Carbon-13 NMR Chemical Shifts in Structural and Stereochemical Analysis [1 ed.] 0895733323

Citation preview

Methods in Stereochemical Analysis Series Editor Alan P. Marchand, Denton, Texas, USA

Advisory Board A. Greenberg, New Brunswick, New Jersey, USA I. Hargittai, Budapest, Hungary A. R. Katritzky, Gainesville, Florida, USA Chen C. Ku, Shanghai, China 1. Liebman, Baltimore, Maryland, USA E. Lippmaa, Tallinn, Estonia S. Sternhell, Sydney, Australia Y. Takeuchi, Tokyo, Japan F. Wehrli, Philadelphia, Pennsylvania, USA D. H. Williams, Cambridge, UK N. S. Zefirov, Moscow, Russia

Selected Titles in the Series John G. Verkade and Louis D. Quin (editors) Phosphorus-31 NMR Spectroscopy in Stereochemical Analysis: Organic Compounds and Metal Complexes Gary E. Martin and Andrew S. Zektzer Two-Dimensional NMR Methods for Establishing Molecular Connectivity: A Chemist's Guide to Experiment Selection, Performance, and Interpretation David Neuhaus and Michael Williamson The Nuclear Overhauser Effect in Structural and Conformational Analysis Eiji Osawa and Osamu Yonemitsu (editors) Carbocyclic Cage Compounds: Chemistry and Applications Janet S. Splitter and Frantisek Turecek (editors) Applications of Mass Spectrometry to Organic Stereochemistry William R. Croasmun and Robert M. K. Carlson (editors) Two-Dimensional NMR Spectroscopy: Applications for Chemists and Biochemists (Second Edition) Louis D. Quin and John G. Verkade (editors) Phosphorus-31 NMR Spectral Properties in Compound Characterization and Structural Analysis Jenny P. Glusker, Mitch Lewis, and Miriam Rossi Crystal Structure Analysis for Chemists and Biologists

Carbon-13 NMR Chemical Shifts in Structural and Stereochemical Analysis Kalevi Pihlaja Erich Kleinpeter

veRt

Kalevi Pihlaja

Erich Kleinpeter

Department of Chemistry University of Turku FIN-20500 Turku Finland

Fachbereich Chemie Universitat Potsdam Am Neuen Palais 10 D- 14415 Potsdam Germany

This book is printed on acid-free paper. @

Library of Congress Cataloging-in-Publication Data Pihlaja, Kalevi. Carbon-13 NMR chemical shifts in structural and stereochemical analysis I Kalevi Pihlaja and Erich Kleinpeter. p. Col. -- (Methods in stereochemical analysis) Includes bibliographical references and index. IS13N 0-89573-332-3 (acid-free) 1. Nuclear magnetic resonance spectroscopy. 2. Carbon--Isotopes-Spectra. I. Kleinpeter, Erich. II. Series. QD272.s6P54 1994 541.2'2'028--dc20 94-5420 CIP © 1994 VCH Publishers, Inc.

This work is subject to copyright. All rights reserved, whether the whole or part of the material is concerned, specifically those of translation, reprinting. re-use of illustrations, broadcasting, reproduction by photocopying machine or similar means, and storage in data banks. Registered names, trademarks, etc., used in this book, even when not specifically marked as such, are not to be considered unprotected by law. Printed in the United States of America ISBN 0-89573-332-3 VCH Publishers Printing History: 10 9 8 7 6 5 4 3 2 Published jointly by VCH Publishers, Inc. 220 East 23rd Street New York, New York 10010

