James S. Fritz and Douglas T. Gjerde Ion Chromatography
Ion Chromatography, 4th Ed. James S. Fritz and Douglas T. Gjerde Copyright © 2009 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim ISBN: 978-3-527-32052-3
Further Reading Miller, JM
Chromatography - Concepts and Contrasts 2e 2006 ISBN: 978-0-471-98059-9
McMahon, Gillian
Analytical Instrumentation A Guide to Laboratory, Portable and Miniaturized Instruments 2007 ISBN: 978-0-470-02795-0
Stuart, Barbara H.
Analytical Techniques in Materials Conservation 2007 ISBN: 978-0-470-01280-2
Dean, John R.
Bioavailability, Bioaccessibility and Mobility of Environmental Contaminants 2007 ISBN: 978-0-470-31967-3
Wang, Joseph
Analytical Electrochemistry 2007 ISBN: 978-0-470-23187-6
Hahn-Deinstrop, Elke
Applied Thin-Layer Chromatography 2006 ISBN: 978-3-527-31553-6
James S. Fritz and Douglas T. Gjerde
Ion Chromatography Fourth, Completely Revised and Enlarged Edition
The Authors Prof. James S. Fritz Ames Laboratory Iowa State University 332 Wilhelm Hall Ames, IA 50011 USA Dr. Douglas T. Gjerde PhyNexus, Inc. 3670 Charter Park Dr./ Suite A San José, CA 95136 USA
&
All books published by Wiley-VCH are carefully produced. Nevertheless, authors, editors, and publisher do not warrant the information contained in these books, including this book, to be free of errors. Readers are advised to keep in mind that statements, data, illustrations, procedural details or other items may inadvertently be inaccurate. Library of Congress Card No.: applied for British Library Cataloguing-in-Publication Data A catalogue record for this book is available from the British Library. Bibliographic information published by the Deutsche Nationalbibliothek The Deutsche Nationalbibliothek lists this publication in the Deutsche Nationalbibliografie; detailed bibliographic data are available in the Internet at http://dnb.d-nb.de. © 2009 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim All rights reserved (including those of translation into other languages). No part of this book may be reproduced in any form – nor transmitted or translated into machine language without written permission from the publishers. Registered names, trademarks, etc. used in this book, even when not specifically marked as such, are not to be considered unprotected by law. Typesetting Kühn & Weyh, Satz und Medien, Freiburg Printing Strauss GmbH, Mörlenbach Bookbinding Litges & Dopf, Heppenheim
Printed in the Federal Republic of Germany. Printed on acid-free paper.
ISBN:
978-3-527-32052-3
V
Contents Preface
13
Acknowledgments
15
1 1.1 1.2 1.2.1 1.2.2 1.2.3 1.2.4 1.2.5 1.2.6 1.3 1.3.1 1.3.2 1.3.3 1.3.4 1.3.5 1.3.6
Introduction and Overview 1 Introduction 1 Historical Development 2 Early Ion-Exchange Separations 2 Cation Separations 3 Separation of Anions 6 On-line Detection 8 The Birth of Modern Ion Chromatography 8 Non-Suppressed-Ion Chromatography 10 Principles of Ion Chromatographic Separation and Detection 13 Requirements for Separation 13 Experimental Setup 13 Performing a Separation 14 Migration of Sample Ions 15 Detection 17 Basis for Separation 17
2 2.1 2.2 2.3 2.4 2.4.1 2.5 2.6 2.6.1 2.6.2 2.6.3 2.6.4
Instrumentation 21 Components of an Ion Chromatograph (IC) Instrument General Considerations 23 Eluent 24 Pump 26 Gradient Formation 29 Sample Injector 30 Columns 31 Column Hardware 31 Column Protection 32 Column Oven 33 Two-dimensional IC 33
Ion Chromatography, 4th Ed. James S. Fritz and Douglas T. Gjerde Copyright © 2009 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim ISBN: 978-3-527-32052-3
21
VI
Contents
2.7 2.8 2.9
Suppressor 33 Detector 34 Data Acquisition and Calculation of Results 35
3 3.1 3.2 3.2.1 3.2.2 3.2.3 3.2.4 3.2.5 3.3 3.3.1 3.4 3.4.1 3.4.2 3.4.3 3.4.4 3.4.4.1 3.5 3.5.1 3.5.2 3.5.3 3.6 3.6.1 3.6.2 3.6.3 3.6.4
