the stationary phase, while the other molecules will flow out of the column. Afterward, the target molecule is released from the entrapment by elution with a proper solvent. Chiral chromatography applies a chromatographic mode (e.g., GC, LC, SFC, and CC based) to separate enantiomers. Enantiomers, which are a type of stereoisomers, cannot be distinguished in achiral environments, such as a classical chromatographic setup. Therefore, either the mobile or the stationary phase should be chiral to allow the separation of enantiomers. Consequently, the socalled chiral selectors are used in the mobile phase or coating or bonded to the stationary phase creating chiral stationary phases (CSPs). The use of these latter CSPs is the most popular option. Another possibility is to derivatize the enantiomers to diastereomers, another type of stereoisomers that can be separated in an achiral environment, for instance, under classical LC conditions. In chiral chromatography, the enantiomers that need to be separated will run down the column at a different speed. One enantiomer will interact more either with the CSP or with the mobile phase (if it contains a chiral agent) than the other, and so separation will be achieved. A given type of CSP is often used in combination with different mobile phases. Here, it is the type of mobile phase used that determines the mode.

10.3 Normal-Phase and Polar Organic Solvents Chromatography

NPLC is an important chromatographic mode besides the more frequently used RPLC. NPLC is preferred to RPLC for some separation problems, including those in which the aqueous solubility of the sample compounds is limited. For instance, NPLC can easily separate both tocopherol isomers that are difficult to separate by RPLC and sugars that are poorly retained in RPLC. Sugars are not soluble when the concentration of water decreases below 70% or 60%. In addition, because of the absence of water in the mobile phase, this technique is ideal for the separation of easily hydrolyzable compounds. In POSC, by using nonaqueous mobile phases, such as methanol or acetonitrile, or their combination, the polar organic solvent accelerates the elution of many hydrophobic substances, which might not elute or elute slowly under reversed phase conditions with aqueous mobile phases. For chiral HPLC separations, the use of different solvent mixtures in NPLC and POSC also have the advantage of offering different selectivities through the use of the possible conformational changes induced on the CSP. For POSC, better UV sensitivity, better solubility of hydrophobic analytes in the mobile phases, improved LC–MS compatibility, and usually faster analysis than in NPLC is observed [8].

233

234

10 Normal-Phase and Polar Organic Solvents Chromatography

10.3.1 Retention Mechanism

NPLC and POSC share the same principles of retention, separation, and elution. In both modes, the difference in polarity between the stationary phase and the mobile phase in relation to that of the mixture components is the driving force for separation. This difference in polarity affects the partition of the mixture compounds over the mobile and stationary phases and thus their retention and separation. Mobile-phase modifiers of the short-chain alcohol type, such as ethanol or isopropanol, are often added to the mobile phase to adjust its polarity. Therefore, in NPLC the least polar compounds elute first while the most polar compounds elute last, that is, the components of a mixture are eluted in order of increasing polarity. In POSC, the nonpolar chromatographic matrix retains nonpolar solutes longer, thus the elution order is opposite to that of the normalphase mode. The distribution of the solute between the two phases depends on the properties of the stationary phase, the hydrophobicity of the solute, and the composition of the mobile phase [9]. 10.3.2 Stationary Phases

Two types of stationary phases are reported for NPLC, nonbonded (inorganic adsorbents), and bonded phases (usually modified silica). In POSC, only bonded phases are used. However, the modifications done on the silica surface for NPLC and POSC are different, as shall be discussed further. 10.3.2.1

Nonbonded Phases

Although several inorganic adsorbents, such as silica, alumina, carbon, titanium oxide, zirconium oxide, magnesium oxide, and various carbonates, have been tested as stationary phases for normal-phase chromatography, silica and alumina are the most popular. Metal oxides are highly stable at all pHs and at high temperatures, which is not the case for silica-based packings. However, silica is widely available with different particles sizes, pore diameters, and surface areas. Metal oxides are available only with limited selections of these parameters. This, together with the complex surface chemistries and lack of knowledge on the properties and selectivities of metal-based packings, make silica-based stationary phases more widely used [10]. Alumina is a polar basic adsorbent and is preferred for the separation of components that are weakly or moderately polar. However, alumina is highly active and should be treated with great caution; a small change in the pH on the alumina stationary phase may allow chemical reactions with the components of the mixture. Moreover, alumina has low theoretical plate numbers (N), variable retention times, and low sample recovery after analysis. Unlike alumina, silica gel is less active and can provide very high selectivity for many applications. Silica is also preferred because of its high sample capacity. Poor reproducibility of retention times, resulting from water adsorption

10.3 Normal-Phase and Polar Organic Solvents Chromatography

Figure 10.1 Structure of silica gel.

to the silica, is one of the main disadvantages of using silica. Improved reproducibility could be achieved by either adding 0.1–0.5% methanol or propanol to the mobile phase to minimize the effects of changes in water content or equilibrating the mobile phase with a certain intermediate concentration of water [11]. Silica consists of silicon atoms bridged three-dimensionally by oxygen atoms (Figure 10.1). Silica has various kinds of functional groups, like the free silanol groups, which are slightly acidic and which give rise to tailing of basic compounds. Silanols that are hydrogen bonded are not acidic and compounds with an OH group tend to adsorb here. Siloxanes are formed from the condensation of associated silanols [11]. There are several types of silica. Type A silicas are known to contain metal impurities, and are more stable in alkaline eluants. Metal impurities in silica largely enhance the acidity of the neighboring silanol groups. Type B silicas are known as high-purity silicas; however, residual metal still exists in them [12,13]. Type C silicas are also available. They exhibit large pore sizes and a surface that is populated by silicone–hydride (Si–H) groups, which are very stable and nonpolar. Type-C silicas have all the advantages of type-B silicas. In addition, they are applicable for the three most common modes of HPLC, namely, NPLC, RPLS, and POSC [14]. 10.3.2.2

Bonded Phases

Bonded phases are generally modified silica. The modification can take place by binding moderately polar groups, such as cyanopropyl (cyano), 1,2 dihydroxypropyl ether or aminopropyl (amino), to the silica surface (Figure 10.2a). Among the bonded phases, cyanopropyl phases are the best for general analysis because they are highly stable. Diol and amino columns can offer different selectivities, but are less stable. These moderately polar bonded phases may be used both in the normal-phase and in the reversed phase modes. However, in POSC, only chemically bonded alkyl chains (C4, C8, or C16 groups) (Figure 10.2b).

235

236

10 Normal-Phase and Polar Organic Solvents Chromatography

Figure 10.2 Bonded silica-based stationary phases: (a) with intermediately polar groups and (b) with apolar alkyl groups.

In the field of chiral separation, normal-phase, reversed phase, and, more recently, also polar organic solvents modes, are widely used. The stationary phases are different from those above. They contain the so-called chiral selectors. Coated or immobilized polysaccharide chiral stationary phases, for instance, are made with high-quality silica support onto which the polymeric chiral selector (amylose or cellulose derivatives) is physically coated or chemically immobilized [15]. The same chiral selectors are often used with different mobile phases, therefore, in different modes. 10.3.2.3

Stationary Phases and Selectivity

The selectivity factor (α) is a parameter reflecting peak spacing between two peaks in a chromatogram (Figure 10.3). The parameter is defined by Equation 10.1, which shows that the more the peaks are separated in time, the higher is the selectivity. αˆ

k 2 t r2 t o ˆ k 1 tr to

(10.1)

During method development, one is changing the factors affecting selectivity to obtain resolution among all the peak pairs. The type of stationary phase

10.3 Normal-Phase and Polar Organic Solvents Chromatography

Figure 10.3 Measuring the selectivity factor for a pair of solutes.

largely influences the selectivity because each type has its own chemical characteristics and interacts differently with the sample components. For instance, silica, which is a polar stationary phase (normal-phase mode), will retain polar compounds more than the nonpolar ones. Introduction of nonpolar groups, such as alkyl groups on silica, shifts the affinity of the column toward nonpolar compounds (polar organic solvents mode). Nonpolar compounds are then retained longer (Figure 10.4). Even a change from one brand to another within a type of stationary phase can modify the chromatographic selectivity. The mobile-phase composition also affects the selectivity considerably. This is discussed further. Column temperatures also influence the selectivity. Higher column temperatures decrease retention times and lower temperatures increase them, but it is difficult in advance to predict whether a given separation will improve or get worse by changing the column temperature. It is important to note that changes in selectivity due to temperature are rather small. However, changes in overall retention time can vary with temperature and thus controlling column temperature may be recommended to achieve reproducible retention times [11,16].

Stationary Phase C18 C8 Deceasing hydrophobicity/ Increasing polarity

C4 Cyano Phenyl Amino Silica

Figure 10.4 Polarity of different stationary phases.

Increasing retention of nonpolar analytes

237

238

10 Normal-Phase and Polar Organic Solvents Chromatography

Table 10.3 Properties of common NPLC mobile-phase solvents (based on Ref. [1]). Mobile phase

Solvent selectivity groupa)

Strength (εo silica)

Polarity index (P´ )

UV cutoff (nm)

Hexane

—

0.00

0.1

190

Toluene

VII

0.22

2.4

285

Chloroform

VIII

0.26

4.1

245

Ethyl acetate

VI

0.48

4.4

256

Acetonitrile

VI

0.52

5.8

190

Tetrahydrofuran

III

0.48

4.0

212

n-propanol

II

0.60

4.0

240

Ethanol

II

0.65

4.3

210

Methanol

II

0.70

5.1

205

Water

VIII

Very large

10.2

a) See Section 10.3.3.2 for Solvent Selectivity Group explanation.

10.3.3 Mobile Phases

For all HPLC separation modes, the choice of the mobile phase is as important as that of the stationary phase because the chemistry of both influences the separation process. A proper mobile phase, for a given separation, is the one that achieves an acceptable separation of the mixture components. In separations based on polarity, the polarity of the mobile phase should be chosen to complement the choice of the stationary phase, for example, polar mobile phases are selected for nonpolar stationary phases and vice versa. Typical solvents used in NPLC separations are listed in Table 10.3. By changing the polarity difference of the mobile phase relative to the stationary phase, retention and selectivity of compounds can be affected. 10.3.3.1

Mobile-Phase Selection

Four factors are often considered in the choice of a mobile phase for polaritybased HPLC separations. These are (1) solvent eluotropic strength (εo), (2) solvent polarity (P´ ), (3) solvent localization, and (4) solvent UV cutoff. The eluotropic strengths (εo) on silica support and the polarity indices (P´ ) for various common solvents are given in Table 10.3. Solvents with high eluotropic strength values are strong solvents, and will elute the more polar analytes more easily in NPLC. A mobile phase with a specific polarity can be obtained by mixing two or more miscible solvents. The resulting mobile phase will have an intermediate polarity. For example, the polarity index (P ´AB ) of a binary mobile phase made by combining solvents A and B is calculated as shown in Equation 10.2. P ´AB ˆ ΦA P ´A ‡ ΦB P ´B ;

(10.2)

10.3 Normal-Phase and Polar Organic Solvents Chromatography

where P ´A and P ´B are the polarity indices for solvents A and B, and ΦA and ΦB are the volume fractions for the two solvents. The polarity will determine the analysis times for a given mixture at a given stationary phase, while the different solvents involved will affect the relative selectivity of the mixture components. The solvent localization is a property that reflects the magnitude of interaction between a solvent and the stationary phase. High interactions cover the stationary-phase surface with a layer of solvent molecules leaving less or even no free adsorptive sites for the analytes and consequently no retention occurs. As water is present in all NPLC solvents under atmospheric conditions, it needs to be controlled or the active sites on the adsorbent need to be deactivated in order to avoid dramatic changes in retention times [16]. However, under ANPC or HILIC conditions, this aqueous or highly polar solvent adsorbed layer is maintained and allows partition of the mixture compounds between the stationary and the mobile phases. There are two main approaches for controlling water content on the NPLC stationary phase [16]. The first involves spiking the mobile phase directly with a little amount of water, acetonitrile, methanol, or propanol. This increases the surface homogeneity of the stationary phase by deactivating the residual silanol adsorption sites present in many silica-based stationary phases; consequently, changes in water content between mobile phases will have a minor influence. This approach faces the problem of low solubility of water in many NPLC solvents. Moreover, the addition of water, acetonitrile, methanol, or propanol changes the elution strength of the eluent. The second approach involves the complete drying of mobile phases prior to their use, thereby removing all traces of water [17]. We can summarize that nonpolar or weakly polar solvents are selected for polar stationary phases (normal-phase mode), while polar solvents are recommended for nonpolar stationary phases (polar organic solvents and reversed phase modes). For the POSC mode, the mobile phases usually are methanol or acetonitrile based. Occasionally, short-chain alcohols (ethanol, propanol, and butanol) are added to change the polarity and selectivity. If a spectrophotometric detector is employed, the solvents used for the preparation of the mobile phase should be transparent at the wavelength of detection, especially when the test method prescribes a low measurement wavelength and gradient elution is performed. Table 10.3 provides UV cutoff values for some common solvents used in HPLC separations. 10.3.3.2

Solvent Strength and Selectivity

The choice of the mobile phase composition depends on the intensity of interactions between the various solutes and the stationary phase. In HPLC polaritybased separations, large versatility in selectivity can be achieved by controlling and varying the mobile phase polarity. For example, in NPLC, if the peaks are overlapping or coeluting, that is, not all the peak pairs are resolved, switching to a mobile phase of lower polarity keeps the solutes on the column for a longer time and provides some more opportunity for their separation. In NPLC, the

239

240

10 Normal-Phase and Polar Organic Solvents Chromatography

Figure 10.5 Elution sequence obtained with different separation modes, for compounds of different polarities. Polarity order of the analytes is 4 > 3 > 2 > 1.

strength of the mobile phase increases with increasing solvent polarity. Thus, by increasing the P´ value, the solvent strength is increased and the retention time is decreased. The opposite scenario holds true for RPLC and POSC; as the solvent P´ values increase, the solvent strength decreases and the retention time increases (Figure 10.5). In addition, at a given polarity or at equal elution strengths, the solvent selectivity groups (Figure 10.6) also influence selectivity. Figure 10.6 shows the Synder’s solvent selectivity triangle with groups of solvents having similar properties located in the same circle. A summary of the solvents found in the different groups is given in Table 10.4. Solvent selectivity is controlled by the proton donor, proton acceptor, and dipole interactions of different solvents. Keeping the solvent strength constant while changing the mobile-phase components

Figure 10.6 Solvent selectivity triangle.

10.3 Normal-Phase and Polar Organic Solvents Chromatography

Table 10.4 Solvents found in the different groups in the solvent selectivity triangle. Group

Solvents

I

Aliphatic ethers, tetramethylguanidine, hexamethylphosphoric acid amide, trialkyl amines

II

Aliphatic alcohols

III

Pyridine derivatives, tetrahydrofuran amides (except formamide), glycol ethers, sulfoxides

IV

Glycols, benzyl alcohol, acetic acid, formamide

V

Methylene chloride, ethylene chloride

VI

(a) Tricresyl phosphate, aliphatic ketones and esters, polyethers, dioxane (b) Sulfones, nitriles, propylene carbonate Aromatic hydrocarbons, halo-substituted aromatic hydrocarbons, nitro compounds, aromatic ethers

VII VIII

Fluoroalkanols, m-cresol, water, chloroform

(using solvents from different circles) leads to alteration in the solvent selectivity, which in turn can lead to the separation of overlapping peaks. Solvent selectivity optimization is best explained with an example. Suppose that a separation in NPLC was carried out using a mixture of hexane and methanol (80 : 20, v/v). From Table 10.3 and Equation 10.2, it follows that the polarity strength of the resulting mobile phase is equal to (0.8 × 0.1) + (0.2 × 5.1) or P´ = 1.1. If one wants to change the selectivity of the mobile phase, it is best to use a modifier that comes from a group having different properties from those of methanol. From Table 10.4 and Figure 10.6, one sees that methanol is in group II, and those solvents in groups distant from group II are likely to result in selectivity differences. For example, solvents chloroform (group VIII) and ethylacetate (group VI) are both miscible with hexane and can be considered as possible alternatives to methanol. Suppose, chloroform is tried first. Once again, Equation 10.2 and Table 10.3 are used to calculate the volumes of hexane and chloroform that need to be mixed while keeping the solvent strength constant. This results in a mixture of 75 : 25, v/v hexane:chloroform. If this mixture does not give an acceptable separation selectivity, the process can be repeated using ethylacetate as a modifier. If mobile phases with different polar modifiers lead to a different selectivity, but not to baseline separation of all mixture compounds, then one can predict and test if mobile phases with two modifiers do not provide improved separation. Changing the nonpolar solvent instead of the polar modifier rarely leads to major changes in selectivity. 10.3.3.3

Isocratic and Gradient Elution

Considering the composition of the mobile phase during a run, two well-known elution modes are distinguished, isocratic and gradient. In the isocratic elution mode, the mobile phase’s composition is kept fixed during the chromatographic

241

242

10 Normal-Phase and Polar Organic Solvents Chromatography

analysis. Isocratic elution is typically effective in the separation of sample components that are not very different in their affinities for the stationary phase. When solutes show markedly different affinities, isocratic elution leads to an increased retention of the later eluting components. In gradient elution, the composition of the mobile phase is varied during analysis, typically from low to high eluting strengths. Gradient elution decreases the retention of the later eluting components so that they are eluted faster. Isocratic elution is characterized by band broadening, that is, the peak width of the later eluting compounds is higher than that of the earlier eluting in a chromatogram. With gradient elution, the peak width for all compounds is rather similar. For a normal-phase separation, the initially applied mobile phase is of low polarity. As the separation progresses in gradient elution, the composition of mobile phase becomes more polar. In reversed phase and polar organic solvents mode, the opposite is true, starting with a more polar mobile phase and the polarity decreases during gradient analysis. A gradient program may include sudden “step” changes in the mobile phase’s composition, as well as different slopes at different times, in order to obtain optimal separation of the mixture in minimal time (segmented gradients). For normal-phase gradient separations on silica columns, one can start with a nonpolar solvent, such as hexane, and then introduce a more polar solvent, such as ethyl acetate, for example, from 5 to 95% ethyl acetate over a gradient time interval. Depending where the analyte(s) of interest elute, the gradient can be modified to improve the resolution of all components.

10.4 Conclusions

Normal-phase and polar organic solvents liquid chromatography are separation modes based on polarity. The stationary phases used in both techniques are versatile. In general, polar stationary phases are applicable in normal-phase mode, while nonpolar phases are selected for polar organic solvents mode. The normalphase mode occasionally may have some advantages over the more widely used reversed phase mode. For instance, NPLC is compatible with hydrophobic samples such as mineral and vegetal oils or with sample extracts obtained with hydrophobic solvents, it can be used for analytes that decompose in water, it is interesting for separating enantiomers preparatively (easier evaporation of mobile phase), and for both very hydrophobic and hydrophilic analytes, it may use higher flow rates due to the low viscosity of the solvent mixtures. Columns tend to be quite stable when using nonaqueous mobile phases and large changes in selectivity are possible by altering the mobile-phase composition. Some of the drawbacks of the normal-phase mode are higher costs of purchase and disposal of solvents, controlling the solvent strength is more difficult in comparison to the reversed phase mode, lower boiling point of solvents makes NPLC systems more susceptible to bubble formation, retention may be more variable, and gradient elution can be difficult because of water uptake by unmodified silanol groups.

References

On the other hand, POSC enables the elution of hydrophobic substances, which do not elute from a reversed phase system. Moreover, the polar organic solvents mode has the advantage of better solubility of many hydrophobic analytes and improved LC–MS compatibility compared to the normal-phase mode. In recent years, NPLC and POSC have regularly been applied in chiral separations [18–21]. Under TLC conditions, the NPLC mode still seems to be the most applied to assay active compounds in pharmaceutical formulations [22].

References 1 Snyder, L.R., Kirkland, J.J., and Dolan, J.W.

2 3

4

5

6

7

8

9

(2010) Introduction to Modern Liquid Chromatography, John Wiley & Sons, Inc., Hoboken, New Jersey. Sakodynsky, K. (1970) M.S. Tswett: his life. J. Chromatogr., 49, 2–17. Martin, A.J.P. and Synge, R.L.M. (1941) Some applications of periodic acid to the study of the hydroxyamino-acids of protein hydrolysates. Biomed. J., 35, 294–315. Dong, M.W. (2006) Modern HPLC for Practising Scientists, John Wiley & Sons, Inc., Hoboken, New Jersey. Russo, R., Guillarme, D., Nguyen, D.T., Bicchi, C., Rudaz, S., and Veuthey, J.-L. (2008) Pharmaceutical applications on columns packed with sub-2 micrometer particles. J. Chromatogr. Sci., 46, 199–208. Ruiz-Angel, M.J., Garcia-Alvarez-Coque, M.C., and Berthod, A. (2009) New insights and recent developments in micellar liquid chromatography (review). Separ. Purif. Rev., 38, 45–96. Matthijs, N., Maftouh, M., and Vander Heyden, Y. (2006) Screening approach for chiral separation of pharmaceuticals IV. Polar organic solvent chromatography. J. Chromatogr. A, 1111, 48–61. Chankvetadze, B., Kartozia, I., Yamamoto, C., and Okamoto, Y. (2002) Comparative enantioseparation of selected chiral drugs on four different polysaccharide-type chiral stationary phases using polar organic mobile phases. J. Pharm. Biomed. Anal., 27, 467–478. Scott, R.P.W. and Kucera, P. (1975) Solute interactions with the mobile in liquid– solid chromatography. J. Chromatogr., 112, 425–442.

10 Nawrocki, J., Dunlap, C., Li, J., Zhao, J.,

11

12

13

14

15

16

17

McNeff, C., McCormick, A., and Carr, P.W. (2004) Part II. Chromatography using ultra-stable metal oxide-based stationary phases for HPLC. J. Chromatogr. A, 1028, 31–62. Meyer, V.R. (2010) Practical High Performance Liquid Chromatography, 5th edn, John Wiley & Sons, Ltd, West Sussex, UK. Claessens, H.A. and Van Straten, M.A. (2004) Review on the chemical and thermal stability of stationary phases for reversed-phase liquid chromatography. J. Chromatogr. A, 1060, 23–41. Nawrocki, J., Dunlap, C., McCormick, A., and Carr, P.W. (2004) Part I. Chromatography using ultra-stable metal oxide-based stationary phases for HPLC. J. Chromatogr. A, 1028, 1–30. Pesek, J.J. and Matyska, M. (2010) Recent developments in type C stationary phases: exploiting the versatility of silica hydride materials. Chromatogr. Today, 3 (Suppl), 24–26. Lammerhofer, M. and Lindner, W. (2000) Recent developments in liquid chromatographic enantioseparation, in Handbook of Analytical Separations: Separation Methods in Drug Synthesis and Purification (ed. K. Valko), Elsevier, Amsterdam, The Nederlands, pp. 337–437. Snyder, L.R., Kirkland, J.J., and Glajch, J. (1997) Practical HPLC Method Development, 2nd edn, John Wiley & Sons, Inc., New York, USA. Jandera, P. and Kucerova, M. (1997) Prediction of retention in gradient-elution normal-phase high-performance liquid

243

244

10 Normal-Phase and Polar Organic Solvents Chromatography

chromatography with binary solvent gradients. J. Chromatogr. A, 759, 13–25. 18 Du, B., Pang, L., Yang, Y., Shen, G., and Zhang, Z. (2014) Chiral liquid chromatography resolution and stereoselective pharmacokinetic study of pioglitazone enantiomers in rats. J. Chromatogr. B, 951–952, 143–148. 19 Yang, S., Li, C., Wang, S., Zhao, L., Hou, Z., Lou, H., and Ren, D. (2013) Chiral separation of two diastereomeric pairs of enantiomers of novel alkaloid-lignan hybrids from Lobelia chinensis and determination of the tentative absolute configuration. J. Chromatogr. A, 1311, 134–139. 20 Ates, H., Younes, A.A., Mangelings, D., and Vander Heyden, Y. (2013)

Enantioselectivity of polysaccharide-based chiral selectors in polar organic solvents chromatography: implementation of chlorinated selectors in a separation strategy. J. Pharm. Biomed. Anal., 74, 1–13. 21 Younes, A.A., Ates, H., Mangelings, D., and Vander Heyden, Y. (2013) A separation strategy combining three HPLC modes and polysaccharide-based chiral stationary phases. J. Pharm. Biomed. Anal., 75, 74–85. 22 Shewiyo, D.H., Kaale, E., Risha, P.G., Sillo, H.B., Dejaegher, B., Smeyers-Verbeke, J., and Heyden, Y. (2011) Development and validation of a normal-phase HPTLCdensitometric method for the quantitative analysis of fluconazole in tablets. J. Planar Chromatogr., 24, 529–533.

245

11 Inline Detectors Ramisetti Nageswara Rao and Pothuraju Nageswara Rao

11.1 Introduction

HPLC involves not only the separation of complex chemical mixtures into their components but also the visualization of the separated components. The earliest forms of liquid chromatography dealt with components of colored pigments visible to the naked eye. Later on, it was realized to be of limited use since many compounds are not colored and difficult to identify them visually. Thus, it became important to detect the colorless substances and considerable efforts were made to develop suitable detectors for LC. The purpose of a detector is not only to sense the presence but also measure the amount of a sample component in the column effluent. It is considered to be one of the most important components of LC and regarded as an input transducer which translates the changes in the chemical composition of the column effluents into electrical signals. The signals are then recorded and processed to get the desired information. In the early days, detection was often carried out by collecting fractions and analyzing them offline using gravimetric or wet methods of analysis. The first online detectors, namely, refractive index (RI), and conductivity were introduced in the 1950/ 60s. They were certainly an improvement over offline approaches; however, both were less sensitive. Over the years, the search for more sensitive detectors led researchers to adapt GC detectors for use in HPLC. However, the removal of the mobile phase solvent through evaporation limited their applicability in reality. Later, the first ultraviolet (UV) detector for HPLC was introduced. Subsequent improvements in design led to variable wavelength and diode array detectors. Over the past three to four decades, many different types of detectors have been developed, but the UV-Vis, fluorescence, refractive index, conductivity, and mass detectors are most successful. Refractive index, for example, is a property of both the solute and the mobile phase. This type of detectors are called bulk property detectors. Alternatively, the detectors such as UV-Vis and electrochemical based on the property of a solute but not solvent are called solute property detectors. The RI detector is widely used in exclusion chromatography but is less sensitive compared to UV. The workhorse detector for HPLC is the 254 nm detector. Analytical Separation Science, First Edition. Edited by Jared L. Anderson, Alain Berthod, Verónica Pino Estévez, and Apryll M. Stalcup.  2015 Wiley-VCH Verlag GmbH & Co. KGaA. Published 2015 by Wiley-VCH Verlag GmbH & Co. KGaA.

246

11 Inline Detectors

Detectors such as fluorescence, conductivity, electrochemical, and so on have been employed for specific applications. Quite a large number of devices have been used as HPLC detectors for a wide variety of applications [1–4]. 11.2 Detector Characteristics

Good detectors exhibit high sensitivity, low noise, a wide range of linear response to all types of compounds. The important characteristics required for an ideal detector include the following: a) b) c) d) e)

Sensitivity Linearity Universal or selectivity Low dead volume Nondestructive

11.2.1 Sensitivity

The sensitivity of a detector is defined as the change in the detector output as a function of analyte concentration. In a calibration curve, the slope is the sensitivity and corresponds to the ability to distinguish between nearly identical amounts of an analyte. It depends greatly on analyte properties, composition, temperature, and pH of the column effluent. While comparing different detectors, one of the most important factors is the limit of detection (LOD), which is the minimum amount of analyte that is detectable. It is determined by both the sensitivity and the detector noise. It is usually defined as the amount of analyte that yields a signal three times higher than the noise level. The value for the noise is the root-mean-square (RMS) variation in the detector output at similar timescales to the signal peak. Sensitivity, noise, and dynamic range together determine the minimum detectable amount of an analyte. Postanalysis processing eliminates the long timescale drifts to obtain accurate peak areas and avoid analytical artifacts. Modern detectors are provided with softwares for correction of baseline drift and peak integration for quantification [5]. 11.2.2 Selectivity

A highly selective detector responds to relatively few analytes. It is extremely useful in identifying all the compounds of interest in a complex mixture, but other components not having the characteristic property cannot be identified by selective detectors. UV-Vis, fluorescence, electrochemical, and radiochemical detectors are all extremely selective. RI and evaporative light scattering detector (ELSD) are universal detectors.

11.3 UV-Visible Absorbance Detector

11.2.3 Linearity

Detector linearity is an important factor in quantitative analysis. Nonlinearity contributes to bias in quantification. A detector using a line source with an interference filter is inherently more linear than a variable wavelength detector. The narrower the spectral band width of a detector, the better the linearity. Linearity can be improved by setting the detector at the absorbance maximum of a compound, where there is little change in absorbance with a change in wavelength. At high absorbance, linearity is usually better for peak area than for peak height. The linear range of the detector exceeds when the response factor deviates by more than 5% from the mean. This usually occurs at lower concentrations for peak heights compared to peak area measurements. 11.2.4 Dynamic Range

It is the range of sample concentrations over which the detector output changes. It is an important consideration while comparing different detectors. The detector response is generally not linear over the entire dynamic range and care is required while extrapolating a calibration curve to the extremes. Typically, the response flattens toward the upper end of the range indicating that the sensitivity of the detector is small in this region. Although the detector signal may be nonlinear in this region, quantification can still be achieved through careful calibration. 11.2.5 Detector Cell Volume

The detector cell volume should be small enough to prevent additional peak broadening. In practice, it is at least 10 times smaller than the volume of the first, most narrow, chromatographic peak. For nano HPLC systems, detection flow cells with a volume of 100 nl or less are appropriate. For conventional HPLC systems, the flow cells are around 5–12 μl. The detection cell should match with the sample and peak volumes. Making it too small will unnecessarily decrease the sensitivity of the detector.

11.3 UV-Visible Absorbance Detector

It is one of the most popular detectors of LC which measures the absorbance of the monochromatic light by the solute according to the well-known BeerLambert’s law. It measures the loss in intensity of ultrviolet or visible light as it passes through the solution exiting an HPLC column. It consists of three

247

248

11 Inline Detectors

Figure 11.1 Flow cell of a UV-Vis detector.

principal components: a source of monochromatic light, a flow cell, and a device to measure the change in the light intensity as it exits from the flow cell. The flow cell is usually a “Z” type with quartz windows covering the ends of a narrow tube of about 1 mm in diameter and 1 cm long (Figure 11.1). The source of monochromatic light allows one to select a single wavelength from a lamp source that emits the light over a wide range of wavelengths. A light measuring device, either a photodiode or a photomultiplier tube, is positioned at the back of the flow cell to measure the exiting light. The output from the photodiode/ multiplier is amplified and sent to a strip-chart recorder or computing integrator. The “peaks” on the chromatogram represent the response of the detector to the individual components as they elute from the column. Either the peak height or the peak area of a particular component can be used to measure its concentration by comparing with the peak obtained by injecting a standard of the same compound. UV-Vis absorbance detectors are simple, reliable, and relatively inexpensive. They are compatible with gradient elution and can be used for preparative liquid chromatography. They are relatively sensitive and selective but not as sensitive as fluorescence and electrochemical detection. Unlike RI, absorbance detectors are not universal. Molecules that do not contain a chromophore (i.e., a functional group that absorbs ultraviolet or visible light) cannot be detected, for example, aliphatic hydrocarbons. In such cases, indirect detection has been used to detect non-UV-absorbing analytes. A UV-absorbing component, such as terphthalic acid, is added to the mobile phase, which elevates the baseline. When non-UV-absorbing components elute, they cause a negative peaks because they absorb less than the mobile phase. Since most chromatographic integrators can be configured to measure negative peaks, quantification is not a problem. There are several methods for improving detector response and chemical labeling is one such technique. In other cases, an additive to the mobile phase changes the physical properties of the analyte of interest leading to a signal at

11.3 UV-Visible Absorbance Detector

Table 11.1 UV cutoff wavelengths of solvents used in HPLC. Solvent

UV cutoff wavelength (nm)

Acetic acid

230

Acetone

330

Acetonitrile

210

n-Butanol

215

Butyl acetate

254

Carbon tetrachloride

265

Chloroform

245

Cyclohexane

210

1,2-Dichloroethane

230

1,2-Dichloromethane

245

Dimethylformamide

310

Dimethyl sulfoxide

268

Dioxane

220

Ethanol

210

Ethylacetate

260

Diethyl ether

220

Heptane

200

Hexane

210

Methanol

210

Methyl-t-butyl ether

210

Methyl ethyl ether

329

Pentane

210

n-Propanol

210

Isopropanol

205

Diisopropylether

220

Tetrahydrofuran

280

Toluene

280

Trichloroethylene

273

Water

200

the detector [6–9]. The chromatographic system must be thoroughly flushed after it has been used in this manner. The UV cutoff wavelengths of different solvents used in HPLC are given in Table 11.1. 11.3.1 Fixed Wavelength Detector

Fixed wavelength detectors are simple, rugged, and less expensive compared to variable wavelength detectors. They consist of a light source, an interference filter, a flow cell, a photodiode detector, and an amplifier. If a low-pressure

249

250

11 Inline Detectors

mercury lamp is used as a line source, an appropriate interference filter is installed to select the desired line very often the 254 nm line of a mercury lamp is used, since at this wavelength many aromatic compounds show a strong absorption. Some of the fixed wavelength detectors use a continuous source such as a deuterium lamp. Low-pressure discharge lamps emit light at lines corresponding to the emission spectrum of the element used in the lamp. Mercury lamps emit at 254, 298, 313, 365, 405, 436, and 546 nm; cadmium lamps emit at 229 and 326 nm; and zinc lamps emit at 214 nm. Linearity is theoretically better compared to variable wavelength detectors because of the very narrow spectral bandwidth of the line sources. The detector “zero” (i.e., zero absorbance) is set by pumping clean mobile phase through the column and flow cell. Another design criterion is single beam versus double beam. In a single-beam instrument, light from the monochromator or lamp/filter passes through a single flow cell into the photosensitive device. Short-term changes in lamp intensity create noise, while long-term changes create drift. In a dual-beam configuration, the light from the light source is split into two beams, using either a beam splitter or a set of fiber optic bundles to send the light to two separate flow cells. The light exiting the flow cell is sent to two separate, matched photodiodes [10,11]. 11.3.2 Variable Wavelength Detector

Variable wavelength detectors have a tungsten and/or a deuterium lamp and a grating to select the wavelength to be changed during the run, so that each peak can be detected at its optimum wavelength (Figure 11.2). A rapidly rotating or vibrating grating is used in the front of optical cell allowing scanning of a range of wavelengths. The signal of the photodetector is correlated with time during measuring cycle, and thus to the wavelength of the transmitted light. Light

Figure 11.2 Optical arrangement for a rapid scanning variable wavelength detector.

11.4 Photodiode Array Detector (PDA)

intensities are usually measured by means of a solid-state photodiode. Optical cells are usually cylindrical with quartz windows at both ends to make detection possible in the 190–700 nm range. With most commercial instruments, cells are available of approximately 8 μl volume for conventional columns and 1 μl for microbore columns.

11.4 Photodiode Array Detector (PDA)

In the PDA, polychromatic radiation, after passing through the sample, is dispersed by a fixed grating and then falls onto an array of photodiodes. Here the full polychromatic light passes through the optical cell before it is dispersed on a grating. The dispersed light impinges on an array of 500–600 equidistant light sensitive diodes, each corresponding to a specific wavelength. Each diode measures a narrow band of wavelengths in the spectrum simultaneously. Figure 11.3 shows the optical arrangement of a PDA detector. It has a number of advantages [12]: a) Allows rapid acquisition, processing and storage of spectra. b) Requires less maintenance than a conventional spectrophotometer. c) Various signal averaging techniques can be used to reduce noise and improve sensitivity. d) A sample can be analyzed simultaneously at different wavelengths. e) UV-Visible spectra could be stored in the spectral libraries for compound identification. f) Compatible with gradient elution. g) Peak purity test can be performed. With PDA three dimensional (wavelength, intensity and retention time) chromatogram can be recorded to identify co eluting peaks (Figure 11.4).

Figure 11.3 Optical arrangement for a photodiode array detector.

251

252

11 Inline Detectors

Figure 11.4 3D-chromatogram by PDA.

11.5 Fluorescence Detector

Many compounds absorb UV radiation and subsequently emit radiation of longer wavelength either instantly (fluorescence) or after a time delay (phosphorescence). Compounds that fluoresce naturally have conjugated cyclic structures, for example, polynuclear aromatic hydrocarbons. Fluorescence detection offers increased sensitivity and selectivity compared to other modes. It is quite common to achieve femtomole detection limits with linearity over three to four orders of magnitude. The advantage of fluorescence detection is selectivity. The two major reasons for greater selectivity are (i) most organic molecules will absorb UV/visible light but not all will fluoresce and (ii) fluorescence utilizes two distinct wavelengths as opposed to one in absorbance. Due to the highly selective nature of fluorescence detection, quantitative analysis can be performed even without complete chromatographic separation. Poor column resolution is not as significant a problem as long as there is detection selectivity to resolve the peaks. Variable slits and wavelength programming, as well as other features available on modern fluorescence detectors, further enhance selectivity and therefore its usefulness. Chemical derivatization allows many nonfluorescent molecules to be detected, thus expanding the number of possible applications. Many nonfluorescent compounds can also be converted to fluorescent derivatives by treatment with suitable reagents. Fluorescence derivatization can be accomplished via either pre or postcolumn methods. In precolumn derivatization, the fluorescent derivatives are formed prior to injection onto the column. In postcolumn derivatization the reagent is mixed with the eluate from the

11.5 Fluorescence Detector

Figure 11.5 Postcolumn derivatization chamber.

column (Figure 11.5). The mixture is then directed through a reactor coil to form the fluorescent derivatives. Ortho-phthaldehyde (OPA) derivatization is selective for primary amines and amino acids, and peptides that have lysine groups [13–15]. The disadvantage of fluorescence detection is the perceived difficulty of its use. The fluorescence excitation and emission wavelengths and fluorescent intensity are affected by changes in the solvent, viscosity, and pH. Various solvent characteristics will shift the excitation and emission wavelength maxima. The presence of other solutes in the solvent can strongly decrease the fluorescence by quenching. Careful purification and degassing of the mobile phase is, therefore, necessary when fluorescence detection is used. Another aspect of concern is the presence of invisible quenchers in the sample. When such compounds coelute with the analytes of interest inaccurate and irreproducible results may be obtained. A block diagram of a fluorescence detector is shown in Figure 11.6. Radiation from a xenon or deuterium source is focused on the flow cell. An

Figure 11.6 Optical arrangement for a fluorescence detector.

253

254

11 Inline Detectors

interchangeable filter allows different excitation wavelengths to be used. The fluorescent radiation is emitted by the sample in all directions but is usually measured at 90° to the incident beam. The 90° optics allow monitoring of both UV absorption and fluorescent emission. In commercial instruments, high-pressure mercury lamps or xenon arcs are used as high-intensity excitation sources. The appropriate wavelength is selected by a band-pass filter or a grating. Higher intensities can be obtained with the former, while with the later selectivity is enhanced. The desired emission wavelength range is also selected with a filter or a grating. LC fluorescence detectors come in many designs from the manufacturers. Therefore, markedly different results are obtained during interlaboratory comparisons using different LC fluorescence detectors. There are three basic designs, filter–filter, grating–filter, and grating–grating, where either a filter or grating is used to select the correct excitation and emission wavelengths. Normally, each manufacturer will specialize in a particular design. Each design can usually accomplish the desired detection, but instrumentation optimization need to be performed for maximum sensitivity. The filter–filter detector is the most sensitive detector, yet the simplest and least expensive. It consists of three basic components: elemental light source, interference filter, and long-wavelength pass filter, or cutoff filter. This workhorse can provide high signal-to-noise levels for certain applications. Because both the excitation and the detection wavelengths can be varied, the detector is highly selective and very useful in trace analysis. For example, polyaromatic hydrocarbons (PAHs) are important air pollutants that have to be detected at very low concentrations. They are barely detectable using UV absorption but easily monitored by fluorescence. Fluorescence detector designs can be either single or dual beam. Single beams are popular since they allow increased light throughput and consequently have the potential for lower detection limits. The common light sources are continuous deuterium, xenon, xenon–mercury, and pulsed xenon. The deuterium lamp is an excellent choice when ultraviolet excitation and long-term stability are needed. This lamp has low noise and limited visible output. Many instruments today utilize xenon lamps. They provide excitation energy throughout the UV/ visible range. The flow cells have illuminated volumes between 5 and 20 μl. Sensitivity is directly proportional to the volume. Fluorescence is normally measured at an angle perpendicular to the incident light. Ninety degrees is the angle that has the lowest scatter of incident light. With the increased presence of microprocessors, fluorescence detectors have become programmable. Wavelength programmability can increase the sensitivity and selectivity. Fluorescence detection is typically linear over three or four orders of concentration. Nonlinearity can occur when the concentration of the sample does not fall within the linear dynamic range of the technique. Sample linearity can fall at both high and low concentrations. Quantitation that is irreproducible may need an internal standard. An ideal internal standard is a compound that has fluorescence properties similar to the analyte.

11.6 Refractive Index Detector (RID)

Figure 11.7 Optical arrangements for (a) diffraction and (b) Fresnel-type RI detectors.

11.6 Refractive Index Detector (RID)

The RI detector is the most universal detector for LC. It was one of the first LC detectors developed. Compared to other detector types, however, its detection limits are poor. Therefore, generally it is used only when other techniques fail. In commercially available instruments, different principles are applied, most often the diffraction type and the reflection or Fresnel-type RI detector. With both types a light source in the visible range is used and photocells are used for the measurement of light intensities. The optical arrangements of (a) diffraction and (b) reflection-type instruments are shown in Figure 11.7. The light passes through two prism-shaped flow cells, one filled with pure mobile-phase solution and one flushed by the column effluent. When the two solutions differ in refractive index, the light beam is deflected at the wall between the two optical cells through an angle depending on the RI difference. Detection can be performed in differential mode by measuring the difference in the light intensity on two photocells. Since refractive indices are strongly dependent on the temperature proper thermosetting is essential to stabilize the temperature of the sample and reference cells. Today, refractive index detectors probably rank third in popularity behind the UV/visible light absorption and fluorescence detectors, but they are by far the most popular type of universal detectors. The major advantage of a refractive index detector is the universal nature of its response. The major disadvantage of a refractive index detector is its lack of sensitivity compared to selective solute property detectors. The second disadvantage of refractive index detectors is the possibility of both positive and negative peaks in the chromatogram. Although the refractive indices of compounds of interest in HPLC are positive, some

255

256

11 Inline Detectors

solutes might have refractive indices that are less positive than the solvent. It is usually possible to find a solvent that gives all positive peaks, but a change in solvent might detract from the chromatographic separation. The bulk property nature of its response makes the refractive index detector impractical for use with gradient elution. The universal response of the refractive index detector allows the economy and convenience of a single detector when sensitivity is not the dominant consideration. It is a popular choice when suitable selective detectors are not available, and it is often used in series with selective detectors to detect unsuspected components that might miss detection by the selective detectors. Refractive index detectors are used as the primary detectors for compounds that do not have strong UV chromophores, fluorophores, electrochemical activity, or ionic conductivity. Traditional areas of application are the detection of carbohydrates and lipids [16–20]. The refractive index detector finds special application in the analysis of polymers by gel permeation or size exclusion chromatography.

11.7 Evaporative Light-Scattering Detector

The ability to nebulize an HPLC column effluent has led to increased utility of light-scattering detectors. The most popular is the evaporative light-scattering detection (ELSD) [21]. It works on the principle of evaporation (nebulization) of the mobile phase followed by the measurement of the light scattered by the resulting particles (Figure 11.8). The column effluent is nebulized in a stream of nitrogen (carrier gas) in a heated drift tube, and nonvolatile particles are left suspended in the gas stream. Light scattered by the particles is detected by a

Figure 11.8 A schematic diagram of an evaporative light-scattering detector.

11.8 Electrochemical Detector

photocell mounted at an angle to the incident light beam. The carrier gas flow rate and drift-tube temperature must be adjusted for whatever mobile phase that is used. Pneumatic nebulizers are generally used for aerosol generation. The column eluent is mixed with a gas stream in a concentrinic tubular nebulizer forming a high velocity jet of liquid droplets. Mobile phase that contains nonvoltile components cause an elevated back ground and decrease sample detectability, as well as a rapid degradation of detector performence, due to their deposition in optical cell. ELSD can be used with gradients as it is not sensitive to temperature or flow rate fluctuations. Mobile phases must be volatile, similar to those used in mass detection. ELSD is compatible with the most volatile solvents used in normal and reversed phase separations. Unlike the UV detector, no chromophores are required and it has orders of magnitude more response than the RID. Its primary uses include the detection of compounds with a weak response to the UV detector, especially carbohydrates, lipids, surfactants, polymers, and petroleum products. Its greater sensitivity and ease of use in gradient elution separations makes it preferable to the refractive index detector.

11.8 Electrochemical Detector

Today, the electrochemical detector (ECD) is widely accepted as a sensitive and selective technique for the analysis of electroactive substances. It responds to substances that are either oxidizable or reducible and the electrical output is an electron flow generated by a reaction that takes place at the surface of the electrodes. The electrochemical detector consists of three electrodes: the working electrode (where the oxidation or reduction takes place), the auxiliary electrode, and the reference electrode (which compensates for any changes in the background conductivity of the mobile phase) (Figure 11.9). Solid electrodes (carbon/glassy carbon, noble metals/gold, platinum/nickel, and copper) are used for oxidations, while mercury is used for reductions. ECD has an enormous linear

Figure 11.9 Schematic representation of an electrochemical detector.

257

258

11 Inline Detectors

dynamic range of more than six orders of magnitude which means that concentrations can be measured as low as 50 pmole/l and as high as 100 μmole/l or more. Electrochemical detection is used mainly in reversed phase or ionexchange separations, where polar mobile phases are used. An inert electrolyte is sometimes added to enhance conductivity. The hydrocarbons are not easily oxidizable or reducible and so would not be suitable for ECD. Phenols and aromatic amines are easily oxidized and are suitable for EC detection. By choosing the appropriate applied voltage for the oxidation/reduction potential and material of the working electrode, a more chemical specific detecting condition can be obtained. HPLC-ECD has the perfect balance of price and performance for routine measurements of biological or environmental samples.

11.9 Charged Aerosol Detection

Recently carona charged aerosol detection (CAD) has started gaining in popularity. Here, the HPLC column eluent is first nebulized with a nitrogen-carrier gas to form droplets that are then dried to remove mobile phase, producing analyte particles. The primary stream of analyte particles is met by a secondary stream of positively charged nitrogen passed through a high voltage platinum carona wire. The charge transfers diffusionally to the opposing stream of analyte particles, further transferred to a collector and is measured by a highly sensitive electrometer. A schematic diagram of CAD is shown in Figure 11.10. CAD is highly sensitive, provides a consistent response, and has a broad dynamic range, which offers advantages when analyzing compounds lacking UV chromophores. Often compared with other universal-type HPLC detection methods, such as RI and ELSD, CAD is much easier to use, and unlike RI, can accommodate gradients.

Figure 11.10 Schematic diagram of a charged aerosol detector system.

11.11 Coupling Detectors

In addition, CAD response does not depend upon the chemical characteristics of the compounds of interest, but on the initial mass concentration of analyte in the droplets formed upon nebulization, providing a much more uniform response. CAD requires only the setting of a few controllable parameters, such as the gas input pressure, the temperature range, and the signal output range. CAD generates a nearly constant response under isocratic conditions for compounds at similar concentrations. This detector also allows the use of a wider variety of mobile phases and buffers. CAD allows detection of all nonvolatile and most semivolatile analytes. CAD is a flexible detection technique and has been used in combination with a variety of different separation modes (reversed phase, ion chromatography, hydrophilic interaction liquid chromatography, supercritical fluid chromatography, and size exclusion chromatography) using normal- and narrow-bore chromatographic columns, for a wide range of different analytes.

11.10 Conductivity Detector

Conductivity detectors are predominately used in ion chromatography. The electrical conductivity of the column effluents can be monitored by measuring the current by applying AC voltage between two electrodes in a flow cell (Figure 11.11). As aqueous buffers are generally used as mobile phases in ionic chromatography, the conductivity changes due to elution of sample ions could be monitored at a high background level. In anion chromatography, a carbonate buffer is used. The column effluent is passed over a small ion-exchange column where buffer cations (Na+, K+) are replaced by hydrogen ions. Owing to the acid– base reaction the background conductivity is suppressed, when anions elute. The conductivity is increased when the buffer cations are replaced by the H+ ions. H‡ ‡ HCO3 ! CO2 ‡ H2 O One major attraction of conductivity detectors is that there is no loss in sensitivity with miniaturization [22]. The prospect of ready miniaturization has led to considerable research interest in using conductivity detectors for capillary and chip-based electrophoresis. Commercial detectors are generally able to detect less than 1 ppb F with 1 s or less integration times.

Figure 11.11 Conductivity detector.

259

260

11 Inline Detectors

11.11 Coupling Detectors

While the detectors discussed above are generally used in routine laboratories, there are other detectors that are used less frequently and often for only one type of application. These are polarimetric, radiochemical and chemiluminence, and so on. The successful coupling of HPLC with mass spectrometry (MS) is one of the most exciting developments in the past two decades. This combination provides un ambiguous identification of various components in complex mixtures. The primary difficulties in combining LC and MS have been the interface, the ionization of analytes in a stream of condensed liquids, and transfer of ions into the high vacuum inside the mass spectrometer. Two common LC–MS interfaces are the electrospray ionization (ESI) and the atmospheric pressure chemical ionization (APCI). During the spraying process, solvent molecules are removed with the aid of heat and drying gases, while the charged analytes are guided into the mass spectrometer. The common types of MS include the quadrupole, ion trap, triple quadrupole, time of flight (TOF), and fourier transform MS (FTMS). Compared to a single-quadrupole LC/MS, the use of LC/MS/MS using a triple-quadrupole analyzer can further increase sensitivity and specificity for quantitation of trace analytes in complex matrices. LC/MS/MS is now the dominant methodology and offers a generic approach for trace analysis in complex mixtures. The hyphenation of LC/NMR is an area of rapid growth due to significant instrument enhancements in recent years. Combining LC and NMR extends the technique for online mixture analysis. The primary problem has been the low sensitivity of NMR, which has been now extended to microgram and even nanograms levels by probe miniaturization, noise reduction, and other innovations in interface technologies. The other hyphenated systems of LC include LC–FTIR and LC–ICP–MS.

11.12 Comparison of HPLC Detectors

A comparison of routinely used HPLC detectors is given in Table 11.2. Most of the LC performed with one of the four types of detectors described in Table 11.2. Generally, the performance of a detector depends on the compounds to be determined, the sample matrix and the separation system used. The UV absorbance detector is used most frequently because of its ease of operation, robustness, reasonable selectivity and its detection limits. For lower detection limits the fluorescence detector is often chosen. The selectivity of fluorescence detection may be an advantage. But on the other hand, it often necessitates derivatization before the separation which may complicate the analytical procedure and introduce experimental error. Compared to UV detection, it requires high purity solvents and the cleanup of samples to obtain adequate reproducibility. For easily oxidizable analytes, electrochemical detection offers the lowest

References

Table 11.2 Comparison of common detectors used in LC. Characteristic

UV-Vis

Fluorescence

ECD

RID

Sensitivity

High

Very high

Very high

Low

Selectivity

Moderate

Very high

High

Universal

Robustness

Excellent

Good

Poor

Good

Suitability for microsystem

Good

Good

Excellent

Not suitable

Destructive/ nondestructive

Nondestructive

Nondestructive

Destructive

Nondestructive

Detectability Min. conc.

1 × 10

1 × 10 g/ml

1 × 10

1 × 10

Peak shape

Positive

9

g/ml

12

Positive

12

Positive

g/ml

7

g/ml

Positive or negative

detection limits. One of the major applications of ECD is the determination of catecholamine in body fluids. The operation and maintenance of an electrochemical detector is much more problematic compared to optical detection. Electrode contamination necessitates frequent cleaning and recalibration. This is much less the case in pulsed amperomatric detection. However, the detection limits with PAD are much higher than with constant potential detection. The RI detector is used when the concentrations to be determined are relatively high and the selectivity requirements are low. Because of its universal nature of detection, the unexpected components in a sample are not readily missed. Both UV and fluorescence detection are routinely combined with gradient elution. ECD and RID are not campatable with gradient elution because of shifting base lines. The performance of the detectors in miniaturized separation systems is quite different. Microcells are not obtainable for RI detection. With UV detection, signal-to-noise ratios decrease when the flow cell size is decreased. The same is true for fluorescence detection. However, this effect can be reversed by the use of a laser as excitation source. With an electrochemical detector, even higher signal-to-noise ratios can be obtained in a miniaturized system.

References 1 Snyder, Lloyd R., and Kirkland, Joseph J.,

3 Simpson, C.F. (ed.) (1984) Techniques

and Dolan, John W. (2010) Introduction to Modern Liquid chromatography, 3rd edn, John Wiley & Sons, Inc., Hoboken, NJ. 2 Hamilton, R.J. and Sewell, P.A. (1982) Introduction to High Performance Liquid Chromatography, Chapman and Hall, London.

in Liquid Chromatography, Wiley-Blackwell. 4 Knox, J.H. (ed.) (1982) High Performance Liquid Chromatography, Edinburgh University Press, Edinburgh, UK. 5 Meyer, Veronika R. (2010) Practical High Performance Liquid Chromatography, 5th edn, John Wiley & Sons.

261

262

11 Inline Detectors 6 Lindsay, S. (1992) High Performance

7

8

9

10

11

12

13

14

Liquid Chromatography, John Wiley & Sons, Inc. Yost, R.W., Ettre, L.S., and Conlon, R.D. (1980) Practical Liquid Chromatography: An Introduction, Perkin Elmer, Norwalk, USA. Corradini, D. and Phillips, Terry M. (2011) Handbook of HPLC, CRC Press Taylor & Francis Group, London. Parriott, Donald (1993) A Practical Guide to HPLC Detection, Academic Press, Inc., New York. Scott, Raymond P.W. (1986) Liquid Chromatography Detectors, Elsevier, Amsterdam, The Netherlands. Poole, Colin F. (2003) The Essence of Chromatography, Elsevier, Amsterdam, The Netherlands. Huber, L. and George, S.A. (1993) Diode Array Detection in HPLC, Marcel Dekker, New York, NY. Poole, C.F. and Schuette, S.A. (1984) Contemporary Practice of Chromatography, Elsevier, Amsterdam, The Netherlands. Katz, E., Eksteen, R., Schoenmakers, P., and Miller, N. (eds) (1998) Hand Book of HPLC, Marcel Dekker, New York.

15 Wilson, I.D., Adlard, E.R., Cooke, M., and

16

17

18

19

20

21

22

Poole, C.F. (eds) (2000) Encyclopedia of Separation Sciences, Academic Press, London. Weston, A. and Brown, P.R. (1997) HPLC and CE: Principles and Practice, Academic Press, San Diego, CA. Snyder, L.R., Kirkland, J.J., and Glajch, J.L. (1997) Practical HPLC Method Development, John Wiley & Sons, Inc., New York. Yeung, E.S. (ed.) (1986) Detectors for Liquid Chromatography, John Wiley & Sons, Inc., New York, NY. Patonay, G. (ed.) (1992) HPLC Detection: Newer Methods, John Wiley & Sons, Inc., New York, NY. Scott, R.P.W. (1996) Chromatographic Detectors, Design, Function and Operation, Marcel Dekker, New York, NY. Christie, William W. (1992) Detectors for High-Performance Liquid Chromatography of Lipids with Special Reference to Evaporative LightScattering Detection, Oily Press, Ayr, U.K. Ishil, D. (1988) Introduction to Microscale High-Performance Liquid Chromatography, Wiley-VCH Verlag GmbH.

263

12 pH Effects on Chromatographic Retention Modes Paweł Wiczling, Łukasz Kubik, and Roman Kaliszan

12.1 Introduction

The pH of a mobile phase is one of the most important parameters affecting retention of ionic analytes in reversed phase high-performance liquid chromatography (RP-HPLC). The ionic analyte contains one or more acidic or basic functional groups in its molecular structure. Acidic compounds lose a proton as pH increases, whereas bases gain a proton as pH decreases: …Acid†HAÛA ‡ H‡

(12.1)

…Bases† B ‡ H‡ ÛBH‡

(12.2)

The dissociation constant, pKa, defines pH at which molecule is half ionized and it also reflects the pH region at which the concentration of the charged and uncharged species varies mostly due to small changes in pH. The often used Handerson–Hasselbalch equation relates to pH, pKa, and the concentration of all analyte forms observed in a solution:
 ‰A Š ; (12.3) …Acids†pK a ˆ pH log ‰HAŠ
…Bases†pK a ˆ pH

log

 ‰BŠ ; ‰BH‡ Š

(12.4)

where [HA], [A ], [B], and [BH+] are concentrations of particular analyte forms. Equations 12.3 and 12.4 clearly show that the higher is the difference between pH and pKa, the more likely is complete dissociation or complete suppression of dissociation of the analyte in a solution. Since the charged and uncharged species are different chemical entities, any change in pH leads to variations in physicochemical properties of an analyte, such as retention, UV/Vis absorbance, electrophoretic mobility, and many others. Specifically, for chromatographic retention, the presence of dissociated group(s) decreases analytes’ hydrophobicity (analytes are more polar or more hydrophilic) that leads to the Analytical Separation Science, First Edition. Edited by Jared L. Anderson, Alain Berthod, Verónica Pino Estévez, and Apryll M. Stalcup.  2015 Wiley-VCH Verlag GmbH & Co. KGaA. Published 2015 by Wiley-VCH Verlag GmbH & Co. KGaA.

264

12 pH Effects on Chromatographic Retention Modes

lower retention of the dissociated form than that of the nondissociated form of an analyte.

12.2 pH Measurements of Mobile Phase

In RP-HPLC, the mixture of water and an organic modifier is usually used as a mobile phase. It complicates the pH measurements, since the proper calibration of any pH meter must be performed upon the addition of an organic solvent. According to the IUPAC, three pH scales can be distinguished for water–organic modifier solutions. They depend on the way the pH measurement and pH meter calibration is carried out and can be summarized as follows: ww pH scale – the pH measurement is performed before the addition of an organic modifier and standard water buffers are used for instrument calibration; sw pH scale – the pH measurement is performed after the addition of an organic modifier and standard water buffers are used for instrument calibration; and the most accurate a ss pH scale – the pH measurement is carried out after the addition of an organic modifier and standard buffers contain the same amount of the organic modifier as a test solution. It is advised to use a sw pH scale that can be easily converted to ss pH as both scales differ by the δ term, which is constant for each solvent (mobilephase organic modifier and its percentage) and has been tabulated for methanol/ water and acetonitrile/water mobile phases [1,2]. The pKa values obtained in any of these pH scales show a direct relationship with the thermodynamic dissociation constants of the compound in the same scales. Thus, the use of ss pH leads to s s s s pK a , and w pH scale leads to w pK a . A detailed description of the procedures for measuring pH can be found in the work of Rosés [3]. Some authors prefer pH measurements in pure water prior to the addition of organic modifier solvent [4]. In many circumstances, it may be sufficient, especially from the practical point of view of routine analysis. However, to properly understand and, what is particularly important, to accurately quantify the retention behavior of ionic analytes, the more exact ss pH scale is mandatory. The most commonly used buffers in chromatography are characterized in Table 12.1. The universal buffers, which ensure approximately constant buffer capacity for a wide range of pHs, are very convenient in controlling pH during chromatographic separation. The equimolar mixture of citric acid/tris/glycine is an example of such a buffer. For mass spectrometry (MS) detection, only volatile buffers can be used and there is no universal buffer available yet. Despite the buffering system used, preliminary pH measurements are required for all the possible mobile-phase compositions, as the pH can substantially change upon the addition of an organic modifier. Figure 12.1 shows the pH of mobile phase for various methanol buffer compositions. The addition of organic modifier dilutes the solution and changes the pKa value of buffer components that consequently leads to a shift in the pH value. For example, the addition of

12.2 pH Measurements of Mobile Phase

Table 12.1 The characteristics of commonly used buffers in chromatography. Buffer

pKa

Buffer range

Trifluoroacetic acid

>2

1.5–2.5

Yes

Phosphoric acid

1.5–3.5 6.0–8.5 11.0–13.5 2.0–7.5

No

Formic acid

2.1 7.2 12.3 3.1 4.7 5.4 3.8

2.5–5.0

Yes

Acetic acid

4.8

3.5–6.0

Yes

Triethanolamine

7.8

6.8-8.8

Yes

Carbonic acid

5.0–7.5 9.0–11.5 7.5–10.5

Yes

Tris(hydroxymethyl)aminomethane

6.4 10.3 6.8 9 8

7.0–9.5

No

Ammonium

9.2

8.0–10.5

Yes

Diethanolamine

8.9

7.9–9.9

Yes

Diethylamine

10.5

9.5–11.5

Yes

Triethylamine

11

9.5–12.5

Yes

Citric acid

Bis-tris propane

MS compatibility

No

No

Source: Reproduced from Ref. [4].

Figure 12.1 The relationship between the methanol content and the pH for three MS-compatible buffers spanning a wide range of pH values. Ammonium formate (squares), ammonium acetate (circles), and ammonium bicarbonate (triangles) are shown in this graph.

265

266

12 pH Effects on Chromatographic Retention Modes

50% of organic modifier might change the pH value of the mobile phase by about 0.5 unit for methanol and even more for acetonitrile [5]. When studying the pH effect on analyte retention, a special emphasis must also be put on column’s properties. The column should have low silanol activity ensuring partitioning as the only operative mechanism of analyte retention. That condition is typically fulfilled for modern chromatographic columns that are allowed to operate at a wide range of pH.

12.3 Effect of pH on Isocratic Retention

Retention factor, k, of a nondissociated form of an acid or a base may be 10–20 times larger than that of the respective dissociated form at any composition of the water–organic mobile phase. It offers a convenient means to rationally modify the separation of ionizable compounds by changing the eluent pH. Retention of ionic analytes in RP-HPLC as a function of the properties is based on the solvophobic theory of Horváth, Melander, and Molnar [6,7]. This theory assumes that the behavior of acidic, basic, and amphoteric compounds depends mainly on two characteristics: the hydrophobicity and the ability to ionize in the mobile phase. The electrically neutral form of an analyte has a relatively high affinity for the stationary phase and hence a greater retention compared to the dissociated form of the same analyte, which in turn is more attracted to the mobile phase. Figure 12.2 shows the relationships between the pH and the experimental retention factor for a weak acid (ketoprofen) and for a weak base (papaverine) at different concentrations of methanol in the mobile phase. Several features can be distinguished in these plots. The relationship (log k versus pH) is sigmoidal with inflection point of each curve corresponding to the pKa value of an analyte. For low pH values, the ketoprofen is present in a nondissociated form (the highest retention), whereas for high pH values, it is present in an ionized form (with the lowest retention). The reverse holds true for papaverine. Thus, for low pH values, papaverine is fully dissociated (the lowest retention), whereas for high pH values, it is completely nondissociated (the highest retention). The retention obviously depends on the methanol content, which is reflected as the shift of the sigmoids toward higher retentions upon decrease in methanol content. The magnitude of that shift is approximately constant as log k is generally linearly related to the organic modifier content in the mobile phase that is usually well accounted for by the Snyder–Soczewiński relationship: log k ˆ log k w

Sφ:

(12.5)

In Equation 12.5, kw denotes the retention in a neat water eluent and S denotes the slope of the relationship between the logarithm of retention factor

12.3 Effect of pH on Isocratic Retention

Figure 12.2 The dependence of retention factor on mobile-phase pH for a weak acid (ketoprofen) and a weak base (papaverine) at different concentrations of methanol in the mobile phase (25% (v/v) 65% (v/v)). The

following conditions were applied: XTerra MS C18 5 μm, 4.6 mm × 150 mm column (Waters, USA), F = 1 ml/min, T = 25 °C, citric/tris/glycine universal buffer of 0.008 M each component.

and the organic modifier content (% v/v). The inflection points, and consequently pKas, are not constant for profiles in Figure 12.2. For acids, an increase and for bases a decrease is generally observed. Ketoprofen and papaverine have pKa of 3.98 and 6.48, respectively, in pure water and much different values upon methanol addition (close to 6.5 and 5.7, respectively, for 80% methanol). The following relationships are usually used to describe pH effects on isocratic retention for weak acids and bases: s

s

k HA ‡ k A 10 s pH s pK a …Acids† k ˆ ; s s 1 ‡ 10s pH s pK a s

…Bases† k ˆ

(12.6)

s

k BH‡ ‡ k B 10 s pH s pK a ; s s 1 ‡ 10 s pH s pK a

(12.7)

where kA, kHA, kB, and kBH denote the retention factors of the dissociated and nondissociated forms of an analyte. The influence of pH on analyte retention is much more complicated for polyprotic or amphoteric analytes (multiple acidic/ basic groups). For compounds with several dissociating groups, an overlay of few sigmoidal shapes is typically observed.

