Paint Analysis: 2nd Revised Edition 9783748604358

Citation preview

Roger Dietrich

Paint Analysis The Handbook for Study and Practice 2nd Revised Edition

1

Cover: chokniti, Adobe Stock

Bibliographische Information der Deutschen Bibliothek Die Deutsche Bibliothek verzeichnet diese Publikation in der Deutschen Nationalbibliographie; detaillierte bibliographische Daten sind im Internet über http://dnb.ddb.de abrufbar.

Roger Dietrich Paint Analysis, 2nd Revised Edition Hanover: Vincentz Network, 2021 European Coatings Library ISBN 3-7486-0425-4 ISBN 978-3-7486-0425-9 © 2021 Vincentz Network GmbH & Co. KG, Hanover Vincentz Network GmbH & Co. KG, Plathnerstr. 4c, 30175 Hanover, Germany This work is copyrighted, including the individual contributions and figures. Any usage outside the strict limits of copyright law without the consent of the publisher is prohibited and punishable by law. This especially pertains to reproduction, translation, microfilming and the storage and processing in electronic systems. The information on formulations is based on testing performed to the best of our knowledge. Please ask for our book catalogue Vincentz Network, Plathnerstr. 4c, 30175 Hanover, Germany T +49 511 9910-033, F +49 511 9910-029 [email protected], www.european-coatings.com Layout: Vincentz Network, Hanover, Germany Printed by: Gutenberg Beuys Feindruckerei GmbH, Hanover, Germany

2

European Coatings Library

Roger Dietrich

Paint Analysis The Handbook for Study and Practice 2nd Revised Edition

3

Foreword I am really glad that you found this book and opened it, curious to find out what to expect. This is the second edition of the book which was first published in 2009 and I can promise a lot of new interesting insights into field analysis and laboratory work with respect to coatings. Compared to the first edition you will find new techniques, additional advanced preparation and sampling methods and a lot of more practical examples. My intention was and is to share my experiences that I have made over the last 30 years in our laboratory and in the field with respect to the analysis of paints and coatings. I am a chemist and my approach to characterize a coating material, a raw material and coating failures is therefore primarily scientifically evidence based. I believe that is very important to have a profound insight into the basic, physical conditions and limitations of the analytical methods in order to comprehend their power and limitations. But I also have worked in the coating industry and know that is not always possible to investigate a problem extensively according to scientific rules. Therefore, I have tried in this second edition to shift the focus more to practical implications of analysing paints and coatings without neglecting the physical and methodical background. I added a section about sampling methods and representativeness because I have made the experience that the implications of proper sampling on the investigation result are too often disregarded. Of course, you know the routine measuring and testing methods that tell something about colour, gloss and other physical conditions of paint films and coating material. These techniques provide the information how a certain material behaves. But they fail when it comes to the question why a tested material exhibits a particular property. This is where this book comes in. I present a toolbox of analytical methods which exist for decades but have not yet found widespread use in the coating industry. However, I would like to demonstrate that the potential of the featured methods is incredibly big and far from being exploited. The possible applications of the methods described here grow with the requirements of the samples to be examined. Almost every month a new question arises which can be answered with a process development based on the methods presented in this book. Although I cannot deny my education as a chemist, the book is aimed at a broad group of users from the coatings industry. It is intended as an aid for the application engineer on site to start solving problems, but as well to provide the laboratory manager with suggestions for new ways of dealing with his tasks. I tried to build a bridge between the scientific view on the reality and the practical requirements on the site and in the applications laboratories. I would like you to hold this book in your hands as a daily reference and tool, and to use it extensively to find out how you can – reveal the causes of coating defects, – investigate paint raw materials with respect to contaminations – or analyse lacquered products in terms of their properties So, if after a short time this manual is lying on your desk or laboratory table with paint stains and marginal notes, I would be more satisfied. In the practical part of the manual I have, therefore, tried to structure and process the topics in such a way that you can draw immediate suggestions for action if, for example, paint craters occur when coating products or if you want to know whether and how two paint batches differ. Of course, this is backed up with theoretical principles in a separate chapter. But the problem solutions described in the practical part should ideally enable

4

you, like a kind of “recipe”, to proceed directly to action with the book in your hand to tackle your daily challenges. I do not see this book only as a handbook documenting the current state of the art, but I would like to use it to stimulate dialogues from which improvements of current procedures or perhaps even new developments can arise. I am therefore always open to suggestions and comments and hope to receive numerous feedbacks, which I will gladly take up and answer as soon as possible. Münster, November 2020 Roger Dietrich [email protected]

Particle characterization | Pigment Dispersibility | Accelerated Paint Stability | Coating Strength

Test our dispersion, separation & adhesion analysers for...

LUMiSizer® Dispersion Analyser

LUMiFrac® Adhesion Analyser

LUMiReader® X-Ray Separation Analyser

Adhesives | Carbon black | Composite bonded joints | Inks | Joint products | Lacquers | Paints & pigments | Plasma coatings | Polymer dispersions | Surface treatment The Next STEP in Dispersion Analysis & Materials Testing

+49-(0)30-6780-6030 [email protected] I www.LUM-GmbH.com

Contents

Contents Part I General information about paint analysis 1 The surface 2 Why paint analysis? 3 Relevance of modern analytical techniques to paint analysis 4 General considerations 5 Chemical mapping 5.1 Infrared microscopy mapping 5.2 TOF-SIMS imaging 5.3 SEM-EDS mapping 6 Depth profiling 7 Instrumentation

11 11 14 15 16 19 20 21 22 22 24

Part II Coating failure analysis 1 The bumpy road to knowledge 2 The analytical procedure 2.1 Inquiry 2.2 Inspection 2.2.1 Macroscopic inspection 2.2.2 Microscopic inspection 2.3 Informed guess 2.4 Instrument selection 2.5 Investigation 2.6 Interpretation 2.7 Iterance 2.8 Implementation 2.9 Incumbency 3 The power of sampling 3.1 The role of sampling in the analytical procedure 3.2 Random sampling and representativeness 3.3 Targeted sampling 3.4 Non-destructive sampling 3.4.1 Wipe sampling 3.4.2 Rinse sampling 3.4.3 Abrasive sampling 3.5 Typical sampling failures 3.5.1 Wrong sample collection 3.5.2 Non-representative sampling 3.5.3 Selecting the wrong amount of samples 3.5.4 Application of a wrong sampling procedure 3.5.5 Inappropriate sampling tools and containers 3.5.6  Insufficient storing and shipping of samples 3.6 Micro­sampling 4 Paint failures and their analytical approach 4.1 Some considerations on failure reasons 4.1.1 Insufficient workpiece preparation 4.1.2 Handling failures

25 27 28 29 29 29 30 32 34 34 37 38 39 40 40 40 45 45 45 46 48 49 49 50 50 51 51 52 53 54 54 55 60

6

Contents

4.1.3 Application conditions 4.1.4 Environment and climate 4.2 Investigation of adhesion failures 4.2.1 Delamination due to substrate contamination 4.2.2 Adhesion defects caused by migration processes 4.2.3 Delamination caused by moulding conditions 4.2.4 Delamination due to application faults 4.2.5 Delamination due to insufficient pretreatment 4.3 Paint cratering and “fisheyes” 4.3.1 Cratering caused by contamination of the paint material 4.3.2  Craters and pinholes caused by substrate contaminants 4.3.3 Craters caused by paint additive agglomeration 4.3.4 Cratering caused by the application conditions 4.4 Bubbles and blisters 4.5 Discolouration 4.6 Hazes and stains 4.7 Paint spots 4.8 Orange peel 4.9 References

Part III Quality control and process analysis 1 Quality control of raw material 1.1 Binders 1.1.1 Identity check 1.1.2 Detection of trace contaminants 1.2 Solvents 1.3 Pigments and fillers 2 Quality control of paint production 2.1 Analysis of filter residues 2.1.1 SEM/EDS analysis of filter residues 2.1.2 FT-IR analysis of filter residues 2.2 Analysis of fogging residues 2.3 Quality check of finished and semi-finished products 2.3.1 ATR-FT-IR screening 2.3.2 TOF-SIMS analysis 2.3.3 Headspace GC-MS analysis 2.4 Paint quality tests 3 Field analysis 3.1 Process analysis of paint shops 3.2 Aerosol analysis 3.3 Operating test 3.4 Sampling of the painting air 3.5 Monitoring of pretreatment steps 3.6 Investigation of the degree of crosslinking in 2-pack paints 3.7 Investigation of paint additive migration 3.8 Marine and aircraft coating inspection 3.9 Handhelds and portables 3.10 References

60 62 63 66 69 71 73 77 78 82 84 86 88 90 93 97 102 108 109 111 112 113 114 115 117 121 124 124 125 127 130 132 133 135 135 135 140 141 144 145 147 149 152 155 157 158 160

7

Contents Part IV Methods of coating analysis 1 Optical light microscopy 1.1 Extended focus imaging (EFI) 1.2 Differential interference contrast (DIC) 2 Fluorescence microscopy 3 Infrared spectroscopy 3.1 Physical background 3.2 Characteristic absorptions 3.3 Instrumentation 3.4 Sample preparation 3.5 Spectrum representation 3.6 Quantification 3.7 Data analysis and evaluation 3.7.1 Data processing 3.7.2 Use of databases 4 Surface infrared spectroscopy 4.1 ATR-FT-IR spectroscopy 4.1.1 Physical background 4.1.2 Depth of penetration 4.1.3 Information depth 4.1.4 Effective path length 4.1.5 Quantification 4.1.6 Detection limit 4.1.7 Instrumentation 4.1.8 Sample preparation 4.2 Reflection infrared spectroscopy 4.2.1 Physical background 4.2.2 External reflection 4.2.3 Instrumentation 4.3 Diffuse reflection spectroscopy 4.3.1 Physical background 4.3.2 Penetration depth 4.3.3 Influences of variable parameters on the spectrum 4.3.4 Sample preparation 4.3.5 Instrumentation 4.3.6 Quantification 4.3.7 Optimization of the measuring parameters 4.3.8 Repeatability 5 Infrared microscopy 5.1 Instrumentation 5.2 Sample preparation 5.3 Infrared microscopy, infrared transmission mode 5.4 Infrared microscopy, reflection mode 5.5 Infrared microscopy, ATR mode 5.6 Line-scan und mapping analyse 6 Raman spectroscopy 6.1 Physical background 6.2 Instrumentation

8

161 161 161 164 164 165 165 166 168 169 170 170 172 173 174 176 177 179 179 180 181 181 182 184 187 188 188 190 196 196 197 199 199 201 202 202 203 203 204 205 207 207 207 210 213 215 216 217

Contents

6.3 Advantages and limitations 6.4 Quantification 6.5 Applications 7 Time-of-flight secondary ion mass spectrometry 7.1 Physical background 7.2 Instrumentation 7.3 Calibration and mass resolution 7.4 Sample preparation 7.5 Spectral evaluation 7.6 Imaging mode 7.7 Quantification 7.8 Summary 7.9 Applications 8 Scanning electron microscopy 8.1 Physical background 8.2 Lateral resolution 8.3 Instrumentation 8.4 Sample condition 9 Electron microanalysis 9.1 Physical background 9.2 Quantification 9.3 Detection limits 9.4 EDS-Imaging 9.5 Applications 10 X-ray photoelectron spectroscopy 10.1 Physical background 10.2 Information depth 10.3 Lateral resolution 10.4 Information retrieval 10.5 Quantification 10.6 Instrumentation 10.7 Applications 10.8 Technical data 11 GC-MS 11.1 Physical background of GC 11.2 Headspace 11.3 Data evaluation 11.4 Application 12 Thin layer chromatography TLC-ATR-FT-IR 12.1 General principle 12.2 Separation procedure 12.3  Identification 12.4 Performance parameters of selected methods

218 219 219 222 223 225 225 226 226 229 229 232 232 235 235 238 239 239 241 242 244 245 245 246 247 247 248 248 249 251 252 253 254 254 255 256 258 259 261 261 261 262 263

13 References Author/Acknowledgements Index

264 266 268

10

The surface

Part I  General information about paint analysis For the investigation of paints, semi-finished products, raw materials and finished coatings, a very extensive “toolbox” of the most diverse analytical methods is available today. This is sometimes a bit confusing for the non-analyst. For an effective use of the individual methods, it is necessary to have a very detailed knowledge of the possibilities and limitations of each method, the excitation modes and the information to be expected. The book that you just opened will try to shed some light on the thicket of methods. A great deal of attention is given to surface processes, as the majority of questions arising in practice are related to surface phenomena in one way or another. But also the so-called “bulk analysis” plays an important role in this book. Classical test methods such as climate-testing, solar simulation etc. are deliberately omitted because these are test methods for a certain parameter but have a completely different objective and are not analysis techniques.

1 The surface This book deals with the application of modern techniques to paint analysis, with a special focus on surface analysis. If one thinks about the word “surface”, it quickly becomes clear what a relative and vague term it is. To a painter “surface” does not mean the same as it does to a surface chemist. To a painter, the surface represents that part of an object which is usually presented to the outside world and can be touched and observed directly, see Figure I.1. A painter considers a brushed

Figure I.1: Solvent droplets on a brushed, hydrophobic steel surface

Roger Dietrich: Paint Analysis © Copyright 2021 by Vincentz Network, Hanover, Germany

11

Part I General information about paint analysis visually clean steel surface. However, it can also be defined as the boundary layer between a solid or liquid material and a surrounding liquid or gaseous phase. A surface physicist would probably refer to it as a phase interface. Alternatively, it could be defined as the area of a solid or liquid thing at which the bulk physical and chemical properties change instantly, a so-called property boundary. A surface chemist, however, is talking about the uppermost molecular layers of a material when he uses the word surface. This is an area that cannot be observed without the help of analytical techniques. In fact, the uppermost layers of an object often determine the quality and behaviour of the material as far as (paint) adhesion is concerned. The uppermost layer of the steel surface is not directly visible without the help of machines. Probably the steel surface exhibits a chemical surface modification which explains the hydrophobic behaviour visible in Figure I.1. Or there is a thin layer of contaminations that produces a hydrophobic property!? – To define this is the task of surface analysis.

Definition of the term surface

So, let’s first define how to use the word surface in this book. A surface is a boundary layer which separates a substrate from the surrounding environment (air, liquid). It is typically 1 nm to 1 μm thick. In contrast, a “thin layer” is defined as being 1 μm to 10 μm thick. The surface plays a significant role in the physical and chemical properties of a material. A contract painter, for example, who paints and coats coils and metal profiles has to rely on the surface quality of the goods he is going to coat. The surface of the raw material that he receives might well look clean. However, the material has a long history before it has been delivered to this company to be painted or coated. During production, storage and transport of a coil, for example, numerous substances may have been adsorbed onto the surface. This surface layer contaminants may not be visible, but it always exists! And sometimes even traces of contaminants can seriously impair the adhesion of a coating to a surface. When it comes to processing of the coil, the chemical composition of the outermost molecular layer plays a significant role. If the coil has been coated with a protective layer of oils to prevent

Figure I.2: AFM (atomic force microscope) image of a paint surface (60x60 μm)

12

The surface corrosion during transport and storage, the paint will exhibit poor adhesion or craters after application. Even a monomolecular layer of some of these oils can have deleterious effects on coating procedures. As these ultra-thin layers are invisible, the unfortunate manufacturer is in fact “blind” as far as the surface quality of his coils is concerned. In most cases, therefore, he will decide to install a cleaning process before applying the coating. But he will do so without knowing if it is necessary and, even worse, without knowing what to remove from the surface. Unfortunately, there is no “magic” process for eliminating all the various kinds of contaminants. His efforts might well produce a surface quality worse than before, due to the presence of oil residues and traces of cleaning chemicals, such as surfactants. Some goods require a pretreatment like e.g. phosphating before coated. The quality of the deposited anti-corrosion phosphate conversion layer is significantly depending on the parameters of the process. Only well crystallized zinc or iron phosphate guarantees a perfect adhesion of the subsequently applied coating. Without analytical methods like e.g. SEM or infrared external reflection spectroscopy, it is impossible to check the performance of the pretreatment process with respect to crystal morphology and percentage of coverage. The author has worked on a lot of adhesion failures issues during the last twenty years which have been caused by poor conversion layer application, although on impulse the painter has been blamed. Another focus is on the coating material itself. As the paint and the painted substrate have to be a chemical match if good adhesion is to be obtained, a few questions need to be asked before the painting process is started. – What is the chemical composition of the substrate surface? – Which pretreatment can be used to improve paint adhesion and what effect will it have? – How do the paint ingredients influence the surface of the material that has to be painted? – What influence do the paint additives have on paint adhesion? Unfortunately, these questions often can’t be answered by simple

Figure I.3: SEM Images of different zinc phosphate conversion layer morphology

13

Part I General information about paint analysis tests or classical chemical analysis, because they require an ability to analyse tiny amounts of substances that have high surface sensitivity. Only the surface analysis techniques described in this book can answer these questions A growing field of application for modern surface analytical techniques is not only paint application but also paint production. Modern high-performance paints have to fulfil many requirements simultaneously that are sometimes hard to match. This does not only create a demand for characterisation of the raw materials and products. The chemical interaction of paint compounds and the reaction between each compound and the ingredients of the substrate (e.g. a polymer) are also key parameters. If, for example, a moulded polymer part has to be coated, it is not just the polymer which is of interest. The manufacturer or supplier of the raw material compounds the polymer to customer demands. In accordance with the requirements imposed on the polymer material, he adds additives to improve flame, light, impact or heat resistance. One parameter the supplier is not concerned about is the paintability of the product made from the granules which he supplies. That is a process which the polymer supplier does not see. However, it has been shown in the past that additives present in polymers “designed” to enhance moulding processes, e.g. offering for example good release from injection moulds, exhibit poor properties with respect to surface finishing by painting. Most of the additives incorporated into a polymer migrate to the surface, driven by temperature, humidity, time or solvents. This sometimes leads to unpredictable results, such as paint adhesion failure, chemical reactions, discolouration, and wetting failure. Many manufacturers of paint for automotive interior parts have therefore discovered, that it is essential not only to know their own paint manufacturing process, but also to learn something about the polymers which have to be painted. This is a task that can easily be fulfilled by the techniques which are described in this book.

2 Why paint analysis? A few years ago, the author was talking to the CEO of a paint manufacturer who had different issues with costumers complaining about the performance of several paint types on a constant level. I asked him if he ever thought about using the powerful toolbox of analytical methods to improve his paint production. He answered: “We sell lacquer, we are not an analytical company!” For him “selling” was the key word. For the quality of the incoming raw material he relied on the factory test certificate of the suppliers, with respect to the quality of the produced paint he banked on his processes. But certificates provided by the supplier do not tell everything, what is essential to know to ensure a safe and high quality production. The best process documentation and compliance cannot guarantee a flawless production. In addition to that, the experience has shown that a lot of customer complaints in the field are not substantiated, because they based on application faults in the coating process. Without knowledge about the details of raw materials, the products and the field application a manufacturer is on a constant “blind flight”. Some paint producers have understood this and profit from a well-equipped laboratory and excellently educated lab staff. But for the commercial department a laboratory and analytical operations are very often simply “lost” costs without profit. On the long run however it pays off to understand the details of a production line. It is a competitive advantage to understand the details of a costumer`s paint shop to avoid unsubstantiated complaints and help him to improve his process. And it is good to know the raw

14

Part I General information about paint analysis tests or classical chemical analysis, because they require an ability to analyse tiny amounts of substances that have high surface sensitivity. Only the surface analysis techniques described in this book can answer these questions A growing field of application for modern surface analytical techniques is not only paint application but also paint production. Modern high-performance paints have to fulfil many requirements simultaneously that are sometimes hard to match. This does not only create a demand for characterisation of the raw materials and products. The chemical interaction of paint compounds and the reaction between each compound and the ingredients of the substrate (e.g. a polymer) are also key parameters. If, for example, a moulded polymer part has to be coated, it is not just the polymer which is of interest. The manufacturer or supplier of the raw material compounds the polymer to customer demands. In accordance with the requirements imposed on the polymer material, he adds additives to improve flame, light, impact or heat resistance. One parameter the supplier is not concerned about is the paintability of the product made from the granules which he supplies. That is a process which the polymer supplier does not see. However, it has been shown in the past that additives present in polymers “designed” to enhance moulding processes, e.g. offering for example good release from injection moulds, exhibit poor properties with respect to surface finishing by painting. Most of the additives incorporated into a polymer migrate to the surface, driven by temperature, humidity, time or solvents. This sometimes leads to unpredictable results, such as paint adhesion failure, chemical reactions, discolouration, and wetting failure. Many manufacturers of paint for automotive interior parts have therefore discovered, that it is essential not only to know their own paint manufacturing process, but also to learn something about the polymers which have to be painted. This is a task that can easily be fulfilled by the techniques which are described in this book.

2 Why paint analysis? A few years ago, the author was talking to the CEO of a paint manufacturer who had different issues with costumers complaining about the performance of several paint types on a constant level. I asked him if he ever thought about using the powerful toolbox of analytical methods to improve his paint production. He answered: “We sell lacquer, we are not an analytical company!” For him “selling” was the key word. For the quality of the incoming raw material he relied on the factory test certificate of the suppliers, with respect to the quality of the produced paint he banked on his processes. But certificates provided by the supplier do not tell everything, what is essential to know to ensure a safe and high quality production. The best process documentation and compliance cannot guarantee a flawless production. In addition to that, the experience has shown that a lot of customer complaints in the field are not substantiated, because they based on application faults in the coating process. Without knowledge about the details of raw materials, the products and the field application a manufacturer is on a constant “blind flight”. Some paint producers have understood this and profit from a well-equipped laboratory and excellently educated lab staff. But for the commercial department a laboratory and analytical operations are very often simply “lost” costs without profit. On the long run however it pays off to understand the details of a production line. It is a competitive advantage to understand the details of a costumer`s paint shop to avoid unsubstantiated complaints and help him to improve his process. And it is good to know the raw

14

Relevance of modern analytical techniques to paint analysis materials in detail to save costs with respect to production failures caused by deficient raw material. Often, when assigning an investigation task to an internal or external laboratory, high expectations are placed on the laboratory: – It should be done quickly (“Best by the day before yesterday”) – There must be no charge – One analysis should, if possible, comprehensively solve the entire possibly complex production problem It should not be surprising, that these wishes will not come true. Solving a production problem normally requires complex procedures, with analytical methods providing the necessary facts. But this process avoids expensive “try and error” actions and aimless activities. The prerequisite for a reliable solution based on measured data are: – A qualified problem analysis – caution during sampling – the selection of appropriate testing methods – specialist knowledge of the laboratory – experience in the evaluation of measurement data – accuracy in the interpretation and formulation of results – sound translation of the results into the process This requires a very intensive exchange with the laboratory in order to effectively combine inhouse knowledge and process know-how with the knowledge generated by the analysis. The dividing line between the internal knowledge of the commissioning company and the external knowledge of the laboratory depends on the personnel structure of the company. A contract coater with a more artisan orientation will usually not know and need many details about paint chemistry, while an automotive supplier specializing in industrial series coating often has its own competence teams on site for all questions of paint chemistry. Both they can also benefit from instrumental analysis methods if these are used in the service and, if necessary, the knowledge of the laboratory service provider is taken into account. Roughly speaking, the less knowledge about the processes and materials in the commissioning company, the more the external knowledge of the service provider can and should be consulted. If this interaction works well and smoothly, in some cases even one analysis costing a few hundred Euro, can save five times as many costs for complaints or faulty batches.

3 Relevance of modern analytical techniques to paint analysis There are hundreds of techniques for analysing paints and coatings. They yield information about viscosity, gloss, haze, hardness, acid value, etc. In other words, they describe the product and its properties. They ensure that the desired level of quality is achieved. On the other hand, standard analytical tools often fall short when failures and production problems arise. The standard techniques are perfect for checking the quality of a product. However, if a product is defective and the root cause has to be investigated, the standard techniques are not very helpful. For example, a monomolecular layer of a release agent on the surface can easily cause severe adhesion failure if

15

Relevance of modern analytical techniques to paint analysis materials in detail to save costs with respect to production failures caused by deficient raw material. Often, when assigning an investigation task to an internal or external laboratory, high expectations are placed on the laboratory: – It should be done quickly (“Best by the day before yesterday”) – There must be no charge – One analysis should, if possible, comprehensively solve the entire possibly complex production problem It should not be surprising, that these wishes will not come true. Solving a production problem normally requires complex procedures, with analytical methods providing the necessary facts. But this process avoids expensive “try and error” actions and aimless activities. The prerequisite for a reliable solution based on measured data are: – A qualified problem analysis – caution during sampling – the selection of appropriate testing methods – specialist knowledge of the laboratory – experience in the evaluation of measurement data – accuracy in the interpretation and formulation of results – sound translation of the results into the process This requires a very intensive exchange with the laboratory in order to effectively combine inhouse knowledge and process know-how with the knowledge generated by the analysis. The dividing line between the internal knowledge of the commissioning company and the external knowledge of the laboratory depends on the personnel structure of the company. A contract coater with a more artisan orientation will usually not know and need many details about paint chemistry, while an automotive supplier specializing in industrial series coating often has its own competence teams on site for all questions of paint chemistry. Both they can also benefit from instrumental analysis methods if these are used in the service and, if necessary, the knowledge of the laboratory service provider is taken into account. Roughly speaking, the less knowledge about the processes and materials in the commissioning company, the more the external knowledge of the service provider can and should be consulted. If this interaction works well and smoothly, in some cases even one analysis costing a few hundred Euro, can save five times as many costs for complaints or faulty batches.

3 Relevance of modern analytical techniques to paint analysis There are hundreds of techniques for analysing paints and coatings. They yield information about viscosity, gloss, haze, hardness, acid value, etc. In other words, they describe the product and its properties. They ensure that the desired level of quality is achieved. On the other hand, standard analytical tools often fall short when failures and production problems arise. The standard techniques are perfect for checking the quality of a product. However, if a product is defective and the root cause has to be investigated, the standard techniques are not very helpful. For example, a monomolecular layer of a release agent on the surface can easily cause severe adhesion failure if

15

Part I General information about paint analysis the material is to be painted. The quantity of substance may be too low to be detected by standard techniques. Poor cleaning procedures in a paint shop can also cause paint defects of a few microns in size. Before this problem can be solved, it is necessary to know what has caused the paint defect. The substance or inclusion particle causing this failure is too small to be characterised by standard techniques. This is an analytical gap that can be closed by the surface analytical techniques described in this book. They will help to answer the question: Why does a product have unexpected properties and why do failures happen? Typical topics in paint analysis are: – What do contaminants in paint layers or wet paint samples consist of? – What is the chemical composition of paint layers at a certain depth from the surface? – How are chemical bonds formed between paint components? – Why does a paint layer peel off a substrate and where does the delamination take place? – What is the reason for paint spots? It should be mentioned that there is no all-embracing technique that can answer these questions. In fact, there are many parameters which influence the decision which technique to employ for the analysis, including: – additional information about the appearance of the defect – preliminary sample investigation by optical light microscopy – chemical and physical properties of the coating – desired detection limit In other words, it takes an experienced user to find the best tool that can answer the questions raised about the sample. These considerations will be discussed later in this book.

4 General considerations Before presenting the techniques discussed in this book, it would be helpful to find a common basic principle to describe them. No matter whether we are talking about infrared spectroscopy or TOF-SIMS or SEM, the main principle consists in probing a sample with radiation. The sample is essentially analysed by radiation that probes for specific properties and characteristics of the material. This radiation, which is called the primary radiation, can consist of electrons, ions, neutral particles and photons, such as infrared waves and X-rays. The primary radiation triggers a reaction specific to the sample that may take the form of the emission of electrons, ions

Figure I.4: General concept of probing the surface of a sample by primary radiation excitation

16

Part I General information about paint analysis the material is to be painted. The quantity of substance may be too low to be detected by standard techniques. Poor cleaning procedures in a paint shop can also cause paint defects of a few microns in size. Before this problem can be solved, it is necessary to know what has caused the paint defect. The substance or inclusion particle causing this failure is too small to be characterised by standard techniques. This is an analytical gap that can be closed by the surface analytical techniques described in this book. They will help to answer the question: Why does a product have unexpected properties and why do failures happen? Typical topics in paint analysis are: – What do contaminants in paint layers or wet paint samples consist of? – What is the chemical composition of paint layers at a certain depth from the surface? – How are chemical bonds formed between paint components? – Why does a paint layer peel off a substrate and where does the delamination take place? – What is the reason for paint spots? It should be mentioned that there is no all-embracing technique that can answer these questions. In fact, there are many parameters which influence the decision which technique to employ for the analysis, including: – additional information about the appearance of the defect – preliminary sample investigation by optical light microscopy – chemical and physical properties of the coating – desired detection limit In other words, it takes an experienced user to find the best tool that can answer the questions raised about the sample. These considerations will be discussed later in this book.

4 General considerations Before presenting the techniques discussed in this book, it would be helpful to find a common basic principle to describe them. No matter whether we are talking about infrared spectroscopy or TOF-SIMS or SEM, the main principle consists in probing a sample with radiation. The sample is essentially analysed by radiation that probes for specific properties and characteristics of the material. This radiation, which is called the primary radiation, can consist of electrons, ions, neutral particles and photons, such as infrared waves and X-rays. The primary radiation triggers a reaction specific to the sample that may take the form of the emission of electrons, ions

Figure I.4: General concept of probing the surface of a sample by primary radiation excitation

16

General considerations Table I.1: List of analytical techniques Primary Secondary radiation radiation Technique Electrons

Infrared radiation

Abbreviation

Analysed area

auger electron spectroscopy

AES

uppermost molecular layer

scanning electron microscopy

SEM

sample surface down to a depth of a few microns

X-rays

electron microanalysis

ESMA EDS WDX

sample surface down to a depth of a few microns

infrared radiation

surface infrared spectroscopy

FT-IR ATR IRRAS

sample surface down to a depth of a few microns

infrared microscopy

IRM XPS ESCA

electrons

X-rays

electrons

X-ray photoelectron

ions

ions

secondary ion mass SIMS spectroscopy TOF-SIMS

sample surface down to a depth of a few nanometres uppermost molecular layer

or X-rays. This “reaction” by the sample is detected by an electronic system composed of an analyser and a detector. The result can be displayed as a spectrum on a computer or be printed on paper. The last step of the process is data evaluation by an experienced analyst. The evaluation must include – plausibility check – comparison with databases – interpretation with respect to the analytical problem The nature of the interaction which occurs between the probing beam and the sample depends on the type, energy and angle of incidence of the probing radiation and, of course, the sample material. The primary radiation interacts with the sample in a specific way. Each type of sample reaction can be detected separately and analysed to reveal the chemical and physical composition of the sample and its surface. The radiation emitted by the sample is called secondary radiation. Each type of primary radiation can produce a different type of secondary radiation. Probing with an electron beam, for example, may lead to the formation of: – secondary electrons – X-rays – back-scattered electrons – fluorescence Each type of radiation conveys different information about the sample that all adds up to a comprehensive understanding of the sample’s properties. Not only the primary radiation, but the secondary radiation emitted by the sample, can consist of electrons, ions, neutral particles and photons that result from sample excitation or reflection of the primary radiation. The latter is a consequence of diffraction and dispersion that change the energy, angle and intensity of the primary radiation in accordance with the topography, structure and chemical

17

Part I General information about paint analysis composition of the sample. The secondary radiation emanating from the sample is detected, analysed and displayed in the form of an angle-, energy- or mass-resolved spectrum, which contains information about the sample and its surface. The various types of probing primary radiation and detected secondary radiation have spawned more than 50 different analytical techniques over the decades. Some of them are useful for solving practical problems and have made their way into routine work. Many of them, however, never passed the experimental stage and have very limited application to technical samples outside of academia. Figure I.5 sums up the abbreviations of a few techniques sorted by the type of primary radiation. A sample excitation by photons (left part) can be performed e.g. by infrared radiation resulting in absorption and the “answer” of the sample can be analysed by reflection absorption spectroscopy (IRRAS), internal reflection spectroscopy (MIR or ATR) or diffuse reflection spectroscopy (DRIFT). A photon excitation can also be done using a laser resulting in the release of ions (LAMMA) or by X-rays that produce photoelectrons analysed by X-ray photoelectron spectroscopy (XPS). In this book, the author covers those techniques which have proven to be very useful for routine work and can deliver data in a reasonable time and at reasonable cost.

Figure I.5: Surface analysis techniques sorted by the excitation radiation type: photons, electrons and ions

Figure I.6: Measuring modes in surface analysis

18

Chemical mapping The techniques mentioned in Table I.1 yield different data about the sample. Each has its particular strengths and weaknesses. It is very important to appreciate this when trying to find the right combination for the given analytical problem. It is commonly said that one technique on its own is no use and so a combination is the best way of achieving the right results. The important parameters of a technique are its – information depth – detection limits – information content – suitability for technical problems Some techniques, for example, allow only very limited sample sizes, which sometimes renders the technique useless for “real world samples”. Others require vacuum conditions, and that excludes liquid or volatile samples. Only a handful of techniques have proven useful for routine work. The limiting features are: – measuring time per sample – suitability for technical samples – comprehensive databases of reference materials – sample preparation Another important question is the sample area to be analysed. If, for example, a paint crater a few microns in diameter has to be analysed for possible surface contaminants capable of causing cratering, the technique to use must allow for spot analysis. This means the investigation of a very small spot with a lateral resolution of a few microns. If, on the other hand, it is the general surface quality of the sample which is of interest, a larger area measurement must be performed in order that a representative image of the surface composition may be obtained. Analysis of the distribution of a specific substance over a certain area calls for a scanning technique that generates a chemical map of the analysed area. On the assumption that not only the chemical composition of a surface area has to be analysed but also the depth distribution, a depth-profiling mode needs to be chosen. That entails sputtering the sample layer by layer and analysing the surfaces as they become exposed.

5 Chemical mapping The analysis of a sample like, e.g. a coating on a polymer surface for the chemical composition answering the question “which substances can be identified in the coating” sometimes leads to the additional question: “where are these substances located?. This means the desired information is the variation of the chemical composition at different points of the surface or inside the bulk material. Imagine a black coating that seems to be homogeneous on the first sight. But there are some strange features in the surface topography like Figure I.7 shows. This might lead to the question if the visible inhomogeneities are correlated with chemical properties!? The investigation of this question asks for a two-dimensional chemical analysis resolving the chemical composition with a high lateral resolution. Two different approaches have to be distinguished with this kind of analysis: mapping and imaging. The so-called mapping can be done by scanning and probing the surface in two dimensions point by point or line by line acquiring a full spectrum for each point and resolving the chemical

19

Chemical mapping The techniques mentioned in Table I.1 yield different data about the sample. Each has its particular strengths and weaknesses. It is very important to appreciate this when trying to find the right combination for the given analytical problem. It is commonly said that one technique on its own is no use and so a combination is the best way of achieving the right results. The important parameters of a technique are its – information depth – detection limits – information content – suitability for technical problems Some techniques, for example, allow only very limited sample sizes, which sometimes renders the technique useless for “real world samples”. Others require vacuum conditions, and that excludes liquid or volatile samples. Only a handful of techniques have proven useful for routine work. The limiting features are: – measuring time per sample – suitability for technical samples – comprehensive databases of reference materials – sample preparation Another important question is the sample area to be analysed. If, for example, a paint crater a few microns in diameter has to be analysed for possible surface contaminants capable of causing cratering, the technique to use must allow for spot analysis. This means the investigation of a very small spot with a lateral resolution of a few microns. If, on the other hand, it is the general surface quality of the sample which is of interest, a larger area measurement must be performed in order that a representative image of the surface composition may be obtained. Analysis of the distribution of a specific substance over a certain area calls for a scanning technique that generates a chemical map of the analysed area. On the assumption that not only the chemical composition of a surface area has to be analysed but also the depth distribution, a depth-profiling mode needs to be chosen. That entails sputtering the sample layer by layer and analysing the surfaces as they become exposed.

5 Chemical mapping The analysis of a sample like, e.g. a coating on a polymer surface for the chemical composition answering the question “which substances can be identified in the coating” sometimes leads to the additional question: “where are these substances located?. This means the desired information is the variation of the chemical composition at different points of the surface or inside the bulk material. Imagine a black coating that seems to be homogeneous on the first sight. But there are some strange features in the surface topography like Figure I.7 shows. This might lead to the question if the visible inhomogeneities are correlated with chemical properties!? The investigation of this question asks for a two-dimensional chemical analysis resolving the chemical composition with a high lateral resolution. Two different approaches have to be distinguished with this kind of analysis: mapping and imaging. The so-called mapping can be done by scanning and probing the surface in two dimensions point by point or line by line acquiring a full spectrum for each point and resolving the chemical

19

Part I General information about paint analysis composition by extracting key features of the spectra. There are a few methods available to perform this kind of analysis which deliver different information: infrared microscopy imaging, Raman microscopy mapping, TOF-SIMS imaging and SEM mapping. Mapping means that a region of interest (ROI) is visually selected and (rather than taking a simple intensity spectrum of the whole area) rastered by the exciting beam point by point and line by line. At each point a whole spectrum is collected and stored. This can be achieved either by moving the sample on a computerized stage (like with an infrared microscope) and keeping the exciting beam fixed. Or the beam of the primary radiation is rastered while the sample is fixed on a stage like scanning electron microscopy (SEM) does. The whole set of spectra of the specified area of interest is the database of this measurement. In a second data analysis process after the measurement special features can be selected and extracted to produce images. Imaging in contrast means taking all spectral intensities simultaneously from the whole ROI and filtering the information according to one selected part of the spectrum.

5.1 Infrared microscopy mapping

a)

b) Figure I.7: a) Optical microscope image (DIC) of a 2K high gloss polyurethane coating surface with unusual topographic features, b) infrared microscopy mapping analysis of the coating surface (marked area); false colour image displaying the peak intensity of the 2238 cm-1 peak of free (unreacted) isocyanate groups of the hardener

20

The imaging mode of infrared microscopy analysis, which will be described in detail in Part IV Chapter 5, delivers molecular information with an information depth of 0.5 to 1 μm into the surface. Figure I.7b shows a false colour 2D image of the coating surface of Figure I.7a. The colour codes the intensity of a specific signal of the NCO group of the isocyanate hardener between dark blue (no free NCO group) and white (high NCO content). This coloured image is the translation of a complex physic-chemical procedure (the excited vibration of the NCO group) into an easy-to-understand picture. The image demonstrates that the content of free isocyanate is not homogeneously distributed in the uppermost layer of the coating. This leads to the conclusion that there must have been something wrong with the mixing of the polyurethane components. Whereas the “back end” (ATR-FT-IR microscopy) is something you cannot understand without knowing vibrational theory, this “front end” picture can transport the key information in a glance.

Chemical mapping

5.2 TOF-SIMS imaging In contrast to the infrared microscopy mapping (IRMM) the TOF-SIMS imaging allows for a higher spatial resolution which is shown in Figure I.8. The cross section of a 25 μm three-layer primer system consisting of a phthalate primer in between two chlorine primer layers has been analysed by a primary ion beam of Ar+ ions in a TOF-SIMS V mass spectrometer.

Figure I.8: Optical microscopy image and TOF-SIMS image of a cross section through a three layer primer system displaying the secondary ion Cl - of chlorine compounds and (C6H5)COO- of phthalate anions

Figure I.9: SEM-EDS mapping result of a paint defect (top left) sowing the false colour intensity images for calcium, silicon and iron (clockwise)

21

Part I General information about paint analysis The false colour images of the Cl- ion and the (C6H5)COO- ion display the intensity of the two characteristic secondary ions in the spectrum of the negative secondary ions of the cross section and thus is an extract from the whole data set of the two dimensional measurement. Whereas IRMM detects vibrational transitions of characteristic bonds and displays the intensity of the absorptions, the peak intensity of a characteristic fragment like C6H5COO- is derived from the integration of the peak area. This is not molecular information because this fragment can originate from a lot of molecules that contain this specific group. However, knowing that the fragment C6H5COO- can be attributed to phthalate compounds and analysing further fragments associated with this secondary ion like e.g. 191 u for PET or 219 u for PBT, the molecular information can be deduced from the mass spectrometry data. The image shows, that there is a phthalate ester-based primer in between a “sandwich” of thin layers of chlorine containing primers.

5.3 SEM-EDS mapping Another method for the visualization of lateral chemical differences which will be described in this book is the EDX (= energy dispersive X-ray analysis) or EDS (= energy dispersive spectroscopy). The sample is scanned point by point and line by line and excited by a beam of electrons between 1 keV and 30 keV. This results in the emission of characteristic X-rays for each element present in the target area. The intensity is detected and gathered by a detector and the result can be extracted from this so-called hypermap as 2d false colour image for each detected element. In contrast to the TOF-SIMS and the infrared microscopy mapping this technique (only) delivers elemental information. If, for example, calcium has been detected (=> Figure I.9) and the lateral distribution is displayed, it cannot be distinguished between calcium carbonate, calcium hydroxide or a calcium soap.

6 Depth profiling The imaging techniques enable the two-dimensional characterisation of a certain sample area in the uppermost plane. But some issues ask for analyses into the depth of the sample. Depending on the so-called information depth or penetrating depth of the applied method this allows for chemical information of the uppermost surface down 1 to 2 μm into the bulk. The information depth is the distance (measured down from the very top of the surface into the bulk) to the area inside the sample, where the information is generated that can be measured by the detector. Everything that lies deeper is not available from the top without sample preparation. But sometimes you might want to go deeper to understand what is happening for example 5 to 10 μm underneath the visible surface. The question is: how to achieve this goal with the minimum influence on the sample. SEM-EDS, for example, offers a limited variation of the information depth by adjusting the acceleration of the primary electrons that excite the sample surface. But the most attractive approach would be sputtering (“slicing”) the sample layer by layer from the top and analysing the surfaces as they become exposed. The TOF-SIMS method offers this mode. But the disadvantage of this process is that each sputtering process changes the surface in a way that most of the organic compounds are cracked or even destroyed. So, the target area is irreversibly modified before the information can be collected. A non-destructive method is offered by the Confocal Raman Microscopy (see Chapter IV). With this method the sample is excited by a laser beam which is focussed to a very narrow so-called

22

Part I General information about paint analysis The false colour images of the Cl- ion and the (C6H5)COO- ion display the intensity of the two characteristic secondary ions in the spectrum of the negative secondary ions of the cross section and thus is an extract from the whole data set of the two dimensional measurement. Whereas IRMM detects vibrational transitions of characteristic bonds and displays the intensity of the absorptions, the peak intensity of a characteristic fragment like C6H5COO- is derived from the integration of the peak area. This is not molecular information because this fragment can originate from a lot of molecules that contain this specific group. However, knowing that the fragment C6H5COO- can be attributed to phthalate compounds and analysing further fragments associated with this secondary ion like e.g. 191 u for PET or 219 u for PBT, the molecular information can be deduced from the mass spectrometry data. The image shows, that there is a phthalate ester-based primer in between a “sandwich” of thin layers of chlorine containing primers.