VCH Verlagsgesellschaft mbH P.O. Box 10 II 61 69451 Weinheim, Germany

VCH Publishers (UK) Ltd. 8 Wellington Court Cambridge CB I I HZ United Kingdom

Series Foreword

Methods in Stereochemical Analysis provides a forum for critical and timely reviews that deal with the applications of physical methods for determining conformation, configuration, and stereochemistry. The term 'stereochemical analysis" is interpreted in its broadest sense, encompassing organic, inorganic, and organometallic compounds, as well as molecules of biochemical and biological significance. The methods include, but are not restricted to, spectroscopic techniques (e.g., NMR, infrared, UV-visible, Raman, mass, and optical spectroscopy), physical techniques (e.g., calorimetry, photochemical, kinetic, and 'direct" methods such as X-ray crystallography, neutron and electron diffraction), and applied theoretical approaches to stereochemical analysis. In establishing the series, the editor and members of the advisory board seek to attract contributions of the highest scientific caliber from outstanding investigators who are actively pursuing research on stereochemical applications of these various techniques and! or applied theoretical approaches. The editor and board members envision contributions in the form either of a monograph or of a multiauthor treatise with individual chapters contributed by a number of outstanding research scientists. Regardless of format, the editor and board members prefer that the contribution consists of critical and timely reviews that place the author's own work in perspective with regard to other important literature in the field, while at the same time retaining the highly personal character of his or her individual contribution. Indeed, rather than necessarily comprehensive, reviews would be critical and timely. Whatever merit the resulting volumes possess necessarily must derive from the excellence of the individual contributions. Accordingly, the editor welcomes suggestions from members of the scientific community of potential topics for inclusion v

vi

SERIES FOREWORD

in the series, and of names of potential contributors. The editor also welcomes suggestions of a critical nature, which might assist him in better fulfilling the stated objectives. It seems fitting, therefore, that the series be dedicated to its readership among members of the scientific community, for ultimately they will gauge the degree to which the series fulfills its objectives. Alan P. Marchand, Editor Denton, Texas

Preface

During the last two decades 13C NMR chemical shifts have become an extremely important source of information for structural analysis. Moreover, instruments permitting the performance of Fourier transforms make such data easily accessible. Useful correlations, invented regularities, and applications to stereochemical analysis, however, often are hidden even from experienced users of techniques behind the vast amount of experimental findings. Thus a critical survey of the applications of 13C NMR chemical shifts to structure elucidation, stereochemistry, and conformational analysis will be very helpful. To maintain the maximum usefulness and to organize the data to bring out regularities and interdependences and to facilitate theoretical treatments, the discussion is confined mainly to alicyclic chemistry and even there to heterocyclanes, although larger molecules are also surveyed to a certain extent. Not included in this volume are the principles of nuclear magnetic resonance and Fourier transform techniques. A few practical matters inherent in the determination of 13C chemical shifts, however, are dealt with in the Introduction. It is further assumed that the reader has at least access to many earlier outstanding reviews; and hence this material is not repeated unless necessary from the present point of view. The main emphasis throughout this book is on the utilization and applicability of experimental results in the everyday practical problems met by ordinary chemists. The second goal is naturally to demonstrate the strength of 13C chemical shifts as sensitive detectors in structure determination. At least an empirical background and justification will also be given for the different shift effects, with special attention to the "i-effects. When planning this book, we soon realized that the number of relevant publicavii

PREFACE

viii

tions was far too large to permit complete coverage. Hence the literature citations are selective; many papers had to be omitted, especially those that are comprehensively referred to in other recent texts or reviews. Thus the absence of any given reference from this book should not be taken as negation of the significance of the material described in it. Thanks are due to Dr. Timo Nurmi, a former student of Professor Pihlaja, who wrote the first draft of Chapter 3. We still regard this work as valid and competent, and it frees us to deal with some more recent, complementary observations (e.g., Chapters 4A and 4B). We also owe thanks to several other former students of Professor Pihlaja, especially Ph.Lic. Maija-Liisa Kettunen (tetrahydro-l,3oxazines) and Dr. Kyllikki Rossi, since in many places unpublished materials based on their research are mentioned. Kalevi Pihlaja Turku, Finland Erich Kleinpeter Potsdam, Germany May 1994

Contents

1. Introduction

1

1. 1 Practical Considerations 1.2 Assignment of 13C NMR Spectra

References

7

25

2. Substituent Effect Correlations 2.1 General Remarks

27

27

2.2 Parametrization and Selected Rules of Additivity

40

2.3 Straightforward Application of (Single) Substituent Effects to Structure Elucidation 42 2.4 Miscellaneous Applications to Stereochemical Analysis and Structure Determination 46 References