267

268

12 pH Effects on Chromatographic Retention Modes

12.4 pH Effect on Organic Modifier Gradients

The organic modifier gradient comprises a separation with a constant increase in eluting power of the mobile phase due to the increase in organic modifier content. Similar to the isocratic mode, the gradient retention of analytes strongly depends on the mobile-phase pH. Figure 12.3 illustrates 10 experimental chromatograms obtained at organic modifier gradients conducted at different but constant mobile-phase pH values for a mixture of acids and bases. Clearly, all the peaks considerably change their position in chromatograms with pH. To better understand the behavior of an analyte during linear organic modifier gradient, the following equations for acids and bases were proposed [8]:
 t0 k 0;HA ‡ k A 10pH pK a chrom ‡ 1 ; (12.8) t R ˆ t 0 ‡ t d ‡ log 2:303b b 1 ‡ 10pH pK a chrom
 t0 k 0;BH ‡ k 0;B 10pH pK a chrom t R ˆ t 0 ‡ t d ‡ log 2:303b ‡ 1 ; (12.9) b 1 ‡ 10pH pK a chrom where tR is the analyte retention time, t0 is the holdup time, td denotes the instrument dwell time, b is the organic modifier gradient steepness (b ˆ Sβt 0 ,

Figure 12.3 The experimental chromatograms of 10 analytes for a series of organic modifier gradient developed at 10 different pHs of the mobile phase. The dotted lines present the retention times of peaks corresponding to the following analytes: (1-barbital, 2-acetylsalicylic acid, 3-ciprofloxacine, 4-piroxicam, 5-indapamid, 6-metoprolol, 7-tramadol, 8-cetirizine, 9-verapamil, and

10-hydroxizine). The following conditions were used: XTerra MS C18 5 μm, 4.6 mm × 150 mm column (Waters, USA), F = 1 ml/min, T = 25 °C, tg = 32 min, φ = 5–80%, citric/tris/glycine universal buffer at 0.008 M of each component. As can be noted, for pH values above 10, a maximal resolution (baseline separation) was achieved.

12.5 pH Gradient

where β is the steepness of the gradient ⠈ …φf φ0 †=t G , φ0 and φf are the initial and final methanol contents, respectively, in the mobile phase, and tG is the gradient duration, and S is a slope of linear log kw versus φ relationship), k0,X is the retention factor for a species X corresponding to initial organic modifier content φ0 , and pKa chrom is the parameter related to the analyte dissociation and reflecting all the resultant ss pK a values occurring during the chromatographic run as a consequence of changes in the organic modifier content. In other words, pKa,chrom can be understood as a thermodynamic ss pK a value that corresponds to the specific organic modifier content φchrom. The value of φchrom can be calculated as an average organic modifier content experienced by a given analyte at pH equal to pKa,chrom: t RZt 0 t e

φchrom ˆ

φ…t† 0

dt ; t 0 k i …t†

(12.10)

where ki is the instantaneous retention factor corresponding to the mobile-phase composition φ. From Equations 12.8 and 12.9, it follows that the relationship between the retention time and the pH of the mobile phase is sigmoidal, with gradient retention time being highest for nondissociated form and the lowest for dissociated form of analytes. However, the inflection point does not correspond to pKa of an analyte. It is higher for bases and lower for acids, as is illustrated in Figure 12.4 for diphenhydramine. To properly account for the effect of pH on analyte retention, the pKa,chrom value must be known. Since it corresponds to the specific organic modifier content, it is different from the usually reported aqueous pKa values. Usually, the RP-HPLC methods are optimized by searching for the most proper organic modifier content in the mobile phase. However, variations in the pH of the mobile phase introduce additional variations in the retention of ionizable analytes. Thus, the simultaneous optimization of both the pH and the organic modifier content provides benefits due to the additional possibility of maintaining reasonable retention times and acceptable selectivity criteria. Since the peak width can be properly predicted for any pH and organic modifier gradient [9], it allows the researcher to find the best solution for a complex and difficult analytical problem involving ionic analytes. In Figure 12.5, an optimized separation of 12 weak acids and bases is presented. It was obtained under double pH/organic modifier gradient conditions. The multistep changes in both pH and methanol allowed to separate this mixture within 20 min with a high resolution [10].

12.5 pH Gradient

It is technically feasible to conduct a pH gradient separation that consists of a pH change at fixed organic modifier content during the chromatographic run.

269

270

12 pH Effects on Chromatographic Retention Modes

Figure 12.4 The diphenhydramine retention times for two series of organic modifier gradients development at different pHs. The following conditions were used: XTerra MS C18 5 μm, 4.6 mm × 150 mm (Waters, USA); F = 1 ml/min; T = 25 °C, tG = 20 min and 60 min, methanol/ buffer cit/tris/gly, φ0 = 5%, φf = 80%. Multistep

pump program was used to ensure constant pH despite increase in methanol content. The vertical line indicates pKa,chrom value of 9. The gray circles show inflection points. The black lines indicate model predictions, as given by Equations 13.8 and 13.9.

The increasing (in case of bases) or decreasing (in case of acids) pH of the eluent provides functional increase in the degree of analyte dissociation and hence, a decrease in its retention during the chromatographic separation. It is similar, in principle, to the well-established conventional gradient HPLC, where the eluting power of the mobile phase is increased with time due to the increasing content of organic modifier. The example of pH gradient program is given in Figure 12.6a. The increase in mobile-phase pH leads to the decrease in the analyte instantaneous (corresponding to particular mobile-phase composition at column inlet) retention factor. The chromatogram corresponding to the situation is given in Figure 12.6b. For a linear pH gradient, the approximate but explicit solution of retention equation allowing the prediction of retention has recently been provided [11,12]

8 for t R < t 1 ‡ t 0 t 0 ‡ t 0 k nion >
  0 > 1 > t1 k nion k ion > > > 1 t 0 ai > 3k nion B C < t 0 k nion 3 t0 ‡ t1 ‡ @1 e A for t 1‡ t 0  t R< t 2 ‡ t 0 ; tR ˆ ai…k nion k ion † > > >
 > > k ion 3k ion k nion > > : t 0 ‡ t 2 ‡ t 0 k ion t 1 ln ‡ for t 2 ‡ t 0  t R k nion ai…k nion k ion † k ion (12.11)

12.5 pH Gradient

Figure 12.5 Chromatograms obtained under optimized conditions for a double pH/ methanol gradient for a mixture of 12 analytes: (1) aniline, (2) 2-amino-5-nitropyridine, (3) N-methylaniline, (4) N-ethylaniline, (5) 2,4,6-collidine, (6) brucine,

(7) p-nitrophenol, (8) diethylbarbituric acid, (9) 2-chloro-4-nitrophenol, (10) 2,6-dimethyl4-nitrophenol, (11) 1-naphtylacetic acid, and (12) N,N-benzyldimethylaniline. (Adapted from Ref. [10].)

where tG is the pH gradient duration, a is the steepness of the pH gradient a ˆ …pHf pH0 †=t G , td denotes the instruments dwell time (time that gradient needs to attain the column inlet), ts is the pH gradient starting time (time at which the pH gradient starts in the pump program), pH0 and pHf are the initial and final pH of the mobile phase, respectively, i = 1 for acids and i = 1 for bases, knion and kion are retention factors corresponding to the nondissociated

271

272

12 pH Effects on Chromatographic Retention Modes

Figure 12.6 Plots illustrating the principles and applications of the pH gradient separation mode. (a) The pH changes during pH gradient and corresponding changes in instantaneous retention factor are presented for a monoprotic acid characterized by knion = 20, kion = 5, and pKa = 7. (b) The chromatogram corresponding to conditions A, along with chromatogram obtained for isocratic conditions at

pH ensuring full dissociation/nondissociation of an analyte. (c) and (d) The experimental chromatograms obtained for optimized pH gradient conditions (solid line) and isocratic conditions for nonionized forms of analytes (broken lines) are given for the examples of ketoprofen and papaverine. (Adapted from Ref. [14].)

12.5 pH Gradient

and dissociated forms of an analyte (both depend on organic modifier content), and t1 and t2 denote time window with the pH-dependent changes in analyte retention: t1 ˆ td ‡ ts ‡

pK a

i
1:5 a

pH0

and t 2 ˆ t d ‡ t s ‡

pK a ‡ i
1:5 a

pH0

:

(12.12) The pH gradient retention time of an analyte will always be within the retention time of its dissociated and nondissociated forms (Figure 12.6b). Hence, simply by varying the pH gradient steepness or pH range, one might modify the analyte retention times. The most interesting feature of pH gradient separation is related to the peak compression phenomenon. The movements of molecules located in the front part (closer to the column outlet) and back part (closer to column inlet) of the peak differ during the gradient separation. When molecules in the back part of the peak move faster than those in the front part, the peak compression is observed. It is a very often exploited phenomenon in organic modifier gradient mode as it leads to the approximately constant peak width as opposed to the isocratic separation, for which a gradual increase of peak width with time is observed (Figure 12.2). For pH gradient mode, the peak width may assume values much varying with pKa of the analyte and with the retention of both the dissociated and the nondissociated forms. It might range from 1 (no compression) to almost fivefold compression and it highly depends on the pH gradient program. Maximum peak compression is expected if the organic modifier content in the eluent is adjusted such that retention of the dissociated form is low and the retention of the nondissociated form is relatively high, whereas the range of pH changes and the duration of gradient provide a relatively lengthy exposure to the analyte to pH, being within ±1.5 unit around its pKa. At such pH range, the effects of changing pH on analyte retention are the strongest. The following relationship has been derived [12] to ensure the maximal peak compression:

 k nion pK a i1:5 pH0 3 k ion ln ‡ : aopt ˆ k nion t 0 t d t s k nion k noin i…k ion k nion † (12.13) Since with a steeper pH gradient, the peak compression is higher, one can anticipate the best conditions to conduct the pH gradient would be for a steep pH gradient developed just before the analyte elution from the column. Figure 12.6c and d illustrate the benefits of pH gradient for the acidic (ketoprofen) and basic (papaverine) compounds, along with the isocratic separation of the nondissociated forms of the analyte. It is clear that the application of pH gradient has led to the improvement in peak shape, a decreased peak width, and an increased peak height. The exact values of peak width are more difficult to describe theoretically than for the organic modifier gradient [13], but some solutions have been recently proposed [9].

273

274

12 pH Effects on Chromatographic Retention Modes

12.6 Determination of pKa, log kw (Hydrophobicity), and S

In order to make practical use of the pH-related effects in chromatography, the parameters describing analyte retention need to be known. It involves the retention factor k, dissociation constant, pKa, and their dependence on the organic modifier content. All these parameters can be directly estimated from the experimental chromatographic data differing in both pH and organic modifier content. The numerous mathematical models and designs have already been proposed for that purpose. They mostly involve isocratic and pH gradient modes [15–25]. However, the most easy and efficient way involves conducting a series of wide-range organic modifier gradients at varying gradient duration and the pH of the mobile phase [8,26]. The example of such a design is given in Figure 12.4. From this type of data, the pKa,chrom, log kw, and S can be estimated by using Equation 12.7 or 12.8 to optimize any pH effect on chromatographic retention, like the requested pH gradient retention time and peak width. RP-HPLC seems to be particularly useful for the determination of both dissociation constant and the (pH-dependent) partition coefficient (log P) related parameters – highly requested parameters in pharmaceutical research [27]. In general, a high correlation between the literature (potentiometric) pKa values and the chromatographic measure of acidity can be expected, as presented in Figure 12.7. The use of a correction factor, which explains the decrease of pKa for acids and increase for bases due to methanol content, is necessary and substantially increases the accuracy of determinations. Also, a high correlation between the retention factor and the lipophilicity can be observed (Figure 12.7). The overall accuracy of log P prediction is about 0.5 unit. This degree of accuracy is generally expected for this type of comparisons. It can be further improved by also considering in calculation the number of hydrogen donor/ acceptor groups in a molecule. In literature, the chromatographic hydrophobicity index (CHI) is occasionally used as a scale closely related to hydrophobicity [28,29]. The determination of CHI is performed by a series of fast (5 min) RP-HPLC gradients conducted at different pH. The retention times are then converted to CHI through calibration procedure with reference to hydrophobicity data. It allows to obtain the CHI versus pH profiles and thus to estimate the hydrophobicity of all the possible species [30,31]. Several LC-based methods of pKa determination were tested in literature [32,33]. Also, a medium-to-high-throughput HPLC pKa assay was proposed and used in the laboratory settings [34]. There are also several methods for a direct simultaneous determination of pKa and chromatographic measures of lipophilicity/hydrophobicity from chromatographic data [8,25,35,36]. The gradient RP-HPLC coupled with time-of-flight mass spectrometry with electrospray ionization source (ESI–TOF-MS) methods was recently shown to be especially applicable for complex mixtures. The use of ESI–TOF-MS detection allowed to achieve the medium-throughput screening rate (100 compounds/day) [37]. All those studies used the pH dependence of analyte retention times and confirmed

12.7 Effect of pH in Normal-Phase Mode

Figure 12.7 The comparison of literature-based lipophilicity parameters log P and pKa values with the chromatographically derived log k values for φ = 0.4 and the pKa,chrom values, respectively.

the applicability of RP-HPLC in medicinal chemistry to characterize physicochemical properties of compounds.

12.7 Effect of pH in Normal-Phase Mode

The influence of the mobile-phase type and properties on chromatographic retention has been extensively studied [38–41]. So far, we have discussed the pH

275

276

12 pH Effects on Chromatographic Retention Modes

effect in the reversed phase mode, that is, when the mobile phase is more polar than the stationary phase. The opposite situation, when the eluent is more hydrophobic than the stationary phase, is known as the normal-phase mode (NP-LC). The pH effect on retention in the NP-LC is less significant as a consequence of several limitations of this mode (e.g., the need of using nonpolar solvents, such as hexane and chloroform, the difficulties in dissolving hydrophilic substances in nonaqueous mobile phases, and general insignificance of pH in nonpolar solvents) [41,42]. One of the interesting and widely used modifications of NP-LC is hydrophilic interaction chromatography (HILIC) [42]. The mobile phase in HILIC has relatively a high water content (5–40%) and is composed of reversed phase-type eluents. HILIC has been designed for polar and hydrophilic analyte separations such as proteins, peptides, and nucleic acids. HILIC is commonly used as a complementary method to RP-HPLC. Apart from the solubility aspect, the other advantages are easier sample processing, the possibility of volatile buffer application which enables avoiding a desalting step, and no need of analyte derivatization [43]. The HILIC analysis is performed in a wide range of mobile-phase pH, mostly under acidic and neutral conditions, due to the instability of silica-based columns at basic pH (which could be avoided by using chromatographic columns with organic polymer backbone instead of silica) [44]. Selectivity in HILIC is determined by polar forces, mainly hydrogen bonds (which depend on the pKa value of analyte, i.e., its acidity/basicity) and dipole– dipole interactions (dependent on polarization, pH of mobile phase, and dipole moment). Hydrophobic and ion exchange interactions are negligible for most of the HILIC chromatographic columns [45]. For this reason, it should be taken into consideration that HILIC, though formally classified as a variant of normalphase chromatography, is governed mostly by polar forces. In general, the retention mechanism in HILIC is opposite to the trends observed in reversed phase mode. The retention time increases with increasing polarity (hydrophilicity) of analytes and stationary phase and with decreasing polarity of the eluent (decreasing the content of water). This retention mechanism is analogous to nonaqueous NP-LC [42,45,46]. During the chromatographic run, a thermodynamic equilibrium is established between eluent and water-enriched solvent layer, adsorbed onto the surface of the column packing material. The analyte is being partitioned between those two phases [47,48]. The HILIC retention time strongly depends on the degree of analyte dissociation and consequently on the pH of the mobile phase. Dissociation of a compound increases its polarization, thus resulting in an increased retention during the chromatographic run. At low pH, acidic compounds possess a neutral charge and basic compounds become protonated with positive charge. At high pH, basic compounds remain nonionized and acidic compounds lose protons becoming anions. For this reason, acids are better separated using high-neutral pH and bases are better separated using low-neutral pH. For the complex mixtures of acids and bases, neutral pH is recommended to obtain the best separation results. Under those conditions, little variations in pH do not significantly

References

affect the retention and separation [4,43,49]. The pH of the mobile phase affects not only the analytes but also the stationary phase. For example, at pH > 4 silica stationary phases are negatively charged [50], thus increasing retention for positively charged compounds and decreasing retention for negatively charged anions. In HILIC, the eluent pH strongly affects the resolution and selectivity of a separation, thus making it an important parameter in the method development process.

12.8 Summary

RP-HPLC and HILIC can be used as a convenient tool for the analysis of ionizable analytes. Especially, by considering the pH changes during the chromatographic separations, one is able to optimize chromatographic conditions to achieve separations that are difficult to obtain by other means. A number of theoretical models and approaches have been proposed to predict retention times and peak widths for specific changes of pH and organic modifier. These models can serve as a basis for the estimation of the parameters, which reflect analyte properties and can be used to optimize chromatographic separations, that is, to find conditions resulting in maximal resolution, selectivity, or peak compression.

References 1 Canals, I., Portal, J., Bosch, E., and

2 3 4

5 6 7 8

9

Roses, M. (2000) Anal. Chem., 72, 1802. Espinosa, S., Bosch, E., and Roses, M. (2000) Anal. Chem., 72, 5193. Rosés, M. (2004) J. Chromatogr. A, 1037, 283. Snyder, L.R. and Dolan, J.W. (2006) HighPerformance Gradient Elution: The Practical Application of the LinearSolvent-Strength Model, John Wiley & Sons, Inc., New York. Subirats, X., Roses, M., and Bosch, E. (2007) Sep. Purif. Rev., 36, 231. Horvath, C., Melander, W., and Molnar, I. (1976) J. Chromatogr., 125, 129. Horvath, C., Melander, W., and Molnar, I. (1977) Anal. Chem., 49, 142. Wiczling, P., Kawczak, P., Nasal, A., and Kaliszan, R. (2006) Anal. Chem., 78, 239. Wiczling, P. and Kaliszan, R. (2010) J. Chromatogr. A, 1217, 3375.

10 Wiczling, P., Markuszewski, M.J., Kaliszan,

11 12 13 14 15

16

17

18

M., and Kaliszan, R. (2005) Anal. Chem., 77, 449. Nikitas, P., Pappa-Louisi, A., and Zisi, C. (2012) Anal. Chem., 84, 6611. Wiczling, P. and Kaliszan, R. (2010) Anal. Chem., 82, 3692. Neue, U., Marchand, D., and Snyder, L. (2006) J. Chromatogr. A, 1111, 32. Kaliszan, R. and Wiczling, P. (2011) Trend. Anal. Chem., 30, 1372. Kaliszan, R., Haber, P., Baczek, T., and Siluk, D. (2001) Pure. Appl. Chem., 73, 1465. Kaliszan, R., Haber, P., Baczek, T., Siluk, D., and Valko, K. (2002) J. Chromatogr. A, 965, 117. Kaliszan, R., Wiczling, P., and Markuszewski, M.J. (2004) J. Chromatogr. A, 1060, 165. Kaliszan, R., Wiczling, P., and Markuszewski, M.J. (2004) Anal. Chem., 76, 749.

277

278

12 pH Effects on Chromatographic Retention Modes 19 Berges, R., Sanz-Nebot, V., and Barbosa, J. 20 21 22 23

24 25 26

27 28 29

30 31 32 33 34

35

(2000) J. Chromatogr. A, 869, 27. Hardcastle, J.E. and Jano, I. (1998) J. Chromatogr. B, 717, 39. Nikitas, P. and Pappa-Louisi, A. (2002) J. Chromatogr. A, 971, 47. Nikitas, P. and Pappa-Louisi, A. (2009) J. Chromatogr. A, 1216, 1737. Pous-Torres, S., Torres-Lapasio, J.R., Baeza-Baeza, J.J., and Garcia-AlvarezCoque, M.C. (2008) J. Chromatogr. A, 1193, 117. Nikitas, P., Pappa-Louisi, A., and Zisi, C. (2012) Anal. Chem., 84, 6611. Zisi, C., Fasoula, S., Nikitas, P., and PappaLouisi, A. (2013) Analyst, 138, 3771. Wiczling, P., Waszczuk-Jankowska, M., Markuszewski, M.J., and Kaliszan, R. (2008) J. Chromatogr. A, 1214, 109. Nasal, A., Siluk, D., and Kaliszan, R. (2003) Curr. Med. Chem., 10, 381. Valko, K. (1997) Anal. Chem., 69, 2022. Du, C.M., Valko, K., Bevan, C., Reynolds, D., and Abraham, M.H. (1998) Anal. Chem., 70, 4228. Fuguet, E., Rafols, C., Bosch, E., and Roses, M. (2007) J. Chromatogr. A, 1173, 110. Valko, K., Bevan, C., and Reynolds, D. (1997) Anal. Chem., 69, 2022. Manderscheid, M. and Eichinger, T. (2003) J. Chromatogr. Sci., 41, 323. Huo, H., Li, T., and Zhang, L. (2013) Springerplus, 2, 270. Chiang, P.C., Su, C.C., Block, K.M., and Pretzer, D.K. (2005) J. Liq. Chrom. Relat. Tech., 28, 3035. Wiczling, P., Nasal, A., Kubik, Ł., and Kaliszan., R. (2012) Eur. J. Pharm. Sci., 30, 1.

36 Chiang, P.C., and Hu, Y. (2009) Comb.

37

38 39 40

41 42 43

44 45 46 47

48 49 50

Chem. High Throughput Screen., 12, 250. Wiczling, P., Struck-Lewicka, W., Kubik, L., Siluk, D., Markuszewski, M.J., and Kaliszan, R. (2014) J. Pharm. Biomed. Anal., 94, 180. Jandera, P. and Churacek, J. (1974) J. Chromatogr., 91, 207. Snyder, L.R. and Quarry, M.A. (1987) J. Liq. Chromatogr., 10, 1789. Schoenmakers, P.J., Billiet, H.A.H., and Galan, L.D. (1979) J. Chromatogr., 185, 179. Espinosa, S., Bosch, E., and Roses, M. (2002) J. Chromatogr. A, 964, 55. Alpert, A.J. (1990) J. Chromatogr., 499, 177. Mora, L., Aristoy, M.-C., and Toldra, F. (2012) Food Anal. Methods, 5, 604. Kirkland, J.J. (1996) J. Chromatogr. Sci., 34, 309. Yoshida, T. (2004) J. Biochem. Biophys. Methods, 60, 265. Churms, S.C. (1996) J. Chromatogr. A, 720, 75. Mant, C.T. and Hodges, R.S. (1991) HPLC of Peptides and Proteins. Separation, Analysis and Conformation, CRC Press, Amsterdam. Alpert, A.J. (1988) J. Chromatogr., 444, 269. Guo, Y. and Gaiki, S. (2005) J. Chromatogr. A, 1074, 71. Takegawa, Y., Deguchi, K., Ito, I., Keira, T., Nakagawa, H., and Nishimura, S.I. (2006) J. Sep. Sci., 29, 533.

279

13 Chemometrics in Data Analysis and Liquid Chromatographic Method Development Biljana Jančic ́-Stojanovic ́ and Tijana Rakic ́

13.1 Introduction

Chromatography is clearly the most useful technique in many areas of sample analysis. Although it has existed in its modern form for many decades, and many books and thousands of papers have been published on the topic, it remains a challenging technique due to the fact that many parameters affect its performance. The steady expansion of its application and continuous improvement in terms of more sensitive detectors present new challenges. Experts always try to obtain faster and more sensitive methods, and better chromatogram characteristics. Competition is always present and leads to the invention of new approaches, which may be faster and better than the previous techniques. Although chemometric techniques have been known for many years, many analytical laboratories do not use them but are still able to successfully resolve their chromatographic problems. Therefore, the question remains: “Why use a chemometric approach if the problem can be resolved without it?” The main advantage of using a chemometric strategy in liquid chromatography enables a more detailed understanding of the investigated system with less effort. In this chapter, two important fields of chemometrics used to resolve chromatographic problems will be presented. The first part of the chapter presents the techniques for managing the large number of data obtained from chromatographic analysis. The origin of the data could be from many chromatographic runs (many samples, many columns, etc.), very complex samples (usually require long runs), or the application detectors which give multivariate signals and thus a lot of data (such as mass spectroscopy or MS). In such cases, it is important to process the data and then organize them to find a pattern in their behavior. To extract the important data in these situations, the analysts should preferably possess required skills and expertise. Even then, very experienced analysts can overlook or lose significant data. To solve these problems, preprocessing is recommended.. After preprocessing, the extracted data can be analyzed using different chemometric tools applied to find patterns of behavior. The most commonly used techniques will be presented here. Analytical Separation Science, First Edition. Edited by Jared L. Anderson, Alain Berthod, Verónica Pino Estévez, and Apryll M. Stalcup.  2015 Wiley-VCH Verlag GmbH & Co. KGaA. Published 2015 by Wiley-VCH Verlag GmbH & Co. KGaA.

280

13 Chemometrics in Data Analysis and Liquid Chromatographic Method Development

The second part of the chapter will present the chemometric applications in liquid chromatography (LC) method development. A step-by-step building of quality into the method will be described, emphasizing all the important aspects of the highly reliable method development. The advantages of chemometric approaches over the traditionally applied trial and error approaches will be observed through the full control gained by the analyst over all important quality characteristics. Thus, the desired goals can be obtained with minimal experimental effort, and the risk of method failure in the validation phase is significantly decreased.

13.2 Chemometrics in Data Analysis

The ideal chromatographic signal (for any separation method or detector) should have well-resolved peaks, an adequate signal-to-noise ratio, no background contribution, and a large linear response range between the analyte concentration and the detector signal for individual samples/runs. In case of complex mixtures, the ideal situation also includes stable retention times and well-defined peak shapes (preferably Gaussian profiles) for all analytes. In general, the success of the chromatographic separation, as well as the robustness and stability, depends on the selection of appropriate chromatographic equipment and experimental conditions (e.g., column type, temperature conditions, gradient of the mobile phase, etc.) [1]. In practice, it is difficult to identify ideal chromatographic conditions, and analysts can encounter various problems during chromatographic analysis. There are a number of sources of variability in chromatographic systems (the pumping system, temperature gradients, stability of stationary phase, etc.). This can be reflected on the chromatographic signals hindering the achievement of perfect separation, detection and quantification of the individual analytes, and obtaining of sample-specific fingerprints [1]. The resulting problems can be solved in two ways. The first method involves changing of the chromatographic conditions with the goal of obtaining a desirable chromatogram. This approach is timeconsuming and often leads to other nonideal chromatographic conditions. The second approach involves applying chemometrics, which produces successful alternatives and provides data preprocessing, thereby removing as many variations as possible from the data set. After data preprocessing, different chemometric techniques could be employed for data analysis. 13.2.1 Data Preprocessing

The most significant problems that appear in chromatographic analysis are baseline and noise, retention time shift, incomplete separation, and data overloading [2]. All these problems can be solved during preprocessing using chemometrics and will be presented briefly in this section.

13.2 Chemometrics in Data Analysis

The first listed problem is baseline and noise. Generally, the overall chromatographic signal is the sum of the analytical signal, baseline and noise. The analytical signal originates from the analyte and depends on many factors. The baseline or, more generally, the background signal is not related to the analyte and presents a type of systematic behavior that depends on the chromatographic conditions. Finally, noise is a nonsystematic variation that depends on the detector sensitivity. Consequently, it is quite obvious that the baseline is the unique part of the overall signal that could be modified. Several baseline correction methods have been proposed in the literature, but two of the most common are curve fitting and the factor model approach. The first approach uses algorithms that fit polynomial functions across segments of the chromatogram, using regions where no analyte peaks elute to determine the coefficients of the polynomial equations and then interpolating the background signal for regions where peaks are eluting. The resulting function is usually a first-order polynomial, but higher order polynomials can also be found in some cases. The fitted line can then be subtracted after finding the overall signal or the area below the fitted line subtracted after finding the overall peak area. It is important to note that almost all chromatographic software packages have some type of baseline handling method [1]. The latter approach (the factor model) is based on the fact that the baseline is a systematic part of the chromatographic signal. That means, baseline should have a similar shape for all samples; only its magnitude should vary. Specifically, the factor model is used to estimate the baseline part of the signal, and the best application mode is to process separately local regions for two reasons: the number of peaks in the intervals should be fairly small and the baseline can change not only in magnitude over the elution time but also in shape during different runs [1]. The second problem in chromatogram analysis is the retention time shift. One of the most important goals in chromatographic analysis is that all peaks should maintain their retention times in different runs under the same chromatographic conditions. In LC, small changes in the composition of the mobile phase, the temperature, or other chromatographic parameters can occur during chromatographic analysis. The influence of variation of these factors can significantly influence the retention time, so the retention time drift complicates peak tracking between chromatographic runs and this problem becomes more severe for a complex sample with many peaks. Consequently, the identification of peaks is not reliable for making comparisons from one run to the next. This problem can be minimized by proper instrument maintenance, by precise control of instrumental conditions, or by the use of some other approaches [2]. This problem can be solved during preprocessing in two ways: integrated peak tables and alignment of the raw signals. Integrated peak tables are the simplest way to ensure that the analytical separation data are properly aligned for chemometric processing, and this approach is well explained in reference [2]. Another method (alignment of the raw signal) is a more complex method than the previous one. The alignment of the chromatographic data can be defined as “a mathematical operation where similar chemical features are repositioned so that they appear at the same

281

282

13 Chemometrics in Data Analysis and Liquid Chromatographic Method Development

elution time in different runs.” The optimal alignment technique should require only minimal or no input from a skilled technician or scientist, should be fast, robust, and applicable for a wide range of analytical situations without extensive customization needed [1]. The selection of the alignment method depends on whether the method will be used for qualitative or quantitative analysis, because some alignment methods can warp peaks directly influencing the quantification. The most commonly used method of alignment is correlation optimized warping (COW). COW is a piecewise or segmented technique that aligns a sample chromatogram with reference chromatograms by stretching or compressing the sample segments. Alternatively, coshift is a fast and simple alignment algorithm. This algorithm is particularly useful when the data require only a single left–right shift in retention time. On the other hand, it is not recommendable when the peaks have moved in different directions. In such situation, it is better to apply an icoshift algorithm (interval correlation shifting), which was derived from the coshift algorithm. Icoshift aligns each data matrix to a target by maximizing the cross-correlation between the sample and the target within a series of user-defined intervals [2]. Shifting using coshift or icoshift does not lead to distortions of the peak shape and, consequently, does not introduce errors into quantitative results. Another approach is the application of a piecewise peak-matching algorithm. In this method, peaks are identified relative to a target signal to which all other signals will be aligned. The chromatograms are aligned by stretching or compressing the regions between the peak apexes. A variant of this algorithm is very useful when MS data are to be analyzed. In this case, the mass spectrum of the apex of each peak in the target signal is compared with the mass spectrum of each peak within a set window on the sample signal. The peaks are matched if their spectra have a satisfactorily high match quality [2]. Generally, the choice of the proper method depends on the homogeneity of the samples, on the degree of missing peaks across the chromatograms, and other factors that should be considered for each individual application [3]. The third listed problem is incomplete separation, which appears when two or more compounds in a mixture have similar chromatographic behavior under the given chromatographic conditions. These situations can be fixed by changing the chromatographic conditions, which is usually time-consuming and, consequently, leads to an increase in analysis cost. In addition, this approach requires an experienced experimentalist. Multivariate detectors can solve this problem and enable quantification of overlapping peaks. Overlapped peaks can be quantified a posteriori, when chromatographic analysis is finished. In this situation, the overlapping peaks are resolved into the contributions from different chemical components by decomposing the original signal to a sum of different profiles by means of modeling [1]. Traditionally, this is performed by fitting Gaussian or Lorentzian curves or a combination of these. Furthermore, for solving this problem, multivariate curve resolution (MCR) could be applied. MCR is an algorithm that can solve the problems associated with the presence of interferents even when there is a high degree of overlapping with the profiles of the analytes [4]. In this approach, the concentration and

13.2 Chemometrics in Data Analysis

response profiles of each analyte are obtained, providing a qualitative and semiquantitative overview of the components in an unresolved mixture without a priori knowledge of the mixture composition [2]. Alternatively, evolving factor analysis (EFA) could also be used. EFA is a noniterative method based on timedependent rank analysis, that is, each row is associated with an increase in the rank by one. However, this method is based on the “first in first out” assumption, that is, the compound that started to elute first will disappear first. EFA includes evolving principal component analysis (PCA) in two directions along the retention point, forward and backward. Eigenvalue from forward PCA shows the retention points where chemical components begin to appear, while backward PCA indicates the retention points where chemical components begin to disappear; by combining information from these data, the elution sequence of each of the components by this model can be identified [5]. A detailed description of this approach was presented in a paper published in 1992 [6]. EFA is traditionally used only with multivariate data, while MCR could be used for both, multivariate and univariate data. Finally, ideally suited technique for interpreting multivariate separation data is parallel factor analysis (PARAFAC). In PARAFAC model, the same factors are used to describe the variation in several matrices simultaneously but with different weights for each matrix. The interpretation in terms of calibration is simple: each analyte has the same chromatographic and spectral profile in all the calibration samples and differs from one another only in the amount in which it intervenes [7]. The PARAFAC model for multivariate data provides three matrices, A, B, and C, which contain the scores and loadings for each component. The residuals, E, and the number of factors, r, are also extracted. PARAFAC finds the best trilinear model minimizing the square sums of the residuals in the model through a procedure of alternating least squares [2]. Generally, ordinary PARACAF requires multivariate detection and several runs with high retention time stability (or proper alignment) to provide a meaningful trilinear decomposition. The PARAFAC model, its comparison with other techniques, and its applications are described in detail in reference [8]. Finally, data overloading is identified as the fourth potential problem. High data acquisition rates combined with a long time required for many separations result in a large number of data points collected for a given separation. Therefore, it is very important to reduce the data used in model systems. One common method to achieve this is to use a table of integrated peaks instead of raw chromatographic data. In this table, the baseline noise is removed and, with the knowledge of irrelevant compounds, the analyst can remove them from the table [2]. Of course, there are some disadvantages, but this is definitely the most straightforward method. When an analyst employs multivariate detection for practical problems, the situation is more complicated. In this case, the analyst must consider plenty of data, and one of the ways to use the important data is to select only the significant peaks. In order to achieve this, the analyst must possess that kind of knowledge and in accordance with that methodically follow only the selected peaks. This approach can solve the problem; however, many well-known advantages of

283

284

13 Chemometrics in Data Analysis and Liquid Chromatographic Method Development

multivariate detectors cannot be used, while some important data for certain analysis cannot be collected. The relevant data from many experiments can be extracted using objective feature selection techniques, which generally have two steps: variable ranking and variable selection. By using this approach, important data are separated from the noise, and the subjectivity of the first method is augmented by the objectivity of the second method. The analysis of variance, the discriminating variable test, and informative vectors can be used as objective features. When the set of data is reduced and the variables have been ranked, they should be included in the appropriate model. This model can be created by using different approaches, and the resulting model should be assessed using necessary tools. 13.2.2 Data Analysis

When preprocessing is finished, different chemometric techniques can be applied for further analysis of the data. Given that finding a behavior pattern is important in chromatography (and in other areas of chemistry), pattern recognition techniques are the most frequently employed tools. According to Brereton [9], pattern recognition methods can be divided into exploratory data analysis (EDA) and unsupervised and supervised pattern recognition. EDA consists mainly of PCA and factor analysis (FA). The other method, unsupervised pattern recognition, involves cluster analysis and the last pattern recognition technique mostly provides classification. The EDA approach relates to the process of revealing hidden and unknown information from data in such form that the analyst obtains an immediate, direct, and easy-to-understand representation of it. Visual graphs are a key element of this approach due to the intrinsic ability of the human brain to obtain a more direct and trustworthy interpretation of similarities, differences, trends, clusters, and correlations through a picture than a series of numbers [10]. Generally, from the data obtained using EDA, it is possible to identify outliers, trends, and patterns in the data and, based on the resulting conclusions, set new theories and hypotheses. By far the most used method is PCA, whose application in chromatography has been confirmed through many published papers. The wide application of PCA relies on its ability to condense a large amount of data into a small number of parameters, called principal components (PCs), which capture the levels, differences, and similarities among the samples and the variables comprising the modeled data. This task is achieved by using linear transformations with the constraint of preserving the data variance and imposing orthogonality of the latent variables [10]. The second tool for EDA is FA, which attempts to express the features using a small number of common factors [11]. By evaluating unsupervised pattern recognition, which mainly encompasses cluster analysis, we arrive at the methods that enable grouping of samples based on mutual similarities or dissimilarities. Cluster analysis is used to classify

13.3 Chemometrics in LC Method Development

objects, characterized by the values of a set of variables, into groups. It is, therefore, an alternative to the principal component analysis for describing the structure of a data table [12]. It can be a very useful tool in chromatography, especially when many stationary phases or substances should be organized into groups. Eventually, supervised pattern recognition is generally used for classification. In contrast to unsupervised pattern recognition, classes of samples are well known and are used for model calibration. Furthermore, the resulting model is used for the analysis of an unknown sample.

13.3 Chemometrics in LC Method Development

Liquid chromatographic method development can be led through many different pathways, but no matter what approach is applied, the primary aim is to obtain reliable and high-quality results. However, using a chemometric strategy provides many benefits compared to the strictly empirical trial and error approach. Chemometric application enables a thorough description of the analyzed system, where all important parameters and characteristics are mathematically related. If the accuracy of these theoretical relations is statistically reliable, the theoretical model of the system behavior can be used for detailed investigation. The chemometric approach provides full control over every segment of the process. In this way, the desired characteristics of the method could be achieved easily using only a theoretical examination of the experimental space. This result will also be observed even when the defined characteristics are not achieved. Furthermore, the weak points of the method and possible critical outcomes will be identified and fully understood. Thus, precautionary measures can be taken to avoid undesired events. The empirical approach consists of method development based on scientific experience and the resulting optimal conditions can be verified through method validation and routine method application. However, this strategy prevents the assessment of all possible causes of method failure. Therefore, there is a possibility that the method will fail to demonstrate the appropriate quality characteristics and that the entire process should be repeated, requiring substantial money and time. More importantly, the empirical approach does not allow for detailed understanding of all analytical method characteristics. Therefore, analytical experts do not clearly understand how the method should be improved to obtain satisfactory results. The essence of the empirical approach is that the quality information is revealed only at the end of development and optimization phase. Considering the shortcomings of the trial and error approach, certain fields of analytical chemistry (such as pharmaceutical analysis) already strongly suggest the implementation of chemometric techniques in method development because this approach enables the direct implementation of quality into the method, that is, a quality by design (QbD) approach [13]. This strategy provides a detailed understanding of the method and its potential risks and uncertainties, which can affect the method. Instead of defining the optimal analysis conditions, the

285

286

13 Chemometrics in Data Analysis and Liquid Chromatographic Method Development

Table 13.1 Chemometric and QbD approach in LC method development. Chemometric approach in LC method development

QbD approach in LC method development

Definition of the objectives of the method

Analytical target profile and critical quality attributes

Definition of investigated factors and their levels

Quality risk assessment and critical process parameters

Selection of appropriate experimental design

Investigation of knowledge space

Creation of mathematical models

Critical quality attribute modeling

Identification of optimal conditions

Design space Selection of working points

Robustness testing

chemometric approach even allows for the definition of several alternative conditions over the entire experimental region for which the method performance would be satisfactory. This space, called the design space, shows the part of the experimental space where the combination of investigated factors and their interactions and process parameters will be set to guarantee a certain quality for the method. The implementation of the quality in such a way significantly decreases the risk of failure for developed methods. The QbD approach for LC method development [14,15] differs only slightly from the chemometric strategy [16], mostly in terminology. The main steps of the two approaches are presented in Table 13.1. Furthermore, the key steps of the QbD approach in the LC method development proposed strategy will be explained in detail. 13.3.1 Analytical Target Profile and Critical Quality Attributes (Definition of the Objectives of the Method)

An analytical target is a set of criteria that define what will be measured and the required performance criteria of the method, which are called critical quality attributes (CQAs) [14,15]. Therefore, the targets for analytical methods can vary, but generally all methods should provide maximal selectivity in minimal analysis time within robust experimental space. CQA yields crucial system responses (outputs) or dependent variables. CQA can be divided into two general classes: those that must be achieved and those that should be achieved. Different elementary and global separation criteria can be used as CQA. Elementary separation criteria include selectivity factor, resolution factor, and the peak-to-valley ratio calculations for adjacent peak pairs. Recently, an S criterion was described that indicates the difference between the retention time of the beginning of the second eluting peak and the end of the first eluting peak [17]. The global separation criteria enable estimation of the

13.3 Chemometrics in LC Method Development

separation quality of all investigated peak pairs. Until now, a variety of chromatographic response functions have been designed for this purpose, and apart from the separation, they can assess other secondary CQA, such as analysis duration, uniformity of peak distribution, robustness, and so on [18,19]. CQA can be directly modeled using an appropriate mathematical model. However, other chromatographic responses, which are occasionally easier to model (such as the retention time or retention factor), can be modeled directly, and complex CQAs can be calculated, that is, modeled indirectly [20]. For example, indirect modeling of the resolution factor is highly recommended because it can show nonlinear or noncontinuous behavior, which prevents reliable prediction. Compared to the resolution factor, the S criterion can be more useful for direct modeling; its computation is simpler and leads to lower prediction uncertainty. 13.3.2 Quality Risk Assessment and Critical Process Parameters (Definition of Investigated Factors and Their Levels)

Critical process parameters (CPPs) are parameters whose variations have an impact on CQA; therefore, they must be monitored and controlled to ensure the appropriate quality of the product. CPPs are denoted as inputs or independent variables. Table 13.2 shows some of the most important CPPs, which should be examined in reversed phased LC method development. Table 13.2 Factors that can be considered in reversed phased LC method development [21]. Factors pH of the mobile phase Amount of the organic modifier Buffer concentration, salt concentrations, or ionic strength Concentration of additives (ion pairing agents, competing amine) Flow rate Column temperature For gradient elution: initial mobile phase composition final mobile phase composition slope of the gradient Column factors: batch of stationary phase manufacturer age of the column Detector factors: UV or fluorimetric detection (wavelength) electrochemical detection (voltage) MS (scan time, geometry of ion source, sheath and auxiliary gas pressure, capillary temperature, collision pressure) Integration factors: sensitivity

287

288

13 Chemometrics in Data Analysis and Liquid Chromatographic Method Development

Among the listed factors, the most critical are the type of stationary phase, type and amount of organic modifier in the mobile phase, pH value of the water phase, temperature, and duration of gradient elution. Some LC methods can be affected by the ionic strength of the water phase and/or the amounts of water additives [22]. The investigated factors can be classified as quantitative, qualitative, or mixed factors. The main characteristic of the quantitative factors are that they can be varied on a continuous scale. This group includes the pH value of the mobile phase, column temperature, buffer concentration, and so on. In contrast, the qualitative factors, such as the reagent batch or the manufacturer of the chromatographic column, vary in a discrete way. The second type is a mixture of factors with a mixture of p components; only p 1 can be changed independently. In LC analysis, the mobile phase is often multicomponent, and the sum of the solvents is always 100%, meaning that one component always depends on other components. The simplest solution to this problem is to select at most p 1 components to be examined as factors in the experimental design. These factors are called mixture-related variables [23]. Alternatively, mixture design – response surface design specially adjusted to this problem – can be used. 13.3.3 Investigation of the Knowledge Space (Selection of an Appropriate Experimental Design)

The goal of the analytical method is to find the appropriate factor combination (i.e., the CPP combination) that will provide the desired CQA. Naturally, at the very beginning, it is impossible to define the exact combination of CPPs, but some preliminary experiments should be performed, and the resulting data combined with the analyst’s experience can aid in defining the factor intervals. The experimental space defined within the factor interval region is called the knowledge space. The factor intervals should be selected in such a way that the expected optimal solution should fall within the knowledge space. In addition, they should be broad enough so that their changes influence the change in the CQA and, thus, provide valuable information about the system behavior. However, extremely broad factor intervals can lead to complex behavior in the monitored responses; therefore, standard mathematical models might fail to describe the system behavior. The appropriate selection of factor intervals has a significant influence on future results; therefore, this phase of the research should be performed with great caution. The next phase involves the creation of an experimental plan. The design of experiments (DoE) methodology consists of performing a relatively small number of well-planned experiments and the associated detailed mathematical and statistical data analysis for maximal extraction of the relevant information [12,16,24,25]. This process consists of two main phases: screening of the potentially important CPPs and their optimization. The screening designs

13.3 Chemometrics in LC Method Development

Figure 13.1 Points representing the experimental plan for (a) 22 full factorial design and (b) 23 full factorial design.

should enable the reduction of the number of relevant CPPs to those that are the most critical and whose values should be tentatively adjusted in the optimization phase. In the majority of cases, the experiments in this phase are designed so that each factor is examined on two levels: a lower (denoted as 1 level) and an upper (denoted as +1) level. The real lower and upper factor values are coded to 1 and +1 to eliminate any differences originating from the various scales used for the investigated factors. 13.3.3.1

Screening Designs

The most frequently applied experimental designs in this phase of LC method development are full factorial design (FFD), fractional factorial design (FrFD), and Plakett–Burman design (PBD). Full factorial design consists of a matrix in which all possible combinations of factors and levels are investigated and, therefore, all important effects and all interactions can be evaluated. The number of experiments for FFD is defined by nk, where n represents the number of levels and k the number of factors. During the screening phase, the factors are generally investigated on two levels. Therefore, for k investigated factors, 2k experiments should be performed. Graphical representation of the required experiments for two and three investigated factors is presented in Figure 13.1. One of the most important properties of FFD is the possibility of the estimation of factors’ interactions, which is generally one of the most important benefits of design of experiments methodology compared to the more traditionally applied one-factor-at-the-time approach. Specifically, the problem of investigating two-factor influences using FFD is resolved by the experimental plan presented in Figure 13.1. However, if it is approached in one-factor-at-the-time fashion, one factor would be held constant, while the other is switched from the lower to the upper level and vice versa. The latter approach does not contain information on the simultaneous change of investigated factors and, therefore, the potential significant interaction between them will remain undetected.

289

290

13 Chemometrics in Data Analysis and Liquid Chromatographic Method Development

Figure 13.2 Points representing the experimental plan for 23

1

fractional factorial design.

However, in LC method development, factor interactions can exist, and the simultaneous effects of the two factors can be different from the isolated, individual factor effects. The fractional factorial design is more suitable than the full factorial design when the number of investigated factors is greater than 4. That is, the number of required experiments for more than four factors using full factorial design is unreasonably high. Thus, the experimental matrix can be systematically reduced into the fractional factorial design. In FrFD, two-level fractional factorial designs are used to reduce the number of experiments by 1/2, 1/4, 1/8, and so on. Therefore, the overall number of experiments can be expressed as nk p, where n represents the number of levels, k the number of factors, and p the size of fraction (p < k), that is, 2k p for two-level FrFD. The schematic representation of the experiments is given in Figure 13.2. The main limitation of FrFD is that the main effects (bN) are confounded by the interaction terms (b(N 1)N). The existence of confounding within FrFD makes individual estimation of aliasing pairs impossible. The Plakett–Burman design presents two-level factorial design for studying N 1 factors in N experiments, where N is a multiple of 4. It has a complex aliasing pattern because all the main factors and the two-factor interactions are partially confounded. Therefore, it is generally applicable if the existence of twofactor interactions can be neglected. Since the number of experiments in PBD is always a multiple of 4, the number of factors of interest can be smaller than the overall required number of factors in the design matrix. In that situation, imaginary factors, called dummy factors, should be added. The dummy factor may be a variable that has no effect on the experiment, and changing from 1 to +1 has no physical meaning. The addition of dummy factors can be valuable for statistical estimation of the effects of factors. The idea of screening designs is to reveal whether the response changes drastically when the factor is shifted from the lower to the upper value [21]. This shift

13.3 Chemometrics in LC Method Development

is denoted as the effect of the factor on the response and can be calculated according to the following equation: P EX ˆ

P Y …‡† Y… † N=2

(13.1)

where E X is the effect of the factor X, ΣY …‡† and ΣY … † are the sums of the responses where the factor X is at (+) or ( ) level, respectively, and N is the number of design experiments. Naturally, different factors will have different effects on different experimental cases. The sign of a factor effect represents the direction of the response change when the factor value is increased. Thus, if the sign of the effect is positive, the increasing factor value will lead to an increase in the response value. In contrast, the negative sign suggests that, by moving the factor to higher values, the response will decrease. The absolute value of the factor effect represents the influence it has on the response. Therefore, the ranking of the factor effects can split all the investigated factors into groups of those with high, medium, or low effects, which is actually the main purpose of the screening design – to present reliable information about the different influential factors and to reduce the number of investigated CPPs to only the ones that are most significant. Only the significant factors will enter the optimization phase, while the remaining CPP values will be held constant. The important question, however, is where and how to set the differentiation line between these two groups of factors. Different statistical analyses can be used to determine the significant factor effects – graphical analysis by half-normal or normal probability plots or comparison of the critical effect value calculated by the algorithm of Dong and error estimation based on a priori declared negligible effects [21]. 13.3.3.2

Optimization Designs

The selection of experimental design depends on the mathematical relation between factors and responses that could be expected. Although there are various theoretical models that can describe the chromatographic system, empirical models are extremely useful when the influence of several factors on the retention behavior of the substances is of interest because, most of the time, there are no theoretical models that include all of the chromatographically relevant parameters [14]. The DoE methodology usually fits a polynomial model to the data. Depending on the selected factors and responses, an adequate polynomial degree is chosen; however, the most commonly applied model is a second-order polynomial model. Therefore, the optimization phase commonly includes the application of response surface designs or response surface methodology (RSM), which essentially refers to the creation of a mathematical dependence of the responses on the investigated factors, which in certain cases can be represented as a surface. The experimental designs, which belong in this category, investigate each factor on at least three levels. The most commonly applied designs of this type are central composite design (CCD), Box–Behnken design (BBD), Doehlert

291

292

13 Chemometrics in Data Analysis and Liquid Chromatographic Method Development

Figure 13.3 Points representing the experimental plan for (a) central composite design (three investigated factors), (b) Box–Behnken design (three investigated factors), and (c) Doehlert design (two investigated factors).

design, D-optimal design, and mixture design. The graphical representations of the experiments required by some of these designs for certain number of factors are presented in Figure 13.3. Central composite design comprises two-level full factorial design, star design, and central point replications. Therefore, the number of required experiments for k factors in CCD is 2k + 2k + central point replications. The distance of experimental points defined by the star design from the design origin can be different, but they are most commonly selected to enable rotability of the design, creating rotatable CCD, or to be equal to 1 or +1 levels of the investigating factors, making face-centered CCD. The Box-Behnken design examines the factors on three levels and provides the same mathematical model as CCD. However, the localization of the experimental points is different, as can be observed in Figure 13.3. BBD consists of a threelevel incomplete factorial design. For k investigated factors, it requires 2k(k 1) points plus a central point, which is a smaller number of experiments than CCD for the same number of factors. However, the lower number of experiments means a lower number of degrees of freedom. Furthermore, the difference between these two designs is that CCD contains the extreme factor combinations, while BBD does not examine the borderline regions. Both characteristics can be advantageous, depending on the experimental problem (in certain cases,

13.3 Chemometrics in LC Method Development

the extreme combinations must be examined, while in others these combinations are physically impossible). The Doehlert design enables the estimation of factors on a different number of levels, which can be useful for certain experimental problems. Therefore, the factor that is expected to have the greatest influence on the system can be examined on more levels. For k factors, the experimental points of this design are represented by one point in the origin and k points on the surface of a sphere constructed around the origin, which provides the equidistant adjacent points. Each of these points creates k new points, leading to a total number of k2 + k + 1 points. D-Optimal design generates design matrices that maximize the determinant of the information matrix. This design is a valuable alternative to the classical designs and particularly useful when the experimental region is not of a regular shape. Furthermore, this design type allows for the creation of nontypical mathematical models (e.g., higher order polynomial models instead of the traditionally applied second-order model). Finally, compared to the classical factorial designs, the number of required experiments is significantly smaller. The creation of this design matrix includes the predefinition of the desired mathematical model and the selection of the optimal number of design points. For each possible number of experimental points, the computer algorithm generates a set of experimental points that satisfy the best the D-optimality criterion and then decides which points represent the best results. The mixture design is generally applied when the investigated factors are not independent. In liquid chromatography, this design is most commonly applied to the investigation of the mobile phase composition, and in that case, the proportions of the constituents are investigated. When this design is applied, the factor levels are not typically defined but reach an overall value of 1 (or 100%). 13.3.4 Critical Quality Attributes Modeling (Creation of Mathematical Models)

After performing the experiments defined by the selected design, the experimentally obtained data are fitted (approximated) using multiple linear regressions and the least squares method of approximation. Depending on the selected factors and responses, the suitable mathematical model is chosen; however, the most commonly applied model is a second-order polynomial model represented by the following equation: Y ˆ β0 ‡

N X iˆ1

βi X i ‡

N X N X

βij X ij

(13.2)

iˆ1 jˆ1

where Y is the response and β0, βi, βij, and βii correspond to the intercept, linear, interaction, and quadratic term coefficients, respectively. This mathematical model provides a quantitative description of the impact of the factors on the response and leads to the creation of linear or curved response surfaces [12]. The estimation of all the parameters included in Equation 13.2

293

294

13 Chemometrics in Data Analysis and Liquid Chromatographic Method Development

requires that the experiments are performed at least on three levels for each factor. The constructed mathematical relation between CPP and CQA or other responses allows for a theoretical examination of the experimental space without performing additional experiments. 13.3.5 Design Space

The design space (DS) represents “the multidimensional combination and interaction of input variables and process parameters that have been demonstrated to provide assurance of quality” [13]. Essentially, the DS represents the region of robustness, that is, the part of the experimental space where the defined CQA will remain within the desired quality level regardless of the changes in the factors, that is, CPPs. Therefore, the optimal analytical method conditions are not defined using a single point anymore but as a domain of experimental space within the knowledge space. This approach to method development provides a reliable confirmation of its quality because all of the potential method instabilities caused by slight changes in the experimental conditions are explained [14,15]. Importantly, the size of the design space directly explains the method robustness, and if it is a small design space, analysts should preserve the experimental conditions. Furthermore, the method changes within the design space are not considered a change, and additional quality controls are required only when the conditions are shifted out of the DS region. Therefore, method transfer between laboratories is significantly easier. In the first phase of the DS definition, the column design space should be determined, and furthermore, the DS for the mobile phase and other process parameters is set [22]. The column design space includes the identification of equivalent or orthogonal columns to define the robustness of the stationary phase. The existence of a column DS ensures that, in case of the destruction of a certain column, the analysis can be performed on another column without quality loss. Equivalent columns can be defined using the column databases. The DS for mobile phase constituents and the remaining process parameters can be defined according to the established mathematical relationship between the CPPs and the CQA. If modeling is performed not for CQA but for simpler chromatographic responses, CQA should be modeled indirectly, that is, calculated a posteriori. Finally, the theoretical changes to CPP should be simulated and the resulting CQA values for each simulation should be recorded. Subsequently, the design space is defined as the experimental region where all CQA values are satisfactory. For low CPP numbers, the design space can be graphically represented and easily visualized. In contrast, if a larger number of CPPs is to be examined, the design space is represented numerically. The multiobjective optimization problems where several CQAs should be achieved simultaneously can be resolved in different ways. One approach is the application of chromatographic response functions that combine several quality criteria into a single numerical value. A variety of objective functions have been

13.3 Chemometrics in LC Method Development

defined in the literature. The Derringer desirability function is one of the most commonly applied. The advantage of this approach is the simplification of the multiobjective problem to a uniobjective problem, which facilitates the identification of the optimal conditions and the definition of the design space. However, if the researcher does not want to apply objective functions, the multiobjective approach should be performed to create design space for each CQA separately to find the regions where all CQAs have satisfactory values. In case of a small number of investigated factors, overlaid contour plots can be created for simpler visualization of the design space. 13.3.6 Selection of the Working Points

A working point can be any point within the design space. It can be selected as the point with the best value for a particular CQA or objective function, or it can be some point that is suitable for experimental work. However, it is desirable to select the most robust design point, as well. Therefore, points located at the edges of the design space should be avoided. Moreover, evaluating the uncertainty of the model prediction is highly recommended. Namely, without examination of this possible source of uncertainty, the design space is called a pseudodesign space, and it becomes the real design space only when this aspect is considered. Recent advances in pharmaceutical science highlight the implementation of model uncertainty as an additional source of variation of method performance. Therefore, the equivalent zone for the working point should be assessed to examine whether the distribution of the selected criterion is greater than a chosen limit [20]. The coefficients of uncertainty can be assessed using Monte Carlo simulation, Bayesian modeling, or bootstrapping techniques [14]. 13.3.7 Robustness Testing

As previously discussed, the chemometric strategies include the robustness concept from the very beginning of the method development process, which significantly decreases the risk of method failure. However, a final confirmation of robustness through experimental testing is required at the end of the optimization phase. Traditionally, the robustness is tested by forcing a narrow variation in each potentially critical factor, while the values of the remaining factors are kept constant. If the method is robust enough, the selected responses will stay within the required limits in all of the experiments. Once again, the DoE methodology provides a different approach to this problem [21]. It suggests the application of designs that are fractional factorial or Plackett–Burman designs for the definition of a robustness testing experimental plan. First, this approach enables the investigation of a large number of factors in a relatively small number of experiments. For example, the PB design requires 12 experiments for the investigation of 11 factors, while the traditional one-factor-at-a-time approach would

295

296

13 Chemometrics in Data Analysis and Liquid Chromatographic Method Development

require 22 experiments. Furthermore, by following the DoE methodology for each experiment, the different factors are simultaneously varied, putting the investigated chromatographic system under realistic conditions. That is, in practice, it is more logical to expect that several potentially critical factors will deviate simultaneously than that only isolated factors will be changed, while the remaining are fixed. Therefore, the traditional robustness testing gives no information about the stability of the system when all goes wrong, that is, when several important factors slightly switch their values in the same direction, which deteriorates the response of interest. After performing the experiments, the interpretation of the factor effects on the important responses can be performed according to the procedure described in Section 3.1. Therefore, it is possible to create a ranking of the factor effects and to identify those that significantly influence the system and whose values should be carefully maintained during routine procedures. In addition, it is possible to calculate the nonsignificant intervals for the significant factors, which determines the upper and lower borders of the significant factor interval within which the response will remain satisfactory. The results obtained in this way should confirm the factors’ borders defined by design space. Finally, in addition to the information about how the system reacts in each of the performed experiments, according to the DoE matrix, knowledge of the system behavior under all possible conditions within the investigated factor intervals can be gained without performing additional experiments. In this way, the theoretical values of the important responses in a worst-case scenario can be calculated, which is a valuable information for the determination of the system suitability limits.

13.4 Conclusions

To obtain reliable and useful data from chromatographic analysis, it is very important to have a comprehensive approach. In this chapter, the role of chemometrics in chromatographic data analysis and LC method development were presented. By taking into account many chemometrics tools and many real problems, a short description of chemometrics method utilization was given. Special attention was given to the application of chemometrics in the development of LC methods in accordance with modern QbD principles explaining in detail the systematic approach to the creation of highly reliable chromatographic methods.

References 1 Amigo, J.M., Skov, T., and Bro, R. (2010)

ChroMATHography: solving chromatographic issues with mathematical models and intuitive graphs. Chem. Rev., 110, 4582–4605.

2 Harynuk, J.H., de la Mata, P.A., and

Sinkov, N.A. (2012) Application of chemometrics to the interpretation of analytical separation data, in Chemometrics in Practical Application

References

3

4

5

6

7

8

9

10

11

12

13

14

(ed. K. Varmuza), In Tech Europe, Rijeka, Croatia, pp. 305–326. Laursen, K., Frederiksen, S.S., Leuenhagen, C., and Bro, R. (2010) Chemometric quality control of chromatographic purity. J. Chromatogr. A, 1217, 6503–6510. Ruckebusch, C. and Blanchet, L. (2013) Multivariate curve resolution: a review of advanced and tailored applications and challenges. Anal. Chim. Acta, 765, 28–36. Kumar, N., Bansal, A., Sarma, G.S., and Rawal, R.K. (2014) Chemometrics tools used in analytical chemistry: an overview. Talanta, 123, 186–199. Keller, H.R. and Massart, D.L. (1992) Evolving factor analysis. Chemometr. Intell. Lab., 12, 209–224. Ortiz, M.C. and Sarabia, L. (2007) Quantitative determination in chromatographic analysis based on n-way calibration strategies. J. Chromatogr. A, 1158, 94–110. Bro, R. (1997) PARAFAC. Tutorial and applications. Chemometr. Intell. Lab., 38, 149–171. Brereton, R.G. (2003) Chemometrics: Data Analysis for the Laboratory and Chemical Plant, John Wiley & Sons Ltd., Chichester, England. Li Vigni, M., Durante, C., and Cocchi, M. (2013) Exploratory data analysis, in Data Handling in Science and Technology (eds S. Rutan and B. Walczak), Elsevier, The Netherlands. Otto, M. (2007) Chemometrics: Staitsics and Computer Application in Analycal Chemistry, Wiley-VCH Verlag GmbH, Weinheim, Germany. Vandeginste, B.G.M., Massart, D.L., Buydens, L.M.C., De Jong, S., and Lewi, P.J., Smeyers-Verbeke (1998) Handbook of Chemometrics and Qualimetrics: Part B, Elsevier, The Netherlands. International Conference on Harmonization (ICH) of Technical Requirements for Registration of Pharmaceuticals for Human Use, Topic Q8 (R2): Pharmaceutical Development, ICH, Geneva, Switzerland (2009). Rozet, E., Lebrun, P., Debrus, B., Boulanger, P., and Hubert, Ph. (2013)

15

16

17

18

19

20

21

22

Design space for analytical methods. Trend. Anal. Chem., 42, 157–167. Orlandini, S., Pinzauti, S., and Furulanetto, S. (2013) Application of quality by design to the development of analytical separation methods. Anal. Bioanal. Chem., 405, 443–450. Cela, R., Ordonez, E.Y., Quintana, J.B., and Rodil, R. (2013) Chemometric-assisted method development in reversed-phase liquid chromatography. J. Chromatogr. A, 1287, 2–22. Mbinze, J.K., Lebrun, P., Debrus, B., Dispas, A., Kalenda, N., Mavar Tayey Mbay, J., Schofield, T., Boulanger, B., Rozet, E., Hubert, Ph., and Marini, R.D. (2012) Application of an innovative design space optimization strategy to the development of liquid chromatographic methods to combat potentially counterfeit nonsteroidal anti-inflammatory drugs. J. Chromatogr. A, 1263, 113–124. Jančic ́ Stojanovic ́, B., Rakic ́, T., Kostic ́, N., Vemic ́, A., Malenovic ́, A., Ivanovic ́, D., and Medenica, M. (2011) Advancement in optimization tactic achieved by newly developed chromatographic response function: application to LC separation of raloxifene and its impurities. Talanta, 85, 1453–1460. Rakic ́, T., Jančic ́ Stojanovic ́, B., Malenovic ́, A., Ivanovic ́, D., and Medenica, M. (2012) Improved chromatographic response function in HILIC analysis: application to mixture of antidepressants. Talanta, 98, 54–61. Lebrun, P., Govaerts, B., Debrus, B., Ceccato, A., Caliaro, G., Hubert, P., and Boulanger, B. (2008) Development of a new predictive modelling technique to find with confidence equivalence zone and design space of chromatographic analytical methods. Chemometr. Intell. Lab., 9, 4–16. Vander Heyden, Y., Nijhuis, A., SmeyersVerbeke, J., Vandeginste, B.G.M., and Massart, D.L. (2001) Guidance for robustness/ruggedness tests in method validation. J. Pharm. Biomed. Anal., 24, 723–753. Kormány, R., Molnár, I., and Rieger, H.J. (2013) Exploring better column selectivity choices in ultra-high performance liquid

297

298

13 Chemometrics in Data Analysis and Liquid Chromatographic Method Development

chromatography using Quality by Design principles. J. Pharm. Biomed. Anal., 80, 79–88. 23 Doornbos, D.A., Smilde, A.K., de Boer, J.H., and Duineveld, C.A.A. (1990) Experimental design, response surface methodology and multicriteria decision making in the development of drug dosage forms, in Scientific Computing and Automation (Europe) (ed. E.J. Karjalainen), Elsevier, The Netherlands.

24 Hibbert, D.B. (2012) Experimental design

in chromatography: a tutorial review. J. Chromatogr. B, 910, 2–13. 25 Ferreira, S.L.C., Bruns, R.E., da Silva, E.G.P., dos Santos, W.N.L., Quintella, C.M., David, J.M., de Andrade, J.B., Breitkreitz, M.C., Jardim, I.C.F.S., and Neto, B.B. (2007) Statistical designs and response surface techniques for the optimization of chromatographic systems. J. Chromatogr. A, 1158, 2–14.

INDEX TO VOLUME 1

Index Terms

Links

A Abraham’s LSER approach

221

absorption detectors

248

UV/visible light

252

abundant fragmentation

101

accuracy mass measurement

38 111

acetaminophen

59

acetone

45

acetonitrile (ACN)

21

34

37

63

75

168

174

187

188

230 organic modifier

138

toxic

168

water mixtures

201

water mobile phases

143

water solutions

143

acid-base behavior

202

acid–base dissociation constant

202

acid-base systems

187

93

acidic compounds, loses

263

acidity-basicity character

170

This page has been reformatted by Knovel to provide easier navigation.

Index Terms ACN/buffer mobile phases

Links 145

active flow technology (AFT) columns

44

51

52–54

58

54

57

illustration of AFT end fitting

51–54

58

parallel segmented flow

51–54

58

curtain flow efficiency

performance of AFT columns

53

sensitivity

53

speed

58

AC voltage

259

additives adsorption chromatography isotherms aerosol generation

54

58

21

183

35 4 26 3

4

257

affinity chromatography (AC)

26

agglomerates

45

air-driven fluid pump

45

alcohols

34

aldicarb sulfoxide

120

alkaline eluants

235

alkaline hydrolysis

75

alkyl chain phases

163

alkyl ether sulfates

115

alkyl ligand

167

alkylsilane

20

surfaces

58

232

35

161

alkyl stationary phases

163

alkyl sulfates

115

This page has been reformatted by Knovel to provide easier navigation.