5.3 SEM-EDS mapping Another method for the visualization of lateral chemical differences which will be described in this book is the EDX (= energy dispersive X-ray analysis) or EDS (= energy dispersive spectroscopy). The sample is scanned point by point and line by line and excited by a beam of electrons between 1 keV and 30 keV. This results in the emission of characteristic X-rays for each element present in the target area. The intensity is detected and gathered by a detector and the result can be extracted from this so-called hypermap as 2d false colour image for each detected element. In contrast to the TOF-SIMS and the infrared microscopy mapping this technique (only) delivers elemental information. If, for example, calcium has been detected (=> Figure I.9) and the lateral distribution is displayed, it cannot be distinguished between calcium carbonate, calcium hydroxide or a calcium soap.

6 Depth profiling The imaging techniques enable the two-dimensional characterisation of a certain sample area in the uppermost plane. But some issues ask for analyses into the depth of the sample. Depending on the so-called information depth or penetrating depth of the applied method this allows for chemical information of the uppermost surface down 1 to 2 μm into the bulk. The information depth is the distance (measured down from the very top of the surface into the bulk) to the area inside the sample, where the information is generated that can be measured by the detector. Everything that lies deeper is not available from the top without sample preparation. But sometimes you might want to go deeper to understand what is happening for example 5 to 10 μm underneath the visible surface. The question is: how to achieve this goal with the minimum influence on the sample. SEM-EDS, for example, offers a limited variation of the information depth by adjusting the acceleration of the primary electrons that excite the sample surface. But the most attractive approach would be sputtering (“slicing”) the sample layer by layer from the top and analysing the surfaces as they become exposed. The TOF-SIMS method offers this mode. But the disadvantage of this process is that each sputtering process changes the surface in a way that most of the organic compounds are cracked or even destroyed. So, the target area is irreversibly modified before the information can be collected. A non-destructive method is offered by the Confocal Raman Microscopy (see Chapter IV). With this method the sample is excited by a laser beam which is focussed to a very narrow so-called

22

Depth profiling focal volume. By focussing the laser this vocal volume (about 1 μm3) is incrementally moved down from the surface into the bulk. This so-called axial profiling scans the chemical composition along the optical axis. So e.g. for a 5 μm layer of a primer on polymer substrate this is the only non-destructive way of characterizing the primer layer and the interface between primer and polymer substrate on a molecular basis. The most commonly used method of gathering information about deeper layers is lateral profiling of a polished cross section. The sample is embedded into a resin, sectioned perpendicular to the surface, grinded and polished achieving a cross section surface that can be analysed by different techniques subsequently.

Figure I.10: Information depth for selected methods

Figure I.11: Axial profiling of a coating layer from the surface down to the substrate by confocal design (focussing the exciting beam along the optical axis) (left) and lateral profiling of a cross section (right)

23

Part I General information about paint analysis

7 Instrumentation Some of the analytical techniques described in this book require vacuum conditions (SEM, TOFSIMS, XPS). Other systems work in ambient air condition (ATR-FT-IR, Raman). Although they differ significantly in detail and in their chemical background, the instrumentation follows a general concept. The instrumentation setup for the majority of techniques consists of an excitation system (the primary system) that generates photons, electrons or ions. The primary beam is directed by a focusing system onto the sample surface and into the desired area. Some techniques have an additional sputtering system (e.g. an ion gun) that allows for subsequent sputtering of layers and thus for depth profiling. After interaction of the primary beam with the sample (surface), the excited secondary radiation is collected by a ray optics system which directs the secondary beam towards the analyser. The analyser separates the secondary radiation spectroscopically by energy, direction or mass. The detector records the separated or resolved signals and measures their intensity. The signals generated by it are displayed as a spectrum, which is a chart of intensity versus wavelength, mass or energy.

Figure I.12: General instrument setup for surface analysis systems

24

The bumpy road to knowledge

Part II

Coating failure analysis

1 The bumpy road to knowledge The main application of surface analytical methods is without doubt that of failure analysis. Types of paint failure are spots, adhesion, wetting and flow problems, craters and stains. Very often, this issue demands the detection and identification of very low quantities of paint components and contaminants in a small sample area (microanalysis). For these tasks, surface analysis provides an extensive set of instruments. But that is not the only prerequisite for a successful problem troubleshooting. Therefore, the author wants to point out which steps are also important before diving deep into the analysis of different failure types: If a production fails and one does not know why, the urgency of beginning a journey of knowledge that makes the mystery familiar is obvious. In order to understand – how a coating product works – why it fails – how it is composed – what can be done if an unexpected coating result occurs – how the quality of raw material can be controlled – how my products can be tested… facts and data are necessary. And in mind: „Furious activity is no substitute for understanding!” Especially, when it comes to paint failure analysis people tend to rely on conjectures more and more, rather than focussing on facts. But only facts are reliable. This book presents various examples of “real world” challenges that have occurred in this or a similar way in the author's professional experience over the last twenty-five years, happening somewhere every day, and shows the ways to solve certain problems. Attention is focused on the analysis of paint defects, as well as on raw material control, process control and production monitoring. It is important to notice

Figure II.1: Powder coating booth

Roger Dietrich: Paint Analysis © Copyright 2021 by Vincentz Network, Hanover, Germany

25

Part II Coating failure analysis that the methods go beyond standard paint quality tests and can be adjusted to the actual situation which makes them flexible for any kind of upcoming problem. If a solution for a problem is based on science it works, if it is based on assumptions it often leads into dead ends, costs a fortune and wastes resources. To achieve facts and data you need machines, analytical equipment and manpower, but it pays off. Facts and data do not solve problems. They must be combined with knowledge and expertise. But if these requirements are fulfilled, the right path is under your shoes. The methods described here, serve as a tool to elucidate the root cause of the failure. They should enable the optimization and improvement of production processes which would otherwise afford a lot of money and time for try and error without having these techniques. On the other hand, the facts generated by measurements and analyses are sometimes used to “share” the costs of production failures with the suppliers. The scientific methods are an important and versatile tool to understand and solve problems, but they are not the universal remedy. If a coating failure happens, it is a good choice to think about the procedure how to solve it before starting into action. Failure analysis includes the search for the cause of a malfunction during production of coating materials as well as the investigation into coating application failures. A paint crater observed at the end of the production line of a product, for example, can be caused by: – coating material production failures (wrong choice of paint components, mixing errors) – storage failures – application failures – insufficient cleaning and pretreatment of the raw product Therefore, a comprehensive search must cover all aspects from raw materials to the finished product. The challenge is the total amount of substances that have to be detected and identified. For example, one droplet of a fluorocarbon lubricant of 10 μm diameter can be the cause for a paint crater. This is a total of a few nanogram of material which has to be found and identified. It is obvious that classical low-cost analytical techniques of routine laboratory processes do not have the power to achieve this. In fact, this task asks for “heavy machinery” like e.g. TOF-SIMS and highly educated manpower. But it is worth it, because a problem solution is nothing without understanding the problem. The efficient failure analysis requires: – the appropriate design of experiment – the appropriate sampling – the appropriate instrumentation – a qualified and educated evaluation of the data and – last but not least the translation and transfer of the achieved knowledge into the process If one of these steps fails, the whole analytical process is worth for nothing. These issues will be described in detail in the following chapters.

26

The analytical procedure

2 The analytical procedure It is a common misbelief that the process of investigation of coating failures starts when the samples arrive in the laboratory. In fact, the process sequence starts earlier on the site of production. The whole process flow of failure analysis comprises: – the investigation goal – the design of experiment – the sampling procedure – the storage of samples – the transport to the laboratory – the first sample inspection – sample preparation – selection of the appropriate measurement technique – sample measurement – data evaluation – data interpretation – report – perception of the results Each step has its challenges and can ruin the whole outcome with respect to the final goal of removing the root cause of the fault. But what is the best path to a sustained solution of production failures? A close look at each step one realizes the obstacles that can appear. It will be demonstrated how internal production knowledge, low cost technical aids, laboratory analyses, expert knowledge and last but not least common sense generate a base on which a reliable solution can grow Normally the trouble starts with suddenly and unexpectedly appearing failure parts. Of course, this requires immediate action in order to find the cause and define measures to eliminate the issue. But how? Quite often the author has experienced a procedure like this when called to help with these issues: – let us have a meeting

Figure II.2: The risk to fail in the analytical process (arbitrary units)

Figure II.3: Coating failure analysis proceeding after sampling

27

Part II Coating failure analysis – – – – – –

let us look for someone guilty let us have a meeting again put pressure on the employees let us have a meeting again ask for different plans to solve the problem then panic

All these “actions” are driven by the understandable desire to get rid of the issue as soon as possible to make sure that the costumer can be served without any delay. The call for a systematic investigation of the root causes is often unheard facing pressure from costumers and financial accountants. Under the pressure of doing “something” plans and measures often suffer from a severe lack of reliable facts but are based on assumptions, rumours or feelings. Nevertheless, sometimes by chance the problems disappear, but nobody can say why. Moreover, nobody can guarantee that it will not pop up again and there is no solution for the future. But what is the value of a removal of the issue instead of a solid solution? For sure, the systematic approach is slower and sometimes does not deliver “results” within a short period of time, but at the end it is more reliable and thus cheaper. Of course, a good plan of what to do is a good start: Figure II.3 shows the favourable path through the whole investigation process. The necessary tool are analytic methods to gather facts and place the latter in opposition to assumptions. But the analytical option is “only” delivering reliable data which are necessary to circle the source of the failure and the area of the production line, where it can be located. From that end to the real solution there is more to do.

2.1 Inquiry Let us assume a paint crater issue appearing in an automotive supplier plant. Someone sampled a failed bezel and presented it. To solve the issue, you would want to know as much as possible about the circumstances of this sudden production failure. A good method is an interview of all the people involved in the project. The questions to be answered with respect to this case are: – Where does the sampled part come from? (directly from your process, somewhere from the supply chain, from the customer) – What is the history of this sample? (which does not mean which way parts of this kind typically go, but what the sample in your hand has “experienced”) – How many parts are affected? (percentage of the lot/production) – When did the issue appear? – When has it been realized? Figure II.4: Visible light microscopy picture of a paint crater

28

The analytical procedure – Are there still good parts which have been produced in the same process? – Are there any changes of production parameters that can be correlated to the failure? – Are there any unusual circumstances around the production line that might have influenced the issue? (e.g. construction works, repair, cleaning procedures) – Are there any (proven) correlations to certain lots, production shifts, material lots, production tools and so on? Please note: You can never enquire too much but always too little. Be curious and vigilant during the whole data acquisition process. Each piece of information can be very precious, when it comes to the data evaluation of the analyses. The history of a sample also includes the exact circumstances of the sampling. Interviews of the people on the site that deal with that kind of product every day and know the machines very well are a versatile tool.

2.2 Inspection 2.2.1 Macroscopic inspection The next step of the analytical process is a personal, visual inspection of the samples. The human visual sense is very sensitive and sometimes realizes details that are hard to measure by machines. It is sometimes quite frustrating to admit that e.g. discolourations are clearly visible but often cannot be detected even by the most sensitive instruments. On the other hand, there are people who call themselves experts and believe that they can classify a failure type just by macroscopic visual inspection. This must be doubted because e.g. a paint bubble of 20 μm diameter in a primer layer of a multicoat system may appear as a crater or a speck in the topcoat upon visual inspection and this is impossible to distinguish without instrumental assistance. But the watchful eye can find out very important details about the samples which can be precious, when it comes to the evaluation of the analytical data: – Do the samples comply with the targeted goal of the analyses and the design of experiment? – How does the failure appear? – Are the failure spots (e.g. paint specks) randomly distributed or located in specific areas? – Are there any optical differences between failed parts and sound parts? – Does the sample exhibit further unusual features which might hint at application failures? – How many failure spots are noticeable? – Are all visible failure spots similar or are there different failure types on the same part? Of course, this list is not comprehensive. Each individual issue requires a customized set of questions. It is good practice that the person who plans the final analytical approach (see Part II Chapter 2.3 “Design of experiment”) should have seen and thoroughly inspected the samples himself before. Please note that touching, wiping, rinsing or other influences on the samples must be avoided during this first inspection, because it can ruin the target area for the analysis.

2.2.2 Microscopic inspection For the first inspection it is very helpful to use a simple mobile microscope (see Figure II.5). These low-cost computer microscopes are easy to transport, easy to use and often deliver surprisingly good pictures that help to improve the sample documentation and support the design of experiment for the next more sophisticated analytical steps. Especially, it is essential when it comes to field investigations at a costumer site or in a production plant. The main task for this inspection

29

Part II Coating failure analysis is to distinguish easy-to-solve issues from more complex problems. If (for example) a paint chip is sampled from a steel object (machine, steel structure) and the backside of this paint chip exhibits microscopic traces of steel dust or if there are suspicious features of the surface underneath the detached paint like grinding or scratching marks that might hint at a sample pretreatment or drying residues that can point at insufficient cleaning. This information can narrow the scope of the analyses which have to be planned. A paint failure in a multilayer coating system can appear as a bubble of the clear coat upon first microscopic inspection but, in fact, is caused by voids or cracks of the base-material. This first tool can help to gather more information about the failure but should not be overestimated on the other hand. If e.g. a paint chip disbonds from a clear primer layer because the primer has not been cured sufficiently, the backside of the paint under first inspection with a simple microscope, might exhibit only a very thin layer of the defective primer which is not visible on microscopic inspection. So, this might lead to the wrong decision concerning the path the analysis has to go. Only a surface infrared spectrum of the backside of the paint chip can reveal the true cause of the failure for this specific example. The first microscopic inspection does not replace a thorough laboratory investigation but can help to avoid wrong approaches to the issue.

2.3 Informed guess Once the initial data of the first inspection and the process parameters have been collected, there should be an informed guess about possible root causes. An informed (or educated guess) must be distinguished from wild speculations which are the result of missing knowledge. The informed guess together with information from interviews directly lead to the design of experiment (DOE) and “prior planning prevents poor performance”. In fact, the design of experiment (DOE) is a key feature of a successful problem solution. If the DOE is poor, the results lead to the wrong direction. The DOE exactly defines the type and amount of analyses to be performed and the expected results. So, at the beginning of a DOE there is the educated guess or let us say the theory what could have happened. To evaluate this question, it is good to have background knowledge about paint technology, paint chemistry and an overview about the analytical methods and their limitations. This limits the group of people that are meant to do the DOE due to their education. Planning an error analysis starts with the question: How do I start? Unfortunately, there is no standard procedure in the sense of: Figure II.5: Microscopic inspection by a computer microscope

30

The analytical procedure – error type A = method A -> solution – error type B = method B -> solution Compare it to the hiking trip onto a mountain. You would want to take the shortest path with less obstacles and hopefully no deviations. At the beginning of the trail a few paths lie in front of you. But how to decide which path is the most promising? When it comes to failure analysis, it is good to keep in mind that coatings fail because something went wrong. You may want to start with an Ishikawa diagram [1] or a FMEA [2] to sort the influences on the coating quality to visualize the possible weak spots in the process. The author himself is not convinced that this approach is the best way to deal with the issue, because these methods operate with probabilities and it has been shown very often that the most improbable root cause has been the key issue of a failure event. Therefore, it is not good advice to rule out one cause just because there is low probability for it. The experience shows that one should beware of the mental mistake of believing exclusively in monocausal connections according to the pattern “if A, then B”. In reality, it is often the coincidence of several factors that influence each other and only in a certain combination lead to a coating defect. For example: The author was called to help in a coating crater issue that appeared erratically with a certain type of polymer parts whereas other parts coated on the same line have not been affected. Following the most probable assumption based on these observations the polymer quality, and the quality of the paint very soon came into the focus. It was argued that there are parts which can be coated without or with a minimum of failures on the same line which seemed to rule out the paint shop as a possible source. After some strategic sampling and using the box-in-box approach (see Part III) it finally has been shown, that the real root cause was a wrong design of the paint shop facilities with respect to the compressed air which was used for spraying the paint: For financial reasons the design of the pressured air supply was limited to a certain air flow volume. But this was not the main problem. In addition to that the contractor, who built the pressured air supply, did not install an oil-free compressor (which is highly recommended for spraying air supply). However, this also did not cause the crater issue, because there were oil traps and filters in the compressed air pipe. The trouble started because the plant operator accepted a contract to paint some parts that demanded a pressured air flow that exceeded the capacity of the supply. The result was that once these parts were to be coated, the compressed air loaded with oil dust flooded the oil traps and filters and finally the oil arrived in the paint booth and was sprayed together with paint onto the parts. So, it was not the paint material, not the cleanliness of the polymer parts, it was not simply the insufficient design of the painting facilities, and not only the failure of the contractor with respect to the not-oil-free compressor. With the “right” parts, the failure of the plant design and the mistake of the contractor would have never been obvious. The painting issue only occurred when the “wrong” parts had to be coated with “too much” air flow and the filters have not been checked and changed for a long time so that they were overloaded. This example shows that the visible failure sometimes has a root cause which is a sum of complex interactions of small hidden anomalies which cannot be described by probability approach. In fact, when it comes to design an experiment/analysis nothing can be ruled out at the beginning of the process. To the authors experience looking over 25 years in failure analysis the “human factor” has always to be drawn into account. If, like the above mentioned case shows, the person planning a compressed air supply for the painting plant tenders the design of the performance of this plant only according to monetary aspects without considering or knowing the technical requirements, then it is possible that compressor oil is continuously or erratically pressed

31

Part II Coating failure analysis through the lines, which then cannot be retained by any filter. As a result, paint failures can occur at irregular intervals and completely erratically, when the next human error (accepting a contract that exceed the parameters of the plant) drives the system to its limits. But back to the planning of the investigations. When a failure disaster such as a crater problem arises, one is faced with many of possible causes. And as always in life there are simple and easy to understand but unfortunately wrong solutions. Very often, supposed connections are then made from observations such as “since we have been using the new batch, craters have been appearing! So, the cause lies in the quality of the paint batch”. As described above, it is basically correct to include such observations in the considerations. However often it turns out that, for example, someone has neglected => that the new batch was also painted on a completely different coating line, under different conditions, – that the components that are painted with the new batch are different from those painted with the “old” batch, – that a new painting robot was installed or similar parameters. Therefore, it is recommended while collecting facts and documenting them, only sufficiently valid information is considered. The next chapters provide information on the options of doing this and how to use them. In short terms, it is about how to go from a well-founded assumption to a certain knowledge and avoid confusing the one with the other.

2.4 Instrument selection The following step in the analytical process is to decide which technical approach must be chosen to prove or disprove the theory. In fact, this step needs a little bit of education because you must have the background which method delivers which results and you must know the limits. Since one single method cannot cover everything, the choice of a method will result in “losing” certain information, either consciously or unconsciously. Therefore, the question: “Was the right measurement method used to answer the question at hand?” plays a very important role. But what is “right” or “wrong”? It is clear to everyone that a hammer is the “right” tool for hammering a nail, for example. Of course, with some practice one can also hammer nails into the wall with a pipe wrench. But the fact that one masters this with virtuosity after a few years of practice still does not make the pipe wrench the “right” tool for this task. Sometimes the aspect of availability of the method plays a bigger role than the suitability of the method. According to the slogan “If you have a hammer, every problem looks like a nail”, questions are treated with methods that offer only a limited suitability for this purpose. For example, it is technically possible to examine a surface for substances that interfere with paint wetting using scanning electron microscopy (SEM), but this does not make sense because the necessary detection sensitivity is not available and organic substances cannot be identified. This means that even if one has a SEM available and handle it with virtuosity, it is not the appropriate method for the problem. The author read a few reports in the past saying, “PWIS (paint wetting impairment substances) were not detected by scanning electron microscopy". This is objectively correct, because PWIS cannot be detected with the SEM method. However, the value of this finding is of the order of “At night it is colder than outside!”. So, to illustrate what I mean, let us look at a paint crater issue. The approach “here we have a crater! Please analyse it.” will not lead to any satisfying results. You must know what can cause a paint cratering and based on this knowledge, you would want to put up the right questions and chose the appropriate methods to investigate the theory.

32

The analytical procedure Table II.1:  Possible applications of analytical methods for paint defects Failure type Possible causes Analytical tools Specks

Adhesion deficiency

Wetting failure

Paint crater, pin-hole

Bubbles and blisters

Stains and precipitations on coating surface

Inclusion of foreign particles

SEM/EDS IR-microscopy

wrong mixing ratio of binder and hardener component

IR (ATR, transmission)

separating agents

TOF-SIMS

oil or grease residues

TOF-SIMS

oil or grease residues

TOF-SIMS

gel particles

SEM/EDS IR-microscopy micro-ATR

oil or grease residues

TOF-SIMS

oily aerosols

TOF-SIMS

outgassing of the substrate material caused by cracks or cavities

metallographic cross section + optical light microscopy and/or SEM

salts

SEM/EDS

water

-

application failure

-

migration of additives to the surface

ATR-FT-IR

external contaminations

TOF-SIMS

For the crater issue e.g., this means: – Are there lacquer wetting disturbing substances (PWIS) in the crater? – Is the enrichment of a filler or a hardener detectable? – Are foreign aerosols detectable in the crater? These questions usually define the type and number of possible investigation procedures. The initial analyses usually serve to confirm or disprove a theory developed based on data from the preliminary assessment and the collection of facts about the cause of the fault. Referring to the above-mentioned example, of course, you want to have the answer to the question “what is the cause?” But this question is not answered by an analysis. Rather, the question must be, for example: “Are there adhesion-disturbing substances in the interface, where the delamination occurs and if so, which ones?” This is a question that can be answered unambiguously with instrumental analysis. Which shows, that a certain amount of prior knowledge about what (theoretically) can cause adhesion problems and what to look for is inevitable. Once the exact question has been formulated, the selection of the appropriate measuring methods is usually based on the desired parameters like e.g. sensitivity, sample size, size of the area to be analysed etc. If, as outlined above, the possible substances that might interfere with adhesion are questioned and consider that even very small quantities of certain substances cause an adhesion deficiency, then the prerequisites for a suitable measuring method are: – very high detection sensitivity and – a high surface sensitivity and – the possibility to deliver molecular information.

33

Part II Coating failure analysis This means that scanning electron microscopy, for example, is completely unsuitable for this type of problem, because, on the one hand, it is not surface sensitive enough and, on the other hand, it only provides elemental information. So, if e.g. silicon is detected on a surface with scanning electron microscopy, it is not possible to decide, whether it is adhesion inhibiting polydimethylsiloxane or non-adhesion inhibiting silicon dioxide.

2.5 Investigation Analytical data acquisition has changed a great deal over the last decades due to progress in computer performance and innovations. The lateral resolution of a SEM, for example, has improved drastically. Sharp pictures without distortion on the nm scale are now available. The main aim of the instrument suppliers is to make data acquisition more convenient and faster. An “easy to use” desktop interface based on standard computer desktop features makes the work much easier and makes people believe that even inexperienced employees are able to perform an analysis. That is wishful thinking. Through modern instrumentation the work seems to be easier, but it also makes it easier to produce wrong results. Besides the skills needed for operating the instrument, a deep understanding of the processes happening during the analytical process is needed to separate artefacts from data. At the very least, the data should be reviewed by an experienced analyst. Another aspect is time. A good analysis needs a certain time frame. To improve the results and get the best information, it is sometimes necessary to repeat the measurements a few times. As to the saying “Grass won’t grow faster if you pull it!”, samples deserve to be treated thoroughly so that the best results will be obtained. In contrast to the tendency to do everything faster, efficiency is not counted in minutes or hours, when it comes to analytical data acquisition. Further methodical details of the measurements will be presented in Part IV of this book.

2.6 Interpretation The evaluation (data oriented) and the assessment (process oriented) of the analytical results are two steps that must be strictly separated. Both steps can and should only be carried out by experts. The data evaluation includes: – the purely technical processing of the data – checking the calibration – checking for plausibility and measurement errors – the comparison with databases, – the correlation of the measured data with tables Whether, for example, an infrared spectrum is “reasonable” or whether a beautiful but completely meaningless spectrum has been produced by preparative errors or previously unknown material properties, can only be decided by someone who understands the technology and has the theoretical materials background. In addition, the comparison with databases (uncritically and automatically used) bears considerable risks and imponderables. One reason for this is, that database comparisons are purely mathematical routines. To put it pointedly, the computer compares zeros and ones. The result is a selection of database spectra with a hit quality index, which reflects the mathematical probability that the measured spectrum matches a stored spectrum. There is no plausibility check. This often leads to very bizarre results. If you rely on this probability indices list without checking them

34

The analytical procedure Table II.2:  TD GC-MS and TOF-SIMS analyses of the same activated carbon filter from a paint shop Laboratory A Laboratory B Sample

carbon filter granules used and loaded

Analytical method thermodesorption GC-MS

extraction + TOF-SIMS

Detected substances

palmitic acid ester, stearic acid ester, tetraethylene glycol dicaprylate, polydimethylsiloxane

cyclohexane, pentane, 1,2,2-trichloro-1, 22-trifluoroethane, 2-methylbutene, ethyl acetate, methylcyclopentane, methylcyclohexane, xylene

and build a theory about the course of the damage and how to remedy it, it is not unlikely that you will not find the cause in a harmless case or, in the worst case, that you will cause great damage by taking the wrong corrective actions. Furthermore, the algorithms only compare the measured spectrum with stored data. However, databases usually become obsolete faster than the seasons change. If, for example, a binder is entered in the database under a trade name X, then the manufacturer may be sold to another company, the product name has been changed or the same binder has been sold under a different trade name with a slight modification. Thus, the corresponding database entry is worthless. Thirdly, databases fail when analysing mixtures! A spectrum of a primer with a high filler content is then “identified” as barium sulphate, for example, because the signals of the barium sulphate stand out prominently in the spectrum. As a result all other paint components are no longer recognized by the database. The evaluation of the measured data is followed by the assessment. The assessment should be carried out in several stages and is process oriented. Finally, it has to answer the core question of whether all outstanding questions have been answered by the analyses carried out, whether correctives actions can be derived from them or if further investigations are necessary: Every analytical method used to test a sample for its composition has certain so-called “blind” spots. This means that each method (due to the physical limitations of the methods) will deliver only an excerpt from reality. No analytical method can determine exactly the complete, comprehensive and “true” composition of a sample. To achieve a maximum of information a concert of different methods is needed. Analysing a sample by SEM-EDS, for example, leads to the detection of the elements in the uppermost μm of the tested area. But it does not tell anything about the molecular structure. Detecting the elements carbon, oxygen and calcium by SEM-EDS e.g. leaves the molecular options of calcium carbonate as well as a calcium soap of an organic acid. To distinguish between these possible structures infrared spectroscopy is needed. This means, when selecting the analysis methods which is used to get closer to the real composition of a sample, it is good to know these “blind spots” (i.e. the information that cannot be detected) and weigh up to what extent the “omission” of this information represents a relevant statement or not. The more preparation steps, analysis steps and evaluation steps are applied from the initial sample to the test result, the more information is lost. It is good to be aware of this fact when evaluating all analysis results (from whatever laboratory). And another fact follows from this insight: Analysis results that are not carried out on the same samples, with the same preparation methods and the same analysis procedures are not comparable in every sense. The following gives an example to illustrate this: Paint craters occurred on a paint shop after the production output was increased. Initial tests and random analyses indicated that the paint air (i.e. compressed air) could possibly be contaminated

35

Part II Coating failure analysis with paint wetting interfering substances (PWIS). The best way to check this is to take a sample and analyse the activated carbon filters used in the compressed air line. For this reason, two independent laboratories examined the same activated carbon filter sample of the compressed air supply. The aim of these tests was to obtain information about whether paint wetting interfering substances (PWIS) are present in the compressed air. The results of both analyses are listed in Table II.2, and seem to be unequal and contradictory. How can this happen? Did one or even both laboratories work inaccurately? The answer is “No!” Rather, the method selection limits the results to one aspect of the real composition. One of the laboratories used the TOF-SIMS method for the analysis of the extracts from the activated carbons, because it has an extremely high sensitivity to PWIS (paint wetting impairment substances). It was deliberately accepted that volatile compounds are not detected by the TOFSIMS method. The second laboratory, however, used a combination of thermo-desorption and GC-MS to examine the activated carbon. Some PWIS cannot be detected by thermo-desorption, because they are simply not thermally desorbable. After thermo-desorption/extraction (which are only separation methods but no analysis methods), this laboratory carried out an analysis and identification using GC-MS. It must be taken into account that GC stands for gas chromatography which is another separation method. By this separation method, the mixture of substances absorbed by the activated carbon is separated on a so-called column into individual components or groups of components (see Part IV, Chapter 5.12.2). Here again, there is discrimination in the sense that not all substances that are injected at the beginning of the column necessarily leave the column at the end. Depending on the choice of column and the temperature program during the chromatographic separation of a mixture, very different results can be obtained. The GC-MS analysis selected for the second laboratory ends with a mass spectrometric analysis of the substances separated from the GC column. Here the result of the identification depends very much on how it was carried out. In many cases, the mass spectra obtained are automatically compared by the computer according to purely mathematical criteria for similarity with stored spectra from a database and this degree of mathematical agreement is output in a so-called “hit quality index”. To our experience, however, a purely mathematical match of 80 % may occur, but the substance “found” above this level may be complete nonsense in the context of the investigation. This means that the quality of the identification depends on whether the substance being searched for is actually present in the database. However, many PWIS are not included in commercially available databases. Based on the originally formulated question, which substances are present in the compressed air, the available results must be assessed as follows: – It can be assumed that the detected substances were present in the compressed air (provided the activated carbon did not already contain these substances at the time of installation). – If a substance was not detected, this does not mean that the compressed air is free of it! The following must also be considered: A conclusion from the composition of the substances adsorbed on the activated carbon to the complete composition of the compressed air can only be drawn if – all substances present in the compressed air are adsorbed on the activated carbon without discrimination, – all adsorbed substances are completely and non-discriminatingly transferred to the analysis, – the chosen analytical method detects all substances without discrimination.

36

The analytical procedure However, precisely these requirements are not met in the present case. This example shows which pitfalls lurk in a supposedly simple analytical task. Of course, this can be applied to all methods. To ensure that the results can be interpreted and implemented in the process, the analytically obtained facts must be combined with knowledge. This intelligence can be the knowledge of internal specialists (process knowledge, expert knowledge) as well as the expertise of specialists outside the company. It is important to carefully distinguish between assured knowledge and assumptions, and it is no good advice to leave the knowledge and intuition of the employees on site at the production lines unused. A qualified brain should generate a quintessence from the information from all these sources, this involves answering the following questions: – What is the desired objective? – Which of the substances found in the defect originates from the process and which is to be classified as an impurity? – Which substances are used in which process step? – Can the detected substances explain the origin of the defect plausibly and without contradiction? – Are there any gaps in knowledge that need to be filled by additional analysis of further samples? If all measurement data and process data can be combined consistently without contradictions, then the solution is not far away. The last step is to check the knowledge gained in practice and to define remedial measures. In the case of surface and material defects, the instrumental analysis of samples is a necessary, helpful but not sufficient method to get to the bottom of the cause of the defects. Even though it is associated with costs, it is always cheaper than shirt-sleeved experiments based on speculation. An effective problem solution can only be generated when internal expertise, external knowledge and measurement data are brought together to form a synthesis.

2.7 Iterance When assessing the results of all analyses carried out, it is necessary to check whether there are contradictions in the statements of the analyses, or if the analytical question has been answered. The comparison of the assessment of the results with the information provided serves to determine whether, for example, the type of coating specified by the client has been verified, or whether a different coating was used, whether the sequence of layers corresponds to the specifications, etc. The comparison of the data evaluation with the failure hypotheses is the step that leads to the actual problem solution. In case of ambiguities it will guide to the definition of a new investigation plan and, if necessary, to the re-evaluation of the damage hypothesis. After developing a good design of experiment, you would expect to perform the projected analysis and find a result that helps to solve the issue. But this is not always how it works. Quite often the first analyses tell that the informed guess was wrong or at least part of the truth. Scientific methods are sometimes frustrating for people which expect a quick, undoubtable and clear answer after doing one analysis. In fact, often the result of the first so called “screening” analysis answers one question and opens the door to several new questions. This is neither the fault of the one who did the analysis nor is it a proof that the analysis is senseless. It is very often the result of our (understandable) desire to have simple and easy answers to complex questions. We do not like complexity, and we do not always see it. Sometimes we find ourselves after the first analyses in a situation, that we do not understand, what we have measured, because we are stuck in the belief that there is one reason for one failure. But the equation: A + B = C ….

37

Part II Coating failure analysis meaning there is a sample A that has suffered an influence B and the result is C is sometimes too simple. We must admit that we neglected influence D and E to get the result C. This challenge requires again the intelligence which means the witness of experienced people. After evaluating their knowledge, the iteration of the analytical process might be necessary. This means with the first results (that did not lead to the quick answer) and the extra “intelligence” we make a new refined (better) informed guess and prove the assumptions by going back to the analytical process and perform additional analyses.

2.8 Implementation Despite all the potential that instrumental analysis methods have for supporting production processes, it is often not possible to implement this concentrated knowledge in the process. This leads to the frustration of the person commissioning the analyses and to the understandable reaction to reject these procedures. But what is the reason for this? The essential factor that contributes to the success of a problem solution with the help of analytical procedures is the flow of information between laboratory and client on the one hand and within the structures of the client on the other. An example: A coating did not pass a climate test, because it developed a deposit on the surface (Figure II.6). The paint shop manager orders a laboratory to investigate this phenomenon. To be more precise, the laboratory is commissioned to investigate the composition of the deposit. Through analysis with TOF-SIMS and the ATR-FT-IR method (Figure II.7), the laboratory determines the composition of the deposit and communicates: The coating consists of 2-[4-[2-hydroxy-3-tridecyloxypropyl]oxy]-2-hydroxyphenyl-4,6-bis(2,4-dimethylphenyl)-1,3,5-triazine. At this point, the non-chemically trained client might put a large question mark: “How does this help me now?” This result can only help, if there is a suitable translator who can put the chemical formula, which is the result of the analysis, into a context with the production in order to answer the actual questions: – What happened in the climatic test? – Why did it happen? – What do we learn from this for prevention? It is therefore necessary at this point to have the knowledge of in-house experts or of the experts in the commissioned laboratory, who can translate the result into knowledge of the products used in the company. This translator has a decisive role to play and at the same time this task requires many complex processes. First of all, this includes knowing what technical substance, raw material etc. is involved. In this specific case, the knowledge that the identified compound is a UV stabilizer with the trade name “Tinuvin” 400. The next step is to check with the coating development laboratory whether this UV stabilizer is actually a formulation component of the coating. Figure II.6: Optical light microscope image of a coating surface with a deposit This requires the knowledge of the

38

The analytical procedure coating developer or, if this is not available or not accessible, a further laboratory analysis on a reference sample of the coating material. If the result of this research or subsequent analysis is: yes, “Tinuvin” 400 is a formulation component, then it is already certain: The climatic test enriches the coating additive on the surface. So, once it is clear what happened, the search for possible causes can begin. Here, the knowledge of the paint specialist or developer is needed to check overdosing, incorrect choice of stabilizer, interaction with other additives, etc. If the evaluation of the research shows: no, “Tinuvin” 400 is not a component of the formulation, however, completely different questions arise regarding errors in the paint production, cross-contamination due to faulty test procedures, etc. This means that the original analysis result, the chemical identification of the substance on the surface, requires a number of evaluation processes, for which well-trained specialists are needed, who have knowledge of the processes and the materials used and can manage the transfer from the analysis result to process improvement. This means that the technical processes presented in the previous chapters can help to better understand a problem and to uncover errors or false assumptions.

2.9 Incumbency For a good outcome of an investigation of production failures it is recommended to have one person in charge (or one fixed team of experts) that deal with the issue. Let us call it a task force (even if the task force is only one person) and all results should be reported to this task force manager (TFM). Why is this so important? Because it saves money and time! If a bunch of people work on the issue without a coordinator gathering and evaluating the information about results of practical tests and/or analytical methods important results may be suppressed and/or wrong results or test outcomes maybe overestimated. This shows that the TFM has a great responsibility and needs certain competences: – The competence to evaluate data and practical production tests as well as statistical impacts. – The competence to design experiments, tests and analytical procedures that lead to effective results. – The support by the management. – Sufficient time and manpower to do his job. This means for a limited time of the project it should be his absolute priority. – Access to all areas which are releFigure II.7: Results of the analysis of the deposit by a) ATR-FT-IR and b) TOF-SIMS vant for the issue.

a)

b)

39

Part II Coating failure analysis Last but not least, the TFM should be encouraged to adopt the role of an inspector having a critical view on each process and production step. The sampling plays an important role in the whole process. Therefore, the role of the sampling process will be discussed in the following chapter.

3 The power of sampling 3.1 The role of sampling in the analytical procedure The importance of the sampling process on the analytical result on the one hand side and the measures derived from the evaluation of the analytical data on the other hand is too often neglected or underestimated. Analytical techniques are much more precise than sampling techniques ever can be. Looking at the possibility to fail during this analytical process the interval for systematic failures during sampling is about 103 times higher (rough estimate) than the error limits of the measurement itself (see Figure II.2). Which means even the most precise analytical technique will not lead to the solution when the sampling is insufficient [3]. If for example a coating failure occurs and the task is to investigate the root causes, it is not a good idea to grab one, not very well defined sample out of a pile of scrap and analyse it without any plan. The sampling must be representative, well designed and perfectly documented. The question of what is representative cannot be answered by production staff. Therefore, well-educated employees have to do that job and take care for a good documentation which includes: – sampling location – sampling date and time – material type and lot – environmental parameters – sampling intention

3.2 Random sampling and representativeness No matter if a production failure or raw material consignment has to be tested, the analysis of the whole population is not an option. Therefore, a random sampling is necessary which guarantees that the sample represents the whole population with respect to the questioned parameters. This is called a representative sampling. But what is a representative sample and how to accomplish it? How big is big enough? In fact, there are two answers to these questions, a statistical and a practical answer. An example will be good to discuss the different aspects.

Example 1: Bernoulli experiment

A paint shop provides high quality decorative coatings for automotive interior parts. After finishing thousands of coated parts, it was found that 1 in 500 parts exhibits a paint defect. It is assumed that poor raw part quality might be the reason for the coating failure. Therefore, a sampling of the next lot of raw parts has been planned to find out if the lot (which consists of thousands of raw parts) contains defective parts. The question is: Which sample size is needed to make sure that one or more defective parts can be found with a confidence level (or confidence interval) of 98.2 %?

40

Part II Coating failure analysis Last but not least, the TFM should be encouraged to adopt the role of an inspector having a critical view on each process and production step. The sampling plays an important role in the whole process. Therefore, the role of the sampling process will be discussed in the following chapter.

3 The power of sampling 3.1 The role of sampling in the analytical procedure The importance of the sampling process on the analytical result on the one hand side and the measures derived from the evaluation of the analytical data on the other hand is too often neglected or underestimated. Analytical techniques are much more precise than sampling techniques ever can be. Looking at the possibility to fail during this analytical process the interval for systematic failures during sampling is about 103 times higher (rough estimate) than the error limits of the measurement itself (see Figure II.2). Which means even the most precise analytical technique will not lead to the solution when the sampling is insufficient [3]. If for example a coating failure occurs and the task is to investigate the root causes, it is not a good idea to grab one, not very well defined sample out of a pile of scrap and analyse it without any plan. The sampling must be representative, well designed and perfectly documented. The question of what is representative cannot be answered by production staff. Therefore, well-educated employees have to do that job and take care for a good documentation which includes: – sampling location – sampling date and time – material type and lot – environmental parameters – sampling intention

3.2 Random sampling and representativeness No matter if a production failure or raw material consignment has to be tested, the analysis of the whole population is not an option. Therefore, a random sampling is necessary which guarantees that the sample represents the whole population with respect to the questioned parameters. This is called a representative sampling. But what is a representative sample and how to accomplish it? How big is big enough? In fact, there are two answers to these questions, a statistical and a practical answer. An example will be good to discuss the different aspects.

Example 1: Bernoulli experiment

A paint shop provides high quality decorative coatings for automotive interior parts. After finishing thousands of coated parts, it was found that 1 in 500 parts exhibits a paint defect. It is assumed that poor raw part quality might be the reason for the coating failure. Therefore, a sampling of the next lot of raw parts has been planned to find out if the lot (which consists of thousands of raw parts) contains defective parts. The question is: Which sample size is needed to make sure that one or more defective parts can be found with a confidence level (or confidence interval) of 98.2 %?