47

3. On the Origin and Stereochemical Connections of the Substituent Effects 51 3.1 a-Effects

52

3.2 Steric Influence on a-Effects

53 ix

x

CARBON-13 CHEMICAL SHIFTS IN STRUCTURAL ANALYSIS

3.3

~-Effects

3.4

~-Effects

57 at Methyl Carbons

3.5 ",-Effects

58

3.6 o-Effects

71

58

3.7 Vicinal and Poly Substitution Effects 3.8 "Buttressing" Effects

72

73

3.9 Influence of the Ring Heteroatoms on the Substituent Effect 73 References

75

4A. Conformational and Configurational Analysis and Substituent Effects on the 13C NMR Chemical Shifts of Alicyclic Compounds 79 4A.l A Complete Conformational Analysis with the Aid of 13C NMR Chemical Shift Correlations: Methyl-Substituted Tetrahydro-l,3-oxazines 79 4A.2 Conformational Analysis of Methyl-Substituted 1,3-Dioxanes 91 and Related Compounds 4A.3 Methyl-Substituted 2-0xo-l,3,2-dioxathianes 4A.4 Oxanes (Tetrahydropyranes) 4A.5 1,3-Dithianes 4A.6 1,3-0xathianes

108

114 118

4A.7 Methyl-Substituted Isochromanes and Related Compounds 124 4A.8 Thianes and Thianium Salts 4A.9 Cyc10hexanes

129

130

4A.I0 Piperidines and N-Methylpiperidines

135

4A.l1 trans- and cis-Decahydroquinolines and Some 141 Related Systems 4A.12 Some Miscellaneous Examples References

153

146

101

xi

CONTENTS

4B. Structural and Stereochemical Analysis of Four-, and Five-, and Seven-Membered Cyclanes and Some Other Systems 157 4B.l Four-Membered Cyclanes

157

4B.2 Influence of Ring Heteroatoms on the 13C Chemical Shifts of Ring Carbons in Five- and Six-Membered Cyclanes 166 4B.3 Five-Membered Cyclanes

168

4B.4 Seven-Membered Heterocyclanes

185

4B.5 Tautomerism and 13C NMR Chemical Shifts

188

4B.6 Alkaloids, Terpene Derivatives, Sugars, and Nucleosides 192 References

201

4C. On the Origin of Stereoelectronic 207 Substituent Effects 4C.l Interplay of Steric and Stereoelectronic Effects 4C.2 The Anomeric Effect 4C.3 The Gauche Effect

209

214

4C.4 The Gauche I Gauche Effect

217

4C.5 Intra- and Intermolecular Hydrogen Bonding References

207

217

218

5. 13C Chemical Shifts as Probes in Configurational and 221 Conformational Analysis 5. 1 General Applications

221

5.2 Stereochemistry of Rigid or Anancomeric Stereoisomers

222

5.3 The Stereochemistry of Flexible Rings and Substituents in Unsaturated Compounds 233 5.4 Factor Analysis of 13C Chemical Shifts

240

5.5 Conformational Analysis of Conjugated Compounds: Dihedral Angle Subject to Steric Hindrance 242 5.6 Stereochemical Analysis of Acyclic Compounds

251

CARBON-13 CHEMICAL SHIFTS IN STRUCTURAL ANALYSIS

xii

5.7 l3C Chemical Shift Subject to Structural Variations to Indicate the Stereochemistry 263 5.8 Stereochemically Relevant Deuterium Isotope Effects on l3C Chemical Shifts 270 5.9 Conformation-Dependent l3C Chemical Shifts in the Solid State 273 5.10 Miscellaneous Applications References

281

288

6. Dynamic NMR Spectroscopy and Frozen Spectra 6.1 The Effect of Dynamic Processes on NMR Line Shape

295

295

6.2 The Determination of Thermodynamic (~GO) and Kinetic Parameters (~G#, 111fT, ~S#) Parameters from Line Shape Variations 297 6.3 Experimental Preconditions

305

6.4 Peculiarities of l3C NMR Spectroscopy

307

6.5 Applications of Dynamic l3C NMR Spectroscopy 6.6 Dynamic l3C NMR Spectroscopy in the Solid State References

307 319

320

7. Application of 13C NMR Spectroscopy to the Configurational and Conformational Analysis of Natural Products, Organometallic Compounds, and Synthetic Polymers 323 7.1 Natural Products