Index Terms alkyne–azide click chemistry

Links 76

allergens

123

alumina

43

stationary phase

234

234

American Society for Mass Spectrometry (ASMS)

89

amide-based stationary phase

75

amines

5

amino acids sequence amino-based materials amino columns, diol columns

123 75 235

amino-modified material

75

amino-modified silica

67

aminopropyl silica

75

ammonium chloride buffer

151

ammonium formate

265

analyte retention

273

pH-dependent changes

273

pH effect

266

time

268

75

analytes behavior chiral chromatography

268 5

dissociation

269

instantaneous

270

pH gradient retention time

273

polarity (hydrophilicity) of

276

analytical chromatograms, for metoprolol 2-phenylbyturic acid

227 20

This page has been reformatted by Knovel to provide easier navigation.

Index Terms

Links

analytical chromatograms, for metoprolol (Cont.) 3-phenyl-1-propanol analytical (linear) chromatography

20 4

analytical method goal of

288

analytical target

286

anionic surfactants

115

anthelmintics

120

anti-Langmuir behavior adsorption isotherm antinutrients APCI/APPI sources

4 123 96

apolar alkyl groups, bonded silica-based stationary phases

236

apolar compounds

114

aqueous buffer solutions

147

aqueous mobile phases

21

173

aqueous normal-phase chromatography (ANPC) aqueous-organic mixtures asymmetry factors

230 161

171

182

88

113

260

89

94

95

20

atmospheric pressure chemical ionization (APCI) interface

103 process

94

source

93

94

atmospheric pressure electron capture dissociation (APECD) source

98

This page has been reformatted by Knovel to provide easier navigation.

Index Terms

Links

atmospheric pressure ionization (API) techniques

87

88

99

88

98

99

88

114

101 atmospheric pressure laser ionization (APLI) atmospheric pressure photoionization (APPI) source

96

average particle size

29

average zone width

13

avermectins axial heterogeneity

120 46

B back pressure band broadening with increasing flow rate

37 7

8

18

54

11

baseline drift

38

batch reactors

116

Bayesian modeling

295

β-blocker chromatograms for

165

benzimidazoles

120

benzylamine, at pH values

223

binodal line

145

bioinformatics

121

biological sample

126

40

biomolecules, characterization of

111

biopesticides

119

This page has been reformatted by Knovel to provide easier navigation.

10–13

Index Terms

Links

bis(triethoxysilyl)ethane

20

Boltzmann’s constant

17

bonded ionic groups

232

bonding type

33

bootstrapping techniques

295

bottom-up approach

123

Box-Behnken design (BBD)

291

Brønsted model

147

292

buffers cations

259

hydro-organic mobile phases

147

ionization constants

148

butocarboxim sulfoxide

120

butyl benzene band profiles

55

C caffeine

59

capacity factors on Click TE-Cys and HILIC columns

77

capillary columns

76

capillary electrophoresis (CE)

93

capillary separated vaporization chamber and nozzle (CSVCN) system

104

carbapenems

76

carbohydrates

63

carbon loading

33

cardiovascular diseases

125

carotenoids

127

carrier gas, flow rate

257

casein variants

123

67

This page has been reformatted by Knovel to provide easier navigation.

Index Terms

Links

αs1-casein

123

catecholamines

261

cationic solutes, on silanols sorption-desorption kinetics of C18-based silanes C18 columns core-shell column

164 163 45

163

113

cell calibration

171

central composite design (CCD)

291

central point injection

44

cephalosporins

76

ceramics

43

charge-based HPLC separation

232

charged aerosol detection (CAD)

258

charge exchange

97

charge transfers

258

292

chemical bonding (grafting) chromatographic ligand chemical interactions chemical stability

161 137 21

chemometrics analytical target profile/critical quality attributes

286

critical process parameters

287

critical quality attributes modeling

293

data analysis

280

design space (DS)

294

investigation of

288

knowledge space

288

in LC method development

285

284

This page has been reformatted by Knovel to provide easier navigation.

Index Terms

Links

chemometrics (Cont.) optimization designs

291

QbD approach in LC method

286

quality risk assessment

287

robustness testing

295

screening designs

289

strategy, advantage of

279

techniques

279

trial and error approach

285

working point

295

chiral chromatography

233

chiral HPLC separations

233

chiral preparative chromatography

3

chiral separations

236

chiral stationary phase (CSP)

233

chloroform

145

146

241

7

8

12

271

272

chromatograms analysis

281

chromatograph analysis

280

behaviors

192

chromatographer columns

282

25

219

6

27

266 conditions

25

data

280

detector

120

molecule, structure of

228

retention

263

281

This page has been reformatted by Knovel to provide easier navigation.

159

Index Terms

Links

chromatographic hydrophobicity index (CHI)

274

chromatographic hydrophobic retention

151

chromatographic objective function

219

chromatographic packing materials

32

chromatographic performance

17

chromatographic practice

214

chromatographic retention

275

behavior of probe solutes

185

mobile phase, pH Measurements of

264

pH effects

263

chromatographic selectivity

34

chromatographic separations

26

conditions

190

229

218

chromatographic signals

280

baseline and noise

281

chromatographic system

25

35

279

280

buffers, characteristics of

281

265

high-pressure liquid chromatography (HPLC)

228

HPLC separation modes charge-based separations

232

other separation mechanisms

232

polarity-based separations

228

size-exclusion chromatography (SEC)

232

normal-phase/polar organic solvents

233

bonded phases

235

isocratic/gradient elution

241

mobile-phase selection

238

This page has been reformatted by Knovel to provide easier navigation.

227

Index Terms

Links

chromatographic system normal-phase/polar organic solvents (Cont.) nonbonded phases

234

retention mechanism

234

solvent strength/selectivity

239

stationary phases

234

stationary phases/selectivity

236

pH-related effects

274

physical methods

227

plant pigments, colorful separation

227

software packages

281

chromatographie separation

256

ChromDream

220

Click Maltose column material

76 76

C18 ligand

183

coccidiostatic drug

120

collision-induced dissociation (CID) mode

88

99

column bed heterogeneity

46

49

axial heterogeneity

46

49

radial heterogeneity and wall effect

49

51

in slurry packed columns

50

column dead time

27

column dead volume

43

column dimensions

32

column efficiency

12–14

16

52 effect of temperature on

36

column effluent ion-exchange column

259

This page has been reformatted by Knovel to provide easier navigation.

19

Index Terms column lengths column loadability column packing

Links 9

16

57

45

163

19

32

188 15

equipment used for downward slurry packed columns

45

material

276

processes

45

techniques

44

column performance

44

column resistance factor

15

columns

9

advancement

39

advantages

37

with core-shell particles

39

diameter

190

properties

266

selection

35

size

32

technology

40

43

temperatures

17

237

column wall commercial GC–MS

52 101

comprehensive two-dimensional liquid chromatography (LC/LC) mode compression

40 39

computer-assisted interpretive optimization strategies conductivity, detectors

219 259

contaminants for data acquisition

111

This page has been reformatted by Knovel to provide easier navigation.

Index Terms

Links

conventional column

55

conventional HPLC

39

conventional pressure

15

core-shell particles

15

39

113 corona charged aerosol detectors (CAD)

151

correlation optimized warping (COW)

282

coshift

282

Coulomb interactions

137

countercurrent chromatography (CCC) technique

136

C18 porous silica microparticles

192

227

critical micellar concentration (CMC) in mobile phase

231

critical process parameters (CPPs)

287

critical quality attributes (CQAs)

286

multiobjective optimization problems cross contamination

287

294 13

cross-linked polymer structures

162

Cucurbita maxima

81

curtain flow column (CF)

51

chromatography columns

52

environment

53

cyano-based silica

75

cyanopropyl phases

32

cyclofructans (CFs) cyclofructan 6 (CF6)-based stationary phases cysteine

76 76

This page has been reformatted by Knovel to provide easier navigation.

43

Index Terms

Links

D data acquisition

114

data analysis, chemometrics

280

284

database for simultaneous analysis using UHPLC–orbitrap MS workflow scheme used to create and apply data-dependent acquisition (DDA) mode data overloading

119 119 115 121 283

data preprocessing chromatographic analysis, problem solving

280

dead time estimation, in chromatographic systems

206

dead volumes

39

Derringer desirability function

295

design of experiments (DoE) methodology

288

desorption electrospray, ionization (DESI)

106

detection

38

limits

116

detectors

44

linearity specifications used in LC deuterium

55

247 39 261 253

lamp

250

ultraviolet excitation and long-term stability

254

deuterium oxide

173

deviations, from linearity

208

dewetting, bonding density

189

This page has been reformatted by Knovel to provide easier navigation.

280

Index Terms dichloromethane (DCM) 3,5-dichlorophenol

Links 46

142

204

diffraction optical arrangements diffusion

255 18

adsorption constant

11

coefficients

10

in liquid phase length

12 18

1,2 dihydroxypropylether

235

N,N-dimethylformamide

207

Dimroth–Reichardt parameter

201

diol-based silica

75

dioxane

138

1,4-dioxane index

169

diphenhydramine

269

retention times

26

dipole interactions

32

dipoles (polarity)

168

disaccharide (Click Maltose)

76

discharge lamps, low-pressure

250

dispersion

36

dispersive solvent

45

distribution equilibrium divinylbenzene (DB) Doehlert design

170

270

dipole-dipole interaction

dispersive SPE (dSPE)

29

65

137

295

296

118 6 113 291–293

DoE methodology

291

D-optimal design

293

This page has been reformatted by Knovel to provide easier navigation.

Index Terms

Links

downward slurry packing methods

45

drug development

77

DryLab

220

dual polarity detection

120

dwell time

217

47

E eco-toxicity eddy diffusion

112 9

10

12

28

30

36 efficiencies

13–15

of AFT column depends on segmentation ratio chromatographic system

55 8

gain of AFT column

57

for 3 µm particle-packed columns

57

trend

15

9

electrochemical (EC) detection (ECD)

257

258

260

261 schematic representation of electrode contamination

257 261

electron affinity (EA)

97

electron capture (EC)

97

electron ionization (EI)

87

100

114

interface

101

102

104

ionization

101

electropneumatic heated nebulizer electrospray

98 113

This page has been reformatted by Knovel to provide easier navigation.

Index Terms electrospray ionization (ESI) analysis APPI interface

Links 88 92 100

ion

88

nebulization process

91

sources

92

TOF-MS detection

260

90

274

electrospray ionization source (ESI-TOF-MS) methods

274

electrospray quadrupole iontrap mass spectrometry

81

electrostatic forces

26

electrostatic interactions

21

eluents

21

nonpolar pH retention factor for

21 277 65

sequence, polarity order

240

strength

210

eluotropic strengths

238

embedded polar group (EPG)

167

empirical equations

148

enantiomers

233

end-capping

33

columns

166

reagent

167

end-point detection processes environmental analysis

65

240

166

44 112

Escherichia coli

81

ethanol

34

113 169

This page has been reformatted by Knovel to provide easier navigation.

67

Index Terms

Links

ethanolamine

75

ethanol index

170

ethanol, toxic

168

ethylene-bridged hybrid silica

67

evaporative light scattering detector (ELSD) schematic diagram evolving factor analysis (EFA)

246

256

256 283

exploratory data analysis (EDA) cluster analysis external standard calibration extra–column volume extracted ion chromatograms

284 38 206 59

F factor analysis (FA)

284

fatty acids

126

filter-filter detector

254

flow cell

55

flow rates

10

11

15

17

19

35–37

39

45

46

59

124

optimum

37

flow sensitive detectors

54

flow velocities

47

49

58 fluid flow

46

flukicides

120

fluorescence (FLD)

254

This page has been reformatted by Knovel to provide easier navigation.

50

Index Terms fluorescence derivatization

Links 252

fluorescence detection

50

disadvantage of

253

fluorescence detectors

254

block diagram of

253

designs

254

optical arrangement for

253

253

fluorescence excitation

253

fluorescence spectroscopies

185

fluorescent emission

254

food bioactive peptides

124

food lipidomics

125

127

food metabolomics

124

125

foodomics

121

approach

121

food proteomics

121

food quality

124

food science

121

food toxicants

117

applications formic acid

124 124

117 80

Fourier transformation (FT) infrared

185

Fourier transform ion cyclotron resonance (FT ICR)

111

Fourier transform MS (FTMS)

260

3–1

fractional factorial design

290

fractional factorial design (FrFD)

289

2

fragmentation collisional-induced dissociation (CID)

99

100

99

This page has been reformatted by Knovel to provide easier navigation.

102

Index Terms

Links

Fresnel-type RI detector

255

optical arrangements

255

FrFD, limitation of

290

FTICR-MS data

123

full factorial design (FFD)

289

one-factor-at-the-time approach

289

properties of

289

fused core particles

17

18

191

35

95

G galactose-modified silica

76

gas chromatography (GC)

11 227

EIMS interface

100

MS, interface

102

gas phase ion–molecule reactions reactions

104 97

gas–vapor mixture

94

gate valve

46

Gaussian distribution Gaussian/Lorentzian curves Gaussian peak shape Gaussian profiles

7 282 5 280

gel filtration chromatography (GFC)

26

232

gel permeation chromatography (GPC)

44

232

ghost peaks

38

glass, electrode glutathione biosynthesis gluten proteins

148 76

81

123

This page has been reformatted by Knovel to provide easier navigation.

Index Terms

Links

glycans

78

glycoconjugates

78

glycosylation

79

gradient elution

37

38

147

174

179

180

216

219

242

49

57

gradient grade solvents

179

gradient of organic modifier

216

gradient program

218

gradient retention times

199

gradient system

37

gradient time

37

grafting density

161

graphitized carbon

113

green chromatography

21

greenest solvent

142

green solvents

142

H Handerson-Hasselbalch equation heart-cut, approach

263 40

height equivalent to theoretical plate (HETP) curves

48 59

helium, molecules

99

heterogeneity of lipids of packed bed

126 53

hexafluorobenzene

97

hexane sulfonate

35

This page has been reformatted by Knovel to provide easier navigation.

Index Terms

Links

high-performance anion-exchange chromatography (HPAEC)

78

high-performance liquid chromatography (HPLC)

2

15

63

159

228

(HPLC)

25

40

228

classification of

228 43

44

93

222

247

256

high pressure liquid chromatography

columns conditions

136

detectors

246

with mass spectrometry (MS)

260

method development, strategy

38

260

pKa assay medium-to-high-throughput polarity-based separations separation modes separations size particles

274 239 59 239 17

ultraviolet (UV) detector

245

UV cutoff wavelengths of solvents

249

vs.UHPLC high-pressure packed columns

238

17 43

high-resolution mass spectrometry (HRMS)

114

analyzers

115

in environmental field

114

high-resolution or tandem (MS/MS) instruments

88

This page has been reformatted by Knovel to provide easier navigation.

Index Terms

Links

high-temperature liquid chromatography (HTLC) separations

111

173

113

high-throughput, UHPLC–ESI–MS/MS analysis Hildebrand

90 138

Δ parameter

141

solubility

138

polarity scales homogeneity, of packing

141 11

homologous series method

208

Horváth, solvophobic theory of

266

HPLC–EI–MS separation

102

hybrids instruments

111

LIT-orbitrap mass analyzer

125

materials

5

19

20

silica core

37

21

26

65

66

75

187

tandem instruments

115

tandem mass spectrometers

121

hydrocarbons hydrogen bonding

258

hydro-organic elution

146

mobile phase

148

reversed mobile phases

152

149

hydrophilic interactions

76

monolithic columns

77

This page has been reformatted by Knovel to provide easier navigation.

Index Terms

Links

hydrophilic (Cont.) partitioning

64

hydrophilic interaction liquid chromatography (HILIC) analysis application of

21

22

230

276

276 77

78

cation-exchange (HILIC/CEX) mixed mode zwitterionic material

76

columns

78

conventional NPLC stationary phases for

67

MS approaches in structural glycomics

79

MS-based metabolomics

80

MS in structural glycomics

78

resources

64

retention mechanism in

22

125

276

role in enrichment/analysis of protein posttranslational modifications

79

separation mechanism in

64

65

stationary phase

67

78

hydrophilicity

76

hydrophobic chains bond

162

retention

221

subtractive approach

221

222

subtractive model

221

222

surface

161

hydrophobic and ion exchange interactions

63

276

This page has been reformatted by Knovel to provide easier navigation.

80

Index Terms

Links

hydrophobic interaction chromatography (HIC)

231

hydrophobic interaction liquid chromatograpy (HILIC)

144

hydrophobicity

26

contribution

222

interactions

5

122 32

231 hyphenation, system

124

hypothetical mobile phase

201

I ICH Q2(R1), testing specified by icoshift

282

algorithm

282

aligns

282

ideal analytical chromatogram

2

ideal chromatographic conditions

280

ideal chromatographic signal

280

ideal Gaussian shape identification, points (IPs)

8 120

immobilization liquid film

182

polysaccharide chiral stationary phases

236

impurities

77

infinite diameter column

44

influence adsorption processes

53

209

information-dependent acquisition (IDA) experiments injection cycle time

115 59

This page has been reformatted by Knovel to provide easier navigation.

36

Index Terms injection techniques

Links 44

inline detectors characteristics

246

detector cell volume

247

dynamic range

247

linearity

247

selectivity

246

sensitivity

246

charged aerosol detection (CAD)

258

conductivity detectors

259

coupling detectors

260

electrochemical detector

257

evaporative light-scattering detector

256

fluorescence detector

252

GC detectors, for HPLC uses

245

HPLC detectors, comparison of

260

photodiode array detector

251

refractive index detector (RID)

245

UV-visible absorbance detector

247

fixed wavelength

249

flow cell of

248

variable wavelength

250

instantaneous retention factor

269

instrument dwell time

268

intermolecular interactions

229

in HILIC, diversity of internal diameter (ID)

66 101

intersystem crossing (ISC)

99

ion chromatography

44

ion-dipole interactions

255

259

137

This page has been reformatted by Knovel to provide easier navigation.

Index Terms

Links

ion-exchange chromatography

26

ion-exchange interaction

36

ion-exchange particles

43

ion formation

95

mechanism of

101

ion fragmentation

91

ionic liquids (ILs)

143

ionic surfactants

231

63

232

ionization analytes retention

269

compounds

150

efficiency

97

mechanism

97

process

97

solutes

170

techniques

88

ion–molecule collisions

101

ion–molecule reactions

104

ion-pair chromatography (IPC)

231

202

100

ion pairing agents

231

reagents

35

ion, suppression

126

isocratic conditions

26

isocratic elution

37

isocratic measurements

176

isocratic mobile phase

181

isocratic mode

268

isocratic retention, effect of pH

266

isocratic retention time

200

231

38

267

This page has been reformatted by Knovel to provide easier navigation.

242

Index Terms isoeluotropic mobile phases isotope-labeled internal standards (ISs) IT-TOF detector

Links 146 92 125

K KCl solution

148

ketoprofen

266

Kinetex and Accucore columns

43

Kinetex HILIC core–shell column

77

kinetics

273

6

KISS principle

25

k value

30

L lactose–modified silica Lambert’s law

76 247

laser beams

98

laser pulses

99

LC-based methods DESI–MS interface

106

effluent

100

EI–MS coupling

100

fluorescence detectors

254

method development

287

parameters MS-based metabolomics

38 124

125

120

121

MS confirmatory analysis recent trends

This page has been reformatted by Knovel to provide easier navigation.

Index Terms

Links

LC-based methods (Cont.) MS instrumentations

93

102

112 MS market

89

MS/MS analysis

122

of pKa determination

274

MS screening analysis

118

MS system

112

MS techniques

111

QqQ system

114

retention “Troika”

66

TOF analysis

115

TOF-MS strategy

123

120 127 67

LC-MS interfaces API sources

88

atmospheric pressure chemical ionization

93

atmospheric pressure laser ionization

98

atmospheric pressure photoionization

95

electrospray interface (ESI)

89

factors influencing

92

modes of operation

92

operation/ion formation, principles of

90

non-API sources

94 96

99

DESI–MS interface

106

direct EI

100

single-photon low-pressure photoionization/EI ionization, combination in supersonic molecular beam (SMB)

104 103

This page has been reformatted by Knovel to provide easier navigation.

111

Index Terms

Links

LC-MS interfaces (Cont.) orientative distribution

89

Lewis acid-base theory

36

ligand-exchange (LE), interaction

36

ligand-solution interface

188

light intensities

255

limit of detection (LOD)

246

linear chromatography

3

linear combinations of predictor variables

215

linear free energy relationships (LFER)

220

linearity

39

linear organic modifier, gradient

268

linear pH gradient

270

linear regression model

65

linear retention model

217

247

215

linear solvation energy relationship (LESR) approach

66

220

model

220

221

solute descriptors

221

linear velocity flow

16

36

55

39

lipid components

126

lipidomics

121

125

126

1

2

11

15

21

35

39

87

227

245

259

liquid chromatography

advances and trends in

113

chemometric applications

280

method

285

This page has been reformatted by Knovel to provide easier navigation.

Index Terms

Links

liquid chromatography (Cont.) modern trends in

14

optimization strategies

199

systems

206

liquid–liquid, chromatography

228

liquid-liquid extraction (LLE)

1

78

145

152 liquid-liquid-like transfer process liquid-liquid partitioning

184 21

liquid mobile phase

227

liquid–solid chromatography

228

liquid solvent state

135

literature-based lipophilicity parameters

275

LIT hybrids

121

LIT-orbitrap over QqTOF

116

localized detector

44

log kw determination

274

longitudinal diffusion

10

low-resolution (LR) mass spectrometers

114

LR mass spectrometers

115

lysine-based zwitterionic material

228

76

M macrolides

120

markers, injection of

207

mass accuracy

124

mass spectrometry (MS)

advances and trends in

54

78

87

113

121

126

136

171

113

117

This page has been reformatted by Knovel to provide easier navigation.

Index Terms

Links

mass spectrometry (MS) (Cont.) analysis

63

chemical ionization

93

detection mass spectroscopy (MS) mass transfer

264 279 19

coefficient

29

kinetics

18

mathematical methods

208

mathematical models

35

MATLAB

36

37

39 274

220

matrix effect

92

matrix solid-phase dispersion (MSPD)

118

maximal peak compression

273

maximum peak compression

273

maximum residue limits (MRLs)

117

mechanical strength

43

MeOH/buffer mobile phase

150

MeOH mobile phase

150

mercury lamp

250

metabolic fingerprinting

124

metabolite profiling

81

metabolomics

76

databases

151

80

121

77

234

125

platform based on UHPLC fluorescence TOF-MS assay

125

metal-based packings

234

metal impurities in silica

235

metal-oxide N-methyl pyrrolidone

21 142

This page has been reformatted by Knovel to provide easier navigation.

Index Terms methanol (MeOH)

Links 21

34

35

37

45

113

148

173

230

methionine

54

method development, strategy

37

methyl-t-butyl ether (MTBE)

141

metoprolol

20

micellar liquid chromatography (MLC)

231

microbial transformation products

116

microelectrodes

50

migration zone

10

milk proteins

123

mobile phase

4–7

10

20

29

34

36

52

78

99

151

178

182

189

229

233

238

257

259

148

174

201

214

239

241

269

293

266 composition

evaporation (nebulization) of

256

flow rates

37

linear velocity

16

modifi ers of the short-chain alcohol type

234

pHs

263

polarity

200

polar solvents

229

purification and degassing of

253

268–271

This page has been reformatted by Knovel to provide easier navigation.

277

Index Terms

Links

mobile phase (Cont.) solvents

35

velocity

11

viscosity

136

18

29

9

10

12

21

36

model polarizability

66

modern UHPLC particle

18

modifier content

214

at solute location

217

in isocratic elution

219

retention factor and protonation constant simultaneous effect molar refraction molecular, dynamics molecular diffusion

205 205 66 185

molecularly imprinted (MI)-SPME techniques

118

monolith, technology

111

monolithic columns

19

37

44 monoliths

19

monomeric surfactant molecules

231

monosaccharide (Click Glucose)

76

Monte Carlo simulation

185

MS-based metabolomics

80

MS-based platforms for lipidomic

295

126

MS-compatible buffers MS detection selectivity

265 80

This page has been reformatted by Knovel to provide easier navigation.

39

Index Terms MS detectors

Links 59

MS/MS fragmentation

121

n

MS sensitivity

116

2

MS operations

111

MS proteomics

121

multicomponent protocols

117

multidimensional chromatography

114

39

multidimensional protein identification technology

122

multimode ionization

114

multiple small isocratic steps

218

multiresidue methods

112

multivariate curve resolution (MCR), algorithm

282

mycotoxins

117

m/z values

115

120

N nano-ESI systems nano HPLC systems

93 247

nano-LC columns

101

technology

100

nano-LC–MS system

101

nano-online HILIC–MS systems

80

nano-RP–LC–MS

80

naphthalene

21

narrow columns

32

AFT columns neutral loss scan (NLS)

58 118

This page has been reformatted by Knovel to provide easier navigation.

Index Terms new material trend

Links 19

NIST library

104

nitrogen (carrier gas)

256

nitromethane

169

index

170

NMR spectroscopy

80

noise contributions, for columns

56

non-API interface

100

nonaqueous reversed phase (NARP)

126

173

nonfluorescent compounds

252

molecules

252

nonionic surfactants nonlinear chromatography

115 3

nonlinear regression

208

nonpolar (hydrophobic) materials

159

nonpolar solvents

241

nonpolar stationary phases non-UV-absorbing components

5

239

248

normal-phase liquid chromatography (NPLC)

21

63

229 normal-phase (NP) mobile phases

98

normal-phase mode (NP-LC) mobile-phase solvents

238

overlapping/coeluting

239

pH effect on retention

276

separations

238

solvents

239

N-values

57

This page has been reformatted by Knovel to provide easier navigation.

201

Index Terms

Links

O octadecylsilane (C18 or ODS) 1-octanol/water partition coefficient

32 138

octyl C8-bonded phases

32

oligosaccharides

76

omic technique

121

174

on-column matched refractive index detection

50

opioid

124

optimal alignment technique

282

optimum linear velocity

29

optimum sample preparation technique

38

orbitrap analyzers

100

orbitrap detectors

111

orbitrap instruments

123

orbitrap mass analyzer

116

drawbacks

117

organic contaminants

112

organic micropollutants

112

organic mobile phase

230

organic modifiers gradient organic polymer backbone organic solvents organochlorine pesticides (OCPs) orthogonal retention behavior

126

116

21

34

269

273

276 37 114 44

ortho-phthaldehyde (OPA) derivatization

253

OSIRIS

220

This page has been reformatted by Knovel to provide easier navigation.

Index Terms

Links

P packing conditions

45

packing factor

11

packing materials

15

33

77

113

composition

36

particles

18

papaverine

266

parallel factor analysis (PARAFAC)

283

model, for multivariate data

283

36

parallel segmented flow column (PSF)

51

parking lot effect

47

particle diameter

14

particle pore structure

45

particle sizes

12

16–18

32

36

39

43

45

55

76

28

38

77 distribution

29

partition chromatography

26

partitioning interaction

65

partitioning process

183

parvalbumin

123

pattern recognition

284

PDA 3D-chromatogram

252

PDA detector, optical arrangement of

251

peak asymmetry

27

factor

27

peak fronting

5

This page has been reformatted by Knovel to provide easier navigation.

Index Terms peak tailing factor PEEK capillary tube pellicular material

Links 27 106 17

peptide-centric approach

123

peptides

124

perfluorinated compounds

113

permeability

19

44

pesticides

113

119

phase diagrams

146

pH control phenolic compounds

35 151

phenylalanine

54

phenyl-bonded phases

32

2-phenylbutyric (PB) acid

20

3-phenyl-1-propranolol (PP)

20

pH gradient conditions

272

pH gradient duration

271

pH gradient mode

273

pH gradient separation

269

phosphocholine

76

phosphorylation

79

photodetector, signal

250

photodiode array detector

251

photodiode detectors

249

photodiodes

251

photodissociation

99

photoionizable compound

97

photoionization

104

principle of

95

pH-sensitive electrode

273

171

This page has been reformatted by Knovel to provide easier navigation.

Index Terms pH shifts

Links 148

pH stability

43

pH-stable hybrid column (XBridge)

20

pH-stable silica

19

-based C18 hybrid packing materials pH values

22 20

physicochemical processes

228

piroxicam

59

PIS scan

116

pKa, determination

274

Plackett–Burman design (PBD)

289

plate height

8

concept

7

Pneumatic nebulizers

257

polar analytes

63

polar compounds

63

polar-embedded alkyl stationary phases

190

polar end-capped stationary phases

167

polar functional group

167

polar interactions polarity

290

5

26

35

75

-based chromatographic system polarities of stationary and mobile phases

230

of different stationary phases

237

model

210

order

240

polar lipids

126

polar metabolites

81

This page has been reformatted by Knovel to provide easier navigation.

Index Terms polar mobile phase

Links 5

polar organic compounds

32

polar organic solvents

21

chromatography

231

mode

237

polar organic solvents chromatography (POSC) polar pharmaceuticals polar pollutants polar stationary phases polyaromatic hydrocarbons (PAHs)

233 78 114 21 254

polyaspartamide-based materials

75

poly(aspartic acid) stationary phase

75

polychromatic light

251

polychromatic radiation

251

polycyclic aromatic hydrocarbons polyether ether ketone (PEEK) tube PolyGLYCOPLEX column

63

99 211 78

poly(2-hydroxyethyl aspartamide) stationary phase polymer-based monolithic columns (polymeric) compounds polymeric packing materials polymeric phases

75 39 228 36 162

polymerization

78

polyphenolics

124

polystyrene

113

poly(succinimide) silica

125

75

poly(2-sulfoethyl aspartamide) stationary phase

75

This page has been reformatted by Knovel to provide easier navigation.

237

Index Terms

Links

pore shape

29

pore sizes

33

distribution

29

porosity

15

19

porous graphitized carbon (PGC)

37

191

porous particles

113

porous polymeric rod

37

porous shell particles

113

porous silica

162

porous silica microparticles

191

post acquisition data processing

117

postcolumn derivations

252

chamber post-translational modifications

253 123

3-P,P-diphenylphosphoniumpropylsulfonate

76

p-p interaction

32

precision

38

precolumn derivatization

252

precursor ion scan (PrIS)

118

PREOPT-W

220

preparative chromatography

121

5

pressure

35

back

58

fluctuations

45

from pump

46

pressure drop

15

pressures

17

pressurized liquid extraction (PLE)

118

principal component analysis (PCA)

283

17

This page has been reformatted by Knovel to provide easier navigation.

66

Index Terms

Links

principal components (PCs)

284

product ion scan (PIS) experiments

115

propanol

34

protein composition

121

protein precipitation

78

proteins

123

analysis of

160

characterization of

123

proteolytic digestion proteomics

123 76

applications

123

workflow

122

protic solvent

148

proton acceptor

169

protonation constant

210

protons, acceptor of

170

proton transfer

97

PSF column

59

pulsations

49

pulsed amperometric detection (PAD)

78

pulsed direct injection

59

pulsed direct injection HPLC-MS

60

pumping clean mobile phase Purnell equation

35

121–123

250 29

30

Q QbD approach for LC method development

286

Q-linear ion traps (QqQLIT)

111

instruments, flexibility

115

This page has been reformatted by Knovel to provide easier navigation.

Index Terms

Links

quadrupole analyzer quadrupole time-of-flight (QqTOF)

106 111

hybrid instruments

119

instruments

116

quality by design (QbD) approach

285

quality risk assessment

287

quantitation quasi-isobaric interfering ions

120

126

39 115

QuEChERS methodology

117

R radial bed heterogeneity

49

compression columns

49

heterogeneity

46

radiation, adsorption

98

radical, molecular ion

101

random walk model rapid scanning variable wavelength

7 250

Rayleigh stability

90

refractive index (RI)

54

137

255

256

detectors refractive index detector (RID)

261

regioisomers

127

regular injection profile through a glass packed column

53

Reichardt ET30 parameter

141

Reichardt scale

142

This page has been reformatted by Knovel to provide easier navigation.

245

Index Terms

Links

relay gradients

176

reproducibility

63

residual silanols

164

resistance in mass transfer resolution

35 12–14

25

28

30

37

39

63

121

of carbohydrates

75

control

35

factor regulating column

29

control

31

of geometrical isomers

37

infl uenced by

30

influence on

14

vs. selectivity, retention, and efficiency

31

resonantly enhanced multiphoton ionization (REMPI) response surface methodology (RSM) retention

98 291 14

30

75

191 accuracy of predictions

203

of analytes

65

behavior

63

dependence of factor

177

204 7

10

11

14

15

27

29

36

66

200 of HILIC stationary phases

64

This page has been reformatted by Knovel to provide easier navigation.

Index Terms

Links

retention (Cont.) NPLC and POSC share retention factors

234 17

66

78

206

213

267

66

215

9

20

269

274

159

199

271 k value

33

uncertainty in prediction

216

retention, in RPLC

181

retention, in ternary solvent systems

200

retention mechanism

276

in HILIC

65

mobility of alkyl chains

186

solutes compete

188

retention models

6 218

retention of acidic species retention time

205 7 276

drift

281

in gradient elution

217

shift

281

reversed phase (RP), analysis

98

reversed phase high-performance liquid chromatography (RP-HPLC)

263 276

reversed phase liquid chromatography (RPLC)

63 230

with alkyl-bonded stationary phases

160

column

182

This page has been reformatted by Knovel to provide easier navigation.

Index Terms

Links

reversed phase liquid chromatography (Cont.) column, in routine laboratory

192

columns

161

184

221

computer-assisted interpretive optimization

218

conventional columns

203

dead time estimation

206

development and trends

160

dimethyl octadecylsilane systems

187

dynamic methods

207

190

fundamental equation for gradient elution

216

general elution problem of

175

gradient elution

216

gradients of modifier

175

176

179

181 isocratic elution

175

199

linear solvation energy relationships (LSER)

220

local models for characterizing columns

221

mobile phase

167

168

model regression process on quality of predictions modifier content, pH, temperature MS-based metabolomics

215 210 80

nonintegrable retention models

217

packing materials

163

pH, as experimental factor

202

polarity models

201

polynomial models

199

81 188

This page has been reformatted by Knovel to provide easier navigation.

170

Index Terms

Links

reversed phase liquid chromatography (Cont.) practical considerations

214

pressure effect

211

process

166

retention factors, deviations of

211

retention mechanisms

181–184

186

188

186

220

189 retention modeling

199

retention prediction, enhancing

214

retention using modifier content

199

reversed mode

159

silanol deactivation

166

silanol effects

164

silica support/chemical bonding

161

solvents

170

static methods

207

stationary-phase characterization

160

temperature, as chromatographic factor

172

temperature, effect of

209

Van’t Hoff equation

209

robustness

39

root-mean-square (RMS) variation

246

RP C18 columns

125

S saccharides

75

Saccharomyces cerevisiae

81

safety assessment

76

123

sample capacity

39

sample carryover

39

This page has been reformatted by Knovel to provide easier navigation.

Index Terms sample preparation LC–MS analysis sample retention

Links 38 112

117

118

55

13

scanning mass spectrometers

120

scanning speed

111

Schiff-base

75

segmentation ratio

52

54

103

114

114

115

118

12–15

19

27

29

31

39

44

59

75

113

114

selected reaction monitoring (SRM) transitions selectivity

factors influencing

33

of HILIC stationary phases

64

semiporous materials

17–19

semiporous particles

18

19

sensitivity

54

97

121 of detector

246

separation methods

39

280

characteristics

40

efficiencies

27

37

factor

14

27

separation science

40

separation selectivity

78

seven ionizable compound, separation of shell particles

150 39

shifting

282

shotgun methodology

126

This page has been reformatted by Knovel to provide easier navigation.

113

Index Terms Si atom Si-C bond signal intensity

Links 162 19 101

silanizing reagent hydrophobic surface silanols

166 164

blockers

167

polar adsorption centers

164

silanophilic interactions

235

164

silica batch-to-batch reproducibility

164

derivatization

166

hybrid material

21

stability of

37

silica backbone of the stationary phase

231

silica-based materials columns

166

monolithic column

212

silica-based monoliths

191

columns silica-based packing silica-based RPLC silica-based stationary phases

213

39

212

36

234

190 5

234

silica columns normal-phase gradient separations silica gel (SiO2)

242 67

silica modification, silanization

161

silica packings

166

silica surface functionalization

161

234

This page has been reformatted by Knovel to provide easier navigation.

239

Index Terms

Links

silicon, atom bridge

235

silicone–hydride (Si–H) groups

235

siloxanes

20

bond hydrolysis

235

170

silver ion chromatography (SIC)

126

SIM mode

104

single-photon ionization (SPI)

104

single photon low-pressure photoionizationelectron ionization interface

105

single-sample analysis

26

size exclusion chromatography (SEC)

26

slurry, reservoir

46

small-angle neutron scattering

185

small-particle stationary phases

17

S/N responses

54

Snyder–Soczewiński relationship

266

SST diagram

170

Snyder’s triangle

34

in Cartesian for choosing solvents for water and water-miscible solvents sodium phosphate

47

141 34 169 151

soft desorption ionization methods for mass spectrometric analyses

89

soft ionization technique

91

soft-walled columns

49

software tools for data acquisition

111

solid core particles

37

solid PEEK ring

52

solid-phase extraction (SPE)

78

117

This page has been reformatted by Knovel to provide easier navigation.

Index Terms solubility

Links 21

of hydrophilic analytes of mobile phase

26

63 200

solute dipolarity

66

hydrophobicity

174

intermolecular interactions of

168

migration efficiency molecules

51 6

polarizability

66

protonation of

203

retention

174

solute/mobile-phase/stationary-phase parameters solute property (SP)

200 221

solute–solvent–stationary-phase interactions

66

solutes pair measuring selectivity factor solvent consumption

237 32

45

35

eluotropic strength

238

fully miscible

143

ions

102

load

59

localization

238

melting point of

135

mixture

143

mobile-phase pH/buffers

147

239

This page has been reformatted by Knovel to provide easier navigation.

46

Index Terms

Links

solvent (Cont.) molecular weight/density/molar volume

136

nomograms

146

nonfully miscible

144

phase diagrams

144

pH, definition

147

pH, in hydro-organic mobile phases

147

physicochemical data of

153

physicochemical properties

135

143

pKa shifts, in hydro-organic mobile phases

148

polarity

138

recycling technologies

168

selectivity

240

selectivity groups

240

selectivity optimization

241

selectivity triangle

240

241

Snyder classification of

138

139

strength

21

40

type

34

viscosity

35

solvent extraction (SE)

117

solvent-rich mobile phases

150

solvent’s solving ability

138

solvent UV cutoff

238

solvophobic theory

182

spectral band width of detector

247

spectrophotometric detector

239

speed

39

136

183

59

This page has been reformatted by Knovel to provide easier navigation.

Index Terms

Links

SPE–HPLC-QqQ system

114

sphingomyelin

127

stability

67

over wide pH range

37

standard HPLC instruments

18

standard isotope-labeled (SIL) peptides stationary phase

123 4

6

7

10

27

28

35

36

233

architecture

185

characterization

199

concentration developed for HILIC environment for HILIC

19

222

3 75

76

184 67

of HTLC columns

113

of monolithic columns

113

particles

2

polarity

201

properties

228

surface

166

yields

36

steeper pH gradient

273

steric hindrance

137

steroids

114

stress, of analytical systems

113

68

strong cation exchange chromatography (SCX) structural protein characterization sub-2 µm particles

79 123 39

This page has been reformatted by Knovel to provide easier navigation.

Index Terms

Links

α/β sugar anomerization

75

sulfobetaine

76

sulfonamides

120

superior mass resolution

121

supersonic EI-LC-MS apparatus cold EI mass spectrum of

105

schematic of

103

supersonic molecular beams (SMB) approach

88 103

surface adsorption

65

surface area

33

surface-modified silica

230

surface tension

137

synchronization

180

system suitability system synchronization

39 180

T Tanaka–Euerby test

222

taurine

75

temperature

21

optimum

37

significant factor

35

temperature gradient

17

temperature-programmed liquid chromatography (TPLC) ternary-phase diagrams

35 145

tetraethoxysilane

19

tetrahydrofuran (THF)

34

141

230 This page has been reformatted by Knovel to provide easier navigation.

167

Index Terms

Links

theoretical plate (N)

8

thermal degradation

104

thermal vaporization

103

thermodynamic dissociation constants

264

thermospray interfaces

100

theroetical plate height (H) thiol-ene click chemistry

8 76

thiols

125

time of flight (TOF)

260

titania

43

pH stability

77

total gradient time

38

total ion current (TIC) chromatogram total particle size total variance total zone broadening

12

115 18 8 11

toxins

123

trial-and-error approaches

218

triethylamine (TEA)

35

triglycerides (TGs)

126

142

triple-quadrupole (QqQ) analyzer

120

mass analyzers

114

mass spectrometers

111

MS scan modes trypsin TSK Amide-80 column Tswett’s original straight-phase mode tungsten filament produces

59 114 122 75 5 101

This page has been reformatted by Knovel to provide easier navigation.

27–30

Index Terms turbulent flow chromatography (TFC) two-dimensional chromatography

Links 112 39

two-dimensional gas chromatography (2D-GC)

39

two-dimensional gel electrophoresis (2DGE)

122

two-dimensional separations

44

two-dimensional systems categories

40

type A silicas

235

type C silicas

235

types of stationary phases

234

types of ternary gradients

177

U ultrahigh-performance liquid chromatography (UHPLC)

15

33

39

87

111

173

228 orbitrap MS particles

119 18

QqLIT analysis

114

QqQ system

120

system

43

ultraperformance liquid chromatography (UPLC) system

19

ultrasonication

45

ultrasonic radiation

47

ultraviolet absorbance detector

260

absorbing component

248

This page has been reformatted by Knovel to provide easier navigation.

Index Terms

Links

ultraviolet (Cont.) absorbing impurities

180

cutoff value

239

detection

54

168

260

261 detectors

135

filters

113

radiation

252

sensitivity

233

transparent

135

vis absorbance

263

wavelengths universal residue analytical method uracil

137

99 112 21

V vacuum UV (VUV) lamp radiation

96 104

validation

38

van Deemter curves

16

35

van Deemter equation

29

30

35

36

39

174

van Deemter plots

29

30

39

van der Waals forces

26

van der Waals interactions

137

Van’t Hoff equation

209

vaporization

103

vegetable oils

126

210

This page has been reformatted by Knovel to provide easier navigation.

Index Terms velocity linear of molecule veterinary drugs

Links 16

45

58 6 117

119

vicinal silanols isolated silanols

164

vinyl silica

76

virtual column

53

internal diameter of viscosities

55 15

17

63 of methanol-water mixtures

168

of water

143

viscous friction

17

volatile buffer

63

volatility

151

W wall effect

49

water buffers

264

water-compatible phases

167

water impurities, of polar solvents

159

water-miscible solvents

160

water-organic mobile phase

266

water-soluble cellular metabolites weak acid (ketoprofen) retention factor on mobile-phase pH

167

81 266 267

weak bases isocratic retention, pH effects on weaker interactions

267 21

This page has been reformatted by Knovel to provide easier navigation.

35

Index Terms weighted regression

Links 215

well-established conventional gradient HPLC Workfl ow scheme

270 119

X xanthophylls Xbridge particles xenon lamps

127 19 253 254

Z zero absorbance

250

ZIC-HILIC column

21

ZIC–HILIC hyphenated offline with RPLC

79

zirconia

43

zirconia-based columns

36

zirconia-based stationary phases

221

zirconia substrate

36

zirconium oxides

113

161

9

11

zone-broadening effects Zorbax XDB columns

165

zwitterionic material

76

zwitterionic stationary phases

76

zwitterions

21

This page has been reformatted by Knovel to provide easier navigation.

Edited by Jared L. Anderson Alain Berthod Verónica Pino Estévez Apryll M. Stalcup

Analytical Separation Science

Volume 02

Editors Prof. Jared L. Anderson

The University of Toledo Department of Chemistry & Biochemistry 2801 W. Bancroft St., MS 602 OH United States

All books published by Wiley-VCH are carefully produced. Nevertheless, authors, editors, and publisher do not warrant the information contained in these books, including this book, to be free of errors. Readers are advised to keep in mind that statements, data, illustrations, procedural details or other items may inadvertently be inaccurate. Library of Congress Card No.: applied for

Prof. Alain Berthod

Uni. Claude-Bernard, Lyon 1 Bat. CPE-Lyon 308-D 69622 Villeurbanne Cedex France Prof. Verónica Pino Estévez

University of La Laguna C/Molinos de Agua 1 38207 San Cristobal la Laguna Spain Apryll M. Stalcup

Dublin City University Irish Separation Science Cluster Glasnevin 9 Dublin Ireland

British Library Cataloguing-in-Publication Data A catalogue record for this book is available from the British Library. Bibliographic information published by the Deutsche Nationalbibliothek The Deutsche Nationalbibliothek lists this publication in the Deutsche Nationalbibliografie; detailed bibliographic data are available on the Internet at http://dnb.d-nb.de.  2015 Wiley-VCH Verlag GmbH & Co. KGaA, Boschstr. 12, 69469 Weinheim, Germany All rights reserved (including those of translation into other languages). No part of this book may be reproduced in any form – by photoprinting, microfilm, or any other means – nor transmitted or translated into a machine language without written permission from the publishers. Registered names, trademarks, etc. used in this book, even when not specifically marked as such, are not to be considered unprotected by law. Print ISBN: 978-3-527-33374-5 Cover Design Formgeber, Mannheim, Germany Typesetting Thomson Digital, Noida, India Printed on acid-free paper

299

Part One Special Liquid Chromatography Modes

XII

Contents

Volume 2

Part One Special Liquid Chromatography Modes 299 1

Chiral Liquid Chromatography: Recent Applications with Special Emphasis on the Enantioseparation of Amino Compounds 301 István Ilisz

2

Chiral Separation of Some Classes of Pesticides by HPLC Method 321 Imran Ali, Iqbal Hussain, Mohd Marsin Sanagi, and Hassan Y. Aboul-Enein

3

Micellar Liquid Chromatography: Fundamentals 371 Maria C. García-Alvarez-Coque, Maria J. Ruiz-Angel, and Samuel Carda-Broch

4

Micellar Liquid Chromatography: Method Development and Applications 407 Maria C. García-Alvarez-Coque, Maria J. Ruiz-Angel, and Samuel Carda-Broch

5

Affinity Chromatography 461 Erika L. Pfaunmiller, Jesbaniris Bas, Marissa Brooks, Mitchell Milanuk, Elliott Rodriguez, John Vargas, Ryan Matsuda, and David S. Hage

6

Immunoaffinity Chromatography: Advantages and Limitations Nancy E. Thompson and Richard R. Burgess

Part Two

483

Capillary Electromigration Techniques 503

7

Capillary Electromigration Techniques: Capillary Electrophoresis Václav Kašička

8

Modern Injection Modes (Stacking) for CE Joselito P. Quirino

9

Capillary Gel Electrophoresis 555 Márta Kerékgyártó and András Guttman

10

Nonaqueous Capillary Electrophoresis Julie Schappler and Serge Rudaz

581

11

Detectors in Capillary Electrophoresis Petr Tůma and František Opekar

607

12

Trends in CE-MS and Applications Anna Tycova and Frantisek Foret

13

Capillary Electrochromatography Kai Zhang and Ruyu Gao

629 653

531

505

Contents

14

Micellar Electrokinetic Chromatography 675 Paolo Iadarola, Marco Fumagalli, and Simona Viglio

15

Chip-Based Capillary Electrophoresis 707 Yuanhong Xu, Jizhen Zhang and Jingquan Liu

16

Chiral Separations by Capillary Electrophoresis 731 E. Sánchez-López, M. Castro-Puyana, M.L. Marina, and A.L. Crego

Index to Volume 2 I1-I24

XIII

301

1 Chiral Liquid Chromatography: Recent Applications with Special Emphasis on the Enantioseparation of Amino Compounds István Ilisz

1.1 Introduction

The phenomenon of optical isomerism discovered by Pasteur has been known for many years. Nowadays, it is evident that the physical, biological, and chemical properties determined by the molecular symmetry and asymmetry play important role in nature. The physiological environment within a living organism is chiral, and the biological activities of enantiomeric forms of molecules can differ dramatically. With the exception of glycine, all of the 20 proteinogenic α-amino acids have a chiral carbon atom adjacent to the carboxyl group. This chiral center allows the existence of enantiomers, that is, two chemically identical molecular species differing only in optical activity (i.e., their ability to rotate the plane of polarized light). The stereoisomers (epimers) of the peptides in which all these amino acids are to be found may possess differences in biological activity in living systems. The past 20 years has seen an explosive growth in the field of chiral technologies, as illustrated by the rapid progress in the various facets of this intriguing field. The impetus for advances in chiral separation has been highest in the past decade and this still continues to be an area of high focus. Many chemical compounds such as drugs, fertilizers, and food additives have been commercialized as racemic mixtures, although in most cases only one of the isomers possesses desirable properties. As the understanding of the biological actions of compounds with respect to stereochemistry has grown, the necessity to investigate the pharmacological and toxicological properties has become more important. In light of the increased awareness concerning biologically important isomers, the US Food and Drug Administration has issued certain guidelines for the marketing of racemic compounds [1,2]. Chirality is now a major theme in the pharmaceutical industry in the design, discovery, and development, and launching and marketing of new drugs. Therefore, there is considerable pressure to develop analytical methods for enantiomeric purity control, pharmacological studies, pharmacodynamic investigations, clinical studies, and so on. The advances in stereoselective bioanalysis have led to a new awareness of the importance of Analytical Separation Science, First Edition. Edited by Jared L. Anderson, Alain Berthod, Verónica Pino Estévez, and Apryll M. Stalcup.  2015 Wiley-VCH Verlag GmbH & Co. KGaA. Published 2015 by Wiley-VCH Verlag GmbH & Co. KGaA.

302

1 Chiral Liquid Chromatography

stereoselective pharmacodynamics and pharmacokinetics, enabling differentiation of the relative contributions of enantiomers to the overall drug process. It is worth to mention that aspects of chirality are also very important in the environment, food and beverages, agrochemical, and petrochemical industries. Initially chiral analyses were relatively difficult and most separations were carried out with derivatization prior to analysis. The main methodologies used for enantioseparation are enzymatic digestion, crystallization, membrane-based spectroscopic, capillary electrophoretic, and chromatographic methods. At analytical level nowadays, the chromatographic techniques are the most popular ones. Various strategies have been devised for the differentiation of enantiomers by chromatography: (i) indirect methods involving sample derivatization by a chiral derivatizing agent (CDA) prior to injection into achiral chromatography columns, (ii) direct methods that use chiral mobile-phase additives with standard stationary phases, and (iii) direct methods where chiral separations are achieved by using a chiral stationary phase (CSP). High-performance liquid chromatography (HPLC) with CSPs seems one of the most promising tools for the chiral resolution of different racemates. The design and development of a CSP capable of effective chiral recognition of a wide range of enantiomers is the key point of the chiral HPLC technique. A number of CSPs for HPLC have been prepared, consisting of either small chiral molecules or chiral polymers as selectors, and many of them have been commercialized. This chapter focuses on the direct HPLC separations applying different types of CSPs; however, indirect methods are also briefly discussed. Even in this restricted field, the limitations do not permit the survey of the large number of studies published on the enantioseparation of various chiral molecules recently. Since two excellent review papers surveying the literature [3,4] and a book on chiral separations [5] have been published within a short period of time, in this chapter publications appeared in the past 2 years (2013–2014) are considered as recent applications with special emphasis on the enantioseparation of compounds containing amino groups.

1.2 Indirect Methods

The reaction of enantiomers with enantiomerically pure reagent-form diastereomeric derivatives that could be separated on achiral columns was the first widely used general method in the field of chiral analysis. The indirect methods are still quite efficient techniques for the enantioseparation of several chiral compounds, including amino acids. The advantages include the commercial availability of a large number of CDAs and a relatively wide choice of chromatographic conditions. The enantiomer molecule and the CDA must possess easily derivatizable and compatible functional groups. The reaction should be comparatively fast, as otherwise a difference in the rate of the formation of diastereomers may cause kinetic resolution. However, it is essential that the chiral derivatization reaction

1.3 Direct Methods

should proceed quantitatively for both enantiomers and that racemization should not occur. Resolution via diastereomer formation is usually improved when bulky (chromophore or fluorogenic) groups are attached to the chiral centers. Furthermore, if the chiral purity of the CDA is not known and/or is not taken into consideration, the chiral purity of the analyte will not be determined precisely. Thus, the indirect method is difficult to apply for the analysis of a standard sample or in pharmaceutical preparations, where a low amount of antipode (at a level of 0.1 or 0.05%) is to be determined. Furthermore, it is difficult to apply for preparative purposes. However, it is suitable for the trace analysis of enantiomers in complex matrices such as biological samples because of the introduction of a highly sensitive UV-visible or fluorescence tag. Following the introduction of new chromatographic and electrophoretic techniques, the importance of chiral derivatization has naturally decreased to some extent. However, this general method is still a method of choice that is widely used, especially in HPLC. The reasons include the large number of commercially available homochiral reagents, well-established reactions, and (if the R and S forms of the reagent are available) the peak of the enantiomeric impurity can be eluted before the main peak. Many biochemically important compounds have at least one functional group in their structure. The major types of reactions for chiral amines, involving amino acids, are mainly based on the formation of amides, carbamates, ureas, and thioureas. The reactions of acid chloride and chloroformate reagents proceed rapidly to furnish the corresponding amides and carbamates. The most important chloroformate possessing a reactive functional group is 1-(9fluorenyl)ethyl chloroformate. Chiral isocyanates and mainly isothiocyanates are good labels with which to produce stable ureas and thioureas. Among the chemically most selective CDAs are isothiocyanates such as 2,3,4,6-tetra-O-acetyl-β-Dglucopyranosyl isothiocyanate and 2,3,4,6-tetra-O-benzoyl-β-D-glucopyranosyl isothiocyanate. Marfey’s reagent, 1-fluoro-2,4-dinitrophenyl-5-L-Ala amide and its chiral variants, in which the L-Ala amide is replaced by some amino acid amide, are among the most important CDAs. Ortho-phthalaldehydes with chiral thiols give highly fluorescent adducts with the primary amino groups of amino acids, ensuring a very low detection limit in enantiomer analysis. CDAs developed earlier are still in use for different new applications. Further information on general chiral HPLC derivatization may be found in the original papers in numerous reviews and monographs [6–9].

1.3 Direct Methods

The first CSP for gas chromatography was described in 1966 [10]. Davankov and Rogozhin [11] introduced chiral ligand-exchange chromatography (CLEC) in 1971, and CLEC became the first direct HPLC method for the resolution of amino acids. The breakthrough in HPLC technology in the 1970s led to the

303

304

1 Chiral Liquid Chromatography

commercial availability of liquid chromatographs appropriate for fast, efficient, and reliable separations. The development of the hardware techniques was accompanied in the 1970s and 1980s by advances related to packing materials. Polymeric materials are not suitable for high-pressure operation, while the relative instability of the physical coating of the support materials proved to be one of the main drawbacks of the early CSPs. With increasing experience and knowledge and continuous development in the field, the disadvantageous properties were progressively overcome. The appearance of mechanically stable, porous, and small-diameter silica particles led to the commercialization of chemically bonded CSPs. The chiral separation techniques have subsequently become a very sophisticated field of analytical chemistry. The number of commercialized CSPs now exceeds 200, and studies focusing on the development of new types of chiral selectors are published virtually daily. Although a large number of CSPs are available nowadays, they tend to be derived from relatively few classes of compounds. Most chiral selectors are based on amino acids (native or derivatized), proteins, oligosaccharides (e.g., cyclodextrins, cyclofructanes, etc.), derivatized linear or branched polysaccharides (e.g., cellulose or amylose), and macrocyclic selectors, such as macrocyclic antibiotics, chiral crown ethers, ionexchangers, and synthetic polymers. Their advantages lie in the ability to operate in different modes of HPLC: in reversed phase (RP), normal-phase (NP), polar organic (PO), or polar ionic (PI) modes. 1.3.1 Ligand-Exchange-Based CSPs

Since Davankov and Rogozhin [11] introduced CLEC for the direct separation of amino acid enantiomers, metal chelate additives have been frequently used in the mobile phase. The mechanism of the chiral recognition process in CLEC is assumed to be based on the reversible formation of diastereomeric complexes between the stereoisomers of analytes, chiral selector, and a metal ion. On the basis of this concept, it can be noted that in case of CLEC, exclusively among the chromatographic techniques, the separation does not require the direct contact between the analyte and the chiral selector; the metal ion, acting as a Lewis acid, coordinates the selector and the analyte through dative bonds leading to the formation of a ternary complex. Chromatographic separation occurs if the complexes of enantiomers have different rates of formation and/or thermodynamic stabilities or different adsorption behaviors. Several chiral ligand-exchange stationary phases have been applied so far with different selectors, mostly chiral amino acids and chiral amino alcohols and their derivatives, for example, proline, leucinol, cysteine, phenylalanine, and penicillamine. Applying copper(II) as metal ion, the formation of fivemembered chelate ring is normally expected. Retention and separation are influenced by the concentration and nature of the mobile-phase components, together with other variables, such as pH and temperature. In all CLEC systems, the eluent pH and ionic strength, specifically the concentration of

1.3 Direct Methods

copper(II), markedly define the stability of the transient metal complexes and hence the chromatographic retention of analytes. New selectors for CLEC, O-benzyl-(S)-Ser, S-benzyl-(R)-Cys, S-diphenylmethyl(R)-Cys, and S-trityl-(R)-Cys were synthetized by Natalini et al. [12,13]. The new selectors have proved to be effective in the enantioseparation of some nonproteinogenic underivatized amino acids, such as phenylglycine, and amino acids containing thienyl-, cyclopentyl-, and tetrahydrofurenyl-rings with fair to good separation and resolution factors. 1.3.1.1

Recent Applications

To develop CLEC-based CSPs for the resolution of chiral drugs, the residual silanol groups on the silica surface of a CSP based on sodium N-[(S)-1-hydroxymethyl-3-methylbutyl]-N-undecylarninoacetate, a (S)-leucinol derivative, were protected with n-octyl groups [14]. Proton pump inhibitors (omeprazole, pantoprazole, lansoprazole, and rabeprazole) were enantioresolved with improved efficiency. The observed chiral recognition ability of the CSP was assumed to be originated from the protection of the nonenantioselective interaction sites on the silica surface and the improved lipophilicity of the stationary phase. Chiral ionic liquids (CILs) with amino acids as cations were applied as novel chiral ligands coordinated with copper(II) for the separation of tryptophan enantiomers [15]. The applicability of CILs with amino acids as cations for chiral separation was demonstrated and a comparative study was also conducted for exploring the mechanism of the CILs as new ligands in CLEC. Indirect chiral ligand-exchange chromatography (ICLEC) was facilitated by substituting chiral ligands from the bidentate Cu(II) complex via racemic analytes in reaction vial [16]. For the resulting Cu(II)–analyte–ligand complexes, higher stability constants and better chromatographic resolutions were found than in the conventional CLEC. The new technique was found to be successful in determining enantiomeric purity of drugs in pharmaceutical formulations. A new chromatographic procedure was developed for the separation of atenolol enantiomers based upon CLEC [17]. The separation was carried out on a C8 column with L-alanine and Cu(II) applied as a chiral selector and central bivalent complexing ion. The optimized HPLC method was utilized in some synthetic and human blood plasma samples. By applying cinchona alkaloids as chiral selectors along with Cu(II) ions, good enantioseparation was obtained for several amino acids using equimolar amounts of Cu(II) and cinchonidine, quinine, or quinidine [18]. The molecular geometry of the diastereomeric complexes formed was modeled and energetic differences between both compounds were calculated by methods based on semiempirical force field. CILs containing imidazolium cations and L-proline anions were applied as chiral selector to separate tryptophan enantiomers on a C18 column [19]. Factors influencing Trp enantiomer separation, such as alkyl chain length of CILs, concentrations of Cu(II) and CILs, pH of the mobile phase, flow rate, organic solvent, and temperature, were studied and some thermodynamic parameters were calculated. Chromatographic properties of isoxazoline-fused 2-aminocyclopentanecarboxylic acid analogs with a sodium

305

306

1 Chiral Liquid Chromatography

N-(R)-2-hydroxy-1-phenylethyl)-N-undecylaminoacetate–Cu(II) complex as a chiral selector were investigated under various conditions [20]. The effects of temperature were studied at constant mobile-phase compositions and thermodynamic parameters were calculated. Some mechanistic aspects of the chiral recognition process were discussed with respect to the structures of the analytes. 1.3.2 Protein-Based CSPs

The stereoselective interactions of proteins with chiral compounds can be a basis for their applications as efficient CSPs. Due to the complexity of the proteinbased selectors, numerous interactions might be used (e.g., hydrogen bonds, π–π, dipole–dipole, and ionic interactions) for achieving enantioselective recognition for a wide range of chiral compounds. However, even small changes under the applied conditions may cause drastic variations in their enantioseparation capacity [21]. The most important commercially available protein-based selectors are human serum albumin (HSA), α1-acid glycoprotein (AGP), ovomucoid (OVM), and cellobiohydrolase I (CBH). HSA is proved to be a useful CSP for acidic and CBH for basic chiral compounds, while AGP and OVM possess wide enantiorecognition ability for a variety of neutral, basic, and acidic pharmaceuticals [3]. Protein-based CSPs played an important role in the 1980s, but the decreased number of publications in the last couple of years indicates a significantly reduced interest. (It is important to note that the protein-based CSPs can serve as a model environment in drug–protein binding studies.) 1.3.2.1

Recent Applications

Chicken α1-acid glycoprotein was immobilized on aminopropyl silica particles and the effects of silica particle diameters on column performances were investigated [22]. Applying different chiral model compounds, the column filled with particles with lowest diameter gave the best performance. Retention of several structurally diverse drugs on α1-acid glycoprotein column was investigated under different chromatographic conditions [23]. AGP retention mechanism was found to be sensitive to changes produced by the percentage and nature of the organic modifier as a result of different shielding degree of AGP charged sites. 1.3.3 Polysaccharide-Based CSPs

The utilization of the ability of polysaccharides to resolve racemic mixtures dates back to 1951, when Kotake et al. reported the resolution of some amino acids by paper chromatography [24] using cellulose as CSP. Early works from Lüttringhaus, Hess, and Rosenbaum [25], Hesse and Hagel [26], and others reported the utilization of different cellulose derivatives without supporting beads. Derivatives

1.3 Direct Methods

of cellulose as beads can be applied for CSPs, offering theoretically higher loading capacity and efficiency, but higher mechanical stability and wider solvent compatibility require the application of support materials. Coating cellulose derivatives onto the surface of silica beads was reported by Okamoto Kawashima, and Hatada in 1984 [27]. Further developments in coating procedures and column technology have led to commercially available, mechanically stable, and state-of-the-art polysaccharide-based CSPs. Among the various polysaccharides, cellulose and amylose have been used for the preparation of commercialized CSPs. Because of their difficult handling and resolution problems, cellulose and amylose could not be utilized as efficient CSPs without modification; however, their ester and carbamate derivatives proved to be very suitable selectors for HPLC applications. Recent developments of the polysaccharide-based CSPs have been focused on the optimization of the substituents (position and quality) on the aromatic moiety of the derivatives. The polymeric chains are built from D-glucose units through ß-1,4 linkage in cellulose and α-1,4 linkage in amylose. In case of polysaccharide-based CSPs, the chiral recognition properties originate from several factors: (i) the presence of chiral centers in the glucose units (molecular chirality), (ii) conformational chirality due to the structure of the separate chains, and (iii) the supramolecular chirality that stems from the alignment of the polymeric chains forming highly ordered structures [28]. The enantiorecognition ability of the polysaccharide-based CSPs is generally assumed to be based on hydrogen bonding and dipole–dipole interactions. For these interactions to take place, the presence of water as a strongly competing species should be avoided. Thus, applying alkanes (e.g., hexane, heptane, etc.) and alcohol (e.g., propan-2-ol) in NP mode can be a good choice as a starting mobile phase for a polysaccharide-based CSP. Recent innovations have led to the availability of immobilized polysaccharidebased CSPs with extended solvent compatibility, and even nonstandard solvents (such as dichloromethane, chloroform, dioxane, toluene, etc.) can be applied. Polysaccharide CSPs recently commercialized can also be applied under RP conditions, but it is important to note that varying the chromatographic modes (column not dedicated to a specific mode) may result in extended equilibration times upon the change of mobile phases and reduction of the observed efficiency. 1.3.3.1

Recent Applications

The enantioseparations of 3,5-disubstituted hydantoins were examined under NP mode using Chiralpak IA containing an immobilized amylose tris-(3,5-dimethylphenylcarbamate) CSP [29]. The effects of polar alcoholic modifier and column temperature on retention and enantioseparation were determined, where both solvent- and temperature-induced reversal of elution order was observed. The enantioseparation of three pyroglutamide derivatives were investigated on amylose and cellulose tris-(3,5-dimethylphenylcarbamate), amylose tris-(S)α-methylbenzylcarbamate), and cellulose tris-(4-methylbenzoate) with various

307

308

1 Chiral Liquid Chromatography

mobile phases under both HPLC and SFC conditions [30]. Performances at semipreparative and analytical levels were compared and validation parameters were calculated. Development of chromatographic methods based on the use of an amylose derivative CSP was achieved to enable enantioresolution of four fully constrained ß-amino acids in NP mode [31]. The results obtained with the polysaccharide-based phase were compared with the results achieved with a glycopeptide-based column employed in PI mode. Chiralpak IC3 and Chiralcel OC columns were applied for the investigation of the enantiomeric composition of oxaliplatin [32]. Experimental results demonstrated the benefits arising from the enantioselective hydrophilic interaction liquid chromatography based on polysaccharide phases. A simple and sensitive HPLC method was developed and validated to determine the enantiomers of propranolol in rat serum [33]. Molecular modeling studies were performed, and binding affinities and interaction distances between propranolol enantiomers and chiral selector were calculated too. Enatioseparations of Koga bases were investigated applying polysaccharide-based CSPs under NP conditions [34]. As a result of temperature changes, alteration in CSP conformation and/or enantioseparation mechanism was observed. A silicabased cellulose 3,5-dimethylphenylcarbamate derivative hybrid material was developed and characterized as CSP [35]. Compared to a commercial Chiralpak IB column, better separation ability was observed for ß-blocker drugs as chiral probes. A systematic study of the effect of basic and acidic additives on HPLC separation of enantiomers of some basic chiral drugs on polysaccharide-based chiral columns in PO mode was conducted [36]. It was concluded that additives may serve, on the one hand, as useful tools for the adjustment of the enantiomer elution order and, on the other hand, they may help in better understanding the chiral recognition mechanisms. A comparative study was done for the evaluation of teicoplanin-based and polysaccharide-based columns applying unusual amino acids as chiral probes [37]. The applied phases were found to be complementary to each other to some extent, while regarding the CSP used the change of elution order of enantiomers was observed. 1.3.4 Cyclodextrin-Based CSPs

Cyclodextrins (CDs) are naturally occurring, nonreducing oligosaccharides. Three of the most widely utilized CDs are composed of 6, 7, and 8 D-glucopyranose units bonded through α-1,4 glycosidic linkages, called α-, β-, and γ-CD, respectively. CDs can be described as toroids with larger and smaller openings. The arrangement of the carbon backbone and the primary and the secondary hydroxyl groups leads to a lipophilic cavity and a hydrophilic surface. Thanks to this unique property, CDs are water soluble and are able to form inclusion complexes. The hydroxyl groups can be easily derivatized forming a large variety of modified CDs possessing uncharged or ionic substituents. Due to their relatively simple production, natural and modified CDs have found several applications.