40

The power of sampling The confidence interval defines the Table II.3:  Approximation for different confidence levels [7] range which includes the true parame- by the “rule of three σ Confidence level Worst case risk ter of an experiment that is repeated infinitely, within a certain probability 86.5 % 2/n (here 98.2 %). For this example, it 95 % 3/n means that the sample size has to have 98.2 % 4/n a minimum size which guarantees with 99.3 % 5/n a probability of 98.2 % that at least one or more defective samples will be found. Well, there is no intuitive way to find out the necessary sample size. This question has to be looked at by statistic methods. Generally speaking, this is a random experiment which allows for two results (defective and not defective) which is also called a Bernoulli experiment. A Bernoulli experiment which is repeated n-times is called a Bernoulli chain [4] [5]. We are looking for the number n of independent trials with a pass or fail outcome of each trial (which means defective part or non-defective part) which are necessary to pull out at least one defective part with a confidence level of 98.2 %. If the result of this Bernoulli experiment follows a binomial distribution and the trials (single sampling) do not alter the population (the total number of parts to be tested) the following notation applies:

X = number of successes of the binomial experiment n = number of trials p = the probability of success on an individual trial P = the binomial probability that an n-trial binomial experiment results in exactly X successes, when the probability of success on an individual trial is p

Suppose that, for this experiment the probability of success on an individual trial p is 1 in 500 parts = 0.2 % = 0.002 The desired probability P for “at least one defective part” is: 0.982 or 98.2 %

P​(X  ≥  1)​  ≥  0.982

in other “words”

1 − P​(X  =  0)​  ≥  0.982​ or



− P​(X  =  0)​  ≥  − 0.018​



P​(X  =  0)​  ≤  0.018

With the formula for binomial distributions: n P​(X  =  k)​  = ​ ​  ​​  ​* ​p​​  k​  * ​​(1 − p)​​​  n−k​ k this leads to n ​​ ​  ​ ​ ​* ​0.002​​  0​  * ​​(1 − 0.002)​​​  n−0​  ≤  0.018​ 0 ​​​(0.998)​​​  n​  ≤  0.018​

()

()

​​​ln ​(​​​(0.998)​​​  n​​)​​  ≤  ln ​(​​0.018​)​​​​

n * ln0.998  ≤  0.018​

41

Part II Coating failure analysis ln ​(​​0.018​)​​

​n  ≥ ​ _  ​​  ln ​(​​0.998​)​​ ​n  ≥  2006.68​ Fortunately this calculation can be approximated by a more “handy” equation [6] shown in Table II.3. For a confidence level of 98.2 % the worst-case risk following the “rule of three σ" [6] can be approximated by 4  ​p  =  ​ _ n

which means solving for 4  ​​ ​n  =  ​ _ p

That n =​​_   4  ​  ≥ ​2000 0.002 So at least 2000 samples have to be randomly inspected to ensure with a confidence level of 98.2 % that at least one defective part is captured. Taking into account that a paint defect (like e.g. a crater, a bubble or a speck) is a local effect which does not necessarily affect the whole part, this mathematical result means that 2000 parts would have to be tested examining the whole surface (because in most cases you cannot figure out where the defect area of a raw part is by visual inspection). Here is the practical (but discouraging) answer: This is not possible! A testing of this size is not reasonable! It has to be mentioned that the above mentioned formula is only valid for randomly selected samples with defects randomly distributed through the population under investigation.

Example 2: Z  ero defect sampling

Figure II.8:  Tank transporter delivering water-based polyester binder

42

How poor can quality be consistent with the finding that a random sample contains zero defects? The first example calculated above implies the knowledge of a defect rate. But often you do not know it. But you want to be sure that your production lot has zero defects without testing the whole population. Therefore, the question is: If you test a certain number of samples on a random basis, how sure can you be that the whole population is ok? Let us assume that 100 samples are selected randomly from a production line and zero defects have been found.

The power of sampling So, n is 100. Following the rule of 3 σ the upper risk level is 3/100 or 0.03 with a confidence level of 95 %. This means that the tested lot can contain up to 3 % defects even if zero defects have been found in a random sample of 100! Note: If nothing goes wrong, it is not necessarily all right!

Example 3: R  andom sampling of binders A water-based polyester paint binder is delivered to a paint manufacturing company by a tank truck. During the unloading process into a storage tank of the manufacturing company representative sample have to be gathered for quality testing. This task cannot be calculated by stochastic methods. Being representative here means taking samples from different areas of the tank from the top, from the bottom and at the outlet pipe. A few parameters have to be realized if sampling a delivery of this type: A water-based binder is a suspension. This suspension can settle into layers of different density inside the tank. During transport over the road it is reasonable to assume that this suspension is well be agitated in the tank by the movement of the vehicle and thus should be homogeneous upon arrival. Therefore, a sampling shortly after arrival of the truck is recommended. Depending on the outside temperature some binder can dry at the top of Figure II.9:  Sampling a water-based polyester binder the tank inside walls. A dried binder consignment at the outlet pipe of a tank transporter suspension is like a rubber material and cannot be homogenized into the suspension. It may fall into the suspension without dissolving and “contaminate” the load (see Figure II.10). This dried raw material can cause hidden failures during paint production ending up in paint specks once the coating is applied. So, you would want to find these rubber pieces during quality testing. Therefore, different levels inside the tank have to be sampled for each delivery. This can be achieved by a dipping vessel which is lowered into the tank down to the desired levels (which have to be measured by a weighted Figure II.10:  Dried film of a water-based polyester binder tape measure!). Once the dipping vessel at the outlet pipe of a tank transporter

43

Part II Coating failure analysis reaches the targeted depth it is opened until it is filled and then hauled up. Three different levels have to be tested at least: the bottom of the tank, the top of the suspension load and one level in between. This is the minimum, but more levels would be desirable. The subsequently drawn sample have to be mixed again to be collected in one aggregate sample which is finally used for the desired testing procedures. Sample numbers depend on the local circumstances: If the tank truck has only one hatch the three samples from the top, an intermediate layer and the bottom must be carried out from that access and from the outlet pipe. If there are more hatches samples should be drawn from all the hatches and the outlet pipe. Sampling over the whole-time scale of the unloading process is recommended. Why? See Figure II.10. This dried, rubber like polyester binder material came out at the very end of the sampling process. A single sampling at the beginning or a double sampling at the beginning and during the unloading would have missed this. The amount of the final aggregate sample is determined by the testing procedures and analytical methods. Another technique to achieve a random sample is stratified sampling. This means the population (e.g. a delivery of a filler) is split into several portions. After that aliquots of each subdivision are randomly drawn and combined to one aggregate sample. If you look at Figure II.11 each tray of ten bezels represents a subdivision of the whole lot. A stratified sample is achieved by e.g. taking out one bezel randomly from each tray. Cluster sampling is a slight variation of this method. This means again dividing the whole population which has to be tested into equal subdivisions. But in contrast to stratified sampling all bezels of a certain amount of randomly selected trays are saved for the analysis. At last representativeness can also be achieved by selecting every nth bezel just by counting (for example every fifth bezel counting from the top of the stack of trays in Figure II.11 to the bottom). This procedure is called systematic random sampling.

Figure II.11: A rack of coated automotive interior parts

44

The power of sampling

3.3 Targeted sampling The previously discussed random sampling serves as a tool to find out if a certain test portion of a population exhibits a feature or failure and quantify the risk of the appearance of certain features. The targeted sampling is a method which is used if a feature is already known and/or visible and portions offering this feature are specifically selected. If you look e.g. at a population of coated polymer parts like shown in Figure II.11 and the task is to find out if there is certain risk that this population may contain an (unknown) percentage of failed parts with poor paint adhesion (that you do not see without testing) you would choose for a random sampling of parts that will be tested by a cross-cut test to find out if the population is o.k. taking account of a certain confidence level (see Example: Zero defect sampling in Chapter 3.2. Part II). If you see the defects (coating peels off or exhibits bubbles) you pick out the failed parts by a targeted sampling and explicitly investigate these failures.

3.4 Non-destructive sampling 3.4.1 Wipe sampling Sometimes the object that has to be analysed cannot be transferred to a laboratory or is too big to be analysed without destroying it. Examples for that are e.g. hazes and discolouration of yacht coatings or aircrafts (see Chapter 3.8 “Field analysis” in Part III) or coating/surface failures of big buildings. Machine parts, transport devices or storage containers are further examples that challenge the way sampling can be performed. If a non-destructive sample of a failure area is needed, wipe sampling is a choice. The material needed for this kind of sampling are: – clean laboratory grade tissue – appropriate analytical grade solvents – clean laboratory grade glass container – laboratory grade disposable gloves

Wiping procedure (wet wiping) In order to achieve a maximum representativeness, the targeted surface is wiped crosswise using a clean laboratory grade tissue soaked with an appropriate solvent of medium polarity. Analytical grade ethanol or isopropanol is suitable for the majority of sampling tasks. The tissue with the sample is then placed into a clean laboratory grade glass container and labelled immediately with an unambiguous sample number. Although ethanol or propanol is suitable in most cases the correct choice of the appropriate solvent has to be adjusted to the sampling task. The solvent has to meet the following requirements:

Figure II.12: Wipe sampling a high-quality exterior coat of a ship

45

Part II Coating failure analysis 1. Inert with respect to the wiped surface (You would not want to dissolve the coating of a ship or plane during sampling!) 2. Clean with respect to a minimum of drying residues. 3. Most effective dissolving the targeted substances.

Wiping procedure (dry wiping):

Figure II.13 shows a wipe sample taken from a conveyor belt of an injection moulding machine. The background was a paint wetting failure of automotive body components. It had been assumed that after injection moulding the raw polymer parts may have been contaminated by substances that are transferred from the conveyor belt between injection moulding machine and transport container. Therefore, the DOE was to investigate the mobile substances that are present on the surface of the belt and check if they can be responsible for the paint defect. Of course, cutting out a section of the belt and transfering it to the laboratory is not a realistic choice. So dry wiping has been chosen to sample the belt surface. This procedure needs analytical grade cellulose or paper which absorbs the particles but also greases, separating agents, additives and residues of cleaning procedures. It should be wiped cross wise over the whole belt. It is important that the target area of the sampling device is never touched by the fingers. Directly after Figure II.13: Wipe sample of a conveyor belt sampling the wiping tissue should be stored in a clean laboratory glass vial for transport to the lab.

3.4.2 

Figure II.14: Rinse sampling a high-quality exterior coat of a ship

46

Rinse sampling

Another choice for non-destructive sampling is rinse sampling. This technique can be used for thin soluble films and precipitations on coating surfaces. The required material for this kind of sampling is the same as for wipe sampling. – clean laboratory grade tissue – appropriate analytical grade solvents

The power of sampling – clean laboratory grade glass container or vials – laboratory grade disposable gloves – disposable glass pipettes The procedure is simple. An amount of 1 ml clean solvent is dropped on the target area making use of a pipette to ensure a confined contact of the solvent. It is allowed to run over the sampling area. The solvent (hopefully) picks up what is soluble and after that is collected and transferred to a sample vial which is immediately closed. The collection can be done by absorbing the solvent by an analytical grade clean tissue or paper or just letting it drop into the sample vial As already mentioned, is has been experienced, that isopropanol or ethanol most commonly are suitable liquids for this kind of sampling, but it has to be checked for each individual site which kind of solvent is appropriate. As there is always a certain risk or probability that the testing solvent extracts components of the coating which are not the target of the investigation, it is highly recommended that a blank feed of a sound area has to be drawn. Sampling failures happening at this stage of investigation are severe, because the sampling cannot be repeated. The targeted area is irreversibly changed by the sampling and thus cannot be used for a second sampling. Therefore, the sampling should be done by experienced staff. Once the samples are in the lab, spectra of the targeted area and the sound area have to be collected exactly the same way (standardised process). The resulting spectra combine the information of the targeted (and hopefully rinsed) substance on the coating surface and the soluble

a)

b)

c)

d)

Figure II.15: The principle and impact of (quasi) non-destructive rub sampling of coatings a) High gloss coating before sampling; b) High diffuse reflectivity abrasive sampling paper c) High gloss coating after sampling; d) and after re-polishing of the sampling area

47

Part II Coating failure analysis substances that are inside the paint film and have been extracted to a certain extent. Therefore, the next step of data evaluation is spectral comparison and/or spectral subtraction in order to separate the desired information from the “background” of the coating. Typical sampling failures, which have to be avoided, are: – allowing the solvent to interact with the coating surface longer than the time necessary for dissolving the target substance (which might result in swelling of the paint) – using non-laboratory equipment – drying of the rinsing solvent during sampling – changing the sampling conditions between target area and reference area

3.4.3 Abrasive sampling As already mentioned, sampling of big and precious objects is a very delicate matter. On the one hand you might wish to have as much sample as possible for a comprehensive evaluation of the issue. A destructive sampling is inevitable: – if, for example, the question is how many coatings have been applied, or – the task is to measure the thickness of each layer of a multilayer system consisting of a few layers of spatula und coatings But on the other hand, the owner would not be pleased, if you cut out a 2x2 cm piece of the object. If it comes to solid precipitations, discolorations, hazing and gloss deficiencies, the design of experiment of the analytical approach focusses on the topmost layers of the coating. To meet the demand of sampling and analysis with minimum impact on the product a minimum destructive sampling has been developed. It is an analysis technique that allows for sampling of the uppermost microns of the coating surface and analysing the sample without any further preparation by infrared spectroscopy. The sampling procedure is performed by a special abrasive paper which is slightly rubbed over the area that has to be investigated. A very low amount of material sticks to the paper and the specimen will be transferred to the lab for the spectroscopic analysis. Of course, this manipulation leaves a visible trace. The sampled area appears a little bit matt after the procedure but can easily be polished leaving no trace of the sampling afterwards, because the amount of material taken off the sample surface is very low but sufficient for trace analyses. Of course, this method has to be performed by experienced persons and the impact has to be checked for each individual object. The amount of sampled material on the hand side and on the other Figure II.16: DRIFT spectrum (Kubelka-Munk) of the abrasive hand the influence on the appearsampled coating showing a typical aliphatic polyester ance of the coating depends on the polyurethane spectrum

48

The power of sampling pressure applied during the procedure and the hardness of the coating. This sampling has to be done very carefully. The rub sample will be analysed by infrared reflection spectroscopy (DRIFT) making use of a specially designed objective and/or by mass spectrometry (TOF-SIMS).

3.5 Typical sampling failures What is the main task after a failure occurs, it has been decided to investigate the root cause and the design of experiment is fixed? You would want to have reliable samples in the first place. Reliable samples are those that ideally represent the issue most comprehensively. A sampling error is the deviation of a parameter of the sample versus the whole population. That means, the essential prerequisite is that the samples suffer a minimum of “bad influence” from the beginning of the sampling to the preparation for analysis, which alter the status of the moment when the failure happened. The sample area to be analysed must be saved. So as simple as it is, the selection of “right” samples is a challenge.

3.5.1 Wrong sample collection Figure II.17 shows a coated polymer part and a reference raw part that has been selected for analysis of the surface. The coating on the left sample exhibited adhesion deficiencies (which is not visible in this picture!). The aim of the analyses was to find out, if there are substances on the raw part surface that can have a negative impact on the coating adhesion. However, a close look reveals, that the reference raw part has a different shape. This means, that the raw part does not fit to the coated part. It is inappropriate, because it most likely has been produced in a different moulding tool and thus does not represent the surface condition of the raw parts of the failed bezel. Although, the moulding conditions might seem to be equal, there are always slight differences that may not be obvious. Similar considerations apply when not the raw part, but the coating is the target of the Figure II.17: Coated polymer part and (wrong) analysis. Figure II.18 show a defective bezel raw part selected for analysis

a)

b)

Figure II.18: a) Coated polymer part showing adhesion deficiencies and b) sound part selected for comparative analysis

49

Part II Coating failure analysis which exhibits significant adhesion deficiencies between the black polymer and the grey coating. For this case the coating quality was questioned and a sound reference part which has been coated without failures has been selected for comparative analysis. But the coated bezels in Figure II.18 (failure part with adhesion deficiency and sound part) are obviously not suited for comparative analysis. The coating adhesion visible with the left bezel can be caused e.g. by residues of an injection moulding agent of the polymer. Whether this polymer additive migrates to the surface is a function of the polymer moulding parameters. As the sound part can obviously not have been produced in the same tool or machine, the result of the moulding process must be different. A suitable reference sample for the comparative analysis of the left bezel would have been a sound sample from the same tool machine and lot. With respect to the coating quality there are also doubts, because it is very likely that the reference bezel has not been coated under the same parameters and in the same batch. The following requirements have to be fulfilled to make sure that analysis of a raw (uncoated) reference part surface delivers reliable information with respect to the above mentioned analytical goal: – the same moulding tool – the same raw material lot – the same moulding parameters – the same day – the same environmental conditions If one of the above mentioned prerequisites are not met, the results are of limited profit for the analytical goal.

3.5.2 Non-representative sampling The representativeness of sampling has already been explained in Chapter 3.2, Part II and it has been shown that representative sampling sometimes is difficult to evaluate and sometimes hard to achieve. Figure II.19 shows an aliquot of a coating material sampled from a hobbock. Upon storage this material separates into three layers. So, the challenge is to draw a representative sample from the hobbock which can be achieved most closely by stirring before sampling. Taking a random sample from the top or the bottom without stirring thoroughly will most likely result in non-representative aliquot sampling. In addition to that aliquots of different container areas should be taken and combined to one laboratory sample.

3.5.3

Figure II.19: Aliquot of a coating sample separating into three phases upon storage

50

 electing the wrong S amount of samples

As shown in Chapter 3.2, Part II it is a challenge to decide how many samples are at least necessary out of a population of parts to meet the

The power of sampling requirements of the design of experiment. This is true for targeted sampling as well as for random sampling. The minimum number of random samples for a given probability of success has been calculated in case study 1 of Chapter 3.2, Part II. It sometimes exceeds the number of samples that can be collected under practical considerations. Taking as an example a paint defect “outbrake” of different shapes and distribution on a coated steel component. The first inspection did not verify unambiguously, if the failure is a speck or a paint bubble or wetting defect. Some parts exhibit one spot, some show up to ten spots. At the beginning of the analytical process when the sampling takes place the number of different defect types, the number of defects of each defect type, and the percentage of affected components is unknown. So, a calculation based on stochastic rules is difficult and might lead to an unrealistic high number of necessary samples. But, for sure, under conditions like these taking out one failed part and analysing one defect on the randomly selected component is not an option.

3.5.4 Application of a wrong sampling procedure The sampling procedure is easy and obvious, when it comes to targeted sampling of products with visible failure. But, as already mentioned in Chapter 3.2, Part II, sampling of deliveries of large lots of liquids, granules or failure parts is more complicated. One risk, which has to be avoided, could be called convenience error. That means samples are preferred which are easily available and others which demand more effort are neglected. Example: A coating failure issue is persisting a whole weak erratically. The process has to be reviewed by taking samples. Although, it is more convenient to pick up just the samples that fall actually out of the machines at the end of the week, it is recommended to select also samples from the beginning of the week although they might be already stowed away in another building. A failed sampling procedure sometimes produces baffling and surprising results. The author asked one of his laboratory personnel to compare three paint samples by ATR-FT-IR spectroscopy that were provided by three different manufacturers. The result was that the three samples were identical. A close look at the spectra instantly revealed what went wrong. All three spectra contained broad and intense absorption peaks of water. The operator had not realized that the three paint samples were water-borne. He had measured the specimen simply by dropping a representative aliquot onto the ATR crystal, as is the standard procedure for liquid samples. However, water (which is the main solvent) exhibits strong and broad absorption peaks which overlay all spectroscopic features of the paint itself. The correct procedure would have been to prepare a dried film of the paint. This kind of sample treatment, however, is not useful for TOFSIMS analysis. Surface-active additives which migrate to the uppermost layer of the film whilst the film is drying, form a homogeneous layer on top that masks the main ingredients. This example shows that sample preparation is one of the key operations for correct results. It should be performed by experienced analysts who have the analytical background to determine whether a standard routine is suitable for the sample to be investigated or whether it has to be modified.

3.5.5 Inappropriate sampling tools and containers If a sudden failure occurs and a quick sampling is required, there is a frequent temptation to grab anything what is available as a tool for sampling. It cannot be stressed enough, that this is not the way it should go. Inappropriate sampling tools or container can ruin the samples or even falsify

51

Part II Coating failure analysis the results of the analyses. Some extreme examples (that have nevertheless been used in the past) are spoons, randomly picked tissues, toilet paper, industrial paper towels, recycling paper, food containers. Once the sample has been drawn, it is very important to save it in an appropriate container. Sampling vessels should be clean, especially when trace analysis is required. Clean does not mean optically clean but analytically free of traceable residues. For example, it is not best practice to use a second-hand cucumber jar for sampling from a ring line of a paint if a trace analysis afterwards has been planned. In such a case, impurities will certainly be found, but they have nothing to do with the actual problem. Sometimes the impurities caused by sampling cannot be distinguished from the substances that in fact caused the failure. The sampling will also serve at the same time for the preservation of evidence, which means that defects are protected against further impairment such as careless damage or contamination, e.g. by covering them.

3.5.6 Insufficient storing and shipping of samples Figure II.20: Some examples of inappropriate sampling tools (that in fact have been used)

Figure II.21: Example of an insufficiently wrapped sample

52

When the samples have been collected correctly according to the standards there is a good chance to ruin it by inappropriate storing and shipping. The storage as well as the sampling must not influence the area which will have to be analysed later in any way. A packaging or storing which is perfect to protect a sample against mechanical damage can be useless with respect to analytical considerations. It is quite common, for example, that samples which have to be analysed by very surface sensitive TOF-SIMS are packed into polymer foam foils. This material is brilliant if you want to protect your samples against scratches or demolition. But these foam foils are made of polyolefins and contain anti-blocking agents. If a sample is in close contact to

The power of sampling this kind of foil, the additives are transferred to the surface of the sample. If the sample surface has to be checked for substances that can interfere with a coating, you will surely find these substances. But you never know if this is a contamination through the foil or has been there before. The same can apply to industrial transport containers. They are designed to fit the product and made to protect it, but as far as samples for trace analysis are concerned, there is always the potential risk to cross-contaminate what you want to analyse. Typical contaminants are polydimethylsiloxane or injection moulding separating agents. According to the experience of the author´s laboratory it is safe and highly recommended to wrap the samples Figure II.22: Polymer transport container equipped with in white (not used) copy paper. injection moulded automotive parts

3.6 Micro­ sampling In coating analysis there are often very small features like crater, paint spots, overspray spots, particle and fibre inclusions and so on. This offers a few challenges with respect to sampling methods. Sometimes a particle of the size of a few tens of micrometres have to be separated and fixed for the analysis. One sampling technique is a scalpel wedge cut of paint spots. This requires a binocular microscope and a fresh, very sharp scalpel blade: There are two options for cutting the spot. One is a wedge cut with the tip sectioning the speck and the layers underneath with an angle of approximately 45°. The second is a top cut parallel to the coating surface without touching the layers underneath. With both methods the particle sectioned from the sample can be saved for further analysis (e.g. infrared microscopy) on self-adhesive tape

Figure II.23: Scalpel wedge or top cut of a paint spot

Figure II.24: Sampling a paint chip in case of adhesion deficiencies using tweezers

53

Part II Coating failure analysis If it comes to adhesion defects the interface between the coating and the substrate will have to be analysed by very sensitive methods like TOF-SIMS for trace contaminations. This means the backside of a paint chip (the side that has been sitting on the substrate) and the surface underneath have to be sampled without contaminating the areas. This can be done by lifting one edge of a loose paint chip and pulling it off with tweezers. It is important that, even with a thoroughly cleaned tweezers, the backside of the paint chip must not be touched except for the edge which is used to pull the paint chip. This operation should have been performed under binoculars.

4 Paint failures and their analytical approach 4.1 Some considerations on failure reasons This is a technical book describing procedures how to measure and evaluate data in order to retrieve facts about the root cause for coating failures. Sometimes the reasons are very simple and obvious, sometimes the process leading to an insufficient product is a complex interaction of a few minor factors that add up to a disaster. Looking over all the defect analyses, the author´s laboratory did during the last 25 years, it can be summed up that behind technical reasons the real root cause of a lot of issues is human error. A (not representative) statistical data evaluation was carried out on 500 randomly chosen damage events that have been investigated in the author´s laboratory during the last years. The surprising results show that: – 27 % of paint failures were caused by poor substrate quality – 27 % by contamination – 27 % by application failures – 16 % by poor paint quality and – 4 % by pretreatment faults In fact, there have been very few truly unpredictable error events, like an unexpected chemical reaction between a new paint additive and a solvents, or a small mechanical problem in a painting robot. But the main reason for production failures is human inadequacies. The author knows that this statement is not very popular, but this is what can be extracted looking why a failure event really happened. In Chapter 2.3, Part II an example of an adhesion problem that occurred erratically is described. The final solution was that the failure was caused by a planning and managing insufficiency. Indeed, failure analysis often deals with hidden human insufficiencies rather than technical, physical and chemical issues. Simple reasons for coating failures that often can be identified with a low-cost equipment are pretreatment failures like insufficient removal of corrosion products (e.g. rust) or dust. A cheap microscope is sufficient for that. But the majority of failure issues asks for a well-equipped laboratory with very sensitive methods to find and identify the causes. If, for example, a conveyor belt of an injection moulding machine is contaminated by a fluorocarbon lubricant, which is transferred to the surface of the produced polymer part resulting in wetting defects upon coating, a very surface sensitive technique like TOF-SIMS is necessary to identify the substance and find the source in the production process.

54

Part II Coating failure analysis If it comes to adhesion defects the interface between the coating and the substrate will have to be analysed by very sensitive methods like TOF-SIMS for trace contaminations. This means the backside of a paint chip (the side that has been sitting on the substrate) and the surface underneath have to be sampled without contaminating the areas. This can be done by lifting one edge of a loose paint chip and pulling it off with tweezers. It is important that, even with a thoroughly cleaned tweezers, the backside of the paint chip must not be touched except for the edge which is used to pull the paint chip. This operation should have been performed under binoculars.

4 Paint failures and their analytical approach 4.1 Some considerations on failure reasons This is a technical book describing procedures how to measure and evaluate data in order to retrieve facts about the root cause for coating failures. Sometimes the reasons are very simple and obvious, sometimes the process leading to an insufficient product is a complex interaction of a few minor factors that add up to a disaster. Looking over all the defect analyses, the author´s laboratory did during the last 25 years, it can be summed up that behind technical reasons the real root cause of a lot of issues is human error. A (not representative) statistical data evaluation was carried out on 500 randomly chosen damage events that have been investigated in the author´s laboratory during the last years. The surprising results show that: – 27 % of paint failures were caused by poor substrate quality – 27 % by contamination – 27 % by application failures – 16 % by poor paint quality and – 4 % by pretreatment faults In fact, there have been very few truly unpredictable error events, like an unexpected chemical reaction between a new paint additive and a solvents, or a small mechanical problem in a painting robot. But the main reason for production failures is human inadequacies. The author knows that this statement is not very popular, but this is what can be extracted looking why a failure event really happened. In Chapter 2.3, Part II an example of an adhesion problem that occurred erratically is described. The final solution was that the failure was caused by a planning and managing insufficiency. Indeed, failure analysis often deals with hidden human insufficiencies rather than technical, physical and chemical issues. Simple reasons for coating failures that often can be identified with a low-cost equipment are pretreatment failures like insufficient removal of corrosion products (e.g. rust) or dust. A cheap microscope is sufficient for that. But the majority of failure issues asks for a well-equipped laboratory with very sensitive methods to find and identify the causes. If, for example, a conveyor belt of an injection moulding machine is contaminated by a fluorocarbon lubricant, which is transferred to the surface of the produced polymer part resulting in wetting defects upon coating, a very surface sensitive technique like TOF-SIMS is necessary to identify the substance and find the source in the production process.

54

Paint failures and their analytical approach

4.1.1 Insufficient workpiece preparation Surface pretreatment is an essential part of the coating process and has dramatic impacts on the appearance and quality of the coatings. Therefore, it is worth to have a close look at the impacts the different methods have on the coating adhesion. Adhesion defects due to insufficient or unsuitable pretreatment are a very common topic both in the painting of metal parts and polymer components. The reason for errors and failures with respect to this topic are often based on the lack of knowledge about the components of the material to be painted and the chemistry and physics behind the process. “Pretreatment” in this context means surface cleaning on the one hand side and surface modification on the other hand. Some very common failures occur when the substrate is cleaned by wiping it manually. Instead of clean (but expensive) cleaning agents it happens that e.g. white spirit is used in combination with recycled cleaning cloths. The agent is an industrial mixture of aliphatic and aromatic hydrocarbons which can leave additional contaminations on the surfaces because it is not clean. In addition to that, it can extract e.g. mould release agents from polymer parts and leave them on the surface after evaporation. Another aspect is cleaning cloths that are recycled and do not meet the requirements of cleanliness in coating environments. And the third obstacle is the risk to cross contaminate the goods if the cloths are used for a long time (which is common practice). Similar considerations apply for sanding media. If a sanding paper for a metal surface is used for multiple objects it is very likely that it spreads cross contaminations if only one oily object is in the population.

Surface treatment of metal substrates There are several well-established surface cleaning procedures for metal parts: – alkaline cleaning – acidic cleaning – dry ice blasting – sand blasting The “wet” cleaning with acidic or alkaline cleaning agents before painting to remove grinding dust, release agents, oils and greases is a critical issue with respect to the chemical surface composition and the impacts are very often not well understood. If an acidic cleaning with organic or inorganic acids and surfactants is chosen, mineral surface contamination, scale, swarf, oxide residues and water-soluble residues can be easily removed. However, acidic cleaning does not reliably remove greases, oils and release agents from metal surfaces.

Figure II.25: View into a surface pretreatment cabin for steel parts

55

Part II Coating failure analysis This task is more likely to be performed by alkaline cleaning, which, on the other hand, does not remove all oils and greases (e.g. silicone oils or perfluorinated polyethers). This means that before deciding which type of pretreatment to choose, it is recommended to know exactly what type of contaminants to remove. In order to guarantee a targeted and process-safe component cleaning, it is necessary to determine beforehand by means of analyses which substances are to be expected on the surfaces. However, this assessment is usually not practiced. But even if a targeted cleaning system is installed at the beginning of a production process, it can later become ineffective if, for example, the composition of the drawing oils, stamping oils, release agents or corrosion protection oils changes without the plant operator being informed. In an automated coating environment, it is desirable to have a one-fits-all cleaning process. But the bad news is, this does not exist. Alkaline cleaners consist of a complex mixture of chemicals that saponify fats into soluble soaps, displace contaminations by surface active substances, neutralize acidic contaminants and rinse the surface. This highly alkaline, caustic cleaners are well suited for steel but are not safe on aluminium and zinc. So, if different materials run in the same process the alkaline solution may perfectly clean steel workpieces. But an aluminium part will react with the alkaline cleaner building up a layer of aluminium oxide/-hydroxide on the surface with low adhesion. If, for example, a powder coating is applied afterwards this coating will exhibit severe adhesion deficiencies. So, the cleaner selection is a vital point for the ability to effectively clean the surface and for the coatability of the metal substrate. It has to be adjusted to the metal type, the amount and chemical composition of the contaminants, the temperature of the bath, the shape of the workpieces, the way they are applied (spraying or immersion) and quality of the available water. There is no one-fits-all solution. One aspect of the surface cleaning by alkaline or acidic cleaner solutions in an automated process is also very often neglected resulting in coating failures like adhesion problems, craters or specks: Rinsing after cleaning is most important. The result of the chemical and physical interaction of the cleaning solutions are saponized fats, emulsified contaminants, cleaners and particles on the surface of the workpieces which have to be rinsed away before they dry. It is recommended to do this in a multistage process and the last step should be demineralized water. Heating of the rinsing water can help to remove the residues of the cleaning process. Again, the effectiveness of the rinsing process depends on a few parameters: the processing time, the shape of the workpieces, the design of the cleaning system (spray or immersion, the chemistry of the cleaning products and the water quality). Another aspect which is related to the above mentioned issue is the monitoring of the baths of a cleaning system. In a well-managed plant, the baths are checked at regular intervals with regard to parameters such as conductivity and pH value, on the assumption that everything is safe if there are no deviations from the target values at these measured values. But this is a mistake! Silicone oils, mineral oils or lubricants containing fluorine have no influence whatsoever on the conductivity or pH value and are, therefore, simply not detected. This means that all measured values may be ok, and yet the bath may be highly contaminated with adhesion inhibiting substances. So, the usual checking routines for pretreatment tanks are insufficient with respect to the load of oils and greases that accumulate and especially for lubricants that cannot be saponized like polydimethylsiloxane or perfluorinated polyethers. In order to test for substances that can inhibit adhesion the TOF-SIMS analysis of samples taken from the pretreatment tanks has been proved to be the best method.

Sand blasting Blasting the surface by sand, glass beads or steel particles is a method which is used to clean the surface of metal workpieces prior to coating application. These abrasive techniques are meant to

56

Paint failures and their analytical approach remove rust, dust and contaminations. They work well and leave in most cases a sufficient result for industrial coatings of e.g. machine parts but do not remove all types of lubricants safely. In addition to that, problems occur if the abrasive material is used several times because oils, lubricants, separating agents removed from a workpiece stay in the blasting material and thus “infect” the next workpiece. The cleanliness of the abrasive material can be checked by extraction with a suitable analytical grade solvent and FT-IR and TOF-SIMS analysis.

Dry ice blasting The cleaning of a metal surface by frozen CO2 has a few advantages. In contrast to methods like sand blasting or glass beads it does not leave behind waist that has to be disposed. The mechanism consists of four steps [8]: 1. The organic material is cooled down by the dry ice and hardens. 2. Caused by the sudden local cooling a thermal tension is created in the boundary of contamination and substrate which loosens the contamination. 3. The pure mechanical impact of the ice pellets is transferred to the loose contamination. 4. The explosive sublimation of the CO2 blasts away the contamination. This is the theory and it works well for loosely bound particles and dust. But in contrast to the marketing promises that oils, lubricants and release agents can be easily removed the author has encountered a lot of cases that showed that even flat steel objects with rough surfaces have not been cleaned sufficiently. The problem is that you cannot rely on visual evaluation because a monolayer of e.g. polydimethylsiloxane remaining on the surface of a workpiece is not visible but nevertheless can cause severe adhesion or cratering problems. The only way to check the success of the dry ice cleaning is a surface sensitive analytical technique like TOF-SIMS. Even the surface infrared spectroscopy (external reflection mode) which serves good for the detection of oil residues in the thickness range between nm to μm range is not sensitive enough. Simple surface tension measurements by testing inks often fail because of surface roughness.

Figure II.26: SEM-BSE (backscattered electron mode) image of the boundary between a correctly formed phosphate layer (B) and an inadequate layer (A) with EDS element spectra of both surface areas

57

Part II Coating failure analysis The effectiveness of other cleaning methods such as CO2 snow, sandblasting or shot peening is also rather unreliable with regard to some adhesion-impeding substances such as “silicone oils” or perfluorinated polyethers. So far as metal substrates are concerned, pretreatment processes serve not only for cleaning but also for corrosion protection. “Bonderizing” of a metal surface is an example of a widely used process which produces a closely packed layer of phosphate crystals on the metal surface. The quality of the phosphate layer is important for the subsequent painting process. Inhomogeneities in the crystal growth have a great impact on the surface micro-structure [9]. The bonderizing process cannot be tested and approved by optical light microscopy because the phosphate crystals are too small. The SEM image of a phosphate layer (Figure II.26) however, clearly shows the difference between a correct phosphate layer (B) and an inadequately treated area (A). Combining this with energy dispersive X-ray (EDS) or EDS element mapping, it can be demonstrated that the failed area is characterised by a lower phosphate concentration and the lack of proper phosphate crystals. The reason for this crystallisation defect could be, for example, organic surface contaminants which can then be detected by surface mass spectrometry (TOF-SIMS).

Polymer surface cleaning and pretreatment

There are a number of pretreatment and cleaning processes for polymer components that are well established and widely used. But sometimes they do not always leave a paintable surface. Example: A “power wash” system is intended to remove surface contamination from polymer components and thus make them paintable. Whether this is successful depends to a large extent on the polymer and its composition. There are always polymer formulations that behave quite unexpectedly. For example, large polymer components made of PC/PET like automotive exterior panels: Due to their complex component geometry they already contain significant amounts of internal release agents in the granulate. Running PC/PET workpieces through a power wash followed by hot air drying some of the release agents are removed from the surface during the washing process. But sometimes a significant amount of these internal release agents migrate back from the inside of the material to the surface during the subsequent drying process. In the end, the component is less clean after cleaning than before the power wash. To approach these effects, which under certain circumstances only occur in the uppermost molecular layer of the polymer, only very sensitive and surface-sensitive methods that also provide

a)

b)

Figure II.27: a) Crystalline deposits of low molecular weight polyamide on a PA6 surface after power wash and b) ATR-FT-IR spectrum identifying crystalline polyamide

58

Paint failures and their analytical approach information about the molecular structure of the interface are possible and thus the TOF-SIMS method is the recommended choice. Similar effects can occur as already mentioned with polyamide, which contains low molecular weight components from the production process. These are readily transported to the surface by water from the power wash and/or during weathering tests. After subsequent drying, they form loosely bound precipitations on the surface that are not paintable (see Figure II.27) that can easily be identified by ATR-FTIR spectroscopy. The power wash is also often used in combination with surface-activating methods such as flame, plasma or corona treatment. However, numerous TOF-SIMS studies over the last 20 years have repeatedly shown that this combination sometimes can be counterproductive. If, for example, a flame treatment is carried out before the power wash, impurities are thermally decomposed and broken down into small, easily removable fragments that can be removed by the power wash. At the same time, all possible surface activations are neutralized. If the flame treatment is applied after the power wash, residues will remain on the surface and the effects of the power wash are counteracted. The flame treatment itself is used to activate polymers that normally do not exhibit sufficient functional groups on the surface to guarantee a good adhesion by chemical bonding or polar bonding. Polyolefins, for example, consist of long C-H chains but do not have active groups like OH, NH or COOH. The effect of the flame treatment is the insertion of oxygen into the hydrocarbon chain and thus provide more active polar groups in the molecule. The effect

Figure II.28: Extract of the TOF-SIMS spectrum of the positively charged secondary ions of a PP compound before and after flame treatment showing a significant increase of oxidized fragments

Figure II.29: Relative intensity of oxygenized species in the uppermost layer of two polyolefins after power wash cleaning

59

Part II Coating failure analysis and the success can be measured by TOF-SIMS. The spectra of pure untreated polyolefins show fragments of the type CxHy+ whereas flame (or plasma) treated surfaces give rise to additional oxygen containing fragments of the type OCxHy+. The peak intensity is a measure of the amount of oxygen activated groups that are present. In the author`s laboratory it has been proved that measuring the intensity of a fragment C2H3O+ (originating from oxidized species) and correlating it with to the intensity of the fragment C3H5+ of regular polyolefin structure (see Figure II.29) delivers a relative value. If this value ​I​C​  ​  ​​​H​  ​​​O​​  +​​​

2 3  ​​ ≥ 0.5 ​​  _ ​I​  +​​

​C3​  ​​​H5​  ​ ​

is 0.5 or higher the activation is sufficient. Figure II.29 shows the results of a pretreatment experiment of “Hifax” and “Hostacom” polymer. Both polyolefin polymers have been analysed before treatment. The relative value of oxygenized fragments was 0.01 for “Hifax” and 0.02 for “Hostacom”. After that surface contaminations have been removed by wiping the surfaces with polar and non-polar solvents. Subsequently the surface has been flame-treated and analysed again directly after the treatment. The relative values (arbitrary units) in Figure II.29 demonstrate that for two of the three “Hifax” batches the degree of activation is above 0.5 which is good. “Hostacom” treated the same way did not show a significant increase of oxygen activation. On the other hand, it must be emphasized that the oxidation degree as measured by this TOFSIMS method is a necessary but not sufficient prerequisite for the paintability of the substrate. Other parameters like e.g. the degree of contamination play also an important role. But the TOF-SIMS method provides the means to check the activation and the cleanliness with only one measurement.

4.1.2 Handling failures Up to this point the technical aspects of coating failures have been explained. But one of the mayor influences on the paint quality are the people that do the job. Especially high-performance coatings require a clean environment comparable to medical workplaces. When coating a high gloss piano black paint on a large object even the smallest hair or skin flake can cause a visible failure in the appearance of the coating. Talking about hair and skin there are a lot of skin care and cosmetic products that contain substances (e.g. cyclomethicones and dimethicones) that can cause e.g. a severe cratering of a coating. This means that the employees, which are directly involved in the coating process, must be well instructed and compliant with respect to not using any of these products. On the other hand, the methods described in this book can easily detect these critical substances in everyday products. Especially the TOF-SIMS method is widely used to analyse series of production material like working clothes, tissues, container, gloves for substances that can interfere with the paintability. The testing of some production material according to critical substances is very important because a production glove, that is surface treated by polydimethylsiloxane from the sewing process, can cause a disaster, when the glove is used for handling of goods that have to be coated.

4.1.3 Application conditions As the mentioned statistical evaluation shows, more than a quarter of the paint failures analysed had been caused by application failures. Typical faults are: – non-compliance with paint drying conditions – non-compliance with “recoating windows”

60

Paint failures and their analytical approach – – – – – –

inappropriate location or construction of paint shops wrong layer thickness incorrect mixing with 2K and 3K systems insufficient cleaning of the spray booth wrong choice of paint solvents incorrect operation of the paint shop

Figure II.30 shows a polymer test panel that was painted with a 2K water-borne polyurethane paint in an automatic painting line. The surface is covered by craters (or better orange peel). The optical inspection led to the initial (but wrong) conclusion, that this might be a paint quality problem. To find the real reason for this painting problem, it was necessary to perform a microanalysis within the paint failures and compare the data to reference samples. So, the analytical method had to combine: – high surface sensitivity – low detection limit – molecular information – high lateral resolution for microanalysis purposes These demands are met by secondary ion mass spectrometry. The TOF-SIMS (Time of flight secondary ion mass spectrometry) analysis of the failed paint surface within a paint failure (Figure II.31) shows an unusual agglomeration of a polypropylene glycol and paraffins inside the craters. As these substances are widely used as paint ingredients as well as in water additives, surfactants and so on, the purpose of the subsequent analytical procedure was to find the source of this chemical. The spectrum of a sample of the application booth water of the paint shop showed the glycol compound being highly concentrated in the booth water. The comparison of the TOF-SIMS spectra of booth water, coagulation additives, batch sample and several materials that had been used in the painting process finally led to a defoamer which was composed of polypropylene glycol. After several cross-checks and further reference analyses it was established that an overdose of the defoamer in the booth water created a fine aerosol of water carrying the defoamer.