Resins and Columns 37 Introduction 37 Polymeric Resins 38 Substrate and Cross Linking 38 Microporous Resins 39 Macroporous Resins 40 Chemical Functionalization 41 Resin Capacity 42 Resins and Columns for Ion Chromatography 43 Monolith Columns 43 Anion Exchangers 45 Porous Anion Exchangers 45 Effect of Functional Group on Selectivity 47 Effect of Spacer Arm Length 52 Latex Agglomerated Ion Exchangers 54 Effect of Latex Functional Group on Selectivity 56 Cation Exchangers 57 Sulfonated Resins 57 Weak-acid Cation Exchangers 61 Other Types 63 Other Resins 63 Chelating Ion-exchange Resins 63 Metal Oxides 64 Multi-purpose Resins 64 Ion-exchange Disks 65
4 4.1 4.2 4.2.1 4.2.2 4.2.3 4.2.4 4.2.5 4.3 4.3.1 4.3.2 4.3.3 4.3.4 4.4 4.5
Detectors 69 Introduction 69 Conductivity Detectors 70 Conductivity Definitions and Equations 73 Principles of Cell Operation 74 Conductance Measurement 75 Conductivity Hardware and Detector Operation 75 Contactless Conductivity Detection 76 Ultraviolet-Visible (UV–Vis) Detectors 77 UV–Vis measurement 77 Direct Spectrophotometric Measurement 78 Post-column Derivatization 81 UV–Vis Hardware and Detector Operation 82 Fluorescence Detector 83 Electrochemical Detectors 85
Contents
4.5.1 4.5.2 4.5.3 4.5.4 4.5.4.1 4.5.4.2 4.5.4.3 4.5.5 4.5.6 4.6 4.7 4.7.1 4.7.2 4.7.3 4.8
Potentiometric Detection 86 Conductometric Detectors 86 Amperometric/Coulometric Detection 87 Pulsed Electrochemical Detection (PED) 89 Pulsed Amperometric Detection (PAD) 91 Integrated Pulsed Amperometric Detection (IPAD) 93 IC–PED 94 Post-column Derivatization 95 Electrochemical Hardware and Detector Operation 95 Refractive Index Detector 97 Evaporative Light Scattering Detector (ELSD) 97 Nebulizer 98 Evaporation Chamber 99 Detection Cell 99 Other Detectors 100
5 5.1 5.2 5.2.1 5.3 5.3.1 5.3.2 5.3.3 5.3.4 5.3.4.1 5.3.4.2 5.3.4.3 5.4
Principles of Ion Chromatographic Separations 105 General Considerations 105 Chromatographic Terms 105 Retention Factor 106 Selectivity 109 Selectivity Coefficients 110 Other Ion-exchange Interactions 112 Selectivity of Sulfonated Cation-exchange Resin for Metal Cations 113 Factors Affecting Selectivity 120 Polymeric Matrix Effect 121 Resin Functional Group 122 Solvation Effects 123 Chromatographic Efficiency 124
6 6.1 6.1.1 6.1.2 6.2 6.2.1 6.2.2 6.2.3 6.2.3.1 6.2.3.2 6.3 6.3.1 6.3.2 6.3.3 6.3.3.1
Anion Chromatography 131 Scope and Conditions for Separation 131 Columns 132 Separation Conditions 135 Suppressed Anion Chromatography 138 Electrolytic Suppressors 140 Solid-Phase Reagents, 1990 [7] 141 Typical Separations 142 Isocratic and Gradient Elution 144 Influence of Organic Solvents 146 Nonsuppressed Ion Chromatography 147 Principles 147 Explanation of Chromatographic Peaks 150 Eluent 150 General Considerations 150
VII
VIII
Contents
6.3.3.2 6.3.3.3 6.3.3.4 6.3.4 6.3.5 6.3.6 6.3.6.1 6.4 6.5 6.5.1 6.5.2 6.5.3 6.6 6.7 6.8 6.9 6.9.1 6.9.2
Salts of Carboxylic Acids 151 Basic Eluents 152 Carboxylic Acid Eluents 153 System Peaks 154 Scope of Anion Separations 155 Sensitivity 155 Conductance of a Sample Peak 158 Coated Columns 160 Optical Absorbance Detection 163 Introduction 163 Direct UV Absorption 163 Indirect Absorbance 164 Detection 166 Pulsed Amperometric Detector (PAD) 168 Evaporative Light Scattering Detector (ELSD) 170 Inductively Coupled Plasma Methods (ICP) 172 Atomic Emission Spectroscopy (AES) 172 Inductively Coupled Plasma Mass Spectrometry (ICP-MS)
7 7.1 7.2 7.2.1 7.2.2 7.2.3 7.2.4 7.3 7.3.1 7.3.2 7.3.3 7.4 7.4.1 7.5 7.5.1 7.5.2 7.5.3 7.5.4 7.6 7.6.1 7.6.2
Cation Chromatography 175 Introduction 175 Columns 176 Historical Development 178 Phosphonate Columns 179 Macrocycle Columns 181 Surfactant Columns 182 Separations 184 Suppressed-Conductivity Detection 184 Non-Suppressed-Conductivity Detection 187 Spectrophotometric Detection 188 Effect of Organic Solvents 191 Separation of Alkali Metal Ions 193 Separation of Metal Ions with a Complexing Eluent Principles 195 Separations 196 Use of Sample Masking Reagents 197 Weak-Acid Ion Exchangers 198 Chelating Ion-Exchange Resins and Chelation Ion Chromatography 201 Fundamentals 201 Examples of Metal-Ion Separation 204