1.3 Direct Methods

Since the introduction of the first high-coverage stable bonded phase CDs [38], they have been successfully applied as CSPs for the separation of various chiral compounds. The CDs are some of the more popular materials used as selector molecules in the field of chiral analysis. One of their advantages lies in their solvent compatibility, they have the ability to operate in different modes of HPLC: RP, NP, PO, and PI modes (they are multimodal CSPs). For the mechanism of the separation on CD selectors, it is generally accepted that under RP conditions, lipophilic solutes interact with CD selectors via inclusion complexation; the hydrophobic part of the guest molecule penetrates the CD cavity and leads to the release of solvent molecules. Van der Waals interactions inside the cavity and additional hydrophilic interactions may also take place. For aromatic CD derivatives π–π stacking increments with analytes containing aromatic moiety may exist as well. In the PO mode, hydrogen bonding, dipolar interactions, and steric repulsion can lead to chiral recognition. In the NP mode, enantioseparation can be achieved via a combination of hydrogen bonding, steric effects, π–π interactions (in the presence of aromatic moieties), and dipole–dipole stacking. 1.3.4.1

Recent Applications

A facile approach to construct a novel native CD CSP bearing cationic imidazolium on the linking bridge via thiol-ene reactions was reported [39]. The enhanced electrostatic interaction of the prepared cationic CSPs with oppositely charged chiral analytes were demonstrated applying dansyl amino acids. Two multiurea-bound ß-CD CSPs were prepared through the Staudinger reactions and the HPLC enantioseparation behaviors were investigated under multimodal elution [40]. The prepared phases exhibited good separation performances for the analytes investigated and also showed some complementary enantioselectivity to each other, due to different electron-donating/withdrawing groups in the phenylcarbamate moieties. Perphenylcarbamoylated CD-clicked CSP was prepared with improved column efficiency and CD surface loading [41]. The enantioselectivity of the prepared CSP was explored with 26 racemates and the characteristics of the column were evaluated in terms of linearity, limit of detection, and limit of quantification. 1.3.5 Cyclofructan-Based CSPs

The cyclofructans (CFs) consist of six or more (usually seven or eight) ß-(2-1)linked D-fructofuranose units. Common abbreviations for these compounds are CF6, CF7, CF8, and so on. Each of the fructofuranose units contains four chiral centers and three hydroxyl groups. These selectors are members of the macrocyclic oligosaccharides such as the cyclodextrins, which are the best known members of this class. CF6, with six D-fructofuranose units, contains 18-crown-6 ether core, similar to the respective crown ethers. The 18-crown-6 ring serves as the skeleton core of CF6, with six fructofuranose units attached on its rim,

309

310

1 Chiral Liquid Chromatography

where the fructofuranose units are alternatively pointing toward and away from the molecular center. The application of native and derivatized CF6 for the enantioresolution of primary amines was reported in 2009 by Armstrong and coworkers [42]. They found that isopropyl-carbamate-functionalized CF6 may serve as an efficient CSP to separate compounds containing primary amino groups (e.g., amino acids and their derivatives), while CF6 functionalized with aromatic moieties can be applied for the enantioseparation of a wide spectra of other chiral compounds. Such columns can be used in NP and PO modes. 1.3.5.1

Recent Applications

R-naphthylethyl-derivatized CF6 was used for direct enantioseparation of novel chiral analogs of spiroindoline phytoalexins under NP conditions [43]. It was found that analytes with electron-withdrawing substituents showed better chiral recognition ability on this CSP. Biaryl atropisomers were screened with CF6based CSPs in NP and PO mode to investigate the interactions governing retention and enantioselectivity [44]. Preparative-scale enantioseparations were also studied. Methods were developed for newly synthesized functional ethanoTröger base racemates applying CD and CF-based CSPs and capillary electrophoretic determinations [45]. CF-based CSPs were most successfully applied under NP conditions in HPLC. The chromatographic performance of a superficially porous CF6-based column was compared with that of the columns packed with 5 μm and 3 μm fully porous particles [46]. Under constant mobile-phase conditions, selectivity and resolution of the separations were found to be comparable between fully porous and superficially porous columns, even though the column filled with superficially porous particles contained lower absolute amounts of chiral selector. 1.3.6 Crown Ether-Based CSPs

Crown ethers with a cavity of specific size are macrocyclic polyethers, where the oxygen atoms can serve as electron donors. Due to this property, alkali metal ions and ammonium ion may be incorporated into their cavity. As a consequence, crown ethers can be used to resolve enantiomers that contain a primary amine functional group. The generated chiral ammonium ions can bind to the macrocyclic crown by inclusion complexation. Enantioselectivity is governed probably by steric factors of the substituents of the chiral ammonium ions and the residues attached to the chiral moieties incorporated into the crown ether. Sogah and Cram introduced the first crown ether-based CSP by immobilizing bis-(1,19-binaphthyl)-22-crown-6 on polystyrene or silica gel [47]. Two decades later Hyun, Jin, and Lee [48] and Machida et al. [49] prepared the first chemically bonded crown ether-based CSPs. Since then, several applications applying crown ether-based CSPs have been published, but because of their rather restricted applicability, these CSPs could not gain a really important position in the field of chiral HPLC analysis.

1.3 Direct Methods

1.3.6.1

Recent Applications

The enantioseparation ability of CSPs containing acridino-18-crown-6 ether selectors was studied by HPLC [50]. The newly prepared CSPs separated the enantiomers of selected protonated primary aralkylamines efficiently. Mexiletine and its analogs were resolved on three different crown ether-based CSPs [51]. The chromatographic behaviors of mexiletine analogs containing a substituted phenyl group at the chiral center were found to be quite dependent on the phenoxy group of analytes. A doubly tethered N-CH3 amide CSP based on (+)-(18crown-6)-2,3,11,12-tetracarboxylic acid was applied to the resolution of tocainide and its analogs [52]. The obtained results were compared with those on a single-tethered N-H amide CSP, a single-tethered N-CH3 amide CSP, and a double-tethered N-H amide CSP. Crown ether-based CSP was utilized for the enantioresolution of aminophosphonic acids and their aminocarboxylic acid analogs [53]. Computer modeling and H1-NMR analyses were performed to gain a better understanding of interactions of the analytes with the CSP. 1.3.7 Macrocyclic Glycopeptide-Based CSPs

The concept of using macrocyclic antibiotics as CSPs was initiated by Armstrong in 1994 [54]. Macrocyclic antibiotics possess several characteristics that allow them to interact with analytes and serve as chiral selectors. There are hundreds of these compounds and, unlike other classes of chiral selectors, they comprise a large variety of structural types. In general, these compounds have molecular masses greater than 600 but less than 2200. There are acidic, basic, and neutral types, and they may display little or no UV-Vis absorbance. The macrocyclic antibiotics used for chiral separations in HPLC include the ansamycins (rifamycins), the glycopeptides (avoparcin, teicoplanin, ristocetin A, vancomycin, and their analogs), and the polypeptide antibiotic thiostrepton. In the past two decades, they have had a rapid and significant impact on the field of enantioseparation. They have unique structural features and functionalities that allow various chiral interactive sites and interactions (i.e., electrostatic, hydrophobic, H-bonding, steric repulsion, dipole stacking, π–π interactions, etc.) between the analyte and the stationary phase. The macrocyclic glycopeptide-based Chirobiotic phases are the latest class of chiral selectors created by the bonding of macrocyclic antibiotics to silica. Because of their favorable properties, these CSPs are able to work in the NP mode with apolar mobile phases, in the RP mode with hydro-organic mobile phases, in the PO mode with 100% polar nonaqueous solvent, and in the PI mode with nonaqueous solvent containing some acid and/or base to adjust the selector (and analyte) ionization state. Most of the time, higher efficiencies are observed in the NP and PO mode; however, applying mass spectrometric (MS) detection PI or RP mode would be more advantageous to start with. The enantioselectivity was found to be different in each of these chromatographic modes. The popularity of Chirobiotic phases is being proven by several hundred papers

311

312

1 Chiral Liquid Chromatography

dealing with the separation of various amino acids, amines, amides, β-blockers, β-agonists, nonsteroidal anti-inflammatory drugs, antineoplastics, and various other biologically important compounds. One of the most characteristic features of the antibiotic-based selectors is their chiral ionic character. All Chirobiotic phases possess analogous ionizable groups that have been proven to play important role in the chiral recognition. Their specific structures and the variety of functionalities give them the power of efficient resolution of almost all types of neutral, acidic, and basic racemates. Before choosing an appropriate glycopeptide-based CSP, it is advisable to examine the structure of the analyte. In chiral chromatography, there are several types of interactions that can occur between solute and CSP. Naturally, not all of these are active or decisive in all mobile phases, so mobile-phase choice can be driven by the type of interaction that is available on a particular CSP or is available on a particular solute. Each Chirobiotic phase has a peptide backbone that provides hydrogen bonding and dipole–dipole interactions, and each, uniquely, has an ionic site of some type. The sugar moieties, if present, may provide hydrogen bonding and some steric effects. Last but not least, the glycopeptide has internal ring structures that will provide inclusion complexation in RP mode. Because of the structural similarities, the macrocyclic glycopeptides are to some extent complementary to one another: where partial enantioresolution is obtained with one glycopeptide, there is a high probability that baseline or better separation can be obtained with another. Each type of interaction has different strength in different mobile phases, so by going from one mobile-phase type to another, on the same column, the mechanism is changing, giving another opportunity for efficient separation. 1.3.7.1

Recent Applications

The chromatographic retention and thermodynamics of the adsorption of enantiomers of α-phenylcarboxylic acids on an eremomycin-based CSP under RP conditions were studied [55]. Relationships between the retention characteristics of the acids, the enantioselectivity, and the concentration of organic modifier were explored. The distribution of L- and D-amino acids in Antarctic lakes were monitored applying a teicoplanin aglycone-based CSP by HPLC–MS method [56]. A vancomycin-based CSP in PI mode was utilized to achieve the enantioseparation of chiral pharmaceuticals from activated sludge inoculum originated from a wastewater treatment plant [57]. The developed method was successfully validated and the enantioselectivity of the biodegradation process was characterized by its application. Macrocyclic glycopeptide-based CSPs were applied for the separation of four bicyclo[2.2.2]octane based 2-amino-3-carboxylic acid enantiomers [58]. The effects of the mobile-phase composition, the structure of the analytes, and the temperature on the separations were investigated. Enantiomeric resolution of atenolol was achieved applying a vancomycinbased CSP [59]. The proposed method was validated and utilized for the determination of atenolol from plasma samples.

1.3 Direct Methods

1.3.8 Ion-Exchanger-Based CSPs

In case of CSPs based on the ion-exchange procedure, retention relies on ionic interactions forming between ionic (or ionizable) solutes and the ionic functional group(s) of the selector. Lindner and coworkers have been particularly interested and active in the development of cinchona alkaloid-based CSPs. CSPs based on quinine, quinidine, epiquinine, and epiquinidine tert-butylcarbamate selectors were synthesized and evaluated under ion-exchange HPLC conditions with a set of racemic N-acetylated and N-oxycarbonylated proteinogenic and nonproteinogenic amino acids [60]. The enantioseparation of quinine- and quinidine-derived CSPs proved to be far superior to that of their C9-epimeric congeners. Cinchona alkaloid-based chiral weak anion-exchangers (WAX) and aminosulfonic acid-based chiral strong cation-exchangers (SCX) have been fused in a combinatorial synthesis approach into single, zwitterionic, and chiral selector motifs [61]. The corresponding zwitterionic ion-exchange-type CSPs allow enantioseparations of chiral acids and amine-type solutes in liquid chromatography using PO mode with largely rivaling separation factors compared to the parent WAX and SCX CSPs. Furthermore, the application spectrum could be remarkably expanded to various zwitterionic analytes such as α- and ß-amino acids and peptides. The new CSP provided strong unequivocal evidence for synergistic effects of the two oppositely charged ion-exchange subunits being involved in molecular recognition of zwitterionic analytes by zwitterionic selectors driven by double ionic coordination. Cinchona alkaloid-based zwitterionic CSPs were recently marketed under the trademark Chiralpak ZWIX(+)TM and ZWIX( )TM by Chiral Technologies Europe. Due to their zwitterionic nature, the chiral selectors in principle allow the enantiomeric separation of a remarkably broad scale of ionizable chiral analytes, ranging from acidic to basic and zwitterionic compounds. Anion-, cation-, and zwitterion-exchange processes have been confirmed to be predominantly involved in selector–selectand interactions and in the enantioselective retention of the analytes. Nonaqueous polar organic solvents in combination with acid and base modifiers (PI mode) proved to be the preferential mobile phase for the separation of zwitterionic solutes on the cinchona alkaloid-based zwitterionic CSPs. The use of MeOH as a protic solvent (which can suppress H-bonding interactions) and MeCN as an aprotic solvent to permit ionic interactions, but to interfere with aromatic π–π interactions, seemed to be the best due to the suppression of nonspecific hydrophobic interactions with the CSP, thereby enhancing enantioselectivity. 1.3.8.1

Recent Applications

A quinine- and a quinidine-based zwitterionic CSPs were applied for the enantioseparation of 27 unusual cyclic secondary α-amino acids [62]. The effects of the nature and concentration of the bulk solvent, the acid and base additives, the structures of the analytes, and the temperature on the enantioresolution

313

314

1 Chiral Liquid Chromatography

were investigated. A large series of diverse amino acids and analogs were applied for elucidation on the chromatographic conditions to be used for generic sample screening and for specific separations [63]. The zwitterionic CSPs were found to offer the features not only of a double ion-pairing with chiral ampholytes in retaining and resolving them but also of a single ion-pairing with acid- or basetype analytes. The enantioseparation of four bicyclo[2.2.2]octane-based 3-amino2-carboxylic acids was reported in PI mode with zwitterionic CSPs [64]. Experiments were performed at constant mobile-phase compositions in order to study the effects of temperature, and thermodynamic parameters were calculated. The effects of temperature on the chiral recognition of cyclic ß-amino acid enantiomers on zwitterionic CSPs were investigated [65]. Unusual temperature behavior was observed, especially on the ZWIX( )TM column, where the application of MeOH/MeCN (50/50 v/v) containing 25 mM triethylamine and 50 mM formic acid as mobile phase led to nonlinear van’t Hoff plots and increasing retention time with increasing temperature. Several free amino acids and analogs were directly resolved into enantiomers by HPLC on zwitterionic CSPs [66]. The systematic investigation was undertaken to gain an insight into the influence of the structural features on the enantiorecognition. Racemic aminophosphonic acids were used as chiral probes for the evaluation of the chiral recognition abilities of zwitterionic CSPs [67]. Mobile-phase characteristics such as acid-to-base ratio, type of counterion, and solvent composition were systematically varied in order to investigate their effect on the separation performance. The enantioseparation of 6-aminoquinolyl-N-hydroxysuccinimidyl carbamate (AQC)-tagged amino acids and other zwitterionic compounds was tested on tert-butyl-carbamate-modified quinine- and quinidine-based CSPs employing PO conditions [68]. The enantioselectivity of proteinogenic AQC-tagged amino acids was found to strongly depend on the chemistry and stereochemistry of the side chain of the amino acid. 1.3.9 Brush-Type (Donor–Acceptor Type, Pirkle-Type) CSPs

When individual synthetic chiral molecules with low molecular mass are attached to a solid support, a brush-type, donor–acceptor phase, or in recognition of his pioneering work [69], Pirkle-type CSP is obtained. In case of Pirkle-type CSPs, the selector molecules are covalently bonded to the chromatographic support, and regularly covering the surface of the inert matrix, they are easily accessible for the solute molecules. Hydrogen bonding, π–π interactions, and dipole–dipole stacking are usually the most important interactions responsible for chiral recognition. These donor–acceptor phases are usually run in NP mode, with mixtures of alkane and alcohol solvents applying basic (e.g., diethylamine) and/or acidic (e.g., trifluoroacetic acid) modifiers. CSPs based on this concept are in widespread use nowadays because of their advantageous characteristics (e.g., good kinetic performance and quite often thermal inertness). Besides these advantageous properties, it is important to note that these CSPs

1.3 Direct Methods

are usually available in both enantiomeric forms, namely, the elution sequence for two enantiomers can be changed by exchange of the column for the one packed with CSP having opposite configuration. Their advantageous characteristics make them a good choice for preparative applications also. 1.3.9.1

Recent Applications

1,2-Diaminocyclohexane and 1,2-diphenyl-1,2-ethylene-diamine were used as selectors for new brush-type CSPs containing a stable bidentate urea-type structure [70]. The totally synthetic CSPs were prepared and characterized in terms of retentivity, selectivity, and permeability, while their efficiency was determined through van Deemter plot analysis. The enantioselectivity of the new columns was tested for a wide range of racemates both in NP and in PO mode. 1.3.10 Molecularly Imprinted CSPs

One of the most attractive synthetic approaches to mimic nature is molecular imprinting. This is a concept of preparing substrate-selective recognition sites in a matrix, based on the application of a molecular template in a casting procedure. Molecularly imprinted polymers (MIPs) can be prepared conceptually applying three steps: (i) prearrangement of the functional monomers with the template molecule in solution phase, (ii) copolymerization of the prepolymer complex forming a rigid network, which holds the templates, and (iii) creating binding cavities in the polymer that are complementary in shape by removing the templates. After grinding and sieving the polymeric material, the obtained particles can be packed into columns and utilized for enantioselective separations. 1.3.11 Synthetic Polymer-Based CSPs

CSPs based on synthetic polymers can also be applied as chiral selectors, and they can mimic the enatiorecognition abilities of the semisynthetic polysaccahrides. These polymer phases are commonly prepared by two different approaches: (i) “grafting-to” approach, in which the chiral monomer is applied to synthesize the initiator polymeric chains in solution phase and then the polymeric chains thus formed are grafted to the surface-activated support by a copolymerization process; (ii) “grafting-from” approach, in which the initiator is immobilized on the surface of the support, thus the growth of the polymer chains starts from the surface, resulting in a well-ordered, surface-confined polymer layer [71]. The synthetic polymer-based CSPs are typically used under NP conditions, where the chiral recognition abilities are mainly due to the hydrogen bonding, π–π interactions, and steric factors.

315

316

1 Chiral Liquid Chromatography

1.4 Summary

Before the availability of stable and efficient CSPs, chiral analyses could be accomplished mainly through derivatization prior to analysis. The developments in column technologies led to the appearance and commercialization of chemically bonded CSPs. By the 1990s, the rapid resolution of optical isomers became routine and commonplace. The number of CSPs nowadays exceeds 200, and enantioseparations based on direct HPLC methods are the most popular ones. The continuous growth and evolution of the knowledge in the field of chiral analysis in the near future will probably allow the design of new chiral stationary phases offering more efficient separations in a shorter analysis time with higher robustness.

Acknowledgments

The author wishes to express his thanks to Prof. Antal Péter for all the useful comments and suggestions.

References 1 US Food and Drug Administration (1992)

2

3

4

5

6

FDA’s policy statement for the development of new stereoisomeric drugs. Chirality, 4 (5), 338–340. Challener, C.A. (2001) Chiral Drugs, 1st edn, John Wiley & Sons, Ltd, Aldershot, UK. Lämmerhofer, M. (2010) Chiral recognition by enantioselective liquid chromatography: mechanisms and chiral stationary phases. J. Chromatogr. A, 1217, 814–856. Cavazzini, A., Pasti, L., Massi, A., Marchetti, N., and Dondi, F. (2011) Recent applications in chiral high performance liquid chromatography: a review. Anal. Chim. Acta, 706, 205–222. Scriba, G.K.E. (ed.) (2013) Chiral Separations: Methods and Protocols, 2nd edn, Springer, New York. Toyo’oka, T. (2005) Fluorescent chiral derivatization reagents possessing benzofurazan structure for the resolution of optical isomers in HPLC: the synthesis,

7

8

9

10

11

characteristics and application. Curr. Pharm. Anal., 1 (1), 57–64. Bhushan, R. and Martens, J. (2010) Amino Acids: Chromatographic Separation and Enantioresolution, HNB Publishing, New York. Bhushan, R. and Brückner, H. (2011) Use of Marfey’s reagent and analogs for chiral amino acid analysis: assessment and applications to natural products and biological systems. J. Chromatogr. B, 879 (29), 3148–3161. Ilisz, I., Aranyi, A., Pataj, Z., and Péter, A. (2012) Recent advances in the direct and indirect liquid chromatographic enantioseparation of amino acids and related compounds: a review. J. Pharm. Biomed. Anal., 69 (SI), 28–41. Gil-Av, E., Feibush, B., and Charles-Sigler, R. (1966) Separation of enantiomers by gas liquid chromatography with an optically active stationary phase. Tetrahedron Lett., 7 (10), 1009–1015. Davankov, V.A. and Rogozhin, S.V. (1971) Ligand chromatography as a novel method

References

12

13

14

15

16

17

18

19

20

for the investigation of mixed complexes: stereoselective effects in amino acid copper (II) complexes. J. Chromatogr., 60 (2), 280–283. Natalini, B., Sardella, R., and Pellicciari, R. (2005) O-benzyl-(S)-serine: a new chiral selector for ligand-exchange chromatography of amino acids. Curr. Anal. Chem., 1 (1), 85–92. Natalini, B., Sardella, R., Macchiarulo, A., and Pellicciari, R. (2008) Cysteine-based chiral selectors for the ligand-exchange separation of amino acids. J. Chromatogr. B, 875 (1), 108–117. Ha, J.J., Han, H.J., Kim, H.E., Jin, J.S., Jeong, E.D., and Hyun, M.H. (2014) Development of an improved ligand exchange chiral stationary phase based on leucinol for the resolution of proton pump inhibitors. J. Pharm. Biomed. Anal., 100, 88–93. Qing, H.Q., Jiang, X.Y., and Yu, J.G. (2014) Separation of tryptophan enantiomers by ligand-exchange chromatography with novel chiral ionic liquids ligand. Chirality, 26 (3), 160–165. Kumar, R. and Bhushan, R. (2014) Indirect chiral ligand exchange chromatography for enantioseparation: a modification of conventional techniques. RSC Adv., 4 (91), 50130–50136. Alizadeh, T. (2014) Enantioseparation of atenolol using chiral ligand-exchange chromatography on C-8 column. (2013). Sep. Purif. Technol., 118, 879–887. Keunchkarian, S., Franca, C.A., Gagliardi, L.G., and Castells, C.B. (2013) Enantioseparation of α-amino acids by means of cinchona alkaloids as selectors in chiral ligand-exchange chromatography. J. Chromatogr. A, 1298, 103–108. Yang, Y.X., Li, J., and Jiang, X.Y. (2013) Chiral ionic liquids as chiral selectors for separation of tryptophan enantiomers by ligand exchange chromatography. J. Cent. South Univ., 20 (5), 1173–1177. Gecse, Z., Ilisz, I., Nonn, M., Grecsó, N., Fülöp, F., Agneeswari, R., Hyun, M.H., and Péter, A. (2013) Highperformance liquid chromatographic enantioseparation of isoxazoline-fused 2aminocyclopentanecarboxylic acids on a

21

22

23

24

25

26

27

28

29

chiral ligand-exchange stationary phase. J. Sep. Sci., 36, 1335–1342. Allenmark, S.G. and Andersson, S. (1994) Proteins and peptides as chiral selectors in liquid-chromatography. J. Chromatogr. A, 666 (1–2), 167–179. Matsunaga, H. and Haginaka, J. (2014) Separation of enantiomers on chiral stationary phase based on chicken α1-acid glycoprotein: effect of silica particle diameters on column performance. J. Chromatogr. A, 1363, 96–100. Chrysanthakopoulos, M., Vallianatou, T., Giaginis, C., and Tsantili-Kakoulidou, A. (2014) Investigation of the retention behavior of structurally diverse drugs on α1-acid glycoprotein column: insight on the molecular factors involved and correlation with protein binding data. Eur. J. Pharm. Sci., 60, 24–31. Kotake, M., Sakan, T., Nakamura, N., and Senoh, S. (1951) Resolution into optical isomers of some amino acids by paper chromatography. J. Am. Chem. Soc., 73 (6), 2973–2974. Lüttringhaus, A., Hess, U., and Rosenbaum, H.J. (1967) Conformational enantiomerism. I. Optically active 4,5,6,7dibenzo-1,2-dithiacy-clooctadiene. Z. Naturforschung B, 22, 1296–1300. Hesse, G. and Hagel, R. (1973) A complete separation of a racemic mixture by elution chromatography on cellulose triacetate. Chromatographia, 6, 277–280. Okamoto, Y., Kawashima, M., and Hatada, K. (1984) Useful chiral packing materials for high-performance liquid chromatographic resolution of enantiomers: phenylcarbamates of polysaccharide coated on silicagel. J. Am. Chem. Soc., 106, 5357–5359. Francotte, E., Wolf, R.M., Lohmann, D., and Müller, R. (1985) Chromatographic resolution of racemates on chiral stationary phases 1. Influence of the supramolecular structure of cellulose triacetate. J. Chromatogr., 347, 25–37. Yang, X., Su, L., Hou, X., Ding, S., Xu, W., Wang, B., and Fang, H. (2014) Highperformance liquid chromatographic enantioseparation of 3, 5-disubstituted hydantoins analogs and temperatureinduced reversals of elution orders on a

317

318

1 Chiral Liquid Chromatography

30

31

32

33

34

35

36

polysaccharide-based chiral stationary phase. J. Chromatogr. A, 1355, 291–295. Baudelet, D., Schifano-Fauxa, N., Ghinet, A., Dezitterd, X., Barbotina, F., Gautret, P., Rigo, B., Chavatte, P., Milleta, R., Furmand, C., Vacchera, C., and Lipka, E. (2014) Enantioseparation of pyroglutamide derivatives on polysaccharide-based chiral stationary phases by high-performance liquid chromatography and supercritical fluid chromatography: a comparative study. J. Chromatogr. A, 1363, 257–269. Sardella, R., Ianni, F., Lisanti, A., Scorzoni, S., Marini, F., Sternativo, S., and Natalini, B. (2014) Direct chromatographic enantioresolution of fully constrained ßamino acids: exploring the use of highmolecular weight chiral selectors. Amino Acids, 46 (5), 1235–1242. Gallinella, B., Bucciarelli, L., Zanitti, L., Ferretti, R., and Cirilli, R. (2014) Direct separation of the enantiomers of oxaliplatin on a cellulose-based chiral stationary phase in hydrophilic interaction liquid chromatography mode. J. Chromatogr. A, 1339, 210–213. Alanazi, A.M., Hefnawy, M.M., AL-Majed, A.A., AL-Suwailem, A.K., Kassem, M.G., Mostafa, G.A., Attia, S.M., and Khedr, M.M. (2014) HPLC-fluorescence method for the enantioselective analysis of propranolol in rat serum using immobilized polysaccharide-based chiral stationary phase. Chirality, 26 (4), 194–199. Guo, H.X., Wu, S., and Sun, J.S. (2014) Influence of temperature on the enantioselectivity of Koga tetraamines on amylose chiral stationary phases. Molecules, 19 (1), 9–21. Weng, X.L., Bao, Z.B., Xing, H.B., Zhang, Z.G., Yang, Q.W., Su, B.G., Yang, Y.W., and Ren, Q.L. (2013) Synthesis and characterization of cellulose 3, 5dimethylphenylcarbamate silica hybrid spheres for enantioseparation of chiral ßblockers. J. Chromatogr. A, 1321, 38–47. Mosiashvili, L., Chankvetadze, L., Farkas, T., and Chankvetadze, B. (2013) On the effect of basic and acidic additives on the separation of the enantiomers of some basic drugs with polysaccharide-based chiral selectors and polar organic mobile

37

38

39

40

41

42

43

44

45

phases. J. Chromatogr. A, 1317 (SI), 167–174. Kucerova, G., Vozka, J., Kalikova, K., Geryk, R., Plecita, D., Pajpanova, T., and Tesarova, E. (2013) Enantioselective separation of unusual amino acids by high performance liquid chromatography. Sep. Purif. Technol., 119, 123–128. Armstrong, D.W. and DeMond, W. (1984) Cyclodextrin bonded phases for the liquidchromatographic separation of optical, geometrical, and structural isomers. J. Chromatogr. Sci., 22 (9), 411–415. Yao, X., Tan, T.T.Y., and Wand, Y. (2014) Thiol-ene click chemistry derived cationic cyclodextrin chiral stationary phase and its enhanced separation performance in liquid chromatography. J. Chromatogr. A, 1326, 80–88. Lin, C., Fan, J., Liu, W., Tan, Y., and Zhang, W. (2014) Comparative HPLC enantioseparation on substituted phenylcarbamoylated cyclodextrin chiral stationary phases and mobile phase effects. J. Chromatogr. A, 98, 221–227. Pang, L., Zhou, J., Tang, J., Ng, S.C., and Tang, W. (2014) Evaluation of perphenylcarbamated cyclodextrin clicked chiral stationary phase for enantioseparations in reversed phase high performance liquid chromatography. J. Chromatogr. A, 1363, 119–127. Sun, P., Wang, C., Breitbach, Z.S., Zhang, Y., and Armstrong, D.W. (2009) Development of new HPLC chiral stationary phases based on native and derivatized cyclofructans. Anal. Chem., 81, 10215–10226. Gondová, T., Petrovaj, J., Kutschy, P., and Armstrong, D.W. (2013) Stereoselective separation of spiroindoline phytoalexins on R-naphthylethyl cyclofructan 6-based chiral stationary phase. J. Chromatogr. A, 1272, 100–105. Woods, R.M., Patel, D.C., Lim, Y., Breitbach, Z.S., Gao, H., Keene, C., Li, G., Kürti, L., and Armstrong, D.W. (2014) Enantiomeric separation of biaryl atropisomers using cyclofructan-based chiral stationary phases. J. Chromatogr. A, 1357, 172–181. Weatherly, C.A., Na, Y.C., Nanayakkara, Y.S., Woods, R.M., Sharma, A., Lacour, J.,

References

46

47

48

49

50

51

52

and Armstrong, D.W. (2014) Enantiomeric separation of functionalized ethanobridged Tröger bases using macrocyclic cyclofructan and cyclodextrin chiral selectors in high-performance liquid chromatography and capillary electrophoresis with application of principal component analysis. J. Chromatogr. B, 955–956, 72–80. Spudeit, D.A., Dolzan, M.D., Breitbach, Z.S., Barber, W.E., Micke, G.A., and Armstrong, D.W. (2014) Superficially porous particles vs. fully porous particles for bonded high performance liquid chromatographic chiral stationary phases: isopropyl cyclofructan 6. J. Chromatogr. A, 1363, 89–95. Sogah, G.D.Y. and Cram, D.J. (1979) Host–guest complexation 14. Host covalently bound to polystyrene resin for chromatographic resolution of enantiomers of amino-acid and ester salts. J. Am. Chem. Soc., 101 (11), 3035–3042. Hyun, M.H., Jin, J.S., and Lee, W.A. (1998) A new HPLC chiral stationary phase for the direct resolution of racemic quinolone antibacterials containing a primary amino group. Bull. Korean Chem. Soc., 19 (8), 819–821. Machida, Y., Nishi, H., Nakamura, K., Nakai, H., and Sato, T. (1998) Enantiomer separation of amino compounds by a novel chiral stationary phase derived from crown ether. J. Chromatogr. A, 805 (1–2), 85–92. Nemeth, T., Levai, S., Kormos, A., Kupai, J., Toth, T., Balogh, G.T., and Huszthy, P. (2014) Preparation and studies of chiral stationary phases containing enantiopure acridino-18-crown-6 ether selectors. Chirality, 26 (10), 651–654. Jin, K.B., Kim, H.E., and Hyun, M.H. (2014) Liquid chromatographic resolution of mexiletine and its analogs on crown ether-based chiral stationary phases. Chirality, 26 (5), 272–278. Lee, K.J., Tak, K.M., and Hyun, M.H. (2013) Resolution of tocainide and its analogues on a doubly tethered N-CH3 amide chiral stationary phase based on (+)-(18-crown-6)-2,3,11, 12tetracarboxylic acid. Bull. Korean Chem. Soc., 34 (10), 2978–2982.

53 Barnhart, W.W., Xia, X.Y., Jensen, R., and

54

55

56

57

58

59

60

Gahm, K.H. (2013) Comparison of chiral separations of aminophosphonic acids and their aminocarboxylic acid analogs using a crown ether column. Chirality, 25 (7), 369–378. Armstrong, D.W., Tang, Y., Chen, S., Zhou, Y., Bagwill, C., and Chen, J.R. (1994) Macrocyclic antibiotics as a new class of chiral selectors for liquid chromatography. Anal. Chem., 66, 1473–1484. Blinov, A.S. and Reshetova, E.N. (2014) Effect of the concentration of organic modifier in an aqueous-ethanol mobile phase on the chromatographic retention and thermodynamic characteristics of the adsorption of enantiomers of α1-phenylcarboxylic acids on silica gel with immobilized eremomycin antibiotic. Russ J. Phys. Ch. A, 88 (10), 1778–1784. Barbaro, E., Zangrando, R., Vecchiato, M., Turetta, C., Barbante, C., and Gambaro, A. (2014) D- and L-amino acids in Antarctic lakes: assessment of a very sensitive HPLC–MS method. Anal. Bioanal. Chem., 406, 5259–5270. Ribeiro, A.R., Afonso, C.M., Castro, P.M.L., and Tiritan, M.E. (2013) Enantioselective HPLC analysis and biodegradation of atenolol, metoprolol and fluoxetine. Environ. Chem. Lett., 11, 83–90. Pataj, Z., Ilisz, I., Grecsó, N., Palkó, M., Fülöp, F., Armstrong, D.W., and Péter, A. (2014) Enantiomeric separation of bicyclo [2.2.2]octane-based 2-amino-3-carboxylic acids on macrocyclic glycopeptide chiral stationary phases. Chirality, 26 (4), 200–208. Hefnawy, M.M., Al-Shehri, M.M., Abounassif, M.A., and Mostafa, G.A.E. (2013) Enantioselective quantification of atenolol in mouse plasma by high performance liquid chromatography using a chiral stationary phase: application to a pharmacokinetic study. J. AOAC Int., 96 (5), 976–980. Maier, N.M., Nicoletti, L., Lämmerhofer, M., and Lindner, W. (1999) Enantioselective anion exchangers based on cinchona alkaloid-derived carbamates:

319

320

1 Chiral Liquid Chromatography

61

62

63

64

65

influence of C-8/C-9 stereochemistry on chiral recognition. Chirality, 11 (7), 522–528. Hoffmann, C.V., Pell, R., Lämmerhofer, M., and Lindner, W. (2008) Synergistic effects on enantioselectivity of zwitterionic chiral stationary phases for separations of chiral acids, bases, and amino acids by HPLC. Anal. Chem., 80 (22), 8780–8789. Ilisz, I., Gecse, Z., Pataj, Z., Fülöp, F., Tóth, G., Lindner, W., and Péter, A. (2014) Direct high-performance liquid chromatographic enantioseparation of secondary amino acids on cinchona alkaloid-based chiral zwitterionic stationary phases. Unusual temperature behavior. J. Chromatogr. A, 1363, 169–177. Zhang, T., Holder, E., Franco, P., and Lindner, W. (2014) Chiral method development and optimization on cinchona and chiral sulfonic acid-based zwitterionic stationary phases for enantiomerseparations of free amino acids by high-performance liquid chromatography. J. Chromatogr. A, 1363, 191–199. Ilisz, I., Grecsó, N., Palkó, M., Fülöp, F., Lindner, W., and Péter, A. (2014) Structural and temperature effects on enantiomer separations of bicyclo[2.2.2] octane-based 3-amino-2-carboxylic acids on cinchona alkaloid-based zwitterionic chiral stationary phases. J. Pharm. Biomed. Anal., 98, 130–139. Ilisz, I., Pataj, Z., Gecse, Z., Szakonyi, Z.M., Fülöp, F., Lindner, W., and Péter, A. (2014) Unusual temperature-induced retention behavior of constrained ß-amino acid enantiomers on the zwitterionic chiral stationary phases ZWIX(+) and ZWIX( ). Chirality, 26 (8), 385–393.

66 Zhang, T., Holder, E., Franco, P., and

67

68

69

70

71

Lindner, W. (2014) Zwitterionic chiral stationary phases based on cinchona and chiral sulfonic acids for the direct stereoselective separation of amino acids and other amphoteric compounds. J. Sep. Sci., 37 (11), 1237–1247. Gargano, A.F.G., Kohout, M., Macikova, P., Lämmerhofer, M., and Lindner, W. (2013) Direct high-performance liquid chromatographic enantioseparation of free alpha-, beta- and gammaaminophosphonic acids employing cinchona-based chiral zwitterionic ion exchangers. Anal. Bioanal. Chem., 405 (25), 8027–8038. Hellinger, R., Horak, R., and Lindner, W. (2013) Enantioseparation of 6aminoquinolyl-N-hydroxysuccinimidyl carbamate tagged amino acids and other zwitterionic compounds on cinchonabased chiral stationary phases. Anal. Bioanal. Chem., 405, 8105–8120. Pirkle, W.H. and House, D.W. (1979) Chiral high-pressure liquidchromatographic stationary phases 1. Separation of the enantiomers of sulfoxides, amines, amino-acids, alcohols, hydroxy-acids, lactones, and mercaptans. J. Org. Chem., 44 (12), 1957–1960. Kotoni, D., Villani, C., Bell, D.S., Capitani, D., Campiglia, P., and Gasparrini, F. (2013) Bidentate urea-based chiral selectors for enantioselective high performance liquid chromatography: synthesis and evaluation of “crab-like” stationary phases. J. Chromatogr. A, 1297, 157–167. Gasparrini, F., Misiti, D., Rompietti, R., and Villani, C. (2005) New hybrid polymeric liquid chromatography chiral stationary phase prepared by surfaceinitiated polymerization. J. Chromatogr. A, 1064, 25–38.

321

2 Chiral Separation of Some Classes of Pesticides by HPLC Method Imran Ali, Iqbal Hussain, Mohd Marsin Sanagi, and Hassan Y. Aboul-Enein

2.1 Introduction

Pesticides are one of the most notorious organic pollutants in our environment. According to pests, they are classified as insecticide, fungicide, herbicide, algaecide, avicide, bactericide, miticide, molluscicide, nematicide, piscicide, and rodenticide. The three major classes of pesticides are insecticide, herbicide, and fungicide, while the rest are classified as miscellaneous pesticides. About 1693 pesticides are available, and the majority of pesticides are organic chemicals, while some are inorganic or biological species [1]. These pesticides control the insect by killing it or by changing its behavior, and are delivered through systems such as spraying, baits, slow-release diffusion, and so on. Insecticides are classified as organochlorine compounds (biphenyl aliphatic, hexachlorocyclohexane, cyclodienes, and polychloroterpenes), organophosphate compounds – esters of phosphorus (phosphates, phosphonates, phosphorothioates, phosphorodithioates, and phosphoramidates), organosulfur compounds that contain two phenyl rings with sulfur atoms, carbamates, formamidines, dinitrophenols, organotins, pyrethroids (first-, second-, third-, and fourth-generation pyrethroids), nicotinoids, spynosins, fiproles or phenyl-pyrazoles, pyrroles, pyrazoles, pyridazinones, quinazolines, benzoyl urea, botanicals (pyrethrum, nicotine, rotenone, limonene, or d-limonene, and neem), synergists or activators, antibiotics, fumigants, insect repellents, inorganics, miscellaneous classes of insecticides (methoxyacrylates, naphthoquinones, nereistoxin analogs, pyridine azomethine, pyrimidinamines, and tetronic acids), and miscellaneous compounds of insecticides: etoxazole, pyridalyl, amidoflumet, pyriproxyfen, buprofezin, and tebufenozide [2]. Fungicides are useful in controlling fungal diseases by specifically inhibiting or killing the fungus that causes the disease. Fungicides are classified on the basis of their chemical compositions that include benzimidazole, dicarboximide, imidazole, piperazine, triazole, phenylamide, oxathiin, anilinopyrimidine, strobilurin, phenylpyrrole, chloronitrobenzene, thiadiazole, cinnamic acid, hydroxyanilide, streptomyces, polyoxin, benzothiadiazole, phosphonate, dithiocarbamate, chloroalkythiols, chloronitrile, phenylpyridinamine, cyanoacetamide oxime, carbamate, Analytical Separation Science, First Edition. Edited by Jared L. Anderson, Alain Berthod, Verónica Pino Estévez, and Apryll M. Stalcup.  2015 Wiley-VCH Verlag GmbH & Co. KGaA. Published 2015 by Wiley-VCH Verlag GmbH & Co. KGaA.

322

2 Chiral Separation of Some Classes of Pesticides by HPLC Method

aldehyde, mineral oils, and some inorganics [3]. Herbicides (weed killers) are useful for weed control and are classified as phenoxy compounds, phenylacetic acid, benzoic acid, phthalic acid, N-1-napthylphthalamic acid, aliphatic acid, substituted phenols, heterocyclic nitrogen derivatives, aliphatic organic nitrogen derivatives, carbamate, metal organic and inorganic salts, and hydrocarbons or oils [4]. Among these pesticides about 28% (482) are chiral in nature, of which 149 are insecticides, 141 are herbicides, 97 are fungicides, and 95 are miscellaneous chiral pesticides [1]. These pesticides are useful for food production since they reduce the rate of diseases in crops. However, pesticides are also harmful both to the environment and to humans. Therefore, it is desirable to reduce the consumption of pesticides to a minimum. Only one enantiomer is biologically active toward the target organisms, while the other enantiomer has less impact (less active), but it may cause adverse effects on some nontarget organisms and, thus, it becomes an unwanted burden for the environment and becomes a potential health hazard. Accordingly, we can reduce the use of pesticides on the basis of chirality [5,6]. Moreover, the enantioselective behavior of chiral pesticides and physiological changes of plants may alter the food chain and become beneficial to the ecological system. Moreover, degradation products of achiral pesticides may be chiral and toxic. The enantiomers can exhibit significant differences both in their biological activity and toxicity and in their environmental behavior [7–14]. Therefore, it is important to investigate the actual impact of chiral pesticides on the environment. In spite of this, a majority of chiral pesticides are being marketed and released into the environment as racemates. Consequently, there is an urgent need to develop analytical methods to determine the optical purity and the enantiomeric resolution of chiral pesticides. Several analytical methods have been used to assess the enantiomeric purity of different classes of pesticides. High-performance liquid chromatography (HPLC) has achieved a good reputation for chiral analysis of pesticides due to the availability of several chiral stationary phases (CSPs), its high speed, sensitivity, and reproducibility. Although the CSPs are generally expensive, they are reliable and provide robust analytical results. A variety of mobile phases, including normal, reversed, and new polar organic phases, are used in HPLC. About 80% of the enantiomeric separations of drugs, pharmaceuticals, and pesticides have been carried out by HPLC [15–22]. As a result, a wide range of applications of HPLC in the chiral resolution of pesticides are presented in this chapter using normal and reversed chiral stationary phases. The mechanism of chiral separation of pesticides using various chiral stationary phases will be discussed.

2.2 Mechanism of Chiral Separation

Several chiral stationary phases (CSPs) have been used for the enantiomeric separation of pesticides that include polysaccharides, cyclodextrins, macrocyclic

2.2 Mechanism of Chiral Separation

glycopeptide antibiotics, proteins, crown ethers, ligand exchangers, Pirkle’s types, and several others [15,23]. Yet, the chiral recognition mechanism(s) at the molecular level on these CSPs is still not fully investigated. Pfeiffer [24] in 1956 explained the three-point model for chiral recognition mechanisms. Furthermore, this model was explained in more detail by Pirkle and Pochapsky [25]. According to the model in question, a minimum of three interactions between the CSP and at least one of the enantiomers play an important role in chiral separation, and at least one of these interactions is stereochemically dependent, that is, one interaction involves the stereogenic center with the CSP. This threepoint model is not applicable for every chiral species. Groombridge et al. [26] postulated a fourpoint model for chiral recognition on some protein CSPs that depends on the formation of transient diastereoisomeric complexes between the enantiomers and the CSP that possess different physical and chemical properties and, as a result, the said enantiomers are resolved at different retention times. Every chiral stationary phase has a different chiral recognition mechanism. However, chiral recognition processes on polysaccharides, cyclodextrins (CDs), macrocyclic glycopeptide antibiotics, proteins, and chiral crown ether-based CSPs are more or less similar. The chiral grooves on polysaccharides, the cavities on CDs, the baskets on macrocyclic glycopeptide antibiotics, the bridges and the loops on proteins, and the cavities on chiral crown ether-based CSPs all provide a chiral environment for enantiomers. Two enantiomers get fitted to different extents and, hence, elute at different retention times. The differences in the stabilities of the enantiomers on these CSPs are due to their different bondings and interactions, the most important of which are hydrogen bonding, dipole-induced dipole interactions, π–π complexation, inclusion complexation, anionic and cationic bonding, and van der Waals forces [15–21]. Steric effects also play a crucial role in the chiral resolution of racemates. Different binding energies of the transient diastereoisomeric complexes result from the various interactions mentioned above. Pirkle-type CSPs contain a chiral aromatic ring; therefore, the formation of a π–π charge transfer diastereoisomeric complexes of the enantiomers (with the aromatic group) with a CSP is considered to be essential. In view of these facts, the π acidic CSPs are suitable for chiral resolution of π donor solutes, and vice versa. However, the newly developed CSPs that contain both π acidic and π basic groups are suitable for the chiral resolution of both types of solutes, that is, π donor and π acceptor analytes. In brief, Pirkle-type CSPs contain a chiral moiety that provides a chiral environment for the enantiomers. Therefore, the enantiomers fit differently to this chiral moiety (due to their different spatial configurations). Accordingly, the two enantiomers form diastereoisomeric complexes that have different physical and chemical properties, along with different binding energies. Therefore, the two enantiomers elute at different retention times with the flow rate of the mobile phase and, hence, chiral separation occurs. On ligand-exchange CSPs, chiral resolution occurs due to the exchange of chiral ligands and enantiomers on specific metal ions through coordinate bonds. The two enantiomers have different exchange capacities because of the

323

324

2 Chiral Separation of Some Classes of Pesticides by HPLC Method

stereospecific nature of the ligand-exchange process and, hence, chiral resolution takes place. Rogozhin and Davankov [27–29] suggested a theoretical model for the mechanism of chiral resolution on these CSPs. In this model, the enantiomers are coordinated to metal ion in different ways depending on their interactions with ligands bonded to the stationary phases that act as the chiral selectors. The authors explained that chiral resolution occurs because of the different bonding along with the steric effects that result in the formation of the transient diastereoisomeric complexes by the two enantiomers. These transient diastereoisomeric complexes are stabilized at different magnitudes by dipole– dipole interactions, hydrogen bonding, van der Waals forces, and steric effects, and elute at different retention times. A general graphical representation of the chiral resolution mechanism of racemates on the aforementioned CSPs is shown in Figure 2.1.

Enantiomers

Steric effect site π-π interaction site

Hydrogen-bonding site

Dipole-induced dipole interaction site

Chiral grooves

CSP

Less binding

More binding

CSP SP

Less stable enantiomer

More stable enantiomer

Chromatogram of separated enantiomers Figure 2.1 Schematic diagram of chiral resolution mechanism.

2.3 Columns and Eluents

2.3 Columns and Eluents

A variety of chiral stationary phases have been developed for the resolution of chiral pesticides. The most important classes of chiral selectors have been explained above. These chiral selectors are available in the form of columns and capillaries, and are marketed under different trade names as shown in Table 2.1. The development of the chiral stationary phases enhanced the utility of highperformance liquid chromatographic technique. These CSPs have been used frequently and successfully in the chiral resolution of many drugs, pharmaceuticals, and other environmental pollutants. Therefore, they may also be used for the enantiomeric separation of chiral pesticides and some reports on the chiral separation of pesticides using the above-mentioned CSPs are to be found in the literature. In a significant proportion of resolutions of chiral pesticides, using these chiral stationary phases, a wide range of mobile phases have been used as eluents. Different compositions of eluents for chiral separation depend on the chiral stationary phase. For the normal-phase mode, various compositions of organic modifiers can be employed, whereas for the reversed phase, polar organic solvents with water and buffer are used. The different optimizing conditions for the chiral resolution include the composition of a mobile phase, its pH, the temperature, the amount injected into the column, the flow rate, and the detection mode. In the case of polysaccharide chiral columns, for normal-phase chromatography, pure ethanol or 2-propanol as an eluent is recommended. In order to decrease the polarity of the mobile phase and increase the retention times of the enantiomers, hexane, cyclohexane, pentane, and heptane are used as the main constituents of the mobile phase. However, other alcohols are also used in the mobile phase. Normally, if pure ethanol or 2-propanol is not suitable for the mobile phase, hexane, 2-propanol, or ethanol with a ratio of 80:20 v/v is used in the mobile phase and the change in the mobile-phase composition is introduced on the basis of observations. Finally, the optimization of the chiral resolution is carried out by adding small amounts of amines or acids (0.1–1.0%). Similarly, the chiral resolution on polysaccharide-based CSPs in the reversed phase mode is carried out using the aqueous mobile phases, and the selection of a mobile phase depends on the solubility and the properties of pesticides to be analyzed. The choice of a mobile phase in the reversed phase mode is very limited. Water is used as the main constituent of the mobile phases. The modifiers used are acetonitrile, methanol, and ethanol. The optimization of the chiral resolution is carried out by adding small quantities of amines or acids (0.1–1.0%). Some of the resolutions are pH dependent and require a constant pH of a mobile phase. Under such conditions, in general, the resolution is not reproducible using mobile phases such as water–acetonitrile or water–methanol and, therefore, a buffer with some organic modifiers (acetonitrile, methanol) have been used as the mobile phase. The optimization of the resolution is carried out by adjusting

325

326

2 Chiral Separation of Some Classes of Pesticides by HPLC Method

Table 2.1 Various normal- and reversed phase chiral columns and their trade names. Columns

Trade name

Polysaccharide-based CSPs Chiralcel OB, Chiralcel OB-H, Chiralcel OJ, Chiralcel OJ-R, Chiralcel CMB, Chiralcel OC, Chiralcel OD, Chiralcel OD-H, Chiralcel OD-R, Chiralcel OF, Chiralcel OD-RH, Chiralcel OG, Chiralcel OA, Chiralcel CTA, Chiralcel OK, Chiralpak AD, Chiralpak AD-R, Chiralpak AR Chiralpak AD-RH, and Chiralpak AS Cyclodextrin-based CSPs

Daicel Chemical Industries, Tokyo, Japan

Cyclobond I, II, and III; Cyclobond AC, RN, SN; ApHpera ACD and BCD Nucleodex β-OH, Nucleodex β-PM ORpak CD-HQ, and Orpak CDB-453 HQ, ORpak CDBS-453, ORpak CDA-453 HQ, and ORpak CDC-453 HQ

Advanced Separation Technologies, Whippany, NJ, USA. Macherey-Nagel, Duren, Germany. Showa Denko, Kanagawa, Japan

Keystone β-OH and Keystone β-PM

Thermo Hypersil, Bellefonte, PA, USA

Β-Cyclose-6-OH

Chiral Separations, La Frenaye, France

YMC Chiral cyclodextrin BR, YMC Chiral NEA (R), and YMC Chiral NEA (S) Macrocyclicglycopeptide antibiotic-based CSPs

YMC, Kyoto, Japan

Chirobiotic R, Chirobiotic T, Chirobiotic V, Chirobiotic TAG, and Chirobioticmodified V

Advanced Separation Technologies, Inc., Whippany, NJ, USA

Protein-based CSPs Chiral AGP, Chiral HSA, and Chiral CBH

Advanced Separation Technologies, Inc., Whippany, NJ, USA

Chiral AGP, Chiral CBH, and Chiral HAS

Chrom Tech, Ltd, Cheshire, UK

Resolvosil BSA-7 and Resolvosil BSA-7PX

Macherey-Nagel, Düren, Germany

Chiral AGP, Chiral CBH, and Chiral HAS

Regis Technologies, Morton Grove, IL, USA

AFpak ABA-894

Showa Denko, Kanagawa, Japan

Keystone HSA and Keystone BAS

Thermo Hypersil, Bellefonte, PA,USA

TSKgelEnantio L1 and TSK gel Enantio-OVM EnantioPac,

Tosoh, Tokyo, Japan

Bioptic AV-1

GL Sciences, Tokyo, Japan

Ultron ES-BSA, Ultron ES-OVM Column, Ultron ES-OGP Column, and Ultron ES-Pepsin

Shinwa Chemical Industries, Kyoto, Japan

LKB Pharmacia, Bromma, Sweden

Crown ether-based CSPs Crownpak CR

Chiral Technologies, Inc., Exton, PA, USA

2.3 Columns and Eluents

Opticrown RCA Chiralhyun CR-1 Chirosil CH-RCA

Separations Kasunigaseki-Chrome, Tokyo, Japan Daicel Chemical Industries, Tokyo, Japan USmac Corporation, Winnetka, Glenview, IL, USA K-MAC (Korea Materials & Analysis Corp.), South Korea Restech Corporation, Daedeok, Daejon, South Korea

Ligand exchange-based CSPs Chirosolve

JPS Chemie, Switzerland

Chiralpak WH, Chiralpak WM, Chiralpak WE, and Chiralpak MA

Separations Kasunigaseki-Chrome, Tokyo, Japan

Nucleosil Chiral-1

Macherey-Nagel, Düren, Germany

Phenylglycine and leucine Chirex types

Regis Technologies, Morton Grove, IL, USA

Orpak CRX-853

Showa Denko, Kanagawa, Japan

Pirkle-type CSPs Opticrown chiralhyun-Leu-1 and Opticrown Chiralhyun-PG-1 Whelk-O 1, Whelk-O 2, leucine, phenylglycine, β-Gem 1, α-Burke 1, α-Burke 2, Pirkle 1-J, naphthylleucine, Ulmo, and Dach Nucleosil Chiral-2

Usmac Corporation, Glenview, IL, USA

Sumichiral OA

Sumika Chemical Analysis Service, Konohana-ku Osaka, Japan

Kromasil Chiral TBB and Kromasil Chiral DMB

Eka Chemicals Separation Products, Bohus, Sweden

Chirex type I

Phenomenex, Torrance, CA, USA

Chiris series

IRIS Technologies, Lawrence, KS, USA

Regis Technologies, Morton Grove, IL, USA

Macherey-Nagel, Düren, Germany

the pH values of the buffers and the amounts of the organic modifiers. The most commonly used buffers are perchlorate, acetate, and phosphate. For the cyclodextrin-based columns, chiral resolutions have been carried out using aqueous mobile phases. Buffers of different concentrations and pH values have been used for this purpose. Triethylammonium acetate (TEAA), phosphate, citrate, and acetate are among the most commonly used buffers [30–33]. Phosphate buffers, such as sodium, potassium, and ammonium phosphate are commonly used. The stability constant of the complexes decreases due to the addition of the organic solvents and, hence, the organic modifiers are used to optimize the chiral resolution. The most commonly used organic solvents are methanol and acetonitrile. Acetonitrile is stronger than methanol. Some other organic modifiers such as ethanol, 2-propanol, 1-propanol, n-butanol,

327

328

2 Chiral Separation of Some Classes of Pesticides by HPLC Method

tetrahydrofuran, and dimethylformamide have also been used for the optimization of the chiral resolution on CD-based CSPs [30,31,34,35]. The effects of the type and concentration of these organic modifiers vary from one analyte to another. Hence, it is very difficult to predict the best strategy for their use as organic modifiers. Sometimes, the use of highly concentrated buffers under the reversed phase mode decreases the lifetime and the efficiency of a column. Therefore, the use of an alternative mobile phase, that is, the normal phase, is advantageous for the chiral resolution on these phases. The most commonly used solvents in the normal-phase mode are hexane, cyclohexane, and heptane. However, some other solvents, such as dichloromethane, acetone, 1-propanol, 2-propanol, ethyl acetate, ethanol, and chloroform, have also been used as the components of the mobile phase. The concentration of the buffer is a very important aspect of the chiral resolution on these phases under the reversed phase mode. The addition of salts to the reversed mobile phase has been found to improve the chiral resolution [33]. The macrocyclic glycopeptide antibiotic columns may be used in normal, reversed, and new polar ionic phase modes. Due to the complex structure of these antibiotics, most of them function equally well in reversed, normal, and modified polar ionic phases. All three modes generally show different selectivity with different analytes. In normal-phase chromatography, the most commonly used solvents are typically hexane, ethanol, and methanol. The optimization of chiral resolution is achieved by adding some other organic acids and bases, such as acetic acid, trifluoroacetic acid (TFA), diethylamine (DEA), or triethylamine (TEA) [36,37]. In the reversed phase mode, buffers are mostly used as mobile phases with small quantities of organic modifiers. The use of buffers as mobile phases has increased the efficiency of the resolution. Ammonium nitrate, triethylammonium acetate (TEAA), and sodium citrate buffers have been used very successfully. A variety of organic modifiers have been used to alter selectivity [38–40] and acetonitrile, methanol, ethanol, 2-propanol, and tetrahydrofuran (THF) have shown good selectivity for various analytes. In the reversed phase mode, the quantities of organic modifiers are typically low, usually on the order of 10–20%. The typical starting composition of a mobile phase is an organic modifier buffer (pH 4.0–7.0) (10:90 v/v). The use of alcohols as organic modifiers generally requires higher starting concentrations, for example, 20% for comparable retention when acetonitrile or tetrahydrofuran is used with a starting concentration of 10%. The effects of the organic solvents on the enantioselectivities also depend on the type of the antibiotic. In fact, better recognition is obtained at lower buffer pH values or close to the isoelectric point of the antibiotic, especially for vancomycin. When vancomycin CSP column is used, the low concentration of organic solvents does not significantly influence the separation, while the enantioresolution is improved for some compounds using ristocetin A and teicoplanin CSPs even when using low organic modifier concentrations [41]. The effects of organic modifiers on the chiral resolution vary from racemate to racemate [42]. The simplified approach has proven to be very effective for the resolution of a broad spectrum of racemates. The first issue to consider here is

2.3 Columns and Eluents

the structure of the analytes. If a compound has more than one functional group that is capable of interacting with a stationary phase and at least one of these groups is on or near the stereogenic center, then the first choice for a mobile phase would be the new polar ionic phase. This is due to the presence of strong polar groups in macrocyclic peptides. Thus, it is possible to convert an original mobile phase to 100% methanol, with the acid/base added to improve the selectivity. The key factor in obtaining a complete resolution still depends on the ratio of acid to base. The actual concentrations of acid and base affect the retention. Therefore, starting with a ratio of 1:1 v/v, some selectivity is typically observed. This is then followed by using ratios of 1:2 v/v and 2:1v/v to monitor change in the resolution that indicates a trend. If an analyte is eluting too quickly, the acid/base concentration is reduced. Conversely, if an analyte is retained too well, the acid/base concentration is increased. The parameters of the concentrations are between 1 and 0.01%. Above 1%, an analyte is too polar and indicates a typical reversed phase system while, when below 0.01%, it indicates a normal phase system. Both trifluoroacetic acid (TFA) and acetic acid have been used as the acid components with ammonium hydroxide and triethylamine as a base. For an analyte/pesticide that has only one functional group, or for the reason of solubility, typical normal-phase solvents (hexane/ethanol) or reversed phase solvents (THF/buffer) are employed. The pH value is an important controlling factor for the enantiomeric resolution in reversed and new polar ionic phases. In general, buffers are used as mobile phases to control a pH in HPLC. The pH value of the buffers ranges from 4.0 to 7.0 in the reversed phase system. It has been observed that when a pH value increases, the values of Rs, k, and α decrease. Therefore, the safest and the most suitable pH values in the reversed phase systems vary from a pH of 4.0 to a pH of 7.0 [38,43]. The protein-based chiral columns were mostly used under a reversed phase, that is, the aqueous mobile phases are frequently used. Buffers of different concentrations and pH values are mostly used for the chiral resolution on these CSPs. The most commonly used buffers were phosphate and borate that were used in the concentration range of 20–100 mM with 2.5–8.0 pH range. However, as with all silica-based CSPs, a prolonged use of an alkaline pH buffer (> 8.5) is not advisable. On the other hand, at a lower pH, irreversible changes are possible in the cross-linked protein phases and, hence, the use of buffers with low pH values for long periods of time is not recommended. Therefore, a buffer that ranges from pH 3.0 to pH 7.0 should be chosen. A pH of 4.5 ammonium acetate buffer may be useful. For the mobile-phase development, any buffer (50 mM, pH 7.0) can be used and the optimization is carried out by changing the concentration and the pH value. The successful use of the organic solvents in the optimization of the chiral resolution on these CSPs has been reported: the hydrophobic interactions are affected by the use of these solvents, the most important of which are methanol, ethanol, 1-propanol, 2-propanol, acetonitrile, and THF. These organic modifiers have been used in the range of 1–10%. Caution must be exercised when using these organic modifiers as they can denature the protein. However, the high concentrations of methanol and acetonitrile have

329

330

2 Chiral Separation of Some Classes of Pesticides by HPLC Method

been used on some of the cross-linked protein CSPs. The selection of these organic modifiers depends on the structure of the racemic compounds and the CSP used. In some cases, charged modifiers such as hexanoic acid, octanoic acid, and quaternary ammonium compounds have also been used for the optimum chiral resolution [44,45]. The aqueous mobile phases containing organic modifiers and acids have been used on the chiral crown ethers (CCEs) based CSPs. In all applications, the aqueous and acidic mobile phases are used, and the most commonly used mobile phases are aqueous perchloric acid and aqueous methanol containing sulfuric, trifluoroacetic, or perchloric acid separately. Compounds with a higher hydrophobicity, generally, have longer retention times on CCE-based CSPs and, therefore, organic modifiers are used to optimize the resolution [46]. This optimization is carried out by adjusting the amounts of methanol, sulfuric acid, and perchloric acid separately. In general, the separation is enhanced by an increase in methanol and a decrease in the acid concentrations. The other organic modifiers used are ethanol, acetonitrile, and THF, but methanol has been found to be the best one [47–49]. In addition to the composition of the mobile phase, the effects of other parameters, such as the temperature, the flow rate, the pH value, and the structure of the analytes, have also been studied, but only a few reports are available in the literature. It has been observed that, in general, lowering the temperature results in a better resolution. The flow rate may be used to optimize the chiral resolution of pesticides on these CSPs. Since all of the mobile phases are acidic in nature, the effect of the pH on chiral resolution is not significant. Ligand exchange columns have been used for the chiral resolution of racemic compounds containing electron-donating atoms and, therefore, its application is confined. In most cases, buffers, sometimes containing organic modifiers, have been used as mobile phase. Therefore, the optimization has been carried out by controlling the compositions of the mobile phases. There are two strategies for the development and the use of the mobile phases on these CSPs. With a CSP that has only a chiral ligand, an aqueous mobile phase containing a suitable concentration of metal ion is used while in the case of a CSP that contains a metal ion complex as a chiral ligand, a mobile phase without a metal ion is used. In the majority of the applications, aqueous solutions of metal ions or buffers have been used as mobile phases. The most commonly used buffers are ammonium acetate and phosphate. However, the use of a phosphate buffer is avoided if a metal ion is being used as the mobile phase additive in order to prevent a complex formation between the metal ion and the phosphate that may block the column. A literature search indicates that these buffers (20–50 mM) have frequently been used for successful chiral resolution, but in some instances, organic modifiers have also been used to improve the resolution. In general, acetonitrile has been used as an organic modifier [50]; however, some reports also deal with the use of methanol, ethanol, and tetrahydrofuran [51–54]. The concentrations of these modifiers vary from 10 to 30%. However, some reports have indicated the use of these organic modifiers with a concentration up to 75% [53,54]. In

2.4 Separation of Enantiomers of Pesticides by HPLC in Normal and Reversed Phase Modes

general, the chiral resolution of highly retained solutes is optimized using organic modifiers. Basically, the organic modifiers reduce the hydrophobic interactions that results in an improved resolution. The pH of the mobile phase has also been recognized as one of the most important controlling factors in the chiral resolution on ligand-exchange chiral phases [55]. The retention and the selectivity of the enantiomeric resolution have also been investigated with respect to the metal ion concentrations on these CSPs. The chiral resolution on CSPs containing only chiral ligands has been carried out using different concentrations of metal ions in the mobile phase. A normal phase mode has frequently been used for the chiral resolution of racemic compounds on Pirkle-type CSPs. Hexane, heptane, and cyclohexane are the preferred nonpolar solvents on these chiral stationary phases. Aliphatic alcohols may be considered hydrogen donors and acceptors and, thus, may interact at many points with the aromatic amide groups of CSPs generating the hydrogen bonds. Therefore, the addition of aliphatic alcohols improves the chiral resolution. Hence, these alcohols are called organic modifiers. The most commonly used alcohol is 2-propanol. However, methanol, ethanol, 1-propanol, and n-butanol have also been used. Some reports have also indicated the use of dichloromethane and chloroform as organic modifiers with hexane. In addition to this, acidic and basic additives improve the chromatographic resolution. A small amount of acetic, formic, or trifluoroacetic acid improves the peak shape and enantioselectivity for acidic and basic solutes. There are reports available that deal with the reversed phase eluents, but a prolonged use of the reversed mobile phase is not recommended. With the development of new, more stable CSPs, the use of the reversed mobile-phase mode has become possible.