Figure II.30: Test panel showing heavy surface defects of the polyurethane coating

Figure II.31: Comparison of positive TOF-SIMS spectra of the failed paint area, application booth water and a reference sample of a defoamer water additive

61

Part II Coating failure analysis This aerosol precipitated on the wet paint layer during the painting process, leading to heavy paint flaws. So, the obvious cause of the failure was an overdose of a defoamer in the paint application booth water, but the reason behind this was the fact that no procedure was installed to check the water quality on a regular basis and exchange the water when necessary. The whole application process offers numerous possibilities to fail. Some examples: – During transport, storage and equipping of coating skids in non-clean room areas, very fine impurities can be deposited on components by aerosols and dust, which are not optically visible, but can nevertheless lead to craters and/or wetting problems. – The cabin air can be loaded with PWIS e.g. by humidification systems, suction of contaminated outside air, outgassing from operating materials or spray mist from water walls. – Lubricants from moving parts can fall onto the goods and cause cratering erratically. – Mould release agents from plastic parts can migrate to the surface due to inappropriate temperature conditions and cause peeling. – Paint craters/wetting faults can also be caused by airborne interfering substances. A distinction must be made between – the ambient air outside the paint booth, – the cabin air, and – the paint shop ventilation (steering air, atomizing air etc.) – A common failure are contaminants that are introduced into the cabin by the compressed air. Possible causes include design faults in the compressed air system and improper maintenance. This type of contamination is particularly “insidious” because even improper maintenance, where a few μg of a PWIS gets into the compressed air flow, can cause irreproducible craters for months. As a rule, a PWIS introduced into the compressed air flow is not continuously transported and discharged but is deposited in dead spots of the compressed air system and is then discharged at irregular intervals, which in this case means that it is mixed with the paint material via the painting air and sprayed onto the components. – Inappropriate programming of robots can lead to paint spots, overspray or wetting defects. – Insufficient design of the cabin air of a paint booth can lead to orange peel or spots. – Insufficient or miss-mixing of multi-component coating systems results in poor adhesion, insufficient chemical and physical resistance. – Excessive paint layer thickness can lead to solvent pops (bubbles), cracks and adhesion defects. – Oven time or temperature mismatch causes poor appearance, insufficient stability or discolouration. – Mismatch of the distance between spray gun and substrate during conventional spraying causes either sags and runs or a granular texture. More details of the evaluation of application processes will be discussed in the field analysis Part III Chapter 3 of this book. But it must be mentioned, however, that some of the listed failures cannot be detected by analytical methods.

4.1.4 Environment and climate One aspect of the application conditions that influences the result of industrial coating processes is the environment. Environment does not only mean the landscape around the plant but the climate and the conditions inside the production plant. But how can the environment influence the process?

62

Paint failures and their analytical approach The first parameter is the ambient air. Most production plants suck the process and production air from outside. It contains (depending on season, daytime and weather) different amounts of components that should not be in the coating. Therefore, the ambient air must be filtered before being used in production. To be save with respect to faultless coating the filters must be checked and changed on a regular basis. But, nevertheless, there are aerosols that can pass the filters and enter the production environment. These aerosols can be found and identified by a special collecting method [10] combined with appropriate trace analysis (see Chapter 3.2 “Aerosol analysis” in Part III). When it comes to coating jobs the whole plant must be free of substances that can cause coating failures. The author is aware of an instance where a sealing of an air filter system installed on the roof of a coating plant has been replaced by a silicone rubber one. This polymer contains a certain amount of low molecular weight short chain silicones which have been transported with the air stream into the paint shop and caused heavy cratering. Secondly the climate and especially the temperature and humidity play also an important role. Some failures only occur in the summer time, some happen when it is cold outside. The reason can be condensation processes when freshly coated goods are transferred into storage buildings that are significantly colder. This slows down the curing on one hand and allows for condensation of volatile compounds of the coating. The latter is enforced when the painted parts are wrapped in foils.

4.2 Investigation of adhesion failures “Without good adhesion, the coating will likely not get the chance to fail by some other mechanism!” [11] The term adhesion describes the attraction between a coating material and a substrate (which means the surface of a workpiece of any material which is meant to be coated) and consists of a combination of different bonding mechanisms: – chemical bonding – polar bonding – mechanical bonding One example of chemical bonding is siloxane or silane-based coatings. They contain active Si-OH or Si-OR (with R representing an organic side chain) groups which react with active sites on glass or metals and generate a Si-O-Si-bond which is quite stable. Polar bonding occurs when dipoles are involved and the coating material as well as the substrate exhibit functional groups with opposite polarity which attract each other. This van der Waals forces or hydrogen bonding are weaker and can easily be disturbed by foreign substances. Mechanical bonding happens when the surface is rough, and the coating and the substrate have the chance to interlock. To improve this hooking, the surface of the substrate is grinded, etched or otherwise treated to increase the number of active sites. On the other hand, there are a number of possible causes for adhesion problems between a paint and the object to be painted: – substrate problems (delamination, hydrolysis) – coating thickness deviations – curing or drying failures – incorrect mixing ratio – impurities – incorrect adhesion test methods – incorrect or insufficient pre-treatment

63

Part II Coating failure analysis – deviations from the process instructions – hydrolysis or chemolysis of the coating material – lack of inter-coat adhesion between successive layers of paint due to application errors

Required information The following questions arise in the case of large-area delamination of coating films both from the substrate and from adjacent coating films: – What has changed from the desired state in the interfaces so that delamination could occur? – In which interface does the detachment occur? – Where is the locus of the failure? Are defect areas statistically distributed or concentrated in certain areas? To collect this information, it is therefore important to characterize the chemical composition of the interfaces and surfaces in the uppermost molecular layer. Depending on the problem and the assumed cause of the defect, different methods are used for this type of damage. The composition of the weak interface, in which the debonding takes place, is of primary interest. It allows conclusions to be drawn about the process of fault generation. But also, the layer thicknesses and the layer sequence are important in order to assess whether the specifications have been fulfilled or, for example, to detect a repair coating.

a)

b)

c)

d)

Figure II.32: Examples of paint adhesion defects of different shapes and causes a) insufficient pretreatment; b) delamination due to injection moulding defects, c) hydrolysis of the base polymer; d) incorrect paint system

64

Paint failures and their analytical approach Table II.4: Examples of coating adhesion issues and their analytical approach Possible root cause Analytical tool Preparation Question Polymer substrate delamination

ATR-FT-IR

Backside of a delaminated paint chip

Low molecular weight polymer components? Chain scission? Degradation?

Paint curing deficiencies or wrong mixing ratio

ATR-FT-IR

Delaminated paint chip compared to a sound reference sample

Deviations compared to the sound sample?

Insufficient film thickness

Metallographic cross section + optical light microscopy

Sample area neighbouring adhesion defect

Compliance with the recommendations of the coating supplier? Deviation from sound reference sample

Contamination of the surface

TOF-SIMS

-

Oils, greases, separating agents, surfactants?

Wrong or insufficient pretreatment

TOF-SIMS

Backside of a delaminated paint chip and substrate surface underneath

Trace residues of pretreatment or cleaning steps? Thermal degradation products after flame or corona treatment?

Deviations from the process parameters

ATR-FT-IR (sometimes difficult to prove)

Delaminated paint compared to a sound reference sample

Deviation from sound reference sample?

Inappropriate testing methods

Difficult to prove

-

-

Hydrolysis or degradation of the coating material

ATR-FT-IR

Ethanol or isopropanol micro-extract of the delaminated paint chip

Degradation products, low molecular weight components?

Insufficient inter-coat adhesion due to application failures

ATR-FT-IR, TOF-SIMS

Both sides of the separating plane

Composition compared to the sound reference sample

Sample preparation If delamination has occurred and a corresponding failure part is on the table, probably it has already passed through many hands. It may have been touched, scratched or wiped in the defect area. Therefore, it is usually not very useful to examine the exposed surface. Analysing a sample like this would finally result in discussions whether the detected substances belong to the problem, or whether they contaminated the examined surfaces due to improper handling during sampling and preliminary examination. Another problem is sometimes the missing piece of paint that has come off. In order to start the investigation with a “fresh” delamination, one should try to remove a fresh piece of paint and thus create a “virgin” interface. It is important to avoid contaminating this interface. This means touching the freshly removed lacquer particle only at the edge and with clean tweezers (see Figure II.24). It is desirable to prepare this samples shortly before the analysis is going to be performed. If there are e.g. droplets or films, underneath the freshly detached paint those may evaporate if the sample is stored too long and you would lose important information.

65

Part II Coating failure analysis Investigation procedure No matter which of the above mentioned possible causes lead to adhesion problems, any interfering substance can be easily detected using surface analysis methods. Important for the selection of the suitable method is the desired detection sensitivity and the depth of information. For sure, a method that provides molecular information should be chosen. This means that the aim must be to identify the interfering substance with its chemical composition. Therefore, three methods are suitable for these problems. These are the methods TOF-SIMS, XPS and ATR-FT-IR. In practice, a combination of ATR and TOF-SIMS has proven to be the method of choice for the determination of the interfacial chemistry in the area of adhesion interferences. When it comes to using the ATR-method there are two instrumental options depending on the size of the area to be examined. Standard ATR-spectroscopy (e.g. using a “GoldenGate” accessory) with a sampling area of 2x2 mm or more is used for “large” paint chips. If the target area is smaller (diameter detection of a monomolecular contamination with polydimethylsiloxane)

67

Part II Coating failure analysis Evaluation Characteristic fragments of a polydimethylsiloxane are detectable on the aluminium surface in the area of delamination. The comparative analysis of the defect-free aluminium surface clearly shows no contamination. The conclusion, therefore, is that the aluminium surface was contaminated just before the painting process. This raises the question as to the source of the contaminants. Indeed, there are many possible reasons: – contamination of the raw material – improper transportation – failure in storage – residues from substrate-treatment operations – aerosol precipitation due to ventilation systems in the production plant – residues of substrate handling

Supplementary process analysis These findings ask for a detailed, critical and comprehensive review of the production process for possible contamination sources. The latter must start by auditing production of the raw material, followed by the storage and transportation conditions, and subsequent processing. This does not mean simply checking the documentation but rather stepping into the production line to take a close look at the processing steps. (see Chapter 3 “Field analysis”, Part III)

Figure II.35: Section of the positive TOF-SIMS spectrum of an aluminium surface after cleaning, a comparative preparation of the cleaning solution, an aerosol sample and a disposable glove used in the production process (selected mass range 125u to 264 u; => detection of polydimethylsiloxane on a disposable glove – red arrow)

68

Paint failures and their analytical approach This preliminary review of the production process and reporting of potential risks is followed by surface analysis helps to confirm the facts. Figure II.35 shows how this procedure might be applied to the above-mentioned aluminium sample. On the metal surface beneath the delaminated paint, a polydimethylsiloxane contaminant was detected. A critical review of the whole process revealed a few potential sources for this contaminant. These areas were sampled and subjected to TOFSIMS analysis. The following samples have been investigated: – an aluminium panel taken from the production line after a cleaning procedure but before the painting process – a sample of the cleaning agent used to treat the aluminium surface – a so-called aerosol target which is meant to detect precipitation by aerosols contaminating the air inside the production plant – a piece of a glove used to handle the aluminium Evaluation This set of reference analyses clearly shows that the polydimethylsiloxane causing the paint failure originated from the glove used to handle the aluminium samples (red arrow in Figure II.35). Control measures Replacement of this glove by a silicone-free one solved the adhesion problem.

4.2.2 Adhesion defects caused by migration processes Up to now, the examples have mainly dealt with external contaminations. But segregation or migration of components from the substrate itself is an issue that must be addressed in the context of polymer substrates. Pigments, for example, are incorporated into polymers by means of dispersing agents. These additives are essential for ensuring that the pigments will be uniformly dispersed. Provided they are present in low concentration, they do not cause problems. But if they manage to migrate to the surface of an injection moulded polymer part, they form a film on the surface that can impair adhesion and wetting.

Example: Paint adhesion failure of a 2K PU paint

The following example shows a paint adhesion failure of a 2K PU paint on a PA 6.6 lamella caused by migration processes of polymer components (see Figure II.36).

Investigation procedure In order to investigate the cause, the paint was removed from the delamination area and analysed on the paint backside facing the polymer using the TOF-SIMS method. As a comparison, the polymer surface exposed underneath was also examined (see Figure II.37). The aim of the analyses was to identify any release agents present.

Data analysis The TOF-SIMS analysis revealed the presence of low-molecular-weight oli-

Figure II.36: Paint adhesion deficiency of a 2K PU coating on PA 6.6

69

Part II Coating failure analysis gomers of the polyamide network on the underside of the delaminated paint, as well as on the polymer surface. The oligomers are short-chain molecules composed of 2, 3 or 4 repeat units of the monomer (see Figure II.38).

Evaluation The low-molecular-weight amides tend to migrate to the surface, where they form a loose boundary which is soluble in different organic solvents and is renowned for preventing paint adhesion. Sometimes they cannot be detected until a weathering test has been performed to verify the long-term quality of a moulded and coated part.

Figure II.37: Overview sketch showing the analysed sample areas of the polymer compound

Figure II.38: Extract from the pos. TOF-SIMS spectrum of the backside of the paint that has been detached from the PA 6.6 surface (detection of silicone traces and oligomers of PA 6.6)

70

Paint failures and their analytical approach These so-called oligomers are loosely bound to the surface, easily soluble and prevent adhesion between the coating and the polymer substrate. Such low-molecular fractions were also detected in the granules. If they are not removed before injection moulding (e.g. by treatment of the granulate) or washed off after injection moulding, they can lead to delamination even after storage in a cold storage tank. Power wash treatment of such components can also be problematic because the polyamide absorbs water during the power wash process, which then escapes during the subsequent drying process, transporting the low-molecular components to the surface. The result are thin layers of loosely bonded polyamide which prevent paint adhesion and bonding.

4.2.3 Delamination caused by moulding conditions For the automotive supplier industry, coating of polymer parts is a critical matter. Uncoated polymers for, e.g. seat faceplates, dashboards, door handles, are only acceptable for low-cost cars. For modern standard and premium cars, a sophisticated surface finish represents the state of the art and acts as major proof of quality. At the same time, the tendency to cut production costs every year has been shown to impair the quality, e.g. of the moulded parts. Each company in the supply chain of any given automotive interior part clearly seeks to turn a profit and must reduce costs. Companies, therefore, sometimes use low-quality materials and recycled polymers or widen the parameters of the injection moulding process to an extent that leads to polymer degradation. While this approach might boost production economics, it very often yields moulded parts that are difFigure II.39: Delamination of a water-borne soft coat ficult to paint. caused by delamination of the uppermost polymer layers One result of inappropriate moulding parameters is a layered structure within the polymer. The polymer forms thin layers that are only loosely bonded together. This fault is not visible but attempts to paint such a part lead to delamination of the uppermost polymer layers. The impression gained is that a paint failure has occurred, but in fact it is the polymer which fails due to improper treatment during the moulding process. An example of this failure is shown in Figure II.39. If such a defect occurs, the question immediately arises where exactly the delamination occurs. Delamination Figure II.40: SEM image of the surface of an ABS/PC of the polymer is not always as clearly polymer after delamination of the coating

71

Part II Coating failure analysis visible under the light microscope as shown above. This is a very severe delamination example that is indeed detectable by optical inspection. However, sometimes it needs more effort (like e.g. a scanning electron microscopy) for the investigation of the polymer surface underneath the detached coating (see Figure II.40). The SEM image of an ABS/PC polymer surface reveals a severe damage of the polymer surface after coating delamination. The polymer exhibits loose strings standing out of the surface. This pattern is not caused by the paint but is a result of the moulding conditions temperature and pressure.

Examination procedure This is a question that can easily be answered by ATR-FT-IR spectroscopy. The underside of the delaminated paint (which is the side facing the polymer surface and formerly formed the paint/ polymer interface) is analysed without any further preparation by forcing it into contact with the ATR crystal. If a cross-cut has been performed for the purposes of quality control, the small pieces of paint that came off make suitable samples for the ATR. As for the substrate side of the delamination area, there are various possible approaches. In most cases, the moulded polymer part is too big for direct contact with the ATR crystal. Thus, a piece of the polymer matching the size of the ATR crystal contact zone must be cut out of the failed area. This approach can fail, if the moulding is too thick, or the surface has a strong surface topography. If sufficient contact cannot be obtained between the polymer surface and the ATR

Figure II.41: 1st ATR spectrum of the backside of the delaminated paint of a failed layer, 2nd reference spectrum of a paint sample taken from an area without failure, 3rd reference spectrum of the polymer material [S = substrate ABS, L=paint]

72

Paint failures and their analytical approach crystal for the above-mentioned reasons, the solution is a thin scalpel-cut parallel to the polymer surface. A 0.5 mm slice of the sample is cut out and can easily be used for ATR spectroscopy. As ATR-FT-IR has a penetration depth of about 1 to 2 μm, it readily answers the questions raised above. In addition to the analysis of the interface, reference measurements of the polymer substrate and the paint may be necessary to ensure that the spectra may be evaluated accurately. An appropriate reference sample of the polymer can be produced by cutting into the substrate or cutting a granule that was used for the moulding process. Figure II.41 shows a spectrum of the underside of the delaminated paint, a reference spectrum of a paint sample taken from an area without defects and a reference spectrum of the polymer.

Evaluation The ATR spectrum of the backside1 of the delaminated paint contains characteristic signals of acrylonitrile-butadiene-styrene (ABS), which is the composition of the polymer. Comparison of the spectra of the paint and the polymer clearly reveals that the delaminated paint has a thin layer of substrate on its underside, this indicating that the failure is caused by a substrate weakness and not by poor paint or painting quality. Paint adhesion to the polymer itself is excellent. However, the instability of the uppermost polymer layers causes delamination of the paint together with a thin layer of the polymer. This polymer layer has been detected on the underside of the paint after delamination. Sometimes the solvents in paint can reinforce this effect.

4.2.4 Delamination due to application faults Cross-linking, curing and/or drying is essential for paint adhesion. Isocyanate-crosslinking polyester polyurethane coatings, for example, make up a large proportion of industrially processed coatings. The mixing ratio, the crosslinking and the drying and curing conditions specified by the manufacturer must be followed in order to ensure optimum adhesion and film stability. In practice, however, it sometimes happens that the plant operator develops a creative interpretation of these conditions, e.g. because he cannot reach certain drying temperatures or wants to save costs. In most cases, however, this shot backfires. Example: Drying and curing a PESPUR paint for example, requires an object temperature of 80 °C for 30 min in order to start and complete the chemical reaction of cross-linking the blocked isocyanates with the polyester binder. These recommendations cannot be replaced with 60 °C for 60 min instead. In the latter case, the obtained coating film may look exactly like a correctly crosslinked coating film, but in fact it certainly is not. This is especially unpleasant if the painted parts have al1 W  hich is the side of the sampled paintchip which is oriented versus the polymer substrate

Figure II.42: Light microscope image of a paint release of a 2K PES PUR decorative lacquer from a PC/PET surface after cross-cut test

73

Part II Coating failure analysis ready been delivered and installed before the lack of adhesion is noticed. Another reason for paint deficiencies can be mixing errors in 2K or 3K systems, caused by defective mixers or human error. Although a 2K PUR coating lacking hardener or with too little hardener being added often has a different gloss level and could therefore be noticed already during the final inspection, this does not happen often enough. With dual cure coatings, on the other hand, incomplete PUR crosslinking or insufficient UV crosslinking result in lower scratch resistance and possibly also negative adhesion properties. How can such cross-linking, mixing or drying errors be detected? Usually the infrared spectroscopic analysis (ATR-FT-IR) is sufficient because with 2K/3K PUR coatings, dual cure coatings and epoxies the drying and curing leave “traces” in the infrared spectrum which can be evaluated qualitatively or sometimes quantitatively [13] [14] [15].

Example: Delamination of a 2K polyester polyurethane paint from a polymer surface Required information For the investigation whether insufficient crosslinking, mixing or drying errors are the root cause, it has to be checked to what extent there are spectral features that reflect these issues. Fortunately, this is the case for many types of crosslinking. For polyester polyurethane lacquers, the bands of the polyester binder at approx. 1728 cm-1 and of the isocyanate hardener at 1683 cm-1 can be integrated, for UV lacquers the signals of free acrylate monomers at 809 cm-1 (± 1 cm-1) and 1405 cm-1 (± 2 cm-1) in the FT-IR spectrum. The actual condition of the defective sample can thus be detected easily by infrared spectroscopy. However, this result can only be evaluated if the target condition is known. For example, with PES-PUR coatings, the binder/hardener ratio can be determined by comparing the intensities of the polyester band at approx. 1728 cm-1 and the isocyanate

Figure II.43: ATR spectrum of the polymer side interface of a delaminated coating film measured at two positions, the resulting difference spectrum and a database spectrum of an aliphatic isocyanate hardener

74

Paint failures and their analytical approach

Table II.5:  Examples of application errors and their analytical approach Application failure Cause Analytical procedure Insufficient layer thickness

wrong robot programming wrong solvent content insufficient positioning of the workpieces blocked pipes or gauges

metallographic cross section + fluorescence microscopy

Wrong curing and drying conditions

insufficient design of the plant non-compliance to the recommended coating conditions

ATR-FT-IR

Wrong hardener choice

negligence, logistic errors, insufficient training of the workers

ATR-FT-IR, TOF-SIMS

Wrong choice of paint

insufficient information, planning errors, inadequate labelling of paint container, negligence

ATR-FT-IR, TOF-SIMS

hardener band at 1683 cm-1. However, this mixing ratio is different for each coating system. In order to assess whether there are any deviations in the mixture, suitable reference samples (e.g. reference panels supplied by the paint manufacturer) and the application specifications of the paint manufacturer are required.

Preparation To investigate whether an incorrect mixing ratio or insufficient crosslinking is the cause of delamination, a sample of the delaminated paint is taken and analysed on both sides by infrared spectroscopy using the ATR-FT-IR method. The spectra are compared with those of a correctly crosslinked system.

Figure II.44: Drying marks on a steel workpiece due to incomplete rinsing

Evaluation The evaluation refers to characteristic bands in the infrared spectrum of the coating system, which reflect the crosslinking and/or the mixing ratio. In the case of 2K polyester polyurethane coatings, for example, intensity of the isocyanate band at 1683 cm-1 is related to the intensity of the polyester band at

Figure II.45: Adhesion deficiency of a cathodic electrocoating on aluminium

75

Part II Coating failure analysis 1728 cm-1. This can be done either by the peak height in the spectrum or by evaluating the peak area. The evaluation programs of modern IR instruments often contain ready-made routines for this task. In addition to the purely visual comparison of the spectra, (dimensionless) numerical values are obtained which give an idea of whether or not the amount of hardener required according to the manufacturer's specifications has been used in the coating pattern in question. In the present case, even the light microscopic observation of the delamination area shows that there are unusual colour inhomogeneities on the polymer surface exposed underneath the delaminating coating (see Figure II.42). In order to assess whether these inhomogeneities are caused by insufficient mixing, a paint particle was freshly delaminated and analysed at two points on the back of the paint using infrared spectroscopy (ATR-FT-IR). The spectra (Figure II.43) show that the peak ratio of the isocyanate band at 1683 cm-1 and the polyester band at 1728 cm-1 varies from position to position. The comparison of the difference spectrum between position 1 and position 2 with a database spectrum of the isocyanate hardener also confirms that more hardener is detectable at position 2 than at position 1. This means that there is an inhomogeneous distribution of binder and hardener in the coating layer that is to bond to the polymer. Complete crosslinking and sufficient adhesion cannot be expected under these conditions. In addition to the above-mentioned errors in mixing the coating, crosslinking and drying (which of course also represent application errors), there are numerous other possibilities for application errors to cause adhesion deficits listed in Table II.5.

Figure II.46: Extract of the TOF-SIMS spectrum (positive polarity) of the aluminium surface underneath the debonded cathodic electrocoat (showing characteristic peaks of the electrocoat BpA and of sodium and potassium chloride)

76

Paint failures and their analytical approach

4.2.5 Delamination due to insufficient pretreatment In Chapter 4.1.1, Part II the implications of different surface pretreatment processes on the adhesion of coatings have already been discussed. A variety of failure risks exist when it comes to pretreatment and cleaning: – inappropriate cleaning bath chemistry – insufficient rinsing due to workpiece geometry (Figure II.44) – insufficient monitoring of the pretreatment baths are the common sources of coating failures. The effect of incomplete rinsing due to inappropriate process parameters is demonstrated in Figure II.44. The example of an aluminium workpiece with a defective electrocoating (Figure

Figure II.47: Excerpts from TOF-SIMS spectra of the negatively charged secondary ions of a PC-PET component before power wash treatment (the 1st and 2nd spectra) and after power wash (the 3rd and 4th spectra)

77

Part II Coating failure analysis II.45) shows large areas of inadequate adhesion after cross-cut test. Figure II.46 displays a spectrum extract of the TOF-SIMS spectrum of the aluminium surface underneath the detached coating. It reveals sodium and potassium chlorides being still present although the workpiece should have been rinsed and dried before the coating was applied. Indeed, the salts are not the real root cause for this adhesion defect but loosely bound aluminium corrosion products which also have been found. But the presence of the water-soluble salts clearly shows that a last de-ionized water rinsing step failed. The salts should have been rinsed down as well as other by-products and residues. The TOF-SIMS spectrum demonstrates what is not visible: an insufficient pretreatment of the aluminium. It is important to note that trace residues must not necessarily be visible but still can be critical. The effects of using a power wash for cleaning products has been also discussed in detail in Chapter 4.1.1, Part II. The experience of the last decade shows that power wash treatment does not necessarily produce a surface ready for coating. Migration of additives and thus recontamination of the freshly cleaned surface from the polymer itself is one issue. Another one is incomplete rinsing and cross contamination between different polymer products.

Example: Delamination of a coating from a PC-PET surface

In a paint shop of an automotive supplier parts made of PC-PET exhibit adhesion problems erratically, even though the bodyshell parts are cleaned with a power wash. The cleaning effect of the power wash was therefore checked by TOF-SIMS analyses.

Procedure One raw workpiece was divided into two halves before cleaning. One half ran through the power wash while the other half remained untreated. Afterwards, the composition of the surface of both halves was examined using the TOF-SIMS method (see Figure II.47).

Evaluation The TOF-SIMS spectrum of the untreated component surface shows only the typical signals of PC and PET (left side in Figure II.47). After the power wash, however, clear signals of phosphates, sulphates and palmitic acid compounds are also detectable. This means that the component was very clean and basically paintable before applying the power wash. After using the power wash, impurities from the power wash were on the surface, which can have a negative influence on adhesion. This shows that the power wash was not operated properly, because at least the water-soluble phosphates and sulphates should have been removed in the last rinsing step with deionised water. The palmitates are internal release agents which obviously migrated to the surface by the power wash. These release agents can impair the paint adhesion.

4.3 Paint cratering and “fisheyes” One of the most frequent painting problems is cratering. As with adhesion problems, there is more than one reason for cratering [16]. Here is a selection of possible root causes: – contaminants on the surface – agglomeration of paint components (e.g. gel particles) – unexpected side reactions of paint components – substrate failures – overdose of a paint component

78

Paint failures and their analytical approach – – – – –

aerosol precipitations into the wet paint film paint bubbles that burst paint spray mist humans (working clothes, gloves, skin or hair products) contaminations of the atomizer or horn air

But the meaning of the term “crater” is sometimes blurred. It is often mixed up with “fisheyes”, pinholes, and wetting defects. A crater definition states “Cratering is the formation of small bowlshaped depressions in a coating film” [17]. Another one “round recesses with a diameter from 0.5 to 3 mm. The appearance ranges from very flat depressions in the final layer to serious wetting problems that go all the way down to the substrate [18]”. A pinhole is referred to as “a small hole in the paint work as a result of improper sanding of solvents boils in lower paint layers” [18]. But on the other hand, pinholes also appear with cavities in the workpiece that open-up to the surface. When it comes to judging a coating failure at the coating line, e.g. by a microscope or magnifying glass, it is often hard to distinguish if it is a pinhole or a crater. A “fisheye” according to the definition [17] are “small, crater like openings in the finish after it has been applied. They are usually caused by oil and grease on a coating substrate”. Or “small depressions with a central mound” [19]. To the author`s opinion this definition is close to the one for crater. All the above-mentioned terms should be summarised under the name “crater”.

a)

b)

c)

d)

Figure II.48: Light microscope image of paint craters caused by a) accumulation of a filler in a powder paint, b) contamination of the paint material with a matting agent and wetting disturbances caused byc) a lubricant from a valve and d) a compressed air oil

79

Part II Coating failure analysis Table II.6:  Paint crater, possible causes and suitable analytical methods Type of crater Possible cause Analytical method Flat dent

contamination by foreign paint

infrared microscopy

agglomeration of paint additives

infrared microscopy/TOF-SIMS

overspray

infrared microscopy/TOF-SIMS

high voltage spark (powder application)

-

agglomeration of paint components like e.g. filler

cross section/infrared microscopy/SEM

side reaction of paint components

SEM/EDS/infrared microscopy/ TOF-SIMS

substrate failures

cross section/optical light microscopy/ fluorescence microscopy

cracks or cavities of the substrate

cross section/optical light microscopy/ fluorescence microscopy

surface contaminations

TOF-SIMS

high voltage spark (powder application)

-

residues (salts) of pretreatment process

SEM/EDS/TOF-SIMS

solvent boils in lower paint layers

cross section/optical light microscopy/ fluorescence microscopy

Wetting defect (round)

contamination of the surface by substances that impair paint wetting (e.g. silicone oil)

TOF-SIMS

Wetting defect (irregularly shaped)

contamination of the surface by infrared microscopy, TOF-SIMS substances that impair paint wetting (e.g. greases, waxes, separating agents)

Pinhole

Required information Paint crater can be caused by – surface active substances as well as – by inclusions underneath the visible coating surface, or – substrate defects (pores, cracks), or – paint component agglomeration The necessary information is, on the one hand, the chemical information about the (surface) composition of the paint inside the crater and outside on the undisturbed surface and, on the other hand, the bulk composition of the paint system.

Examination procedure The depth information can be obtained from a cross section (microtome or scalpel section) or a cross section through the centre of the crater, which must be examined for agglomerations or foreign contamination by infrared microscopy or scanning electron microscopy. Using spatially resolved methods such as TOF-SIMS, IR microscopy or SEM/EDS, a chemical analysis of the bottom of the crater can be carried out under visual control. This allows a specific

80

Paint failures and their analytical approach chemical characterization of gel particles, dust particles or oil droplets in the crater centre. By comparative analysis of the crater centre and the sound paint surface chemical differences are determined. In particular, if the preliminary light microscopic examination of the crater indicates that the bubble may have burst and collapsed, a microanalysis of the crater inner walls may also be useful in order to detect any residual substances that may indicate the cause of the bubble formation (and the subsequent cratering). Modern surface analytical methods offer a variety of tools for discovering the root causes, see Figure II.50). TOF-SIMS, IR microscopy and SEM/EDS permit chemical microanalysis with visual inspection. Thus, focused characterization of gel particles, dust particles or oil droplets in the

Figure II.49: SEM pictures of paint craters of different sizes, shapes and origin

Figure II.50: Investigation of paint craters by modern analytical methods

81

Part II Coating failure analysis centre of the crater is possible. Comparing the chemical composition of the paint crater centre and the paint surface outside the crater provides the chemical differences that may have caused the cratering. If the microscopic inspection of the crater hints at burst paint bubbles, a microanalysis of the inside walls of the crater might produce residues of substances that may have caused degassing. Finally, a cross-cut of the crater zone can be performed to answer the following questions: – Is there an inclusion particle in the paint system? – How many layers have been applied and where in a multi-layer system is the origin of the crater? The sample dictates whether a simple scalpel cut, or a microtome section is enough, or whether a metallographic cross section is needed. The following chapter explains the most common causes of cratering problems and how they can be identified using modern analytical tools.

4.3.1 Cratering caused by contamination of the paint material The whole production chain of painted parts, starting from the production of raw materials and ending with the painting process, offers plenty of scope for contamination. Internal quality control should discover most of the contaminants of raw materials before a paint is delivered to the customer. Aspects of quality control are discussed in Part III of this book in detail. At various steps in paint production, the product can be contaminated by – machines and tools – storage containers – humans – aerosols – transportation – storage conditions Sometimes small and hidden details can cause a severe crater issue. The author is aware of a big failure issue with a powder coating caused by a siliconized sewing yarn that has been used in a production plant for big bags in India. The big bags have been produced for the European market and contaminated tons of coating powder in Germany with silicone. As the coating powders were used for large scale industrial coating a big crater failure issue resulted from this apparently irrelevant detail. Where cratering is caused by external contaminants, such as dust and dried foreign paint, it is obvious that only micro-analytical techniques promise success. In most cases, this sort of crater is caused by very small solid particles. This means that there is a very low quantity of a substance or a mixture of substances that must be identified and that it is concentrated in a microscopic area of about a few microns. As demonstrated before, the micro-analytical methods of TOF-SIMS, micro-ATR and SEM-EDS are suitable for this task. However, there are a few practical considerations that narrow the choice of technique. As far as TOF-SIMS is concerned, its high surface sensitivity is a drawback for this special analytical task even though it permits microanalysis. If, for example, a dust particle has fallen into the wet paint and caused a crater (see Figure II.51), it might be covered with a thin layer of paint or surface-active paint additives. This “protective coating” on the particle renders the particle “invisible” to TOF-SIMS.

82

Paint failures and their analytical approach Thus, SEM/EDS is more suitable in this case as it delivers images of the topography and chemical composition as well as elemental information about the composition of the dust particle (see Figure II.51). Elemental analysis of the inclusion particle in the centre of the paint crater (Figure II.52) shows that it consists of aluminium, silicon and calcium as well as trace components. It can be concluded, that fillers from a foreign paint have been introduced into this clear coat due to improper handling. This conclusion is confirmed by the reference analysis of the paint, which shows that it does not contain these elements on a regular basis (Figure II.53). After the cratering substance has been detected and identified, the production steps can be sampled individually. The sampling procedure must be selected depending on the analytical method and differs for TOF-SIMS, scanning electron microscopy (SEM) or infrared spectroscopy (ATR-FT-IR). Figure II.51: SEM/BSE image of a clear coat paint Abraded particles or lubricants for machines crater centre with particle inclusions and equipment can easily be sampled by wiping the target areas. For this purpose, a clean paper tissue or paper filter should be used which can be analysed by TOF-SIMS or ATR-FT-IR for organic substances or SEM/EDS to identify inorganic material, without any further sample preparation. Wipe samples are also suitable for containers and tools, such as stirring devices. Aerosols or dusts are collected by “passive sampling” in which cleaned target material, such as aluminium and silver, is exposed for a defined period at certain positions. The aerosol or dust precipitates on the target material and is subse- Figure II.52: EDS spectrum of a crater centre with particle inclusion quently analysed by optical light microscopy, mass spectrometry (TOF-SIMS) and scanning electron microscopy (SEM). The next step is to inspect the paint shop. As well as during paint production, each machine, container or transportation device, such as a conveyor belt, are contamination sources that must be inspected when cratering problems occur. With conveyor belts, for example, there have been numerous cases of paint cratering due to the use of fluorocarbon lubricants to grease the transportation device or of fluoro Figure II.53: EDS reference spectrum of the paint outside the crater polymers to coat the belt.

83

Part II Coating failure analysis Table II.7: Source and type of contaminants during paint production (sampling and detection methods) Source of contamination

Type of contamination

Sampling technique

Analytical technique

Machines and equipment

abrasion

wipe sample sample of product or intermediate products filter sample

SEM/EDS ATR/FT-IR IR microscopy

lubricants

wipe sample sample from oil separators sample from tubes

TOF-SIMS ATR-FT-IR

Production and transportation containers

paint residues residues of cleaning agents release agents

wipe sample

TOF-SIMS ATR-FT-IR

Ventilation/ air conditioner

aerosol dust

aerosol targets [10]

TOF-SIMS SEM/EDS

Transportation

aerosol dust

aerosol targets

TOF-SIMS SEM/EDS

Storage

inside coating of containers additives migrating from seals and filler necks

sample of stored product SEM/EDS sample from inside ATR/FT-IR coating of container IR microscopy

The sampling procedures listed in Table II.7 therefore apply to paint shops, too. The following areas must be sampled and analysed step by step: – paint delivery – storage – preparation (stirring devices, pumps, containers) – painting (robots, tubes, spray guns etc.) – paint drying (machines, oven, transportation devices) Besides machines and equipment, there is also the “human factor” to be taken into consideration when it comes to paint defects. Many failures are attributable to wrong treatment of the paint or the raw substrate, inappropriate storage or faulty operation of painting installations.

4.3.2 Craters and pinholes caused by substrate contaminants The ways how workpieces can be contaminated before coating are manifold. No matter if one talks about polymer parts, metal products, wood or paper, they all went through different production steps, transportation, storing, handling processes. Each step leaves chemical traces on the surface. Example: A powder coating contains deep pinholes as shown in Figure II.54. As the majority of the pinholes penetrate the coating down to the substrate, surface contamination had been assumed. The procedure for detecting and identifying this contaminant is as follows:The TOF-SIMS spectrum of the centre of the crater (Figure II.55) contains characteristic peaks of a release agent (polydimethylsiloxane) and a fluorocarbon lubricant (perfluorinated polyether, typical trade names: “Krytox” or “Barrierta”). Both are substances which can cause adhesion problems as well as pinholes and craters. This information raises the question how this contamination found its way into the paint layer. In general, there are several possibilities:

84

Paint failures and their analytical approach – contaminants in the paint material – malfunction of the painting installation – contaminants on the substrate These questions can be answered by reference measurements on: – sample film of the paint on a neutral panel (e.g. aluminium), – random sample from different paint lots – random samples from the painting installation – random samples from raw, unpainted parts As all these samples are random, the sampling method must ensure that they are representative. The results of this analytical series help to narrow the problem and link it to a certain production line. In this case, the reference measurements showed that the root cause of the cratering was a punching machine which was emitting an oil aerosol that left microscopic oil droplets on the surface. After the machine was cleaned thoroughly, the cratering stopped.

a)

b)

Figure II.54:  a) Optical light microscopy image and b) EFI-3D image of a powder coating pinhole

Figure II.55: TOF-SIMS spectrum of the centre of a crater on a painted metal panel

85

Part II Coating failure analysis

4.3.3 Craters caused by paint additive agglomeration Not only foreign particles and wetting disturbing substances can cause craters. Insufficient mixing of paint components can also lead to massive craters.

Example 1: KTL flat craters due to additive enrichment Problem definition

Figure II.56: Flat crater in an anticorrosion paint

When coating a steel substrate with an anticorrosive paint, craters occurred which were very flat and had the shape of a depression.

Investigation method

Figure II.57: Light microscopic image of the cross section through the crater in the anti-corrosion coating (fluorescence/reflected light)

First, an investigation of the paint surface in the crater and the undisturbed paint surface was carried out using the TOF-SIMS method (not displayed). This analysis revealed no differences in the chemical composition of the uppermost coating layer. Thus, foreign contaminations could be excluded. If the cause of the crater is not detectable at the surface, it might be located in a deeper layer. The analytical techniques that have been discussed cannot sufficiently penetrate a paint layer system of 10 to 100 μm. A metallographic cross-cut of the paint layer system helps to identify the cause. This cross-cut offers various analytical techniques to be used. Optical light and fluorescence microscopy of the cut (Figure II.57), for instance, could be followed by SEM/EDS.

Evaluation

Figure II.58: Scanning electron microscope image of the metallographic cross section through the crater in the corrosion protection paint (BSE, 20 kV)

86

The optical microscopy shows that the centre of the crater is inhomogeneous. The additional SEM/EDS analysis reveals the nature of this inhomogeneity. The BSE image (Figure II.58) shows that there is a round particle inclusion

Paint failures and their analytical approach beneath the paint surface in the centre of the crater. The material contrast indicates chemical differences compared with the surrounding paint. As discussed before, EDS delivers the chemical information about the defect zone and the paint (Figure II.59). Comparison of the elemental compositions of the particle inclusion and the paint reveals that the inclusion is qualitatively of the same composition as the paint, but that there is significant agglomeration of tin compounds. In the absence of any knowledge of the paint’s composition, this information cannot be evaluated. However, the knowledge that organo-tin compounds are added as catalysts adds the essential information to the analytical result: local agglomeration of the catalyst is causing the cratering. Not only catalysts but also levelling agents and fillers or surfacers can cause craters if they are not dissolved properly.

a)

b)

Figure II.59: EDX spectrum a) of the electrocoat and b) of the particle

a)

b)

c)

d)

Figure II.60: Infrared microscopy analysis of a powder coating crater: the false-colour display shows on a scale from dark red (low intensity) to white (high intensity) the intensity distribution of a characteristic signal of the filler aluminium hydroxide (b) and of two polyesters of the paint binder (c and d) in the measuring field over the coating crater (a)

87

Part II Coating failure analysis

Example 2: Powder coating craters caused by a filler Problem definition In a white powder coating system, flat craters are formed whose edge and centre differ slightly in colour from the surrounding coating.

Examination method The crater (see marking in Figure II.60) was examined with the infrared microscope (IRM). A raster scan grid was placed over the marked area of the crater and an analysis was performed line by line, point by point (IRM mapping). The spectra of the individual components were extracted from the collected data set and characteristic signals have been integrated. The result is displayed in a false-colour image on a scale from dark red (low intensity = little substance) to white (high intensity = much substance).

Evaluation The false-colour representations in Figure II.60 show the distribution of aluminium hydroxide (b), polyester 1 (c) and polyester 2 (d) in the area of the crater. It is obvious that in the centre of the crater a polyester component of the paint is concentrated. The darker crater rim however is caused by an agglomeration of the filler aluminium hydroxide.