8 8.1 8.1.1
Ion-Exclusion Chromatography Principles 207 Equipment 209
207
195
173
Contents
8.1.2 8.1.3 8.2 8.2.1 8.2.2 8.3 8.4 8.5 8.5.1 8.6 8.7 8.7.1 8.8 8.8.1 8.8.2 8.8.3 8.8.4
Eluents 209 Detectors 210 Separation of Organic Acids 211 Effect of Alcohol Modifiers 214 Separation of Carboxylic Acids on Unfunctionalized Columns 216 Simultaneous Determination of Anions and Cations 217 Conclusions 220 Determination of Carbon Dioxide and Bicarbonate 222 Enhancement Column Reactions 222 Separation of Bases 223 Determination of Water 226 Determination of Very Low Concentrations of Water by HPLC 229 Separation of Saccharides and Alcohols 230 Introduction 230 Separation Mechanism and Control of Selectivity 230 Detection 235 Contamination 235
9 9.1 9.2 9.3
Ion Pair Chromatography 239 Principles 239 Typical Separations 242 Mechanism 246
10 10.1 10.2 10.3 10.4 10.5
Zwitterion Stationary Phases 251 Introduction 251 Simultaneous Separation of Anions and Cations 253 Separation of Anions 255 Separation of Cations 256 Mechanism 259
11 11.1 11.1.1 11.1.1.1 11.1.1.2 11.1.1.3 11.1.1.4 11.2 11.2.1 11.2.2 11.2.3 11.3 11.3.1 11.3.2 11.4
Capillary Electrophoresis 263 Introduction 263 Steps in Analysis 264 Capillary Pretreatment 264 Sample Introduction 264 Sample Run 265 Detection 265 Principles 265 Terms and Relationships 265 Zone Broadening 267 Sample Injection 267 Electroosmotic Flow (EOF) 268 Effect of EOF on Separations 270 Control of EOF 271 Separation of Ions 274
IX
X
Contents
11.4.1 11.4.1.1 11.4.2 11.4.3 11.4.4 11.5 11.5.1 11.5.2 11.5.3 11.5.4 11.5.4.1 11.6
Separation of Anions 274 Separation of Isotopes 276 Separation of Cations 278 Separations at Low pH 279 Capillary Electrophoresis at High Salt Concentration 280 Capillary Electrophoretic Ion Chromatography 283 Micellar Electrokinetic Chromatography (MEKC) 284 Partial Complexation 285 Effect of Ionic Polymers 287 Effect of Alkylammonium Salts 291 Separation Mechanism 294 Summary 294
12 12.1 12.1.1 12.1.2 12.2 12.3 12.3.1 12.3.2 12.3.3 12.4 12.4.1 12.4.2 12.4.3 12.5 12.5.1 12.5.2 12.5.3 12.5.4 12.6 12.6.1 12.6.2 12.6.2.1 12.6.2.2 12.6.2.3 12.7 12.7.1 12.7.2
DNA and RNA Chromatography 299 Introduction 299 Importance of DNA and RNA Chromatography 299 Organization of this Chapter 300 DNA and RNA Chemical Structure and Properties 301 DNA and RNA Chromatography 303 Development of DNA and RNA Chromatography 303 Column Properties 305 Ion-pairing Reagent and Eluent 306 Temperature Modes of DNA and RNA Chromatography 307 Nondenaturing Mode 307 Fully Denaturing Mode 308 Partially Denaturing Mode 309 Instrumentation 310 Effect of Metal Contamination 310 The Column Oven 313 UV and Fluorescence Detection 313 Fragment Collection 314 Applications of DNA Chromatography 314 DHPLC 314 Nucleic Acid Enzymology 315 Telomerase Assays 315 Polynucleotide Kinase Assays 316 Uracil DNA Glycosylase Assays 317 Applications of RNA Chromatography 317 Separation of Messenger RNA from Ribosomal RNA 318 Analysis of Transfer RNA 319
13 13.1 13.2 13.3
Sample Pretreatment 323 Dilute and Shoot or Pre-treat the Sample? 323 Particulate and Column-contaminating Matter 324 Preconcentration 325
Contents
13.3.1 13.3.2 13.4 13.4.1 13.4.2 13.4.3 13.4.3.1 13.4.3.2 13.4.4
Collection of Ions from Air 325 Preconcentration of Ions in Water 326 Sample Pretreatment 328 Anions in Acids 328 Neutralization of Strongly Acidic or Basic Samples Dialysis Sample Preparation 329 Passive Dialysis 330 Donnan (Active) Dialysis 330 Isolation of Organic Ions 333
14 14.1 14.1.1 14.1.2 14.1.3 14.2 14.3 14.3.1 14.3.2 14.3.3 14.4 14.4.1