2.4 Separation of Enantiomers of Pesticides by HPLC in Normal and Reversed Phase Modes

High-performance liquid chromatography is the most popular and most widely applicable technology in the field of chiral analysis of pesticides due to the availability of a large number of chiral stationary phases in the form of normal- and reversed phase modes. Over the course of time, various liquid chromatographic approaches have been developed and applied in this field. HPLC remains the most suitable modality as a result of its advantages such as its high speed, sensitivity, and reproducibility. A variety of mobile phases, including normal, reversed, and new polar organic phases, are used in HPLC. About 80% of the chiral resolutions of pharmaceuticals, drugs, agrochemicals, and other compounds have been carried out using HPLC [15–20]. HPLC has also been used for the chiral separation of pesticides [21,22]. In spite of a variety of CSPs that are available for HPLC, they have not been used very frequently in the analysis of chiral environmental pollutants. This is due to the fact that some

331

332

2 Chiral Separation of Some Classes of Pesticides by HPLC Method

organochlorine pollutants are transparent to UV radiation and, hence, the most popular UV detectors cannot be used in HPLC for detecting such xenobiotics. However, HPLC can be used jointly with MS, polarimetry, and other optical detection techniques for the chiral resolution of such types of pollutants. Moreover, a large number of reports are available for the chiral resolution of some UV-absorbing pesticides by HPLC in normal- and reversed phase modes. 2.4.1 Enantiomeric Separation of Pesticides in a Normal-Phase Mode

The enantiomeric separation of pesticides has been carried out using various types of chiral stationary phases under normal-phase conditions. Among all chiral stationary phases, polysaccharide-based CSPs are the most popular CSPs because of their versatility, durability, and loading capacity. They are effective not only under normal-phase conditions but also under reversed phase conditions when the appropriate mobile phases are used. The majority of the polysaccharide-based CSPs employed have been cellulose- and amylose-based polysaccharide columns [56,57]. Thus, Caccamese and Principato [58] separated the enantiomers of vincamine alkaloids using the Chiralpak AD column with the hexane-2-propanol and hexane-ethanol as mobile phases separately. Ellington et al. [59] described the successful separations of the enantiomers of various organophosphorus pesticides (crotoxyphos, dialifor, fonofos, fenamiphos, fensulfothion, isofenphos, malathion, methamidophos, profenofos, crufomate, prothiophos, and trichloronate) using Chiralpak AD, Chiralpak AS, Chiralcel OD, Chiralcel OJ, and Chiralcel OG chiral columns with different mixtures of heptane and ethanol as eluting solvents. They also studied the effects of concentrations of ethanol on the chiral resolution of organophosphorus pesticides and reported poor separation of the enantiomers of fenamiphos, fensulfothion, isofenphos, profenofos, crufomate, and trichloronate pesticides at the higher concentrations of ethanol. The effects of temperature on the chiral resolution of organophosphorus pesticides had also been reported and the maximum chiral resolution was achieved at low temperatures. The effects of temperature (from 20–60 °C) on the chiral resolution of fensulfothion on Chiralcel OJ column is shown in Figure 2.2. It may be concluded from this figure that chiral resolution is enhanced at lower temperatures and reaches its maximum at 20 °C. On the other hand, the retention times increase when the temperature decreases. Ali and Aboul-Enein [60] studied the effects of various polysaccharide chiral stationary phases on the chiral resolution of o,p-DDT and o,p-DDD on Chiralpak ADR, Chiralcel OD-R, and Chiralcel OJ-R. The results of these findings are presented in Table 2.2, which shows that the best resolution of these pesticides has been obtained on Chiralpak AD-R under the normal-phase mode. Li et al. [61] developed a fast and precise HPLC method for the chiral resolution of phenthoate in soil samples. The authors used Chiralcel OD column for the chiral resolution with hexane-2-propanol (100:0.8, v/v) as the mobile phase. AboulEnein and Ali [62] reported that the chiral resolution on the polysaccharide-

2.4 Separation of Enantiomers of Pesticides by HPLC in Normal and Reversed Phase Modes

Figure 2.2 Effect of temperature on chiral separation of fensulfothion with Chiralcel OJ column [58].

based CSPs is pH dependent under the normal-phase mode. Only a partial resolution of certain antifungal agents was achieved at lower pH, while the resolution improved by increasing the pH using the triethylamine on amylose and cellulose chiral columns. Champion et al. [63] performed the enantiomeric separation of five polychlorinated compounds (trans-chlordane, cis-chlordane, heptachlor, heptachlor epoxide, and α-HCH) with different polysaccharide CSPs in the normal phase mode and the values of chromatographic parameters measured are shown in Table 2.2. As seen in Table 2.2, baseline separation was achieved for the enantiomers of trans-chlordane, cis-chlordane, heptachlor on Chiralcel OD column, α-HCH on Chiralcel OJ column, and heptachlor epoxide on Chiralpak AD column. The effects of the concentration of isopropanol (IPA) on the chiral resolution of cis-chlordane, trans-chlordane on Chiralcel OD column are shown in Figure 2.3. It may be concluded that a concentration of 0% of isopropanol is the most appropriate one as it enables the maximum resolution of cis- and trans-chlordane on Chiralcel OD column. Wang et al. [64] studied the effect of alcohols (ethanol, n-propanol, isopropanol, isobutanol, and n-butanol) on the resolutions of fipronil, isocarbophos, and carfentrazone-ethyl pesticides on the cellulose-tris(3,5-dimethylphenylcarbamate) chiral stationary phase. The concentration range for each alcohol has been 2–20%; however, in the case of carfentrazone-ethyl, it was 0.1–20% as shown in Table 2.2. The best results for the resolution of fipronil and isocarbophos was obtained using 5% isobutanol and 5% isopropanol, respectively. At the same time, the best separation of carfentrazone-ethyl was obtained with 0.5% isopropanol, while no resolution was observed with ethanol, n-propanol, n-butanol, and isobutanol from 20 to 2 %. Liu, Gan, and Qin [65] carried out the enantiomeric resolution of synthetic pyrethroid (bifenthrin, permethrin, cypermethrin, and cyfluthrin) on Sumichiral OA-2500-I and the two chained Chirex 00G-3019-DO columns with hexane-1,

333

334

Chiral separation of pesticides on different columns with normal and reversed mobile-phase conditions.

Pesticides

Column/CSPs

Mobile phase (v/v)

k1

k2

α

Rs

References

trans-Chlordane

Chiralcel-OD

Hexane/isopropanol (99–1)

0.83

0.93

1.1

—

[62]

Hexane (100)

2.4

2.9

1.2

—

Hexane/isopropanol (99–1)

0.87

1.0

1.2

—

Hexane (100)

1.7

3.2

1.9

—

cis-Chlordane

Chiralcel-OD

Heptachlor

Chiralcel-OD

Hexane (100)

0.80

0.97

1.2

—

α-HCH

Chiralcel-OJ

Hexane/isopropanol (90-10)

1.0

1.4

1.4

—

Heptachlor epoxide

Chiralpak-AD

MeOH

0.37

0.64

1.7

—

Fipronil

Cellulose-tri (3,5-DMPC)

Hexane/ethanol (80–20)

0.78

—

1.00

0.0

Hexane/ethanol (85–15)

0.99

—

1.12

0.25

Hexane/ethanol (90-10)

1.69

—

1.14

0.63

Hexane/ethanol (95–05)

4.27

—

1.18

1.01

Hexane/ethanol (98–02)

10.70

—

1.22

1.27

Hexane/n-propanol (80–20)

0.90

—

1.00

0.0

Hexane/n-propanol (85–15)

1.40

—

1.00

0.0

Hexane/n-propanol (90–10)

2.18

—

1.13

0.46

Hexane/n-propanol (95–05)

4.94

—

1.19

1.06

Hexane/n-propanol (98–02)

11.83

—

1.21

1.27

Hexane/isopropanol (80–20)

0.90

—

1.26

0.0

Hexane/isopropanol (85–15)

1.38

—

1.28

0.46

Hexane/isopropanol (90–10)

2.54

—

1.31

0.91

Hexane/isopropanol (95–05)

6.06

—

1.36

1.37

[63]

2 Chiral Separation of Some Classes of Pesticides by HPLC Method

Table 2.2

Cellulose-tris (3,5-DMPC)

15.56

—

1.37

1.69

Hexane/n-butanol (80–20)

1.47

—

1.00

0.0

Hexane/n-butanol (85–15)

1.38

—

1.00

0.0

Hexane/n-butanol (90–10)

2.14

—

1.12

0.38

Hexane/n-butanol (95–150)

4.85

—

1.17

1.05

Hexane/n-butanol (98–02)

11.28

—

1.20

1.22

Hexane/isobutanol (80–20)

1.16

—

1.29

0.0

Hexane/isobutanol (85–15)

1.83

—

1.35

1.26

Hexane/isobutanol (90–10)

3.15

—

1.37

1.47

Hexane/isobutanol (95–05)

7.42

—

1.38

1.81

Hexane/isobutanol (98–02)

—

—

—

—

Hexane/ethanol (80–20)

0.77

—

1.17

0.30

Hexane/ethanol (85–15)

0.97

—

1.23

0.51

Hexane/ethanol (90–10)

1.21

—

1.34

0.78

Hexane/ethanol (95–05)

2.13

—

1.25

1.32

Hexane/ethanol (98–02)

3.69

—

1.27

1.61 0.96

Hexane/n-propanol (80–20)

0.87

—

1.34

Hexane/n-propanol (85–15)

1.05

—

1.33

1.14

Hexane/n-propanol (90–10)

1.56

—

1.37

1.40

Hexane/n-propanol (95–05)

2.61

—

1.40

1.81

Hexane/n-propanol (98–02)

4.49

—

1.41

2.63

Hexane/isopropanol (80–20)

0.96

—

1.52

1.47

Hexane/isopropanol (85–15)

1.43

—

1.60

1.79

Hexane/isopropanol (90–10)

1.90

—

1.67

2.18 (continued)

2.4 Separation of Enantiomers of Pesticides by HPLC in Normal and Reversed Phase Modes

Isocarbophos

Hexane/isopropanol (98–02)

335

336

Pesticides

Carfentrazone-ethyl

λ-cyhalothrin

Column/CSPs

Cellulose-tri (3,5-DMPC)

Chiralpak AD

Mobile phase (v/v)

k1

k2

α

Rs 2.66

Hexane/isopropanol (95–05)

3.26

—

1.89

Hexane/isopropanol (98–02)

5.93

—

1.66

2.42

Hexane/n-butanol (80–20)

0.94

—

1.28

0.87

Hexane/n-butanol (85–15)

1.09

—

1.28

0.91

Hexane/n-butanol (90–10)

1.50

—

1.30

1.27

Hexane/n-butanol (95–05)

2.42

—

1.32

1.38

Hexane/n-butanol (98–02)

4.90

—

1.39

1.86

Hexane/isobutanol (80–20)

1.01

—

1.40

1.23

Hexane/isobutanol (85–15)

1.32

—

1.43

1.56

Hexane/isobutanol (90–10

1.65

—

1.43

1.61

Hexane/isobutanol (95–05)

1.34

—

2.70

1.65

Hexane/isobutanol (98–02)

6.27

—

1.68

2.56

Hexane/isopropanol (80–20)

0.78

—

1.00

0.0

Hexane/isopropanol (85–15)

0.91

—

1.00

0.0

Hexane/isopropanol (90–10)

1.08

—

1.10

0.24

Hexane/isopropanol (95–05)

1.47

—

1.10

0.33 0.45

Hexane/isopropanol (98–02)

2.49

—

1.10

Hexane/isopropanol (99–01)

11.72

—

1.08

0.52

Hexane/isopropanol (99.5–0.5)

5.45

—

1.07

0.58

Hexane/isopropanol (99.9–0.1)

3.53

—

1.09

0.48

Hexane/ethanol (95–05)

0.56

0.67

1.20

1.60

Hexane/ethanol (96–04)

0.60

0.72

1.24

1.81

References

[67]

2 Chiral Separation of Some Classes of Pesticides by HPLC Method

Table 2.2 (Continued)

0.63

0.81

1.28

2.06

Hexane/ethanol (99–01)

0.84

1.10

1.30

2.20

Hexane/ethanol (98–02)

1.10

1.35

1.32

2.35

Hexane/1, 2-dichloroethane (95–05)

1.05

1.94

1.85

4.96

Hexane/1, 2-dichloroethane (96–04)

1.72

3.19

1.86

4.90

Hexane/1, 2-dichloroethane (97–03)

2.98

5.65

1.90

4.54

Hexane/1, 2-dichloroethane (98–02)

4.56

8.44

1.85

4.18

Hexane/1, 2-dichloroethane (99–01)

6.70

2.19

1.82

3.17

Hexane/isopropanol (95–05)

1.18

1.75

1.49

5.95

Hexane/isopropanol (96–04)

1.25

1.92

1.53

6.54

Hexane/isopropanol (97–03)

1.79

2.34

1.55

7.31

Hexane/isopropanol (98–02)

2.24

3.59

1.60

8.41

Hexane/isopropanol (99–01)

3.33

5.76

1.73

10.41

Hexane/ethanol (95–05)

1.56

1.99

1.27

3.22

Hexane/ethanol (96–04)

1.82

2.34

1.28

3.50

Hexane/ethanol (97–03)

2.24

2.90

1.30

3.82

Hexane/ethanol (98–02)

2.92

3.87

1.33

4.30

Hexane/ethanol (99–01)

4.52

6.13

1.36

4.79

Chiralcel OJ

Hexane/isopropanol (95–05)

8.47

9.32

1.10

1.56

Chiralcel OD

Hexane/isopropanol (99.5–0.5)

6.76

7.40

1.09

1.42

Chiralpak AS

Chiralcel OD

Chiralcel OJ

Salithion

Neonicotinoid 1

Chiralpak AD

Hexane/isopropanol (99.5–0.5)

4.15

5.03

1.21

4.14

Chiralpak OT (+)

Methanol (100)

2.02

2.52

1.25

2.28

Chiralcel OD-H

Hexane/ethanol (40–60)

1.40

1.97

1.41

2.74

Hexane/ethanol (50–50)

1.92

2.65

1.38

2.75

Hexane/ethanol (60–40)

2.98

4.05

1.36

2.77

[70]

[73]

(continued)

2.4 Separation of Enantiomers of Pesticides by HPLC in Normal and Reversed Phase Modes

Hexane/ethanol (97–03)

337

338

Pesticides

Neonicotinoid 2

Neonicotinoid 3

Metalaxyl

Myclobutanil

Imazalil

Malathion

Column/CSPs

Chiralcel OD-H

Chiralcel OD-H

Amylose tris-(S)-1-(PEC)

Amylose tris-(S)-1-(PEC)

Amylose tris-(S)-1-(PEC)

Amylose tris-(S)-1-(PEC)

Mobile phase (v/v)

k1

k2

α

Rs

Hexane/ethanol (70–30)

5.25

7.01

1.34

2.87

Hexane/ethanol (80–20)

11.75

15.46

1.32

3.15

Hexane/ethanol (40–60)

0.69

1.13

1.62

1.98

Hexane/ethanol (50–50)

1.02

1.67

1.64

2.15

Hexane/ethanol (60–40)

1.70

2.82

1.67

2.42

Hexane/ethanol (70–30)

3.26

5.57

1.71

2.84

Hexane/ethanol (80–20)

7.79

13.69

1.76

3.43

Hexane/ethanol (50–50)

0.41

0.49

1.18

0.98

Hexane/ethanol (60–40)

0.64

0.75

1.17

1.09

Hexane/ethanol (70–30)

1.18

1.39

1.17

1.19

Hexane/ethanol (80–20)

2.50

2.90

1.16

1.27

Hexane/ethanol (90–10)

8.54

9.88

1.57

1.49

Hexane/isopropanol (85–15)

3.01

4.66

1.55

2.51

Hexane/isopropanol (90–10)

4.28

6.68

1.56

2.80

Hexane/isopropanol (95–05)

7.56

12.27

1.62

3.75

Hexane/isopropanol (85–15)

5.08

5.96

1.17

0.86

Hexane/isopropanol (90–10)

8.01

9.59

1.20

1.06

Hexane/isopropanol (95–05)

17.49

21.74

1.24

1.49

Hexane/isopropanol (85–15)

2.50

2.70

1.08

0.52

Hexane/isopropanol (90–10)

3.71

4.08

1.10

0.70

Hexane/isopropanol (95–05)

7.56

8.44

1.12

0.86

Hexane/isopropanol (90–10)

2.49

2.49

1.00

0.0

References

[76]

2 Chiral Separation of Some Classes of Pesticides by HPLC Method

Table 2.2 (Continued)

Triadimefon

Fipronil

Napropamide

Paclobutrazol

Metalaxyl

Hexaconazole

Amylose tris-(S)-1-(PEC)

Amylose tris-(S)-1-(PEC)

Amylose tris-(S)-1-(PEC)

Amylose tris-(S)-1-(PEC)

Amylopectin-tris-(PC)

Amylopectin-tris-(PC)

3.19

3.48

1.09

0.77

Hexane/isopropanol (95–05)

4.53

5.02

1.11

0.87

Hexane/isopropanol (85–15)

1.81

2.33

1.29

1.34

Hexane/isopropanol (90–10)

2.48

3.27

1.32

1.43

Hexane/isopropanol (95–05)

3.74

5.07

1.35

1.59

Hexane/isopropanol (98–02)

5.97

8.47

1.42

1.84

Hexane/isopropanol (90–10)

4.71

4.71

1.00

0.0

Hexane/isopropanol (95–05)

6.66

6.92

1.04

0.40

Hexane/isopropanol (98–02)

9.95

10.58

1.06

0.63

Hexane/isopropanol (85–15)

2.83

3.45

1.22

1.03

Hexane/isopropanol (90–10)

7.72

9.88

1.28

1.35

Hexane/isopropanol (95–05)

14.63

19.42

1.33

2.03

Hexane/isopropanol (85–15)

1.55

1.66

1.07

0.38

Hexane/isopropanol (90–10)

2.04

2.25

1.10

0.55

Hexane/isopropanol (95–05)

3.31

3.66

1.11

0.72

Hexane/isopropanol (98–02)

7.65

8.76

1.14

1.14

Hexane/isopropanol (85–15)

2.03

2.36

1.16

0.81

Hexane/isopropanol (90–10)

3.57

4.29

1.20

1.30

Hexane/isopropanol (95–05)

6.38

7.74

1.21

1.66

Hexane/isopropanol (80–20)

4.57

5.34

1.17

0.94

Hexane/isopropanol (85–15)

5.74

6.69

1.16

0.97

Hexane/isopropanol (90–10)

9.11

10.77

1.18

1.37

Hexane/isopropanol (80–20)

1.78

2.38

1.34

1.49

Hexane/isopropanol (85–15)

2.33

3.19

1.36

1.83

Hexane/isopropanol (90–10)

3.60

4.99

1.39

2.03

[79]

(continued)

2.4 Separation of Enantiomers of Pesticides by HPLC in Normal and Reversed Phase Modes

Ethofumesate

Amylose tris-(S)-1-(PEC)

Hexane/isopropanol (98–02)

339

340

Pesticides

Column/CSPs

Myclobutanil

Amylopectin-tris-(PC)

Tebuconazole

Uniconazole

Paclobutrazol

Benalaxyl

EPN

Amylopectin-tris-(PC)

Amylopectin-tris-(PC)

Amylopectin-tris-(PC)

Amylopectin-tris-(PC)

Chiralpak AD

Mobile phase (v/v)

k1

k2

α

Rs

Hexane/isopropanol (95–05)

7.53

10.60

1.41

2.45

Hexane/isopropanol (70–30)

3.72

4.55

1.22

0.86

Hexane/isopropanol (80–20)

5.63

6.93

1.23

1.10

Hexane/isopropanol (85–15)

8.17

10.04

1.23

1.20

Hexane/isopropanol (90–10)

13.11

16.27

1.24

1.47

Hexane/isopropanol (80–20)

2.07

2.38

1.15

0.63

Hexane/isopropanol (85–15)

2.80

3.32

1.19

0.92

Hexane/isopropanol (90–10)

4.37

5.39

1.23

1.34

Hexane/isopropanol (95–05)

9.84

12.01

1.22

1.54

Hexane/isopropanol (80–20)

2.11

2.13

1.01

0.05

Hexane/isopropanol (85–15)

2.31

3.02

1.31

1.39

Hexane/isopropanol (90–10)

3.86

5.17

1.34

1.48

Hexane/isopropanol (95–05)

9.33

13.09

1.40

2.05

Hexane/isopropanol (85–15)

2.30

3.41

1.48

1.61

Hexane/isopropanol (90–10)

3.79

5.88

1.55

2.19

Hexane/isopropanol (95–05)

8.16

12.65

1.55

2.42

Hexane/isopropanol (85–15)

2.59

2.91

1.12

0.66

Hexane/isopropanol (90–10)

3.63

4.12

1.13

0.77

Hexane/isopropanol (95–05)

4.75

5.59

1.18

0.83

Hexane/isopropanol (97–03)

8.79

10.32

1.17

1.01

Hexane/isopropanol (90–10)

0.86

0.96

1.11

1.13

Hexane/isopropanol (95–05)

0.90

1.04

1.16

1.41

References

[80]

2 Chiral Separation of Some Classes of Pesticides by HPLC Method

Table 2.2 (Continued)

0.89

1.05

1.18

1.61

Hexane/isopropanol (97–03)

0.94

1.08

1.15

2.31

Hexane/isopropanol (98–02)

1.17

1.46

1.25

3.22

Hexane/isopropanol (99–01)

1.86

2.61

1.40

5.39

Hexane/isopropanol (90–10)

1.23

1.45

1.18

1.89

Hexane/isopropanol (95–05)

1.33

1.59

1.19

2.11

Hexane/isopropanol (96–04)

1.56

1.87

1.19

2.28

Hexane/isopropanol (97–03)

1.48

1.70

1.15

1.87

Hexane/isopropanol (98–02)

1.93

2.24

1.16

2.13

Hexane/isopropanol (99–01)

2.27

2.68

1.18

2.50

Acetonitrile/water (50–50)

15.41

19.77

1.24

2.47

Acetonitrile/water (50–50)

—

—

—

—

Chiralpak AD-RH

Acetonitrile/isopropanol (50–50)

4.74

8.00

1.69

1.00

Acetonitrile/isopropanol (50–50)

3.26

4.11

1.26

0.60

Chiralcel OD-RH

Acetonitrile/water (50–50)

4.54

10.28

2.27

2.03

Acetonitrile/water (50-50)

—

—

—

—

Chiralcel OJ-R

Acetonitrile/isopropanol (50–50)

—

—

—

—

Chiralpak AS

o,p-DDT

Chiralpak AD-RH

o,p-DDD o,p-DDT o,p-DDD o,p-DDT o,p-DDD o,p-DDT o,p-DDD

Acetonitrile/isopropanol (50–50)

—

—

—

—

o,p-DDT

Acetonitrile/water (50–50)

3.49

8.80

2.52

0.80

o,p-DDD

Acetonitrile/water (50–50)

—

—

—

—

o,p-DDT

Acetonitrile/isopropanol (50–50)

—

—

—

—

o,p-DDD

Acetonitrile/isopropanol (50–50)

—

—

—

—

Epoxiconazole

Cellulose-tris-(3,5-DMPC)

Methanol/water (75–25)

4.38

8.26

1.88

5.54

Methanol/water (80–20)

2.79

5.40

1.93

5.45

Methanol/water (85–15)

1.74

3.34

1.92

5.32

[59]

[81]

(continued)

2.4 Separation of Enantiomers of Pesticides by HPLC in Normal and Reversed Phase Modes

Hexane/isopropanol (96–04)

341

342

Pesticides

Terallethrin

Pyriproxyfen

Column/CSPs

Cellulose-tris-(3,5-DMPC)

Cellulose-tris-(3,5-DMPC)

Mobile phase (v/v)

k1

k2

α

Rs

Methanol/water (90–10)

1.16

2.22

1.91

5.05

Methanol/water (95–5)

0.90

1.75

1.96

5.00

Methanol/water (100–0)

1.01

1.62

1.60

3.41

Acetonitrile/water (50–50)

4.18

8.39

2.00

9.23

Acetonitrile/water (60–40)

1.85

3.86

2.08

7.71

Acetonitrile/water (70–30)

1.03

2.16

2.09

6.08

Acetonitrile/water (80–20)

0.65

1.39

2.12

5.14

Acetonitrile/water(90–10)

0.56

1.09

1.96

4.21

Acetonitrile/water (100–0)

0.95

1.75

1.84

4.83

Methanol/water (65–35)

15.09

15.91

1.05

0.66

Methanol/water (70–30)

8.18

8.58

1.05

0.54

Methanol/water (75–25)

4.52

4.74

1.05

0.53

Methanol/water (80–20)

2.63

2.74

1.04

0.42

Methanol/water (90–10)

1.01

1.01

1.00

—

Methanol/water (100–0)

0.70

0.70

1.00

—

Acetonitrile/water (50–50)

5.66

6.28

1.11

1.56

Acetonitrile/water (60–40)

2.18

2.43

1.12

1.22

Acetonitrile/water (70–30)

1.03

1.15

1.12

0.89

Acetonitrile/water (80–20)

0.53

0.60

1.13

0.66

Acetonitrile/water (90–10)

0.33

0.33

1.00

—

Acetonitrile/water (100–0)

0.51

0.51

1.00

—

Methanol/water (80–20)

12.01

13.05

1.09

0.93

References

2 Chiral Separation of Some Classes of Pesticides by HPLC Method

Table 2.2 (Continued)

Lactofen

Cellulose-tris-(3,5-DMPC)

Cellulose-tris-(3,5-DMPC)

5.99

6.43

1.07

0.66

Methanol/water (90–10)

3.20

3.38

1.06

0.61

Methanol/water (95–05)

2.04

2.04

1.00

—

Methanol/water (100–0)

1.13

1.13

1.00

—

Acetonitrile/water (55–45)

12.38

13.09

1.06

0.91

Acetonitrile/water (60–40)

7.38

7.82

1.06

0.84

Acetonitrile/water (70–30)

3.20

3.39

1.06

0.72

Acetonitrile/water (80–20)

1.52

1.58

1.04

0.47

Acetonitrile/water (90–10)

0.75

0.75

1.00

—

Acetonitrile/water (100–0)

0.45

0.45

1.00

—

Methanol/water (75–25)

4.02

4.82

1.20

1.59

Methanol/water (85–15)

1.54

1.83

1.19

1.35

Methanol/water (90–10)

1.04

1.22

1.17

1.13

Methanol/water (95–05)

0.76

0.88

1.16

0.92

Methanol/water (100–0)

0.64

0.72

1.14

0.76

Acetonitrile/water (40–60)

14.20

14.20

1.00

—

Acetonitrile/water (80–20)

0.54

0.54

1.00

—

Acetonitrile/water (100–0)

0.38

0.38

1.00

—

Methanol/water (75–25)

14.71

6.26

1.11

1.07

Methanol/water (80–20)

7.01

7.75

1.10

0.92

Methanol/water (85–15)

3.13

3.46

1.11

0.83

Methanol/water (90–10)

1.52

1.69

1.11

0.66

Methanol/water (95–05)

0.79

0.88

1.11

0.63

Methanol/water (100–0)

0.77

0.77

1.00

—

Acetonitrile/water (50–50)

15.19

15.19

1.00

— (continued)

2.4 Separation of Enantiomers of Pesticides by HPLC in Normal and Reversed Phase Modes

Benalaxyl

Methanol/water (85–15)

343

344

Pesticides

Quizalofop-ethyl

Diclofop-methyl

Profenofos

Column/CSPs

Cellulose-tris-(3,5-DMPC)

Cellulose-tris-(3,5-DMPC)

Cellulose-tris-(3,5-DMPC)

Mobile phase (v/v)

k1

k2

α

Rs

Acetonitrile/water (80–20)

0.57

0.57

1.00

—

Acetonitrile/water (100–0)

0.44

0.44

1.00

—

Methanol/water (75–25)

21.48

22.61

1.05

0.69

Methanol/water (80–20)

11.40

12.02

1.05

0.59

Methanol/water (85–15)

6.06

6.38

1.05

0.54

Methanol/water (90–10)

3.40

3.56

1.05

0.45

Methanol/water (95–05)

2.11

2.11

1.00

—

Methanol/water (100-0)

1.20

1.20

1.00

—

Acetonitrile/water (50–50)

13.04

13.04

1.00

—

Acetonitrile/water (80–20)

1.02

1.02

1.00

—

Acetonitrile/water (100–0)

0.45

0.45

1.00

—

Methanol/water (75–25)

14.43

14.43

1.00

—

Methanol/water (85–15)

4.43

4.43

1.00

—

Methanol/water (100–0)

0.77

0.77

1.00

—

Acetonitrile/water (50–50)

12.73

13.95

1.10

1.53

Acetonitrile/water (60–40)

4.35

4.76

1.09

1.22

Acetonitrile/water (70–30)

1.87

2.05

1.10

0.97

Acetonitrile/water (80–20)

0.87

0.95

1.10

0.72

Acetonitrile/water (90–10)

0.86

0.86

1.00

—

Acetonitrile/water (100–0)

0.34

0.34

1.00

—

Methanol/water (70–30)

12.90

12.90

1.00

—

Methanol/water (80–20)

4.25

4.25

1.00

—

References

2 Chiral Separation of Some Classes of Pesticides by HPLC Method

Table 2.2 (Continued)

Methamidophos

MCPA-isooctyl

Isofenphos-methyl

Cellulose-tris-(3,5-DMPC)

Cellulose-tris-(3,5-DMPC)

Cellulose-tris-(3,5-DMPC)

Cellulose-tris-(3,5-DMPC)

0.93

0.93

1.00

—

Acetonitrile/water (50–50)

8.47

8.47

1.00

—

Acetonitrile/water (70–30)

1.83

1.83

1.00

—

Acetonitrile/water (100–0)

0.62

0.62

1.00

—

Methanol/water (70–30)

6.90

6.90

1.00

—

Methanol/water (80–20)

2.41

2.41

1.00

—

Methanol/water (100–0)

0.77

0.77

1.00

—

Acetonitrile/water (70–30)

0.85

0.85

1.00

—

Acetonitrile/water (100–0)

0.49

0.49

1.00

—

Methanol/water (60–40)

1.28

1.28

1.00

—

Methanol/water (80–20)

1.82

1.82

1.00

—

Methanol/water (100–0)

0.76

0.76

1.00

—

Acetonitrile/water (40–60)

0.97

0.97

1.00

—

Acetonitrile/water (70–30)

2.35

2.35

1.00

—

Acetonitrile/water (100–0)

0.96

0.96

1.00

—

Methanol/water (80–20)

12.57

12.57

1.00

—

Methanol/water (90–10)

2.88

2.88

1.00

—

Methanol/water (100–0)

1.12

1.12

1.00

—

Acetonitrile/water (60-40)

10.37

10.37

1.00

—

Acetonitrile/water (80–20)

1.88

1.88

1.00

—

Acetonitrile/water (100–0)

0.79

0.79

1.00

—

Methanol/water (70–30)

5.49

5.49

1.00

—

Methanol/water (80–20)

1.99

1.99

1.00

—

Methanol/water (100–0)

0.65

0.65

1.00

—

Acetonitrile/water (50–50)

5.21

5.21

1.00

— (continued)

2.4 Separation of Enantiomers of Pesticides by HPLC in Normal and Reversed Phase Modes

Malathion

Methanol/water (100–0)

345

346

Pesticides

Phenthoate

Fluroxypyr-meptyl

Acephate

Trichlorphon

Column/CSPs

Cellulose-tris-(3,5-DMPC)

Cellulose-tris-(3,5-DMPC)

Cellulose-tris-(3,5-DMPC)

Cellulose-tris-(3,5-DMPC)

Mobile phase (v/v)

k1

k2

α

Rs

Acetonitrile/water (70–30)

1.07

1.07

1.00

—

Acetonitrile/water (100–0)

0.53

0.53

1.00

—

Methanol/water (75–25)

7.54

7.54

1.00

—

Methanol/water (90–10)

1.39

1.39

1.00

—

Methanol/water (100–0)

0.87

0.87

1.00

—

Acetonitrile/water (50–50)

6.57

6.57

1.00

—

Acetonitrile/water (70–30)

1.26

1.26

1.00

—

Acetonitrile/water (100–0)

0.57

0.57

1.00

—

Methanol/water (75–25)

14.85

14.85

1.00

—

Methanol/water (90–10)

1.94

1.94

1.00

—

Methanol/water (100–0)

0.80

0.80

1.00

—

Acetonitrile/water (50–50)

17.75

17.75

1.00

—

Acetonitrile/water (70–30)

2.33

2.33

1.00

—

Acetonitrile/water (100–0)

0.65

0.65

1.00

—

Methanol/water (60–40)

0.97

0.97

1.00

—

Methanol/water (80–20)

0.88

0.88

1.00

—

Methanol/water (100–0)

0.76

0.76

1.00

—

Acetonitrile/water (50–50)

0.76

0.76

1.00

—

Acetonitrile/water (80–20)

0.84

0.84

1.00

—

Acetonitrile/water (100–0)

0.93

0.93

1.00

—

Methanol/water (60–40)

1.33

1.33

1.00

—

Methanol/water (80–20)

0.87

0.87

1.00

—

References

2 Chiral Separation of Some Classes of Pesticides by HPLC Method

Table 2.2 (Continued)

Fenamiphos

Acetochlor

P-tefuryl quizalofop-

Cellulose-tris-(3,5-DMPC)

Cellulose-tris-(3,5-DMPC)

Cellulose-tris-(3,5-DMPC)

Cellulose-tris-(3,5-DMPC)

1.07

1.07

1.00

—

Acetonitrile/water (80–20)

0.61

0.61

1.00

—

Acetonitrile/water (100–0)

0.99

0.99

1.00

—

Methanol/water (80–20)

13.38

13.38

1.00

—

Methanol/water (90–10)

3.13

3.13

1.00

—

Methanol/water (100–0)

1.17

1.17

1.00

—

Acetonitrile/water (60–40)

10.06

10.06

1.00

—

Acetonitrile/water (80–20)

1.84

1.84

1.00

—

Acetonitrile/water (100–0)

0.78

0.78

1.00

—

Methanol/water (70–30)

3.17

3.17

1.00

—

Methanol/water (90–10)

0.85

0.85

1.00

—

Methanol/water (100–0)

0.62

0.62

1.00

—

Acetonitrile/water (40–60)

6.45

6.45

1.00

—

Acetonitrile/water (70–30)

0.81

0.81

1.00

—

Acetonitrile/water (100–0)

1.69

1.69

1.00

—

Methanol/water (70–30)

4.33

4.33

1.00

—

Methanol/water (90–10)

1.25

1.25

1.00

—

Methanol/water (100–0)

0.82

0.82

1.00

—

Acetonitrile/water (50–50)

4.22

4.22

1.00

—

Acetonitrile/water (70–30)

1.13

1.13

1.00

—

Acetonitrile/water (100–0)

0.66

0.66

1.00

—

Methanol/water (85–15)

9.46

9.46

1.00

—

Methanol/water (90–10)

1.75

1.75

1.00

—

Methanol/water (100–0)

5.59

5.59

1.00

—

Acetonitrile/water (60–40)

5.55

5.55

1.00

— (continued)

2.4 Separation of Enantiomers of Pesticides by HPLC in Normal and Reversed Phase Modes

2,4-D-ethylhexyl

Methanol/water (100–0)

347

348

Pesticides

Hexaconazole

Column/CSPs

Cellulose-tris-(3,5-DMPC) (5 μM)

Cellulose-tris-(3,5-DMPC) (3 μM)

Cellulose-tris-(3,5-DMPC) (5 μM)

Cellulose-tris-(3,5-DMPC) (3 μM)

Flutriafol

Cellulose-tris-(3,5-DMPC) (5 μM)

Cellulose-tris-(3,5-DMPC) (3 μM)

Cellulose-tris-(3,5-DMPC) (5 μM)

Mobile phase (v/v)

k1

k2

α

Rs

Acetonitrile/water (70–30)

0.98

0.98

1.00

—

Acetonitrile/water (100–0)

2.76

2.76

1.00

—

Acetonitrile/water (50–50)

3.20

—

1.15

3.94

Acetonitrile/water (70–30)

0.87

—

1.17

2.98 2.25

Acetonitrile/water (90–10)

0.44

—

1.19

Acetonitrile/water (50–50)

3.41

—

1.16

3.09

Acetonitrile/water (70–30)

0.88

—

1.17

1.71

Acetonitrile/water (90–10)

0.48

—

1.17

1.26

Methanol/water (70–30)

5.45

—

1.14

2.59

Methanol/water (80–20)

1.75

—

1.13

2.12

Methanol/water (90–10)

0.66

—

1.13

1.44

Methanol/water (70–30)

4.86

—

1.12

1.50

Methanol/water (80–20)

1.71

—

1.09

0.76

Methanol/water (90–10)

—

—

—

—

Acetonitrile/water (50–50)

1.45

—

1.12

2.55

Acetonitrile/water (70–30)

0.45

—

1.14

1.99

Acetonitrile/water (90–10)

0.27

—

1.16

1.35

Acetonitrile/water (50–50)

1.52

—

1.12

1.44

Acetonitrile/water (70–30)

0.49

—

1.07

0.50

Acetonitrile/water (90–10)

—

—

—

—

Methanol/water (70–30)

2.13

—

1.08

1.39

Methanol/water (80–20)

0.87

—

1.08

1.06

References

[82]

2 Chiral Separation of Some Classes of Pesticides by HPLC Method

Table 2.2 (Continued)

Cellulose-tris-(3,5-DMPC) (3 μM)

Cellulose-tris-(3,5-DMPC) (5 μM)

Cellulose-tris-(3,5-DMPC) (3 μM)

Tetraconazole

Cellulose-tris-(3,5-DMPC) (5 μM)

Cellulose-tris-(3,5-DMPC) (3 μM)

Epoxiconazole

—

—

—

—

—

—

—

—

Methanol/water (80–20)

—

—

—

—

Methanol/water (90–10)

—

—

—

—

Acetonitrile/water (50–50)

3.71

—

1.11

3.09

Acetonitrile/water (70–30)

0.92

—

1.12

2.31

Acetonitrile/Water (90–10)

0.4

—

1.11

1.22

Acetonitrile/water (50–50)

3.98

—

1.34

6.71

Acetonitrile/water (70–30)

0.93

—

1.13

1.30

Acetonitrile/water (90–10)

—

—

—

—

Acetonitrile/water (50–50)

4.15

—

1.29

7.37

Acetonitrile/water (70–30)

0.86

—

1.31

5.04

Acetonitrile/water (90–10)

0.30

—

1.39

3.35

Acetonitrile/water (50–50)

4.49

—

1.29

5.79

Acetonitrile/water (70–30)

0.87

—

1.30

2.85

Acetonitrile/water (90–10)

0.34

—

1.33

1.66

Cellulose-tris-(3,5-DMPC) (5 μM)

Methanol/water (70–30)

—

—

—

—

Cellulose-tris-(3,5-DMPC) (3 μM)

Methanol/water (70–30)

—

—

—

—

Cellulose-tris-(3,5-DMPC) (5 μM)

Acetonitrile/water (50–50)

4.22

—

2.04

20.90

Acetonitrile/water (70–30)

1.03

—

2.10

16.89

Acetonitrile/water (90–10)

0.42

—

2.27

12.3

Acetonitrile/water (50–50)

4.57

—

2.00

16.71

Acetonitrile/water (70–30)

1.06

—

2.03

10.45

Acetonitrile/water (90–10)

0.47

—

2.08

7.06

Methanol/water (70–30)

8.71

—

1.29

5.86

Cellulose-tris-(3,5-DMPC) (3 μM)

Cellulose-tris-(3,5-DMPC) (5 μM)

(continued)

2.4 Separation of Enantiomers of Pesticides by HPLC in Normal and Reversed Phase Modes

Diniconazole

Methanol/water (90–10) Methanol/water (70–30)

349

350

Pesticides

Column/CSPs

Cellulose-tris-(3,5-DMPC) (3 μM)

Penconazole

Cellulose-tris-(3,5-DMPC) (5 μM)

Cellulose-tris-(3,5-DMPC) (3 μM)

Cellulose-tris-(3,5-DMPC) (5 μM)

Cellulose-tris-(3,5-DMPC) (3 μM)

Myclobutanil

Cellulose-tris-(3,5-DMPC) (5 μM)

Cellulose-tris-(3,5-DMPC) (3 μM)

Mobile phase (v/v)

k1

k2

α

Rs

Methanol/water (80–20)

2.93

—

1.62

9.56

Methanol/water (90–10)

1.16

—

1.75

9.42

Methanol/water (70–30)

7.72

—

1.49

6.51

Methanol/water (80–20)

2.78

—

1.55

5.90 5.19

Methanol/water (90–10)

1.19

—

1.67

Acetonitrile/water (50–50)

4.20

—

1.22

7.58

Acetonitrile/water (70–30)

1.34

—

1.05

1.18

Acetonitrile/water (90–10)

0.61

—

1.06

1.00

Acetonitrile/water (50–50)

5.29

—

1.05

1.12

Acetonitrile/water (70–30)

—

—

—

—

Acetonitrile/water (90–10)

—

—

—

—

Methanol/water (70–30)

6.57

—

1.18

3.87

Methanol/water (80–20)

2.31

—

1.17

3.20

Methanol/water (90–10)

0.95

—

1.17

2.29

Methanol/water (70–30)

5.95

—

1.15

2.54

Methanol/water (80–20)

2.23

—

1.15

1.68

Methanol/water (90–10)

0.98

—

1.15

1.15

Acetonitrile/water (50–50)

3.44

—

1.42

9.99 7.58

Acetonitrile/water (70–30)

0.88

—

1.44

Acetonitrile/water (90–10)

0.38

—

1.51

5.10

Acetonitrile/water (50–50)

3.69

—

1.42

7.72

Acetonitrile/water (70–30)

0.9

—

1.44

4.28

References

2 Chiral Separation of Some Classes of Pesticides by HPLC Method

Table 2.2 (Continued)

Cellulose-tris-(3,5-DMPC) (5 μM)

Fenbuconazole

Cellulose-tris-(3,5-DMPC) (5 μM)

Cellulose-tris-(3,5-DMPC) (3 μM)

Cellulose-tris-(3,5-DMPC) (5 μM)

Cellulose-tris-(3,5-DMPC) (3 μM)

Triadimefon

Cellulose-tris-(3,5-DMPC) (5 μM)

Cellulose-tris-(3,5-DMPC) (3 μM)

0.42

—

1.45

2.85

5.07

—

1.26

4.47

Methanol/water (80–20)

1.91

—

1.31

4.75

Methanol/water (90–10)

0.87

—

1.41

4.91

Methanol/water (70–30)

4.53

—

1.24

3.06

Methanol/water (80–20)

1.83

—

1.27

2.43

Methanol/water (90–10)

0.89

—

1.37

2.25

Acetonitrile/water (50–50)

6.86

—

1.33

9.03

Acetonitrile/water (70–30)

1.48

—

1.34

7.38

Acetonitrile/water (90–10)

0.56

—

1.37

4.79

Acetonitrile/water (50–50)

7.42

—

1.33

7.48

Acetonitrile/water (70–30)

1.49

—

1.34

4.46

Acetonitrile/water (90–10)

0.59

—

1.34

2.80

Methanol/water (70–30)

16.1

—

1.21

3.96

Methanol/water (80–20)

5.09

—

1.24

4.35

Methanol/water (90–10)

1.97

—

1.29

4.80

Methanol/water (70–30)

13.1

—

1.21

3.23

Methanol/water (80–20)

4.76

—

1.23

3.17

Methanol/water (90–10)

1.96

—

1.28

2.89

Acetonitrile/water (50–50)

2.72

—

1.15

3.82

Acetonitrile/water (70–30)

0.63

—

1.17

2.43

Acetonitrile/water (90–10)

0.22

—

1.20

1.45

Acetonitrile/water (50–50)

2.87

—

1.16

2.67

Acetonitrile/water (70–30)

0.65

—

1.17

1.26

Acetonitrile/water (90–10)

—

—

—

— (continued)

2.4 Separation of Enantiomers of Pesticides by HPLC in Normal and Reversed Phase Modes

Cellulose-tris-(3,5-DMPC) (3 μM)

Acetonitrile/water (90–10) Methanol/water (70–30)

351

352

Pesticides

Column/CSPs Cellulose-tris-(3,5-DMPC) (5 μM)

Cellulose-tris-(3,5-DMPC) (3 μM)

Mobile phase (v/v)

k1

k2

α

Rs

Methanol/water (70–30)

3.70

—

1.29

5.15

Methanol/water (80–20)

1.28

—

1.29

4.13

Methanol/water (90–10)

0.51

—

1.28

2.73

Methanol/water (70–30)

3.22

—

1.27

2.66

Methanol/water (80–20)

1.26

—

1.26

1.59

Methanol/water (90–10)

0.55

—

1.23

1.00

DMPC: dimethylphenylcarbamate; PEC: phenylethylcarbamate; PC: phenylcarbamate.

References

2 Chiral Separation of Some Classes of Pesticides by HPLC Method

Table 2.2 (Continued)

2.4 Separation of Enantiomers of Pesticides by HPLC in Normal and Reversed Phase Modes

Figure 2.3 Effect of the concentration of isopropanol (IPA) on chiral resolution of cischlordane and trans-chlordane on Chiralcel OD column. (a) Hexane–isopropanol (97:3, v/v), (b) hexane–isopropanol (99:1, v/v), and (c) hexane (100%) [62].

2-dichloroethane (500:1, v/v) and hexane-1,2-dichloroethane-ethanol (500:10:0.05, v/v/v), respectively. Lin et al. [66] performed the chiral separation of methamidophos on Chiralcel OD column with the mobile phase of n-hexaneisopropanol (80:20, v/v) at the 0.5 ml/min flow rate on OR and CD detectors and confirmed that R-(+)-methamidophos eluted before S-( )-methamidophos at 230 nm. Similarly, the same group of researchers also studied the chiral separation of fosthiazate that possesses two stereogenic centers on Chiralpak AD column with a mobile phase consisting of n-hexane-ethanol (95:5, v/v) and the 1.0 ml/min flow rate. The identification of peaks has been confirmed with OR and CD detectors at 230 nm and the values of capacity, separation, and resolution factors varies with a change of concentration of ethanol as shown in Table 2.3 [67]. Furthermore, Xu et al. [68] carried out the enantiomeric resolution of pyrethroid insecticides (λ-cyhalothrin) with Chiralpak AD (amylase tris [3,5-dimethylphenyl carbamate]), Chiralpak AS (amylose tris[(S)-1-phenyl carbamate]), Chiralcel OD (cellulose tris[3,5-dimethylphenyl carbamate]), and Chiralcel OJ (cellulose tris[4-methyl benzoate]) chiral stationary phases. In this study, they concluded that all the chiral stationary phases are good for the resolution when different ratios of eluting solvents are used as shown in Table 2.2 and confirmed that 5% ethanol is a good modifier in all the cases examined. They also studied the effect of temperature on the chiral resolution with two different

353

354

Effect of concentration of ethanol on the chiral resolution of R/S methamidophos [65,73].

Mobile phase (hexane-ethanol)

Capacity factor (k)

Separation factor (α)

Resolution (Rs)

k1

k2

k3

k4

α12

α23

α34

α14

α13

α24

Rs12

Rs23

Rs34

(95:05, v/v)

2.69

3.10

8.60

13.82

1.15

2.77

1.61

5.13

3.19

4.45

1.56

14.85

9.21

(90:10, v/v)

1.37

1.55

4.16

6.73

1.13

2.68

1.62

4.92

3.04

4.33

1.12

12.24

8.05

(85:15, v/v)

0.90

1.01

2.62

4.25

1.13

2.59

1.62

4.73

2.91

4.2

0.92

9.99

7.58

2 Chiral Separation of Some Classes of Pesticides by HPLC Method

Table 2.3

2.4 Separation of Enantiomers of Pesticides by HPLC in Normal and Reversed Phase Modes

Figure 2.4 Effect of temperature on chiral (99:1, v/v), 0.40 ml/min; (b) Chiralpak amylose separation of λ-cyhalothrin with two chiral sta- tris-([S]-α-methylbenzyl-carbamate), n-hexanetionary phases. (a) Chiralpak amylose tris-(3,5- ethanol (97.5:2.5, v/v), 0.040 ml/min [67]. dimethylphenyl-carbamate), n-hexane-ethanol

columns as shown in Figure 2.4 and confirmed that 20 °C is the optimum temperature for the baseline separation. Li et al. [69] developed methods for the chiral resolution of five organophosphorus compounds (Figure 2.5) on Chiralpak AD, Chiralpak AS, Chiralcel OD, and Chiralcel OJ columns. The baseline separations of all these compounds were obtained on Chiralpak AD column by using different ratios of hexane-ethanol and hexane-isopropanol as the mobile phase. Compound 1 was separated with hexane-ethanol (90:10, v/v), while compounds 2, 3, 4, and 5 had eluted with hexane-isopropanol (90 : 10, v/v) and (95: 5, v/v). Zhou et al. [70] also studied the enantiomeric resolution of salithion on different chiral stationary phases (Chiralcel OD, Chiralcel OJ, and Chiralpak AD) with different concentrations of isopropanol with hexane at the 1.0 ml/min flow rate, and the best separation was observed on Chiralpak AD column with hexane-isopropanol (99.5/0.5, v/v), and various chromatographic values measured are shown in Table 2.2. The same group of scholars resolved a long series of triazole fungicides (hexaconazole, triadimefon, tebuconazole, diniconazole, flutriafol, propiconazole, and difenoconazole) on Chiralcel OD and Chiralcel OJ columns. These authors reported that four compounds, namely, hexaconazole, triadimefon, tebuconazole, and diniconazole had been separated on Chiralcel OD column. The enantiomers of flutriafol had been obtained by changing the mobile

355

356

2 Chiral Separation of Some Classes of Pesticides by HPLC Method

R1

O

O

Y

P R2

O

O

O

R3 X

Organophosphorous Compound 1 2 3 4 5

R1 C2H5 CH3 CH3 CH3 CH3

R2 C2H5 CH3 CH3 CH3 CH3

R3

X

CH3 2-Cl C2H5 2-Cl CH3 – CH3 2-Cl CH3 –

Y 4-Cl 4-Cl 4-Cl – 4-F

Figure 2.5 Chemical structure of organophosphorus compounds [68].

phase from hexane-2-propanol (90:10, v/v) to hexane-ethanol (90:10, v/v). In the case of propiconazole, only three enantiomers could be separated on Chiralcel OD column using hexane-2-propanol (90:10, v/v) with the flow rate of 0.6 ml/ min at 15 °C. A satisfactory resolution of difenoconazole was achieved on Chiralcel OJ column using hexane-ethanol (90:10, v/v) as the mobile phase. They had also studied the effects of the temperature with linear Van’t Hoff plots at 10–35 °C, and concluded that the enantiomers of these triazole fungicides, with the exceptions of diniconazole and triadimefon, could be separated by altering the temperature [71]. Zhang et al. [72] resolved the enantiomers of metalaxyl and metalaxyl acid on Chiralcel OJ-H [cellulose-tris(4-methylbenzoate)] column with n-hexane-2-propanol-acetic acid (95:5:0.1, v/v/v) at the 0.5 ml/min flow rate. Zhang et al. [73] carried out the chiral separation of three nematicide fosthiazate insecticides with three different polysaccharide chiral stationary phases (Chiralcel OD-H, Chiralpak AD-H, and Chiralpak IB) by HPLC and SFC. In this study, these researchers also observed the effects of the temperature and the organic modifiers on the resolution of analytes and concluded that the best separation of all three analytes was achieved using Chiralcel OD-H column with different ratios of hexane–ethanol as the eluting solvents at the 1.0 ml/min flow rate as shown in Table 2.2. In the case of Chiralcel AD-H and Chiralpak IB columns, all the analytes did not resolve under good conditions with the exception of two compounds that were resolved on both chiral stationary phases. Emerick et al. [74] performed the enantioseparation of organophosphorus compound (methamidophos) on four different polysaccharide CSPs – [amylose tris(S)1-phenylethyl carbamate)] as (CSP-1), amylose tris(3,5-dimethylphenylcarbamate) as (CSP-2), cellulose tris(3,5-dimethylphenylcarbamate) as (CSP-3), and the Chiralcel OD [cellulose tris(3,5-dimethylphenylcarbamate)] as (CSP-4) with different compositions of hexane-2-propanol as the mobile phase. A mixture of n-hexane-2-propanol (80:20, v/v) was used initially and the chromatographic parameters k, α, and Rs obtained for CSP-1 and CSP-3 indicated poor resolution

2.4 Separation of Enantiomers of Pesticides by HPLC in Normal and Reversed Phase Modes

with Rs less than 1.0. Furthermore, different ratios of 2-propanol (30, 10, 8, and 5%) were also evaluated with CSP-1, CSP-3, and CSP-4, respectively. Under the chromatographic conditions studied, the highest resolutions Rs were 1.51 and 1.56 when n-hexane-2-propanol (95: 5, v/v) and (98: 2, v/v) was used with CSP-3 and CSP-4, respectively. The replacement of 2-propranol by ethanol was also studied and it was concluded that the use of ethanol at 10–20% reduces the retention factor with the poor enantiomeric separation of methamidophos. Similarly, when of n-heptane was used instead of n-hexane with 1% acid additives, no improvements in the chromatographic parameters for the enantioseparation of methamidophos was observed. Sun et al. [75] studied the chiral resolution of uniconalzole and their enantioselective effects on the growth of rice seedlings and cyanobacteria. These authors performed the chiral separation on Chiralpak AD column using n-hexane-isopropanol (85:15, v/v) as the mobile phase at the 1.0 ml/min flow rate and the peaks were confirmed with CD and OR detectors at 250 nm as shown in Figure 2.6. Recently, Lao and Gan [76] studied the enantioselective degradation of warfarin in soil samples and resolved the compound with triproline chiral stationary phase and a fluorescence detector. The eluting solvent was n-hexane(0.1% TFA)-2-propanol with a ratio of (92:8, v/v) and (96:4, v/v). The baseline separation was achieved using the latter mobile phase.

Figure 2.6 Chiral separation of uniconazole on Chiralpak AD column with CD and OR detectors [74].

357

358

2 Chiral Separation of Some Classes of Pesticides by HPLC Method

Wang et al. [77] synthesized the chiral selector (amylose tris-(S)-1-phenylethyl carbamate) for the enantiomeric resolution of 32 pesticides, in which 10 pesticides showed an interaction with the CSP used, and resolved these pesticides using different compositions of hexane-isopropanol as mobile phases. They also studied the effects of temperature on the resolution and concluded that as the temperature increases, the capacity and separation factors decrease as shown in Table 2.4. Similarly, Tan et al. [78] synthesized a new chiral stationary phase

Table 2.4 Effect of temperature on capacity, separation, and resolution factors of pesticides [76]. Pesticides

Temperature °C

Capacity factor k1

Metalaxyl

Myclobutanil

Fenoxapropethyl

Imazalil

Malathion

Triadimefon

Separation factor α

Resolution (Rs)

k2

0

4.04

6.78

1.68

1.67

10

3.31

5.42

1.64

2.22

20

3.01

4.66

1.55

2.51

30

2.49

3.73

1.50

2.36

40

2.22

3.19

1.43

2.22

0

24.53

31.53

1.29

1.27

10

20.58

25.91

1.26

1.39

20

17.49

21.47

1.24

1.49

30

15.70

19.10

1.22

1.33

40

14.06

16.85

1.20

1.18

0

9.42

12.29

1.30

0.83

10

6.79

8.92

1.28

0.80

20

6.07

7.53

1.24

1.02

30

4.88

5.86

1.20

1.01

40

4.34

5.15

1.19

1.04

0

9.97

11.25

1.30

0.81

10

8.61

9.69

1.13

0.89

20

7.56

8.44

1.12

0.86

30

6.47

7.16

1.11

0.85

40

5.59

6.12

1.09

0.76

0

6.22

7.10

1.14

0.84

10

5.05

5.70

1.13

0.80

20

4.53

5.02

1.11

0.76

30

3.72

4.03

1.08

0.58

40

3.24

3.41

1.05

0.44

0

8.29

12.32

1.49

1.63

10

6.83

9.81

1.44

1.74

2.4 Separation of Enantiomers of Pesticides by HPLC in Normal and Reversed Phase Modes

Ethofumesate

Fipronil

Napropamide

Paclobutrazol

transpermethrin

cispermethrin

20

5.97

8.47

1.42

1.84

30

5.20

7.07

1.36

1.50

40

4.83

6.40

1.33

1.36

0

14.01

15.02

1.07

0.73

10

11.02

11.78

1.07

0.62

20

9.95

10.58

1.06

0.63

30

8.21

8.66

1.06

0.58

40

7.21

7.54

1.05

0.46 0.84

0

10.11

13.09

1.29

10

8.33

10.68

1.28

0.92

20

7.72

9.88

1.28

1.35

30

6.00

7.64

1.27

1.61

40

4.97

6.28

1.26

1.81

0

11.16

13.38

1.20

0.97

10

9.22

10.85

1.18

1.11

20

7.65

8.76

1.14

1.14

30

6.77

7.62

1.13

1.04

40

5.54

6.18

1.12

1.07

0

3.98

4.87

1.22

1.14

10

3.81

4.64

1.22

1.26

20

3.57

4.29

1.20

1.30

30

3.29

3.88

1.18

1.34

40

3.06

3.54

1.16

1.31

15

14.16

15.67

1.11

1.18

22

12.01

14.60

1.21

1.26

23

11.01

12.46

1.13

1.34

25

9.67

11.19

1.16

1.32

30

7.64

8.88

1.16

1.12

35

6.64

7.64

1.15

1.02

38

4.36

4.99

1.14

0.93

43

3.27

3.74

1.14

0.78

15

23.91

26.34

1.10

2.02

22

23.96

26.47

1.10

2.20

23

20.57

22.76

1.11

2.21

25

19.51

21.75

1.11

2.27

30

17.78

20.27

1.14

2.25

35

16.83

19.68

1.17

2.31

38

11.38

14.14

1.24

2.00

43

7.82

10.26

1.31

1.87

359

360

2 Chiral Separation of Some Classes of Pesticides by HPLC Method

[(S)-valine-(R)-1-phenyl-2-(4-methylphenyl)ethylamine] for the resolution of pyrethroid insecticides (fenpropathrin, fenvalerate, brofluthrinate, cypermethrin, and cyfluthrin). They also compared it with the resolution on Pirkle-type 1-A chiral stationary phase and achieved a satisfactory baseline separation on the synthesized chiral selector. Oda et al. [45] studied the chiral resolution of warfarin on avidin and ovomucoid CSPs using methanol, ethanol, propanol, and acetonitrile as organic modifiers. In general, the chiral recognition behavior of these modifiers on avidin and ovomucoid CSPs had been of the following order: methanol>ethanol>propanol>acetonitrile. Recently, Pan et al. [79] compared the enantioresolution of seven triazole fungicides (tebuconazole, hexaconazole, myclobutanil, diniconazole, uniconazole, paclobutrazol, and triadimenol) on the Pirkle-type (S,S)-Whelk O1 chiral column and four different cellulose derivative columns, namely, cellulose tribenzoate (CTB), cellulose tris-(4-methylbenzoate) (CTMB), cellulose triphenylcarbamate (CTPC), and cellulose tris(3,5-dimethylphenyl carbamate) (CDMPC), in the normal-phase mode with ethanol, n-propanol, isopropanol, and n-butanol, respectively, as the polar modifier in hexane mobile phase, and concluded that only two triazole fungicides (hexaconazole and triadimenol) had been resolved with different compositions of alcohols in hexane. Among all these cellulose derivative columns, the best separation was achieved on CDMPC and no resolution was achieved on CTB. Some other pesticides were also resolved with different CSPs and mobile phases, and many scholars compared the enantiomeric resolution with various organic mobile phases. The details of the chromatographic parameters are presented in Table 2.2 [77,80,81]. 2.4.2 Enantiomeric Separation of Pesticides in a Reversed Phase Mode

The enantiomeric resolution of pesticides using reversed phase chromatography has been carried out using aqueous mobile phases with different nonpolar CSPs. Among all chiral stationary phases, the polysaccharide chiral stationary phases have been widely used as shown in the literature. The majority of the researchers performed the chiral separation of pesticides on polysaccharide CSPs using polar organic modifiers with water or buffers. Ali and Aboul-Enein [60] reported the resolutions of o,p-DDT and o,p-DDD enantiomers on Chiralpak AD-RH, Chiralcel OD-RH, and Chiralcel OJ-R chiral stationary phases. The following mobile phases were used: acetonitrile–water (50:50, v/v) and acetonitrile-2-propanol (50:50, v/v) at the 1.0 ml/min flow rate. The detection was carried out at 220 nm for both pesticides. Tian et al. [82] developed an HPLC method for the enantiomeric resolution of 20 chiral pesticides on cellulose tris-3,5-dimethyl carbamate (CDMPC) chiral stationary phase under reversed phase mode using different compositions of acetonitrile–water and methanol–water as eluents with the 0.8 ml/min flow rate and the detection at 210 and 230 nm, respectively. The parameters of these findings are presented in Table 2.2. Qiu et al. [83] carried out the chiral separation of nine triazole fungicides (hexaconazole, flutriafol,

2.4 Separation of Enantiomers of Pesticides by HPLC in Normal and Reversed Phase Modes

Figure 2.7 HPLC chromatogram of nine triazole fungicides. Mobile phases: methanol/ water (75:25) on 3.0 μm column for flutriafol, tetraconazole, penconazole, myclobutanil,

fenbuconazole, and triadimefon; acetonirile/ water (60:40) on 3.0 μm column for epoxiconazole and acetonitrile/water (60:40) on 5.0 μm column for diniconazole [82].

diniconazole, tetraconazole, epoxiconazole, penconazole, myclobutanil, fenbuconazole, and triadimefon) on the Lux Cellulose-1 columns with different particle sizes (3.0 and 5.0 μm) and with cellulose-tris-(3,5-dimethylphenylcarbamate). In this analysis, the scientists used various ratios of acetonitrile and methanol with water as the organic modifier. Moreover, the effects of temperature on the chiral resolution were also studied. The best separation of all fungicides with different chromatographic conditions is shown in Figure 2.7. Examining this figure makes it clear that the polysaccharide chiral stationary phase is good for the resolution of these fungicides using different ratios of acetonitrile and methanol with water. Li et al. [84] also studied the enantiomeric separation of eight triazole fungicides (tetraconazole, fenbuconazole, epoxiconazole, diniconazole, hexaconazole, triadimefon, paclobutrazol, and myclobutanil) in soil samples on polysaccharide chiral stationary phases with the aqueous mobile phase, and a mass spectrometer (MS) detector was used to distinguish the enantiomers. These authors concluded that the best chromatographic separation of all 16 enantiomers of 8 triazole fungicides had been achieved using Chiralcel OD-RH column, with a mixture of

361

362

2 Chiral Separation of Some Classes of Pesticides by HPLC Method

acetonitrile-2 mM ammonium acetate in water (55:45, v/v) as the mobile phase. The baseline separation of all peaks was achieved with more than 1.0 resolution factor Rs. These authors also tested methanol as the organic modifier instead of acetonitrile in reversed phase liquid chromatography. However, unsatisfactory results were obtained. Different concentrations of the ammonium acetate buffer (0.5, 1, 2, 5, and 10 mM) were used to obtain better peaks A good response of the MS/MS detector was observed at the 2 mM buffer concentration. Moreover, other chromatographic conditions were optimized and the best results were obtained at the 0.45 ml/min flow rate at 25 °C. Dong et al. [85] performed the enantioselective analysis of myclobutanil (triazole fungicides) in cucumber and soil sample through an LC–MS method. Using this technique, the researchers used the cellulose-based column (Chiralcel OD-RH, 150 mm × 4.6 mm i.d., 5 μm particle size) and the two amylose-based columns (Chiralpak AD-RH and ASRH, 150 mm × 4.6 mm i.d., 5 μm particle size) with a variety of reversed mobilephase combinations, and the best chromatographic separation of myclobutanil enantiomers was achieved on Chiralcel OD-RH column with acetonitrile–water (70:30, v/v) as the mobile phase and the 0.5 ml/min flow rate at 40 °C. Vetter et al. [86,87] described the separation of enantiomers of toxaphene on the tert-butyl-dimethylsilylated-β-CD-based CSPs using HPLC. Blessington and Crabb [88] performed the chiral separation of aryloxypropionate herbicides on the Chiral AGP column. In this study, the phosphate buffer (10 mM, pH 6)-2propanol (94:4, v/v) was used as the mobile phase with the detection at 240 nm. Recently, Shishovska and Trajkovska [89] carried out the chiral separation of permethrin enantiomers using the β-cyclodextrin-based chiral stationary phase. The separation of all enantiomers is shown in Figure 2.8. The retention times for the trans-enantiomers were 19.8 and 22.7 min, and for the cis-enantiomers 38.2 and 42.3 min. They also studied the effects of temperature on the chiral resolution and observed that at high temperatures, the values of chromatographic parameters k, α, and Rs decrease with an increase of temperature as shown in Table 2.4. Schneiderheinze, Armstrong, and Berthod [90] studied the chiral resolution of phenoxyalkanoic acid herbicides in plant and soil samples on Chirobiotic T CSPs with methanol-1% triethylammonium acetate as the mobile

Figure 2.8 Chiral separation of permethrin enantiomers on β-cyclodextrin-based chiral stationary phase [88].