4.3.4 Cratering caused by the application conditions In Chapter 4.1.3 we already mentioned the impact of application conditions on the coating result: – inappropriate location or construction of paint shops – incorrect mixing of 2K and 3K systems – insufficient cleaning of the spray booth and the robot – wrong choice of paint solvents – incorrect operation of the paint shop are some of the prominent failure reasons. But paint craters/wetting problems can also be caused by airborne interfering substances. A distinction must be made between – the ambient air outside the spray booth – the cabin air and – the paint shop ventilation (steering air, atomizing air, etc.). During transport, storage and equipping of skids in non-clean room areas, aerosols and dust can deposit on components. Although these are not optically visible, they can still cause craters and/ or wetting problems. The cabin air can be loaded with PWIS e.g. by humidification systems, suction of contaminated outside air, outgassing from operating materials or spray mist from water walls. Far more common, however, are contaminants that are introduced into the cabin by the compressed air. Possible causes include design faults in the compressed air system and improper maintenance. This type of contamination is particularly “insidious” because even improper maintenance, when a few μg of PWIS get into the compressed air flow, can cause irreproducible craters for months. As a rule, a PWIS introduced into the compressed air stream is not continuously transported and discharged but is deposited in dead spots of the compressed air system and is

88

Paint failures and their analytical approach then discharged at irregular intervals, which in this case means that it is mixed with the paint material via the painting air and sprayed onto the components. The following example, focusses on a failure type caused by the contamination of the air supply (Figure II.61).

Investigation procedure Since certain PWIS can cause craters already in concentrations of one molecule layer, the most sensitive surface method (=> TOF-SIMS) should generally be preferred. This is especially true if already strong, educated guesses hint at a compressed air contamination. However, in many cases very good results can also be achieved with infrared microscopy, as in the example in Figure II.61.

Evaluation Figure II.61 shows the infrared microscopy analysis of a wetting deficiency in a metallic paint on a polyurethane. The ATR measuring crystal was guided over the crater in a two-dimensional scanning grid of 8x8 measuring points and a spectrum was recorded at each measuring point. The false-colour display illustrates the evaluation of this data set with respect to a characteristic signal line at 729 cm-1 which is typical for aliphatic hydrocarbons, which are the main constituents of paraffins, waxes and mineral oils. This distribution correlates with the location of the contaminant optically detectable in the crater. Due to their non-polar properties, aliphatic hydrocarbons have a highly detrimental effect on wetting, especially when it comes to a (polar) water-based coating system. Since this analysis clearly shows the chemical reason for the wetting disturbance, it is necessary to look for the origin of these aliphatic hydrocarbons by comparative investigations in the operating process. In the present case the compressed air (more precisely the compressor oil) could be identified as the source.

Figure II.61: Two-dimensional infrared microscopy ("mapping") analysis of a wetting deficiency in a water-based metallic paint on a polymer substrate (top right) The false-colour representation at the top left shows the intensity distribution of a characteristic signal line of an aliphatic hydrocarbon (=> mineral oil) in a colour scale from white (high concentration) to dark blue (low concentration)

89

Part II Coating failure analysis

4.4 Bubbles and blisters A bubble is a round cavity in a paint layer which is empty (except for vapour or air) whereas a blister is a dome like cavity which contains substances like moisture, liquids, salts or other particles. The reasons for the formation of paint bubbles and blisters are as numerous as for cratering. Sometimes both problems are connected if a bubble of the paint bursts and forms a paint crater. Here are several reasons: – rust or water on the workpiece – salt residues – high heating rates – layer thickness too high – microscopic cracks in the workpiece – foaming in the paint – (corrosion) creep of a paint layer – degassing of the substrate – infiltration and sub-corrosion of a paint layer – outgassing from the substrate or adjacent coatings – agglomeration of low molecular weight polymer components on the surface The recommended first step of the examination is to get an overview of the topography by optical light and fluorescence microscopy. If the focusing depth maybe low for a topographic image like this, it is a good idea to enhance it by EFI method2. The other option, even for low resolutions, is SEM/EDS which has a better focusing depth and allows for additional chemical analysis. But often the cause of a bubble or blister is not visible upon first macroscopic and microscopic inspection as shown in Figure II.63 (left). From the top view you would never real-

2 EFI= Extended focus imaging see Chapter 1.1, Part IV

Figure II.62: SEM/BSE image of a metallographic cross-cut of a primer paint blister (caused by a substrate failure and a particle inclusion in the basecoat)

90

Paint failures and their analytical approach

Figure II.63: Bubbles in a single-layer coating on PC-ABS (left) and metallographic cross section through the bubble (caused by a crack in the substrate)

Figure II.64: EFI light microscopy image of a bubble cut parallel to the surface (left) a foreign particle is visible at the bottom of the bubble (right)

Figure II.65: SEM/BSE image of a cross section through a basecoat paint bubble in a multilayer coating

91

Part II Coating failure analysis ize that the root cause is a crack in the polymer substrate, which has been revealed by a metallographic cross section (right).

Preparation Therefore, the bubble/blister must be opened to see if there are residues of substances that hint at the possible cause of the paint defect. This can be achieved by cutting the bubble parallel to the surface under optical binocular microscope control (see Figure II.64).

Figure II.66: Overview picture of blisters in a black coating after a weathering test (left) and optical microscopy image of the bottom of the blister showing a crystalline residue (right)

Figure II.67: Infrared microscopy spectrum of the crystalline residue inside the blister (green) and the reference spectrum of the bottom of the bubble (black)

92

Paint failures and their analytical approach Another option is a metallographic cross-cut combined with scanning electron microscopy or infrared microscopy. Thus, a selected area of the bubble can be analysed directly with respect to the chemical composition. By this proceeding any foreign substances detected in a bubble can then be subjected directly to a microanalysis. Figure II.65 show the SEM-BSE image of a cross section through the centre of a bubble in a three-layer coating. The SEM-image not only reveals that the failure is located in the middle of the basecoat layer but also that it is empty (bubble). That hints at degassing of basecoat components. If a bubble is opened by a scalpel cut parallel to the coating surface like the one in Figure II.64 and the microscopic inspection shows that there are particles inside the microanalysis methods SEM-EDS or infrared microscopy are recommended to determine the chemical composition.

Example The infrared microscopy spectrum of the crystalline particle inside the blister of a coating that appeared after a weathering test (Figure II.66) shows characteristic peaks of polyamide as well as the spectrum of bottom of the bubble (Figure II.67). But some additional peaks at 1300, 1275, 1248 and 976 cm-1 qualify the particle as crystalline polyamide [20]. The root cause for this blister lies in the insufficient treatment of the polymer (raw material): It is widely known that polyamide contains low molecular weight (short chain) oligomers which tend to migrate to the surface under the influence of water-vapour and temperature. On the surface these oligomers like to crystallize looking like salt crystals. If this happens before the polymer is coated, the PA crystals (which are hygroscopic) cause blisters during the paint drying process. To avoid this failure, polyamide should be conditioned and dried before it is injection moulded.

4.5 Discolouration Slight colour changes, like e.g. the yellowing of a white paint or coating, are sometimes very frustrating when it comes to analyse for the root causes. The reason is that our eyes are very sensitive for colour changes as long as the discoloured area is close to the not affected area. The possible reasons for a discolouration are diverse: – pigment particle size [21] – pigment particle structure – chemical changes of the binder or additives (oxidation, photodegradation) [22] [23] – unexpected chemical side reactions between paint components and/or with environmental substances – contaminations – migration effects What makes the analytical approach challenging is the fact that a very low amount of chromophoric groups (sometimes below the detection limit of any available analytical method) is sufficient to give a significant visual effect. Therefore, there are two concepts that must be discussed with respect to the choice of instruments for the investigation of such an issue. The limit of detection is the lowest amount of analyte (here chromophore) in a sample which can be detected (not necessarily quantitated) as an exact value. This LoD is not a fixed value for each method but is dependent on the sample composition. The lowest detection limit can be realized by TOF-SIMS (about fractions of one molecular layer). On the other hand, the high surface sensitivity of TOF-SIMS can disturb the analytical goal if the yellowing takes place in layers underneath the uppermost molecular layers. If for example the yellowing of a coating is analysed

93

Part II Coating failure analysis which contains modified polydimethylsiloxane surfacers (which do not contribute to the discolouration), only the latter are detectable if the sample is analysed as received because these surfacers cover the uppermost coating layer. The detection limit is also not a fixed value. The standard procedure of investigating a discoloured material is to compare the failed area to an area without discolouration by spectroscopic methods. Very often there is not only one single difference between these two areas but a couple of them which makes it necessary to decide which one is responsible for the discolouration. Look at the sample shown in Figure II.68: The white structured coating exhibits a significant yellowing on the left-hand side. The task to find the root cause for this effect focusses on the question “what is different comparing the discoloured area to a not affected area?”. As this difference might be caused by a very, very low amount of substance, the analytical approach has to focus on methods that have a low detection limit. The following procedures can be considered: – Extraction and mass spectrometric analysis by GC-MS. The drawback of this procedure is that the solvent which will safely extract the chromophores is unknown. The only key to this parameter is experience and trial and error. Once the appropriate solvent has been found, so much of the low concentrated chromophores might have been lost for the trials, that there might be nothing left for characterisation. In addition to that, it is not sure that only the chromophores will be extracted, which makes it difficult to evaluate the spectra. A mixture of extracted substance might occur in the chromatogram which make it difficult to determine which of the substances extracted from the failed area is responsible for the discolouration. And due to the low concentration of the eluate there is only one shot for GC-MS screening with one column and one or two temperature programs. And that’s it. If the substance does not pass the column because it was not appropriate for the target molecule, one faces the risk to have nothing in hand. – Scanning electron microscopy (SEM): By scanning electron microscopy the discoloured area can be compared to a not affected area visually making use of SE and BSE images. In addition to that the elemental composition of both areas can be compared. This is useful for pigment and filler characterization, for particulate inorganic contaminations. Particle size can be compared as well as particle structure. What cannot be found (by design of the method) are organic substances. This means the above mentioned oxidation/degradation processes of the polymer backbone of paint binders are not accessible by SEM. But let’s now have a look at the implications of the SEM analysis applied for yellowing effects: As mentioned before human eyes are very sensitive and they average the colour impression over a large area. If for example the yellowing of a white paint is caused by contaminations by very small iron oxide particles, a very low concentration is sufficient to give the visual impression of a yellow surface. If such an area is analysed by Figure II.68: White coating showing partial yellowing

94

Paint failures and their analytical approach SEM/EDS making an overview analysis of a larger area (which means a few mm2 in terms of SEM/EDS analysis) it is very likely that the iron will be below the detection limit of SEM/EDS. In fact, it is necessary to go into detail pictures with high magnification and search manually in order to find the foreign particles. – TOF-SIMS: With respect to low concentrated organic substances (contaminations) or changes of the polymer backbone of the binder, the TOF-SIMS method is the best choice because it has a very low detection limit and delivers not only elemental spectra like SEM but also structural information. The challenge is to make a representative measurement. The standard area size of a TOF-SIMS measurement is a few mm2 which of course cannot be representative if you analyse for example the coating on a door. So, a significant amount of measurements must be averaged to ensure representativeness. – Infrared spectroscopy and microscopy: The infrared method is a versatile tool to detect oxidation or photo-degradation processes because it delivers molecular information. On the other hand, the poor detection limit can be a drawback. Whether this method is suitable for the target sample depends on the complexity of the chemical composition. The yellowing of a polyurethane clear coat for example caused by oxidation of the central methylene group to a diquinone imide can easily be detected by infrared spectroscopy [23]. If the coating is more complex (e.g. a filled primer with barium sulphate, calcium carbonate, talc) the overlay of characteristic signals of all the paint ingredients makes it sometimes difficult to identify small changes of the binder polymer backbone. This in fact depends on the sample and therefore it is not straightforward to recommend the infrared spectroscopy for each kind of discolouration issue. Back to the sample shown in Figure II.68. For this sample a migration of aromatic substances from the polymer (a styrene polymer) into the coating was assumed. To prove or falsify this the detec-

Figure II.69: ATR-FT-IR spectrum of the discoloured coating (green) and the coating without discolouration (red)

95

Part II Coating failure analysis

Table II.8:  Possible reasons for discolouration and their analytical approach Cause Analytical target

Analytical method

Coating production and formulation faults: Oxidation of binder due to storage conditions

functional groups like e.g. ATR-FT-IR, TOF-SIMS carbonyl groups

Wrong raw material like e.g. aromatic binders instead of aliphatic: Unclean equipment

functional groups, contaminations (e.g. iron)

ATR-FT-IR, TOF-SIMS SEM-EDS

Wrong stabilizer concentration

stabilizers

TOF-SIMS

Too high temperatures during curing

functional groups

ATR-FT-IR

Contamination by foreign material

functional groups and/or contaminations

ATR-FT-IR, TOF-SIMS, SEM-EDS

Unsuitable stabilizers Coating processing failures:

tion of conjugated double bonds is necessary which can exclusively be achieved by infrared spectroscopy. Therefore, samples of the discoloured area and the area without discolouration have been compared by ATR-FT-IR measurements. Indeed, the comparison of both spectra reveal differences (Figure II.69). At 2264 cm-1 the signal of free isocyanate shows that the yellowed surface seems to exhibit a lower amount of free (unreacted) isocyanate which means that there seems to be a higher level of cross linking. At 3674 cm-1 and 1009 cm-1 characteristic peaks of talc are significantly higher in the discoloured area. But no conjugated aromatic group has been found. Therefore, the question is: can a lower amount of free isocyanate and/or a higher amount of talc cause the observed colour change and how can talc be increased in a cured coating film. As the answer to these questions is not straightforward the point measurement of the surface (the sampled size in diamond ATR is 2x2 mm) has been replaced by an infrared microscopy line-scan of 87 measurements over the borderline between discoloured and not discoloured area to achieve more statistical significance. Figure II.70 displays the results of this measurement. It shows that the intensity of the isocyanate signal varies a lot in both areas. But averaging the signals over the whole line-scan there is indeed a slight Figure II.70: Result of an infrared microscopy line-scan increase of free unreacted isocyanate over the borderline between discoloured area the false in the not affected area. Which means colour contour image displays the intensity of a that either the isocyanate/polyester characteristic signal of the NCO group of free isocyanate -1 crosslinking to polyurethane has hapat 2264 cm ), colour coded between dark blue pened to a higher extent in the dis(low intensity) and white (high intensity)

96

Paint failures and their analytical approach coloured area, or unexpected post curing side reactions of the isocyanate took place. So besides statistical effects that affected the initial point measurements due to very inhomogeneous coating composition, there is an affect that correlates with the discolouration.

4.6 Hazes and stains A special type of discolouration is caused by hazes, stains or deposits on paint surfaces. Sometimes these failure types are hard to distinguish. These faults can be caused by: – Inhomogeneous coating material – Migration of polymer compounds to the surface – Contaminations – Side reactions of the paint – Inappropriate storing and handling of the coating material – Wrong paint quality – Mixing errors – Noncompliance of application parameters – Topcoat degradation by aggressive solvents or cleaners – Migration of components of spatula, primer, basecoat into the topcoat – Residues of a cleaning or polishing process Many aspects that have already been discussed in Chapter 4.2.2, Part II apply here too. Migration effects from the substrate, which lead to adhesion problems of the coating and are e.g. due to insufficient cleaning or pretreatment, can also be responsible for the appearance of stains and deposits on the cured coating surface. Examples are low molecular weight components of polyamides or polyesters that migrate through paint layers to the visible surface. Migration of additives or binder components from one of the paint layers to the very surface is also a common source of stains. Figure II.71 shows a black coating on a polymer substrate which exhibits white stains caused by crystalline deposits on the surface.

Investigation procedure For optically visible deposits like crystalline particles or droplets exhibiting a

Figure II.71: Painted polymer surface with white deposits (top) and ATR spectra of the surface of the PU paint with white deposits, the paint surface without deposits, and a reference sample

97

Part II Coating failure analysis thickness of a few microns or less the ATR is the preferable tool for identification. The method has an information depth of a few microns. For good quality results, it is important that the paint surface with the stains is brought into intense optical contact with the ATR crystal and that the selected area is representative of the whole sample. At least two measurements are necessary: one of the surface with precipitation and one of a reference area. That can be either an area of the coating surface without precipitation or (if not available) a section of the coating layer). Figure II.71 shows the results of such an ATR-FT-IR investigation of a hydro soft coating surface with an unknown crystalline deposit. By comparing an ATR spectrum of the paint surface in the area of the deposit with an ATR spectrum of the paint surface in a sound area and a reference spectrum of a paint additive, the deposit has been identified as an additive that migrated to the surface of the paint system. A further preparative possibility is given if the deposit is optically visible and can be removed from the paint surface with a small drop of solvent. This drop (with the dissolved deposit) is transferred directly to the ATR crystal and the solvent is allowed to evaporate. Then the spectrum of the residue dried on the surface of the crystal is recorded. With this method, extremely small amounts of substance are sufficient to obtain a proper spectrum. As background the spectrum of the pure ATR crystal is measured. In this way, reasonable spectra with substance quantities of a few micrograms or less are possible. Fortunately, there are also portable and/or handheld infrared spectrometer available that can be used directly in the field (see also Chapter 4 Part IV of this book). They offer the choice of either measuring in the non-contact mode via the physical process of external reflection spectroscopy or

Figure II.72: Comparison of the infrared reflection spectra of a high gloss yacht coating showing a haze (red) compared to a spectrum of a reference area of the hull without haze (blue)

98

Paint failures and their analytical approach contacting the surface by an ATR module. This spectroscopic measurement should exclusively be done by skilled and educated spectroscopists. The performance of the measurement is easy to learn but the physical background of external specular reflection und attenuated total reflection is complex and, depending on the measurement conditions and the sample condition, there can be physical effects in the spectra (e.g. “Reststrahlenbanden”) which have to be understood to be distinguished from real sample absorption features that provide the desired information (Part IV Chapter 4.2). In addition to that a deep understanding of the physical background is needed to decide on the site which surface can be analysed by external reflection and which surface must be analysed by attenuated total reflection method. By infrared reflection spectroscopy the chemical composition of the uppermost 1 to 2 μm of the coating can be analysed with respect to: – binder type (e.g. polyurethane, acrylate, epoxy resin) – binder/hardener/accelerator ratio of 2K or 3K-systems – hardener side reaction products – contaminations – migration effects – paint type identification The key question of the infrared spectroscopy analysis is “how does the discoloured paint differ from the unaffected coating?”

Example: Polyurethane high gloss coating showing a haze Figure II.72 shows the infrared reflection spectra of a high gloss polyurethane topcoat consisting of a polyester binder (1732 cm-1) and an isocyanate hardener. It was measured in an area showing a strong haze effect and the reference spectrum of the topcoat without haze was collected far from the failure site. By comparison it is obvious that the chemical composition of the polyurethane is different with respect to polyester binder and the hardener (1684 cm-1) ratio. The intensity of the signals correlates with the concentration in the mixed and cured polyurethane. As the signal of the polyester at 1732cm-1 is significantly higher for the haze area compared to the reference sample without haze, it can be concluded that the haze area contains more polyester binder. Considering that the reference area displays the correct mix-

Figure II.73: Paint surface showing white stains (top) and scanning electron microscopy image of the surface (bottom)

99

Part II Coating failure analysis ing ratio of binder and hardener the final conclusion is that there is an overdose of polyester which has no isocyanate reaction partner and thus remains uncured in the coating film. So, the root cause of this hazing effect is a mixing ratio failure i.e. a human error. Sometimes hazing effects are caused by ultrathin contamination layers that are too thin to be detected by infrared reflection spectroscopy. Then the infrared reflection delivers the information that there is no contamination. Unfortunately, this does not mean that there is really no contamination. But it means there is no contamination within the detection limit of the method. For these ultrathin layers the low detection limit of the TOF-SIMS method is needed. However, a TOF-SIMS analysis which is a vacuum method cannot be done on the site. This requires a wipe sampling combined with a laboratory measurement. Although this example seems to be very simple with respect to the analytical procedure, in fact detecting the cause of the formation of veils on paint surfaces is sometimes very complex. The fatal thing for the analyst is the discrepancy between the optical visibility of a change and its analytical detectability. In addition to the migration effects of components of the paint or the substrate mentioned above, this can also be purely physical effects such as an accumulation of optical scattering centres such as topographic defects on or in the surface. Figure II.73 shows the image of a dark painted surface with a white veil or, if you like, stains. A chemical comparison of the area with veil and an area without veil using TOF-SIMS and ATRFT-IR (not displayed here) did not exhibit differences. This proves, that in fact no chemical reason for the failure is measurable. Scanning electron microscopic analysis finally revealed that the veil was in fact an accumulation of scattering centres in the form of craters in the paint, probably caused by gas bubbles (solvent pops) of the paint.

Figure II.74: Optical light microscopy image (DIC = differential interference contrast) of a surface with a veil (consisting of discrete particles with inhomogeneous distribution)

100

Paint failures and their analytical approach A colour inhomogeneity like a veil can also be caused by very few colour centres in a homogeneous matrix. If the population of these colour centres is too low, they are not accessible by infrared spectroscopy. This means that a technique is needed that provides the necessary detection sensitivity. The TOF-SIMS method is recommended for detecting small quantities. But there are some drawbacks to keep in mind with this special problem. The high detection and surface sensitivity are actually an advantage of this method in case of discoloration, haze formation and deposits. But for the same reasons it can be an obstacle. Let's take as an example the deposit shown in Figure II.74. It shows a light microscopy image of a surface that exhibits a veil upon visual inspection. The light microscopy image reveals that this appearance is caused by discrete particles with an inhomogeneous distribution. The total amount of substance that causes the visual effect is very low. When applying the TOF-SIMS method in this case it is important to be aware of the measurement field. On a regular basis an area of 100x100 or 200x200 μm is selected to collect a sum spectrum. If a visual haze or deposit consists of an accumulation of discrete particles inhomogeneously distributed like visible in Figure II.74, the measuring field contains only a few particles. More than 95 % of the information in the spectrum must be attributed to the “gaps” between the particles. Thus, the actual signal of the substance(s) causing the haze or veil may be very small in the spectrum and is superimposed by the signals of the substrate. For the spectrum evaluation such a spectrum poses difficulties to assign the real interfering substance. Haze is a coating appearance issue which happens with high gloss coatings. From the physical point of view, it is an optical effect of light diffraction and scattering that is caused by very small discontinuities in the paint or on the paint surface. So, the analytical aim is to characterize the nature of these scattering centres.

Figure II.75: Photographic image of a yacht coating exhibiting a multi-coloured haze

101

Part II Coating failure analysis This can happen in the uppermost molecular layers of the topcoat, in the topcoat bulk and at the interface to underlying coating layers. So, the main questions to be answered are – What kind of discontinuities cause the optical effect (bubbles, particles, droplets)? – Where are these spots located (surface, bulk, interface)? – How does the chemical composition of the discoloured layer differ from a coating area without the haze? – So, what can be done on the site to meet these requirements without cutting out a reasonable piece of the coating? The initial analysis step is, as already mentioned, the microscopic inspection. We already stated that the limit of this method is the lateral resolution and the depth of information. The colours showing in Figure II.75 hint at very small scatterers in the order of a few hundred nm or a few μm. In this dimension it cannot be clearly distinguished by optical methods if the scatterer is a bubble (filled with a gas), a particle or a droplet. A scanning electron microscope (SEM) would be needed to magnify and resolve these failures. However, this kind of instrument is not portable. So, without destructive sampling of a piece of at least 1x1 cm of the coating and subsequent SEM analysis or ATR-FT-IR spectroscopy there is no chance to finally evaluate the question which kind of scatterer causes the optical effect.

4.7 Paint spots Paint spots, along with craters and adhesion problems, are the most common types of failure. There are as many reasons for this kind of failure as there are for cratering: – inclusions of foreign particles – agglomerations of paint components, such as fillers and structure additives – paint spray caused by wrong application – substrate failure – residues from pretreatment processes

First inspection The evaluation of the failure spot by optical light microscopy inspection can reveal if there are particles or fibres sticking out of the speck, whether the material of the speck has a different colour and structure than the surrounding paint layer and what shape the specks have. This allows initial conclusions to be drawn as to the direction in which the search for the cause should go. An important aspect that determines the further procedure is the location of the fault in the paint system. If the light microscopic examination proves that the inclusion is located immediately below the surface or even protrudes from it, then an analysis can be carried out directly from the top (SEM/EDS or infrared microscopy) without further preparation. If the defect is more likely located in a deeper layer (e.g. a speck visible in the clear coat but caused by an inclusion in the basecoat) then an analysis as received from the top is not possible without further preparation. However, the light microscopic evaluation sometimes does not provide absolute and unambiguous criteria and requires a bit of experience. Unfortunately, sometimes, the size and shape of a paint spot provide no indication as to its origin. The paint spot shown in Figure II.76, for example, seems to have been caused by a fibre. Yet ATR-FT-IR analysis shows that it consists of agglomerations of a paint additive.

102

Paint failures and their analytical approach As this example illustrates, an optical examination which produces a result of the kind “it looks like a...” is very often misleading. Such a preliminary evaluation sometimes has to be revised. Optical inspection does not always tell the story of the paint spot, but instead furnishes hints as to the best technique for analysing the spots. There is no standard procedure for analysing paint spots like “If it looks like A, use method B and you will be successful”. Actually, a combination of techniques will deliver the best results. The choice of method is due to the experienced analyst and his intuition.

Preparation If the pre-inspection hints at an inclusion in a layer underneath the very surface and the defect is not accessible analytically from the top (or the results are not satisfying), the next step is to cut or grind through the defect area to make the actual inclusion accessible for analysis. There are several options regarding the choice of the sectional planes. The simplest preparation is a scalpel cut parallel to the surface through the spot. This method has the advantage that it is quick and requires no equipment. However, there is only one attempt. If this fails, this area is lost for analysis. In addition, what is cut off with the scalpel is often lost although it may still contain important information. A professionally performed metallographic cross section or microtome section (see Figure II.80) through the centre of the specimen perpendicular to the lacquer surface is a better choice,

Figure II.76: Optical light microscopy image of a paint spot of a metallic polymer paint (left) and ATR-FT-IR spectrum of the spot compared with a reference measurement of a paint additive

a)

b)

Figure II.77: Optical light microscopy image (EFI) of a paint spot (a) 3D view and (b) plan view

103

Part II Coating failure analysis

Figure II.78: Flowchart of paint spot analysis

Figure II.79: Schematic representation of the sectional view of a cross section, wedge section and parallel section through a defect in a multi-layer system

104

Paint failures and their analytical approach because it does not produce as many artefacts as the scalpel blade. In some cases, it may also be useful to make a wedge cut (see Figure II.79) through the entire coating system either with the microtome or as a micro-section after embedding in resin. Optical light microscopy or, better, scanning electron microscopy, of the cross section can answer the following questions: – Is a substrate failure causing the visible spot? – Are there agglomerations of paint components? – Is there an inclusion particle of foreign material? – Are there cavities visible in the paint spot? Although the choice of the correct method or methods cannot always and exclusively be made according to a fixed scheme (see Figure II.78), the boundary conditions of the analysis (very small amount of substance in a microscopically small sample area) already limit the choice of methods. As microanalysis methods must be used, the choice is limited to SEM/EDS, infrared microscopy, Raman microscopy and

Figure II.80: Fluorescence microscopy image of a cross section through the centre of a paint spot showing a particle inclusion

Figure II.81: Scanning electron microscopy image (BSE) of a metallographic cross-cut through the centre of a paint spot in a paint layer on a phosphated steel

Figure II.82: EDS spectrum of the surrounding paint (left) and EDS spectrum of the particle inclusion (right)

105

Part II Coating failure analysis Table II.9:  Summary of EDS results EDS spectrum EDS spectrum of the Element of the paint inclusion particle Carbon

+

+

Oxygen

+

+

Iron

+

+

Sulphur

+

-

Chlorine

+

-

Tin

+

-

Zinc

-

+

TOF-SIMS. In many cases a combination of several methods will be necessary to find the cause. The flow chart in Figure II.78 illustrates a basic procedure that has proven to be particularly effective.

Example: Paint spot on a polymer substrate

Figure II.80 displays the fluorescence microscopic image of a cross section through a paint spot on a polymer subAluminium + + strate. What does this picture tell? Silicon + There is an almost round inclusion of Phosphorus + 15 µm. Thus, even micron-sized particles or fibres in a sample residue can be characterised selectively. Figure III.20 demonstrates this with a screen residue containing differently coloured and shaped particles. The infrared microscopy point analysis allows for the selective analysis of each particle under visual control. Even different areas of multi component particles (e.g. abraded coating particles of hobbocks) can be resolved. The absolute detection limit is a few nanograms or micrograms, depending on the type of sample. Nevertheless, ATR microscopy is not a trace-analysis technique. For trace analysis of a sample residue, which is not very common, TOF-SIMS must be chosen.

Figure III.21: Optical light microscopy image of fogging by a binder precipitated on a glass target

129

Part III Quality control and process analysis

2.2 Analysis of fogging residues Environmental and consumer safety reasons have boosted the importance of analysis and detection of gaseous emissions from paints and paint ingredients. In particular, the automotive supplier industry is very sensitive about emissions from paints and polymers. All organic substances and, especially, paints and polymers emit gaseous compounds to a certain extent. This is called “fogging”. This fogging can consist of – solvent residues – monomers and oligomers of the polymer or binder – additives – reaction products and by-products of the paint drying process The quantity of volatile compounds emitted by a solid material depends on the – temperature – age of the material – production process From a chemical point of view, it is almost impossible to produce a “zero emissions” polymer or coating. The question is not if there are any emissions but rather how many emissions can be expected from a material and (more important) what kind of emissions does the consumer have to face. Evaluation of fogging with respect to smell, nuisance and health risks requires chemical characterization of the components. If the level of fogging reaches a critical value, the analysis must be extended to the raw material and intermediate compounds. Expected emissions from paint, polymer part or raw material are simulated by treating the sample as follows: The material is placed in a glass container, which is then closed and heated for

Figure III.22: TOF-SIMS spectrum of a fogging film emitted by a painted automotive polymer panel compared with a reference spectrum of a binder used to produce the paint

130

Quality control of paint production a certain time to a specified temperature (e.g. 80 °C) [8]. The glass container is closed by a glass plate, which is cooled during heating. The gaseous compounds emitted by the sample precipitate on the surface of this glass plate. The film of precipitated substances is analysed photometrically or gravimetrically at regular intervals. The result is a percentage or microgram value which shows the expected degree of fogging from this material. What these analytical data do not tell is the composition of the emissions, knowledge of which is necessary for any evaluation of possible hazards to consumer health and the environment. The infrared spectroscopic and mass spectroscopic analysis of the precipitated film yields the information about the components and serves as a basis for certifying the analysed material in respect of safety data. The TOF-SIMS spectrum of fogging from a painted automotive polymer panel shown in Figure III.22 is compared to a reference TOF-SIMS spectrum of the binder which was used in the making of the paint. This mass-spectroscopic comparison shows that the main component of the precipitation is a trimethylolcaprylic acid ester, which is a major ingredient of the binder. Thus, the fogging can be unambiguously attributed to a raw material. Armed with this knowledge, it is possible to reduce or eliminate the fogging problem by changing the recipe or pretreatment of the binder. Detailed studies of the fogging problem with the aid of modern analytical techniques thus lead to improved quality and that benefits the paint manufacturer, the consumer and the environment. The same is true of raw materials which have been contaminated by unexpected by-products due to production problems or transportation in a contaminated container. As well as TOF-SIMS, ATR-FT-IR can be useful for identifying fogging emissions. This is demonstrated in Figure III.23, where a volatile contaminant emitted by a binder at 90 °C has been identified as sodium lactate.

Figure III.23: ATR-FT-IR spectrum of a fogging film emitted from a binder compared with a database spectrum of sodium lactate

131

Part III Quality control and process analysis

2.3 Quality check of finished and semi-finished products In addition to production control and raw material control during the manufacturing process, the methods discussed here can also be used to examine semi-finished products, finished products and competitive samples (reverse engineering). Reasons for this can be: – A paint already used for a certain time, exhibits unexpected problems with a new batch. It must be checked whether the paint manufacturer has made changes to the formulation that have not been reported. – The difference between the formulation of a paint of manufacturer A and manufacturer B must be investigated.

Figure III.24: Analytical options to identify the components of a lacquer

Figure III.25: ATR-FT-IR spectrum of a white automotive paint measured as wet film (top) and measured as dry film (bottom)

132

Quality control of paint production Table III.3: Overview of the analytical options for the examination of semi-finished and finished products Analytical Compound method Preparation Information Limitations Pigments and fillers

Solvents

Binders

SEM-EDS

none

elemental composition

no information about molecular composition

GC-MS

-

-

not applicable

ATR-FT-IR

separation from binder

molecular identification

low concentrated compounds are not detectable spectral interference with other oxidic compounds

TOF-SIMS

thin film on metal target

molecular identification

spectral interference with surfactants

GC-MS

minimum

molecular identification

problems with water-based paints multiple runs necessary for unknown matrices

TOF-SIMS

-

non volatile contaminants

volatile compounds are not accessible

ATR-FT-IR

separation from binder and fillers

molecular identification

only screening results (sometimes incomplete)

SEM-EDS

-

-

not appropriate for this matrix

TOF-SIMS

thin film on metal target

molecular composition

without pre-separation incomplete results no information about chain length and details of the structure

GC-MS

derivatisation pyrolysis

molecular composition

very complex with unknown matrices

ATR-FT-IR

separation of pigments and fillers

molecular composition

without pre-separation not unambiguous no information about details of the structure

It is important to realize that even the most sophisticated analytical methods are not able to determine the complete qualitative and quantitative composition of a highly complex mixture such as a ready-mixed paint in such a way that it is possible to develop a recipe that corresponds to the original paint. The following chapters will focus on the possibilities and limits of such a task: Assuming there is a finished sample of a paint and the composition should be determined as precisely as possible. Which opportunities are available in the analytical tool box?

2.3.1 ATR-FT-IR screening Assuming a semi-finished sample or a finished sample of an unknown paint with unknown composition is available. The aim should be to get as close as possible to the “true” composition by analytical means. The first step is usually a so-called ATR-FT-IR screening analysis. This means that the coating sample is applied to the ATR crystal of a routine instrument as received and a spectrum is recorded as a wet film. Then the coating sample is dried according to a suitable procedure and

133

Part III Quality control and process analysis the measurement is performed again. From these two measurements, one can already get first important information: – What kind of main components (binders, solvents, pigments, fillers) does the paint contain? – What is the relative concentration of these compounds?

Example: White automotive paint

A white automotive paint is homogenized intensively and then applied to the ATR crystal of a routine spectrometer as obtained (see part IV ATR-FT-IR spectroscopy). The ATR-FT-IR spectrum of the wet film shows intense absorptions of water at 3359 cm-1 and 1644 cm-1, which is the first important information: that it is a water-based system. However, other (co-)solvents cannot be reliably identified from the sum spectrum. In order to determine these, further methods such as Headspace GC-MS are required. The spectrum of the air-dried film (not to be confused with a cured film, which would have required a hardener in this case!) shows an intensive line at 1728 cm-1, just like the spectrum of the wet film. From the peak position in the spectrum, it can be concluded that it is a carbonyl compound such as typical polyester binders. As this signal does not decrease during the air drying of the film, it must be a low volatile or non-volatile compound. This supports the assumption that it is the (polyester) binder component(s) of the coating. The lack of aromatic absorption features above 3000 cm-1 and below 1000 cm-1 hint at some sort of aliphatic polyester. The fingerprint bands that determine the more exact composition are superimposed by those of another binder component. This component shows intensive signals at 1550 cm-1, 1385 cm-1, 1079 cm-1, 1016 cm-1, 911 cm-1,814 cm-1. A comparison with database shows that it is a melamine. Finally, the spectrum of the dry film below 800 cm-1 shows a strongly rising flank, which can be an indication (not a proof!) of titanium dioxide. This section of the spectrum is a hint that

Figure III.26: ATR-FT-IR spectrum of a white automotive paint measured as a dry film (top) compared with a database spectrum of a melamine binder (bottom)

134

Quality control of paint production further analysis of the dry film with Raman spectroscopy or scanning electron microscopy may be useful to find out about the titanium dioxide pigment. Conclusion: The two (relatively easy to perform) ATR-FT-IR analyses of the wet paint and the dry film already give important information about the main components of the paint: – water, – aliphatic polyester, – melamine, – pigment white as well as hints on how to proceed analytically to obtain additional information: – SEM-EDS (for determination of inorganic components such as pigments and fillers) – Headspace GC-MS (for identification of the volatile component) – TOF-SIMS (for identification of additives)

2.3.2 TOF-SIMS analysis With the help of TOF-SIMS analysis, dry films of coatings can be examined, especially with regard to additives such as light stabilizers, flow agents and the type of binders. The spectra of binders do show the binder base (e.g. phthalic acid esters, TMP esters, melamines, etc.) and, in individual cases, the chain structure of the polymer backbone, but not the exact composition or trade name.

2.3.3 Headspace GC-MS analysis In Headspace GC-MS analysis (see also Part IV, Chapter 5.13.2) , a sample of the paint is heated in a closed vessel and then the volatile compounds enriched in the gas space above the liquid sample are separated by gas chromatography and analysed by mass spectrometry. This means that this technique is suitable for the solvent content of the paint [9-10] as already explained in Chapter 1.2, Part III. The solvents can be determined with this technique if they can be evaporated at the temperature selected in the headspace preparation, pass through the column of the gas chromatograph where they can be separated cleanly, and can be identified with sufficient certainty by mass spectrometry.

2.4 Paint quality tests The delivery of a paint lot is checked by the paint manufacturer on a regular basis according to the parameters given in a quality certificate. This certificate covers physical characteristics such as colour, gloss, coverage, dry film density, VOC content and flash point; chemical characteristics, and general technical data. But beyond this information there are “hidden risks” that are not covered by the quality certificate and cannot be revealed by standard testing methods [11]. Some examples are: – an unsuitable combination of solvents can cause cratering problems; – inappropriate storage and transport conditions can cause agglomeration of paint components or hydrolysis of binders; – inadequate dispersion will cause segregation of fillers or pigments and/or variations of the colour shade; – the replacement of one paint binder by another can cause unexpected fogging; – the wrong choice of paint components can cause hydrolysis and agglomeration.

135

Part III Quality control and process analysis This reveals that some questions about the quality of paint that go beyond the standard quality certificate must be asked. Again, one of the main questions is: “Did I get what I have ordered?” This means that the analytical task is to prove that each paint delivery is effectively identical compared to a standard. This is a question that can be investigated by ATR-FT-IR (attenuated total reflection Fourier transform infrared) spectroscopy. Samples of a paint delivery can be directly analysed both as wet films (for example, to identify the solvents) as well as dried films (to check the binder, filler, pigment and additive content)

Figure III.27: Comparison of the ATR-FT-IR spectra of defective paint sample (blue) and a reference sample (red) revealing differences (grey)with respect to a surface additive (green) and a solvent (orange)

Figure III.28: Quantification of a siloxane surface additive by ATR-FT-IR and regression analysis

136

QC Rheometers anyone can use From development to quality control – reliable analyzes with versatile, innovative instruments increase your productivity and maintains quality standards. Thermo Scientific Rheometers meet these requirements: Ease of use, measurement and evaluation software for beginners and professionals, extensive accessories for every application.

Thermo Scientific™ HAAKE™ MARS™ iQ rheometers for comprehensive QC needs

Thermo Scientific™ HAAKE™ MARS™ rheometers for advanced QC and applied R&D demands

Entry level Thermo Scientific™ HAAKE™ Viscotester™ iQ, the portable rheometer for flexible QC tasks

Find out more at thermofisher.com/qcrheometers

For Research Use Only. Not for use in diagnostic procedures. © 2021 Thermo Fisher Scientific Inc. All rights reserved. All trademarks are the property of Thermo Fisher Scientific and its subsidiaries unless otherwise specified. AD53379E 0121

BUILD UP NETWORKS AND BE PREPARED FOR FUTURE CHALLENGES www.european-coatings.com/events We are providing a wide range of comprehensive training and exchange opportunities to deepen your competence and expertise. Get new impulses, meet the industry face-to-face or profit from our virtual offers. Conferences & congresses Digital bundles Exhibitions Webcasts And more

ABOUT T OP ICS INF ORMAT ION AT E AVAILABLE AND D ATE S AR S.COM/EVENTS

WWW.EUROP EA

N-C OATING

EUROPEAN

C OATINGS events

Vincentz Network · P.O. Box 6247 · 30062 Hannover · Germany www.european-coatings.com/events T +49 511 9910-033 · [email protected]

Quality control of paint production without any further sample preparation. Figure III.27 shows an ATR-FT-IR spectrum of an air-dried film of a defective paint sample that exhibited poor performance, although all the parameters of the quality certificate had been proved to be within specification. Compared to a reference spectrum of a delivery without any complaints, the spectrum of the deficient sample seems to be identical. But detailed data evaluation by spectral subtraction reveals that there is a difference. The deficient paint sample contains a siloxane surface additive and a solvent that has not been detected in the reference sample. This shows that the ATR-FT-IR technique is a versatile tool to perform a quick “identity check”. Together with compact instrumentation that can be placed almost anywhere the analysis leaves the laboratory and can be placed directly in the production line. It not only answers the question “what is in a paint?” but can also be used for quantification purposes, as Figure III.28 shows. If a paint component, such as the surface additive quoted above, shows spectral features that can be distinguished from the rest of the paint formulation, a quantitative analysis is possible. In this case, the amount of the siloxane surface additive in the deficient paint sample was determined by making a standard addition of this additive to the paint followed by ATR-FT-IR analysis. The intensity of a characteristic peak of the surface additive, that does not interfere with other paint ingredients, was measured and displayed in a diagram showing intensity versus addition. The regression analysis revealed that the additive had a concentration of 5 % in the paint instead of the expected 0.5 %. It must be added that this ATR-FT-IR analysis technique cannot be used for low concentrations of additives because of the detection limits of the method. If low additive concentrations must be detected and identified, the use of TOF-SIMS (time of flight secondary ion mass spectrometry) is recommended.