Method Development and Validation 335 Choosing a Method 335 Define the Problem Carefully 335 Experimental Considerations 336 Example of Method Development 338 Some Applications of Ion Chromatography Statistical Evaluation of Data 341 Common Statistical Terms 341 Distribution of Means 344 Confidence Intervals 345 Validation of Analytical Procedures 347 Analytical Control 349
15 15.1 15.2 15.3 15.4 15.5 15.5.1 15.5.2 15.5.3 15.5.4 15.5.5 15.5.6 15.5.7 15.5.8 15.5.9
Chemical Speciation 353 Introduction 353 Detection 355 Chromatography 356 Valveless Injection IC 357 Speciation of some Elements Chromium 359 Iron 360 Arsenic 361 Tellurium 362 Selenium 363 Vanadium 364 Tin 364 Mercury 365 Other Metals 365 Index
369
359
340
328
XI
XIII
Preface The first edition of this book in 1982 described the emergence of ion-exchange chromatography as a practical and rapid method of separation and analysis. The initial and somewhat restrictive definition of the subject was broadened to 'ion chromatography' in order to describe the efficient chromatographic separation of anions or cations using any form of automatic detection. In the intervening years ion chromatography, or IC as it is sometimes called, has undergone impressive development and has attracted an ever-growing number of users throughout the world. Over the years, IC has proved to be extremely rugged, sensitive and reliable. It is the technology of choice to measure ions in such diverse situations as plating baths, drinking water, nuclear power plant water, foodstuffs and so on. Although IC is a well-established technique that is widely used throughout the world, it continues to enjoy vibrant growth and development. An annual International Ion Chromatographic Symposium (IICS) is held at various locations. This 4th edition of Ion Chromatography has been expanded and extensively revised. Some new features are listed below. . With some help from William La Course (see Acknowledgements), Chapter 4 on Detectors has been expanded and completely rewritten. New features include a strong section on pulsed electrochemical detection and an extensive table of dyes for tagging for the fluorescent detection of bio ions. . Chapter 5 (Principles of IC) has been expanded to include a discussion of factors that affect selectivity and chromatographic efficiency. . Although IC is the preferred method for inorganic ions and smaller organic ions, the need for chromatographic determination of larger organic and bio ions has been growing rapidly, as for example in the pharmaceutical industry. Here, Ion Pair Chromatography (IPC) is often the preferred technique. A new chapter (Chapter 9) on IPC is now included. . Chapter 10 on Zwitterion Stationary Phases describes a fascinating variation of IC. Pure water can often be used as the eluent to separate sample ions when a zwitterion stationary phase is employed. Ions elute as cation-anion pairs. Ion Chromatography, 4th Ed. James S. Fritz and Douglas T. Gjerde Copyright © 2009 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim ISBN: 978-3-527-32052-3
XIV
Preface .
.
. .
Ions are separated on a truly micro scale on an open capillary based on differences in their electrophoretic migration under the influence of an applied electrical potential (Chapter 11). Migration rates of the analyte ions are modified by interaction with organic ions of opposite charge in the electrolyte. Analyte peaks are unusually sharp because the zone broadening due to mass transfer between two distinct phases is eliminated. The reader is introduced to the fascinating world of the chromatography of large bio ions in Chapter 12. Structural features, common terms and chromatographic techniques are explained in a manner that the non-expert can easily understand. DNA molecules are extremely large ions relative to what can normally be separated by a chromatographic process. The new Chapter 13 is devoted to Sample Pretreatment. Method development, statistical evaluation of data, and validation and control of analytical procedures are covered in a new chapter (Chapter 14).