324

7.2 Organometallic Compounds 7.3 Synthetic Polymers References

327

335

336

8. Application of Other NMR Parameters in Combination with the 13C Chemical Shift in Conformational and Configurational Analysis 339 8. 1

IH

Chemical Shift

339

8.2 Homonuc1ear Coupling Constants: and Jlong-range 342

2JH ,H' 3 JH,H'

CONTENTS

xiii

8.3 Heteronuclear Coupling Constants

lle,H

and

31c .H

346

8.4 Nuclear Overhauser Effects and Spin-Lattice Relaxation Times 349 8.5 Strategies for the Assignment of Medium-Sized and Larger Ring Natural Products by Multidimensional NMR Spectroscopy 352 References

Index

361

358

CHAPTER

1 Introduction

Despite the existence of several texts l - 5 that discuss the application of NMR techniques, including the two-dimensional (2-D) methods 3 to the determination of DC chemical shifts, it was felt necessary to briefly summarize some important factors influencing one's choice of approaches.

1.1 Practical Considerations 1.1.1 Sample Preparation The sample, which should not be overly concentrated, must be dissolved in a deuterated solvent with some tetramethylsilane (TMS) as an internal reference. Very clean sample tubes of 5 or 10 mm o.d. are to be used (the height of the solution in the tube can be verified in the manual accompanying the instrument; should be only slightly higher than the receiver coil). A deuterated solvent is needed for field frequency stabilization by means of a special lock system: any detected drift in the deuterium resonance is automatically compensated by a correcting voltage to the magnet power supply, to stabilize the correct field value. The solution must be carefully filtered to remove any solid particles. Degassing from paramagnetic oxygen is also necessary, by • bubbling completely dry nitrogen gas (or any dry rare gas) through the solution for approximately 10 minutes, or • by several freeze-pump-thaw cycles.

2

CARBON-13 CHEMICAL SHIFTS IN STRUCTURAL ANALYSIS

For routine spectra the latter method is not strictly recommended, but it is highly commendable for relaxation time measurements.

1.1.2 Careful Tuning of the Spectrometer Equipment After the sample has been prepared, the tube is carefully lifted into the probe head, the field is locked on the deuterium signal of the NMR solvent used, and the resolution of the spectrometer is optimized by shimming the deuterium signal to its maximum intensity. In general, all functions of the spectrometer should be optimized, including tuning of the probe head and that of the decoupler power with a standard sample, and running the 13C NMR spectrum of the latter under standard conditions. One should start with a real sample only when the signal-to-noise ratio (51 N) and the line width after a given number of scans (n) are good enough. Finally the receiver gain (now under computer control) must be set to prevent saturation and overload of the analog-digital converter (ADC).

1.1.3 Recording the Free Induction Decay (flO) Before the FlO can be recorded, the pulse width and the tip angle (TI/4 is sufficient for routine runs) must be set; the TI-pulse, which should be of minimum intensity, needs to be checked regularly. Then, the delay between the scans must be set (being sure that the signal intensity is independent of relaxational effects); 5 seconds is enough for low molecular weight compounds. If the carbon atoms (especially the quaternaries) relax too slowly, and therefore do not appear in the spectrum, it is necessary to use longer delays or reduce the pulse width and tip angle. The spin system then recovers in shorter time, although more scans will be needed. The next step is to select the spectral width (SW) and the number of data points to detect the signal (see digital resolution, Section 1.1.5). In 13C NMR spectroscopy, the chemical shift range is 250 ppm or less; thus at 50 MHz we expect all resonances to lie within 12,500 Hz (this is the spectral width, sometimes called also the spectral window). Signals lying outside SW are detected inside too but folded (consequently at incorrect frequency, often with incorrect phase and therefore easily assignable). Also the noise outside SW is folded into the spectral area studied and reduces the 51 N ratio. Hence computer-controlled filters were designed which suppress the noise and signals outside SW. Quadrature detection, where the excitation frequency is centered in SW, further reduces the noise and improves 51 N by the square root of n, the number of scans. Phase cycling is usually recommended by the spectrometer manufacturers, especially for 20 runs, to (1) suppress ghost and phantom peaks, (2) suppress the main signals in special pulse sequences as in INADEQUATE (see Section 1.2.6), (3) destroy residual magnetizations, and (4) be able to repeat the experiments fast. In running special 20 pulse sequences, it is strongly advisable to take into account the experience of an NMR specialist. Often phase cycling greatly lengthens multidimensional NMR experiments.