139

Part III Quality control and process analysis

3 Field analysis Laboratory failure analysis is a powerful tool to solve production and quality problems. But sometimes the puzzle of analytical data and background information does not fit together, or the conclusions are not unambiguous. From inside laboratory it is possible to specify for example that a wetting problem is caused by a certain separating agent and the chemical structure of the separating agent can be identified. With good reference samples it is feasible to attribute it to a group of process chemicals or external contaminations. But when it comes to the question “How did this substance get into the process?” there is little chance to answer this without having seen the site. Sometimes little seemingly unimportant observations on the site help to put pieces of data and background information together to a uniform and conclusive result. For example, a few years ago, the author had to deal with a problem of coating appearance malfunction (stains) happening only in the winter time. The laboratory analysis by TOF-SIMS and ATR-FT-IR revealed that the stains have been caused by the agglomeration of a paint component on top of the cured coating. Although, the substance had been identified and clearly attributed to the coating material no unusual amount of this additive could be found in the original coating material. Up to that point, no explanation was obvious why the coating material can cause this effect. It had been proved by analysis to be chemically identical with reference material. Visiting the site, it was observed that the painted and cured parts have been stored (still warm after leaving the curing oven) in plastic bags and moved from the production hall into a cold storage building. A short inspection of the bags with the stored parts then revealed that due to the temperature difference outgassing, volatile components of the paint have condensed inside the bags and “contaminated” the surface of the parts. So, the failure was the treatment of the work pieces directly after the curing of the paint which only leads to appearance failures during winter time when the temperature difference between “inside the production hall” and storage building are significant. There is no chance to find that out from the distance. On the other hand, the plant´s routine workers do not realize the problem because they follow their daily practice and do not consider the difference between winter and summer. Even the production leaders did not realize because they did not deal with every detail of production and storage. So, besides laboratory or better in addition to laboratory analysis it is sometimes to have an “external eye” on the process. On the site it is good advice to ask as many questions as you should do on the first sample inspection at Figure III.29: The coating of a wind power plant requires special field analysis techniques the beginning of the failure root

140

Field analysis Table III.4:  Sources of contaminations during paint application (sampling and identification method) Suspected source of contamination Sampling Analytical tool Coating material

random sample of the homogenized material sieve residue

SEM/EDS ATR/FT-IR IR microscopy headspace GC-MS

Coating booth

wipe sample of the cabin walls sample from separators and/or filter sample of tubes and/or pipes

TOF-SIMS ATR-FT-IR

Coating robot

exhausting the pipes through turbine and bell material samples of the robot rinsing of the tubes

TOF-SIMS ATR-FT-IR

Coating preparation

wipe sample of production bins and stirring units samples from material feeders

TOF-SIMS ATR-FT-IR

Ventilation system

aerosol targets filter samples

TOF-SIMS SEM/EDS

Compressed air supply sample from oil and water separators sample of tubes

TOF-SIMS ATR-FT-IR

Conveyor system

TOF-SIMS SEM/EDS

aerosol targets

cause investigation process. There are no stupid questions because nobody can know all details of any plant and sometimes a simple question can lead to surprising results.

3.1 Process analysis of paint shops As mentioned above, one result of the analysis of coating defects may be that further questions regarding the process arise. If, for example, the result of an analysis of paint craters is that they were caused by “silicone oil” (chemically polydimethylsiloxane or abbreviated PDMS), the question follows where in the process this interfering substance was introduced and why. These questions can only be answered by intensively studying the process. The reasons for this can be manifold: – planning error – errors during installation – maintenance errors – unsuitable coating material – defective paint material – insufficient knowledge about the function of the system – insufficient communication with the client of the painting – incorrect use of operating resources – system malfunctions – entry of impurities by external service providers – … If such a foreign substance, which is used in many technical products, has been detected in the painting process, it seems difficult to locate the source. Theoretically, the substance can be located

141

Part III Quality control and process analysis in any small seal, in any valve or even in distant parts of the plant, from which it is introduced, for example, via the painting air. So where to start the search? It is not practical to dismantle the entire plant and examine each component individually. The whole plant is a complex and sometimes “organically” and individually grown system which is not easy to survey. Good news is: It is not necessary to investigate each valve, tube or bin. The magical tool is to (virtually) reduce the complexity of the system by breaking it down into appropriate units. One possibility is to apply a principle that I named “box in box”. This means that one imagines the whole plant as a system of closed “boxes”, each of which is a closed system in itself: The exterior production building in which the coating plant is installed can be considered as a box. Inside this box there is another one which is the paint shop (box 2). Inside the paint shop are the coating line and/or a paint booth (box 3) which houses a paint robot (box 4) and so on (Figure III.30). As a rule, in so far as it has a clean room, the paint shop is already a “closed” system in purely technical terms. Closed does not mean hermetically sealed but separated from the “unclean” environment. This box “painting plant” has entrances through which (pressurized) air, water, paint, work pieces, operating materials and people are brought in (interconnection points). Within this box there are further smaller boxes which in turn can be regarded in a certain way as a purely schematic closed system, namely the painting cabins, the air circulation system and the control room. These also have technically defined entrances through which, for example, air, solvents, rinsing media, paint and paint products are introduced. Within the painting cabins there

Figure III.30: “Box in box” principle for sampling a paint shop

142

Field analysis are further “boxes”: the painting robots. If certain foreign substances like e.g. a polydimethylsiloxane PDMS has been detected on the surface of a work piece, it is obvious that it must have entered the closed system of the boxes through “entrances” or interfaces. PDMS for example may have entered through the “main door” by contaminated raw parts. Or it was sucked from ambient air into the air supply system. Or it came in as a “stowaway” inside the material container of hardener, base components or solvents as a residue of the coating material production. Or it is entered via process chemicals. Or it was brought in by humans. As a result, the entries for work pieces, air, process chemicals, employees, flushing agents have to be sampled to “trap” the contamination. This means that strategic sampling is carried out at the entry points or interconnections to the “boxes” (see circles in Figure III.30) to check where the interfering substance has entered the closed system. A look at the outer box “paint shop” with regard to the example of polydimethylsiloxane mentioned above shows that the substance is detected on the components surface, then it might have entered the closed system via the “inputs” – compressed air – paint material – coating goods – production air supply – solvents – flushing media – employees This means that if it has to be checked if the compressed air that is transferred to the painting system, is free of this interfering substance, a representative sample has to be taken at the transfer point from one box into the next one and examine it for PDMS. The details of the sampling procedure have to be decided on site in each individual case based on plant conditions. This can be done, for example, by blowing the compressed air stream over a reflective surface (see Figure III.32) or by examining a compressed air filter. The same procedure is used for the other “inputs” to the box. To ensure that the rinsing medium is free of PWIS1, a representative sample is taken and examined before it is transferred to the spray booth. This can be taken e.g. from the storage tank or the direct supply line. In this way, all inputs are checked one by one. What is the advantage of this “box in box” strategy? Of course, one could also start at any point in the spray booth and, for example, open a rinsing agent line to test the rinsing agent for PWIS. If something is found, it is not possible to determine whether the rinsing medium itself is contaminated or whether it has taken the PWIS from a part of the plant, e.g. a valve. If, for example, the compressed air is sampled on the robot and PWIS is found there, then both the compressed air may have been contaminated and any valve or line through which the compressed air has flowed. Appropriate sampling at the strategic interfaces mentioned above can now be used to isolate the source of the fault before going into detail with the investigation of plant components. For the evaluating of the application conditions some of the aspects are: – interim storage – preparation for painting (agitators, pumps, containers) – insufficient cleaning – painting (robots, hoses, spray guns etc., turbines) 1 PWIS = Paint wetting impairment substances

143

Part III Quality control and process analysis – – – – – – – –

drying (plants, conveyor belts etc.) humans (working clothes, gloves, skin or hair products) lubricants from moving parts mould release agents from plastic add-on parts overloaded air filters, and check of oil and water separators inappropriate design of the compressed air system insufficient programming of the paint robot mould release agents from cleaning cloths, unsuitable cleaning agents, sanding media, adhesive tapes – seals and insulating material from the building – aspiration of contaminated air from outside (silicone spray, polishing agent, dust), condensation

3.2 Aerosol analysis Aerosols are airborne solid and liquid particles or droplets. The typical size is between a few nm and more than 100 µm. Therefore, these aerosols are invisible. But they can influence the surface quality of an object that has been coated or is ready to coat. These can occur in a paint shop in different forms: – paint spray mist – oil droplets from compressed air lines – aspirated components of the ambient air of the plant – human sources (e.g. hairspray or skin care products)

Figure III.31: Paint preparation stage for coating material observed in a paint shop

144

Field analysis These aerosols can have a considerable influence on the quality of the coating. The author is aware of failure incidents where whole painting lines have been contaminated with PWIS through the use of silicone-based hairspray. Aerosols usually do not remain where they are created but are distributed over a large area due to the small particle size by pressure differences and ventilation systems. Aerosols play a decisive role in many cases, especially in the formation of paint craters. In order to test whether an aerosol load is present in an operating section and identify the composition, a process has been developed in the recent past which has since proven itself hundreds of times in practical use [7]. Specially cleaned metal samples of etched silver (so-called silver targets or aerosol targets) are placed in the area to be investigated. These passive collectors remain in the paint shop for the selected investigation period (e.g. one week) and are passively exposed to the suspended particles and gases in the room during this time. It traps the aerosol such that it can be identified by surface analysis and it separates that fraction of aerosols which can contaminate the products from those which do not. This results in an irreversible accumulation or binding to the metal surface. After the sampling time, the targets are collected and placed in an airtight container for transportation to the lab and analysis in a TOF-SIMS instrument which analyses the surface with regard to the accumulated foreign substances. Following this procedure, even discontinuous aerosols can be detected and characterized. This special sampling technique was developed to answer a very special analytical question. Contaminants on surfaces can originate from different sources as mentioned above. One source is the ambient air. A production plant equipped with some fifty injection moulding machines, for example, has hundreds of sources of oil contaminants. Many machine parts are driven by air pressure which produces a fine oil dust that is expelled into the ambient air. This dust is invisible because the droplets are microscopic. But the oil can be smelled in such production plants. This “aerosol” can cause major adhesion problems. If, e.g. an automotive interior part that has to be finished afterwards by painting or chemical vapour deposition is stored in this contaminated air environment, the oil dust precipitates on the material surface to form a very thin, invisible film that can cause severe adhesion problems. Furthermore, the aerosol can and will be transferred to other sections of the production plant that are linked up to the ventilation system.

3.3 Operating test When a production failure occurs, practical tests are often carried out which are supposed to help solve the problem. The production conditions are changed; parameters are modified, but very often without an expedient design of experiment (DOE). This is the engineer approach while more scientifically oriented people rely on analyses first to fix facts before starting different action plans. The truth lies, as often in life, rather in the middle. Without a doubt, operational tests are an important means to localize the source of error, but they do not replace the recording of facts and measurement data. Therefore, analysis and field tests should always go hand in hand. A field test serves to vary a parameter that is assumed to influence the problem in a targeted manner and to evaluate the outcome of the test to see if the assumption is correct. The following points are important to ensure that these field tests also lead to reliable results: – careful planning – complete documentation Before planning the operating tests (DOE) the first question is: Which kind of information is expected from the experiment?

145

Part III Quality control and process analysis Table III.5: Protocol of a coating test to localize the root cause for a cratering issue Power Base Clear Part DOE (6 parts) wash Primer coat coat Result Side bezel part 1 – 3 series treatment parts 4 & 5 flame treated after power wash

yes

yes

test primer removal of part 1 & 4 after primer application

red

yes

• part 1 & 3 without crater • part 2 crater

red

yes

• part 4 & 5 without crater

The total expected expenditure depends on the answer to this question as well as the determination how often the experiment has to be repeated so that the result is statistically relevant. This should be illustrated by a specific example: the case of cratering in a multi-layer coating. There are many possible causes for this (as already described in Part II): – contamination of the paint material – contamination of the painting air – contamination of the cabin air – contamination of the oven air and condensates in the oven – impurities in paint lines – contamination on the outside of the robot or its attachments – flushing media with which robot parts or plant components were flushed – contaminated cabin walls – incorrect layer thickness – insufficient mixing of paint components – insufficient component cleaning – …. If all these options have to be checked by analysis (which would undoubtedly be possible) this would mean that every single part, no matter how small, would have to be tested. This is not a practical approach with respect to at least hundreds of components of e.g. a painting robot. Here, well planned practical tests backed by selective analyses according to the “box in box” principle help to limit the problem to one (or more) plant components before going on to examine individual parts in “suspicious” areas. It is important that practical tests are planned in such a way that the conclusions to be drawn from them are clear. This means – to be aware of all parameters that have an influence on the outcome of the experiment and – that only one parameter is changed and at the same time – the number of repetitions of the test has to be chosen in such a way that random results are excluded. Table III.5 shows the results of a test painting of components done with the purpose to identify the causes of craters. In the course of the test painting, components were painted with and without flame pretreatment. A close look at the content of this table reveals significant errors that can lead to doubtful conclusions: The initial plan was to coat bezels with different production parameters in order to verify the influence of – flame treatment – power wash – primer application … on the appearance of craters.

146

Field analysis According to the records of the coating trial, six components were painted. However, only five parts appear in the protocol of the test and the evaluation. So, in the first place that means the recording is not complete. The occurrence of craters happens with an (unknown) failure rate. Six samples being coated with different parameters in order to evaluate if a certain parameter can be attributed to the crater issue is far too little. This experiment shows that two of three parts painted under normal series conditions are okay and only one part had craters. So, there is a 66 % probability that under normal production conditions parts are produced without defects. However, only two parts were additionally flame-treated under conditions other than those of series production. These also did not show any craters. This number of test parts is therefore not sufficient to evaluate at the end of the test whether the paint job is coincidentally defect-free within the statistical probability or whether the result was improved by the additional flame treatment. In addition, one more component was taken from each test group after primer application, which changing another parameter. Summing that up, this means that the results of this painting test are completely useless in terms of the original task to find out whether an additionally introduced flame treatment would improve the painting result.

3.4 Sampling of the painting air A frequently encountered source of error, which leads to cratering in coating, is compressed air. Errors made in the early planning stage of the air supply system can cause problems that develop later under certain production conditions. When it comes to painting air supply, an oil-free compressor is recommended. If this is not possible, a well-engineered air cleaning system including oil mist filters has to be installed and maintained. The dimensions of the system and flow rate of the compressed air have to be adjusted to the maximum load that has to be supplied if the production runs on a 100 % level. If one of these points fails, lubricating oils, abrasion and grease can be accidently introduced into the painting air. If a contamination has entered the system, it can spread unnoticed from the compressor to the robot or the paint spray gun. It requires considerable effort to find and remove this source of failures. For this reason, sampling the compressed air or the paint spray air is one of the first tasks when, after the source of the error has been narrowed down, there is a suspicion that the compressed air could be contaminated. Sampling of the compressed air,has to be done very carefully by experienced and well educated persons because sampling errors can lead to false conclusions and make it impossible to solve the problem. The easiest way of sampling is overblowing a clean mirror surface (e.g. silicon wafer pieces) (see Figure III.32). The frequency of testing is important for a reliable result. Interior contaminations of an air supply system are not necessarily detectable on the first test. In addition, it must be taken into account that impurities can be present in different forms in the compressed air line. Compressor oils, for example, can be present as oil vapour or oil droplets in the compressed air, whereas viscous fats can be present as particles or plugs or bound to abrasion particles. All these components behave differently in the air flow and can accumulate at different points in the system. So, if the compressed air line is opened at any point to take a sample, a negative result (no detectable contamination) can mean that either:

147

Part III Quality control and process analysis – there is no contamination in the compressed air or – that no pollution is measurable at the chosen sampling point but that there is nevertheless a load in the system, or – there is a contamination which is not visible, but which could be detected analytically, or – no contamination is detected under the plant conditions prevailing at the time of sampling, but nevertheless a contamination exists that would have been detectable under other sampling conditions, or – the absolute sampling time was too short. A qualitative sampling according to Figure III.32 by overblowing a reflecting surface aims to check visually whether there are liquid or solid contaminants in the compressed air that can be deposited on the test surface. However, this method fails with respect to low-concentrated vapours or aerosols of substances that can cause a crater problem even in small quantities. The absence of visually recognisable spray mists on the testing mirror like in Figure III.32 does not mean that there is none! The sampling procedure for trace analysis to detect those low concentrated pollutants needs a carrier material suitable for the analysis. Depending on the analysis method to be applied, different carrier materials are used. The following procedures are suitable, depending on the analysis method to be applied: Headspace GC-MS: For this analytical method the sampling is carried out on a porous absorber material on which the oil vapours, aerosols and fats are adsorbed on site. For this purpose, an air stream is branched off at a suitable point and passed over the porous medium which is located

Figure III.32: Result of a sampling of the atomizing air of a painting robot (the separated oil mist is clearly visible here)

148

Field analysis in a closed container for a defined period of time. After completion of the sampling, the air flow is stopped; the vessel is hermetically sealed and brought to the laboratory. There the collected material is desorbed by passing a clean gas flow. TOF-SIMS: Especially if the analysis of the craters was carried out with the TOF-SIMS method, it makes sense for reasons of comparability of the measurement results to also carry out the air sampling in such a way that the deposited substances can be examined in the TOF-SIMS. The sampling technique for GC-MS is not suitable for this purpose. Rather, in this type of sampling the air to be investigated is guided over a cleaned metal surface (e.g. aluminium or silver) and the substances contained therein are deposited on the metal surface. This procedure is closer to the actual mechanisms involved in the deposition of impurities in the painting process than sampling on a porous absorption medium.

3.5 Monitoring of pretreatment steps To achieve perfect coating, it must be ensured that no surface-active residues interfere with the coating material. There may be: – oils or greases, – antistatic agents, – anticorrosion additives, – mould release agents, – dust, – environmental contamination, on the surface depending on the circumstances the substrate encountered prior to coating. Therefore, in most cases some sort of pretreatment or cleaning process such as a power wash unit should be installed. The relevance and the results of different pretreatment methods have already been discussed before. As already explained in Chapter 4.1.1 of Part II, pretreatment and cleaning processes prior to painting are both a blessing and a curse. Unfortunately, there is no easy, on-line testing device available that is able to detect and identify surface contamination. Sometimes test inks are used, but they give only a rough measure of the surface tension without showing specific contaminants and their chemical structure. In addition to that this test fails with rough surfaces. Optical or spectroscopic methods such as fluorescence spectroscopy that have been advocated as fast on-line testing methods suffer from a lack of sensitivity and omission of any molecular identification of the contaminants, which is required in order to choose the right pretreatment procedure. Neither the amount nor the chemical composition of surface contamination nor the result of the pretreatment process can be checked without surface-sensitive analytical methods. It has to be remembered that a mono-molecular layer of some contaminants (for example, fluorocarbon greases) on the surface can disturb a coating process and so lead to craters, pinholes or adhesion problems. A potential testing method would have to meet the following requirements: – high surface sensitivity – low detection limit for organic and inorganic contaminants – molecular identification. Fortunately, time of flight secondary ion mass spectrometry allows for the surface detection of contaminations with very high sensitivity. Figure III.33 shows a TOF-SIMS analysis of a steel panel taken from a coil compared to a reference spectrum of a corrosion protection oil, revealing that the surface is contaminated by this oil which can be identified and named exactly. TOF-SIMS is the method to detect and identify all surface contaminants with a detection limit down to parts

149

Part III Quality control and process analysis per billion using only a single measurement without any sample preparation (except for cutting out an appropriate sample). Cleaning before painting is not always a guarantee for good adhesion, indeed sometimes it is counterproductive. The following list shows only a few examples of processes and their effects on the wettability and adhesion properties of a coating material. In order to remove surface contamination from substrates and to activate the surface, a few surface treatment techniques have been established: – power wash – plasma treatment – CO2 snow jet – Corona treatment – fluorination – blasting with pellets are some examples. The use of one of these methods is believed to ensure that the surface is in a condition that allows a high quality coating to be obtained. But sometimes it does not. One common reason is that a pretreatment process, once installed, is used for a variety of different parts that have to be cleaned. However, this is very often done without knowing: – Whether the surface is contaminated? – How the surface is contaminated? – How the pretreatment process influences the material? If, for example, an injection moulded polymer part carries a surface contamination of hydrocarbon release agents and silicon oils, the chemistry of power washing may completely remove the hydrocarbon contamination but not the silicon oil. In addition to that, residues of the surface-active agents, antifoam additives and coagulants added to the power wash liquid can leave the surface more contaminated after treatment than it was before.

Figure III.33: TOF-SIMS spectrum of a steel panel taken from a coil (left), compared to a reference spectrum of a corrosion protection oil (right)

150

Field analysis Table III.6: Possible applications of surface analytical methods in the pretreatment process of steel Production step Analytical task Method Delivery of steel

absence of oils and greases

TOF-SIMS

Blasting

topography

LM, SEM-EDS

cleanliness

TOF-SIMS

Degreasing

freedom from oils and greases

TOF-SIMS

Rinsing

test of the rinsing bath

TOF-SIMS

surface cleanliness Pickling

micro-roughness

SEM-EDS

Acidic rinsing

test of the bath

TOF-SIMS

Metal plating

test of the bath

TOF-SIMS

Plasma or flame treatment sometimes leaves smaller fragments of contaminants on the surface and results in chain scission of the uppermost polymer layers if the parameters are not adjusted properly. Hence there could be a major difference between the expected and the real surface condition, leading to surprising and unexpected coating results. Another aspect is the influence of a cleaning process on the substrate itself. A plasma treated polymer surface which has been cleaned from surface contamination represents an activated surface. This activation means that the surface attracts organic substances from the ambient air and it also drives additives to migrate from the polymer bulk to the surface within minutes, hours or days. The TOF-SIMS analysis of a polyamide surface exhibiting painting and printing problems (Figure III.34) illustrates that before plasma treatment there is only a flame protection additive present on the surface. After plasma treatment, the migration and recontamination by release agents from the polymer bulk onto the surface is quite obvious. In this sense the pretreatment did a bad job with respect to paintability and printability. It has been proved in the authors laboratory and elsewhere that the ef- Figure III.34: Positive TOF-SIMS spectrum of a polyamide fect of fluorination of a polymer sur- 6/6.6 surface before (above) and after (bottom) plasma treatment (showing that the plasma treatment caused face is strongly dependent on the re-contamination of the surface by release agents amount of surface contaminants and migrating from within the polymer)

151

Part III Quality control and process analysis the composition of the polymer substrate [12]. During fluorination, several chemical processes are running in parallel on the surface. Surface analysis has proved the occurrence of insertion of oxygen and fluorine into the polymer network of the uppermost layers, chain scission, and rearrangement of polymer fragments and fragmentation of contaminants. This results in a chemical “cocktail” that has to be carefully controlled in order to achieve the expected results. Wipe cleaning of polymeric coating material with isopropanol can lead to the extraction of components from the coating material, which then spread on the surface as a thin film. When cleaning with power wash and similar washing systems, incorrect programming of the system, unsuitable chemistry or inadequate rinsing with deionized water can leave residues of cleaning agents or carry-over residues of contamination of other components on the painted material. Corona treatment, flame treatment or fluorination of components often leads to a chemical degradation of the uppermost substrate layers if the process is not correctly controlled. The resulting fragments then only adhere loosely to the substrate and in some cases lead to wetting disorders, but also in other cases to large-area delamination (see delamination). Conversely, depending on the component geometry and system, areas can also occur which are not or only insufficiently reached by surface treatment. Cleaning and pretreatment of the components to be coated, whether made of metal or polymers, must always be tailor-made. It has been shown several times that the pretreatment in a plant which has been adjusted to a specific material type at the start of production can be counterproductive if the material type is changed. If, for example, a plant is set up for coating PP-EPDM and flame treatment is advisable in this case, then a switch to polycarbonate components is made, then flame treatment is not only not advisable but can have exactly the opposite effect to that desired. Therefore, it is not good advice to blindly trust the promises of the system manufacturers. It makes sense to check the actual success (or failure) of the pretreatment steps analytically on a random basis. Since the paintability is usually decided in the uppermost molecular layers of the paint material, a correspondingly sensitive measuring method is required to check the composition in this area. The TOF-SIMS method is therefore best suited for this purpose.

3.6 Investigation of the degree of crosslinking in 2-pack paints During paint application there are some issues which have been investigated successfully by means of analytical techniques. One of those aspects is the degree of crosslinking in 2-pack polyurethane (or 2K-polyurethane) paints. Although the dosage of masterbatch and hardener in modern paint plants is very precise, malfunctions and human error sometimes lead to a wrong mixing ratio. One goal of failure analysis is to verify and prove the binder/hardener ratio of the paint film. This has been done successfully with ATR-FT-IR. The paint film is pressed onto the ATR crystal and measured without any further sample preparation (for more details see Part II Chapter 4.2.4). As Figure III.35 shows, there is a relationship between the intensity of the characteristic peaks of the isocyanate hardener (marked by an arrow) and the ratio of master batch/binder and hardener. The relationship is linear if all the parameters of the measurement (optical contact, pressure, coverage of the ATR crystal surface) are identical. This is a prerequisite for quantitative determination of the binder/hardener ratio by measuring the intensity of the peak intensity of hardener and binder (see Figure III.36).

152

Field analysis

The procedure is as follows: 1. Prepare a set of dried paint samples of different masterbatch/hardener ratios. 2. After the paint has dried, measure the ATR-FT-IR spectra of these reference samples. 3. Determine the intensity of one characteristic peak of the binder (e.g. the carbonyl stretching vibration of polyester) and the hardener. The majority of instrument suppliers offer a software module for this digital operation. The intensity can be determined via the peak height or the peak area. As the peaks of the carbonyl stretching vibrations of binder and hardener overlap, determining the peak height is the method of choice for this task. uation Part III  4. Divide the measured intensity of the selected characteristic vibrations of the binder by the intensity of the corresponding peak of the hardener to obtain a relative peak intensity value:

uation III.1:   

ூ

‫ܫ‬௥௘௟ ൌ ್   ூ೓

Equation III.1

where Ib is the intensity of the C=O stretching vibration of the binder, Ih is the intensity of the C=O stretching vibration of the hardener and Irel is the relative peak intensity – For each spectrum of a reference paint sample produced with a given binder/hardener ratio, the result is a relative peak intensity value which is displayed in a diagram of Irel vs binder/ hardener ratio (see Figure III.36).

Figure III.35: Comparison of ATR-FT-IR spectra of 2-pack polyester-polyurethane paint films of different binder/hardener ratio

153

Part III Quality control and process analysis Table III.7: Allocation of evaluated peak and peak intensity of IR spectroscopy Wavenumber of evaluated peak Peak assignment

Peak intensity [peak height]

1725 cm-1

ν, C=O stretching vibration of polyester binder

0.31

1685 cm-1

ν, C=O stretching vibration of isocyanate hardener

0.26

– From these readings, mathematical software calculates a linear regression line which serves as a calibration curve for determining the binder/hardener ratio of an unknown sample. – After this calibration process, determine the binder/hardener ratio of each unknown sample by measuring Irel of the sample and reading the corresponding value off the calibration line. During acquisition of the sample spectrum, ensure that the same measuring parameters apply as for the reference samples.

Example: Binder/hardener ratio of 2-pack polyester-polyurethane paint

Evaluation of the intensities of the characteristic binder and hardener peaks of 2-pack polyester-polyurethane paint yields the values shown in Table III.7. These data yield rel. peak intensity Irel of 1.182. Taking this value, a binder/hardener ratio of 100 : 7.2 can be read off the calibration line. As the rel. peak intensity Irel does not change during paint drying if the paint components are non-volatile, this method can be used to determine the binder/hardener ratio with an accuracy that is sufficient for technical purposes. From the viewpoint of chemometric analysis, the method is not precise enough, but it has worked very well for those technical applications which don not need that level of precision. It has to be mentioned that the values of the binder/hardener ratio reveal nothing about the real degree of crosslinking. If, for example, inhibition of reactivity has occurred because a catalytic additive was forgotten, this is not apparent from the infrared spectrum.

Figure III.36: Calibration diagram displaying the rel. peak intensities Irel vs. binder/hardener ratio for 2-pack polyester-polyurethane paint

154

Field analysis Where reaction inhibition of binder and hardener is suspected, some additional preparative operations can help to find out. First, a spectrum of the paint layer is acquired, and the rel. peak intensities are determined as described above. Then, the paint layer is extracted with ethanol or propanol to wash out the fraction of unreacted binder. After drying, the paint layer is measured again and the resultant binder/hardener ratio after rinsing is expressed in terms of the value determined before rinsing. This procedure yields the amount of binder that has actually reacted with the hardener.

3.7 Investigation of paint additive migration Paint formulations feature a lot of additives which are included to enhance the performance of the paint with respect to appearance, adhesion, wetting, etc. As these additives do not take part in the crosslinking process, they are mobile and tend to migrate within the paint system. Some of

Figure III.37: Pos. TOF-SIMS image of the distribution of a polysiloxane additive in a multi-layer system

Figure III.38: Arrangement of analysis spots along a wedge cut of a multi-layer paint system

155

Part III Quality control and process analysis them can be found at the surface, whereas others will accumulate at the interface of the paint and substrate or between two paint layers. In a multi-layer system, additives can migrate between two, three or more paint layers. A detailed knowledge of the distribution of an additive can thus help to improve the technical properties of the polymer coating. For example, a light-stabilizer additive incorporated into a clear coat unfolds its greatest effect if it is homogeneously distributed in the paint layer or exhibits a gradient to the surface, whereas surface-flow-control compounds in water-borne systems form self-assembling, highly ordered structures in the uppermost paint layer and could prove problematic for repair coats applied as a second layer. For these issues, the main question is: How does an additive migrate and distribute itself in a coating layer? And: Which information is needed to be able to answer this question? a) The chosen analytical technique must yield molecular chemical information for unequivocal identification of the additive. b) The distribution has to be examined along a cross-section of a single-layer coating or a multilayer coating. c) Given that the absolute quantity of an additive in a coating layer is very low, the chosen technique for detecting and identifying it must have very high sensitivity.

Figure III.39: Arrangement of analysis spots along a wedge cut of a multi-layer paint system (schematic side view)

Figure III.40: Analysis result: Distribution of different paint components across the cut through a multilayer paint system as shown in Figure III.38

156

Field analysis Therefore, what we need is a molecular-depth profile of an organic molecule in an organic (multi-layer) system. This seems to be a huge, if not impossible, challenge. Depth-profiling methods utilise a sputtering beam that removes one layer at a time but have the great disadvantage that the sputtering process destroys organic substances. On the other hand, non-destructive sputtering is a technical contradiction. Yet non-sputtering techniques have been developed that show the distribution of certain organic substances in a multi-layer system. One is TOF-SIMS imaging, as demonstrated in Figure III.37. This figure shows a TOF-SIMS image of the distribution of a siloxane additive in a multi-layer system. First, a microtome section or a metallographic cross-section through the paint system has to be performed. After this mechanical preparation step, the area of interest in the cross-section is scanned by a micro-focused pulsed ion beam. A mass spectrum is collected at every pixel point of a, say, 300x300 pixel array. Once this collection of mass spectra has been stored, an image is generated by extracting the characteristic peak(s) of the substance to be analysed. In this example, the characteristic mass peaks of a siloxane additive (which are 73 u, 147 u, 207 u, 221 u, 281 u, corresponding to fragments of the siloxane network) have been extracted from the overall data set to generate images of their intensity over the scanned area. All five images of the intensity distribution for masses 73 u, 147 u, 207 u, 221 u and 281 u in Figure III.37 display the distribution of the polysiloxane additive over the scanned area of the cross-section and thus help to detect migration and segregation processes. The second technique is that of TOF-SIMS analysis of a wedge cut of the paint system. Again, mechanical preparation consists in making a microtome section. However, this cut does not slice rectangularly through the layer(s) but rather at a very small angle nearly parallel to the paint surface. The result is a cross-cut area which presents the whole layer system from the paint surface down to the substrate. This cutting zone is readily accessible to surface analytical methods and thus can be analysed by TOF-SIMS. The test procedure is as follows: Starting on the paint surface, a set of spot analyses is performed along a line that crosses the paint layers down to the substrate as shown in Figure III.38. The analysis spot must be equidistant, to allow for the size of the cutting zone, to make sure that there are enough points for each layer. Each spot analysis returns a full set of information about the paint composition at that specific point. The full data set of spot analyses lined up from the paint surface to the substrate offers a complete image of the distribution of paint components over the paint layer system. From this mass spectroscopic data set, characteristic peaks for a certain additive can be extracted. The intensity of these characteristic peaks is evaluated by integrating the peak area and normalising it to the peak area of an internal reference substance. The concentration or peak intensity of this reference substance must be independent of the quantity of additive. As a rule, good results are achieved by taking the sum of integral intensities of hydrocarbons as an internal reference. The outcome is that a normalized intensity for this additive is achieved for each spot on the cutting zone. The analytical result, as shown in Figure III.40, is an image of the distribution of several additives along a virtual cut through the paint system.

3.8 Marine and aircraft coating inspection Marine coatings cover several requirements. For commercial vessels corrosion protection is the main issue. Coating failures have a strong impact on the corrosion resistance that is to say on the life time of the ship. For leisure ships, boats or super-yachts the coating also expresses the personality and the social status of the owner. When it comes to coating failures super-yacht owners and leisure boat owners make a point of appearance issues that are not known in the commercial ship

157

Part III Quality control and process analysis market. The inspection and analysis of marine coatings is a very special challenge because it is not possible to transfer a whole laboratory into the shipyard and destructive sampling is not appreciated in most cases. On the other hand, there are some portable instruments and (non-) destructive sampling methods that allow for a lot of investigations that is presented here: Failure analysis of coatings on ships and aircrafts is often limited by the fact that the majority of analytical methods to investigate failures like discolorations, hazing and gloss deficiencies cannot be performed on the site. In general, samples have to be taken out of the finished surface and transferred to a laboratory in order to be analysed for the chemical composition. This is already mentioned in Part II of this book. However, there are some methods that do not require a vacuum measurement and are portable to the shipyard (see Chapter 3.9). With specular reflection infrared spectroscopy and Raman spectroscopy on-site analysis is feasible. The surface composition of a ship coating (which means the chemical composition of the uppermost nm to µm) can measured without contacting the object as long as the portable instrument can be placed a few cm in front of the target surface. Figure III.42 demonstrates this for a polyester polyurethane coating which has been analysed without any preparation by external reflection of infrared radiation. The top spectrum is the reflectance spectrum which exhibits the characteristic features of a polyurethane but is distorted by dispersion of the peak intensity of strong bands. After Kramers-Kronig transformation the spectrum resembles absorbance spectra that can easily be compared to database spectra. This method is useful if for example hazes or precipitations are visible that cannot be sampled.

3.9 Handhelds and portables

Figure III.41: Picture of a high gloss yacht coating

158

Once an expert is called on the site the client expects a lot. Very often a lot of field tests have been performed and failed before the expert arrive. So, everyone is looking forward what the expert is doing in the field and the expert find himself in a situation that there is the implicit expectation that he arrives with a huge amount of equipment make some measurements, print the results and all the mysterious problems are solved. Arriving just with a sample kit and taking some pictures and samples often disappoints these expectations. But in fact, the options to analyse on the site with the necessary methods are sometimes limited. A lot of methods are vacuum techniques and thus are not portable. But nevertheless, there are some techniques that can be utilised on the site and that is sometimes very helpful because it saves time and it increases the belief in expert knowledge.

Field analysis – As already described the first inspection by portable computer microscopes is very helpful to guide the following examinations into the right direction. – A portable infrared spectrometer (which are now available from different manufacturers) can be placed close to the object that has to be analysed and allow for non-contact measurements using external specular or diffuse reflection techniques (see Figure IV.26 in Part IV of this book). Even if the direct measurement is not applicable, having the portable instrument on the sites can be helpful when it comes to detailed analysis of different machine parts. The author e.g. has use it for detailed investigation into the parts of painting robots. It is more convenient to disassemble a robot on the site, perform the analysis of seals, pipes, hoses and valves and immediately re-assemble it instead of sending all the parts to a laboratory and wait for days for the results. Some manufacturers even offer handheld instruments that are smaller than just portable ones [13]. They are used in a way similar to a laser speed pistol and measure in external reflection mode. – Portable Raman spectrometers are also available nowadays. Although they are not surface sensitive these instruments are helpful when an identification of unknown substances is required without touching the sample. In contrast to infrared spectroscopy Raman can even been used to analyse samples behind e.g. glass or other obstacles like storage container [14].

Figure III.42: Infrared reflectance and Kramers-Kronig transformed absorbance spectrum of a high gloss yacht coating measured by non-contact infrared specular reflection analysis

159

Part III Quality control and process analysis

3.10 References [1] [2]

[3]

[4]

[5] [6]

[7]

160

R. Dietrich, „Über den Sinn und Unsinn von Rohstoffkontrollen,“ Journal für Oberflächentechnik, p. 74ff, 2014 J. La Nasa, I. Degano, F. Modugno, M. Perla Colombini, „Industrial alkyd resins: characterization of pentaerythritol and phthalic acid esters using integrated mass spectrometry,“ Rapid communications in Mass Spectrometry, Bd. 29, Nr. 3, 23 12 2014 C. Jiang, M. Driffield, E. L. Bradley et al., „Studies of the aging effect on the level of isocyanate residues in polyester-based can coating systems,“ J. Coat. Technol. Res., Bd. 6, p. 437 ff, 2009 M. P. Colombini, F. Modugno, S. Giannarelli, R. Fuoco, M. Matteini, „GC-MS characterization of paint varnishes,“ Microchemical Journal, Bd. 67, pp. 385–396, 2000 I. Brown, „Identification of Organic Compunds by Gas Chromatography,“ Nature, Bd. 188, pp. 1021–1022, 1960 Shimadzu, „Paint Thinner Analysis Utilizing Headspace GC-MS,“ [Online]. Available: https:// www.shimadzu.com/an/literature/gcms/ jpo212072.html M. Feld, J.-U. Deimel, H. Riedel. , „Kontaminationen von Oberflächen über die Luft,“ Metalloberfläche, Bd. 54, Nr. 11, pp. 48–52, 2000

[8] [9] [10]

[11] [12]

[13]

[14]

D. Lee Hicks et al. „Automotive Fogging Analysis by Xenon UV Exposure“. Patent US 9,506,863 B2, Novembre 2016 J. Haken, „Gas Chromatography in Coatings Analysis,“ Encyclopedia of Analytical Chemistry, Bd. 15, p. 1738ff, 2000 C. J. Schwarz, D. Clay, Monitoring Solvent Levels in Waterborne Coatings by Static Headspace Gas Chromatography, T. Finnigan, Hrsg., Austin, Texas, pp. 1–3 R. Dietrich, „Knowledge versus risk (Part 2),“ Europeen Coating Journal, p. 34, 07/08 2009 A. Kharitonov, „Direct fluorination of polymers – From fundamental research to industrial applications,“ Progress in Organic Coatings 192–204, Bd. 61, p. 192–204, 2008 A. R. Leung Tang, „Non-Destructive Testing (NDT) of an Industrial 2K Epoxy Resin-coated panel undergoing accelerated weathering,“ Agilent Technologies, USA, 2016 F. Fromm, „New Possibilities in Raw Material ID using Handheld Raman Spectroscopy,“ Pharmaceutical Business Review, 2015

Optical light microscopy

Part IV Methods of coating analysis 1 Optical light microscopy One of the main topics of this book is failure analysis. Years of experience in failure analysis show that a great deal of quality claims of the coating industry is due to very small defects, particles, fibres, cracks, or material defects. The first step in failure analysis is to find out what kind of problem exists, and in most cases the tool to use is (optical) light microscopy. This basic technique can be used to carry out a preliminary sample inspection to gain an overview of the problem. Light microscopy reveals initial, basic answers to such questions as – What might be causing craters and spots in paint layers? – Where does paint delamination originate in a multi-layer system? – What does a residue in a raw material look like? The term microscopy is derived from the Greek words “mikros” (= small) and “skopein” (= view). The competence of the human eye to resolve very small objects is limited. Usually we achieve magnification by approaching the object (i.e. an extension of the angle of vision) and/or by accommodation. Since the clear visual range is 250 mm, the rays of two object details must be 0.0167° apart to be perceived separately. This means that two object points are only just perceived separately by the unarmed eye if they are at least 73 μm apart (provided that the observer is not suffering from defective vision). For all distances below this limit we need aids. Light microscopy is, therefore, the first and most important instrument, especially in defect analysis, for examining paint defects such as specks, craters or paint detachments. The theory of light microscopy will not be discussed in detail here. A sufficient number of specialist books have been written on this subject [1, 2]. However, special procedures which open-up interesting possibilities in damage and failure analysis should be mentioned here.

1.1 Extended focus imaging (EFI) A main disadvantage of conventional light microscopy is the lack of depth of focus. At high resolution, rough material surfaces, such as those of structured polymers or metals, cannot be inspected both very sharply and at high resolution at the same time. Therefore, in the past, a scanning electron microscope had to be used to scan the surface topography of rough and structured samples, even for low-resolution purposes. Thanks to the development of powerful digital cameras and corresponding software, the physically limited depth of field can now be considerably extended by the EFI (extended focus imaging) method. In this process, a data set of x images with different levels of sharpness is taken from a

Roger Dietrich: Paint Analysis © Copyright 2021 by Vincentz Network, Hanover, Germany

161

Part IV Methods of coating analysis strongly structured object. From this data set, special image processing software calculates a depth-of-field image of the entire object. The software module EFI (extended focus imaging) extracts the sharp details of different focal planes of an image series and combines them into a single image with infinite depth of field.

Figure IV.1: Painted key panel showing paint adhesion failure after laser treatment; A= light microscopy image of the border between lasered symbol and paint, B= EFI-3D image of the same area

Figure IV.2: Light microscopy image of a paint crater (left) and calculated EFI-3D image of the same paint failure

162

Optical light microscopy

Figure IV.3: Reflected light microscopy picture of a defect in a high gloss coating at 200x magnification (right bright field picture, left DIC picture)

Figure IV.4: Optical ray diagram of the differential interference contrast in reflected light microscopy

163

Part IV Methods of coating analysis It is thus an easy matter to obtain topographical images of extremely rough surfaces or paint failure without the help of scanning electron microscopy for magnifications of less than x1000. This additionally permits the layer thickness and topography to be measured.