As in the previous editions, our goal has been to describe the materials, principles and methods of ion chromatography in a clear concise style. Whenever possible the consequences of varying experimental conditions are considered. Because commercial products are constantly changing, the equipment used in ion chromatography is described in a somewhat general manner. Our approach to the literature of IC is selective rather than comprehensive. Key references are given together with the title so that the nature of the reference will be apparent. Our goal is to explain fundamentals, but also to provide more detailed information in the form of figures and tables. We have tried to write a book that is enjoyable to read as well as one that is informative. Although it has been hard work, writing this book has also been a stimulating experience. We hope that we have been able to convey this enthusiasm to our readers. James S. Fritz, Ames, Iowa Douglas T. Gjerde, San Jose, California
October 2008
XV
Acknowledgments In the preparation of our 1st Edition of Ion Chromatography, published in 1982, we received valuable help from many sources, and we have continued to benefit from the skill, knowledge and willingness of many individuals over the years. Now with this 4th Edition of the book, we are again pleased to acknowledge some valuable contributions. We thank William R. LaCourse, who wrote Section 4.5 on Electrochemical Detection. Bill, who is Professor of Analytical Chemistry at the University of Maryland, Baltimore County, 1000 Hilltop Circle, Baltimore, MD 21250,
[email protected], is well known for his expertise in pulsed-amperometric detection, and we are grateful for his willingness to contribute to the section which deals with this subject. We also thank Peter Scott, currently a university student, for creating the original design used on the front cover of this book. Peter also helped with index and drew several of the figures that appear in the text. Jie Li, Shelly Li and Jeff Sun provided us with a clearer insight into ion analysis as used in the pharmaceutical industry, particularly with regard to the validation of analytical methods. Wenzi Hu, a world leader in ion chromatography using zwitterion stationary phases, kindly supplied reprints of his papers. We also received figures and other information from several companies who are in the business of providing tools for Ion Chromatography. They include, in no particular order, Dionex, Hamilton, Metrohm, Zellweger Analytics, Alltech (product line of Grace David) and Transgenomic. Our assistants played a major part in preparing and proofing the manuscript. In Ames, Marilyn Kniss did most of the typing, patiently checking the manuscript for consistency and clarity of expression. She also handled the correspondence. In San Jose, Tiffany Nguyen performed additional valuable and necessary work. We would like to extend our grateful thanks for the love and support of our respective families, for they, more than anything else, make life enjoyable, worthwhile and meaningful. James S. Fritz Ames, Iowa
Ion Chromatography, 4th Ed. James S. Fritz and Douglas T. Gjerde Copyright © 2009 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim ISBN: 978-3-527-32052-3
Douglas T. Gjerde San Jose, California
1
1 Introduction and Overview 1.1 Introduction
The name ‘ion chromatography’ applies to any modern method for chromatographic separation of ions. Normally, such separations are performed on a column packed with a solid ion-exchange material, but if we define chromatography broadly as a process in which separation occurs by differences in migration, capillary electrophoresis may also be included. Ion chromatography is considered to be an indispensable tool in a modern analytical laboratory. Complex mixtures of anions or cations can usually be separated and quantitative amounts of the individual ions measured in a relatively short time. Higher concentrations of sample ions may require some dilution of the sample before introduction into the ion-chromatographic instrument. ‘Dilute and shoot’ is the motto of many analytical chemists. However, ion chromatography is also a superb way to determine ions present at concentrations down to at least the low parts per billion (lg L–1) range. Although the majority of ion-chromatographic applications have been concerned with inorganic and relatively small organic ions, larger organic anions and cations may be determined as well. As in the three previous editions, our goal has been to describe the materials, principles and methods of ion chromatography in a clear, concise style. The following résumé is intended as a kind of road map to guide the reader through the contents of this book and to highlight some of the changes made in this fourth edition. In the first chapter we recount some of the historical milestones and briefly cover the most basic principles of ion chromatography, or IC as it is often called. The various components and hardware of IC instruments are described in Chapter 2, but it is not our intention to discuss specific commercial instruments. Chapter 3 has been updated to include advances in column technology and promising new columns, such as monolithic columns. Chapter 4 on detectors has been expanded to include new material on the contactless conductivity detector (CCD) and pulsed electrochemical detectors. Chapter 5 has been completely rewritten and now includes detailed sections on the factors that influence selectivity and efficiency. An updated and detailed treatIon Chromatography, 4th Ed. James S. Fritz and Douglas T. Gjerde Copyright © 2009 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim ISBN: 978-3-527-32052-3