INTRODUCTION

3

Thus, if enough sample is available, the new gradient pulses, 6a which make phase cycling unnecessary, promise substantial reductions in the spectrometer time needed. This property is of special importance for 3D experiments, which still have excessively long run times. The next step is to accumulate the number of scans n necessary for a sufficient 51 N: one must check from time to time by Fourier transformation (FT). The spectrum is obtained in the time domain as free induction decay (FlO) and must be transformed into the more familiar frequency domain by FT; the SIN improves as a function of V;;. Two facts should be kept in mind: 1. Long-term accumulations will lower the resolution, because even smallest disturbances change the field homogeneity even though the system is under computer shimming. 2. As a result of the improvement in SIN through V;; , there is a point after which further improvement of this ratio becomes very inefficient and time-consuming. It should be mentioned that other methods (i.e., other than the common FT) can be used to calculate the regular frequency NMR spectra,6b including the following. 1. The maximum entropy method (MEM): any incidental test spectrum is Fouriertransformed into the FlO and the FlO so obtained is compared with the experimental FlO; iteratively the test spectrum approaches the experimental spectrum, and finally, the correct spectrum of maximum entropy will be obtained. 2. Linear prediction (LP): since the front part of the FlO already contains all necessary information about the spectrum, only this part is considered; the corresponding decay is then calculated, and both the resol ution and the SINenhanced frequency spectrum are obtained. To successfully apply these two methods, dramatically larger amounts of computer time often are necessary.

1.1.4 Precision and Accuracy of Data To select the digital resolution (DR) for obtaining the data with sufficient accuracy, the number of data points (Sf) must be 211 (from the FT algorithm); for IJC NMR spectra, 16K or 32K spectra are recommended. The number of data points is divided (also from the FT algorithm) into DPI2 real and DPI2 imaginary data points; only DPI2 points therefore are useful for detecting the IJC NMR spectrum.

DR = 2SW DP

Thus for a 32K spectrum at 50 MHz:

DR =

2 x 12,500 32,768 = 0.76 Hz (ca. 0.015 ppm)

a result that has the following consequences:

4

CARBON-13 CHEMICAL SHIFTS IN STRUCTURAL ANALYSIS

l. No resonance position can be measured better than ±0.76 Hz (±0.015 ppm). This limitation is no problem if only the chemical shift data are wanted, but it becomes critical when C,H coupling constants for structural assignments are to be determined. 2. Two resonances having chemical shifts that differ by 1.53 Hz (ca. 0.03 ppm) cannot be separated but appear as one absorption in the l3C NMR spectrum. The acquisition time (AQ) of any NMR run is inversely proportional to the DR (DR = l/AQ). Thus, the better the digital resolution, the longer the acquisition time. After the final FID of 211 data points has been obtained, the precision can be improved by adding to the FID another 211 data points with zero value; this is called zero filling. Now we have 2" instead of 2 11 - 1 data points in the real part of the 13C NMR spectrum, and the precision is improved accordingly.

1.1.5 Sensitivity and Resolution Enhancement Mathematical treatment of the FlD before Fourier transformation makes it possible to improve the sensitivity or the resolution of the final spectrum. Sensitivity enhancement is achieved by multiplication of each data point of the FlD by an exponential decay (EM); the larger the selectable line-broadening term (LB), the more rapid the decay. The SIN is improved, but the lines become broader and the resolution poorer. Resolution enhancement can be accomplished by EM multiplication with negative LB or by multiplications with various other functions, such as sine bell and Gauss. The resolution will be improved at the expense of a decrease in SIN, which is not so important in routine 13C NMR spectroscopy (but may be of importance in producing research quality 1D and 2D NMR results).

1.1.6 Spectrum Manipulations

1.1.6.1 Baseline Correction The baseline should be completely flat. Otherwise, it is called a rolling baseline, and the first few data points of the FlD will be falsified for instrumental and operational reasons and can simply be removed. Other solutions to prevent rolling baselines are • special filters or transformation software (recommended by the spectrometer manufacturers), • a short delay between the excitation pulse and the acquisition of the data, and • zero filling at the beginning of the FlD.