1.2 Differential interference contrast (DIC) The DIC method (DIC or DIK = differential interference contrast) is used for the opposite case of very poorly structured samples where, especially at higher magnification, sufficient image contrast cannot be achieved with simple light microscopy. This principle goes back to an invention by Georges Nomarski and a patent from the 1950s [3]. The beam path is split by a so-called “Nomarski prism” into two coherent wavefronts of equal amplitude but polarization directions oriented perpendicular to each other. These hit the sample with weak surface structures and gain a path difference due to the small height differences (see before). After passing the sample, these wavefronts are again brought to interference. This creates a relief contrast in the result image, which makes it possible to make even weak surface structures visible, as shown in Figure IV.3a. The right image taken with bright field illumination does not show any contrast whereas the DIC image reveals, that there is a slight topographical difference. The DIC image creates the impression of an oblique object illumination. This technique is particularly suitable for the investigation of – very flat paint crater, – of scratches, scores, depressions and – minimal surface distortions.

2 Fluorescence microscopy A very useful technology with light microscopy for paint analysis makes use of the different fluorescence of different materials when illuminated with UV light. This is of particular interest when an unknown foreign component in a coating system needs to be localized. Instead of using visible light, the sample is illuminated with a mercury vapour lamp at wavelengths between 300 nm and 800 nm. This excitation raises molecules from a low-energy ground state to a more energetic excited state. When this excited molecule falls back to the ground state, a photon is emitted a few nanoseconds after the absorption of the UV light photon. The resulting fluorescent light is then focused and processed in the light microscope as usual. For example, black foreign inclusions in a black pigmented coating can be detected provided, that they exhibit a deviation in the Figure IV.5: Fluorescence microscopic image of a cross secfluorescence behaviour compared to tion of a defect in a single-layer coating system on a PCABS injection moulded polymer part the surrounding coating. Other effects

164

Part IV Methods of coating analysis It is thus an easy matter to obtain topographical images of extremely rough surfaces or paint failure without the help of scanning electron microscopy for magnifications of less than x1000. This additionally permits the layer thickness and topography to be measured.

1.2 Differential interference contrast (DIC) The DIC method (DIC or DIK = differential interference contrast) is used for the opposite case of very poorly structured samples where, especially at higher magnification, sufficient image contrast cannot be achieved with simple light microscopy. This principle goes back to an invention by Georges Nomarski and a patent from the 1950s [3]. The beam path is split by a so-called “Nomarski prism” into two coherent wavefronts of equal amplitude but polarization directions oriented perpendicular to each other. These hit the sample with weak surface structures and gain a path difference due to the small height differences (see before). After passing the sample, these wavefronts are again brought to interference. This creates a relief contrast in the result image, which makes it possible to make even weak surface structures visible, as shown in Figure IV.3a. The right image taken with bright field illumination does not show any contrast whereas the DIC image reveals, that there is a slight topographical difference. The DIC image creates the impression of an oblique object illumination. This technique is particularly suitable for the investigation of – very flat paint crater, – of scratches, scores, depressions and – minimal surface distortions.

2 Fluorescence microscopy A very useful technology with light microscopy for paint analysis makes use of the different fluorescence of different materials when illuminated with UV light. This is of particular interest when an unknown foreign component in a coating system needs to be localized. Instead of using visible light, the sample is illuminated with a mercury vapour lamp at wavelengths between 300 nm and 800 nm. This excitation raises molecules from a low-energy ground state to a more energetic excited state. When this excited molecule falls back to the ground state, a photon is emitted a few nanoseconds after the absorption of the UV light photon. The resulting fluorescent light is then focused and processed in the light microscope as usual. For example, black foreign inclusions in a black pigmented coating can be detected provided, that they exhibit a deviation in the Figure IV.5: Fluorescence microscopic image of a cross secfluorescence behaviour compared to tion of a defect in a single-layer coating system on a PCABS injection moulded polymer part the surrounding coating. Other effects

164

Infrared spectroscopy that are not visible under the light microscope, such as flow lines, can also be detected with fluorescence microscopy (see Figure IV.5).

3 Infrared spectroscopy Infrared spectroscopy (IR) is an analytical tool that has been well known for decades and found its way into routine work with the application of Fourier transform operations to data processing. The basic principle of IR spectroscopy is the structural characterization of materials through the absorption of infrared radiation by inter-atomic bonds. A defined wavelength range is scanned with infrared light to yield a collection of absorption information which can be displayed as bands in an infrared spectrum. The spectrum is evaluated by comparison with reference spectra and by examining the individual peaks to identify the various functional groups in the molecule or material, such as esters, hydrocarbons, acids, amines and the like.

3.1 Physical background The probe used in IR spectroscopy is radiation from the infrared region of the electromagnetic spectrum. This corresponds to energies between 0.001 and 1.6 eV. These photons excite characteristic vibrations of the inter-atomic bonds in a molecule. The energy needed to excite the vibrations is absorbed from the incident infrared radiation. As the bonds between different atoms have distinct bond energies, it takes characteristic energies to excite them; this is known as the “chemical shift”. These energies correspond to certain wavelengths of the infrared beam. To understand this mechanism, consider a simplified model of a chemical bond between two atoms in which two masses m1 and m2 are joined by a spring. This is called an “harmonic oscillator”. (Of course, a real inter-atomic bond is not a harmonic oscillator, but the model illustrates the basic principles very well.) The spring between the two masses (atoms) has a force constant k. If the two masses oscillate against each other, the frequency of the vibration can be described by Hooke’s law.

Figure IV.6: Basic principle of infrared spectroscopy

Figure IV.7: Simplified model of an inter-atomic bond consisting of two masses M1 and M2 joined by a spring

165

Infrared spectroscopy that are not visible under the light microscope, such as flow lines, can also be detected with fluorescence microscopy (see Figure IV.5).

3 Infrared spectroscopy Infrared spectroscopy (IR) is an analytical tool that has been well known for decades and found its way into routine work with the application of Fourier transform operations to data processing. The basic principle of IR spectroscopy is the structural characterization of materials through the absorption of infrared radiation by inter-atomic bonds. A defined wavelength range is scanned with infrared light to yield a collection of absorption information which can be displayed as bands in an infrared spectrum. The spectrum is evaluated by comparison with reference spectra and by examining the individual peaks to identify the various functional groups in the molecule or material, such as esters, hydrocarbons, acids, amines and the like.

3.1 Physical background The probe used in IR spectroscopy is radiation from the infrared region of the electromagnetic spectrum. This corresponds to energies between 0.001 and 1.6 eV. These photons excite characteristic vibrations of the inter-atomic bonds in a molecule. The energy needed to excite the vibrations is absorbed from the incident infrared radiation. As the bonds between different atoms have distinct bond energies, it takes characteristic energies to excite them; this is known as the “chemical shift”. These energies correspond to certain wavelengths of the infrared beam. To understand this mechanism, consider a simplified model of a chemical bond between two atoms in which two masses m1 and m2 are joined by a spring. This is called an “harmonic oscillator”. (Of course, a real inter-atomic bond is not a harmonic oscillator, but the model illustrates the basic principles very well.) The spring between the two masses (atoms) has a force constant k. If the two masses oscillate against each other, the frequency of the vibration can be described by Hooke’s law.

Figure IV.6: Basic principle of infrared spectroscopy

Figure IV.7: Simplified model of an inter-atomic bond consisting of two masses M1 and M2 joined by a spring

165

Part IV Methods of coating analysis _

√ ​

1   ​  k  ​ ​​ ​ϑ =  ​ _   ​ ​   _ 2𝝅c 𝝁

(Equation IV.1)

​m​  ​​  * ​m​  ​​

(Equation IV.2)

with 1 2    ​​​​ ​μ =  ​ _ ​m​ 1​​+ ​m​ 2 

If one mass of this harmonic oscillator is changed (i.e. one atom is replaced by a different one), clearly the frequency of the vibration must change. The molecular structure governs the type of vibrations which can be excited – stretching, rocking, or bending. The excitation of these vibrations follows certain selection rules: – Interaction of an inter-atomic bond with the radiation is only observed if the excited bond is a dipole, e.g. an -N-H bond. If a bond or molecule is completely symmetrical, it cannot be excited by infrared radiation. – Chemical bonds that represent strong dipoles absorb very strongly in the infrared spectrum. – The number of degrees of freedom and therefore the number of fundamental vibrations of a molecule consisting of n atoms is n = 3N-6 for a nonlinear molecule and n = 3N-5 for a linear molecule. Infrared methods can be categorised by the wavelength of the exciting beam, see Table IV.1. These different wavelength or wavenumber regions in the electromagnetic spectrum are used for specific applications in paint analysis. NIR mainly serves quality control purposes whereas MIR spectroscopy is used for structural identification. FIR doesn’t play a role in paint analysis. Having passed through the material, the infrared beam is analysed, and the transmitted portion of the infrared radiation is detected. The level of absorption at each infrared wavelength is recorded to yield the infrared spectrum. The infrared spectra of many organic and inorganic compounds are unique patterns of peaks and serve as a fingerprint for identifying such compounds. In mixtures, the intensity of the pattern produced by each compound is proportional to the concentration of that compound.

3.2  Characteristic absorptions For routine structural organic determinations, the most important absorptions in the infrared region are the simple stretching vibrations. The stretching vibrations of typical organic molecules tend to fall within distinct regions of the infrared spectrum, as shown below. To illustrate the above-mentioned rules, consider the spectrum of methanol: Starting at 4000 cm-1, which is very typical of mid-infrared spectra, the spectrum displays a broad peak between 3600 cm-1 and 3000 cm-1 that can be attributed to the O-H stretching vibrations of the hydroxyl groups. The C-H stretching vibrations of the methyl group appear at a lower wavenumber between 3000 cm-1 and 2800 cm-1. They split up into asymmetric and Figure IV.8: Stretching and bending vibrations of a symmetric vibrations that can be dis– CH2-group

166

Infrared spectroscopy Table IV.1: Wavenumber range of the infrared Infrared Wavelength

Wavenumber

Near infrared, (abbr.: NIR)

λ = 760 to 2550 nm

λ = 13,200 to 4000 cm-1

Mid infrared, (abbr.: MIR)

λ = 2.55 μm to 25 μm

λ = 4000 to 400 cm-1

Far infrared, (abbr.: FIR)

λ = 0.025 mm to 1 mm

λ = 400 to 10 cm-1

Table IV.2: Examples of characteristic IR absorptions (selection) Characteristic Substances Functional group absorption (cm-1)

Typical paint compounds

Alkyl C-H stretch

2950 to 2850 (m or s)

alkanes or alkyl side chains

waxes, hydrocarbon solvents

Alkenyl C-H stretch alkenyl C=C stretch

3100 to 3010 (m) 1680 to 1620 (v)

unsaturated hydrocarunsaturated alkyd resins bons or substances with unsaturated side-chains

Aromatic C-H stretch

~3030 (v)

aromatic hydrocarbons

solvents such as toluene, aromatic resins

O-H stretch C-O stretch

3550 to 3200 (broad, s) 1000 to 1200 (s)

alcohols, polyesters, polyols

polyester resins, butanols, PEG, PPG

N-H stretch

3500 to 3300 (m)

amines, polyurethanes

HALS, DABCO, 2-pack PUR resins

Figure IV.9: Infrared spectrum of methanol

167

Part IV Methods of coating analysis Table IV.3: Assignment of characteristic signals in the infrared spectrum of methanol Type of vibration Description Peak assignment [cm-1] ν, CHx

stretching vibration of the methyl group

2943 2831

δ, CH2

H-C-H deformation vibration, bending of the methyl group

1448

ν, OH

stretching vibration of the hydroxyl group

3316

ν, OC

stretching vibration of the C-O group

1010

tinguished in the spectrum. Further down the wavenumber scale, the bending vibrations of the methyl group can be detected at about 1448 cm-1. Finally, the C-O stretching vibration of the hydroxyl group is to be found at 1010 cm-1. This simple example demonstrates how the structure of a molecule (which may be a solvent) correlates with its infrared absorption spectrum.

3.3  Instrumentation The technical principle behind infrared spectroscopy is to compare the intensity of an infrared beam which has passed through a sample with a second one (called the reference beam) which has had no interaction with the sample. In the past, this was achieved with a two-beam alignment. Nowadays, the technique is based on a Fourier transform of one beam that has been divided by a beam splitter before passing through the sample. The device used for this is called a Michelson interferometer and it works as follows: An interferometer is an optical device that splits a light beam into two beams and then recombines them after each has travelled along a unique path. One of the split beams travels to a fixed mirror and runs back to the beam splitter, the second beam is reflected by a movable mirror which is tuned. Both partial beams interfere at the position of the beam splitter and are guided to the sample position (see Figure IV.10). When this radiation recombines in the FT-IR spectrometer, a complex beam of oscillating intensity is generated. This periodic change in beam intensity happens because one light path in the interferometer is constantly changing. Application of a Fourier transform to this time-varying intensity produces the infrared spectrum. The sample is placed in a sample chamber of the FT instrument. The sample chambers are

Figure IV.10: Principles of Fourier transform infrared spectroscopy

168

Infrared spectroscopy designed to fit a variety of sampling tools, such as the above-mentioned transmission experiment, ATR-FT-IR and IRRAS (infrared reflection absorption spectroscopy) accessories. Data acquisition, processing and evaluation are normally performed with the aid of software solutions which are distributed by the instrument suppliers.

3.4  Sample preparation The classic transmission infrared spectroscopy is characterized by transmitting the infrared beam perpendicularly through a thin plane-parallel layer of the analyte in the mid-infrared region (MIR). With solid material, a very small amount of the sample is mixed with a surplus of a matrix substance that exhibits little interaction with the infrared beam, e.g. KBr or PE, and pressed into a disc. The so called KBr disc is placed in the beam such that the beam passes through the disc in a rectangular shape. Liquids such as solvents or diluted binders are placed between two thin discs of KBr or NaCl or filled into special cuvettes which are transparent for infrared light (glass totally absorbs infrared radiation). The liquid in the cuvette forms a thin layer which can transmit the infrared beam. This preparative technique can be used for: – analysing solvents and binders, – characterizing clear coats, – identifying spots in clear coats, – identifying contamination or segregation layers on paint surfaces. These old-fashioned methods are not state of the art. The above mentioned sample preparation methods have been nearly completely replaced by reflection techniques like ATR-FT-IR, IRRAS1 or DRIFT2 (see in the following).

1 IRRAS: infrared reflection absorption spectroscopy 2 DRIFT: diffuse reflection Fourier transform spectroscopy

Figure IV.11: General setup for transmission infrared spectroscopy

169

Part IV Methods of coating analysis For the analysis of pigmented and filled paints (which in fact represent the majority of samples), ATR-FT-IR (an infrared reflection method) is ideal. The advantages are obvious. No sample preparation is needed, and absorbing samples can be analysed without difficulties.

3.5  Spectrum representation The acquired sample spectrum I(ν) displays the – emission spectrum of the IR source, – transmission of the optical components of the spectrometer, – characteristic of the detector and, of course, – absorption by the sample. So, there is a great deal of information in the spectrum that has nothing to do with the chemical composition of the material which is wanted to analyse. To get rid of this data, the measured spectrum I(ν) is divided by a reference spectrum I0(ν). This reference spectrum, which is normally measured with a beam which does not pass through a sample, includes all the non-sample specific information. The resulting transmission spectrum T(ν) contains information on the chemical composition of the sample: ​T(ν) = I(ν)/I0(ν)

(Equation IV.3)

A more versatile plot of an infrared spectrum is its absorbance spectrum A(ν), which is derived from the transmission spectrum by: A(ν) = log 1/T(ν)

(Equation IV.4)

The absorbance A or extinction E is proportional to the sample concentration and the sample thickness and therefore can be used for quantification purposes.

3.6  Quantification Quantitative determination of the concentration of a substance requires a linear relationship between the sample concentration and sample thickness on one hand and the detected extinction or

Figure IV.12: Transmission and absorbance spectra of a paint matting additive

170

Infrared spectroscopy absorbance on the other. The linear relationship between absorbed radiation and substance concentration for the transmission experiment is given by the Lambert-Beer law: ​I​  ​​

​​  _I0 ​   =   αcd = ​Eλ​  ​​​

(Equation IV.5)

where: Eλ is the absorbance and α is the extinction coefficient Quantitative analysis means connecting the concentration c of an analyte to the intensity of a characteristic peak of the analyte by the absorbance Eλ. The thickness of the analysed layer is defined, for example, by the width of the infrared cell as far a classical transmission experiment on a solvent is concerned or the thickness of the KBr disc when it comes to solids. But what does this mean for the actual analysis? For example: the concentration of a paint solvent such as butyl acetate in a mixture of paint solvents has to be determined. The proceeding starts with a calibration curve. At least ten samples (better more) containing a known concentra-

Figure IV.13: Quantification of the solvent content (isopropanol) of a clear coat: a) Infrared spectra after repeated addition of IPA b) normalized intensity of the 950 cm-1 peak of IPA after repeated addition of 100 ml solvent

171

Part IV Methods of coating analysis tion of the analyte have to be thoroughly prepared. The second step is to record the absorbance spectra of these ten standards. From the averaged spectra, a characteristic peak of reasonable intensity is picked which does not interfere with peaks from the rest of the sample. The intensity of this peak is measured by integrating the peak area or the peak height. The intensity data can now be plotted versus the known concentration of the analyte in an x/y chart. Linear regression of the data points should yield a straight line indicating a linear relationship between concentration and peak intensity. The correlation between the slope and the intercept of this line fit can now be used to determine the concentration of the solvent in an unknown mixture by relating the peak intensity of the analyte of the unknown mixture to the calibration curve [4]. Example: Figure IV.13a shows the spectra of 3.84 g of a polyester clear coat which has been added equal portions of isopropanol between 5.4 % and 31.4 % (weight). The solvent can be identified by the characteristic signal of isopropanol at 950 cm-1 which has only a small overlap with the peaks of the polyester which exhibits very characteristic absorption at 1059 cm-1. The peak intensity of the 950 cm-1 signal of IPA and the 1059 cm-1 peak of the clear coat has been determined by integration of the peak area. The normalized intensity of the 950 cm-1 IPA band 3 has been correlated to the amount of IPA added repeatedly (Figure IV.13b). The determination coefficient R2 of 0.985 proves a sufficient linear correlation of this quantification.

3.7 Data analysis and evaluation Nowadays, instrumentation for performing all kinds of analysis, and especially for infrared analysis, is easy to handle. It is easy to produce a spectrum after brief instruction and without any knowledge of the underlying principles. On the other side of the coin, it is easy to get the wrong result. Correct interpretation and evaluation of spectra requires fundamental understanding and experience. The spectrometer industry provides customers with plenty of computer assistance and software that supports to do the job. But, in the end, an experienced spectroscopist is needed. The easier automatic interpretation by software tools seems to be, the more likely it is that the result is wrong.

3

Figure IV.14: Topcoat test panel showing adhesion deficiencies (top): ATR-FT-IR spectra of the backside of a detached paint chip compared to a spectrum of the paint bulk (bottom)

172

​  ​​ IPA IPA I n​  IPA ​  ​ =  ​​I_ ​I​ ref​​ ​​  (with ​​I​ n​  ​​= normalized intensity of the 950 cm-1 signal of IPA; ​​ I​ IPA​​ ​= integration of the intensity of the 950 cm-1 IPA signal; I​ ​​ ref​​ ​integration of the intensity of the 1059 cm-1 signal of the clear coat)

Infrared spectroscopy But what is the right way to properly evaluate a spectrum? There are several. The first one is to generate a table of all peaks and correlate it with peak tables that are available in book form or computerised format. This peak table can normally be generated automatically by a software tool. Comparison with the acquired spectrum of the sample leads to the identification of major functional groups, such as esters, acrylates, urethanes, hydrocarbons, and so on. This is actually quite general information that will not lead very far in most cases. Peak assignments are quite often very ambiguous. For example, polyesters and acrylates share several characteristic peaks, although, they are different as far as their chemical backbone is concerned. To learn more from the spectra, it is necessary to compare the data to databases of infrared spectra either bought from a supplier or collected by the company itself.

3.7.1

Data processing

The evaluation of infrared spectra quite often involves accessing small spectral differences that are not visible on the first view. Fortunately, modern spectroscopy software allows for data processing to work out and extract the spectral features that are needed. Figure IV.14 shows a test panel with a black coating that obviously exhibits adhesion problems. The question is why this happened and where the separating plane is located. This means that the analytical task is to distinguish if the paint suffers a cohesion deficiency or if the coating lifts off the polymer for some reason. The comparison of an ATR-FT-IR spectrum of the backside a delaminated paint chip and the paint bulk is not very perceptive on the first view. Both spectra seem to be identical. Not until a spectral subtraction has been calculated (absorbance spectrum of the backside of the delaminated paint chip – paint bulk reference spectrum) (see Figure IV.15) it becomes visible that there is spectral difference that can be attributed to the polymer.

Figure IV.15: Result of the spectral subtraction (blue) compared to a reference spectrum of the polymer (red)

173

Part IV Methods of coating analysis This spectral subtraction revealed that a very thin layer of the polymer has been delaminated together with the paint. So, the weak boundary is not the paint but the uppermost layers of the test panel. Without this information the coating manufacturer might change the recipe of his paint for no reason assuming, that the coating formulation is not appropriate. This example demonstrates how helpful data processing can be, but it has to be mentioned that a rash use can lead to false results and evaluations. So, some experience is helpful to perform mathematical operations on spectra without falsifying the results or producing artefacts. Other operations offered by spectroscopy software are e.g. spectral addition, averaging, atmospheric corrections, Kramers-Kronig (see Chapter 4.2.2, Part IV) or Kubelka-Munk conversions.

3.7.2 Use of databases One advantage of infrared spectroscopy is the extensive use and knowledge fixed in numerous scientific publications. The experience of decades with this technique and the wide use as a routine tool in the chemical and coating industry lead to very comprehensive databases of reference substances that are commercially available. In addition to that companies using infrared spectroscopy on a regular basis often have collected their proprietary databases (which is a good choice). Databases can help to identify unknown substances or to check the quality of incoming goods. However, a database, even if supplied with elaborate software, cannot do everything. It is simply a collection of spectra that someone has measured and collected at a certain time. These spectra have been rendered searchable by means of digitization. Therefore, they are nothing more than digits in a data file. The search software uses a mathematical algorithm to compare the digitized data of the reference spectrum with the digitized data of the unknown spectrum. This might, for example, take the form of a comparison of peak intensity via the correlation coefficient of a linear regression. The result of this mathematical operation is a probability that the unknown spectrum has something to do with the saved reference spectrum. It is very important to realize that this mathematical result has nothing to do with the actual chemical composition of the sample. This is a common source of confusion because it is easier to present a list of probabilities than to think about the problem and analyse the probability that the mathematical result is the desired answer. The computer-generated hit list displays the degree of mathematical correspondence between the saved spectra and your data. So, what you get from computer analysis is a hit list of probabilities that has to be checked thoroughly for plausibility. All available information about the sample therefore must be taken into account. This includes the following: – Is the sample a liquid, a solid, a paste or a powder? – What circumstances led to the problem? – Is the sample a mixture or a pure substance? A chemical background is therefore needed for an understanding of the spectra. If, for example, a water-borne binder has to be analysed by horizontal ATR (H-ATR) (see Figure IV.25) and it is analysed as received, more than 90 % of the ATR-FT-IR spectrum will be dominated by the characteristic peaks of water. The database will tell you that there is a 90 % probability that the sample is water. An inexperienced spectroscopist might therefore conclude that this is the wrong sample unless he knows that he has to wait until the water dries off before the polymer itself is measured. Figure IV.16 shows an example of a 2-pack polyurethane paint. The commercially available database yields a list of suggestions. The first three of these are shown in Figure IV.16. The “best” search result from the standpoint of mathematical compliance shows the spectrum of a vinylidene

174

Infrared spectroscopy chloride polymer, which is completely misleading. The spectrum of a compound consisting of polyether-polyurethane is ranked only as second best and even the third proposal has nothing to do with the sample’s real chemical composition. This, of course, is a dreadful outcome and if the result of such a research would to be published, that would cause a furore or laughter. This demonstrates that a computer, which is unable to evaluate the search result, is no substitute for a reasonable amount of experience. A second problem with commercial databases is their “expiry date”. As quickly as commercially available products or raw materials are measured and stored in the database, they are either taken off the market or their brand name is changed just as the database is published. This means

Figure IV.16: Example of a database search: sample spectrum showing a polyester polyurethane (above) and the database results derived from a commercial database below

175

Part IV Methods of coating analysis that a search result is useless for identifying an unknown compound. Companies that produce the spectra often take purchased products from the market, measure the ATR spectrum and save it in the database under the brand name without knowing anything about the chemical composition. For practical use in a production laboratory, it might therefore be more useful to create a database of the products and raw materials used in the process rather than to buy one.

4 Surface infrared spectroscopy Standard absorption/transmission infrared spectroscopy of thin layers (KBr disc or liquid cell) is useful for identifying and characterizing even low concentrations of substances. However, it has one disadvantage: it is not surface sensitive, since the infrared beam passes through the sampling area in a rectangular manner. When it comes to paint samples very often surface phenomena are important, which cannot be resolved by transmission techniques. For these cases, surface-sensitive infrared reflection techniques, such as ATR-FT-IR, permit analysis of very thin surface layers, of thin layers of filled and pigmented paints, and of the surface quality of the substrate to be painted. They all have a low detection limit and a high surface sensitivity. When switching from transmission to reflection spectroscopy, the intensity of an infrared beam reflected by a sample surface is compared with that of a reference beam that has not interacted with a sample. However, in contrast to the transmission method, there are some physical processes and parameters (such as refraction, multiple reflection, and polarisation) that influence the appearance of an infrared reflection spectrum. To distinguish between chemical effects in the infrared spectrum from physical effects such as asymmetric peaks and peak inversions, it is useful to know something about the “behaviour” of infrared radiation passing through the interface between two media of different refractive index. The methods which will be described now are based on external reflection (such as RAS and infrared microscopy) and internal reflection (ATR-FT-IR). It is therefore necessary to deal with the

Figure IV.17: The general principle of surface infrared spectroscopy (see excitation and vibration of surface related chemical bonds)

176

Part IV Methods of coating analysis that a search result is useless for identifying an unknown compound. Companies that produce the spectra often take purchased products from the market, measure the ATR spectrum and save it in the database under the brand name without knowing anything about the chemical composition. For practical use in a production laboratory, it might therefore be more useful to create a database of the products and raw materials used in the process rather than to buy one.

4 Surface infrared spectroscopy Standard absorption/transmission infrared spectroscopy of thin layers (KBr disc or liquid cell) is useful for identifying and characterizing even low concentrations of substances. However, it has one disadvantage: it is not surface sensitive, since the infrared beam passes through the sampling area in a rectangular manner. When it comes to paint samples very often surface phenomena are important, which cannot be resolved by transmission techniques. For these cases, surface-sensitive infrared reflection techniques, such as ATR-FT-IR, permit analysis of very thin surface layers, of thin layers of filled and pigmented paints, and of the surface quality of the substrate to be painted. They all have a low detection limit and a high surface sensitivity. When switching from transmission to reflection spectroscopy, the intensity of an infrared beam reflected by a sample surface is compared with that of a reference beam that has not interacted with a sample. However, in contrast to the transmission method, there are some physical processes and parameters (such as refraction, multiple reflection, and polarisation) that influence the appearance of an infrared reflection spectrum. To distinguish between chemical effects in the infrared spectrum from physical effects such as asymmetric peaks and peak inversions, it is useful to know something about the “behaviour” of infrared radiation passing through the interface between two media of different refractive index. The methods which will be described now are based on external reflection (such as RAS and infrared microscopy) and internal reflection (ATR-FT-IR). It is therefore necessary to deal with the

Figure IV.17: The general principle of surface infrared spectroscopy (see excitation and vibration of surface related chemical bonds)

176

Surface infrared spectroscopy physics of transmission of infrared radiation in different optical media and the reflection and refraction of light at the interface of the sample and, for example, air.

4.1 ATR-FT-IR spectroscopy Paints are normally very strong absorbers of infrared waves. Consequently, they do not transmit infrared radiation and therefore are not amenable to analysis by infrared transmission techniques. The best way to characterize these layers is to use internal reflection infrared spectroscopy, called MIR, ATR, or FTR. Attenuated total internal reflectance (ATR) spectroscopy is a versatile and powerful technique for infrared sampling. It makes for rapid analysis as little or no sample preparation is usually required. ATR is ideal for those materials which are strong absorbers. In addition, ATR provides useful information about the surface properties or conditions of a material. The infrared light is transmitted by a crystal made of infrared-transmitting material which is in physical contact with the sampling area. The interface between the sample and the ATR-FT-IR crystal is where interaction of the infrared beam takes place. The sample molecules of the contacted surface absorb infrared radiation and characteristic vibrations of the inter-atomic bonds are excited in line with the selection rules (see Chapter 3.1 ) discussed earlier. In contrast to the classical transmission technique, only a few microns of the surface of the material are involved in this process, not the whole material. Thus, even strongly absorbing samples can be characterized without difficulty, a fact which is very useful for all kinds of paints, and especially for water-borne paints that cannot be analysed by infrared transmission spectroscopy. The theoretical background for this method was described by Harrick in the 1960s [5]. In contrast to classical infrared spectroscopy which transmits a thin layer of the target material, this technique is a reflection method, which does not irradiate the entire sample material but only areas near the surface. For this purpose, the sample area to be examined is brought into contact with the surface of a crystal, which acts as a light guide for the

Figure IV.18: Basic principle behind ATR-FT-IR spectroscopy

177

Part IV Methods of coating analysis

Figure IV.19: Internal (left) and total internal reflection (right) of an IR beam at the interface between ATR crystal and sample

Figure IV.20: Evanescent infrared wave in the case of total reflection at the sample/ATR crystal interface

Figure IV.21: Absorption and ATR-FT-IR spectrum of a polyethylene layer

178

Surface infrared spectroscopy IR radiation. At the contact area between the surface of the optically denser material (crystal) and the target surface the infrared beam interacts with a thin layer (thickness > dp, is the effective path length de is. All parameters which cause an increase in the effective path length lead to a better detection limit. – If the thickness of the detected layer is a lot lower than the effective path length (d 15 μm

Detection limit

0.1 to 1 At.%/fractions of a molecular layer

Quantification

possible, but not easy

Samples

solid/vacuum paint

Sample size

8 to 20 mm

Analysis of insulating samples

possible with charge compensation

Time per measurement

2 min to 1 h

Time per analysis

4 h to 1 d

Availability

good

Data evaluation

advanced, commercial databases available

Available information

molecular and atomic composition of surfaces

Vacuum

UHV

Sample preparation

minimum

photoelectrons. If the energy distribution is wider than the chemical shift of the peak caused by neighbouring atoms, it makes no sense to measure it. As mentioned above, an important parameter that influences the energy distribution of the XPS peak is the X-ray source. In order to ensure that the energy distribution of the source is as narrow as possible, aluminium or magnesium is used as anti-cathodic material (FWHM 0.85 and 0.7 eV). The most important part of the XPS instrumentation is, of course, the analyser, whose task is to determine the electron energy. Two different designs on the market are called CMA and CHA. This analyser consists of two panels (CMA) or two concentric hemispheres held at different potential. Through focusing and retarding lenses and a slit, the photoelectrons enter between the two hemispheres and are retarded by the potential of the inner surfaces. Only photoelectrons that have a specific kinetic energy, the so-called pass energy, can travel through the gap between the hemispheres. Varying the retarding potential enables the whole XPS spectrum to be recorded. However, a minimum energy distribution of each measured signal cannot be avoided. The accuracy of the energy determination is given by the FWHM of the peak dE, which should not exceed 0.5 eV.

10.7 Applications Due to their high surface sensitivity, XPS and TOF-SIMS are ideal for studying surface phenomena, such as adsorption, desorption, and corrosion, along with catalytic processes and surface reactions. They also serve as additional techniques for thin surface layers, such as paints. Like the other techniques, XPS only needs small amounts of analyte. One of the practical disadvantages is

253

Part IV Methods of coating analysis the fact that it is often necessary to prepare a significant amount of standard or reference samples to solve more complicated problems. In paint analysis, XPS has been successfully used to: – study the surface treatment and coating of effect pigments – detect and identify surface contaminants of substrates – identify and quantify additives in the uppermost layer of paints and coatings

10.8 Technical data see Table IV.16: Technical data of XPS methods

11

GC-MS

GC-MS is not a surface sensitive method but widely used in the coating industry. Especially Headspace GC-MS analysis is a standard technique used to detect and identify the volatile components (VOC) of a paint. Coatings contain a large number of components with different properties, many of which are not accessible to direct mass spectrometric analysis due to molecular weight, polarity or volatility. While the non-volatile substances can be partially identified by time-of-flight mass spectrometry and infrared spectroscopy, the volatile components are not detectable in TOF-SIMS analysis, for example, due to the fact that this is a vacuum technique. Headspace GC-MS fills this analytical gap. Due to the large number of details in the application, only the essential aspects of the headspace GCMS method and its application to Figure IV.100: The GC-MS method combination paints are described below. Further literature can be found in detailed process descriptions [60]. It is important to mention that GC-MS in fact are two methods GC and MS. The first one is a separation technique that delivers the components of a mixture separated into different fractions. It cannot be used for identification unless the second one, the molecular identification by mass spectrometry (or other detectors like flame ionization detector) is combined with it. In the last 25 years, the author has often seen a GC chroma­ Figure IV.101: General principle of the GC experiment togram presented in a report that

254

Part IV Methods of coating analysis the fact that it is often necessary to prepare a significant amount of standard or reference samples to solve more complicated problems. In paint analysis, XPS has been successfully used to: – study the surface treatment and coating of effect pigments – detect and identify surface contaminants of substrates – identify and quantify additives in the uppermost layer of paints and coatings

10.8 Technical data see Table IV.16: Technical data of XPS methods

11

GC-MS

GC-MS is not a surface sensitive method but widely used in the coating industry. Especially Headspace GC-MS analysis is a standard technique used to detect and identify the volatile components (VOC) of a paint. Coatings contain a large number of components with different properties, many of which are not accessible to direct mass spectrometric analysis due to molecular weight, polarity or volatility. While the non-volatile substances can be partially identified by time-of-flight mass spectrometry and infrared spectroscopy, the volatile components are not detectable in TOF-SIMS analysis, for example, due to the fact that this is a vacuum technique. Headspace GC-MS fills this analytical gap. Due to the large number of details in the application, only the essential aspects of the headspace GCMS method and its application to Figure IV.100: The GC-MS method combination paints are described below. Further literature can be found in detailed process descriptions [60]. It is important to mention that GC-MS in fact are two methods GC and MS. The first one is a separation technique that delivers the components of a mixture separated into different fractions. It cannot be used for identification unless the second one, the molecular identification by mass spectrometry (or other detectors like flame ionization detector) is combined with it. In the last 25 years, the author has often seen a GC chroma­ Figure IV.101: General principle of the GC experiment togram presented in a report that

254

GC-MS served as proof that a particular substance was detected. But this is wrong as discussed in the following chapters.

11.1 Physical background of GC Gas chromatography (GC) like other chromatographic methods (e.g. liquid chromatography or thin layer chromatography) is a technique to separate a substance from a complex mixture. This is achieved by selective partitioning the gas phase of the mixture of volatile substances between a mobile phase which is a carrier gas and a stationary phase in a so-called column while this column is heated. The separation in the column is driven by the different partition equilibrium of the analyte between the two phases. The partition coefficient α is the equilibrium constant of the partition equilibrium according to the Nernst law: ​C​  ​​

a   ​​ ​α  = ​ _ ​C​  ​​ b

(Equation IV.36)

Figure IV.102: Separation of an analyte by the distribution between two phases A and B

Figure IV.103: The chromatogram

255

Part IV Methods of coating analysis α depends on the kind of phase, the temperature, the solubility in the phases, the vapour pressure and the pressure. At the end of the column the detector writes a transcript of the chromatographic process the socalled chromatogram. The separated substances which sequentially leave the column can either just be detected (e.g. by a flame ionization detector or a thermal conductivity detector). This type of detectors log the different fractions leaving the column, the retention time and the intensity. They do not identify the fractions. For the molecular identification the GC has to be coupled to a mass analyser (GC-MS) or an infrared spectrometer (GC-FT-IR). The following chapters deal in particular with headspace GC-MS, a technique to separate, identify, and quantify the solvent components of a liquid as well as cured coating films.

11.2 Headspace In fact, the headspace GC MS is a combination of three methods: – an evaporation process based on vapor pressure (headspace) which separates the volatile from the non-volatile components, – a chromatographic separation of the gas phase (GC) based on different polarities and vapor pressure between a mobile phase and a stationary phase, and – an identification of the components released after the separation from the so-called separation column by mass spectrometry (MS). It is very important to understand what happens to the paint during this investigation. From the entire paint sample, the portion that is volatile under the given analytical conditions is transferred to the gas phase. However, this is not necessarily the total content of volatile components. From this selection, a further separation step then “chromatographs” the portion that runs over the column under the given analysis conditions and leaves the column again. Depending on the column and temperature, this is again only a part of the substance that has entered the gas phase in the headspace. In the last step, the substances leaving the chromatographic separation column are identified by mass spectrometry (as far as the components are available in the database). The result is, therefore, by no means an exact image of the true composition of the paint.

Instrumentation

Figure IV.104: Setup of headspace GC-MS measurement

256

GC-MS

Sample preparation

In the first step, a small amount of the paint sample (e.g. 1 g) is placed in a glass vial (headspace vial) and diluted if necessary (e.g. with acetonitrile) and, if necessary, provided with an internal standard. The glass vial is closed with a cap with septum and brought to an elevated temperature (e.g. 150 °C). During this process the volatile components change into the gas phase and an equilibrium is established. The amount of analyte (here for example the solvent) in the gas phase is determined by the partition coefficient K and the relative volume of the gas space above the paint sample. The partition coefficient is different for each substance. The lower the K value, the easier it is for a solvent to enter the gas phase. For example, the K value for toluene in an air/water system at 40 °C is 2.82 for toluene but 1355 for ethanol, which means that at a given temperature significantly more toluene than ethanol will enter the gas phase from a water-based paint. However, this means that the detection limit for toluene is significantly lower than for ethanol. By increasing the temperature, the K value can be reduced, which means that a larger proportion of the volatile substance enters the gas phase and the detection sensitivity is improved. After the equilibrium is set, a sample is taken from the gas space above the liquid with a syringe. This can be done in two ways. 1. Manual sampling by a gas-tight syringe and manual injection on the separation column: After the equilibrium between liquid/solid phase and gas phase is established in the closed headspace glass, a gas-tight syringe is inserted through the septum and an aliquot of the gas phase is taken. The syringe is withdrawn and inserted into the port of the gas chromatograph. There the sample of the gas phase is injected as quickly as possible. Care must be taken that the gas phase does not condense in the syringe due to temperature differences between the syringe and the headspace oven. This can be ensured by using a heated syringe, for example. 2. Automatic sampling: With this method, a needle is inserted through the septum into the gas space of the headspace glass during the equilibrium adjustment. This needle is used to pressurize the gas space with a carrier gas after equilibrium is set. The gas sample consisting of carrier gas and gas phase to be analysed is then introduced into the column of the gas chromatograph via a valve which opens and a transfer bridge. This procedure is called “Balanced pressure system”. This automatic sampling can also be done via a so-called “Pressure loop” [59].

GC separation

This sample of the gas phase saturated with the volatile components of the paint is injected into a heated separation column in the next step (GC). This is a thin tube, e.g. made of quartz glass, which is coated on the inside with a so-called stationary phase. This phase takes over the actual task of separating the individual components of the mixture of the gas phase. After addition the substances to be analysed, the mixture to be separated run over the stationary phase in a gas stream. As it passes through this column, the individual components of the mixture to be separated are distributed between the stationary and mobile phases. Depending on polarity and vapor pressure, different components remain on or in the stationary phase for different lengths of time and are thus separated from each other. Numerous parameters influence the result of the separation [61]. These include, for example: – injection speed, – injection pressure, – needle temperature or – injector temperature.

257

Part IV Methods of coating analysis – – – – –

choice of stationary phase column internal diameter column length film thickness column temperature…

Not all substances in the gas phase can be separated with a stationary phase. For example, the stationary phase – methyl polysiloxane is used for the separation of hydrocarbons, amines, waxes and phenols, – polyethylene glycol for amines, acids and alcohols and – a mixture of methylpolysiloxane and phenylpolysiloxane for alcohols, acetates and phenols. The injected sample is transported over the column by an inert carrier gas while the separation takes place on the stationary phase. During the separation the temperature of the oven in which the column is located can be varied. Certain separations require successive temperature stages of different duration. Some components such as alcohols or amines are not so easy to measure simply because of their reactive groups, e.g. because they can react with the surface of the injection port or the column of the gas chromatograph [62]. Here a chemical reaction (derivatisation) may have to be used.

Detection

When the separated individual components leave the column, they are identified one after the other in a detector. This can be a: – flame ionization detector (FID) or a – flame photometric detector (FPD), – electrolytic conductivity detector (ELCD) – electron capture detector (ECD), – thermal conductivity detector (TCD) – mass spectrometer. The method is quantifiable. But the methods and prerequisites for quantification and the advantages and parameters of the above mentioned detector types would lead too far for this book. A good overview can be found in the book of Bruno Kolb and Leslie S. Ettre [59].