2
1 Introduction and Overview
ment of anion chromatography and cation chromatography is presented in the next two chapters. Selection of appropriate columns, detectors and eluents is discussed and numerous examples of typical separations are given. Ion exclusion chromatography, Chapter 8, continues to be a popular and useful method for the separation of hydrophilic sample components, such as carboxylic acids, amines and carbohydrates. Ion chromatography is generally defined as an analytical method in which anions or cations are separated by differences in the rate at which they pass through a column packed with either an anion- or cation-exchange particles. However, excellent separations of ionic analytes can also be obtained by ion chromatography on a standard reversed-phase HPLC column. With this technique, separation of cations or anions is achieved by using an aqueous-organic eluent together with an ion-pairing reagent. Ion-pair chromatography, covered in the new Chapter 9, is particularly advantageous for separation of organic ions. Chapter 10 on zwitterion stationary phases is another new addition to this book. Separations are generally performed on an HPLC column coated with a zwitterion surfactant that contains both positive and negative sites. In some cases cations and anions can be separated in a single run using pure water as the eluent! Resolution of the sample peaks obtainable in either liquid or ion chromatography is limited by two factors. One is due to mechanical pumping, which gives a curved flow profile. The second limiting factor stems from a slow rate of equilibration of solutes between the mobile and stationary phases. Use of very small ionexchange particles can reduce but never eliminate this source of peak broadening. Capillary electrophoresis (CE), which is covered in Chapter 11, addresses both of these issues. No eluent is required for CE; sample ions are separated by differences in their electrophoretic flow rates through open capillary containing no packing material. The electrophoretic mobilities can be modified by selective interactions with electrolyte ions of the opposite charge. Chapter 12 on Separation of DNA/RNA highlights a trend in IC toward a greater emphasis on analytical separations of bio ions. Sample pretreatment is discussed in Chapter 13. Chapter 14 on Method Development and Validation includes tips for selecting appropriate conditions for an IC analysis, and this final chapter covers chemical speciation.
1.2 Historical Development 1.2.1 Early Ion-Exchange Separations
Modern ion chromatography is built on the solid foundation created by extensive work in classical ion-exchange chromatography. Columns containing ionexchange resins have been used for many years to separate various cations and
1.2 Historical Development
anions from one another. Cations are separated on a cation-exchange resin column, and anions on an anion-exchange resin column. The most used types are as follows:
For example, Na+ and K+ can be separated on a cation-exchange resin (Catex) column with a dilute solution of a strong acid (H+) as the eluent (mobile phase). Introduction of the sample causes Na+ and K+ to be taken up in a band (zone) near the top of the column by ion exchange: Resin–SO–H+ + Na+, K+ > Resin SO–Na, K+ + H+ Continued elution of the column with an acidic eluent (H+) introduces competition of H+, Na+ and K+ for the exchange sites (–SO3–), causing the Na+ and K+ to move down the column. K+ is more strongly retained than Na+, and thus the Na+ zone moves down the column faster than the K+ zone. As originally conceived and carried out for many years, fractions of effluent were collected from the end of the column and analyzed for Na+ and K+. Then a plot was made of concentration vs fraction number to construct a chromatogram. All this took a long time and made ion-exchange chromatography slow and awkward to use. However, it soon was realized that under a given set of conditions all of the Na+ would be in a single fraction of several milliliters and all of the K+ could be recovered in a second fraction of a certain volume. Thus, under pre-determined conditions, each ion to be separated could be collected in a single fraction and then analyzed by spectroscopy, titration, etc., to determine the amount of each sample ion. The ability to collect a single fraction that contains all of the separated sample ion permits the use of step gradients. In this mode, conditions are adjusted so that an ‘all-or-nothing’ situation prevails. A sample ion either sticks onto the ionexchange column or it passes quickly through. Conditions are selected so that only one ion type will pass through the column while the other sample ions are strongly retained and form a tight band at the top of the column. Then the eluent is changed so that a second ion is rapidly eluted, while the others remain tightly stuck. Frequently, several gradient steps can be performed to elute different sample ions at each step. 1.2.2 Cation Separations
Early studies on the separation of metal cations included separations based on affinity differences and some specific separation with complexing eluents.