1.1.6.2 Phase Correction The spectroscopist wants pure absorption mode spectra; dispersion mode spectra should be produced only if it is desired to obtain exact measurements of the chemi-

INTRODUCTION

5

cal shift of very broad lines. Phase errors after FT are simply removed by phase correction. Magnitude mode and power mode spectra, which include the imaginary data points in the calculation of the line shapes, are less often used for work in one dimension, but often for 20 NMR spectroscopy, where they can simplify the calculation procedure.

1.1.6.3 Referencing the Spectrum The 13C NMR chemical shifts are usually referenced to internal TMS (8 = 0 ppm); 8-values to lower field have positive sign. Peak picking is strongly recommended for broad band decoupled DC NMR spectra, to permit the chemical shifts of the DC resonances of the carbon atoms involved to be found more easily. External referencing, in which the reference compound sealed in a capillary and inserted coaxially into the NMR tube, is possible as well. If compounds other than TMS are used for internal referencing, the following chemical shifts (ppm) with respect to TMS are recommended: CS 2 = 192.8 C6H6 = 128.6 CHCl 3 = 77.2 OMSO = 40.5 (CH 3 hCO = 30.4 Solvent effects are lowest on nonpolar references « 1 ppm) as TMS and cyclohexane (8 = 27.5 ppm) but should be taken into account as concentration effects in cases of reference compounds that are polar (intermolecular interactions, as e.g., hydrogen bonds) and aromatic (formation of collision complexes). It was found that when cyclohexane is used as the internal reference, the relative DC chemical shifts could be measured three times more precisely than with TMS.6c Also the more handy hexamethyldisiloxane (HMOS) was used as internal reference compound; the conversion into the TMS scale of the chemical shifts so measured in different solvents was reported. 60 If deuterated solvents are used for referencing, deuterium isotope effects up to 1 ppm may be present.

1.1.7 Quantitative

13C

NMR Spectroscopy

In IH broad band (BB) decoupled 13C NMR spectra, the signal intensity is, in addition to the number of resonating nuclei C i in the sample, a function of their relaxation times TI and the nuclear Overhauser effect (NOE) enhancement exhibited by C i due to BB IH decoupling. To gain results independent of the latter two effects, the following are necessary: • careful suppression of all NOE enhancements and • selection of a delay between scans that is long enough to allow complete recovery of the signals (ca. 5 times the TI of the slowest relaxing carbon atom).

CARBON-13 CHEMICAL SHIFTS IN STRUCTURAL ANALYSIS

6

The suppression of the NOE enhancements can be managed by (1) running the spectrum with inverse gated decoupling (BB decoupling off during acquisition; the relaxation delay for an error < 1% must be 5T 1 ) 7 and (2) addition of a relaxation reagent, such as chromium triacetylacetonate [Cr(acach): (0.1 M for routine use, but higher concentrations are possible as long as no line broadening occurs J. x If very exact quantitative results are wanted, it is advisable to consult the paper of Cookson and Smith,9a which deals with the dependence of the error in quantitative 13C NMR spectroscopy on • the ratios of T1(carbon) to T1(hydrogen), • the relationship of the length of the relaxation delay without BB decoupling to that of the acquisition with BB decoupling (a question of spectrometer time), and • the concentration of the relaxation reagent and its preferred interaction with special sites of the substrate. The peak height also was used for quantitation of DC NMR spectra. This NMR parameter proved to be less dependent on line asymmetries, small overlaps, baseline humps, and small drifts in frequency or phase of the line than the ordinary integration procedure (see Chapter 4A and Table 4.7). % Beside the factors already mentioned, which are inherent in the FT method, the dynamic range of the ADC of the spectrometry equipment used is crucial for determining the intensity ratio of two peaks of very different intensity. If the bit number is too low, the weakest peak will not be stored at all. Because at least 16-bit ADCs are standard today, dynamic range is not a limiting factor for the majority of quantitative IJC NMR problems. According to the data handling considerations just noted, trustworthy results are most likely to be obtained by integrating nearly noise-free spectra with a completely flat baseline in the pure absorption mode (with good phase correction). Also, the digital resolution of the spectrum should be high enough in relation to the given line width of the detected signals to ensure that the data points available are sufficient to describe the line shapes. Otherwise the probability of noncoincidence between the top of the signal and a data point increases. Completely absent saturation effects, as well as optimal sensitivity of the spectrometer system, are further and deciding preconditions for representative quantitative measurements. OjJ'let effects (on signals more remote from the excitation frequency than others) cannot be prevented but usually are small compared with other uncertainties. If only an approximate quantitative estimate is expected, one can compare the intensities of usually rapidly relaxing, proton-bearing carbon atoms, bearing in mind that the resonances compared arise from carbon atoms having the same number of directly attached protons. This way, quantitative results of sufficient accuracy can be obtained in reasonably short spectrometer runs.