11.3 Data evaluation As a rule, the mass spectra obtained are automatically compared with stored database spectra according to a defined algorithm. The (mathematical) degree of agreement is output in the form of a so-called “hit-quality” index as a digit between 1 % and 100 % or 0 and 1. This often makes it difficult to evaluate and use the analysis results. But does it mean when a compound is identified with a probability of 72 %, for example? This is left to the opinion of the analyst and is therefore subjective. Furthermore, there is no plausibility check. In paint samples, for example, components are often “identified” that are completely nonsensical and are never used in a paint. So much for the theory of this procedure. In detail, however, there are a few pitfalls that must be considered, especially when testing coatings. In the first place, it is important to understand that the result of a headspace GC-MS analysis does not reflect the complete composition of a paint as well as any other method. Rather, the result reflects which of the (volt-

258

GC-MS Table IV.18: Examples of GC-MS parameters of typical paint components [64]

Temperature program

Substance

Carrier gas Column

Injection

Acrylates

hydrogen

cross-linked PEG-TPA

1 µl, directly on the column

1 min at 35 °C, annealing to 60 °C, rate 10 °C/min, annealing to 160 °C rate 15 °C/min

Glycoles/dioles helium

cross-linked methyl silicone

on the column, 1 µl, inlet program 50 °C -> 180 °C during 8.45 min, heating rate 12 °C/min

3 min at 50 °C, annealing to 180 °C heating rate 8 °C/min

Esters

cross-linked PEG

split (25:1), 1 µl, inlet 250 °C

1 min at 45 °C, annealing to 200 °C heating rate 5 °C/min

helium

ile) fractions of the coating arrive at the back of the mass spectrometer after headspace separation and chromatographic separation. And this proportion of the total sample depends on various factors: – Headspace temperature: Here the author would like to add a quote from an application note of a GC-MS manufacturer, which makes it very clear how the performance of a headspace GC-MS analysis influences the result. It says with respect to the detection of plasticizers in wall paint and the influence of the headspace temperature: “Depending on the range of volatiles requested, the user can select the headspace temperature to achieve the desired results.” [63] This means that the choice of headspace temperature e.g. can influence the final result of the analysis in almost any way. Conversely, this means that when analysing an unknown mixture of substances without prior knowledge of the sample, one analysis is not sufficient, but rather an attempt is made to limit the required temperature range by varying the headspace temperature. – GC separation: It is mentioned above, that not all substance classes can be reliably separated with one column and a temperature program. In the case of a mixture of substances, this may require several runs with different parameters to get closer to the true composition. In addition, coatings often contain components such as glycols, which are difficult to separate using a column due to their properties [65]. Polyethylene glycols, for example, also serves as a stationary phase and must therefore not be mobile along the column. This means that for a complete GC-MS analysis, it is important to know in advance what to look for and then be able to select the column and temperature program. And it means that the analysis of a completely unknown sample will be more or less difficult or expensive. This method is very well suited to compare e.g. two batches of a lacquer, a preliminary product or an additive.

11.4 Application Example: Comparison of two paint batches Task: Two batches of a base coat that behave differently during paint application should be compared with regard to their volatile components.

259

Part IV Methods of coating analysis Table IV.19: Results of the identification of volatile organic compounds of a base coat by headspace GC-MS screening (i: Fit >90 %: excellent match; Fit, >80 % to 90 % good match; Fit 90 %: excellent match; Fit, >80 % to 90 % good match; Fit 2

Imaging of element distribution

no

yes

no

no

yes

yes

Topography imaging

yes

no

no

yes

no

no

Chemical (molecular) information

no

no 10

yes

no

no

yes

Lowest grasped sampling depth

upper0.2 to 1 μm a few 3 nm most 100 nm 11 mono­layer

0.5 nm

0.5 to 1 nm

Biggest available sampling depth 12

a few mono­ layers

2 μm

a few μm

10 nm

5 to 10 nm

Spatial resolution

better 1 nm

1 μm

~ 2 mm 13 3 nm

better 1μm

approx. 5 μm

Quantification

no

yes

yes 14

no

yes14

yes

Depth profiling of elemental composition

no

no

no

no

yes

yes

Typical measuring area

0.01 to 10,000 μm2

100 μm2

2x2 mm 15

100 μm2

50 μm2

100 μm2

1 μm

10

only in special cases depending on the refractive indices of the ATR crystal, substrate and the angle of incidence means the depth of information directly accessible with the method. With additional etching devices, test cuts, etc., in principle any depths can be reached 13 depending on the contact area of the ATR crystal being used 14 with limitations, a relative quantification is possible 15 with single bounce diamond ATR 11 12

263

Part IV Methods of coating analysis

13 References [1] S. Bradbury, B. Bracegirdle, Introduction to Light Microscopy, Oxford: BIOS Scientific Publishers Ltd, 1998. [2] D. Gerlach, Das Lichtmikroskop. Eine Einführung in Funktion, Handhabung und Spezialverfahren für Mediziner und Biologen, Stuttgart: Georg Thieme verlag, 1976. [3] G. Nomarski, „Interferential polarizing device for study of phase objects“. USA Patent US2924142, 11 Mai 1953. [4] J. d. P.R. Griffith, in Fourier Transform Infrared Spectroscopy, New York, Wiley, pp. 338-367. [5] N. Harrick, Internal Reflection Spectroscopy, New York: Harrick Scientific Corp., 1967. [6] N. Harrick, „Reflection of Infrared Radiation from a Germanium-Mercury Interface,“ J. Opt. Soc. Am., Bd. 49, pp. 376-379, 1959. [7] J.D.E. McIntyre, D.E. Aspnes, (p. 417, „Differential Reflection Spectroscopy of very thin surface films,“ Surface Science, Bd. 24, p. 417 ff, 1971. [8] A. Udagawa, T. Matsui, S. Tanaka, Appl. Spectrosc., Bd. 40, Nr. 6, p. 794 ff, 1986. [9] R.G.Greenler, J.Vac.Sci.Technol., Bd. 12, Nr. (6), p. 1410 ff, 1975. [10] S. A. Francis, A. H. Ellison, J.Opt.Soc.Am., Bd. 49, Nr. 2, p. 131 ff., 1959. [11] R. d. L. Kronig, „On the theory of dispersion of X-rays,“ Journal of the Optical Society of America, Bd. 12, Nr. 6, pp. 547-556, 1926. [12] H. Kramers, „La diffusion de la lumiere par les atomes,“ in Atti Cong. Intern. Fisici, (Transactions of Volta Centenary Congress), Como, 1927. [13] M. Handke, M. Milosevic, N. J. Harrick, Vibrational Spectroscopy, Bd. 1, p. 251 ff, 1991. [14] J.H.Lambert, Photometria, seu de mensura et gradibus luminis colorum et umbras, Augsburg, 1760. [15] P. Kubelka, F. Munk, Zeitung für technische Physik, Bd. 12, p. 593, 1931. [16] M. Gurevic, Physikalische Zeitung, Bd. 31, p. 753, 1931. [17] D. B. Judd, J. Res. Nat. Bur. Std, Bd. 12, p. 354, 1934. [18] T. Smith, Trans. Opt. Soc. London, Bd. 33, p. 150, 1931. [19] D. J. Fraser, P. R. Griffiths, Applied Spectroscopy, Bd. 44, p. 193, 1990. [20] H. Hecht, Analytische Chemie, Bd. 48, p. 1775, 1976. [21] M. P. Fuller, P. R. Griffiths,, Analytical Chemistry , Bd. 50, p. 1906, 1978. [22] K. Moradi, C. Depecker, J. Corset, Applied Spectroscopy, Bd. 48, Nr. 12, pp. 1491-1497, 1994.

264

[23] F. Cindy, T. J. Baulsir, J. Tague jr., „Introduction to Diffuse Reflectance Infrared Fourier Transform Spectroscopy,“ Spectra Tech Technical Note T2. [24] S. A. Yeboah, S. Wang, P. R. Griffiths, Applied Spectroscopy, Bd. 38, p. 259 ff, 1984. [25] Z. Krivacsy, J. Hlavay, Bd. 41, Nr. 7, p. 1143 ff, 1994. [26] R. S. Shreedhara Murthy, D. E. Leyden, Analytical Chemistry , Bd. 58, p. 1228, 1986. [27] Z. Krivacsy, J. Hlavay, J. Mol. Struct., Bd. 294, p. 251, 1993. [28] K. Schwarzschild, Astrophys. J., Bd. 40, p. 317, 1907. [29] W. Steel, „The design of reflecting microscope objectives,“ Australian Journal of Scientific research A Physical Sciences, Bd. 4, Nr. 1, 1951. [30] C. V. Raman, K. S. Krishnan, „A new type of secondary radiation,“ Nature, Bd. 121, p. 501, 1928. [31] N. J. Everall, „Raman Spectroscopy in Coatings Research and Analysis Part 1. Basic Principles,“ JCT Coatingstech, Bd. 2, Nr. 19, pp. 38-44, 2005. [32] N. J. Everall, „Raman spectroscopy in coatings research and analysis Part II Practical Applications,“ JCT Coatings Tech, Bd. 2, Nr. 20, pp. 46-52, 2005. [33] T. Dieing, O. Hollricher und J. Toporski, Confocal Raman Microscopy, Berlin, Heidelberg: Springer Verlag, 2011. [34] Department of Geoscience University of Arizona, „RRuFF Database of RAMAN spectra,“ [Online]. Available: https://rruff.info/ anatas/display=default/R060277. [35] J. Zieba-Palus, A. Michalska, A. WeseluchaBirczynska, „Characterisation of paint samples by infrared and Raman spectrocopy,“ Journal of Molecular Structure, Bd. 993, pp. 134-141, 2011. [36] S. Parnell, K. Min und M. Cakmak, Kinetic studies of polyurethane polymerization with Raman spectroscopy Polymer, Bd. 44, No. 18, pp. 5137-5144, (2003) [37] M. Cregut, M. Bedas, A. Assaf, M. J. DurandThouand, G. Thouand, „Applying Raman spectroscopy to the assessment of the biodegredationof industrial polyurethane waist,“ Environmental Science and Pollution Research, Bd. 16, Nr. 21, pp. 9538-9544, 2014. [38] G. A. Pitsevich, M. Shundalau M. A. Ksenofontov, D. S. Umreiko, „Vibrational analysis of 4,4-methylene diphenyldiisocyanate,“ Global Journal of Analytical Chemistry, Bd. Volume 2, Nr. Issue 3, pp. 114-124, 2011.

References [39] National Institute of Advanced Industrial Science and Technology, „https://sdbs. db.aist.go.jp/sdbs/,“ [Online 31 01 2020]. [40] National Institute of Advanced Industrial Science and Technology, „https://sdbs. db.aist.go.jp/sdbs,“ [Online am 31 1 2020]. [41] John Wiley & Sons Inc., „https://spectrabase.com,“ [Online am 31 1 2020]. [42] S. J. Thomson, „LXXXIII. Rays of positive electricity,,“ The London, Edinburgh and Dublin Philosophical Magazine and Journal of Science, Bd. 20, Nr. 118, pp 752-767, 1910. [43] A. Benninghoven, F. G. Rüdenauer, H. W. Werner, SIMS, New York: John Wiley & Sons, 1987. [44] D. v. Leyen, Dissertation Univerität Münster, Münster, 1993. [45] H. Feld, „Dissertation,“ Münster, 1991. [46] K. Meyer, Diplomarbeit Universität Münster, Münster, 1994. [47] R. Dietrich, „Surface Analysis of Silicon Wafers (Diplomarbeit),“ 1990. [48] D. v. Leyen, J. Vac. Sci. Technol., Bd. A 7, p. pp 1790, 1989. [49] R. Dietrich, „Möglichkeiten und Grenzen der Fehleranalytik,“ Journal für Oberflächentechnik, p. 52ff, 2015. [50] R. Dietrich, „Submonolayer detection of polymer additives at the surface of industrial products,“ Fresenius Journal of Analytical Chemistry (361), pp. 692694, 1998. [51] R. Dietrich, „Krater im Lack Lackierungs­störungen analysieren,“ Metalloberfläche, pp. 35-39, 1999. [52] H. Bonaduce, A. Andreotti, „Py-GC-MS of Organic Paint Binders,“ in Organic Mass Spectrometry in Art and Archaeology, J. W. a. S. Ltd, Hrsg., 2009, pp. 303-326. [53] P.F.Schmidt, „Praxis der Rasterelektronenmikroskopie und Mikrobereichsanalyse,“ 1994. [54] M .P. Seah, D. Briggs, Practical Surface Analysis, Chichester: John Wiley& Sons, 1990. [55] E. Zambonin, P. G. Desimoni, „Spectroscopies for Surface Characterization,“ in Characterization of Advanced Polymers, Weinheim, VCH Verlagsgesellschaft mbH, 1993, pp. 6-19.

[56] Physical Electronics Perkin Elmer Corporation, Handbook of X-Ray Photoelectron Spectro­scopy, Eden Prairie Minnesota, 1992. [57] National Institute of Standards and Technology, „NIST X-ray Photoelectron Spectroscopy Database Number 20,“ 2000. [Online].Available: https://srdata. nist.gov/xps.. [Online 3 2018]. [58] G. B. &. D. Briggs, The XPS of Polymers (CD rom version), SurfaceSpectra Limited. [59] M. Seah, „Quantification of AES and XPS,“ in Practical Surface Analysis Second Edition Volume 1, Chichester West Sussex, John Wiliey and Sons Ltd., 1990, pp. 201-251. [60] Bruno Kolb, Leslie S. Ettre, Static Headspace-Gas Chromatography: Theory and Praktice, Wiley-Interscience, 2006, p. pp 83 ff. [61] J. d. Zeeuw, „Impact of GC Parameters on the Separation,“ Restek Corporation, [Online]. Available: www.restekgmbh.de/ node/33874. [Online 2020]. [62] Restek GmbH, „A Technical Guide for Static Headspace Analysis Using GC,“ [Online]. Available: www.restek.com. [Online 31 1 2020]. [63] J. J. Manura, „Note 28: Analysis of Volatile Organics In Latex Paints By Automated Headspace Sampling,“ Scientific Instrument Services Inc., [Online]. Available: www.sisweb.com/referenc/applnote/app28h.htm. [Online 20 10 2017]. [64] Agilent, „www.agilent.com,“ 2018. [Online]. Available: www.agilent.com/cs/ library/catalogs/public/5991-5213EN_ GC_Catalog_Applications.pdf. [Online 3 2018]. [65] W. Goodman, „VOC Analysis of Water-Based Coatings by Headspace-Gas Chromatography,“ 1 3 2020. [Online]. Available: www.perkinelmer.com/ lab-solutions/resources/docs/App_VOCanalysisofwaterbasedcoatings.pdf. [66] T. Kowalska, K. Kaczmarski, W. Prus, „Theory and Mechanism of Thin-Layer Chromatography,“ in Handbook of Thin-Layer Chromatography, J. S. a. B. Fried, Hrsg., Marcel Dekker, 2003, pp.4780.

THE RHEOLOGY HANDBOOK

Thomas G. Mezger THE RHEOLOGY HANDBOOK

5th Revised Edition

5th Revised Edition

The Mission: Extremely practical coverage of the mathematical and physical principles behind rheology – from flow properties to viscoelastic behavior to taking actual measurements and interpreting them correctly. To provide a sound basis for learning about the principles of rheology and successfully applying them. The Audience: Beginners with no previous knowledge of rheology, as well as advanced users seeking to refresh their knowledge and find out about the latest developments. For all those who seek a deeper understanding of rheology or who simply want a reference book for their daily work. The Value: This book describes the principles of rheology clearly, vividly and in practical terms. The third edition of this standard work has been expanded to include the rheology of additives in water-based dispersions and surfactant systems. Not only is it a great reference book, it can also serve as a textbook for studying the theory behind the methods.

Thomas G. Mezger is a wellknown expert and trainer in the field of rheology and rheometry and its useful transfer into industrial practice.

2020, 528 pages, hardcover, 199 € order no. 20473, eBook: 20474

ISBN 978-3-86630-5328

www.european-coatings.com/shop Bodo Müller + Ulrich Poth

Wilhelm Kettler et al.

Bodo Müller + Ulrich Poth

Albert Rössler

KeTTler | Colour TeChnology of CoATings

Design of Experiments for Coatings The Mission: A single book covering the practical and scientific basics underpinning the strategic formulation of modern paint and coatings systems – from physicochemical concepts to the recipes themselves. This book explains and elaborates in some depth on the key principles of coatings formulation. Indispens-able for formulators. The Audience: Trainees, students and newcomers to the profession who are seeking to acquire a solid grounding in coatings formulation, along with experienced formulators wishing to deepen, extend or refresh their knowledge. A knowledge of chemistry and basic knowledge of binders, pigments and additives are required. The Value: Coatings formulation explained step by step. The book opens with a look at the composition of coatings, placing special emphasis on the base binder in each type. Advice on specific formulations is then given before formulation guidelines are analysed. Throughout, the focus is on coatings formulation and how to arrive at the final recipe. A special feature of the book is its detailed index, which allows the reader to conduct targeted searches for specific aspects of coatings formulation.

Albert Rössler · Design of Experiments for Coatings Ulrich Poth Automotive Coatings Formulation

ehensive state-of-the-art surions and explains the various nt optics. Colour problems s and chemistry of solid-colour on, optical microscopy of fication, methods of elemenement and visual assessment tolerances and acceptability, a special chapter devoted to re assessment of effect colour

THE RHEOLOGY HANDBOOK Already in its 5th edition, this standard work describes the principles of rheology clearly, vividly, and in practical terms. It presents the use of rheology in different industrial areas and provides the reader with various useful instructions for experiments on material characterization.

Thomas G. Mezger

MÜLLER + POTH | COATINGS FORMULATION

newcomers to the profession nding in colour technology, as nd or refresh their knowledge.

Discover our other bestsellers: THOMAS G. MEZGER | THE RHEOLOGY HANDBOOK

he colour technology underthe fundamentals of colour colour-order systems, through plication of pigments. ng a solid grounding in the ion of colour technology to erts and those with aspirations

THERE‘S MORE!

ISBN 978-3-86630-126-9

ISBN 978-3-86630-885-5

9 783866 308855

European EuropeanCoatings Coatings Symposium Library

Wilhelm Kettler et al.

Bodo Müller + Ulrich Poth

Albert Rössler

COLOUR TECHNOLOGY OF COATINGS

COATINGS FORMULATION

DESIGN OF EXPERIMENTS FOR COATINGS

Gain the latest knowledge about the colour technology of coatings in a single book – from the fundamentals of colour perception, to colour measurement and colour-order systems, through to the characterisation and practical application of pigments.

Your comprehensive knowledge base when it comes to the formulation of paints and coatings: already in its 3rd edition, this book imparts the composition of coatings clearly, placing special emphasis on the base binder in each type.

In order to efficiently develop and improve coatings formulations, it is essential to analyse the several factors affecting their properties. This book offers a comprehensive overview of the statistical approach of design of experiments, pointing out its benefits for coatings development.

2016, 300 pages, hardcover, 189 € order no. 769, eBook: 769_EBOOK

2017, 312 pages, hardcover, 189 € order no. 20031, eBook: 20032

2014, 168 pages, hardcover, 169 € order no. 601, eBook: 601_EBOOK

EUROPEAN

C OATINGS library

www.european-coatings.com/shop

Author

Author Dr Roger Dietrich (OFG-Analytik GmbH) studied chemistry at the University of Muenster, Germany. During his doctoral thesis he concentrated on applications and advancement of surface analytical methods for the characterisation of technical surfaces. In 1993 he founded the company OFG-Analytik GmbH together with two fellow partners which have also been engaged in the development of surface analytical methods. OFG-Analytik GmbH offers service analyses and expertise concerning failure analyses and quality control for industrial customers. The author is one of the CEOs of the company and he has focussed his activities on surface and materials analysis of paints and coatings. Besides he has been assistant lecturer for “Applied Surface Analysis” at the University of Applied Science in Muenster and hosts analytical training courses for industrial and institutional customers. His expertise is in demand with clients all over the world and he has written several publications about analysis of paints and coatings.

Acknowledgements My intention to write this book was to let people participate in my experiences of the last 25 years that I was involved in paint and materials analysis and inspire progressive ways to deal with coating issues. But writing a book being a full-time CEO and scientist is sometimes challenging with respect to time management. Looking back I have to emphasize that this book, the preceding publications and my experiences of the last decades of course would not have been possible without the support and trust my former teacher Prof. Dr Josef Grobe (University of Münster, Germany) gave me at the beginning of my scientific career in the 1990s. Thank you for promoting my ideas and opening new ways. My appreciation goes to Dr Jürgen Sawatzki of Bruker Optics who supported me in the Raman section of this book. We know each other now for nearly thirty years and he has always been an inspiring dialog partner with respect to Microspectroscopy. My special thanks go to Dr Konrad Völkel of the University of Osnabrück, Germany who made sure, that I did not go wrong with the statistics of sampling. That was a great assistance to me, and I really appreciate that he found the time to do this. To conclude I would like to mention gratefully my chemist colleague Dr Heinz-Wilhelm Wilde. We discuss coating issues nearly every week for more than twenty years now and I learned a lot about the refinements of coating formulation from our scientific debates. We are united in the constant “battle” to replace wild guesses by knowledge and I really appreciate our weekly brainstorming sessions.

267

Index

Index

Symbols 2-[4-[2-hydroxy-3-tridecyloxypropyl]oxy] -2-hydroxyphenyl-4,6-bis(2,4-dimethyl­phenyl)-1,3,5-triazine  38 2-pack polyester-polyurethane  153, 154 2-pack polyurethane paints  152

A AAS, atomic absorption spectroscopy  125 abraded particles  83 abrasion 84 abrasive sampling  48 absorbance spectrum  170 absorption 166 acidic cleaning  55 acrylonitrile butadiene styrene (ABS)  73 activation 60 additive  13, 14, 111, 113, 115–117, 120, 121, 125–130, 135, 139, 149–151, 157, 170, 225, 227, 228, 230, 234, 235, 249, 254 additive, agglomeration  86 additive, migration  155 adhesion  117, 120, 145, 149, 150, 155 adhesion, deficiencies  49, 50 adhesion, failure  14, 15, 63, 67, 69, 232 adhesion, promoters  192 adhesive tapes  144 adipic acid  117 aerosol  62, 83, 124, 144, 145, 148 aerosol precipitation  66, 68, 79 aerosol targets  84, 141, 145 AES (Auger electron spectroscopy)  17, 66 AFM (atomic force microscope)  263 agglomeration  124, 135, 140 air supply system  143, 147 aircraft 45 aircraft coating  157 aircraft industry  113 alkaline cleaning  55 alkyl, bond energy  250 ambient air  124, 143, 144, 145, 151 analyser, sample detection  17, 24 analysis, insulating samples  253 analysis, isolators  232 analyte  169, 171, 172, 198–203, 218, 219, 253, 255, 257 angle of incidence  180, 182 antioxidant  123 appearance  16, 112, 117, 140, 146, 155, 157 application, conditions  60 application, failures  26 application, faults  73

268

aromatic 250 assessment  34, 37 atomic absorption spectroscopy (AAS)  125 atomizer 79 atomizing air  62 ATR (internal reflection spectroscopy)  17, 33, 177, 179, 181 ATR crystal material  184 ATR crystal  72, 115, 120, 128, 133, 134, 152, 184 ATR infrared microscopy  66 ATR microscope  129 ATR, horizontal  174, 185 ATR-FT-IR  38, 66, 72, 75, 84, 111, 112, 114, 115, 118–120, 122, 124, 127–129, 131–136, 139–141, 152, 153, 169, 170, 174–179, 182–184, 186, 187, 210, 239, 241, 247, 261, 263 attenuated total internal reflectance  177 automotive  71, 131 automotive coatings  234 automotive interior parts  14 availability, instruments  244 available information  253 axial profiling  23, 218

B back-scattered electrons  17, 236 back-scattering coefficient  236 Bernoulli chain  41 Bernoulli experiment  40 binder  113, 114, 118, 121, 126, 128, 129, 134, 135, 232, 234, 239 binder/hardener ratio  152, 154 binder-production 232 binomial distribution  41 blasting 56 blisters  90, 91, 92, 93 bond energy  247 booth water  61 box in box tool  143, 146 BSE (back-scattered electrons)  236 bubbles  45, 62, 63, 90, 91, 100, 102 bulk  11, 12, 19, 22, 23 by-product  116, 117, 130, 235

C cabin air  62, 146 calibration process  114, 154 carbon black  186, 187, 208, 211, 218, 219, 221

Index catalyst  87, 113 CDS (wavelength-dispersive spectrometer)  241, 243 CHA analyser  253 chain scission  65, 151, 152 characteristic vibrations  165 characteristic X-ray radiation  237 chemical bonding  63 chemical composition  12, 13, 16, 18, 19, 20, 23 chemical information, molecular  263 chemical map  19, 218, 237 chemical maps/imaging  232 chemical molecular information  263 chemical reactions  14 chemical shift  165, 248, 249, 252 chromatogram  254, 255, 256 chromatographic process  117, 118 cleaner selection  56 cleaning process  65, 149, 150, 151 cleanliness 151 clear coat  156, 208 climate 62 cluster sampling  44 CMA analyser  253 CO2 57 CO2, snow jet  150 coagulation additives  61 coating appearance  140 coating booth  141 coating failure analysis  111 coating material production failures  26 coating material  111, 141 coating robot  141 compressed air  36, 143, 144, 147, 148 compressed air supply  141 compressed air system  62 confidence interval  40 confidence level  40 confocal Raman microscopy  22 construction of paint shops  61 contaminants  16, 114, 117, 121, 207, 226, 232–234, 254 contamination  33, 39, 52–58, 60, 62, 66–69, 79, 80, 82–84, 86, 88, 89, 93–96, 99, 100, 108, 109, 111, 140, 141, 147, 149 conveyor belts  144 conveyor system  141 corona treatment  59, 150, 152 corrosion  112, 149, 150, 157 corrosion resistance  157 co-solvent content  112 crater  4, 13, 19, 25, 26, 28ff, 42, 53, 56, 61, 78ff, 100, 117, 141ff, 148, 162, 205, 211, 229, 246 critical angle  179 cross section  21, 22, 23, 80 cross-cut test  78 crosslinking process  155

cross-reaction 114 crystalline polyamide  93 curing  74, 232 cuvettes 169 cyclic compound  117

D data evaluation  27, 172, 252, 253 data interpretation  27 databases  17, 19, 34, 174 defective paint material  141 defoamer 61 degassing 90 degradation process  114 degradation  65, 71, 94, 95, 97 degreasing 151 degree of crosslinking  152, 154 delamination  16, 64, 71, 72, 112, 152, 246, 112 deposit  38, 58, 97, 101 depth of focus  161 depth of penetration  179 depth profiling  19, 22, 232, 244, 253, 263 depth resolution  232, 244, 253 design of experiment  27, 30, 145 detection limit  19, 61, 66, 93–95, 100, 115, 125, 129, 149, 176, 182, 183, 225, 231, 232, 241, 245, 252, 253, 257, 263 detection of elements  232, 244, 253, 263 detection sensitivity  33 detector  17, 22, 24 diamond 184 DIC (differential interference contrast)  163, 164 diffraction  101, 243 diffuse reflection  18,159, 169,188, 195, 196ff digital separation  128 discolouration  14, 29, 93, 113, 114 discontinuities  101, 102 dispersing agents  69 distribution 156 DOE (design of experiment)  145, 146 drawing oil  111 dried film  134, 139 DRIFT  49, 169, 187, 188, 196, 199, 200–204 dry ice blasting  55, 57 dry wiping  46 drying conditions  60 dual cure coating  74 dust  83, 124 dust particles  81 dynamic mode  227 dynamic SIMS  222

269

Index

E ECD (electron capture detector)  258 EDS (electron microanalysis)  17, 21, 22, 241 EDX (electron microanalysis)  22, 66, 235, 237, 238, 240, 241, 243–246 effect pigment  254 effective path length  181 EFI (enhanced light microscopy)  90, 91, 161, 236 EI mass spectrometry  227 ELCD (electrolytic conductivity detector)  258 electrocoating 78 electrodeposition primer  249 electrons  16, 17, 18, 22, 24 element ions  227, 228 elemental analysis  83 EMA  241 energy dispersive X-ray analysis  22 environment 62 environmental contamination  149 ESCA (electron spectroscopy for chemical analysis)  17, 247 ESMA(electron microanalysis)  17, 241, 263 ester 250 evaluation 34 evanescent wave  179 expenditure of time for measurement  244 expert knowledge  37 extended focus imaging  161 external knowledge  15, 37 external reflection mode  159 external reflection spectroscopy  13, 98 external reflection  176, 187, 190–193, 195, 196, 205–207, 211 extinction coefficient  181 extraction  94

F failure analysis proceeding  27 failure analysis  25, 161, 218, 219, 235, 246 failure spot  246 false colour image  20, 22 far infrared  167 fatty acid  123 ferrous oxide  123 fibres  205 field analysis 7 field test  145 filler  87, 88, 121, 126, 127, 129, 133–135 filter residue  124, 127, 128 first sample inspection  27 fisheyes  78 flame ionization detector  254, 256, 258 flame treatment  59, 146, 147, 151, 152 flat dent  80

270

flow additive  112 fluorescence  17 fluorescence microscopy  75, 80, 106, 164 fluoride  66 fluorinated polymer  66 fluorination  150, 151, 152 fluorocarbon lubricant  84 fluoro-surfactant  66 flushing media  143, 146 FMEA  31 foaming 90 focal volume  23 fogging  111–113, 129, 130, 131, 135, 235 fogging residue  130 foreign aerosol  33 foreign particles  86, 87, 95, 102 Fourier transform  168 FPD (flame photometric detector)  258 fragment ion  227, 228 fragmentation pattern  116, 117, 120, 227 free isocyanate  20 FT-IR spectrometer  17, 168 FT-IR spectroscopy  17, 20, 24, 33, 38, 39, 51, 57, 65ff, 72, 74ff, 84, 95, 98, 103, 107, 111, 112, 114, 118ff, 133 ff, 177ff

G gas 115 gas-chromatographic analysis  117, 255 GC-MS  36, 94, 108, 109, 115–117, 133, 149, 226, 233, 253, 254, 256, 258–260 gel particles  78, 81 germanium 184 glass beads  56 glove 69 grease  55, 65 grease residues  33 grinding dust  55

H handheld infrared spectrometer  98 handheld instruments  158, 159 handheld Raman spectrometer  217, 218 handling 113 handling failures  60 hardener 152 hardener ratio  153 harmonic oscillator  165 H-ATR 119 haze  97, 158 haze effect  99 headspace GC-MS  111, 112, 134, 135, 141 148, 254 hexanediol  117 hit quality index  36

Index Hooke’s law  165 horizontal ATR  174, 185 hoses  143, 159 HPLC-MS  115 hypermap 22

I ICP (inductively-coupled plasma)  125 identification of chemical bonds  253 identity checks  232 identity control  114 imaging of element distribution  263 implementation 38 improper transportation  68 improvement of production processes  26 impurities  112, 124, 141, 146, 147, 149 inappropriate sampling tools  51 inclusions  53, 80, 83, 102, 106 incomplete rinsing  77 incorrect mixing ratio  75 incorrect storage  66 inductively-coupled plasma  125 information 244 information content  19 information depth  19, 20, 22, 180, 182, 211, 218, 232, 241, 247, 248, 253 informed guess  30 infrared microscopy  53, 80, 89, 93, 96, 102, 105, 106, 176, 193, 204, 206–210, 213, 215, 218, 221 infrared microscopy imaging  20 infrared radiation  179, 247 infrared spectroscopy  16, 35, 48, 57, 74–76, 83, 95, 96, 99, 101, 165 infrared waves  16 inhomogeneities 76 instrument selection  32 insufficient cleaning of the raw product  26 insufficient crosslinking  74 insufficient layer thickness  75 insufficient mixing  146 insufficient pretreatment  64 insufficient storing  52 inter-atomic bonds  165 inter-coat adhesion  65 internal expertise  37 internal reflection  176, 177, 179, 183–186, 204, 205 internal release agents  112 internal separating agents  111 internal specialist  37 internal standard  230, 257 investigation goal  27 investigation of chemical bonds  244 ions 16 IR microscopy  81, 84

IRM (infrared microscope)  17, 204, 206, 207 IRRAS  17, 169, 176 Ishikawa diagram  31 isocyanate hardener  152

K KBr disc  169, 171, 176 kinetic energy  247 KK treatment  70 Kramers-Kronig  174, 195, 208, 209, 210 Kramers-Kronig transformation  158 KRS-5  184, 185 Kubelka-Munk  48, 174, 198–200, 202

L laboratory analysis  140 lacquer wetting disturbing substances  33 Lambert-Beer law  171 lateral profiling  23 lateral resolution  34, 61, 125, 232, 248, 252, 253 layer thickness  122, 146 levelling agent  87 light-stabilizer additive  156 limit of detection  93 line-scan  206, 209, 210, 213, 215 liquid cell  176 liquid chromatography  115, 255 loose boundary  70 lowest grasped sampling depth  263 low-molecular weight compounds  113 lubricants  83, 84

M macroscopic inspection  29 maintenance errors  141 mapping  206, 212–215, 246 marine coating  157 mass range  232 mass resolution  225–229 mass spectrometer  258–260 mass spectrometry  227 masterbatch  152, 153 matrix effect  227 mechanical bonding  63 melamine  134, 135 metallographic cross section  65, 75, 82, 83, 86, 91, 92, 103, 240 meta-stable ions  227 Michelson interferometer  168 microanalysis 81 micro-extract  65 micro-extraction  122 microsampling 53

271

Index microtome cross section  207 microtome section  82, 103, 240 mid infrared  167 migration  66, 69, 232 MIR  166, 169, 177, 185 mirror optics  206 mixing ratio  33 mobile microscope  29 mobile phase  255, 256, 261, 262 molecular identification  133, 149 molecular information  61, 232 molecular ions  227, 228 molybdenum oxide  123 monomer  113, 130, 235 mould release agents  62 moulding conditions  71 multi-layer system  156

N near infrared  167 near infrared spectroscopy  114 neopentyl glycol  117 neutral particles  16 NIR  114, 166 non-contact measurement  218 non-contact mode  98 non-destructive sampling  45 non-representative sampling  50 non-volatile  226, 254, 256, 261 number of internal reflections  182

O oil droplets  81 oil trace  192 oils 65 oligomer  71, 113, 130, 235 operating test  145 optical contact  182 optical light microscopy  161 orange peel  108 organic contaminant  126 organo-tin compound  87 overspray 80 oxidation  60, 93, 94, 95 oxide residue  55

P packaging 113 paint adhesion failure  14, 162 paint analysis, analytical technique  11, 14-16 paint bubbles  82 paint crater  28, 88, 205, 232, 246 paint flaws  246 paint material  124, 142, 143, 146, 152

272

paint production  7, 14, 111, 113, 119 paint raw materials  232 paint setting impairment substances, PWIS  32 paint shop ventilation  62 paint spray mist  79 paint spray  79, 102 paint wetting  32 paintability  14, 151, 152 painting air  142, 146, 147 painting robots  143, 159 particle size deviation  112 partition coefficient  255, 257 PC-PET 78 PDMS  141, 143, 228, 234 PE 235 penetration depth  22, 73, 179, 180, 181, 199, 211, 236, 241, 248 perception of the results  27 perfluorinated polyethers  56 perpendicular 207 PES PUR  73 pH value  112 phosphate 123 photodegradation 93 photons 16 PIDD 222 pigment  111, 121, 133 pinhole  79, 80, 84 pipes  141, 159 plasma treatment  150, 151 plausibility checks  17 polar bonding  63 polishing process  97 polyaddition 113 polyamide 70 polycondensation  113, 232 polydimethylsiloxane  113, 116, 117, 141, 143, 228, 234 polyester  116, 117, 128, 134, 135, 153, 154, 158 polyester binder  128 polyester polyurethane lacquers  74 polyether 250 polyethylene glycol  250, 259 polymerisation 113 polypropylene glycol  250 polyurethane  61, 73, 74, 75, 89, 91, 95, 96, 98, 99 polyurethane clear coat  95 poor paint quality  54 portable computer microscopes  159 powder coating  82 power wash  58, 67, 71, 146, 149, 150, 152 precipitation 46 pretreatment  56, 65, 115, 119, 128, 131, 146, 149–152 pretreatment faults  54

Index primary electrons  235 primary ion dose  222 primary radiation  16 printability 151 problem analysis  15 process analysis  141 process control  111 process knowledge  37 processing auxiliaries  122 production air supply  143 production problems  15 PWIS, paint wetting impairment substances  32, 36, 143, 145 pyrolysis GC-MS  234

Q qualitative analysis  125 quality certificate  135, 136, 139 quality control  7, 112, 218, 232, 234, 246 quantification  136, 170, 181, 202, 219, 229, 232, 244, 251, 252, 253, 263 quantitative analysis  125 quasi-molecular ions  227

R radiation  16, 17, 18, 20, 24 Raman microscope  217, 218, 220, 222 Raman microscopy mapping  20 Raman microscopy  105 Raman spectroscopy  135, 158, 215 Raman technique  215 random sampling  40, 43 raw material  7, 12, 14, 15, 111–113, 124, 130–132 ray optics system  24 reaction product  130, 235 rearrangement 114 reference spectrum  170 refractive index  180, 182 region of interest  20 regression analysis  136, 139 release agent  15, 55, 84 representative measurement  95 representative sampling  40 residue  68, 120 resins  232, 234 Reststrahlenbanden 99 reverse engineering  132 rinse sampling  46 rinsing 56 robot  141-144, 146-148, 159 robot programming  75 ROI  213, 217 role of sampling  40

S sample preparation  19, 22, 27, 252, 253 sample properties  232 sample size  232, 244, 253 sample spectrum  170 sample  226, 244, 253 sampling depth  263 sampling failures  49 sampling procedure  27 sampling process  40 sand blasting  55, 56 sanding media  55 scalpel cut  103 scanning electron microscopy  17, 32, 124, 235 scattering 101 SE image  235 seals  144, 159 secondary beam  24 secondary electron  17, 235 secondary ion mass spectrometry  222 secondary ions  222, 224, 225, 227, 228, 230–233 secondary neutral mass spectrometry  224 secondary radiation  17 sectional planes  103 segregation 69 SEM  13, 16, 17, 20–22, 24, 72, 120, 221, 225, 235, 236, 238, 239, 241, 243, 245–248, 252, 263 SEM image  71 SEM-BSE image  126 SEM-BSE  83, 90, 91 SEM-EDS  35, 80–87, 90, 95, 102, 105, 106, 111, 112, 125, 129, 133, 135, 151 SEM-EDX  86, 241 separating agent  33, 65 separating column  117, 118 shipping of samples  52 side reaction  78, 80, 97, 114 side-chain modification  234 sieve residue  141 silicate 123 silicon 184 silicone spray  144 silicone-free 69 SIMS  17, 222 SNMS  224 solid content  112 solvent blend  117–120 solvent boils  80 solvent residue  128, 130, 235 solvents  112, 113, 115, 117–121, 134–136, 142, 143, 239 spatial resolution  21, 210, 211, 218 spatial 263 specks 33

273

Index spectral addition  174 spectral subtraction  115, 128, 129, 139 specular reflection infrared spectroscopy  158 spherical aberration  218 spin-coating 226 spot analysis  127, 157 spots 205 spray guns  143 sputtering process  222 sputtering system  24 SSIMS  223 stabiliser 122 stage of development  244 stains  25, 97, 98, 100, 112, 140 standard analytical tools  15 static SIMS  223 stationary phase  255–259, 261, 262 storage  84, 113, 124, 135, 140, 143, 159 storage failures  26 stratified sampling 44 stretching vibration  166 structural analysis  232, 234 structural characterization  165 structural identification  166 structural information  232 structural studies  253 structure determination  244 study of insulating material  244 substrate 66 substrate contaminant  84 substrate contamination  66 substrate failures  80 substrate handling  68 suitability 19 surface 11 surface analysis  11, 12, 14, 18, 24 surface charging  247 surface cleaning  58 surface contaminant  254 surface contamination  149, 150, 151 surface infrared spectroscopy  176 surface modification  121 surface pretreatment  77 surface sensitivity  14, 61, 149, 176, 253 surface topography  161 surfacer 87 surfactant 65, 113 suspension 43 systematic random sampling  44

T tank truck  44 targeted sampling  45 task force manager  39 TCD 258 TD GC-MS  35

274

thermal conductivity detector  256 thermal degradation  65 thermo-desorption 36 time per analysis  232, 253 time per measurement  232, 253 time-of-flight analyser  225 TiO2 246 TOF-SIMS  16, 17, 20–22, 24, 26, 33, 35, 36, 38, 39, 49, 51, 52, 54, 56–62, 65–70, 75–80, 82–86, 89, 93, 95, 96, 100, 101, 106, 108, 109, 111, 112, 115–120, 122–125, 129–133, 135, 139–141, 145, 149–151, 225, 248, 263 TOF-SIMS image  157 TOF-SIMS imaging  20, 21, 157 TOF-SIMS spectrum  67, 68, 70, 78 top cut  53 topography imaging  263 trace analysis  205 trace contaminant  115, 116, 118–120, 232 transflectance  188, 193, 195, 206 transmission infrared spectroscopy  169 transmission 205 transmission spectrum  170 transport  113, 124, 135 transportation 84

U uncured paint film  239 uppermost layer  12, 20 UV lacquers  74 UV stabilizer  38

V vacuum  252, 253 vacuum technique  239 valves 159 VOC 254 volatile components  120, 140 volatile compound  130 volatile substances  255

W water separators  141, 144 wavelength  167, 180, 182 wavelength-dispersive spectrometer  243 wavenumber 167 WDS  241, 243 WDX  17, 238, 241, 243 weak interface  64 weathering test  93 wedge cut  155, 156, 157 wet wiping  45 wetting failure  14

Index wetting problem  88 wind power plant  140 wipe cleaning  152 wipe sample  45, 83-85, 124, 141 workpiece preparation  55 wrong sample collection  49 wrong sampling procedure  51

X

Y yacht coating  45, 98 yellowing  93, 94, 95, 113

Z ZnSe  184

XPS  17, 66, 120, 247, 248, 263 XPS method, availability  253 XPS survey spectrum  249 X-rays  16, 17, 18, 22 X-ray excitation  248 X-ray photoelectron spectroscopy  67, 247

275