2.4 Separation of Enantiomers of Pesticides by HPLC in Normal and Reversed Phase Modes

phase at a pH of 4.1 (60:40, v/v), and the detection was carried out by the chiroptical detector. Möller et al. [91] resolved the enantiomers of α-HCH in brains of seals on the Chiraldex column using methanol–water (75:25, v/v) as the eluting solvent. Dondi et al. [92] used terguride-based CSP for the enantiomeric resolution of chrysanthemic acid [2,2-dimethyl-3-(2-methylpropenyl)-cyclopropanecarboxylic acid] and its halogen-substituted analogs. In this analysis, UV diode array and chiroptical detectors were used for the identification of enantiomers, and it was observed that isomers with (1R) configuration always elute before those with (1 S) configuration. The elution sequence of cis- and trans-isomers has been strongly affected by the mobile-phase pH, whereas the enantioselectivity has remained the same. Ludwig, Gunkel, and Hühnerfuss [93] carried out the enantiomeric analysis of 2-(2,4-dichlorophenoxy) propionic acid (dichlorprop) and its degradation products in a marine microbial community on Chiral AGP column with water–2propanol–phosphate buffer (10 mM, pH 4.85) (94:4: 2, v/v/v) as mobile phase at 230 nm. Armstrong et al. [94] developed a method for the enantioseparation of warfarin, coumachlor, coumafuryl, bulan, crufomate, fonofos, anacymidol, napropamide, and 2-(3-chlorophenoxy)-propionamide pollutants on Chiral AGP, and other cyclodextrin-based CSPs. Chu and Wainer [95] separated the enantiomers of warfarin in serum samples using chiral HPLC. Weber, Kreuzig, and Bahadir [96] used the permethylated β-CD CSP for the chiral separation of phenoxypropionates. Recently, Malakova et al. [97] studied the chiral resolution of two enantiomers, (R)-(+)-warfarin and (S)-( )-warfarin, and their determination in hepatoma HepG2 cell line using a glycopeptide-based chiral stationary phase with the acetonitrile–methanol–ammonium acetate buffer (10.0 mM, pH 4.1) (31:5:64, v/v) mobile phase at 1.2 ml/min flow rate, and the LOD was 0.121 μmol/l for (S)-warfarin and 0.109 μmol/l for (R)-warfarin. Guillén-Casla and Pérez-arribas [98] analyzed the enantiomeric separation of mixture of aryloxyphenoxypropionic herbicides (diclofop acid and diclofop-methyl) with α1acid glycoprotein CSP using one- and two-dimensional liquid chromatographic methods. In these methods, the achiral separation of samples containing diclofop acid and diclofop-methyl racemate mixtures was carried out by injecting 20.0 μl of sample on the C18-LUNA column. The isocratic mobile phase containing methanol–phosphate buffer (30 mM, pH 7) (73:23, v/v) at the 1.0 ml/min flow rate was used. The chiral resolution of diclofop acid enantiomers and the diclofop-methyl enantiomers was performed with the phosphate buffer (70 mM, pH 7.0)-2-propanol (99.5:0.5, v/v) and the phosphate buffer (30 mM, pH 7.0)-2propanol (91:9, v/v), respectively, at the 0.8 ml/min flow rate. Recently, the same group of researchers carried out the chiral separation of aryloxyphenoxypropionic acid herbicides (fluazifop-butyl, quizalofop-ethyl, and mefenpyr-diethyl) on the α1-acid glycoprotein chiral stationary phase. In this study, the optimization of the chromatographic conditions was performed through the factorial experimental design with the phosphate buffer (pH 6.5–7.0) and propanol (5–10%) at the 15–25 °C column temperature. Mathematical deconvolution was also studied to determine peak areas and to calculate Rs, the enantiomeric ratio

363

364

2 Chiral Separation of Some Classes of Pesticides by HPLC Method

(ER), and the enantiomeric fraction (EF), and it was concluded that peak deconvolution provided a simple, effective, and reproducible method for the herbicide determination in soil matrices at low levels of micrograms per gram [99]. Mano et al. [100] used the flavoprotein conjugated silica CSP for the chiral resolution of warfarin. They also studied the effects of pH on chiral resolution and observed that the best results were obtained at pH of 4.0–4.8. The effects of different compositions of the mobile phases and column particle size on capacity, separation, and resolution factors under the reversed phase mode using different chiral stationary phases is also presented in Table 2.2 [83].

2.5 Conclusions

The introduction of racemic pesticides into our environment is a serious issue as achiral analyses do not determine the exact dose and toxicity of pesticides. Moreover, different toxicities of the two enantiomers confuse farmers when they measure the dose lethal for pests. Sometimes, the pest control remains ineffective as a result of the illusive dose of racemic pesticides used. Furthermore, it is not possible to determine the impact of the chiral pesticide degradation product without considering the chiral aspect of racemic pesticides. Chiral HPLC methods have proved their effectiveness in the analysis of racemic pesticides and resolution of the corresponding enantiomers. Accordingly, HPLC should be used to study the implications of chiral pesticides before they are introduced into the environment. The enhanced knowledge of chiral pesticides may help determine the required dose and may also save the environment from an unnecessary load of pesticides. Agricultural scientists should think in terms of chiral pesticides and ought to provide information to farmers.

References 1 Ulrich, E.M., Morrison, C.N., Goldsmith,

2 3

4

5

M.R., and Foreman, W.T. (2012) Chiral pesticides: identification, description, and environmental implications. Rev. Environ. Conta. Toxicol., 217, 1–74. http://ipmworld.umn.edu/chapters/ware .htm) (accessed January 12, 2013). (http://www.apsnet.org/edcenter/ intropp/topics/Pages/Fungicides.aspx) (accessed January 12, 2013). (www.agriinfo.in/default.aspx? page=topic&superid=1&topicid=804) (accessed January 12, 2013). Richardson, S.D. and Ternes, T.A. (2005) Water analysis: emerging contaminants

and current issues. Anal. Chem., 77, 3807–3838. 6 Kurihara, N., Miyamoto, J., Paulson, G.D., Zeeh, B., Skidmore, M.W., Hollingworth, R.M., and Kuiper, H.A. (1997) Chirality in synthetic agrochemicals: bioactivity and safety consideration. Pure Appl. Chem., 69, 2007–2025. 7 Ward, T.J. and Hamburg, D.M. (2004) Chiral separations. Anal. Chem., 76, 4635–4644. 8 Gasparrini, F., Misiti, D., and Villani, C. (2001) High-performance liquid chromatography chiral stationary

References

9

10 11

12

13

14

15

16

17

18

19

20

21

phases based on low-molecular-mass selectors. J. Chromatogr. A, 922, 35–45. Garrison, A.W. (2006) Probing the enantioselectivity of chiral pesticides. Environ. Sci. Technol., 40, 16–23. Sekhon, B.S. (2009) Chiral pesticides. J. Pestic. Sci., 34, 1–12. Dong, F.S., Liu, X.G., Zheng, Y.Q., Cao, Q., and Li, C.J. (2010) Stereoselective degradation of fungicide triadimenol in cucumber plants. Chirality, 22, 292–298. Ye, J., Zhao, M.R., Liu, J., and Liu, W.P. (2010) Enantioselectivity in environmental risk assessment of modern chiral pesticides. Environ. Pollut., 158, 2371–2383. Liu, W.P., Ye, J., and Jin, M.Q. (2009) Enantioselective phytoeffects of chiral pesticides. J. Agric. Food Chem., 57, 2087–2095. Wang, Y., Tai, K., and Yen, J. (2004) Separation, bioactivity, and dissipation of enantiomers of the organophosphorus insecticide fenamiphos. Ecotoxicol. Environ. Saf., 57, 346–353. Aboul-Enein, H.Y. and Ali, I. (2003) Chiral Separations by Liquid Chromatography and Related Technologies, Marcel Dekker, New York, NY. Beesley, T.E. and Scott, R.P.W. (1998) Chiral Chromatography, John Wiley & Sons, Inc., New York. Aboul-Enein, H.Y. and Wainer, I.W. (eds) (1997) The Impact of Stereochemistry on Drug Development and Use: Chemical Analysis, vol. 142, John Wiley & Sons, Inc., New York. Subramanian, G. (ed.) (1994) A Practical Approach to Chiral Separations by Liquid Chromatography, Wiley-VCH Verlag GmbH, Weinheim. Allenmark, S. (1991) Chromatographic Enantioseparation, Methods and Applications, 2nd edn, Ellis Horwood, New York. Krstulovic, A.M. (1989) Chiral Separations by HPLC: Applications to Pharmaceutical Compounds, Ellis Horwood, New York. Huhnerfuss, H. and Kallenborn, R. (1992) Chromatographic separation of marine pollutants. J. Chromatogr. A, 580, 191.

22 Kallenborn, R. and Hühnerfuss, H. (2001)

23

24

25

26

27

28

29

30

31

32

Chiral environmental pollutants, in Trace Analysis and Toxicology, Springer, Berlin. Ali, I. and Aboul-Enein, H.Y. (2004) Chiral Pollutants: Distribution, Toxicity and Analysis by Chromatography and Capillary Electrophoresis, John Wiley & Sons, Ltd., Chichester, UK. Pfeiffer, C.C. (1956) Optical isomerism and pharmacological action: a generalization. Science, 124, 29. Pirkle, W.H. and Pochapsky, T.C. (1989) Considerations of chiral recognition relevant to the liquid chromatography separation of enantiomers. Chem. Rev., 89, 347–362. Groombridge, J.J., Jones, C.G., Bruford, M.W., and Nichols, R.A. (2000) Ghost alleles in the Mauritius kestrel. Nature, 403, 616. Rogozhin, S. and Davankov, V.A. (1970) German Patent 1932190, 1969, cited in Chem. Abstr., 72, 90875c. Rogozhin, S.V. and Davankov, V.A. (1971) Ligand chromatography on asymmetric complex-forming sorbents as a new method for resolution of racemates. J. Chem. Soc. Chem. Commun., 490. Davankov, V.A. (1994) Chiral selectors with chelating properties in liquid chromatography: fundamental reflections and selective review of recent developments. J. Chromatogr. A, 666, 55–76. Han, S.M. and Armstrong, D.W. (1989) HPLC separation of enantiomers and other isomers with cyclodextrin bonded phases: rule for chiral recognition, in Chiral Separations by HPLC (ed. A.M. Krstulovic), Ellis Horwood, Chichester, p. 208. Stalcup, A.M. (1994) Cyclodextrinbonded chiral stationary phases in enantiomer separations, in A Practical Approach to Chiral Separations by Liquid Chromatography (ed. G. Subramanian), Wiley-VCH Verlag GmbH, Weinheim, p. 95. Feitsma, K.G., Bosmann, J., Drenth, B.F.H., and De Zeeuw, R.A. (1985) A study of the separation of enantiomers of some aromatic carboxylic acids by high-

365

366

2 Chiral Separation of Some Classes of Pesticides by HPLC Method

33

34

35

36

37

38

39

40

41

performance liquid chromatography on a β-cyclodextrin-bonded stationary phase. J. Chromatogr., 333, 59–68. Han, S.M. (1997) Direct enantiomeric separations by high performance liquid chromatography using cyclodextrins. Biomed. Chromatogr., 11, 259–271. Piperakis, S. Kakoulidou-Tsantili, A., and Parissi-Poulou, M. (1995) Solvent selectivity in chiral chromatography using a β-cyclodextrin-bonded phase. Chirality, 7, 257–266. Hargitai, T. and Okamoto, Y. (1993) Evaluation of 3,5-dimethylphenyl carbamoylated α-, β-, and γ-cyclodextrins as chiral stationary phases for HPLC. J. Liq. Chromatogr., 16, 843–858. Armstrong, D.W., Tang, Y., Chen, S., Zhou, Y., Bagwill, C., and Chen, J.R. (1994) Macrocyclic antibiotics as a new class of chiral selectors for liquidchromatography. Anal. Chem., 66, 1473–1484. Advanced Separation Technologies, Inc. (1999) Chirobiotic Handbook: A Guide to Using Macrocyclic Glycopeptide Bonded Phases for Chiral LC Separations, 2nd edn, Advanced Separation Technologies, Inc., Whippany, NJ. Armstrong, D.W., Rundlett, K.L., and Chen, J.R. (1994) Evaluation of the macrocyclic antibiotic vancomycin as a chiral selector for capillary electrophoresis. Chirality, 6, 496–509. Tesarova, E., Bosakova, Z., and Pacakova, V. (1999) Comparison of enantioselective separation of N-tert-butyloxycarbonyl amino acids and their non-blocked analogues on teicoplanin-based chiral stationary phase. J. Chromatogr. A, 838, 121–129. Ekborg-Ott, K.H., Liu, Y., and Armstrong, D.W. (1998) Highly enantioselective HPLC separations using the covalently bonded macrocyclic antibiotic, ristocetin A, chiral stationary phase. Chirality, 10, 434–483. Armstrong, D.W., Liu, Y., and EkborgOtt, K. (1995) A covalently bonded teicoplanin chiral stationary phase for HPLC enantioseparations. Chirality, 7, 474–497.

42 Peter, A., Torok, G., and Armstrong,

43

44

45

46

47

48

49

50

D.W. (1998) High-performance liquid chromatographic separation of enantiomers of unusual amino acids on a teicoplanin chiral stationary phase. J. Chromatogr. A, 793, 283–296. Aboul-Enein, H.Y. and Ali, I. (2002) Optimization strategies for HPLC enantioseparation of racemic drugs using polysaccharides and macro cyclic glycopeptide antibiotic chiral stationary phases. IL Farmaco, 57, 513–529. Allenmark, S. (1994) Separation of enanatiomers by protein-based chiral stationary phases by liquid chromatography, in A Practical Approach to Chiral Separations by Liquid Chromatography (ed. G. Subramanian), Wiley-VCH Verlag GmbH, Weinheim, pp. 183–216. Oda, Y., Mano, N., Asakawa, N., Yoshida, Y., Sato, T., and Nakagawa, T. (1993) Comparison of avidin and ovomucoid as chiral selectors for the resolution of drug enantiomers by highperformance liquid chromatography. Anal. Sci., 9, 221. Shinbo, T., Yamaguchi, T., Yanagishita, H., Kitamoto, D., Sakaki, K., and Sugiura, M. (1992) Improved crown ether-based chiral stationary phase. J. Chromatogr. A, 625, 101–108. Hyun, M.H., Jin, J.S., and Lee, W. (1998) Liquid chromatographic resolution of racemic amino acids and their derivatives on a new chiral stationary phase based on crown ether. J. Chromatogr. A, 822, 155–161. Hyun, M.H., Jin, J.S., Koo, H.J., and Lee, W. (1999) Liquid chromatographic resolution of racemic amines and amino alcohols on a chiral stationary phase derived from crown ether. J. Chromatogr. A, 837, 75–82. Hyun, M.H., Han, S.C., Jin., J.S., and Lee, W. (2000) Separation of the stereoisomers of racemic fluoroquinolone antibacterial agents on a crown-etherbased chiral HPLC stationary phase. Chromatographia, 52, 473–476. Feibush, B., Cohen, M.J., and Karger, B.L. (1983) The role of bonded phase composition on the ligand-exchange

References

51

52

53

54

55

56

57

58

59

chromatography of dansyl-D,L-amino acids. J. Chromatogr., 282, 3–26. Davankov, V.A., Bochkov, A.S., and Belov, Y.P. (1981) Ligand-exchange chromatography of racemates: XV. Resolution of α-amino acids on reversedphase silica gels coated with n-decyl-lhistidine. J. Chromatogr., 218, 547–557. Roumeliotis, P., Unger, K.K., Kurganov, A.A., and Davankov, V.A. (1983) Highperformance ligand-exchange chromatography of α-amino acid enantiomers: studies on monomerically bonded 3-(L-prolyl)- and 3-(Lhydroxyprolyl)propyl silicas. J. Chromatogr., 255, 51–66. Roumeliotis, P., Kurganov, A.A., and Davankov, V.A. (1983) Effect of the hydrophobic spacer in bonded [Cu(lhydroxyprolyl)alkyl]+ silicas on retention and enantioselectivity of α-amino acids in high-performance liquid chromatography. J. Chromatogr, 266, 439–450. Shieh, C.H., Karger, B.L., Gelber, L.R., and Feibush, B. (1987) Ligand exchange chromatography of amino alcohol enantiomers as Schiff bases. J. Chromatogr., 406, 343–352. Remelli, M., Fornasari, P., Dandi, F., and Pulidori, F. (1993) Dynamic columncoating procedure for chiral ligandexchange chromatography. Chromatographia, 37, 23–30. Perez, S. and Barcelo, D. (2008) Applications of LC–MS to quantitation and evaluation of the environmental fate of chiral drugs and their metabolites. Trends Anal. Chem., 27, 836–846. Ward, T.J. and Ward, K.D. (2010) Chiral separations: fundamental review 2010. Anal. Chem., 82, 4712–4722. Caccamese, S. and Principato, G. (2000) Separation of the four pairs of enantiomers of vincamine alkaloids by enantioselective high-performance liquid chromatography. J. Chromatogr. A, 893, 47–54. Ellington, J.J., Evans, J.J., Prickett, K.B., and Champion, W.L. Jr. (2001) Highperformance liquid chromatographic separation of the enantiomers of organophosphorus pesticides on

60

61

62

63

64

65

66

67

68

polysaccharide chiral stationary phases. J. Chromatogr. A, 928, 145–154. Ali, I. and Aboul-Enein, H.Y. (2002) Determination of chiral ratio of o,p-DDT and o,p-DDD pesticides on polysaccharides CSPs by HPLC under reversed phase mode. Environ. Toxicol., 17, 329–333. Li, Z.Y., Zhang, Z.C., Zhou, Q.L., Gao, R.Y., and Wang, Q.S. (2002) Fast and precise determination of phenthoate and its enantiomeric ratio in soil by the matrix solid-phase dispersion method and liquid chromatography. J. Chromatogr. A, 977, 17–25. Aboul-Enein, H.Y. and Ali, I. (2001) A comparison of chiral resolution of econazole, miconazole and sulconazole by HPLC using normal phase amylose CSPs. Fresenius J. Anal. Chem., 370, 951–955. Champion, W.L., Jr., Lee, J., Wayne, G.A., DiMarco, J.C., Matabe, A., and Prickett, K.B. (2004) Liquid chromatographic separation of the enantiomers of transchlordane, cis-chlordane, heptachlor, heptachlor epoxide and hexachlorocyclohexane with application to small-scale preparative separation. J. Chromatogr. A, 1024, 55–62. Wang, P., Jiang, S., Liu, D., Jia, G., Wang, Q., Wang, P., and Zhou, Z. (2005) Effect of alcohols and temperature on the direct chiral resolutions of fipronil, isocarbophos and carfentrazone-ethyl. Biomed. Chromatogr., 19, 454–458. Liu, W., Gan, J.J., and Qin, S. (2005) Separation and aquatic toxicity of enantiomers of synthetic pyrethroid insecticides. Chirality, 17, S127–S133. Lin, K., Zhou, S., Xu, C., and Liu, W. (2006) Enantiomeric resolution and biotoxicity of methamidophos. J. Agric. Food Chem., 54, 8134–8138. Lin., K., Zhang, F., Zhou, S., Liu, W., Gan, J., and Pan, Z. (2007) Stereoisomeric separation and toxicity of the nematicide fosthiazate. Environ. Toxicol. Chem., 26, 2339–2344. Xu, C., Wang, J., Liu, W., Shengg, D., Tu, Y., and Ma, Y. (2008) Separation and aquatic toxicity of enantiomers of the pyrethroid insecticide lambda-

367

368

2 Chiral Separation of Some Classes of Pesticides by HPLC Method

69

70

71

72

73

74

75

76

77

78

cyhalothrin. Environ. Toxicol. Chem., 27, 174–181. Li, L., Zhou, S., Zhao, M., Zhang, A., Peng, H., Tan, X., Lin, C., and He, A. (2008) Separation and aquatic toxicity of enantiomers of 1-(substituted phenoxyacetoxy)alkylphosphonate herbicides. Chirality, 20, 130–138. Zhou, S., Lin, K., Li, L., Jin, M., Ye, J., and Liu, W. (2009) Separation and toxicity of salithion enantiomers. Chirality, 21, 922–928. Zhou, Y., Li, L., Lin, K., Zhu, X., and Liu, W. (2009) Enantiomer separation of triazole fungicides by high-performance liquid chromatography. Chirality, 21, 421–427. Zhang, X., Xia, T., Chen, J., Huang, L., and Cai, X. (2010) Direct chiral resolution of metalaxyl and metabolite metalaxyl acid in aged mobile phases: the role of trace water. J. Agric. Food Chem., 58, 5004–5010. Zhang, C., Jin, L., Zhou, S., Zhang, Y., Feng, S., and Zhou, Q. (2011) Chiral separation of neonicotinoid insecticides by polysaccharide-type stationary phases using high-performance liquid chromatography and supercritical fluid chromatography. Chirality, 23, 215–221. Emerick, G.L., Oliveira, R., Belaz, K.R.A., Goncalves, M., and Deoliveira, G.H. (2012) Semipreparative enantioseparation of methamidophos by HPLC-UV and preliminary in vitro study of butyrylcholinesterase inhibition. Environ. Toxicol. & Chem., 31, 239–245. Sun, J., Zhang, A., Zhang, J., Xie, X., and Liu, W. (2012) Enantiomeric resolution and growth-retardant activity in rice seedlings of uniconazole. J. Agric. Food Chem., 60, 160–164. Lao, W. and Gan, J. (2012) Enantioselective degradation of warfarin in soils. Chirality, 24, 54–59. Wang, P., Liu, D., Lei, X., Jiang, S., and Zhou, Z. (2006) Enantiomeric separation of chiral pesticides by high-performance liquid chromatography on an amylose tris-(S)-1-phenylethylcarbamate chiral stationary phase. J. Sep. Sci., 29, 265–271. Tan, X., Hou, S., Jiang, J., and Wang, M. (2007) Preparation of a new chiral

79

80

81

82

83

84

85

stationary phase for HPLC based on the (R)-1-phenyl-2-(4-methylphenyl) ethylamine amide derivative of (S)-valine and 2-chloro-3,5-dinitrobenzoic acid: enantioseparation of amino acid derivatives and pyrethroid insecticides. J. Sep. Sci., 30, 1888–1892. Pan, C., Shen, B., Xu, B., Chen, J., and Xu, X. (2006) Comparative enantioseparation of seven triazole fungicides on (S,S)Whelk O1 and four differentcellulose derivative columns. J. Sep. Sci., 29, 2004–2011. Wang, P., Liu, D., Jiang, S., Gu, X., and Zhou, Z. (2007) The direct chiral separations of fungicide enantiomers on amylopectin based chiral stationary phase by HPLC. Chirality, 19, 114–119. Sun, J., Liu, J., Tu, W., and Xu, C. (2010) Separation and aquatic toxicity of enantiomers of the organophosphorus insecticide O-ethyl O-4-nitrophenyl phenylphosphonothioate (EPN). Chemosphere, 81, 1308–1313. Tian, Q., Lv, C., Wang, P., Ren, L., Qiu, J., Li, L., and Zhou, Z. (2007) Enantiomeric separation of chiral pesticides by high performance liquid chromatography on cellulosetris-3,5-dimethyl carbamate stationary phase under reversed phase conditions. J. Sep. Sci., 30, 310–321. Qiu, J., Dai, S., Zheng, C., Yang, S., Chai, T., and Bie, M. (2011) Enantiomeric separation of triazole fungicides with 3-lm and 5-lm particle chiral columns by reverse-phase high-performance liquid chromatography. Chirality, 23, 479–486. Li, Y., Dong, F., Liu, X., Xu, J., Li, J., Kong, Z., Chen, X., Liang, X., and Zheng, Y. (2012) Simultaneous enantioselective determination of triazole fungicides in soil and water by chiral liquid chromatography/tandem mass spectrometry. J. Chromatogr. A, 1224, 51–60. Dong, F., Cheng, L., Liu, X., Xu, J., Li, J., Li, Y., Kong, Z., Jian, Q., and Zheng, Y. (2012) Enantioselective analysis of triazole fungicide myclobutanil in cucumber and soil under different application modes by chiral liquid chromatography/tandem mass

References

86

87

88

89

90

91

92

93

spectrometry. J. Agric. Food Chem., 60, 1929–1936. Vetter, W., Klobes, U., Luckas, B., and Hottinger, G. (1997) Enantiomeric resolution of persistent compounds of technical toxaphene (CTTs) on t-butyldimethylsilylated β-cyclodextrin phases. Chromatographia, 45, 255–262. Vetter, W. and Luckas, B. (2000) Enantioselective determination of persistent and partly degradable toxaphene congeners in high trophic biota. Chemosphere, 41, 499–506. Blessington, B. and Crabb, N. (1989) Proposed primary reference methods for the determination of some commercially important chiral aryloxypropionate herbicides in both free acid and ester forms. J. Chromatogr., 483, 349–358. Shishovska, M. and Trajkovska, V. (2010) HPLC-method for determination of permethrin enantiomers using chiral β-cyclodextrin-based stationary phase. Chirality, 22, 527–533. Schneiderheinze, J.M., Armstrong, D.W., and Berthod, A. (1999) Plant and soil enantioselective biodegradation of racemic phenoxyalkanoic herbicides. Chirality, 11, 330–337. Möller, W.A., Bretzke, C., Hühnerfuss, H., Kallenborn, R., Kinkel, J.N., Kopf, J., and Rimkus, G. (1994) The absolute configuration of (+)-α-1,2,3,4,5,6hexachlorocyclohexane, and its permeation through the seal blood–brain barrier chromatographic separation of enantiomers of.(Durchlässigkeit der BlutHirn-Schranke von Seehundenfür das Enantiomer (+)-α-1,2,3,4,5,6Hexachlorcyclohexan und dessen absolute Konfiguration). Angew. Chem. Int. Ed. Engl., 33, 882–884. Dondi, M., Flieger, M., Olsovska, J., Polcaro, C.M., and Sinibaldi, M. (1999) High-performance liquid chromatography study of the enantiomer separation of chrysanthemic acid and its analogous compounds on a terguridebased stationary phase. J. Chromatogr. A, 859, 133–142. Ludwig, P., Gunkel, W., and Hühnerfuss, H. (1992) Chromatographic separation of

94

95

96

97

98

99

100

the enantiomers of marine pollutants. Part 5. Enantioselective degradation of phenoxycarboxylic acid herbicides by marine microorganisms. Chemosphere, 24, 1423–1429. Armstrong, D.W., Reid, G.L. III, Hilton, M.L., and Chang, C.D. (1993) Relevance of enantiomeric separations in environmental science. Environ. Pollut., 79, 51–58. Chu, Y.Q. and Wainer, I.W. (1988) The measurement of warfarin enantiomers in serum using coupled achiral/chiral, high-performance liquid chromatography (HPLC). Pharm. Res., 5, 680–683. Weber, K., Kreuzig, R., and Bahadir, M. (1997) On enantioselective separation of phenoxypropionates using permethylated β-cyclodextrin HPLC and GC columns. Chemosphere, 35, 13–20. Malakova, J., Pavek, P., Svecova, L., Jokesova, I., Zivny, P., and Palicka, V. (2009) New high-performance liquid chromatography method for the determination of (R)-warfarin and (S)-warfarin using chiral separation on a glycopeptide-based stationary phase. J. Chromatogr., B, 877, 3226–3230. Guillén-casla, V. and Pérez-arribas, L. (2008) One- and two-dimensional direct chiral liquid chromatographic determination of mixtures of diclofopacid and diclofop-methyl herbicides. J. Agric. Food Chem., 56, 2303–2309. Guillén-Casla, V., Magro-Moral, J., Rosales-Conrado, N., Pérez-Arribas, L.V., León-González, M.E., and Polo-Díez, L.M. (2011) Direct chiral liquid chromatography determination of aryloxyphenoxypropionic herbicides in soil: deconvolution tools for peak processing. Anal. Bioanal. Chem., 400, 3547–3560. Mano, N., Oda, Y., Asakawa, N., Yoshida, Y., and Sato, T. (1992) Development of a flavoprotein column for chiral separation by high-performance liquid chromatography. J. Chromatogr. A, 623, 221–228.

369

371

3 Micellar Liquid Chromatography: Fundamentals Maria C. García-Alvarez-Coque, Maria J. Ruiz-Angel, and Samuel Carda-Broch

3.1 Background and Development 3.1.1 Use of Additives in Reversed Phase Liquid Chromatography

The performance of reversed phase liquid chromatography (RPLC) results from the physical and chemical properties of both the substrate and the molecules bound to the surface of the stationary phase. This is the basis of the preparation of a variety of bonded phases. A simpler approach is the incorporation of a small concentration of an additive into a conventional polar mobile phase, which is able to reversibly alter the stationary-phase surface [1]. With this approach, diverse secondary reactions take place inside the column, whose intensity can be modulated by varying both the chemical nature and the concentration of the additive. The objective of incorporating additives in the mobile phase is to enhance the chromatographic performance, changing the absolute and relative retention (i.e., analysis time and selectivity) to convenient values, and improving the peak profile and resolution. The addition of different types of reagents has given rise to new chromatographic modes and an impressive increase in the number of compounds that can be analyzed by RPLC. 3.1.2 Ion-pair RPLC

In this context, in 1976, Knox and Laird developed ion-pair chromatography (IPC) by adding a small amount of an ionic surfactant (an amphiphilic compound with a hydrophobic chain and a hydrophilic head group) to the mobile phase in RPLC [2]. Coating with ionic surfactant is a simple and inexpensive way of converting a silica-based packing into an ion-exchanger, since the adsorbed surfactant is able to associate with ionic solutes of opposite charge. Meanwhile, the amount of free monomers in the mobile phase remains small [1,3]. This results in increased retention, proportional to the amount Analytical Separation Science, First Edition. Edited by Jared L. Anderson, Alain Berthod, Verónica Pino Estévez, and Apryll M. Stalcup.  2015 Wiley-VCH Verlag GmbH & Co. KGaA. Published 2015 by Wiley-VCH Verlag GmbH & Co. KGaA.

372

3 Micellar Liquid Chromatography: Fundamentals

of counterion adsorbed on the stationary phase. The nature of the stationary phase, the nature and concentration of ionic surfactant and organic solvent, the ionic strength, and pH are the main factors that affect the chromatographic behavior. Ionic surfactants with alkyl chains ranging from six (hexyl) to 16 (hexadecyl) methylene units are usual. Methanol, acetonitrile, and 1-propanol are the most common organic solvents added to the aqueous phase to elute the adsorbed compounds according to their polarity. One of the problems of IPC is that the partial adsorption of surfactant on the stationary phase makes column equilibration critical, which may yield nonreproducible results. Also, retention times may drift, owing to coating leakage, with a need of periodic column regeneration. This has belittled the use of this chromatographic mode for routine separations. Also, care should be taken to avoid exceeding the critical micellar concentration (CMC), since the retention behavior changes drastically when micelles start to form. 3.1.3 Birth of Micellar Liquid Chromatography

A few years after the development of IPC, in 1980, looking for new mobile phases in RPLC with particular properties, Armstrong and Henry suggested the possibility of using aqueous solutions of surfactants of different nature (with either ionic or nonionic head groups), at concentrations exceeding the CMC [4,5]. Above the CMC, stationary-phase saturation is reached or the changes in surfactant coating are small at increasing concentration of surfactant in the mobile phase. Beyond this threshold, the retention, instead of further increasing, decreases progressively due to a number of new secondary effects, such as the displacement of the adsorbed solutes by the surfactant, the formation of ion pairs in the mobile phase between solutes and surfactant monomers, and the interaction of solutes with micelles, which are dynamic aggregates formed by surfactant monomers with the nonpolar hydrocarbon chain oriented toward the core, and the neutral or ionic head group toward the surface (Figure 3.1). Micelles added to the mobile phase introduce new sites of interaction, notably modifying the solubility and transfer of solutes between mobile phase and stationary phase with respect to conventional RPLC with hydro-organic mixtures and IPC. The combination of surfactant-modified stationary phase and micelles has profound implications with regard to selectivity, analysis time and efficiency, and consequently, chromatographic resolution [6,7]. The possibility of using surfactants with ionic, zwitterionic, and nonionic head groups increase the capability of RPLC to separate ionic or neutral solutes that are able to interact with the surfactants. The RPLC mode with surfactant above the CMC has been called micellar liquid chromatography (MLC). Since micelles behave as a pseudo-phase within the mobile phase, the technique is classified among the pseudo-phase liquid chromatographic modes (where the mobile phase contains different types of entities that interact with solutes, such as micelles, cyclodextrins, vesicles, or nanometersized oil droplets in oil-in-water microemulsions) [1]. The impact of MLC has

3.1 Background and Development SO3

O3 - S O S 3

O

O3 - S SO 3

O

O

O

O

SO 3 SO 3

SO 3 SO SO 3 3

3 SO O

O

O

O

O

SO O3 S O 3

O

SO 3

O

O

O

O3 - S SO 3

O

O

O

O

SO 3

SO3

SO 3 O

SO 3 O

SO

O

O

SO 3 SO 3

SO 3 O

(a)

SO 3

O

O

O

Si

O

OSi

O

Si

O

O

Si

O

Si

O

Si

O

Si

O

Si

O

Si

N+

N+

N+

(b)

OSi

SO3

O

O

N+

Si

N+

O

SO3

SO3

O

O

O

Si

SO3

SO3

SO3

N+

N+

N+

+ N

N+

N+

N+ N+

N+

N+

N+ N+

N+

N+

N+

N+

+ N + N

+ N

+ N

+ N

Si

O

Si

O

OSi

O

Si

O

Si

+ N

O

Si

O

N + OSi

O

Si

O

Si

O

N + OSi

O

OSi

O

Si

Figure 3.1 Simplified illustration of surfactant-mediated environments in a C18 chromatographic system, with mobile phases of: (a) SDS and (b) CTAB.

373

374

3 Micellar Liquid Chromatography: Fundamentals

been greater than for other pseudo-phase modes. Its unique selectivity is attributed to the ability of micelles to compartmentalize and organize solutes at the molecular level. However, the association of surfactant monomers with the bonded phase, forming a structure similar to the exterior of open micelles (Figure 3.1), allows similar interactions with the solutes with deep implications with regard to analysis time, selectivity, and efficiency. The adsorption of an approximately fixed amount of surfactant monomers on the stationary phase is an important feature with regard to robustness. This gives rise to a stable modified stationary phase, with properties remarkably different from those of the underlying bonded phase [5,8]. In the first reports on MLC, the mobile phase contained only water, buffer, micelles, and a very low amount of surfactant monomers (the so-called “pure micellar mobile phases”), with micelles playing the role of an organic modifier. The idea is attractive, but solutions containing only surfactant are too weak and yield poor peak shape. For this reason, as early as 1983, Dorsey, DeEchegaray, and Landy suggested the addition of a small amount of organic solvent to the mobile phase (usually a short chain alcohol) to enhance the efficiency [9]. The term “hybrid micellar mobile phases” (and “hybrid MLC”) was given to the ternary mobile phases of water, surfactant, and organic solvent, where the concentration of organic solvent is maintained low enough to allow the formation of micelles. However, the micellization process and the secondary equilibria between solute, micelle, and modified stationary phase are altered.

3.2 The Mobile Phase in MLC 3.2.1 Suitable Surfactants

The type and concentration of a surfactant used in MLC are determined by practical limitations of the chromatographic system. The lower concentration of the surfactant must be well above the CMC to guarantee the formation of micelles, whereas the upper concentration is determined by the surfactant solubility and viscosity of the resulting mobile phase. A suitable surfactant for MLC should have low aggregation number, low CMC, and for ionic surfactants, low Krafft point (defined as the temperature at which the ionic surfactant solubility equals its CMC), which should be below the room temperature. A high CMC would imply operating at high surfactant concentration, which would result in viscous solutions with undesirable high system pressure and background noise in UV detectors. Since these detectors are the most common in MLC, a suitable surfactant must have also a small molar absorptivity at the operating wavelength. Several surfactants of diverse nature fulfill these conditions, but the surfactants that are routinely used in MLC are often limited to the anionic sodium dodecyl sulfate (SDS), the cationic cetyltrimethylammonium bromide (CTAB, the

3.2 The Mobile Phase in MLC

Table 3.1 Characteristics of the most common surfactants in MLC [5]. Surfactant

Molecular weight

CMC (mol/l)

SDS

288.4

8.2 × 10

CTAB

364.5

Brij-35

1198 (avg.)

3

Aggregation numbera)

Krafft or cloud point (°C)

υ (l/mol)c)

60

15

0.246

9 × 10

4

61

20–25

0.364

9 × 10

5

40

100b)

1.084

a) In water. b) Cloud point in 1–6% solutions. c) Partial specific volume.

chloride CTAC is also used), and the nonionic polyoxyethylene(23)dodecyl ether (Brij-35), whose main characteristics are summarized in Table 3.1. It should be noted that both SDS and Brij-35 have a C12 chain, differing in their head group. SDS is, by far, the most common surfactant in MLC used in at least 90% of the analytical reports [8]. The reason for this preference may be its commercial availability in high purity with a relatively low cost and its ability to efficiently dissolve proteins in biological matrices allowing direct injection of the samples in the chromatograph. Also, SDS has interesting effects on acidic solutes, which become weaker (i.e., the region of dominance of the neutral species increases) [10,11], and on basic compounds, whose peak tailing is eliminated due to the effective masking of silanol groups [12]. However, SDS is often chosen simply because it has been used in hundreds of MLC studies and applications. This has relegated the exploration of the possibilities offered by other surfactants, including the nonionics, such as Brij-35 and Triton X-100, and zwitterionics as n-dodecyl-N,N-dimethylamino-3-propane-1-sulfonate. Although, in general, SDS is preferred for analytical applications, Brij-35 has drawn attention because RPLC with solutions containing micelles of this surfactant seem to emulate in vitro the partitioning process in biomembranes better than conventional RPLC [13]. Also, due to the coating with polyoxyethylene chains, the polarity of the underlying alkyl-bonded phase is increased. The retention capability of a column modified with Brij-35 is appreciably smaller compared to SDS. Thus, Brij-35 seems appropriate to elute moderately hydrophobic basic compounds, without the need of adding an organic solvent (i.e., chromatography with water and detergent) [14], whereas an organic solvent must be added to SDS micellar mobile phases to elute basic compounds in practical times [15]. However, polar compounds are weakly retained with mobile phases of Brij-35. In general, the use of nonionic surfactants in MLC is a chromatographic field hardly explored. Mixed micelles of Brij-35 or Triton X-100 and SDS have also been suggested to improve selectivity [16,17]. An example is given by the separation of the diuretic hydrochlorothiazide and its degradation product 5-chloro-2,4-disulfamoylaniline in urine [16]. Although these compounds were well separated from the matrix background with a Brij-35 mobile phase, they remained unresolved. In contrast, baseline separation was obtained between the two compounds with SDS mobile phases, but the separation of the diuretic from the peak of an

375

376

3 Micellar Liquid Chromatography: Fundamentals

endogenous compound in urine was unsatisfactory. A mobile phase of 0.004 M SDS (below its CMC) and 0.02 M Brij-35 instead provided a good resolution, without compromising the separation of the drug from the matrix background. 3.2.2 Sites of Interaction in Micelles

An important feature of surfactants is that they form micelles in solution at a concentration above the CMC. Being the size of micelles, a few nanometers, their macroscopic properties in solution resemble those of a truly homogeneous solution. Thus, they do not cause light scattering, a property that renders them useful for spectrophotometry. Microscopically, however, micellar solutions are heterogeneous, being composed of two distinct media: the amphiphilic micellar aggregates (micellar pseudo-phase) and the surrounding bulk water or hydroorganic mixture that contains surfactant monomers in a concentration approximately equal to the CMC. Micelles provide hydrophobic and, for ionic surfactants, also electrostatic sites of interaction. Three sites of solubilization can be identified in a micelle: the core (hydrophobic), the surface (hydrophilic), and the palisade layer (the region between the surfactant head groups and the core). Depending on polarity and steric factors, solutes can remain in bulk solution, be associated with the free surfactant monomers or with the micelle surface (i.e., the surfactant polar head), be inserted into the micelle palisade, or penetrate the micelle core. Therefore, different solutes associated with micelles in the mobile phase experience particular microenvironments. These are reflected by perturbations in solute physical– chemical properties, including changes in solubility, acidity, photophysical properties, and reaction rates. The ability of micelles to cosolubilize hydrophobic and hydrophilic analytes in complex matrices, such as serum, is particularly interesting. 3.2.3 Organic Solvent Addition to Enhance the Chromatographic Properties

As explained, pure micellar mobile phases have two serious problems: the weak elution strength and the poor efficiencies. Shorter retention times are obtained by increasing the surfactant concentration, but the chromatographic efficiency deteriorates significantly compared to conventional RPLC. These problems were remediated by adding an organic solvent, which decreases the polarity of the aqueous solution, alters the micelle structure and acts on the stationary phase changing the amount of adsorbed surfactant [18]. Also, organic solvent molecules wet the bonded phase changing its physical–chemical structure (rigidity) and polarity. Pure SDS micellar solutions are in general useless as mobile phases, except for analyzing highly polar compounds. For this reason, hybrid micellar mobile phases that contain a small amount of an organic solvent, usually a short- or mediumchain alcohol (in a smaller amount than needed in conventional RPLC), are used in most applications [8]. The organic solvent decreases the analysis time

3.2 The Mobile Phase in MLC

to acceptable values (especially important for the most nonpolar solutes), changes the selectivity, and often improves the shape of chromatographic peaks with favorable effects on resolution. 3.2.3.1

Type and Concentration of Organic Solvent

The amount of organic solvent that can be added to RPLC is limited by its solubility. Miscibility with water is increased by the presence of surfactant in the mobile phase, allowing the potential use of a wider range of organic solvents at concentrations larger than those in aqueous solution. Organic solvents that would normally not be considered in conventional hydro-organic RPLC can form stable mobile phases with micelles. Thus, for instance, the molar solubility in water at 25 °C for 2-methyl-1-butanol, 1-pentanol, 1-hexanol, and pentane, which is 6.1 × 10 3, 4.5 × 10 3, 1.2 × 10 3, and 9.5 × 10 6, respectively, increases to 0.46, 0.92, 0.79, and 0.095 in 0.285 M SDS [19]. In spite of the wide range of compatible solvents, only the aliphatic alcohols 1-propanol, 2-propanol, 1-butanol, and 1-pentanol are routinely used in MLC to develop analytical methods. 1-Propanol is the most common modifier, whereas 1butanol and 1-pentanol are used to decrease the retention of moderately and highly nonpolar solutes, respectively. Surprisingly, there are only a few reports on MLC with acetonitrile, which is the solvent of choice in conventional RPLC and offers improved peak shape in MLC [12], or with ethanol, which has attracted considerable attention in recent time owing to its low toxicity. The reason is probably the smaller elution strength of acetonitrile and ethanol (also of methanol), together with the high popularity of 1-propanol among MLC users. It should be added here that aliphatic carboxylic acids have been proposed as useful alternatives [20]. The potential use of a second organic solvent in the hybrid micellar mobile phase has also been suggested. Thus, for example, the addition of 1-propanol was checked to improve the MLC separation of PAHs with micellar mobile phases of SDS, using 1-pentanol as the main organic solvent [21]. The beneficial effect of 1-propanol on the separation was explained by the decreased polarity of the aqueous phase, which increased the solubility of surfactant monomers and 1-pentanol. This would lead to a more dynamic system with more 1-pentanol available to displace the adsorbed surfactant molecules and modify the stationary phase. Also, the attraction of the nonpolar solutes toward the less polar mobile phase is increased. Methanol and acetonitrile were found less effective than 1-propanol. 3.2.3.2

Interaction of Organic Solvent with Micelles and Effects on the CMC

There is a practical limit to the amount of organic solvent added to a given micellar mobile phase, since the viscosity and, consequently, the pump back pressure, increase. Also, the addition of too much organic solvent can disrupt the micelles or, if not, can give rise to changes in the micelles affecting the surfactant aggregation number and CMC, so that conventional micelles are no longer formed in the mobile phase [15]. Organic solvents partition to micellar aggregates, and the degree of association increases with their hydrophobicity. This modifies the shape of the micelles. Similar to solutes, organic solvent molecules

377

3 Micellar Liquid Chromatography: Fundamentals

30

25

20 CMC (mM)

378

15

10

5

0

0

5

10 15 Organic solvent (%, v/v)

20

Figure 3.2 Critical micellar concentration of SDS in solutions containing acetonitrile (filled circle), methanol (rhombus), ethanol (filled rhombus), 1-propanol (square), 1-butanol (filled square), and 1-pentanol (blank circle). (Data taken from Ref. [22].)

can be located outside the micelles, into the micelle palisade, or into the micelle core. The incorporation of the organic solvent in the micelle can result in additional interactions with solutes. However, as long as the integrity of micelles is maintained, the addition of an organic solvent will not create a hydro-organic system, even though the interactions between solutes and micelles are weaker, and the stationary phase is more similar to that in conventional RPLC. Figure 3.2 shows the changes observed in the CMC for SDS upon addition of the most common organic solvents. Methanol, with the shortest carbon chain, is more polar and soluble than the other alcohols. The association of this alcohol with micelles is weak, but it solvates surfactant monomers efficiently, increasing their concentration in water. This hinders the interaction of the monomers during the formation of micelles, and consequently, a greater amount of surfactant is required to form the micelles (i.e., the CMC progressively increases with added methanol). A similar behavior is observed for acetonitrile, in spite of the different polarities and structures (acetonitrile is dipolar and nonprotic, while methanol is polar and protic). The effect of ethanol and 1-propanol is opposed to methanol. These alcohols remain mainly outside the micelles, dissolved in the bulk liquid, but interact more strongly with the micelle surface than methanol, reducing the repulsions among the ionic heads of the surfactant monomers. This benefits the formation of micelles and reduces the aggregation number and CMC. As the length of the alcohol alkyl chain increases, its affinity for the SDS micelle increases. 1-Butanol and 1-pentanol are inserted in the intermonomer spaces of the micelle palisade, owing to their particular structure that combines

3.2 The Mobile Phase in MLC

a polar group with a nonpolar chain, similarly to surfactant molecules. These alcohols align with the surfactant molecules in the micelle palisade, with the polar hydroxyl group of the alcohol oriented toward the bulk liquid. This gives rise to swollen mixed micelles. Such micelles are geometrically hindered to allocate additional surfactant monomers, which is translated into a CMC reduction that slows down above 4% butanol and 1.5% pentanol. This can be explained by considering that above a given concentration, the amount of alcohol entering the palisade is nonsignificant, and the excess is solubilized in the swollen micelle core. Further additions of alcohol lead to a dramatic change in the mobile-phase microstructure, yielding a microemulsion. On the other hand, the organic solvent molecules may disrupt the micellar interface by increasing the distance between adjacent groups (which decreases the charge density for ionic surfactants) and decreasing the distinction between the hydrophilic and the hydrophobic domains. The organic solvent fluidizes the micelle, which becomes more labile. Solute retention changes with the concentration of organic solvent in the mobile phase, similar to the CMC. This means that the collateral effects that change the CMC in hybrid mobile phases are, at least partially, those that induce shorter retention: the modification of bulk water and micelle. Thus, using SDS and 1-pentanol as modifier, both CMC and solute retention were observed to decrease strongly for 1-pentanol concentrations up to 1.5–2.0%. The change in retention upon addition of 1-propanol and 1-butanol was smaller, as occurred with the CMC. 3.2.3.3

Micelle Breakdown at High Organic Solvent Content

A high concentration of organic solvent prevents the formation of micelles. This means that the mobile phase will contain only free surfactant monomers. The maximal concentration to avoid micelle breakdown depends on the nature of the surfactant and organic solvent. Thus, SDS micelles have been found to disrupt at concentrations (v/v) above 30–40% methanol, 30% ethanol, 30% acetonitrile, 22% 1-propanol, 10% 1-butanol, and 6% 1-pentanol [15,23]. Also, CTAB micelles do not exist in solutions with more than 20% methanol. These maximal concentrations are still low in comparison with those needed in conventional RPLC. When the concentration of organic solvent in the mobile phase is increased, the transition to a situation where micelles do not exist is gradual (there is no sudden breakdown of micelles), with a progressive reduction in the aggregation number. This fact, and the absence of remarkable changes in the chromatographic behavior when micelles are disrupted, does not allow knowing exactly if micelles still exist. It is thus not surprising that, in the literature, some authors using high amounts of organic solvent have claimed to be working under MLC conditions, without being aware that no micelles were formed [24]. Nevertheless, there is no reason to neglect the potentiality of mobile phases containing surfactant above its CMC in water and a high amount of organic solvent (without micelles) [25,26]. In this situation, the interactions between solutes and a rather large amount of free surfactant monomers (instead of micelles) may coexist in the bulk solvent with the interactions with the still surfactant-modified stationary phase. This chromatographic mode in between MLC and

379

380

3 Micellar Liquid Chromatography: Fundamentals

conventional RPLC (first addressed by Li and Fritz [25]) has been called high submicellar liquid chromatography (HSLC). Although the chromatographic performance under micelle breakdown conditions in the mobile phase is still scarcely studied, several reported procedures show that the combination of reduced analysis time, larger selectivity, and improved peak shape, with respect to MLC and conventional RPLC, makes HSLC a promising chromatographic mode, to achieve in practical times separations of compounds unresolved or highly retained with other RPLC modes (Figure 3.3). However, it has the disadvantage that a larger amount of organic solvent is needed with regard to MLC. (a) 1 2 3 5

4 6

7

0 (b)

10

20

8

30

1

2 35 8

0 (c)

20

10

764

30

40

1 2 3 5 8 76

0

10

4

20 30 Time (min)

9

40

Figure 3.3 Separation of a mixture of β-blockers. Column: Kromasil C18. Mobile phases: (a) 15% v/v acetonitrile (hydro-organic conditions), (b) 0.1125 M SDS/25% acetonitrile (micellar), and (c) 0.1125 M SDS/35%

acetonitrile (high sumicellar). Compounds: (1) atenolol, (2) carteolol, (3) pindolol, (4) timolol, (5) acebutolol, (6) metoprolol, (7) esmolol, and (8) celiprolol. (Data taken from Ref. [24].)

3.3 The Stationary Phase in MLC

The organic solvent is seen as a secondary modifier, which can affect the micelle nature and displace the analyte partition equilibrium toward the bulk mobile phase. However, the role of the organic solvent is not far from that in a hydro-organic mixture [23]. The loss of protagonism can be explained by its association with the micelles or surfactant monomers, which decreases its capability to interact with analytes. Since the stabilization with an organized structure (as the micelles) is stronger, the disruption of micelles at high concentration of organic solvent is translated into a significant increase in the elution strength, which may become similar to that observed with a hydro-organic mobile phase in the absence of surfactant.

3.3 The Stationary Phase in MLC 3.3.1 Surfactant Coating of Bonded Phases

Most analytical procedures in the MLC literature involve conventional C18 columns, and in a much lesser extent, C8 columns (see Chapter 9). In some specific applications or studies, cyanopropyl columns have been used. The use of other columns is anecdotic. All types of columns are modified when a surfactant is incorporated into the mobile phase [27]. In conventional RPLC with hydro-organic mobile phases, the surface composition of bonded silica is more relevant to chromatographic processes than bare silica, since solute retention results from interaction with the few outer nanometers of the bonded silica. This can be extended to the surfactant-coated RPLC modes. The bonded stationary phase may be totally or partially coated by the surfactant [28]. A full similar coating would render the stationary phases all similar, but this seems not be the case. Indeed, coating with surfactant changes radically the column properties, but the subjacent stationary phase (the bonded moiety) still plays a role in the interaction with solutes. Surfactant adsorption on the porous RPLC packing affects chromatographic retention, owing to the change of diverse surface properties of the stationary phase. With nonionic surfactants, only the polarity of the stationary phase changes (in the absence of specific interactions with head groups), whereas with ionic surfactants, an asymmetric bilayer with a net charge (positive or negative) appears on its surface, which acts as a dynamic ionexchanger for ionic analytes (Figure 3.1). Solid-state nuclear magnetic resonance has been used to investigate the structure of the surfactant layer on alkyl- and cyano-bonded stationary phases [29]. It appears that the chain of SDS and Brij-35 surfactant monomers is adsorbed on the alkyl-bonded stationary phase through hydrophobic interaction, affecting the penetration depth of solutes into the bonded phase. The surfactant head group is, therefore, in contact with the polar solution, oriented away from the surface.

381

382

3 Micellar Liquid Chromatography: Fundamentals

Solutes can experience hydrophobic interactions with the nonpolar chain of the adsorbed surfactant and the nonmodified bonded stationary phase, and polar or ionic interactions with the head of the adsorbed surfactant. In the case of SDS, the sulfate head group creates a negatively charged hydrophilic layer, which attracts cationic solutes (Figure 3.1a). Meanwhile, the adsorption of Brij-35 with a hydrophilic polar end increases the stationary-phase polarity, which remains neutral. CTAB gives rise to two kinds of interactions with the stationary phase (Figure 3.1b): hydrophobic association with the alkyl-bonded layer as anionic surfactants, creating a positive charge that attracts anionic solutes, and electrostatic attraction of the positively charged surfactant head to the residual free silanols on the support, which are buried inside the thickness of the bonded layer [30]. This explains why alkyl-bonded phases covered with the anionic SDS are more polar than the phases covered with the cationic CTAB. In contrast, on cyano-bonded phases, both charged surfactants (SDS and CTAB) are adsorbed head down with their tails projected outward, creating thus pseudo-alkyl bonded phases. The smaller surface charge of the modified cyano-bonded phases is responsible for the more important role of solute–micelle interactions for charged solutes (see Section 3.4.2). 3.3.2 Stationary-Phase Saturation by Surfactant in Aqueous Medium

The adsorption isotherms of stationary phases of different nature, in contact with mobile phases containing SDS, CTAB, or Brij-35, have been extensively studied. The results have provided information about column conditioning in MLC, which is useful to achieve reproducible results. In a comprehensive work, Berthod and coworkers studied the adsorption isotherms for SDS and CTAB on five stationary phases of various polarities: three nonpolar silica (C1, C8, and C18) and two polar silica (cyanopropyl and naked silica) [28]. The authors found that the adsorbed amount for CTAB was larger with respect to SDS, which was explained by the attraction of the CTAB cation to the ionized silanol group. Maximal adsorbed SDS was found close to one surfactant molecule per bonded moiety for C1 and C8 bonded phases, and close to two SDS molecules for a C18 phase. Examination of the hysteresis loop for a C18 material, modified with surfactant, provided additional information about the extent of stationary-phase modification [18]. The BET surface area (surface available to the nitrogen molecules at 77 K) was found to decrease about 60% for both nonionic (Brij-35) and anionic (SDS) surfactants. These studies indicated that the surfactant molecules fill part of the silica pore volume, producing a thick continuous film on the interior walls, rather than completely filling the pores. On doing so, the stationary phase surface area is reduced. The amount of adsorbed anionic and cationic surfactants on the alkylbonded column increases rapidly at increasing surfactant concentration in

3.3 The Stationary Phase in MLC

the mobile phase. An increase in the ionic strength also increases surfactant adsorption. However, there is some disagreement on the conditions needed to reach stationary phase saturation by surfactant. Some authors have found a constant amount of adsorbed surfactant for ionic surfactants just above the CMC. This has been explained by the fact that a change in surfactant concentration is translated into an increase in the concentration of micelles in the mobile phase, whereas the number of surfactant monomers remains constant and equal to the CMC. However, the assumption that the column is saturated with surfactant just above the CMC seems to be not the case for all surfactant/stationary-phase combinations. For SDS, the adsorbed amount in C18 columns may increase as much as 20% after reaching the CMC (8.1 × 10 3 M) [28]. Also, the adsorption of Brij-35 on these columns continues largely after its CMC (which is rather low: 9.0 × 10 5 M), exceeding it as much as about 50 times [18]. However, under the usual experimental conditions in MLC (surfactant concentration above 0.05 M), the stationary phase is saturated with the monomers of both surfactants. The existence of a plateau of adsorbed ionic surfactant above the CMC with some stationary phases allows rapid analyses using a micelle gradient, since the initial conditions can be recovered without re-equilibration time [31]. However, the particular elution strength behavior of MLC makes gradient elution often unnecessary [32]. The effective removal of highly hydrophobic solutes from the stationary phase transported by the micelles gives rise to a “gradient effect” in MLC under isocratic conditions (Figure 3.4). 3.3.3 Use of Large-Pore Stationary Phases and Monolithic Columns

Shorter columns and shorter chain length stationary phases have been suggested to overcome the problem of the inability of MLC to effectively elute highly retained nonpolar solutes. However, although retention is reduced, resolution often decreases. Other solutions are the use of alkyl-bonded stationary phases with large pores [33], and monolithic columns [34]. The lack of strength of pure micellar mobile phases (i.e., without organic solvent) has been attributed partially to the exclusion of micelles from the pores, within nearly all (>99%) of the stationary-phase resides and solutes spend most of their time. Since the excluded micelles (whose dimensions are typically 30–60 Å, commensurate with the pore diameter of typical small pore C18 columns) do not have direct access to the solutes associated with the stationary phase (except when these have diffused out of the pores), even high concentrations of micelles are not sufficient to elute moderately to highly hydrophobic compounds. In the case of nonionic surfactants that form large micelles, steric effects are the most likely reason for micelle exclusion from small-pore stationary phases. With ionic surfactants that form smaller charged micelles, both electrostatic and steric effects may limit the ability of micelles to penetrate the pores.

383

384

3 Micellar Liquid Chromatography: Fundamentals (a) 1 2 3

5

0

1

2

4

3

6 7 8 0

40 Time (min)

80

(b) 5

6 7 1

8 2 3

0

20 Time (min)

4

40

Figure 3.4 “Gradient effect” observed in the analysis of a mixture of β-blockers and phenols. Column: 150 mm × 4.6 mm Zorbax Eclipse-XDB C18. Mobile phases: (a) 30% acetonitrile, and (b) 0.15 M SDS/15% acetonitrile.

Compounds: (1) carteolol, (2) acebutolol, (3) metoprolol, (4) celiprolol, (5) 2,4-dinitrophenol, (6) 3-bromophenol, (7) 2,4-dichlorophenol, and (8) 2,4,6-trichlorophenol.

When large-pore stationary phases are employed, the micelles are able to penetrate the pores and interact with the solutes [33]. This would benefit mainly nonpolar solutes that participate in a direct transfer mechanism between micelles and stationary phase (see Section 3.4.3). In order to determine whether large-pore stationary phases overcome the lack of strength in MLC, several C8 and C18 phases ranging from 100 to 4000 Å pore size were investigated with micellar mobile phases of nonionic (Brij-22), anionic (SDS), and cationic (dodecyltrimethylammonium bromide, DTAB) surfactants. It should be noted that as the pore size of a porous material is increased, the specific surface area is reduced. As a consequence, the volume of the bonded stationary phase is also decreased, and thus the stationary phase-to-mobile phase ratio. Therefore, under equal mobile-phase conditions, the retention time for a solute with a large-pore stationary phase will necessarily be shorter than with an otherwise identical small-pore stationary phase. For this reason, in order to get a valid conclusion, the authors compared the behavior of micellar and hydro-organic mobile phases

3.3 The Stationary Phase in MLC

using stationary phases of different pore sizes and found that large-pore stationary phases really allow better penetration of the micelles into the pores, such that they are able to reach the solutes at the internal surface of the stationary phase better, and elute them in less time. On the other hand, methods for conventional RPLC on silica-based particle columns are easily transferred to silica-based monolithic columns [34]. These are highly porous and allow for high mobile-phase flow rate associated with low column back pressure, which gives rise to short analysis times (fast methods). These columns were used to speed up quantitative structure–retention relationship (QSRR) studies to estimate the membrane permeability of drugs in a wide range of polarities. The performance was compared with a silica-based microparticle column, both with hydro-organic and micellar mobile phases. Micellar mobile phases and a monolithic column posed no practical problems. However, due to the larger viscosity, maximal flow rate was below 7 ml/min with the MLC method, whereas 9 ml/min can be reached in conventional RPLC with hydroorganic mobile phases. For the isocratic hydro-organic RPLC method with the conventional column, a rather wide retention time window was obtained, while the peaks of the least retained solutes were not completely separated from the solvent peak. Flow rates of 9 ml/min with the monolithic column and hydroorganic mixtures did shorten the analysis times, but differentiation from the solvent peak of the least retained solutes and among the peaks became almost impossible, resulting in the loss of information. With MLC, all drugs could be clearly distinguished and were retained even at high flow rate. Compared to isocratic hydro-organic RPLC, much shorter retention times for the most retained drugs and, in general, longer retention times for the least retained were observed. Thus, for MLC, a smaller retention window was found with a remarkably shorter analysis time on both microparticle-based and monolithic columns. 3.3.4 Removal of Surfactant from the Stationary Phase in the Presence of Organic Solvent

The main topic in the MLC literature is the role of micelles in the chromatographic behavior. Certainly, micelles increase the solubility of analytes and contribute to their desorption from the stationary phase, with an elution strength often larger than that of the organic solvent. However, as commented, besides the formation of micelles, the most relevant difference between RPLC without and with surfactant is the surfactant layer associated with the alkyl-bonded chains in MLC. As in IPC, the surfactant added to the mobile phase covers the stationary phase. Upon the addition of organic solvent, this coverage is reduced, since the solvent dissolves the adsorbed surfactant at least partially [18,35]. A relatively small amount of organic solvent added to an SDS micellar mobile phase significantly reduces the coating thickness, with a clear trend that depends on the concentration and polarity of the organic solvent. This decreases the retention in the order, methanol < acetonitrile < ethanol < 1-propanol < 1-butanol < 1-pentanol, which correlates with the extent of surfactant desorption from

385

3 Micellar Liquid Chromatography: Fundamentals

4.6

Adsorbed SDS (μmol/m2)

386

4.2

3.8

3.4

3.0

0.0

0.5 1.0 1.5 2.0 Organic solvent mole fraction

2.5

Figure 3.5 Effect of alcohols on maximal adsorbed surfactant for a mobile phase containing 0.2 M SDS. Alcohols: methanol (filled circle), 1-propanol (square), and 1-pentanol (blank circle). (With permission from Ref. [36].)

the stationary phase (stronger for 1-pentanol). While the addition of 5% methanol may reduce the amount of SDS by about 10%, 5% 1-pentanol will reduce it by about 50%. In general, these changes depend on the surfactant-to-organic solvent ratio and affect both retention and efficiency. The organic solvent competes with the surfactant for adsorption on the stationary phase: the longer the alkyl-chain of an alcohol, the stronger the adsorption on C18 stationary phases. The alkyl chains of 1-propanol and longer n-alcohols have been found to interpenetrate the alkyl chains of the bonded phase to form a single monolayer, adopting a structure similar to that of the surfactant monomers in MLC, with the hydroxyl group oriented toward the aqueous phase. This competition between alcohols and surfactant molecules for adsorption sites on the stationary phase explains the linear reduction in the amount of adsorbed surfactant with increasing concentration of alcohol in the mobile phase (Figure 3.5). If the addition of organic solvent to the mobile phase reduces the amount of adsorbed surfactant, in the limit, the BET surface area and the cumulative pore volume will reapproach that of the unmodified stationary phase [18]. High concentrations of organic solvent can sweep out completely the adsorbed surfactant molecules from the surface of the bonded phase. In HSLC, where the concentration of organic solvent does not allow the formation of micelles, the retention mechanism will depend on the amount of surfactant that still remains adsorbed on the alkyl-bonded phase, and on the interaction of solutes with surfactant monomers in the bulk solvent, which replace the micelles [26]. It can be expected that, without surfactant adsorption, the observed effect on retention and resolution would be solely a result of the interactions with the surfactant

3.4 Retention Mechanisms

monomers in the mobile phase, in addition to the interaction with the nonpolar bonded phase and organic solvent in the mobile phase.