3
4
1 Introduction and Overview
Strelow and his coworkers have published extensive data relating to the selectivity of a sulfonated polystyrene cation exchanger for various cations in acidic solution [1]. The equilibria of cations in hydrochloric, nitric or sulfuric acid solutions with a cation exchanger involves complexation in some cases as well as competition between H+ and the metal cation for the exchange sites. For example, mercury(II) and cadminium(II) form chloride complexes even in dilute solutions of hydrochloric acid. Selectivity data in perchloric acid probably give the best indication of true ion-exchange selectivity, because the perchlorate anion has almost no complexing properties with metal cations. In general, cations with a 3+ charge are more strongly retained by a cation exchanger than cations with a 2+ charge, and ions with a 2+ charge are retained more strongly than those with a 1+ charge. Fritz and Karraker [2] were able to separate metal cations into groups according to their charge. Most divalent metal cations were eluted with a 0.1 M solution of ethylenediammonium perchlorate. Bismuth(III) and zirconium(IV) remained quantitatively o n the cation-exchange column. The use of the 2+ ethylenediammonium ion permitted a lower concentration to be used than would have been the case with an H+ eluent. Several inorganic acids exhibit a complexing effect for metal ions. The complexing acids include HF, HCI, HBr, HI, HSCN and H2SO4. The complexed metal ions are converted into neutral or anionic complexes and are rapidly eluted, while the other cations remain on the cation-exchange column. The data for hydrochloric acid [3] indicate selective complexing between metal cations and the chloride ion. For example, cadmium(II) has a distribution coefficient of 6.5 in 0.5 M hydrochloric acid, but a D = 101 in 0.5 M perchloric acid. Calcium(II), which shows no appreciable complexing, has a distribution coefficient of 147 in 0.5 M perchloric acid and 191 in 0.5 M hydrochloric acid. Strelow, Rethemeyer and Bothma [3, 4] also reported data for nitric and sulfuric acids that showed complexation in some cases. Mercury(II), bismuth(III), cadmium(II), zinc(II), and lead(II) form bromide complexes and elute in the order given in 0.1 to 0.6 M hydrobromic acid [5]. Most other metal cations remain on the column. Aluminum(III), molybdenum(VI), niobium(V), tin(IV), tantalum(V), uranium(VI), tungsten(VI) and zirconium(IV) form anion fluoride complexes and are quickly eluted from a hydrogen-form cation-exchange column with 0.1 to 0.2 M HF [6]. An eluent containing only 1% hydrogen peroxide in dilute aqueous solution will form stable anionic complexes with several metal ions. Fritz and Abbink [7] were able to separate vanadium(IV) or (V) from 25 metal cations, including the separation of vanadium(V) from 100 times as much iron(III). Strewlow [8] used hydrogen peroxide and sulfuric acid to separate titanium(IV) from more than 20 cations by cation exchange. Fritz and Dahmer [9] separated molybdenum(VI), tungsten(VI), niobium(V) and tantalum(V) as a group from other metals by adding dilute hydrogen peroxide to the sample solution and passing it through a cation-exchange column. Most of the eluents listed above are volatile upon heating and do not interfere with colorimetric, titrimetric or other methods for chemical determination of the metal ions separated. For the most part, group separations, rather than separation
1.2 Historical Development
of individual metal ions, are obtained, and only a short ion-exchange column is needed. Another valuable ‘all-or-nothing’ group separation uses an eluent consisting of 0.1 M tartaric acid and 0.01 M nitric acid [10]. Antimony(V), molybdenum(VI), tantalum(V), tin(IV) and tungsten(VI) form tartrate complexes in this acidic medium, but lead(II) and many other metal cations are not complexed and are retained by the cation exchanger. Samples containing tin(IV) must be added to the column in the tartrate solution. In a few cases an eluent containing an organic complexing reagent has been used successfully for the chromatographic separation of several metal ions. A notable example is the separation of individual rare earth ions with a solution of 2-hydroxylisobutyric acid as the eluent [11]. However, such separations necessitate careful equilibration of the column to maintain a desired pH. Sometimes gradient elution is used, and either the pH or the eluent concentration is changed. Metal cations usually form complexes with inorganic anions much more readily in organic solvents than in water. For example, the pink cobalt(II) cation requires around 4 M or 5 M aqueous hydrochloric acid to be converted to a blue cobalt(II) chloride anion. In a predominantly acetone solution, the intensely blue cobalt(II) is formed in very dilute hydrochloric acid. Thus, the scope of ionexchange group separations is increased greatly by carrying out separations in a mixture of water and an organic solvent. Fritz and Rettig [12] showed that zinc(II), iron(III), cobalt(II), copper(II) and manganese(II) can be separated from each other on a short cation-exchange column with eluents containing a fixed, low concentration of HCl, increasing the acetone concentration from 40% to 95% in steps. Later Strelow et al. [13] published extensive lists of metal-ion distribution coefficients in water/acetone/hydrochloric acid systems. Korkisch and coworkers have studied the effect of ethanol, acetic acid, ethylene glycol and many other solvents upon the ion-exchange behavior of metal ions in systems containing hydrochloric and other complexing acids [14]. The selectivity of low-capacity cation columns for monovalent ions can be adjusted by the addition of an organic modifier to the eluent. Using a nitric acid eluent of pH 2.5, for example, the elution order for monovalent ions is Li+, Na+, NH4+, K+. Simple amines elute in the order of the carbon number, after NH4+, with the result that (CH3)NH3+ (methylammonium) can co-elute with potassium. In most cases, this co-elution is of little significance, because potassium and methylammonium are not often in the same sample. However, where the analysis of either of these species in the presence of the other is desired, the selectivity can be modified by the addition of 40% methanol to the eluent [15]. The methanol causes the potassium to elute later but does not affect the elution time of methylammonium.