1.1.8 Solid-State

13C

NMR Spectra

The line widths of ordinary IJC NMR spectra, obtained on solid samples, are much too broad to permit the extraction of relevant structural information. The conditions

INTRODUCTION

7

responsible for these broad line widths-namely, (1) the chemical shift anisotropy, (2) the dipolar couplings (now lying in the range of several kilohertz), and (3) the extremely long (under these conditions) spin-lattice relaxation times T1-can be avoided by special techniques of measurement. The application of high power decoupling (HPD), cross-polarization (CP), and magic angle ""pinning (MAS) results in CP-MASJ3C NMR spectra with line widths of only few hertz, to be discussed in the regular manner. The MAS technique entails the extremely rapid (several kilohertz) spinning of the sample about the so-called magic angle of 54°44'.

1.2 Assignment of

13C

NMR Spectra

Routine 13C NMR spectra are run under proton BB decoupling and consequently consist of a number of lines (see Fig. 1.1); the data obtained allow the number of different carbon atoms in the molecule studied to be determined supposing the symmetry arguments are provided. In the 13C NMR spectrum of compound 1, which is shown in Fig. 1.1, the intensity of the resonances of ortho- and metacarbon atoms, by reason of this symmetry, is approximately doubled compared with the other CH I1 carbon atoms of the molecule.

9 8

1

In this detection mode, quaternary carbon atoms are usually easy to recognize because of their lower intensity, since • their Tl relaxation times are longer (and they cannot completely recover within usual relaxation delays) and • the intensity increase of the signal due to NOE enhancement is smaller (as a result of the absence of directly attached protons). These sensitivity disadvantages can be removed by longer relaxation delays, lower tip angles, a more viscous solvent, and lower recording temperatures (the latter two effects shorten T), as do relaxation reagents).

8

CARBON-13 CHEMICAL SHIFTS IN STRUCTURAL ANALYSIS E

Cl Cl

:~~~: \ _ - - - - - - - - - - - - - - - - - - 9£Z·;2 ~------------------

:=>:\\.=================:---------.;;========;:;;c

ti8l '2 ~ 898·,Z lOO 92

-

...

o

1009' - - - - - - - - - - - - - - - - - - -

'0;·[' - - - - - - - - - - - - - - - - - -

o

to

9ZZ Z9 - - - - - - - - - - - - - - - - - -

'0'·[6 - - - - - - - - - - - - - - - - - - - - - - - - -

[09·921

----

PIg tiZI

--------

;~~~:

-

-:::======================== II

~iii ~i~~~!~~~5S;

oal·till

-------------------------

-

INTRODUCTION

9

The effect of simultaneously decoupling all protons is also obtained by a train of composite pulses (WALTZ-I6 decoupling), which improves the SIN and avoids the troublesome dielectric heating of the sample at the same time. If the sample contains magnetically active nuclei other than 1H or 13C (as is the case with organometallic compounds), the corresponding coupling patterns become more complicated; however, they can be decoupled by special equipment-triple resonance.

1.2.1 Distinguishing CH, CH 2 , and CH 3 Several methods for distinguishing the CH, CH 2 , and CH 3 groups of carbon atoms are available.