3.4 Retention Mechanisms

A surfactant adsorbed on the stationary phase exposes an extreme of its molecule to the mobile phase, changing the stationary phase polarity and type of interactions, which can be hydrophobic and electrostatic for charged surfactants, or more specific as is the case with Brij-35, which interacts strongly with hydroxyl groups in phenols and polyphenols. Solutes also still experience the underlying bonded phase. This, in combination with shape and steric constraints, affects the retention and selectivity [36,37]. Several of the interactions with surfactant monomers are also found with the micelles in the mobile phase. The variety of interactions in MLC does not exist in any homogeneous hydro-organic mobile phase. This makes the RPLC mode compatible with a wide range of solutes (ionic to water insoluble). The main strength of MLC lies precisely in the capability of performing and controlling the separation of mixtures of cationic, anionic, and uncharged polar and nonpolar solutes, using isocratic elution, provided appropriate surfactant, organic solvent, and experimental conditions are chosen. To understand the partitioning behavior in MLC, it should be considered that the attraction of solutes to the surface of the surfactant-modified stationary phase is usually stronger than the attraction to micelles in the mobile phase. However, since the amount of adsorbed surfactant on the stationary phase remains constant, an increase in the concentration of surfactant in the mobile phase results in decreased retention. Peak efficiency also usually deteriorates appreciably. In contrast, the addition of more organic solvent decreases the retention and improves the peak shape, as explained below. It may be asked if the improved performance is really competitive with respect to conventional RPLC. Unfortunately, in the literature, only a few reports compare directly conventional hydro-organic RPLC and MLC, in spite of the fact that such kind of studies are also interesting to reveal the retention mechanisms that take place inside an RPLC system in the presence of additives [38]. 3.4.1 The Three-Phase Model

In MLC, solutes are partitioned between the mobile phase and the stationary phase, which are modified by the presence of micelles and surfactant monomers, respectively (Figure 3.1). In pure micellar systems, the retention behavior is explained by considering three phases or environments: surfactant-modified stationary phase, bulk aqueous solvent, and micellar pseudo-phase [36,37]. Solutes are separated on the basis of their differential partitioning among the three phases. Therefore, three different equilibria are possible: (i) the distribution of solutes

387

3 Micellar Liquid Chromatography: Fundamentals

388

Si

O

Si

O

OSi

O

Si

O

OSi

O

Si

O

Si

O

Si

O

Si

O

Si

O

Si

O

Si

O

Si

O

Si

O

OSi

O

Si

O

OSi

O

Si

O

Si

O

Si

O

Si

O

Si

O

Si

O

Si

O

OSi

O

Si

O

Si

O

Si

O

Si

O

Si

O

Si

O

Si

O

Si

O

Si

Figure 3.6 Simplified illustration of the direct transfer mechanism between micelle and surfactant-modified stationary phase.

between the micelle and the aqueous solvent, (ii) the partition of solutes between the stationary phase and the aqueous solvent, and (iii) the direct transfer between the surfactant-modified stationary phase and the micelle (Figure 3.6) [37]. These equilibria are described by three partition coefficients: PAS (between aqueous solvent and stationary phase), PAM (between aqueous solvent and micelles), and PMS (between micelles and stationary phase). The coefficients PAS and PAM account for the solute affinity to the stationary phase and micelles, respectively, and have opposite effects on solute retention: as PAS increases retention increases, whereas as PAM increases retention is reduced due to the stronger association with micelles. In hybrid MLC, the equilibria are significantly displaced away from the micelle and the stationary phase toward the bulk hydro-organic mixture, which has decreased its polarity. Both PAS and PAM coefficients decrease, especially PAS. This makes retention to decrease. The transitions that may happen among the three environments in a pure micellar chromatographic system (i.e., aqueous solvent, micelles, and stationary phase) can be described by the following equation [37]: Ve

V0 k P AS ˆ ˆ VS Φ 1 ‡ υ…P AM 1†‰MŠ

(3.1)

where Ve represents the total volume of mobile phase needed to elute the solute from the column, VS and V0 are the volume of the active surface on the

3.4 Retention Mechanisms

stationary phase and the column void volume, respectively, Φ ˆ V S =V 0 is the phase ratio, and υ the partial specific volume of surfactant monomers in the micelle (Table 3.1). When micelles are not present in the mobile phase, Equation 3.1 is reduced to the partitioning equation in conventional RPLC: V e ˆ V 0 ‡ V S P AS

(3.2)

Equation 3.1 allows the measurement of the strength of solute–stationary phase and solute–micelle interactions [39]. Since in MLC, the stationary phase is modified by surfactant adsorption, PAS in conventional RPLC and MLC will be different for the same column. MLC is a nice example of the use of secondary equilibria in liquid chromatography, which can be altered by a variety of factors, such as the nature and concentration of surfactant and organic solvent, temperature, ionic strength, and pH for ionizable compounds [10,40]. A decrease in retention time is usually observed when the concentration of the surfactant or the organic modifier is increased. However, different analytes in a mixture respond in different ways to changes in concentration of surfactant and/or organic solvent, resulting in a change in selectivity. 3.4.2 Separation of Charged Solutes

Partitioning equilibria in MLC are governed by hydrophobic forces, but electrostatic attraction and repulsion for ionic surfactants are also important. Neutral solutes eluted with nonionic and ionic surfactants and charged solutes eluted with nonionic surfactants will be affected only by hydrophobic, dipole–dipole, and proton donor–acceptor interactions with micelles and stationary phase. Charged solutes will also interact electrostatically with ionic surfactants (i.e., the charged surfactant layer on the stationary phase and the charged outer-layer of micelles). With ionic surfactants, two situations are possible: solute charge is opposite to or the same as the charge of the surfactant head and, therefore, can be attracted to or repelled, respectively, by the surfactant monomers in both surfactant-modified stationary phase and micelles. In the case of electrostatic repulsion, charged solutes cannot be retained by the stationary phase and will elute at the dead time, unless significant hydrophobic interaction with the modified bonded layer exists. In contrast, combined electrostatic attraction and hydrophobic interactions with the modified stationary phase may be sufficiently large to offset the micelle attraction, and retention will be strong, but this can be decreased to practical values by adding an organic solvent. This is the case of positively charged protonated basic compounds eluted with SDS micellar mobile phases. Instead, these compounds elute at the dead time with CTAB. The function of the micellar pseudo-phase in MLC has been compared with that of the organic modifier in conventional RPLC, since for most solutes,

389

390

3 Micellar Liquid Chromatography: Fundamentals

an increased surfactant concentration in the mobile phase results in decreased retention. This behavior contrasts to that found in IPC, where the addition of ionic surfactant increases solute retention, owing to the electrostatic attraction to the stationary phase covered progressively with ionic surfactant [3]. However, it should be noted that in MLC, the elution strength increases with surfactant concentration only if the solute interacts with micelles. According to their retention behavior, three groups of solutes can be recognized: binding, nonbinding, and antibinding [41]. Binding solutes associate (bind) with micelles and show a decreased retention at increasing concentration of micelles in the mobile phase. Both nonbinding and antibinding solutes do not associate with micelles, and their retention remains unaltered (nonbinding) or is increased (antibinding) with increasing micelle concentration. The most frequent behavior is binding to micelles, while antibinding is quite uncommon. It is evident that electrostatic repulsion is an important issue in antibinding behavior, since antibinding solutes with anionic or cationic surfactants are negatively or positively charged, respectively. This behavior is not observed between a charged solute and an oppositely charged surfactant. It is also never observed with C8- or C18-bonded phases modified by adsorption of ionic surfactants, since solutes elute at the dead time due to repulsion between the solutes and the charged surfactant layer on these stationary phases. The antibinding behavior has been observed only either with C1 phases, which do not adsorb large amounts of surfactant, or cyano-bonded phases where ionic surfactant molecules are adsorbed with their tails projected upward. In both cases, charged solutes are strongly excluded or repelled from the micelle, which forces them to bind to the stationary phase, where they are retained due to hydrophobic forces. The smaller surface charge of the SDS- and CTAB-modified cyanopropyl-bonded phases is also responsible for solute–micelle ionic interactions playing a more important role in the retention of weak bases and weak acids, respectively. By increasing the ionic strength, antibinding solutes can adopt nonbinding or even binding behavior, proving again the electrostatic origin of the observed behavior. This effect has been explained as similar to the flocculation of colloidal systems: the electrical double layer surrounding the micelle is narrowed in a mobile phase containing higher concentration of ions, which facilitates the approximation of the solute to the micelle assembly to establish interactions between the solute and the hydrophobic micelle core. Therefore, for the transition from antibinding to nonbinding, and further to binding, solute polarity must be sufficiently low for the solute to associate with the nonpolar portion of the micelle, once the electrostatic repulsion has been minimized by salt addition. Thus, in the absence of salt, bromophenol blue eluted with SDS from a cyanopropyl-bonded phase behaves as antibinding, whereas in the presence of as little as 0.02 M NaCl, it appears as nonbinding, and at slightly higher salt concentration, it is strongly bound to SDS micelles [42].

3.4 Retention Mechanisms

3.4.3 Direct Transfer of Solutes from the Micelle to the Stationary Phase

In pure micellar systems, solute migration can occur either via the exit of the solute from the micelle, followed by its diffusion in the bulk aqueous solvent and subsequent association with another micelle or to the surfactant-modified stationary phase, or by direct transfer through collisions and transient merging of intact micelles with the surfactant-modified stationary phase (Figure 3.6) [36,43]. The latter path can be neglected in most situations. Also, in principle, it should be applicable only to nonionic micellar systems, since in the case of ionic surfactants, ionic head groups in the micelles and on the surface of the modified stationary phase will be similarly charged and an electrostatic repulsive barrier will exist, impeding solute mass transfer across this interface. However, it has been reported that for hybrid micellar mobile phases, which contain organic solvents, micelles, and a variety of micellar fragments and surfactant-stabilized clusters, this additional migration mechanism can become operative as is the case of cationic solutes eluted with anionic SDS micelles [24]. Thus, the micelle can lose a fragment or incorporate a fragment, eventually mediating the transfer of a solute either between micelles or clusters in the mobile phase or between a micelle in the mobile phase and the surfactant-coated (micellar-like) stationary phase. It seems, thus, that in pure micellar mobile phases, water-insoluble nonpolar solutes or species strongly bound to surfactant monomers do not participate significantly in the aqueous environment of the three-phase partitioning scheme [43]. That is, they spend most of their time in the stationary phase or bound to the micelle, and very little time in the bulk aqueous phase. These solutes exhibit large affinity for the surfactant-coated stationary phase and the micelles, and consequently, can be transported only between them by direct transfer. This explanation has been called the solubility limit theory. For such solutes (for which large PAS and PAM values are expected), only a single equilibrium will exist, described by the partition coefficient between micelles and stationary phase (PMS = PAS/PAM). From Equation 3.1, kˆ

ΦP MS ΦP MS  1=P AM ‡ υ…1 1=P AM †‰MŠ υ‰MŠ

(3.3)

Therefore, the selectivity coefficient for a pair of compounds experiencing only direct transfer will not depend on the concentration of surfactant in the mobile phase [39]: αˆ

P MS;a 1=P AM;b ‡ υ…1 P MS;b 1=P AM;a ‡ υ…1

1=P AM;b †‰MŠ P MS;a  1=P AM;a †‰MŠ P MS;b

(3.4)

By addition of organic solvent to the mobile phase, the solubility of nonpolar solutes increases and this transfer mechanism loses importance with regard to the usual two-step mechanism (micelle-bulk solvent and bulk solventstationary phase).

391

392

3 Micellar Liquid Chromatography: Fundamentals

1.7

(a)

1.8

(c)

log k

log k

1.6 1.6 1.5 1.4

1.4

1.2

1.3

1.0

1.2

0.8

0

1

2

3

4

5

0

1

2

3

4

5

50

80 k

1.1

(b) 70

k

(d) 45

60

40

50

35

40

30

30

25

20

20

10

15 0

1 2 3 4 Carbon number in homolog, nC

5

0

1 2 3 4 Carbon number in homolog, nC

5

Figure 3.7 Correlations of log k and k versus the number of carbon atoms for n-alkylbenzenes. Mobile phases: (a, b) 35% 2-propanol and (c, d) 0.2 M SDS. (Data taken from Ref. [44].)

3.4.4 Effect of Polarity on MLC Retention 3.4.4.1

Retention Behavior of Homologous Series

The regular linear increase in the logarithm of the retention factor, log k, for a homologous series is a recognized measurement of hydrophobic interactions in an RPLC system with hydro-organic mobile phases (Figure 3.7a) [44]: log k ˆ log α…CH2† nC ‡ log β

(3.5)

where the intercept log β corresponds to the contribution to the retention from the functional group common to the series, nC is the number of carbon atoms or repeat units in the homologs, and α…CH2† is the nonspecific selectivity of a methylene group (called methylene or hydrophobic selectivity), which is the ratio of

3.4 Retention Mechanisms

the retention factors of two compounds in a homologous series differing only in a methylene group: α…CH2 † ˆ

k n‡1 kn

(3.6)

The linear dependence between log k and nC (Equation 3.5) has been attributed to the direct proportionality between log k and the free energy of retention, which is in turn a linear combination of the free energy increments associated with the different parts of the molecule. Solute polarity should play also an important role in governing the retention in micellar systems. However, in the early work in MLC, it was already observed that micelles in the aqueous mobile phase have an important effect on the retention of homologous series, being often k (and not log k) that is linearly related to nC (Figure 3.7d) [44–46]: k ˆ a nC ‡ b

(3.7)

where a and b are fitted coefficients. This behavior has been observed for SDS, CTAB, and Brij-35 with C8 and C18 stationary phases, and different homologous series, such as n-alkylbenzenes, 2-alkylanthraquinones, and n-alkylphenones. Also, it has been used to demonstrate that the separation mechanism in hybrid MLC is more similar to pure MLC than to conventional RPLC. In conventional RPLC, the measurement of selectivity (Equation 3.6) should be independent of the successive pairs of homologs and type of series, for a given system of mobile phase and stationary phase. In micellar eluents, this is not the case, since the relationship between log k and nC is nonlinear. Thus, in the n-alkylbenzene homologous series, α…CH2 † between n-pentylbenzene and n-butylbenzene is smaller than between ethylbenzene and toluene. Also, the difference in α…CH2 † between two types of homologs increases with the number of carbon atoms. Thus, it is consistently greater for n-alkylphenones with respect to n-alkylbenzenes. The observed curvature in the log k versus nC plots in MLC was first attributed to the different location (with different microenvironment polarities) in the micelle for different members in a homologous series: the more hydrophobic (larger) homologs tend to be located in a less polar environment of micelles. It can be, thus, assumed that these compounds experience a smaller change in their microenvironment polarity upon being transferred from the micelle to the alkyl-bonded stationary phase. Meanwhile, owing to the more polar specific group (carbonyl) for n-alkylphenones, these are located in a more polar location of micelles than are n-alkylbenzenes. This situation does not exist in hydro-organic systems. The solubility limit theory also provides an interesting explanation of the retention behavior of homologous series. At a low homolog number, the solutes are relatively water soluble; thus, aqueous solvent-stationary phase partitioning plays its largest role: the slope of the log k versus nC curve reaches its largest value. As the homologs become larger, they are less water soluble. In the limit, the homologs partition directly between the micelles and the modified stationary phase, and log k becomes independent of nC. The smaller α…CH2 † coefficients in MLC with regard to conventional RPLC also constitute an evidence of the closer

393

394

3 Micellar Liquid Chromatography: Fundamentals

environment of a methylene group in the micellar mobile phase to the modified stationary phase, which is micelle-like. 3.4.4.2

Correlation between Retention and Octanol–Water Partition Coefficient

The observed behavior for homologous series has also been found for other compounds [46,47]. Excellent correlations for k versus the logarithm of octanol– water partition coefficient (log Po/w) have been reported in MLC with compounds representing a relatively broad range of molecular interactions, molecular shapes and sizes, and covering more than four orders of magnitude in polarity: k ˆ a log P o=w ‡ b

(3.8)

A practical advantage of the linear k versus log Po/w relationships (instead of log k versus log Po/w), in MLC, is that a larger number of compounds are eluted per time unit in the isocratic mode with regard to conventional RPLC: chromatographic peaks are more evenly distributed with longer retention times for the least retained solutes and shorter for the most retained ones (see Figures 3.3 and 3.4). However, the type of general relationship (k or log k versus log Po/w) seems to depend on the selected set of compounds and the nature of both mobile phase and stationary phase. Also, the linearity of both correlations seems to improve by addition of organic solvent to the micellar mobile phase. The organic solvent apparently provides an environment closer to the octanol–water mixture than pure micellar systems. On the other hand, MLC has demonstrated to be useful for estimating log Po/w, based on the good correlation between retention and log Po/w for series of neutral compounds of diverse polarities. When the ionization degree is the same for structurally related ionizable compounds, the difference in retention is also a result of the differences in polarity. Thus, acceptable correlation has been found for groups of basic compounds in acidic mobile phases of SDS, where the positively charged protonated species dominate. For ionizable compounds with different ionization degrees, the correlations are poor due to the extra ionic interactions, unless a correction is applied. Equation 3.8 can also be used for ionizable compounds if apparent log Po/w values measured for the ionized species are considered [48]. 3.4.4.3

Effect of Organic Solvent Polarity

The addition of organic solvent to the micellar mobile phase reduces the polarity of the aqueous bulk solvent. This yields a decreased retention (even relatively small amounts of organic solvent will produce dramatic effects) and influences the selectivity. The effect is attenuated as the surfactant concentration is increased. Depending on the strength of the interactions, the organic solvent or the surfactant has a prevalent effect on the elution strength. The observed elution strength in hybrid MLC approximately correlates with the polarity (log Po/w) of both solutes and organic solvent. Less polar solvents are able to shorten the retention to a larger extent. Also, the decrease in retention is more intense for less polar solutes. In a comprehensive study, 21 organic solvents (alkanols, alkanediols, dipolar aprotic solvents, and alkanes) were added to an SDS

3.4 Retention Mechanisms

(a)

3

2-methyl-1-butanol cyclohexanol hexanol 3-methyl-1-butanol 2-pentanol 1-pentanol

2 1 log Po/w

1-butanol

2-butanol 2-methyl-1-propanol

1-propanol

0

acetonitrile

ethanol methanol 2,3-butanediol

-1

1,2-propanediol formamide

-2 -3 5 (b)

6 7 Retention factor

8

9

6 cyclohexanol 1-hexanol

4

log Po/w

2-pentanol

2

1-pentanol 2-methyl-1-butanol

2-butanol 2-methyl-1-propanol

1-butanol 1-propanol

0 ethanol

-2

methanol

2,3-butanediol

1,2-propanediol

-4 1500

2000

acetonitrile

formamide

2500 3000 3500 Efficiency or plate number

Figure 3.8 Correlation between the retention factor (a) and the efficiency (b), for benzene and the octanol–water partitioning coefficient for several organic solvents added to a

4000

micellar mobile phase. In all cases, the mobile phase contained 0.285 M SDS and 5% organic solvent. (Figure 3.8a with permission from Ref. [36].)

micellar mobile phase to determine their effect on two solutes of widely differing polarity (benzene and 2-ethylanthraquinone) in a C18 column [19]. The retention decreased as the organic solvent polarity decreased, but the effect of alkane additives on retention was too small. The elution strength was observed to parallel the organic solvent log Po/w values (i.e., polarity) (Figure 3.8a). Also, the correlation between the solute retention factor and log Po/w for the organic solvents was similar to the correlation between the solute retention factor and the binding constants of solvents with micelles (i.e., their ability to bind to micelles).

395

3 Micellar Liquid Chromatography: Fundamentals

3.4.5 Effect of pH on Retention

Chromatographic retention of ionizable compounds depends on the mobilephase pH, being a weighted mean of the behavior of the basic (kA) and acidic (kHA) species, as follows [10,11]: k ˆ kA

1 K Hh k A ‡ k HA K H h ‡ k HA k HA ˆ 1 ‡ K Hh 1 ‡ K Hh 1 ‡ K Hh

(3.9)

where KH is the protonation constant. Since the intrinsic retention for each species is different, a sudden change in retention will happen at pH values close to log KH in the mobile phase medium (Figure 3.9). The dependence is sigmoidal. log KH

log KH

(a)

B

Retention factor

Retention factor

HA

HA

A–

BH+

(b)

BH+

(d)

(e)

log KH

log KH

A–

B

A– (c)

B

(f)

log KH

log KH

Retention factor

396

BH+

HA

pH

pH

Figure 3.9 Effect of pH on retention in a C18 column and on the protonation constant KH for acidic (HA/A ) and basic (BH+/B) compounds. Mobile phases: (a, d) hydro-organic and (b, c, e, f) micellar with anionic surfactant (b, e) and cationic surfactant (c, f).

3.4 Retention Mechanisms

However, for conventional alkyl-bonded stationary phases (with working pH in the range 2.5–7.5), the full protonation process (which covers several pH units) is not observed. Therefore, the plots of retention versus pH show different patterns. In hybrid micellar systems, the protonation equilibria are shifted by the presence of both organic solvent and surfactant, due to the modification of the bulk solvent polarity, and the association of solutes with micelles and the adsorbed surfactant monomers on the stationary phase. Changes in the acid–base behavior (i.e., apparent constant K ´H ) and retention of ionizable solutes are observed with both nonionic and ionic surfactants, but the most interesting effects correspond to ionic surfactants, due to the electrostatic interaction between solutes and charged micelles and stationary phase (Figure 3.9). The shift in K ´H depends on the concentration of organic solvent and surfactant. Usually, K ´H decreases as the concentration of organic solvent increases. For ionic surfactants, solute–surfactant electrostatic interactions are responsible for the trends in K ´H and k versus pH curves at increasing surfactant concentration. In alkyl-bonded columns, the surfactant head group is exposed, so that the stationary phase surface is charged. Consequently, stationary phases modified with the anionic SDS will cause repulsion of anionic species, and stationary phases modified with the cationic CTAB will repel cationic species. This results in different elution behavior for weak acids and bases as a function of micelle concentration and pH in the mobile phase. For weak acids having neutral acidic and anionic basic species, and for weak bases with cationic acidic and neutral basic species, eluted both with an anionic surfactant, K ´H increases with the concentration of surfactant, due to the preferential attraction of the acidic species (neutral or cationic) toward the modified surface of the stationary phase with respect to the micelles (Figure 3.9b and e). The magnitude of the shift in log K ´H is larger for compounds experiencing larger interactions with the surfactant, and can amount to more than one unit. The shift in K ´H to higher pH benefits the observation of the maximal retention (the retention of the acidic species) within the operable limits of silica-based columns. Also, column life is extended as working at less stressing pH is possible, and measurements are more reproducible, since these can be made in a region of constant retention, where only one species dominates. On the other hand, for weak acids, the largest k-values are observed in acidic solution, where the neutral species dominate, and the smallest in more basic media where the anionic deprotonated species prevails with a weaker association with the surfactant-modified stationary phase (the k versus pH curves are descending). The same shape will be observed for weak bases: the dominant cationic protonated species at low pH is more strongly retained than the neutral basic species. In the recommended MLC procedures for weak acids, such as amino acids and phenols, using SDS as surfactant, the pH is often fixed at 2.5–3, where the protonated species dominates. Under these conditions, the separation space is wider, which favors chromatographic resolution. Although basic drugs (with log KH = 9–10) do not change their retention in the working pH

397

398

3 Micellar Liquid Chromatography: Fundamentals

range of conventional alkyl-bonded columns, low pH is also selected to enhance the efficiencies due to the protonation of silanols. For micellar mobile phases of a cationic surfactant, such as CTAB, which stabilize the basic species, the changes in both K ´H are opposite to those observed for anionic surfactants (Figure 3.9c and f). Weak acids are stronger and weak bases are weaker in the CTAB micellar system. Also, a mirror shape of the k versus pH curves with respect to SDS is observed (ascending curves) for both weak acids and bases, since the basic species experience higher retention. The difference between the shape of the k versus pH curves in MLC and conventional RPLC should be remarked. In MLC with SDS, the curves for weak acids and bases are both descending, and for CTAB these are both ascending. In contrast, for conventional RPLC, the k versus pH curves are descending for weak acids and ascending for weak bases (Figure 3.9a and d). In general, the different distribution of the K ´H values for ionizable compounds in micellar mobile phases, with respect to hydro-organic mixtures, contributes to the changes in selectivity observed between both chromatographic modes [11].

3.5 Efficiency 3.5.1 Chromatographic Efficiency in MLC: Should it be a Topic of Concern?

The MLC literature contains many examples on the reduced efficiency. The problem becomes progressively more acute as solute polarity decreases [49]. Both organic solvent addition and temperature raise were soon given as solutions to improve the efficiency [9,50], but the use of organic solvents to yield the hybrid MLC mode is preferred. However, as commented in Section 3.4, it is not often easy to appraise the chromatographic performance in MLC, since most reports deal exclusively with the observed changes upon addition of organic solvent to a pure micellar mobile phase; the performance with regard to conventional hydro-organic mixtures is usually not checked. Only in a few studies both chromatographic modes have been compared using the same C18 column and probe compounds, with interesting results: for polar or relatively polar compounds, hybrid MLC often yields similar or even improved efficiency relative to conventional RPLC, although the efficiency for highly nonpolar compounds is clearly inferior [49]. Therefore, comments such as “In spite of the advantages of MLC and the fervor of its proponents, this separation technique has not seen widespread application because it tends to be less efficient than conventional RPLC”, “Despite the advantages, MLC has not found widespread use due to poor column efficiency”, or “The problem of reduced efficiency in MLC still remains, despite extensive study”, which question the validity of MLC, are not justified.

3.5 Efficiency

3.5.2 Understanding the Reduced Efficiency with Pure Micellar Mobile Phases

Dorsey, DeEchegaray, and Landy believed that the reason of the reduced column efficiency in MLC with pure micellar mobile phases was the “poor wetting” of the nonpolar stationary phase, due to the high water content in the mobile phase [9]. According to these authors, “If micellar mobile phases are ever to be widely accepted as a viable chromatographic technique, the efficiency achieved must at least approach that of conventional RPLC”. These authors proposed the addition of some organic solvent to the mobile phase. In that study, the effect of 10% methanol, ethanol, 1-propanol, and acetonitrile with benzene as probe compound was checked, and certainly, an improvement in both efficiency and peak symmetry was observed as the polarity of the organic solvent decreased, with the best results corresponding to 1-propanol. Other authors disagreed with the hypothesis of Dorsey, DeEchegaray, and Landy, and offered other reasons for the reduced efficiency in MLC. Thus, Yarmchuck and coworkers referred to a possible slow solute exit rate from the micelle in the mobile phase and from the surfactant-modified stationary phase, which gives rise to poor mass transfer between the micelles, bulk aqueous phase, and stationary phase [7], and Borgerding and coworkers more accurately suggested that the thick surfactant layer on the stationary phase was the reason of the poor efficiencies [18]. Solute diffusion in the stationary phase coated with surfactant is very difficult and slows down the solute mass transfer (transfer to, in, or from the stationary phase). Certainly, alkyl-bonded stationary phases in the presence of surfactant are quite different from the original phase because of adsorption of surfactant in amounts approximating that of the bonded hydrocarbon, creating a double layer. The adsorbed surfactant changes the stationary phase (its surface, polarity, carbon load, and porosity), and consequently, the solutestationary phase and micelle-stationary phase interactions and kinetics. The comparison of Van Deemter plots for C18 columns using acetonitrile– water mixtures and pure micellar mobile phases of Brij-35 revealed that the term relating the theoretical plate height directly to the square of the stationary-phase film thickness, and inversely to the solute stationary-phase diffusion coefficient, was increased in MLC [18]. For well-designed column packings, this term is negligible, but it is significant for column packings with a thick stationary phase or poor stationary-phase diffusion, such as the original bonded phase materials. The Brij-35-coated column is somewhat analogous to these materials. This should impact adversely the MLC efficiency. This is especially problematic for highly nonpolar solutes that have a high affinity for the stationary phase: the higher the solute-stationary phase coefficient (which gives information about the affinity of the solute for the surfactant-coated stationary phase), the smaller the efficiency. Also, the greater the fraction of partitioning of nonpolar solutes that occurs via direct transfer mode, the poorer the chromatographic efficiency. Watersoluble solutes, which do not have to undergo such transfer process, exhibit larger efficiencies.

399

400

3 Micellar Liquid Chromatography: Fundamentals

Accordingly, the reason for the improved efficiency with organic solvent addition should be rationalized in terms of its effect upon the surfactant-modified stationary phase [18,35]. Organic solvents produce a thinner surfactant layer, which permits better solute diffusion. The achieved efficiency is parallel to the ability of organic solvents to desorb SDS monomers from the bonded stationary phase. Organic solvents are also expected to influence the fluidity/rigidity of the surfactant/bonded ligand structure, enhancing the solute diffusion, just as their presence alters the fluidity of the micellar aggregate structure in the mobile phase. 3.5.3 Effect of the Type of Organic Solvent and Organic Solvent/Surfactant Ratio on the Efficiency

The enhanced efficiency depends on the chemical characteristics of the bonded stationary phase, the type of surfactant and organic solvent, and their relative concentration in the mobile phase [49]. Most studies on this topic have been performed with SDS and, in much lesser extent, with Brij-35, which offers poorer efficiencies. In fact, with this surfactant broad peaks are usually obtained even close to the dead time. In a study on the efficiency of MLC with SDS using a large set of organic solvents as modifiers (alkanols, alkanediols, formamide, dipolar aprotic solvents, and n-hexane) and benzene and 2-ethylanthraquinone as probe compounds (made parallel to the study on the elution strength commented in Section 3.4.4.3), an increase in the efficiency was observed with increasing organic solvent log Po/w (Figure 3.8b) [19]. As can be observed (see also Figure 3.8a), efficiency and elution strength are correlated in MLC to some extent. It seems that there is no single additive that will be the best in MLC. The optimal choice is dictated by solute polarity: a less polar organic solvent is needed to attain the maximal efficiency when working with less polar solutes. Interestingly, the addition of dipolar aprotic solvents, such as acetonitrile, and the weakly protic formamide, to the micellar mobile phase results in more plate counts than those predicted from their respective log Po/w values based upon the alkanol and alkanediol data (Figure 3.9b). Although evaluated in the early days of MLC, acetonitrile was not selected as a useful candidate to enhance the efficiency because, at similar concentration, it did not improve the elution strength as much as short-chain alcohols did. However, recent work has shown the advantage of using acetonitrile as a modifier in MLC: the improved efficiencies result in better resolution for complex mixtures [12]. A fact checked consistently is that the organic solvent to surfactant ratio controls the efficiency in MLC, since it dictates the amount of surfactant coating on the stationary phase. The greater the concentration of organic solvent, the greater the efficiency and the amount of surfactant desorbed from the stationary phase, giving rise to smaller carbon loading and film thickness. Also, at fixed organic solvent concentration, the efficiency decreases as the surfactant concentration in the mobile phase increases. However, the interpretation of plate count data with added organic solvent is complex, since in parallel, the organic solvent changes the structure of micelles and the mobile phase polarity, affecting the retention.

3.5 Efficiency

3.5.4 The Case of Basic Compounds

The RPLC analysis of basic compounds (including many drugs of interest with protonable nitrogens), using conventional silica-based packings, is problematic due to poor peak shape (broad and tailed peaks) and long retention. The reason of this behavior is the ion-exchange interaction of the cationic protonated species with free silanols on the support, which is a slow process [51]. This interaction can be at least partially avoided by decreasing the pH of the mobile phase below 3.5 to suppress silanol ionization. Several manufacturers have developed a variety of deactivated packings (with protected silanols). However, because of their high cost, conventional RPLC packings are still common, and silanols are usually blocked with amine modifiers added to the mobile phase, which associate with silanol sites blocking ion-exchange processes. In recent years, the anionic surfactant SDS added to hydro-organic mixtures (in the submicellar or micellar modes) has been reported as an effective silanol suppressor, yielding for basic compounds increased efficiency, superior to that achieved with amines, and almost symmetrical peaks, while in the hydro-organic mode peak deformation is significant (and even deactivated packings do not yield symmetrical peaks) [12]. Reported examples are found for β-blockers, phenethylamines, tetracyclines, and tricyclic antidepressants [49]. The suppression of the silanol effect with SDS is not caused by direct electrostatic interaction with free silanols (case of amines), but it is a result of the protective coating of surfactant monomers on the stationary phase, which prevents very efficiently the cationic solutes penetrate the bonded alkyl chains to interact with the buried silanols. Cationic solutes instead interact electrostatically with the anionic sulfate group in the adsorbed surfactant monomers through an ion-exchange mechanism, which seems to be a fast process (easier than ion-exchange processes involving the silanols on the silica surface). Pure micellar mobile phases yield, however, poor peak shape due to the thick surfactant layer on the column. The organic solvent in the hybrid mobile phase reduces this layer and permits better diffusion of the cationic solutes, still preventing their penetration in the bonded phase to reach the silanols. The net effect seems to be more effective than the direct electrostatic interaction of amines with free silanols to improve the peak shape. Unfortunately, although the retention times may be shorter with respect to hydro-organic RPLC, these can remain still long due to the electrostatic attraction of the basic compounds to the modified stationary phase. This forces the addition of more organic solvent to reduce the retention times to practical values, reaching often HSLC conditions (Figure 3.3) [26]. Comprehensive studies have been conducted on the chromatographic behavior of a group of basic compounds (β-blockers) in mobile phases containing SDS in the submicellar and micellar modes in comparison with conventional RPLC with acetonitrile or alcohols [15,23,24]. It was found that acetonitrile can give rise to significantly better peak shape, which is interpreted by a larger reduction of the adsorbed surfactant layer on the stationary phase. The poorest efficiencies were obtained for the hydro-organic mode (N = 800–1700). Efficiencies improved

401

402

3 Micellar Liquid Chromatography: Fundamentals

in the micellar mode (N = 1000–3300). However, the most outstanding enhancements were observed in IPC (low surfactant concentration and high organic solvent content) and the high submicellar mode (high surfactant concentration and high organic solvent content), with efficiency values frequently in the range of N = 4000–9000. In the hydro-organic mode, tailed peaks were obtained (B/A = 1.5–3.0, A and B being the left and right half-widths, respectively). Meanwhile, chromatographic peaks were almost symmetrical in both micellar and high submicellar modes (B/A = 1.0–1.3), with small changes with mobile-phase composition, except by the observation of a significant deterioration in the presence of high organic solvent content (>50% acetonitrile). This can be interpreted by the loss of a significant portion of the surfactant layer from the stationary phase, which favors solute penetration and interaction with the buried silanols. As observed, the interpretation of the interactions that take place inside alkyl-bonded silica columns, in the presence of SDS (and other additives) using hybrid mobile phases, is relatively simple by following the changes in retention and peak performance. The enhancement in peak shape itself offers a measurement of silanol suppression. Also, retention and peak performance allows probing the surfactant layer on the stationary phase: good peak shape (nearly symmetrical narrow peaks) confirms the coating of the stationary phase by the anionic surfactant, which gives rise to negligible interaction with free silanols. Thus, the long retention times and high efficiencies found with a Kromasil C18 column and mobile phases containing SDS and 50–60% methanol suggested that a significant amount of surfactant still covered the stationary phase. For up to 35% 1-propanol or 50% acetonitrile, the surfactant layer was not either desorbed totally [23]. This agrees with a previous observation on the tight insertion of the surfactant alkyl chains in the alkyl moieties of the bonded layer of the densely grafted phases.

3.6 Significance of MLC

From its beginnings in 1980, MLC has evolved into becoming a real alternative in some analyses (and a complement in others) to conventional RPLC, owing to its peculiar features. The first developments in MLC were based on the works of Armstrong, Berthod, Cline-Love, Dorsey, Khaledi, Hinze, and Foley, among other researchers. After these efforts, MLC was more or less forgotten, but since the 1990s, further work demonstrated its usefulness in modern pharmaceutical and food analyses. Today, there is more than three decades of MLC experience with a great volume of scientific production, which amounts to several hundreds of reports. A number of qualities have been revealed as the technique has developed. First, the biodegradability of surfactants and the small amount of organic solvent in the mobile phase makes MLC an RPLC mode with lower toxicity and reduced environmental impact, which may be classified inside the group of “green analytical

References

techniques”. Also, the stabilization of the organic solvent in the micellar medium reduces the risk of evaporation. This property yields highly reproducible retention facilitating the optimization of the separation conditions, since the mobile phases are preserved in the laboratory for a long time without significant changes in their composition. In addition, the mobile phases can be recycled during the analysis, as long as a sufficiently small number of injections are made, which reduces the cost. Also remarkable is the large variety of interactions between solutes, micelles, and the adsorbed surfactant monomers on the stationary phase, which produce particular selectivity and give rise to a highly versatile technique, appropriate for a wide range of solutes (from hydrophilic to hydrophobic). The analyses can be usually carried out within a single isocratic run in adequate analysis times. Finally, the feasibility of direct injection of physiological fluids is really attractive, since it avoids the tedious sample pretreatment required when hydro-organic mobile phases are used [52]. This topic will be discussed in detail in Chapter 4. Despite these features and others that are discussed in this chapter or in Chapter 4, the applicability of MLC in the analytical laboratories is still limited. The reason may be, partially, the main drawbacks early described in the first reports on the technique: the weak elution strength and reduced efficiency of pure micellar mobile phases (containing only surfactant). Some of the solutions proposed along the years to overcome these limitations, such as the addition of a small amount of organic solvent, the use of large-pore stationary phases or high temperature, have favored the potential of the technique, but they have not resulted in an extensive use. However, it seems that the only real limitation of this chromatographic mode to date is related to the impossibility of direct online coupling to mass spectrometric detection, due to the presence of high concentrations of surfactant in the mobile phase. References 1 García-Alvarez-Coque, M.C. and Torres-

4 Armstrong, D.W. and Henry, S.J. (1980)

Lapasió, J.R. (2013) Secondary chemical equilibria in reversed-phase liquid chromatography, in Liquid Chromatography: Fundamentals and Instrumentation, Chapter 4 (eds S. Fanali, P. Haddad, C.F. Poole, P.J. Schoenmakers, and D. Lloyd), Elsevier, Amsterdam, pp. 87–104. 2 Knox, J.H. and Laird, G.L. (1976) Soap chromatography. A new HPLC technique for separation of ionizable materials: dyestuff intermediates. J. Chromatogr., 122, 17–34. 3 Cecchi, T. (2009) Ion-Pair Chromatography and Related Techniques, CRC Press, Boca Raton, FL.

Use of an aqueous micellar mobile phase for separation of phenols and polynuclear aromatic hydrocarbons via HPLC. J. Liq. Chromatogr., 3, 657–662. 5 Berthod, A. and García-AlvarezCoque, M.C. (2000) Micellar Liquid Chromatography, Chromatographic Science Series, vol. 89 (ed. J. Cazes), Marcel Dekker, New York, NY. 6 Yarmchuck, P., Weinberger, R., Hirsch, R.F., and Cline-Love, L.J. (1982) Selectivity in liquid chromatography with micellar mobile phases. Anal. Chem., 54, 2233–2238.

403

404

3 Micellar Liquid Chromatography: Fundamentals 7 Yarmchuck, P., Weinberger, R., Hirsch,

8

9

10

11

12

13

14

15

R.F., and Cline-Love, L.J. (1984) Effect of restricted mass transfer on the efficiency of micellar chromatography. J. Chromatogr., 283, 48–60. Ruiz-Angel, M.J., García-Alvarez-Coque, M.C., and Berthod, A. (2009) New insights and recent developments in micellar liquid chromatography. Sep. Purif. Rev., 38, 45–96. Dorsey, J.G., DeEchegaray, M.T., and Landy, J.S. (1983) Efficiency enhancement in micellar liquid chromatography. Anal. Chem., 55, 924–928. Torres-Lapasió, J.R., Baeza-Baeza, J.J., and García-Alvarez-Coque, M.C. (1997) Description of the retention behaviour in micellar liquid chromatography as a function of pH, surfactant and modifier concentration. J. Chromatogr. A, 769, 155–168. Ruiz-Angel, M.J., Torres-Lapasió, J.R., and García-Alvarez-Coque, M.C. (2004) Effects of pH and the presence of micelles on the resolution of diuretics by reversed-phase liquid chromatography. J. Chromatogr. A, 1022, 51–65. Ruiz-Angel, M.J., Carda-Broch, S., and García-Alvarez-Coque, M.C. (2010) Peak half-width plots to study the effect of organic solvents on the peak performance of basic drugs in micellar liquid chromatography. J. Chromatogr. A, 1217, 1786–1798. Escuder-Gilabert, L., Martínez-Pla, J.J., Sagrado, S., Villanueva-Camañas, R.M., and Medina-Hernández, M.J. (2003) Biopartitioning micellar separation methods: modelling drug absorption. J. Chromatogr. B, 797, 21–35. Fernández-Navarro, J.J., Ruiz-Angel, M.J., and García-Alvarez-Coque, M.C. (2012) Reversed-phase liquid chromatography without organic solvent for determination of tricyclic antidepressants. J. Sep. Sci., 35, 1303–1309. Ruiz-Angel, M.J., Torres-Lapasió, J.R., García-Alvarez-Coque, M.C., and CardaBroch, S. (2008) Retention mechanisms for basic drugs in submicellar and micellar reversed-phase liquid chromatographic modes. Anal. Chem., 80, 9705–9713.

16 Cline-Love, L.J. and Fett, J.J. (1991)

17

18

19

20

21

22

23

24

Optimization of selectivity in micellar chromatographic procedures for the determination of drugs in urine by direct injection. J. Pharm. Biomed. Anal., 9, 323–333. Nakao, R. and Halldin, C. (2013) Mixed anionic and non-ionic micellar liquid chromatography for high-speed radiometabolite analysis of positron emission tomography radioligands. J. Chromatogr. A, 1281, 54–59. Borgerding, M.F., Hinze, W.L., Stafford, L.D., Fulp, G.W. Jr., and Hamlin, W.C. Jr. (1989) Investigations of stationary phase modification by the mobile phase surfactant in micellar liquid chromatography. Anal. Chem., 61, 1353–1358. López-Grío, S., García-Alvarez-Coque, M.C., Hinze, W.L., Quina, F.H., and Berthod, A. (2000) Effect of a variety of organic additives on retention and efficiency in micellar liquid chromatography. Anal. Chem., 72, 4826–4835. Boichenko, A.P., Kulikov, A.U., Loginova, L.P., and Iwashchenko, A.L. (2007) Aliphatic carboxylic acids as new modifiers for separation of 4-dinitrophenyl amino acids by micellar liquid chromatography. J. Chromatogr. A, 1157, 252–259. Mao, C., McGill, K.E., and Tucker, S.A. (2004) Optimization of micellar liquid chromatographic separation of polycyclic aromatic hydrocarbons with the addition of second organic additive. J. Sep. Sci., 27, 991–996. López-Grío, S., Baeza-Baeza, J.J., and García-Alvarez-Coque, M.C. (1998) Influence of the addition of modifiers on solute–micelle interaction in hybrid micellar liquid chromatography. Chromatographia, 48, 655–663. Ruiz-Angel, M.J., Carda-Broch, S., and García-Alvarez-Coque, M.C. (2010) Effect of short-chain alcohols on surfactantmediated reversed-phase liquid chromatographic systems. J. Chromatogr. A, 1217, 7082–7089. Ruiz-Angel, M.J., Torres-Lapasió, J.R., García-Alvarez-Coque, M.C., and

References

25

26

27

28

29

30

31

32

33

34

Carda-Broch, S. (2009) Submicellar and micellar reversed-phase liquid chromatographic modes applied to the separation of β-blockers. J. Chromatogr. A, 1216, 3199–3209. Li, X. and Fritz, S. (1996) Novel additives for the separation of organic compounds by high-performance liquidchromatography. J. Chromatogr. A, 728, 235–247. Ruiz-Angel, M.J., Carda-Broch, S., and García-Alvarez-Coque, M.C. (2014) High submicellar liquid chromatography. Sep. Purif. Rev., 43, 124–154. Berthod, A. and Roussel, A. (1988) The role of the stationary phase in micellar liquid chromatography. J. Chromatogr., 449, 349–360. Berthod, A., Girard, I., and Gonnet, C. (1986) Micellar liquid chromatography: adsorption isotherms of two ionic surfactants on five stationary phases. Anal. Chem., 58, 1356–1358. Lavine, B.K. and Hendayana, S. (1996) Band broadening in micellar liquid chromatography. J. Liq. Chromatogr. Rel. Technol., 19, 101–123. Ruiz-Angel, M.J., Torres-Lapasió, J.R., Carda-Broch, S., and García-AlvarezCoque, M.C. (2003) Improvement of peak shape and separation performance of β-blockers in conventional reversed-phase columns using solvent modifiers. J. Chromatogr. Sci., 41, 350–358. Dorsey, J.G., Khaledi, M.G., Landy, J.S., and Lin, J.L. (1984) Gradient elution micellar liquid chromatography. J. Chromatogr., 316, 183–191. Ruiz-Angel, M.J., Caballero, R.D., SimóAlfonso, E.F., and García-Alvarez-Coque, M.C. (2002) Micellar liquid chromatography: a suitable technique for screening analysis. J. Chromatogr. A, 947, 31–45. McCormick, T.J., Foley, J.P., Riley, C.M., and Lloyd, D.K. (2000) The effect of stationary-phase pore size on retention behavior in micellar liquid chromatography. Anal. Chem., 72, 294–301. Detroyer, A., Vander Heyden, Y., Reynaert, K., and Massart, D.L. (2004) Evaluating “fast” micellar monolithic liquid

35

36

37

38

39

40

41

42

43

44

chromatography for high-throughput quantitative structure–retention relationship screening. Anal. Chem., 76, 1903–1908. Berthod, A. (1997) Causes and remediation of reduced efficiency in micellar liquid chromatography. J. Chromatogr. A, 780, 191–206. Ruiz-Angel, M.J., Carda-Broch, S., TorresLapasió, J.R., and García-Alvarez-Coque, M.C. (2009) Retention mechanisms in micellar liquid chromatography. J. Chromatogr. A, 1216, 1798–1814. Armstrong, D.W. and Nome, F. (1981) Partitioning behavior of solutes eluted with micellar mobile phases in liquid chromatography. Anal. Chem., 53, 1662–1666. Torres-Lapasió, J.R., Ruiz-Angel, M.J., and García-Alvarez-Coque, M.C. (2008) Micellar versus hydro-organic reversedphase liquid chromatography: a solvation parameter-based perspective. J. Chromatogr. A, 1182, 176–196. Marina, M.L. and García, M.A. (1997) Evaluation of distribution coefficients in micellar liquid chromatography. J. Chromatogr. A, 780, 103–116. Baeza-Baeza, J.J., Dávila, Y., FernándezNavarro, J.J., and García-Alvarez-Coque, M.C. (2012) Measurement of the elution strength and peak shape enhancement at increasing modifier concentration and temperature in RPLC. Anal. Bioanal. Chem., 404, 2973–2984. Armstrong, D.W. and Stine, G.Y. (1983) Selectivity in pseudophase liquid chromatography. Anal. Chem. 55, 2317– 2320. Armstrong, D.W. and Stine, G.Y. (1983) Evaluation and perturbation of micelle– solute interactions. J. Am. Chem. Soc., 105, 6220–6223. Borgerding, M.F., Quina, F.H., Hinze, W.L., Bowermaster, J., and McNair, H.M. (1988) Investigation of the retention mechanism in nonionic micellar liquid chromatography using an alkylbenzene homologous series. Anal. Chem., 60, 2520–2527. Khaledi, M.G., Peuler, E., and NgehNgwainbi, J. (1987) Retention behavior of homologous series in reversed-phase

405

406

3 Micellar Liquid Chromatography: Fundamentals

45

46

47

48

liquid chromatography using micellar, hydro-organic and hybrid mobile phases. Anal. Chem., 59, 2738–2747. Hinze, W.L. and Weber, S.G. (1991) Why the relationship between the logarithm of k and homologue number in MLC is not linear? Anal. Chem., 63, 1808–1814. García-Alvarez-Coque, M.C. and TorresLapasió, J.R. (1999) Quantitation of hydrophobicity in micellar liquid chromatography. Trend. Anal. Chem., 18, 533–543. Dorsey, J.G. and Khaledi, M.G. (1993) Hydrophobicity estimations by reversedphase liquid chromatography. Implications for biological partitioning processes. J. Chromatogr. A, 656, 485–499. Ruiz-Angel, M.J., Carda-Broch, S., GarcíaAlvarez-Coque, M.C., and Berthod, A. (2004) Micellar versus hydro-organic mobile phases for retention– hydrophobicity relationship studies with

49

50

51

52

ionizable diuretics and an anionic surfactant. J. Chromatogr. A, 1030, 279–288. Ruiz-Angel, M.J., Carda-Broch, S., and García-Alvarez-Coque, M.C. (2013) Chromatographic efficiency in micellar liquid chromatography: should it be still a topic of concern? Sep. Purif. Rev., 42, 1–27. Tomasella, F.P., Fett, J., and Cline-Love, L.J. (1991) Effects of organic modifier and temperature on micellar liquid chromatography. Anal. Chem., 63, 474–479. Nawrocki, J. (1997) The silanol group and its role in liquid chromatography. J. Chromatogr. A, 779, 29–71. García-Alvarez-Coque, M.C. and CardaBroch, S. (1999) Direct injection of physiological fluids in micellar liquid chromatography. J. Chromatogr. B, 736, 1–18.

407

4 Micellar Liquid Chromatography: Method Development and Applications Maria C. García-Alvarez-Coque, Maria J. Ruiz-Angel, and Samuel Carda-Broch

4.1 Reasons for Using Micellar Mobile Phases

In spite of the problems found in its initial development, after more than three decades of experience, micellar liquid chromatography (MLC) seems to be an alternative to conventional RPLC with hydro-organic mobile phases, with increasing interest in “green” chemistry [1–3]. Most MLC procedures use hybrid micellar mobile phases containing a surfactant above the critical micellar concentration (CMC) and a relatively small amount of organic solvent to increase the elution strength and improve the efficiencies. There are several reasons for using micellar mobile phases in RPLC: i) The variety of interactions between solutes, stationary phase, hydroorganic phase, and micelles, which give rise to unique selectivities [4]. ii) The possibility of separating charged and neutral solutes in a single run or solutes in a wide polarity range with retention time windows narrower than in conventional RPLC [5]. Chromatographic peaks in MLC appear usually more evenly distributed, which makes the use of gradients less necessary. However, in case gradient elution is needed, equilibration times are shorter than those with conventional RPLC [6]. iii) The high solubilization capability of micelles, which facilitates dissolution of most matrices. This saves time in sample preparation, since the direct on-column injection of physiological fluids, or other liquid samples containing proteins, is possible, eliminating the tedious sample pretreatment required in conventional RPLC [7,8]. iv) Hybrid eluents still have the advantage of requiring significantly smaller amounts of organic solvent with respect to conventional RPLC, especially for nonpolar compounds. This means lower cost, and lower toxicity and environmental impact of hazardous wastes. This is attractive considering the increasing restriction on the use of organic solvents in the laboratories.

Analytical Separation Science, First Edition. Edited by Jared L. Anderson, Alain Berthod, Verónica Pino Estévez, and Apryll M. Stalcup.  2015 Wiley-VCH Verlag GmbH & Co. KGaA. Published 2015 by Wiley-VCH Verlag GmbH & Co. KGaA.

408

4 Micellar Liquid Chromatography: Method Development and Applications

v) The stabilization of organic solvents by micelles in the mobile phase reduces the risk of evaporation, making the mobile phases stable for a longer time. vi) The retention is highly reproducible and can be modeled accurately to predict changes in retention times with mobile-phase composition (pH, and concentration of surfactant and organic solvent) [9]. This facilitates the optimization of the separation conditions. vii) Several types of detection are enhanced due to the compartmentalization of organic compounds by micelles [10–13]. All these features have allowed the development of multiple MLC analytical applications, which are competitive or complementary to conventional RPLC. All basic knowledge related to MLC is described in Chapter 3. This chapter is devoted to the practical aspects of this chromatographic mode in method development.

4.2 How to Work in MLC

MLC uses the same hardware and columns as conventional RPLC. However, the routine work with hybrid micellar mobile phases requires keeping some cautions usually not described in scientific reports, which should be followed to preserve the column performance for long time periods of intensive use. Practical indications a chromatographer should consider in MLC when developing an analytical procedure are described next. If these are followed, the life of the hardware and column can be comparable or even longer than those of conventional RPLC. Coating of columns with surfactant yields stable stationary phases that can be used for other purposes, after proper regeneration. 4.2.1 About the Mobile Phase 4.2.1.1

The Surfactant

The most usual surfactants in MLC are sodium dodecyl sulfate (SDS), cetyltrimethylammonium bromide (CTAB), and polyoxyethylene(23)dodecyl ether (Brij-35) [1,2]. In MLC, the range of surfactant concentration in the mobile phase is usually narrow. The lowest concentration should be well above the CMC to form the micelles (8.2 × 10 3 M for SDS, 9 × 10 4 M for CTAB, and 9 × 10 5 M for Brij-35), and concentrations exceeding 0.20 M are not convenient due to the high viscosity of the mobile phase and degradation of the efficiency. Usual ranges are 0.05–0.15 M for SDS, 0.04–0.1 M for CTAB, and 0.01–0.05 M for Brij-35. For ionic surfactants, the chromatographic work should be conducted above the so-called Krafft point (temperature at which the solubility of surfactant monomers equals the CMC) to avoid surfactant precipitation and possible ruin of the column and chromatographic system The Krafft point

4.2 How to Work in MLC

for SDS and CTAB is around 15 °C and 20–25 °C, respectively, but these values are affected by the presence of salts. Nonionic surfactants have also a temperature above which phase separation occurs, which is called the cloud point. However, this seems not to be a problem, since the cloud point for the most common nonionic surfactant in MLC (Brij-35) is about 100 °C for aqueous 0.008–0.05 M solutions. Most separations are performed in buffered media, with phosphoric and citric acids being the most appropriate acid–base systems. The pH of the mobile phase should be measured after the addition of the organic solvent. For mobile phases containing SDS, potassium salts cannot be used, since potassium dodecyl sulfate presents a high Krafft point and precipitates from aqueous solutions at room temperature. 4.2.1.2

The Organic Solvent

Organic solvents, such as alcohols, added to the micellar mobile phases enhance the efficiency of chromatographic peaks (which is rather low when only the surfactant is present) and permit the accurate control of retention and selectivity [9,14] (see also Chapter 3). The elution strength of alcohols depends on their chain length. 1-Propranol is the most usual choice. Methanol and ethanol are rarely used due to their low elution strength. Strongly retained compounds require a small amount of 1-butanol or 1-pentanol. The elution strength of acetonitrile is below that of 1-propanol, but efficiency enhancements are often larger with acetonitrile [15]. The concentration of organic solvent must be low enough to guarantee the integrity of micelles. This depends on the type of surfactant and organic solvent. For SDS, the maximal volume fractions of methanol, acetonitrile, 1-propanol, 1-butanol, and 1-pentanol are approximately 30–40, 30, 22, 10, and 6% v/v, respectively [16]. However, analytical reports where these maximal values are exceeded (i.e., micelles do not exist) are not unusual. In fact, the weak elution strength of methanol forces the use of high concentrations of this alcohol to achieve sufficiently short analysis times, giving rise to submicellar conditions (i.e., working in the high submicellar liquid chromatographic mode, HSLC see Chapter 3). Usually, analysts working in MLC assay several organic solvents before selecting the most appropriate. However, the selection of the organic solvent is relatively simple, since it depends primarily on solute polarity. The following guidelines have been given for SDS mobile phases [5]: a low volume fraction of propanol (∼1%) is needed to separate polar compounds with octanol–water partition coefficients, log Po/w < 1; a larger concentration of this solvent (∼5–7%) should be added for compounds in the range 1 < log Po/w < 2; and other alcohols ( 3. For positively charged basic drugs with 1 < log Po/w < 3, 1-propanol is relatively weak due to the association of the protonated solutes with SDS adsorbed on the stationary phase. In this case, a high concentration of 1-propanol (∼15%) or 1-butanol ( 30 °C and pH > 6. Therefore, the mobile phase should be saturated with silica by placing a short precolumn packed with bare silica, or alternatively, the same packing as the analytical column located before the injection valve. With convenient experimental cautions, hundreds of injections can be made without modification of the chromatographic system or pressure buildup [8]. 4.2.2.1

Column Conditioning

Before starting column conditioning with the micellar mobile phase, methanol should be replaced by water to avoid crystallization of the surfactant inside the system (at least 30 column volumes of water are required to assure complete removal of methanol). For this operation, the eluent should be pumped at low flow rate (< 0.5 ml/min). Once the pressure decreases, the flow rate may be increased. Now, the system is ready to be flushed with the micellar mobile phase to assure equilibration of the column. Different studies on column coating using surfactant breakthrough patterns have revealed that most surfactant is adsorbed in less than one hour on the bonded stationary phase by pumping the mobile phase at 1 ml/min. Column saturation is often reached at the CMC or close to it. However, some additional surfactant may be adsorbed on the column above the CMC, as is the case of the nonionic Brij-35 [17,18]. 4.2.2.2

Mobile Phase Flushing

The micellar solution should never stay motionless in a chromatographic system to avoid the formation of surfactant crystals that could ruin the column and instrument. Therefore, if it is stopped during several hours, the micellar solution should be replaced by water. If the pump is working, the micellar mobile phase can be kept inside the chromatographic system along several hours, and even overnight (also, along several days). In this case, the flow can be reduced to a minimal value (e.g., 0.1–0.25 ml/min). This is even advisable, since this avoids daily cleaning and re-equilibration. In order to reduce the cost, the mobile phase can be recycled between analyses. Recycling can be carried out also during the analyses, as long as a sufficiently small number of injections are made. This is possible due to the low evaporation of organic solvents in micellar media. 4.2.2.3

Column Regeneration

After one or two weeks of continuous experimental work, the elimination of strongly retained compounds may be convenient by substitution of the micellar

4.3 Optimization of Experimental Conditions

mobile phase with methanol, which is also required when the experimental work is finished. This should be made following the procedure explained in Section 4.2.2.2 in reverse: the micellar mobile phase should be removed first with 10–20 column volumes of water to assure that no surfactant remains in the system. Next, methanol (where most surfactants are highly soluble) or a 75 : 25 methanol–propanol mixture is pumped at low flow rate to remove the adsorbed surfactant. In order to assure complete surfactant desorption, at least 10 column volumes of organic solvent should be passed through the column at a low flow rate. There is some concern about complete surfactant desorption from the stationary phase to recover the original column conditions, existing some inconsistency in the data available in the literature. This is the reason of the recommendation of keeping a column for the exclusive use of a given surfactant. It seems that the layer of SDS monomers on C8 and C18 stationary phases can be removed completely, but this is not the case for CTAB and Brij-35 [17,18]. Therefore, a third final step can be included to check the recovery of the initial stationary-phase surface, consisting in the measurement of the retention times of a solute mixture with a hydroorganic mobile phase before and after using the micellar mobile phase.

4.3 Optimization of Experimental Conditions 4.3.1 Need of Interpretive Strategies

In MLC, the retention behavior (elution strength and selectivity) can be quite different from conventional RPLC with hydro-organic mixtures, even after the addition of an organic solvent. The range of interactions provided by MLC is far superior, which is very attractive. However, the need to control at least two factors (the concentrations of surfactant and organic solvent), which interact with each other, gives rise to several local optima. This may result in hard optimization protocols of the chromatographic conditions, especially if a sequential strategy based on the retention observed with previous mobile phases is applied. The solution is to optimize simultaneously all selected factors using a computerassisted interpretive optimization strategy (based on the description of the retention behavior and peak shape of solutes inside a selected factor space) [19,20]. This solution is much more efficient and reliable, and requires fewer experiments than trial-and-error strategies to derive an acceptable separation. An interactive computer program (MICHROM) was developed and marketed in 2000 to help in method development in MLC [21,22]. To achieve good predictions of the retention behavior, accurate models describing the retention are needed. Fortunately, although the retention and selectivity may strongly change with pH and the concentrations of surfactant and organic modifier, the observed changes are highly regular, being well described by relatively simple models [9,20,23]. In the first step of the

411

412

4 Micellar Liquid Chromatography: Method Development and Applications

optimization protocol, the chromatographer develops a limited number of experiments covering a reasonably wide factor region, to gather information about the chromatographic behavior of the compounds in the sample. This is used to fit equations that allow the prediction of retention and peak shape under any other condition. In the second step, chromatograms are simulated by adding the predicted signals for the compounds in the analyzed mixture, under a large number of separation conditions for a prefixed distribution of the experimental factors. The condition offering maximal resolution or an acceptable analysis time is selected. The tools for modeling the peak shape and resolution in MLC are the same as those for conventional RPLC [20]. However, the retention in MLC needs particular models that are described next. 4.3.2 Modeling the Retention Behavior in MLC

A three-phase (stationary phase, water, and micelle) model was established early in the development of MLC (Figure 4.1a) [24], which was further extended to hybrid micellar mobile phases (Figure 4.1b) [25]. These models gave rise to the proposal of empirical and mechanistic equations that describe accurately the changes in solute retention at fixed and variable pH. 4.3.2.1

Mechanistic Models for Pure Micellar Mobile Phases

Armstrong and Nome [24], Arunyanart and Cline-Love [26], and Foley [27] suggested three different approaches to describe the retention in MLC as a function of the concentration of surfactant monomers involved in micelle formation ([M], total concentration of surfactant minus the CMC). The main objective of the authors in proposing such equations was to measure the strength of solute– micelle and solute–stationary-phase interactions (Figure 4.1a). This kind of models is also useful to predict the retention for optimization purposes. The approach of Armstrong and Nome, published in 1981, is based on the transitions among the three environments that exist in a micellar chromatographic system [24]: bulk aqueous solvent (A), micelles (M), and stationary phase (S), described by the partition coefficients of solute between the stationary phase and the bulk aqueous solvent (PAS), and between the micelle and the bulk aqueous solvent (PAM): kˆ

ΦP AS 1 ‡ υ…P AM 1†‰MŠ

(4.1)

where Φ = VS/V0 is the phase ratio (VS being the volume of active surface on the stationary phase and V0 the column void volume), and υ is the partial specific volume of surfactant monomers in the micelle. Arunyanart and Cline-Love considered instead the association equilibria of solute in bulk aqueous solvent with the stationary phase binding sites or the micelle, governed by the solute–stationary phase (KAS) and solute–micelle association (KAM) constants, respectively (Figure 4.1a) [26]:

4.3 Optimization of Experimental Conditions

O

O

O

SO 3 O

O

SO 3

SO 3

O

O

SO 3

O Si

O Si

O3 - S SO 3

O

O

Si

Si

O

O

O

SO 3 O

O

Si

SO 3 O

O SO3 S 3

O

O

O - S SO 3

O

O

O

3– C N

O

O

SO 3

SO3-

OSi

SO3-

O

O

–C

N SO 3 O

SO 3 O

KSD

SO3-

O

Si

SO 3

CH 3

SO3-

O Si

SO 3

KAD

SO 3 O

CH

SO 3

KMD

KAS KAD SO3O

O SO 3

O

KAM

SO 3 SO 3

O

O

O

O

CH3 – C N

Si

O

O

SO 3

SO 3

SO3-

O3 - S SO 3

O Si

O

SO 3

O

SO3O

O

O

O

OSi

SO3-

O

O Si

SO 3

SO 3 O

SO3-

O

SO 3

O

O

O

SO3-

SO 3

O

SO3-

O

O

O - S SO 3

O

SO3-

SO 3

OSi

SO 3

SO 3

O Si

O

SO 3

(b)

SO 3

SO 3 O

O

SO3-

O

O

O O

O

SO3-

SO3-

O

SO 3

KAS

Si

O

O

SO3

KAM

O

O

O

SO 3 O

SO 3

SO 3 O

SO3-

O3 - S SO 3

SO 3

SO 3

SO 3

(a)

SO3-

O

O

SO3O

SO3O

SO3O

OO

Si

O

Si

O

Si

O

Si

O

Si

O

Si

O

Si

O

Si

O

Si

Figure 4.1 Simplified illustration of the three-phase chromatographic systems using a C18 stationary phase and SDS pure micellar mobile phase (a), and SDS/acetonitrile hybrid micellar mobile phase (b).

413

414

4 Micellar Liquid Chromatography: Method Development and Applications

kˆΦ

‰ASŠ ΦK AS ‰S Š ˆ ‰AŠ ‡ ‰AMŠ 1 ‡ K AM ‰MŠ

(4.2)

Since the stationary-phase activity usually does not change by changing the concentration of surfactant, the parameters in the numerator of Equation 4.2 are usually combined as kˆ

K ´AS 1 ‡ K AM ‰MŠ

(4.3)

Finally, Foley considered the association between the solute in bulk aqueous solvent and the micelles as a secondary equilibrium, which affects the solute retention in the absence of micelles (kA) [27]: k ˆ kA

1 1 ‡ K AM ‰MŠ

(4.4)

Based on different perspectives, the three parabolic models (Equations 4.1, 4.3 and 4.4) are similar. For convenience, these are rewritten to express a linear dependence. From Equation 4.3, 1 1 K AM 1 ‡ K AM ‰MŠ ‡ ˆ c0 ‡ c1 ‰M Š ˆ ‰M Š ˆ k K AS K AS K AS

(4.5)

where KAS is used instead of K ´AS . A zero intercept (high KAS value) suggests a direct transfer between the micelles and the modified stationary phase. Equation 4.5 has been verified experimentally for a variety of solutes (ionic, neutral, polar, and nonpolar), surfactants (anionic, cationic, and nonionic), and column packings (C8, C18, and cyano), with experimental errors usually below 3–5%. To obtain these results, the following assumptions should be valid: i) The surfactant concentration in the mobile phase is above the CMC and not affected by the small amount of solute in the chromatographic system. ii) The aggregation number and geometry of micelles is not altered with an increasing surfactant concentration, which would affect the KAM values. iii) The stationary phase is saturated with the adsorbed surfactant. In fact, a linear plot of 1/k versus [M] is an evidence of saturation. An important factor that should be considered to calculate KAM and KAS from MLC data is the correct subtraction of the CMC from the total concentration of surfactant. Otherwise, significant errors will be obtained. However, there is no significant difference between using the total or the micellar concentration of surfactant to predict retention factors for optimization purposes.