1. Running proton-coupled 13C NMR spectra. This simplest experiment is very time-consuming (splitting of the signals, no NOE enhancements) and less easy to survey because the long-range C,H coupling constants are more complicated. 2. Off-resonance spectra (SFORD). Splittings due to long-range C,H couplings collapse, and the magnitude of the direct C,H couplings is reduced parallel to the position of a single decoupler frequency just outside the detected spectral window. In the case of several resonances, the doublets (CH), triplets (CH 2 ), and quartets (CH 3 ) are less easy to distinguish. 3. Attached proton test (APT IO , i-modulated spin echo 11a ). The pulse sequence to be used produces the usual BB-decoupled 13C NMR spectrum with the CH and CH 3 resonances as positive signals but the CH 2 resonances as negative ones or vice versa (this depends on the APT sequence applied); see Fig. 1.2. Even in narrow multiline 13C NMR spectra, the differentiation of CH 2 and CH/CH 3 is unequivocal. The differentiation of CH and CH 3 remains unsatisfactory even if the more highly substituted carbon atoms are expected to resonate at a lower field. This problem can be overcome with more sophisticated APT experiments. 4. Signal editing. Pulse sequences, which use nonselective polarization transfer from protons to directly bound carbon atoms, are called by the acronyms DEPT [distortionless enhancement (by) polarization transfer] 12 and INEPT [insensitive nuclei enhanced (by) polarization transfer]. 13 In the case of DEPT, three differ-

Figure 1.1. 13C NMR spectrum of compound 1. The routine BB-decoupled 13C NMR spectrum consists of 16 lines in accordance with the number of different carbon atoms (16 in 1). Quaternary carbon resonances (C=O at 179.18 ppm; C-14 and C-17 at 137.67 and 136.15 ppm, respectively; C-2 at 93.50 ppm) are usually of lower intensity; carbon atoms, which are equivalent for symmetry reasons, are correspondingly higher in intensity (doubled for C-15,15' and C-16,16' at 126.6 and 129.6 ppm, respectively). From the other carbon resonances (using the editing technique given in Section 1.2.1: see Fig. 1.3), the methyl carbon C-18 (at 21.0 ppm), C-13 (at 62.2 ppm), C-ll (at 37.65 ppm), C-6 (at 53.5 ppm), and C-5 (at 46.0 ppm) can be assigned unequivocally. With these assignment techniques alone, it is impossible to differentiate between the remaining CH 2 groups (at 26.0,25.9,25.8,25.2, and 23.2 ppm, respectively). Experimental recording conditions: 6 mg of 1 in 0.6 mL of CDCI 3 , 5 mm o.d. tube; sweep frequency, 125.697 MHz; sweep width, 25 kHz; 64K data points; 32 scans; relaxation delay 3 seconds; pulse width (90°) 4 f.Ls; filter function; exponential weighting; line broadening factor 0.5 Hz.

Q

_ ________ J

_ _________________ ~__~ ______ l __ J__

1

llT-

1

(")

~

o ?

__________ .________________~________________ ~_____ LJJ

w

~ II

I

(")

:r: tTl

s:

§ r

--'I

---,-----,,.--.,---r-...,.--.,..--,----..,.----,,.---~...,__,__....

130

120

110

100

90

eo

70

60

50

40

30

ppm

Figure 1.2. Attached proton test (APT)'o for the IlC NMR spectrum of compound 1 (quaternary carbon resonances at lowfield > 130 ppm excluded). The CH, and the aromatic CH carbon resonances at 31.02, 129.61 (C-IS, IS'), and 126.60 ppm (C-16, 16'), respectively, and those ofC-S and C-6 at S3.S0 and 46.00 ppm, respectively, are obtained as positive signals; the CH 2 resonances at 62.23 (C-13), 37.65 (C-II), and 26.01, 25.87, 2S.79, 2S.24, and 23.24 ppm (C-7-C-10 and C-12, respectively) are displayed as negative ones. Experimental recording conditions: 6 mg of 1 in 0.6 mL CDCI" 5 mm o.d. tube (DEPT pulse sequence used); sweep frequency, 125.697 MHz; sweep width, 25 kHz; 64K data points; 32 scans; pulse sequence optimized on 'JCH = 140 Hz (3.57 ms); pulse width (90°), 9 f..ls.

C/l

:r:

~

C/l

2 C/l ...., :::0

c: ...., c:

(")

~

r

~

E