4.3 Optimization of Experimental Conditions

4.3.2.2

Empirical and Mechanistic Models for Hybrid Micellar Mobile Phases

Equation 4.5 describes also the effect of micelles with hybrid micellar mobile phases containing surfactant and a fixed concentration of organic solvent. The organic solvent added to the micellar mobile phase shifts the association equilibria away from the stationary phase and micelle toward the bulk solvent, which is less polar (Figure 4.1b). This decreases KAM and KAS, especially for highly hydrophobic solutes. However, the KAM/KAS ratio increases, due to the significant reduction of the surfactant layer on the stationary phase. Therefore, the mobile phase is stronger. The organic solvent effect on retention depends on the ratio between the concentrations of surfactant and organic solvent, and is different for different solutes. The linear and quadratic relationships between log k and volume fraction of organic modifier, φ, followed in conventional RPLC, are also valid in hybrid MLC at fixed micellar concentration: log k ˆ log k 0

Sφ ˆ c0 ‡ c1 φ

log k ˆ c0 ‡ c1 φ ‡ c11 φ2

(4.6) (4.7)

where c0, c1, and c11 are regression coefficients with characteristic values for a given solute and column/solvent system. Obviously, the simultaneous optimization of surfactant and organic solvent needs a model containing both factors [9]. The models correlating log k with [M] and φ have been demonstrated to be poorer than those correlating 1/k with both factors. The simplest model is 1 ˆ c0 ‡ c1 ‰MŠ ‡ c2 φ k

(4.8)

which allows good descriptions only in small regions of the factor space. In wider ranges, new terms including both surfactant and modifier concentrations are needed. Good predictions are obtained for polar and moderately polar compounds with 1 ˆ c0 ‡ c1 ‰MŠ ‡ c2 φ ‡ c12 ‰MŠφ k

(4.9)

Highly hydrophobic compounds need an additional term: 1 ˆ c0 ‡ c1 ‰MŠ ‡ c2 φ ‡ c12 ‰MŠφ ‡ c22 φ2 k

(4.10)

The regression coefficients in these empirical equations have been related to physicochemical constants [19,25,28]. Based on Equation 4.5, and considering the effects of the organic solvent on the micelles and stationary phase modified

415

416

4 Micellar Liquid Chromatography: Method Development and Applications

by adsorption of surfactant (see Chapter 3), the following model was proposed: 1 1 ‡ K AM …1 ‡ K MD φ†=…1 ‡ K AD φ†‰MŠ 1 ‡ K φAM ‰MŠ ˆ ˆ k K AS …1 ‡ K SD φ†=…1 ‡ K AD φ† K φAS

(4.11)

where K φAM and K φAS are apparent constants that depend on the concentration of organic solvent. The constants KMD and KAD account for the shift in the equilibrium between micelle and solvent, and KMD and KAD describe the shift in the equilibrium between stationary phase and solvent, due to the decrease in solvent polarity and the change in the nature of micelle and stationary phase, in the presence of organic solvent, with respect to the pure micellar solution (Figure 4.1b). The KSD term is needed only for highly nonpolar solutes, which are strongly associated with the stationary phase. In other cases, KSD = 0, and Equation 4.11 is simplified to (compare with Equation 4.5): 1 1 K AM …1 ‡ K AD φ† ‡ …1 ‡ K MD φ†‰MŠ ˆ c0 ‡ c1 ‰MŠ ‡ c2 φ ‡ c12 ‰MŠφ ˆ K AS k K AS (4.12) The parameters in Equations 4.11 and 4.12 should be obtained by fitting the data from at least four and five mobile phases, respectively (distributed in the corners and center of a two-dimensional space defined by surfactant and modifier), taking into account the practical limitations of the chromatographic system.

4.3.2.3

Modeling in a Three-Factor Space of Surfactant, Organic Solvent, and pH

For the separation of weak organic acids and bases with micellar mobile phases, it is usual to fix the pH at a preselected value and only optimize the concentration of surfactant and modifier. However, when the analyzed mixture contains one or more ionizable compounds, pH tuning can offer unique opportunities to improve the resolution. The drawback is that when optimizing this factor in MLC, the complexity of the optimization increases, since changes in the concentrations of organic solvent and surfactant affect the acid–base behavior of solutes, due to the different partitioning of the acidic and basic species [23]. For compounds exhibiting acid–base behavior, the retention in pure micellar medium can be obtained following the Arunyanart and Cline-Love approach (see Equation 4.2) [29]: kˆΦ

‰ASŠ ‡ ‰HASŠ ‰AŠ ‡ ‰HAŠ ‡ ‰AMŠ ‡ ‰HAMŠ

(4.13)

where [A], [AM], and [AS] refer to the basic species, and [HA], [HAM] and [HAS] to the acidic species. For a monoprotic system, the apparent protonation constant in the micellar medium is expressed by

4.3 Optimization of Experimental Conditions

KH ˆ

1 ‰HAŠ ˆ ‰AŠh K a

(4.14)

h being the concentration of hydrogen ion and Ka the acidic constant. By combining Equations 4.13 and 4.14, and considering the association equilibria relating all involved species to the stationary phase and micelles, the following results (see also Equation 4.3) are obtained: K AS ‡ K HAS K H h (4.15) 1 ‡ K H h ‡ K AM ‰MŠ ‡ K HAM K H h‰M Š By dividing numerator and denominator in Equation 4.15 by (1 + KH h) [23,30]: kˆ

kˆ

…K AS ‡ K HAS K H h†=…1 ‡ K H h† KH AS ˆ 1 ‡ …K AM ‡ K HAM K H h†‰MŠ=…1 ‡ K H h† 1 ‡ K H AM ‰M Š

(4.16)

H where K H AS and K AM are apparent binding constants that depend on the pH. Alternatively, by dividing Equation 4.15 by (1 + KAM [M]):   K AS K HAS 1 ‡ K HAM ‰MŠ ‡ K Hh k A ‡ k HA K M 1 ‡ K AM ‰MŠ 1 ‡ K HAM ‰MŠ 1 ‡ K AM ‰MŠ Hh   kˆ ˆ 1 ‡ K HAM ‰MŠ h 1 ‡ KM H 1‡ K Hh 1 ‡ K AM ‰MŠ (4.17) is an apparent protonation constant that depends on the concentrawhere K M H tion of micellized surfactant. In the right-side term of Equation 4.17, the retention factor appears as a weighted mean of the retention of the basic (A) and acidic (HA) species. A further step is to describe the simultaneous effect of the three factors (surfactant, organic solvent, and pH) on the retention of ionizable compounds. Based on both Equations 4.17 and 4.11, after eliminating the KSD term due to the significant polarity of weak acids and bases, we can obtain

K AS =…1 ‡ K AD φ† K HAS =…1 ‡ K HAD φ† ‡ κK H h 1 ‡ K AM ‰M Š…1 ‡ K MD φ†=…1 ‡ K AD φ† 1 ‡ K HAM ‰M Š…1 ‡ K HMD φ†=…1 ‡ K HAD φ† kˆ 1 ‡ κK H h (4.18) (4.18)

where KAS, KAM, KAD, and KMD are equilibrium constants associated with the basic species; KHAS, KHAM, KHAD, and KHMD, correspond to the acidic species; and κ is given by κˆ

1 ‡ K HAM ‰MŠ…1 ‡ K HMD φ†=…1 ‡ K HAD φ† 1 ‡ K AM ‰MŠ…1 ‡ K MD φ†=…1 ‡ K AD φ†

(4.19)

Polynomial models can also be suitable to describe the retention, as the following: 1 ˆ a0 ‡ a1 ‰MŠ ‡ a2 φ ‡ a3 pH ‡ a12 ‰MŠφ ‡ a13 ‰MŠpH ‡ a23 φpH k ‡ a123 ‰MŠφ pH ‡ a33 pH2

(4.20)

Rapid convergence of the retention models in the optimal solution can be achieved by previous partial fitting of the data, measured at extreme pH values to obtain the retention factors for the acidic and basic species. The protonation

417

418

4 Micellar Liquid Chromatography: Method Development and Applications

constant can be estimated using data measured at intermediate pH values. Experimental data from five mobile phases at three different pH values (e.g., 3, 5, and 7) are needed, but extra data are often used to improve the reliability of the predictions. Relative errors below 6% are usually obtained.

4.4 Analytical Procedures in MLC

Hundreds of reports have been published describing the successful use of MLC in the determination of a variety of compounds in several types of samples. However, applications mainly focus on the determination of drugs in physiological fluids and pharmaceutical formulations. The direct injection capability of MLC after simple filtration (with no noticeable damage after repetitive injections) is especially appreciated. Table 4.1 lists MLC procedures for the determination of drugs in physiological fluids, oriented to follow the parent drug and/or its metabolites. Table 4.2 lists procedures applied to formulations and Table 4.3 reports MLC procedures for compounds of industrial, environmental, veterinarian, and nutritional interest, among others. The type of column, nature, and concentration of surfactant and organic solvent, and pH, for the different procedures, are given. Most MLC analyses are performed in the isocratic mode, with sufficiently short analysis times and good selectivity. The vast majority of reports employ hybrid micellar mobile phases, where the amount of organic solvent is limited to preserve the formation of micelles. Next, some details are given about the analysis of drugs in physiological fluids and formulations. 4.4.1 Analysis of Physiological Samples 4.4.1.1

Need of Previous Separation Steps in Conventional RPLC

The control of drugs and their metabolites in physiological fluids is of great interest in clinical chemistry, doping, toxicology, and pharmaceutical research. The therapeutic efficacy of many drugs is often related to their concentration in physiological fluids, which depends on their dosage, route, and frequency of administration. However, analytical assays to control drug administration are problematic, since the drugs are frequently at very low concentration in the physiological fluids, strongly bound to proteins, and suffer from the interference of numerous endogenous (and exogeneous) compounds in a complex matrix [225–226]. Some drugs are rapidly metabolized after oral administration, and therefore, proper monitoring of the target drug should be performed by detection of one or several of its metabolites, since the excreted amount of the parent drug is very low. In addition, compounds excreted as conjugates in urine require hydrolysis.

4.4 Analytical Procedures in MLC

Table 4.1 Experimental characteristics of MLC procedures for the analysis of physiological fluids. Compounds

Sample (number of analytes)/mobile phase/stationary phase/elution mode/ temperature

Ref.

Alkaloids: Cotinine, nicotine, nicotinic acid, and nicotinamide

Urine (2)/0.2 M SDS, 3% v/v 2-propanol, pH 4.6/Econosphere CN-bonded silica/ isocratic/40 °C

[31]

Serum and urine (2)/0.15 M SDS, 6% 1-pentanol, pH 3/Kromasil C18/isocratic

[32]

Serum (1)/0.15 M SDS, 6% 1-pentanol, 0.001 M KCl, pH 6/Kromasil C18/isocratic

[33]

Amino acids: Proline

Urine (1)/0.03 M SDS, 8% 1-propanol, pH 5.3/C18/isocratic

[34]

Analgesics, antibiotics, anticonvulsants, and antiarrhythmics: Acetaminophen, acetylsalicylic acid, chloramphenicol, carbamazepine, phenobarbitone, phenytoin, and procainamide

Serum (7)/0.02–0.10 M SDS, pH 3/Supelcosil LC-18, Supelcosil LC-CN/isocratic

[7]

Serum (3)/column switching/0.02–0.04 M SDS, 4–14% acetonitrile (extraction)/ 0.02–0.04 M SDS, 28–70% methanol, pH 3–4.6/Spherisorb-5 RP-18 (separation)

[35]

Analgesics: Acetaminophen

Serum and urine (1)/0.02 M SDS, pH 7/ Kromasil C18/isocratic

[36]

Anesthetics: Procaine and tetracaine

Plasma (2)/0.15 M SDS, 10% 1-propanol, 0.5% triethylamine, pH 2.5/Spherisorb ODS-2/isocratic

[37]

Antiarrithmics: Acebutolol, N-acetylprocainamide, atenolol, celiprolol, disopyramide, labetalol, lidocaine, metoprolol, nadolol, procainamide, propranolol, and quinidine

Urine (7)/0.1 M SDS, 15% 1-propanol, pH 3/Spherisorb ODS-2/isocratic

[12]

Plasma (2)/0.05 M SDS, 1% 1-butanol, 0.9% NaCl, pH 3/Kromasil C18/isocratic/ 25 °C

[38]

Serum (3)/0.15 M SDS, 7% 1-butanol, 0.9% NaCl, pH 7/Kromasil C18/isocratic/ 25 °C

[39]

Antiarrythmics and diuretics: Acebutolol, atenolol, labetalol, metoprolol, nadolol, propranolol, and amiloride, bendroflumethiazide, piretanide, and triamterene

Urine (10)/0.11 M SDS, 15% 1-propanol, pH 3/Spherisorb ODS-2/isocratic

[40]

Antiarrythmics and metabolites: Desisopropylpropranolol, α-naphtoxylactic acid, α-naphtoxyacetic, propranolol, and propranolol glycol

Urine (5)/0.15 M SDS, 10% 1-propanol, pH 3/Spherisorb ODS-2/isocratic

[41]

(continued)

419

420

4 Micellar Liquid Chromatography: Method Development and Applications

Table 4.1 (Continued) Compounds

Sample (number of analytes)/mobile phase/stationary phase/elution mode/ temperature

Ref.

Antibacterial agents: Amoxicillin, ampicillin, ciprofloxacin, cloxacillin, danofloxacin, dicloxacillin, difloxacin, enoxacin, flumequine, levofloxacin, lomefloxacin, marbofloxacin, moxifloxacin, norfloxacin, ofloxacin, sulfacetamide, sulfadiazine, sulfaguanidine, sulfamerazine, sulfathiazole, sulfamethazine, sulfamethizole, sulfamethoxypyridazine, sulfachlorpyridazine, sulfamonomethoxine, sulfabenzamide, sulfamethoxazole, sulfadimethoxine, sulfaquinoxaline, sulfathiazole, and sulfisomidinetinidazole

Human urine and cow milk (12)/0.07 M SDS, 6% 1-propanol/YMC-Pack ODS/isocratic/40 °C

[42]

Urine (5)/0.05 M SDS, 2.4% 1-pentanol/ Spherisorb ODS-2/isocratic

[43]

Milk (11)/0.019 M SDS, 5.8% acetonitrile, pH 3/Hypersil ODS-2/isocratic

[44]

Serum (2)/0.075 M SDS, 3% 1-propanol, pH 3/Licrospher 100RP18/isocratic

[45]

Milk (2)/0.075 M SDS, 3% 1-propanol, pH 3/isocratic

[46]

Milk (6)/0.08 M SDS, 8.5% 1-propanol, pH 7/Kromasil C18/isocratic/25 °C

[47]

Milk (14)/0.04 M SDS, 2% 2-propanol, pH 2.8–3.5/Lichrospher C18/isocratic/ 40 °C

[48]

Urine (1)/0.10 M SDS, 4% 1-butanol, pH 3/Hypersil phenyl/isocratic/25 °C

[49]

Urine (5)/0.15 M SDS, 12.5% 1-propanol, 0.5% triethylamine, pH 3.0 or 0.05 M SDS, 12.5% propanol, 0.5% triethylamine, pH 3.0/Kromasil C18/isocratic

[50]

Urine and serum (2)/0.01 M SDS, pH 6/ C18/isocratic

[51]

Plasma (2)/0.15 M SDS, 5% 1-propanol, 0.3% triethylamine, pH 4/Simmetry C18/ isocratic

[52]

Urine (4)/0.11 M SDS, 6% 1-propanol, pH 3/Zorbax C18/isocratic/25 °C

[53]

Milk (4)/0.05 M SDS, 10% 1-butanol, 0.5% triethylamine, pH 3/Kromasil C18/ isocratic/25 °C

[54]

Serum (5)/0.08 M SDS, pH 3/LiChrosorb RP-18/isocratic/30 °C

[55]

Serum and urine (1)/0.1 M SDS, pH 5.7/ LiChrospher 100 RP-18/isocratic

[56]

Plasma and urine (1)/column switching/ 0.02 M SDS, 10% methanol, pH 6 (extraction)/30% methanol, pH 6 (separation)/ RP-18

[57]

Plasma (1)/column switching/0.02 M SDS, 10% methanol, pH 6.8 (extraction)/ 30% methanol, pH 6.8 (separation)/ RP-18

[58]

Anticancer drugs: Endoxifen, hexamethylene bisacetamide, 6-mercaptopurine, 6-thioguanine and their metabolites, 6-mercaptopurine riboside, methotrexate, mitomycin-c, tamoxifen, 6-thioguanine riboside, and 6-thioxanthine

4.4 Analytical Procedures in MLC

Plasma (1)/0.15 M SDS, 7% 1-butanol, pH 3/Kromasil C18/isocratic/40 °C

[59]

Plasma (2)/0.15 M SDS, 7% 1-butanol, pH 3/Kromasil C18/isocratic/40 °C

[60]

Serum (1)/0.02 M SDS/Supelcosil LC-CN/isocratic

[61]

Serum (1)/column switching/0.01 M SDS (extraction)/65% methanol (separation)/ Adsorbosphere ODS

[62]

Serum (8)/0.06 M SDS, 5% 1-butanol, pH 7/Kromasil C18/isocratic/25 °C

[63]

Serum (3)/0.05 M SDS, 7% 1-butanol, pH 7/Kromasil C18/isocratic

[64]

Urine (2)/0.15 M SDS, 10% 1-propanol, 0.3% triethylamine, pH 3.5/Hypersil phenyl column/isocratic

[65]

Serum (3)/0.13 M SDS, 2.4% 1-pentanol, 0.1% triethylamine, pH 7/Kromasil C8/ isocratic/25 °C

[66]

Serum (2)/0.15 M SDS, 6% 1-pentanol, pH 7/Kromasil 5 C18 /isocratic

[67]

Urine (1)/0.2 M SDS, 1% 1-butanol, pH 3/ Kromasil C18/isocratic/25 °C

[68]

Antihistamines: Cetirizine

Plasma (1)/0.135 M SDS, 11% 1-propanol, 0.3% triethylamine, pH 3.3/EC Nucleosil C18-SN /isocratic

[69]

Anti-inflamatory drugs: Diclofenac, flunixin, 5-lipoxygenase inhibitor zileuton, and phenylbutazone

Urine (2)/0.025 M SDS, 9% 1-butanol, pH 3/CN-bonded silica/isocratic

[70]

Equine serum (3)/C18/isocratic

[71]

Antipyrine metabolites: 4-Aminoantipyrine, 4-methylaminoantipyrine, and 4-formylaminoantipyrine

Plasma (3)/0.1 M SDS, 2.5% 1-pentanol/ Nucleosil C18/isocratic

[72]

Plasma (3)/0.1 M SDS, 2.5% 1-pentanol/ Spherisorb RP-18/isocratic

[73]

Simulated physiological fluids (4)/0.05 M SDS, 1% 1-butanol, pH 3.0/Synergi C18 /isocratic/30 °C

[74]

Serum (5)/0.05 M SDS, 2.5% 1-propanol or 6% 1-pentanol, pH 7/Kromasil C18/ isocratic

[75]

Plasma and urine (1)/0.02 M CTAB, 15% 1-propanol, pH 7.5/Spherisorb ODS-2/ isocratic

[76]

Urine (5)/0.07 M SDS, 0.3% 1-propanol, pH 7.4/Spherisorb ODS-2/isocratic

[77]

Plasma (1)/0.03 M CTAB, 3% 1-propanol, pH 7/Spherisorb ODS-2/isocratic

[78]

Anticonvulsant agents: Bromazepam, carbamazepine, diazepam, flunitrazepam, halazepam, medazepam, nitrazepam, oxazepam, phenobarbital, phenytoin, tetrazepam, and zopiclone

Antidepressants: Desipramine, imipramine, and trazodone

Antivirals: Efavirenz, zidovudine, lamivudine, stavudine, tenofovir, zidovudineleucine, and zidovidine-valine

Barbiturates: Amobarbital, barbital, butabarbital, diallybarbituric acid, hexobarbital, pentobarbital, phenobarbital, and secobarbital

(continued)

421

422

4 Micellar Liquid Chromatography: Method Development and Applications

Table 4.1 (Continued) Compounds

Sample (number of analytes)/mobile phase/stationary phase/elution mode/ temperature

Ref.

Plasma (2)/0.04 M CTAB, 3% 1-propanol, pH 7.5/Spherisorb ODS-2 /isocratic

[79]

Serum (4)/0.10 M SDS, 4% 1-butanol, pH 7/Spherisorb ODS-2/isocratic/25 °C

[80]

Biogenic amines and metabolites: Dopamine, homovanilic acid, hydroxyindoleacetic acid, serotonin, and tyramine

Serum (5)/0.15 M SDS, pH 3/Kromasil C18/isocratic

[81]

Bronchodilators: Caffeine, theobromine, and theophylline

Serum (1)/0.02–0.10 M SDS/μ-Bondapak C18/isocratic

[7]

Serum (1)/0.05 M SDS/Supelcosil LC-18/ isocratic

[61]

Serum (1)/0.001 M 3-(dimethyldodecylammonium) propanesulfonate, 3% 1-propanol/μ-Bondapak/isocratic

[82]

Urine (3)/0.075 M SDS, 1.5% 1-propanol/ Spherisorb ODS-2/isocratic

[83]

Serum (2)/0.05 M SDS, 2.5% 1-propanol, pH 7/Kromasil C18/isocratic

[84]

Urine and serum (1)/0.125 M SDS, 3% 1-pentanol, pH 3/Kromasil C18/isocratic

[85]

Urine and serum (1)/0.15 M SDS, 5% 1-pentanol, pH 7/Kromasil C18/isocratic

[86]

Cathecolamines and metabolites: Epinephrine, normetanephrine, norepinephrine, and metanephrine

Serum (4)/0.075 M SDS, 1.6% 1-butanol, pH 7/C18/isocratic/25 °C

[87]

Cephalosporins: Cefmenoxime hemihydrochloride, cefotaxime, cefotiam dihydrochloride, cefradine, and cephalexin

Serum (2)/0.08 M SDS, 8% 2-propanol, pH 3/Nucleosil C18/isocratic

[88]

Serum (4)/0.02 M SDS, pH 6.1 or 0.15 M SDS, pH 3.1/Develosil ODS/isocratic/ 40 °C

[89]

Desferroxiamine and chelates with Al and Fe

Serum (3)/0.2 M SDS, 5% acetonitrile or 0.5% Brij-35, pH 7.4/Spherisorb ODS-2/ isocratic

[90]

Diuretics: Althiazide, amiloride, bendroflumethiazide, benzthiazide, bumetanide, canrenoic acid, canrenone, chlorthalidone, chlorthiazide, clopamide, cyclothiazide, dichlorphenamide, ethacrynic acid, furosemide, hydrochlorthiazide, hydroflumethiazide, indapamide, piretanide, polythiazide, probenecid, spironolactone,

Plasma (1)/0.05 M SDS, 5% 1-propanol, pH 5.8/C8/isocratic

[91]

Serum and urine (1)/0.10 M SDS, 3% 1-propanol, pH 3.5/Nucleosil RP-18/ isocratic

[92]

Urine (1)/0.02 M Brij-35, 0.004 M SDS, pH 6.5/Hypersil C18/isocratic

[93]

Urine (2)/0.05 M SDS, 5% methanol/ Spherisorb ODS-2/isocratic/50 °C

[94]

Calcium channel blockers: Nifedipine and verapamil

4.4 Analytical Procedures in MLC

torasemide, triamterene, trichlormethiazide, and xipamide

Urine (10)/0.042 M SDS, 4% 1-propanol, pH 4.5/Spherisorb ODS-2/isocratic

[95]

Urine (8)/0.055 M SDS, 10% 1-propanol/ Spherisorb ODS-2/isocratic

[96]

Urine (7)/0.055 M SDS, 8% 1-propanol/ Spherisorb ODS-2/isocratic

[97]

Urine (15)/0.055 M SDS, 6% 1-propanol, pH 3/Spherisorb ODS-2/isocratic

[98]

Urine (19)/0.040 M SDS, 4% tetrahydrofuran, pH 3.2/Hypersil C18/isocratic/50 °C

[99]

Urine (1)/0.05 M SDS, 6% 1-propanol, pH 3/Spherisorb ODS-2/isocratic

[100]

Urine (1)/0.05 M SDS/C18/isocratic

[101]

Diuretics, steroids, and stimulants: Amiphenazole, amiloride, amphetamine, clostebol, ephedrine, phenylpropanolamine, methandienone, methoxyphenamine, nandrolone, and spironolactone

Urine (10)/0.1 M SDS, 3% 1-pentanol/ Spherisorb-5 RP-18/isocratic

[102]

Insecticides: Carbaryl and 1-naphthol

Urine and serum (2)/0.15 M SDS, 6% 1-pentanol, pH 3/C18/isocratic

[103]

Melamine

Milk (1)/0.05 M SDS, 7.5% 1-propanol, pH 3/Kromasil C18/isocratic

[104]

Plasma and urine (1)/0.2 M SDS, pH 3/ Kromasil C18/isocratic

[105]

Nucleoside/nucleotide reverse transcriptase inhibitors: Lamivudine and metabolites

Gastric and intestinal fluids (8)/0.15 M SDS, 4% 1-butanol, pH 7/Kromasil C18/ isocratic/30 °C

[106]

Omeprazole and metabolites: Hydroxyomeprazole, omeprazole, and omeprazole sulphone

Serum and urine (3)/0.08 M SDS, 10% 1-propanol, pH 7/Kromasil C18/isocratic

[107]

Opiates: Benzoylecgonine, codeine, heroin, 6-monoacetylmorphine, morphine, and thebaine

Serum (3)/0.15 M SDS, 7% 1-butanol, pH 7/Kromasil C18/isocratic/25 °C

[108]

Serum (4)/0.1 M SDS, 4% 1-butanol, pH 7/Kromasil C18/isocratic

[109]

Phenols: 2-Aminophenol, 4-nitrophenol, and phenol

Urine (3)/column switching/0.03 M CTAB, 7% acetonitrile, pH 5 Nucleosil C4 (extraction)/0.03 M CTAB, 20% acetonitrile, pH 5 Kromasil or Nucleosil C18 (separation)

[110]

Radioligands: [11 C]AZ10419369, [11 C] AZD2184, [11 C]clozapine, [11 C]deprenyl, [11 C]diazepam, [11 C]doxepin, [18 F]FEDTBZ, [18 F]FE-PE2I, [11 C]flumazenil, [18 F]LBT-999, [18F]MCL-524, [11 C]MePPEP, [11 C]MNPA, [11 C]PBR28, [11 C]Ro15-4513, [11 C]rolipram, [11 C]venlafaxine, and [11 C]verapamil

Plasma (9)/0.1 M SDS, 1–2% to 10–25% 1-butanol, pH 7.2/Atlantis Prep OBD T3 C18/gradient

[111]

Plasma (5)/column switching/ 0.02–0.025 M SDS, 10% acetonitrile (cleanup)/0.02–0.025 M SDS, 10–50 or 70% acetonitrile (separation: MLC and HSLC)/XBridge OST C18/gradient

[112]

(continued)

423

424

4 Micellar Liquid Chromatography: Method Development and Applications

Table 4.1 (Continued) Compounds

Sample (number of analytes)/mobile phase/stationary phase/elution mode/ temperature

Ref.

Plasma (4)/1–2% v/v Triton X-100, 0.1 M SDS, 0–5% 1-butanol, pH 7/monolithic C18/isocratic

[113]

Plasma (6)/0.05 M SDS, 5–68% acetonitrile, pH 7/monolithic C18/gradient

[114]

Urine (1)/column switching/0.02 M SDS, 18% methanol, 25% 1-propanol, pH 6 (extraction)/0.02 M SDS, 38% methanol, 2% 1-propanol, pH 6/RP-18 (separation)

[115]

Urine (6)/0.05 M SDS, 9% 1-butanol/ Spherisorb-5 RP18/isocratic

[116]

Urine (16)/0.036 M SDS, 1.9% 1-butanol/ Hypersil C18/isocratic/50 °C

[117]

Urine (13)/0.12 M SDS, 7% 1-pentanol/ Spherisorb ODS-2/isocratic

[118]

Urine (2)/0.018 M SDS, 8.3% tetrahydrofuran/Hypersil C18/isocratic

[119]

Stimulants: Amphetamine, ephedrine, methoxyphenamine, phenylephrine, and phenylpropanolamine

Urine (5)/0.15 M SDS, 3% 1-pentanol, pH 3/Spherisorb ODS-2/isocratic

[120]

Vitamins: B6 (pyridoxal, pyridoxine, and pyridoxamine)

Serum (3)/0.15 M SDS, 2% 1-pentanol, pH 3/Kromasil C18/isocratic

[121]

Steroids: Acetonide, betamethasone, corticosterone, cortisol, cortisone, deflazacort, dehydrotestosterone, dexamethasone, dydrogesterone, fludrocortisone, fludrocortisone acetate, hydroxycorticosterone, 21-hydroxydeflazacort, 11α-hydroxyprogesterone, medroxyprogesterone, medroxyprogesterone acetate, methandienone, methelonone enanthate, methylprednisolone, methyltestosterone, nandrolone, nandrolone decanoate, norhisterone, prednisolone, prednisone, progesterone, testosterone, testosterone enanthate, testosterone propionate, triamcindone, and triamcinolone

Particularly troublesome is the presence of high molecular mass proteins in the samples when injected directly into a conventional RPLC system, since they tend to denature and precipitate in the injection valve or at the column head. This produces clogging of the system and irreversible adsorption on the stationary phase, giving rise to a rapid degradation of chromatographic performance and an increase in back pressure. In a conventional RPLC system using hydro-organic mixtures, the harmful proteinaceous material must be usually removed from the sample prior to injection, to prevent irreversible adsorption on the stationary phase. These procedures are tedious, repetitive, time-consuming, and expensive, and introduce additional sources of error because of incomplete recoveries. Proteins can either be insolubilized by organics (acetone, acetonitrile, or trichloroacetic acid) and sodium hydroxide or be removed by ultrafiltration. Isolation of the analyte from the matrix is usually still needed by liquid–liquid or solid–phase extraction (and

4.4 Analytical Procedures in MLC

Table 4.2 Experimental characteristics of MLC procedures for the analysis of pharmaceutical formulations. Compounds

Sample (number of analytes)/mobile phase/stationary phase/elution mode/ temperature

Ref.

Alkaloids: Nicotinic acid and nicotinamide

Diverse formulations (2)/0.15 M SDS, 6% 1-pentanol, pH 3/Kromasil C18/isocratic

[32]

Aminoacids: Glycine, lysine, threonine, and methionine

Pills, capsules, drops, and powders (4)/ 0.05 M SDS, 3% 1-propanol, pH 3/Spherisorb ODS-2/isocratic

[122]

Analgesics, antihistamines and phenethylamines: Acetaminophen, acetylsalycilic acid, carbinoxamine, chlorpheniramine, dexbrompheniramine, dexchlorpheniramine, diphenhydramine, doxylamine, pheniramine, phenyltoxolamine, tripolidine, azatadine, ephedrine, methoxyphenamine, phenylephrine, phenylpropanolamine, and pseudoephedrine

Tablets (3)/1.8% Brij-35, 0.012 M SDS, 4.5% 1-propanol/Zorbax CN/isocratic/65 °C

[123]

Tablets (1)/0.02 M SDS, 15% 1-propanol, pH 3/C18/isocratic

[124]

Diverse formulations (15)/0.05 M SDS, 6% pentanol, pH 7/Eclipse XDB C8/ isocratic

[125]

Cough and cold preparations (4)/0.15 M SDS, 6% pentanol, pH 7/Spherisorb ODS-2 C18/isocratic

[126]

Anesthetics: Bupivacaine, lidocaine, mepivacaine, procaine, propanocaine, and tetracaine

Diverse formulations (6)/0.15 M SDS, 1-propanol (9 : 1), pH 3/Spherisorb ODS-2/isocratic

[127]

Anesthetics and muscle relaxants: Lidocaine and tolperisone

Diverse formulations (2)/0.075 M SDS, 7.5% 1-propanol/Zorbax C18/isocratic

[128]

Antianginals: Diltiazem, nadolol, nifedipine, propranolol, and verapamil

Diverse formulations (5)/0.05 M SDS, 5% pentanol, pH 7/Kromasil C18/isocratic

[129]

Antiarrithmics: Atenolol, carteolol, celiprolol, labetalol, metoprolol, nadolol, oxprenolol, propranolol, and timolol

Tablets, capsules, and ophthalmic solutions (9)/0.15 M SDS, 15% 1-propanol, pH 3/Spherisorb ODS-2/isocratic

[130]

Antiarrithmics, diuretics, and/or vasodilators: Atenolol, metoprolol, oxprenolol, amiloride, bendroflumethiazide, chlortalidone, hydrochlorthiazide, hydralazine, indapamide, and pindolol

Tablets and capsules (8)/0.15 M SDS, 7% 1-propanol, pH 3/Spherisorb ODS-2/ isocratic

[131]

Tablets (2)/0.15 M SDS, 18% methanol, 0.3% triethylamine, pH 3/Symmetry C18/ isocratic

[132]

Diverse formulations (2)/0.07 M SDS, 15% 1-propanol, pH 3/C18/isocratic

[133]

Tablets, pills, capsules, suspensions, and drops (8)/0.05 M SDS, 2.4% 1-pentanol/ Spherisorb ODS-2/isocratic

[134]

Diverse formulations (6)/0.04 M SDS, 2% 2-propanol/Nucleosil C18/isocratic/40 °C

[135]

Human and veterinary formulations (7)/ SDS, 6% acetonitrile, pH 3/Hypersil ODS/isocratic

[136]

Tablets and capsules (1)/0.1 M SDS, 15% 1-butanol, pH 7/Hypersil C18/isocratic/ 60 °C

[137]

Antibacterial agents: Amoxicillin, ampicillin, azithromycin, cefuroxime, cloxacillin, dicloxacillin, levofloxacin, moxifloxacin, norfloxacin, pipemidic acid, sulfacetamide, sulfadiazine, sulfadimethoxine, sulfaguanidine, sulfamerazine, sulfamethazine, sulfamethizole, sulfamethoxazole, sulfanilamide, sulfathiazole, and trimethoprim

(continued)

425

426

4 Micellar Liquid Chromatography: Method Development and Applications

Table 4.2 (Continued) Compounds

Anticonvulsant agents: Bentazepam, carbamazepine, clorazepate, chlordiazepoxide, diazepam, diltiazem, ethosumixine, halazepam, oxazepam, phenobarbital, phenytoin, pinazepam, tetrazepam, and zoplicone

Antidepressants: Amitriptyline, clomipramine, doxepin, imipramine, maprotiline, nortriptyline, trazodone, and trimipramine

Antidiabetic drugs: Glicazide, glipizide, metformin, and nateglinide

Antihistamines: Azatadine, carbinoxamine, cetirizine, cyclizine, cyproheptadine, diphenhydramine, doxylamine, tripelennamine, brompheniramine,

Sample (number of analytes)/mobile phase/stationary phase/elution mode/ temperature

Ref.

Tablets (1)/0.02 M SDS, 8% acetonitrile/ XTerra C18/isocratic/50 °C

[138]

Diverse formulations (2)/0.1 M SDS, 3% 1-butanol/Hypersil ODS/isocratic/35 °C

[139]

Diverse formulations (4)/0.025 M SDS, 2% 1-butanol, pH 5/Ultra C18/isocratic

[140]

Diverse formulations (4)/0.15 M SDS, 2.5% 1-propanol, 0.5% triethylamine, pH 3/C18/isocratic

[141]

Tablets and capsules (4)/0.11 M SDS, 6% 1-propanol, pH 3/Zorbax C18/isocratic/ 25 °C

[53]

Diverse formulations (3)/0.1 M SDS, 3% 1-butanol, pH 3/Spherisorb ODS-2/ isocratic

[142]

Pills and capsules (6)/0.1 M SDS, 3% 1-butanol, 0.1% triethylamine, pH 3/Spherisorb ODS-2/isocratic

[143]

Capsules, pills, tablets, injections, drops, and suppositories (7)/CTAB/C18/ isocratic

[144]

Capsules, pills, tablets, and injections (7)/ 0.04 M CTAB, 5–15% 1-propanol, pH 3–4/Kromasil C18/isocratic

[145]

Tablets and capsules (7)/0.075 M SDS, 6% 1-pentanol, pH 3/Eclipse XDB C18/ isocratic

[146]

Tablets and injectables (1)/0.2 M SDS, 1% 1-butanol, pH 3/Kromasil C18/isocratic/ 25 °C

[68]

Tablets and capsules (7)/0.02 M Brij-35, pH 3/Zorbax C18/isocratic/25 °C

[3]

Diverse formulations (3)/Zorbax XDB C18/isocratic

[147]

Diverse formulations (3)/0.12 M SDS, 10% 1-propanol, 0.3% triethylamine, pH 5.6/Nucleosil C18/isocratic

[148]

Tablets, capsules, powders, solutions, and syrups (7)/0.15 M SDS, 6% 1-pentanol/ Spherisorb ODS-2/isocratic

[149]

4.4 Analytical Procedures in MLC

chlorcyclizine, chlorphenamine, flunarizine, hydroxyzine, promethazine, terfenadine, tripelennamine, and tripolidine

Tablets, capsules, suppositories, syrups, and ointments (11)/0.02 M CTAB, 3% 1-propanol, pH 6; 0.02 M CTAB, 3% 1-propanol, pH 7; 0.04 M CTAB, 3% 1-butanol, pH 3; 0.04 M CTAB, 3% 1-butanol, pH 5/Spherisorb C18/isocratic

[150]

Capsules (1)/0.135 M SDS, 11% 1-propanol, 0.3% triethylamine, pH 3.3/EC Nucleosil C18-SN/isocratic

[69]

Antihistamines and local anesthetics: Bupivacaine, chlorpheniramine, diphenhydramine, lidocaine, pheniramine, phenylpropanolamine, prilocaine, procaine, pseudoephedrine, and tripolidine

Formulations (10)/0.05 M SDS, 10% 2-propanol/Zorbax SB C18/isocratic/30 °C

[151]

Antiplatelet agents: Clopidogrel

Tablets (1)/0.15 M SDS, 10% 1-propanol, 0.3% triethylamine, pH 3/Nucleodur MNC18/isocratic

[152]

Antivirals: Nelfinavir

Tablets (1)/0.5 M Tween 20, 2% 1-butanol, pH 4.2/LiChrosphere C18/isocratic/25 °C

[153]

Benzyl alcohol and benzaldehyde

Injectable formulations (2)/0.07 M SDS, 10% 1-propanol, 0.3% triethylamine, pH 7.5/ Apex ODS-2/isocratic

[154]

Bronchodilators: Caffeine and theophylline

Tablets, capsules, and pills (1)/0.05 M SDS, 1.5% 1-propanol/Spherisorb ODS-2/isocratic

[155]

Tablets, capsules, and syrups (1)/0.05 M SDS, 3% 1-propanol/Spherisorb ODS-2/isocratic

[156]

Calcium channel blockers: Flunarizine, nifedipine, verapamil, and their degradation products

Diverse formulations (6)/0.15 M SDS, 10% 1-propanol, 0.3% triethylamine, pH 4 or 6.8/cyanopropyl-bonded/isocratic

[157]

Catecholamines: L-Dopa, 2-methyldopa, epinephrine, dopamine, and isoprotenerol

Tablets, capsules, injections, suspensions, and gels (5)/0.1 M SDS, 5% 1-propanol, pH 3/Spherisorb ODS-2/isocratic

[158]

Diuretics: Acetazolamide, althiazide, amiloride, bendroflumethiazide, bumetanide, chlortalidone, chlorthiazide, cyclothiazide, furosemide, hydrochlorthiazide, hydroflumethiazide, spironolactone, triamterene, trichlormethiazide, and xipamide

Tablets and suspensions (9)/0.05 M SDS, 3% 1-propanol/Spherisorb ODS-2/ isocratic

[159]

Tablets (5)/0.07 M SDS, 0.5% 1-pentanol/ Spherisorb ODS-2/isocratic

[160]

Tablets and capsules (7)/0.02 M SDS or 0.15 M SDS, pH 7/Spherisorb ODS-2/ isocratic

[161]

Tablets, capsules, injectables, and drops (1)/0.06 M SDS, 8% 1-propanol, pH 3/ Spherisorb ODS-2/isocratic

[162]

Diverse formulations (1)/0.05 M SDS/ C18/isocratic

[102]

Tablets (1)/3 M Tween-20, 8% 1-butanol, pH 5.9/Luna C18/isocratic/25 °C

[163]

Histamine H2 receptor antagonist: Ranitidine

[164] (continued)

427

428

4 Micellar Liquid Chromatography: Method Development and Applications

Table 4.2 (Continued) Compounds

Sample (number of analytes)/mobile phase/stationary phase/elution mode/ temperature

Opiates: Codeine, morphine, noscapine, and papaverine

Injection solution (4)/0.10 M SDS, 5% 1-butanol, pH 2.5/Kromasil C18/isocratic/ 40 °C

Parabens and active ingredients: Benzoic acid, methyl paraben, propyl paraben, paracetamol, caffeine, and guaifenesin

Syrups (6)/0.04 M SDS, 0.1% trichloroacetic acid/Kromasil C18/isocratic/ 40 °C

[165]

Phenethylamines: Amphetamine, arterenol, ephedrine, phenylephrine, phenylpropanolamine, mephentermine, methoxyphenamine, pseudoephedrine, and tyramine

Capsules, tablets, pills, powder, syrup, and drops (9)/0.15 M SDS, 5% 1-pentanol, pH 7/Spherisorb ODS-2/isocratic

[166]

Risedronate

Tablets (1)/0.02 M SDS, 10% 1-propanol, 0.3% triethylamine, pH 6/Simmetry C18/ isocratic

[167]

Nonsteroideal anti-inflammatory drugs: Acemetacin, diclofenac, indomethacin, ketoprofen, nabumetone, naproxen, tolmetin, and piketoprofen

Diverse formulations (8)/0.06 M CTAB, 10% 1-butanol, pH 7/Kromasil C18/ isocratic

[168]

Diverse formulations (6)/0.15 M SDS, 10% 1-propanol, pH 3/Spherisorb ODS-2/isocratic

[169]

Selenium (IV)

Tablets and syrups (1)/0.05 M SDS, 10% 1-butanol/Hypersil ODS/isocratic

[170]

Steroids (anabolics and corticoids): Beclomethasone, bethamethasone, budesonide, danazol, dexamethasone, dydrogesterone, fludrocortisone, fluocinolone, hydrocortisone, medroxyprogesterone, methyltestosterone, nandrolone, nandrolone decanoate, progesterone, stanozolol, testosterone enanthate, testosterone proprionate, and triamcinolone

Tablets, injections, suspensions, and gels (9)/0.1 M SDS, 7% 1-pentanol/Spherisorb ODS-2/isocratic

[171]

Creams, gels, and ointments (7)/0.1 M SDS, 4% 1-butanol, pH 7/Kromasil C18/ isocratic

[172]

Pills (1)/0.04 M SDS, 10% 1-propanol/ Hypersil ODS/isocratic/60 °C

[173]

Capsules (1)/0.04 M SDS, 2% 1-pentanol/ Hypersil ODS/isocratic/60 °C

[174]

Tablets and cocktails (2)/0.032 M CTAB, 0.24% 1-pentanol/Hypersil C18/isocratic/ 60 °C

[175]

Capsules, pills, and syrups (5)/0.1 M SDS, 4% 1-pentanol, pH 3/Kromasil C18/ isocratic

[176]

Multivitamin tablets (7)/0.016 M SDS, 3.5–10% 1-butanol, pH 3.6/Particil ODS-2 (250 mm × 4.6 mm i.d.)/gradient/ 35 °C

[177]

Multivitamin syrup (2)/0.077 M SDS, 12% 1-butanol, pH 7/Spherisorb ODS-2/ isocratic/30 °C

[178]

Vitamins: A, B3 (nicotinamide), B1 (thiamine), B2 (rivoflavin), B6 (pyridoxal, pyridoxine, and pyridoxamine), B9 (folic acid), B12 (cyanocobalamin), C (ascorbic acid), and E (tocopherols)

Ref.

4.4 Analytical Procedures in MLC

Table 4.3 Experimental characteristics of MLC procedures for the analysis of other samples. Compounds

Sample (number of analytes)/mobile phase/stationary phase/elution mode/temperature

Ref.

Alkaloids: Berberine, coptisine, jatrorrhizine, matrine, nicotine, oxymatrine, sophocarpine, oxysophocarpine, sophoridine, oxysophoridine, and palmatine

Coptis rizhome, phellodendron bark, and Chinese parent medicines (4)/ 0.12 M SDS/Bondapak phenyl/isocratic

[179]

Sophora medicinal plants (6)/0.02 M Brij-35/0.01 M SDS, 5% acetonitrile, pH 6/Diamonsil C18/isocratic/40 °C

[180]

Tobacco and chewing gums (1)/0.15 M SDS, 6% 1-pentanol, 0.001 M KCl, pH 6/Diamonsil C18 (250 mm × 4.6 mm i.d.)/isocratic

[33]

Amino acids and biogenic amines: Tryptophan, tyrosine, tryptamine, and tyramine

Wines (4)/0.15 M SDS, 5% 1-propanol, pH 3/Kromasil C18/isocratic

[181]

Antibacterial agents: Blasticidin S, carbadox, danofloxacin, difloxacin, enrofloxacin, flumequine, kasugamycin, marbofloxacin, olaquindox, oxolinic acid, sarafloxacin, sulfacetamide, sulfachloropyridazine, sulfadiazine, sulfadimethoxine, sulfafurazole, sulfaguanidine, sulfamerazine, sulfamethazine, sulfamonomethoxine, sulfanilamide, sulfaproxyline, sulfapyridine, sulfaquinoxaline, sulfathiazole, sulfathiocarbamide, sulfisomidine, and sulfisoxazole

Honey (11)/0.019 M SDS, 5.8% acetonitrile, pH 3/Hypersil ODS-2/isocratic

[44]

Honey, eggs, and meat (14)/0.04 M SDS, 2% 2-propanol, pH 2.8–3.5/ Lichrospher C18/isocratic/40 °C

[48]

Fish muscles (5)/0.065 M SDS, 12.5% 1-propanol, 0.5% triethylamine, pH 3/ Kromasil C18/isocratic

[182]

Eggs (4)/0.05 M SDS, 10% 1-butanol, 0.5% triethylamine, pH 3/Kromasil C18/ isocratic/25 °C

[54]

Standard samples (2) 0.69 M SDS/ Aqua C18/isocratic

[183]

Chicken muscles and liver, bovine meat and liver, milk, and baby food (2)/0.1 M SDS, 10% acetonitrile, 0.3% triethylamine, pH 4/Supelco Discovery HS C18 column/isocratic

[184]

Anti-infective: Ethopabate

Chicken muscles and liver, eggs, and baby food (1)/0.1 M SDS, 10% 1-propanol, 0.3% triethylamine, pH 4/Nucleodur C18/isocratic

[185]

Antioxidants: Anserine, butylated hydroxyanisole, butylated hydroxytoluene, carnosine, docecyl gallate, hydroxytyrosol, nordihydroguaiaretic acid, octyl gallate, propyl gallate, tertbutylhydroquinone, 3-tert-buytl-4-

Edible oil (7)/0.1 M SDS, 2.5% 1-propanol, pH 3/Spherisorb ODS-2/isocratic

[186]

Sunflower, corn, and olive oils; margarine, lard, and butter oil (4)/0.1 M SDS, 2.5% 1-propanol, pH 3/Spherisorb ODS-2/isocratic

[187]

(continued)

429

430

4 Micellar Liquid Chromatography: Method Development and Applications

Table 4.3 (Continued) Compounds

Sample (number of analytes)/mobile phase/stationary phase/elution mode/temperature

Ref.

hydroxyanisole, and 2,4,5trihydroxybutyrophenone

Powered and liquid milk, cream of milk, and dietetic supplements (7)/ 0.05–0.15 M SDS, 1–9% 1-propanol, pH 3/Spherisorb ODS-2/isocratic

[188]

Olive oil (5)/0.01 M SDS, 30% 1-propanol acid, pH 2/Lichrosorb RP-18/isocratic

[189]

Meat samples (2)/0.10 M SDS, pH 7/ Kromasil amino/isocratic/25 °C

[190]

Olive extracts (1)/0.05 M SDS, 4% methanol, pH 7/Kromasil C18/isocratic

[191]

Aromatic amines: Benzidine, o-anisidine, o-phenylenediamine, o-nitroaniline, o-toluidine, p-chloroaniline, p-cresidine, and p-toluidine

Waste waters (8)/0.085 M SDS, 3.2% 1-pentanol, pH 7/Princeton Sphere-100 C18/isocratic

[192]

Biogenic amines and metabolites: Agmatine, cadaverine, dopamine, homovanilic acid, hydroxyindoleacetic acid, histamine, 2-phenylethylamine, putrescine, serotonin, spermidine, spermine, tryptamine, tyramine, and agmatine sulfate

Trout samples (9)/0.4 M SDS, 58–70% acetonitrile, pH 3/LiChrospher RP-18/ gradient

[193]

Food substrates (9)/0.4 M SDS, 58–70% acetonitrile, pH 3/LiChrospher RP-18/ gradient

[194]

Fish sauce (4)/0.15 M SDS, pH 7/Kromasil C18 column/isocratic

[195]

Carbamates: Carbaryl, carbofuran, desmedipham, methiocarb, and propoxur

Commercial pesticide formulations and water (5)/0.07 M Brij-35/Kromasil C18/ isocratic

[196]

Water samples (3)/0.15 M SDS, 6% 1-pentanol, pH 3/C18/isocratic

[197]

Cholesterol

Food (1)/0.03 M Brij-35, 10% 1-propanol, pH 7.2/Zorbax CN/isocratic

[198]

Disulfiram

Herbal preparations (1)/0.1 M SDS, 4% 1-butanol, pH 7/C8/isocratic

[199]

Flavonoids: Chrysin, hesperetin, and quercetin

Honey (3)/0.124 M SDS, 7.8% ethanol, 5% acetic acid/Spherisorb C18/isocratic

[200]

Folylpolyglutamate hydrolase activity

Crude tissue extracts (1)/0.2 M SDS/ Whatman PXS 10/25 ODS/isocratic

[201]

Food additives: Curcumin, capsaicin, and piperine

Spice samples (3)/0.15 M SDS, 12.5% 1-propanol, pH 7/Kromasil C18/ isocratic/25 °C

[202]

Fungicides: Ammonium tetramethylenedithiocarbamate, disodium ethylenebisdithiocarbamate, sodium N,Ndiethyldithiocarbamate, sodium N,N-

Pond water (5)/0.0125 M CTAB, 30% methanol, pH 6.8/μ-Bondapak CN/ isocratic

[203]

4.4 Analytical Procedures in MLC

dimethyldithiocarbamate, sodium N-methyldithiocarbamate, and thiram

River water (1)/0.01 M CTAB, 20% acetonitrile, pH 6.3/Spherisorb ODS-2/ isocratic

[204]

Human growth hormone

Escherichia coli fermentation broth/ 0.035 M SDS, 20–30% 1-propanol, pH 6.4/Nucleosil C4/60 °C

[205]

Insecticides: Deltamethrin and imidacloprid

Shampoo (1)/0.12 M SDS, 9% 1-butanol/Kromasil C18/isocratic

[206]

Fruit juices (1)/0.10 M SDS, 2.5% 1-propanol, pH 7/Kromasil C18/ isocratic

[207]

Maleic hydrazide

Tobacco (1)/0.004 M CTAB, pH 7/ Hypersil ODS/isocratic

[208]

Melamine

Dietetic supplements (1)/0.15 M SDS, pH 3/Kromasil C18/isocratic/25 °C

[209]

Preservatives: 4-Aminobenzoic acid, benzoic acid, benzylalcohol, 2-hydroxybenzoic acid, 4-hydroxybenzoic acid; 4hydroxybenzoic acid esters or parabens (methyl, ethyl, propyl and butyl) and sorbic acid

Cosmetics (5)/0.1 M SDS, 2.5% 1-propanol, pH 3/Spherisorb ODS-2/isocratic

[210]

Cosmetics and food samples (7)/2% Brij-35, 20% 1-propanol, pH 3/Lichrosorb ODS/isocratic

[211]

Cranberry juice (14)/0.045 M SDS, 1.5% 1-pentanol, pH 2.5/Kromasil C18/ isocratic/40 °C

[212]

Cosmetics (5)/0.045 M SDS/Zorbax SB C18/isocratic/30 °C

[151]

Phenolic compounds and phospholipids: Caffeic acid, p-coumaric acid, oleuropeina, and tyrosol; Phosphatidylcholine and phosphatidylethanolamine Phenothiazine drugs: Oxopromethazin, dioxopromethazin, diethazin, promethazin, promazin, chlorphenethazin, chlorpromazin, perphenazin, and fluphenazin

Virgin olive oil (6)/0.07 M SDS, 2.5% 2-propanol, pH 3/Nucleosil 120 C18/ isocratic

[213]

Standard (9)/0.005 M CTAB, pH 4.6/ LiChrosorb C8/isocratic

[214]

Polyphenols: Chlorogenic acid, rutin, and scopoletin

Tobacco (3)/0.022 M SDS, 0.45% 1-propanol, pH 5/Eclipse XDB C18/ isocratic/25 °C

[215]

Proteins: Bovine serum albumine, chymotrypsinogen, β-lactoglobulin, lysozime, myoglobin, ovalbumin, and thyroglobulin

Mixtures of proteins and beef heart (5)/ Neodol 91-6, pH 7/Supelcosil LC-8/ isocratic

[216]

Steroids (anabolics and corticoids): Androsterone, bolasterone, boldenone, corticosterone, cortisone, cortisol, dehydroepiandrosterone, deoxycorticosterone, epitestosterone,

Standard solutions (9)/0.075 M SDS, 12.4% acetonitrile or 0.13 M SDS, 4.5% pentanol with 0.01 M Tb(III)/Hypersil C18/isocratic

[217]

(continued)

431

432

4 Micellar Liquid Chromatography: Method Development and Applications

Table 4.3 (Continued) Compounds

Sample (number of analytes)/mobile phase/stationary phase/elution mode/temperature

Ref.

hydroxyprogesterone, 11β-hydroxyprogesterone, 11-ketotestosterone, nandrolone, nandrolone decanoate, medroxyprogesterone, medroxyprogesterone acetate, methyltestosterone, progesterone, testosterone, testosterone enanthate, and testosterone propionate

Standard solutions (13)/0.04 M SDS, 5% 1-propanol/Hypersil C18/isocratic/ 60 °C

[218]

Sugars: Arabinose, glucose, lactose, maltose, and xylose

Infant formula and syrups (5)/0.017 M SDS, 7.7% ethanol, pH 6.7/Kromasil 100-10NH2/isocratic/35 °C

[219]

Sunscreen agents: 2-Ethylhexyl-4-dimethylaminobenzoate, 2-ethylhexyl-4methoxycynnamate, and 2-hydroxy-4methoxybenzophenone

Cosmetic products (3)/0.1 M SDS, 10% 2-propanol, 0.3% triethylamine, pH 3/ Hypersil C8/isocratic

[192,220]

Tetracyclines: Chlortetracycline, doxycycline, minocycline, oxytetracycline, and tetracycline

Animal feeds (5)/0.05 M SDS, 5% 1-butanol, pH 3/Hypersil ODS/ isocratic

[193,221]

Skin-lightening agents: Arbutin and hydroquinone

Plant extracts and cosmetics (2)/ 0.006 M Brij-35, 1% acetonitrile, pH 6/Nova-Pack C18/isocratic

[222]

Vanillin and ethylvanillin

Smoking tobacco (2)/6% Brij-35/ Radicalpack C18/isocratic

[223]

Vitamins: A and E

Standard (2)/3% (w/v) SDS, 15% butyl alcohol, pH 7/C18/isocratic

[224]

re-extraction), using appropriate conditions (pH and solvent composition), followed by evaporation of the solvent. In these operations, a suitable internal standard should be used to correct possible errors in the global procedure (even a simple deproteinization step may cause analyte losses due to drug protein binding). The complexity of the procedures does not allow processing of a large number of samples in a short period of time, which may be needed. An additional problem is the use and disposal of toxic solvents and chemicals, which are dangerous not only to the analyst but also to the environment. The most problematic fluids are those that contain a high protein content, mainly blood, plasma, and serum. The protein content of urine and cerebrospinal and interstitial fluids is smaller. This is the reason of the large effort dedicated to the development of systems that can tolerate the direct injection of physiological fluids, avoiding all previous preparation steps. Advantages are found related to the simplification of the procedures, the reduction in the analysis time with the consequent increase in sample throughput, and the improved accuracy and precision in the quantification of the drugs.

4.4 Analytical Procedures in MLC

4.4.1.2

MLC: A Direct Injection Chromatographic Mode

MLC provides a solution to the direct injection of physiological samples through the solubilization of the protein components in the physiological matrix by interaction with the micelles in the mobile phase, and protection of the stationary phase by coating with surfactant monomers to avoid clogging [8,225] (see Table 4.1). The proteins are, therefore, swept harmlessly away, eluting with or shortly after the solvent front (Figure 4.2). Micelles also release protein-bound drugs, which results in higher concentrations in the mobile phase for partitioning to the stationary phase and detection, although displacement by micelles can be incomplete, depending on the nature of the drug, protein binding, and mobile-phase composition. Interestingly, doubled peaks ascribed to the protein-bound (the peak at shorter retention time) and unbound drug were observed for cephalosporins in serum samples, injected directly into a C18 column and eluted with an SDS mobile phase [89]. The results suggested that the protein binding of a drug might be evaluated, and for a strongly bound drug, the alteration of drug–protein binding by changing the conditions, such as the pH, is required for the recovery and quantitation of the total drug. The anionic SDS is the most common surfactant in the analyses of physiological fluids by MLC. The nonionic Brij-35 can also be used but has the disadvantage of its higher adsorption on C18-bonded stationary phases. Cationic surfactants are not compatible, since they give rise to protein precipitation [7]. Sufficiently surfactant should be used in the mobile phase to obtain an adequate coating of the reversed phase packing to prevent protein adsorption and keep the endogenous proteins solubilized during a chromatographic run. Thus, the surfactant concentration should be well above the CMC, and the organic solvent content should be kept as low as possible. In a recent interesting methodology, the physiological fluid was injected into a chromatographic system using initially micellar conditions. Once the protein band was eluted, a gradient of organic solvent was applied, which disrupted the micelles (going to the HSLC mode), in order to elute highly retained compounds in a short time [112]. Conventional SDS-modified C18 columns can accommodate hundreds of injections of physiological matrices without any increase in back pressure or decrease in column performance (noticeable column damage). The use of the same column for a large array of analyses eliminates the risk of nonreproducibility of offline procedures previous to the chromatographic separation. Another advantage of the direct injection is the low sample demand of only a few microliters, which is interesting in pediatric applications. The first application of MLC to the assay of drugs in physiological fluids was reported in 1985 [7]. In 1989, a kit for drug monitoring using MLC was patented [227]. Today, the versatility of MLC has been demonstrated through the numerous procedures reported for a wide range of drug classes normally monitored, such as analgesics, anticancer drugs, antidepressants, bacteriostats, β-blockers, bronchodilators, catecholamines, diuretics, and steroids, among others (Table 4.1). Analytical procedures have been developed in urine, plasma, serum, and cow milk samples (Figures 4.2–4.5). MLC seems to be also useful in

433

434

4 Micellar Liquid Chromatography: Method Development and Applications

Figure 4.2 Chromatograms of a standard aqueous mixture of 50 ng/ml of three metabolites of propranolol (a), and urine samples of a volunteer before (b) and after (c) the ingestion of a pharmaceutical preparation containing propranolol. Mobile phase: 0.15 M SDS-10%

v/v propanol at pH 4, flow rate 1 ml/min, UV detection at 210 nm. Column: Spherisorb ODS-2 (125 mm × 4.6 mm i.d.). Peak identity: NLT, α-naphthoxylactic acid; NAC, α-naphthoxyacetic acid; PPG, propranolol glycol. (Reproduced with permission from Ref. [41].)

4.4 Analytical Procedures in MLC

Figure 4.3 Chromatograms of: (a) urine spiked with four penicillins (10 μg/ml), (b) cloxacillin excreted in urine as an unchanged drug 2 h after oral ingestion, and (c) cloxacillin excreted 12 h after oral ingestion. All the samples were 1 : 50 diluted with 0.05 M SDS at pH 3. Mobile phase: 0.11 M SDS, 6% propanol

at pH 3, flow rate 1 ml/min, UV detection at 210 nm. Column: Zorbax C18 (150 mm × 4.6 mm i.d.). Peak identity: AMO, amoxicillin; CLO, cloxacillin; DIC, dicloxacillin; and AMP, ampicillin. (Reproduced with permission from Ref. [53].)

435

436

4 Micellar Liquid Chromatography: Method Development and Applications

Figure 4.4 Chromatograms of urine: (a) and serum blanks (b), omeprazole excreted in urine at 3 h (c) and 10 h (d) after oral administration, and omeprazole in serum at 2 h (e) and 6 h (f) after oral administration. Mobile phase: 0.08 M SDS,10% propanol at pH 7.

Column: Kromasil C18 (150 mm × 4.6 mm i.d.). Peak identity: HOME, 5-hydroxyomeprazole; OME, omeprazole; and OMES, omeprazole sulphone. (Reproduced with permission from Ref. [107].)

the screening of illegal drugs in sport (stimulants, narcotic analgesics, anabolic steroids, β-blockers, and diuretics in urine samples) [228]. UV detection is usual, but enhanced detection has been reported by measuring the absorbance in the visible region of drug derivatives formed precolumn, and with a variety of other techniques, such as fluorimetry, amperometry, inductively coupled plasma-mass spectrometry (ICP-MS), and immunoassay (see Section 4.5).

4.4 Analytical Procedures in MLC

Figure 4.5 Chromatograms obtained from milk samples: (a) blank milk at 260 nm, (b) blank milk treated with the diazotization and coupling reagents at 490 nm, and (c) milk spiked with 25 ng/ml of several sulfonamides. Mobile phase: 0.08 M SDS, 8.5% propanol at pH 3, flow rate 1 ml/min, UV detection at

4.4.1.3

490 nm. Column: Kromasil C18 (250 mm × 4.6 mm i.d.). Peak identity: ACE, sodium sulfacetamide; MET, sulfamethiazole; GUA, sulfaguanidine; MER, sulfamerazine; THI, sulfathiazole; and THO, sulfamethoxazole. (Reproduced with permission from Ref. [47].)

Matrix Background Signal

As commented, the micelles in the mobile phase solubilize the high molecular mass proteins. Meanwhile, the underlying alkyl-bonded silica phase is protected by a constant layer of surfactant monomers. The proteins are not retained probably due to the formation of protein–surfactant complexes, which are excluded from the pores of the stationary-phase support, appearing as a broad band at the solvent front (Figures 4.2–4.4). This band is similar, although larger, to the background signal observed in conventional RPLC with physiological matrices, even after carrying out a sample cleanup. The matrix also contains endogenous compounds with peaks at diverse retention times, some of them standing out among other smaller peaks that may represent a serious interference, overlapping the peaks of the analytes and resulting in a useless region in the chromatogram. Unfortunately, the high background

437

438

4 Micellar Liquid Chromatography: Method Development and Applications

signal of the matrix at the beginning of the chromatogram and the peaks of endogenous compounds can affect the detection of early-eluting compounds, representing a limitation of the method. Only when the elution of the analyte occurs after the protein band, the determination will be possible [93,229]. Drugs eluting at shorter times will need extraction procedures. It is, thus, convenient to decrease the retention of the protein band as much as possible. The profile of the band depends on the mobile-phase composition and pH. It can be reduced if UV detection is carried out at larger detection wavelength, or by dilution of the physiological samples with 1% NaCl solution [38]. On the other hand, especial attention should be paid to the endogenous compound yielding the largest peak in the chromatogram of urine matrix, whose retention decreases at increasing concentration of surfactant and organic solvent (this effect is less noticeable for the protein band). The endogenous compound exhibits acid–base behavior in the usual working pH with pKa = 4.5–5.0. Its retention is thus minimized in the pH 5.5–7.5 range. However, most analyses are carried out with mobile phases at pH 2.5–3.5. Therefore, both drugs and endogenous compounds should be considered in the optimization of the separation conditions to avoid overlapping. The reliability of the direct injection approach requires reproducible background signals. The position of the peak of the main endogenous compound in the MLC chromatograms of urine matrix has been measured for male and female volunteers of different ages, diets, and weights [229]. The variation in the retention time of the peak among volunteers was