5
6
1 Introduction and Overview
1.2.3 Separation of Anions
Since most metal ions are cationic, it may sound strange to discuss their separation by anion-exchange chromatography. However, Kraus and Nelson, working at Oak Ridge National Laboratory in the USA, found that in aqueous hydrochloric acid solutions a number of metal ions form anionic complexes and are strongly taken up by anion-exchange resins. For most of the metal ions, a plot of the D value of several thousand is attained. An illustration of such plots for most of the metallic elements in the periodic table was published by Kraus and Nelson in 1956 [16]. Separations are generally achieved by adding the sample to an anion-exchange column in rather concentrated hydrochloric acid and eluting the nonsorbed metal ions with the same HCl concentration. Then the sorbed metal ions are eluted one at a time by stepwise reduction of the HCl strength of the eluent. Figure 1.1 illustrates one of the many practical separations published by Kraus and his coworkers [17].
Figure 1.1 Separation of metal ions on Dowex 1 × 10 anion exchange resin. (From Ref. [18] with permission.)
In a similar manner, elements that form anionic fluoride complexes can be separated from others and from each other on an anion exchanger by eluting with eluents containing HF plus HCl [18, 19]. Extensive studies of metal ion behavior on anion-exchange columns have also been carried out with eluents containing mixed H2SO4/HF [20, 21].
1.2 Historical Development
Anion-exchange distribution coefficients for most metallic elements in sulfuric acid solution have been measured [22, 23]. Uranium(VI), thorium(IV), molybdenum(VI) and a few other elements are retained selectively by anion-exchangers from solution in approximately 6 M nitric acid [24]. Operating in a predominantly organic solvent greatly improves the ability of metal ions to form complexes with halide and pseudo-halide anions. Such complexes generally are taken up strongly by an anion-exchange resin. Korkisch, Fritz, Strelow and others have published extensively on anion-exchange separations in partly nonaqueous solutions. Korkisch and Hazan [25] describe a method to separate metal ions that form chloride complexes from those that do not. The method uses an eluent consisting of 90–95% methanol in 0.6 M hydrochloric acid and requires only a short anion-exchange column. The metal ions studied are either retained as a sharp band or quickly pass through the column. Thus, we have an ‘all-or-nothing’ situation, and excellent group separations are obtained. Chromatographic separations of individual ions are also possible, and many have been published. An example is shown in Figure 1.2. Ion exchange in nonaqueous and mixed media has been reviewed [26].
Figure 1.2 Separation of nickel(II) and manganese(II) on a 6.0 × 2.2 cm column containing Dowex 1 × 8 resin, with partly nonaqueous eluents. (From Ref. [25] with permission.)
7
8
1 Introduction and Overview
Systems containing dimethylsulfoxide, methanol and hydrochloric acid have been studied for the anion-exchange behavior of 26 elements [27]. Numerous separations of two- to four-component mixtures of metal ions were carried out with quantitative results. 1.2.4 On-line Detection
At this stage in the development of ion-exchange chromatography, separation of cations or anions was still a slow and laborious process. It was becoming apparent that widespread use of ion-exchange chromatography as an analytical tool would require a system that gave fast separations with automatic recording of chromatograms. In 1971 an apparatus for ‘forced-flow chromatography’ was described in which the eluent was pushed through the analytical column by compressed nitrogen [28]. Detection of eluted ions was by UV-Vis spectrophotometry using a 30 mm × 2 mm flow cell. Iron(III) (10–90 lg) could be separated from most other metal ions and measured quantitatively in only 6 min. Forced-flow methods were soon developed for the chromatographic separation of a number of other metal ions [29–32]. The chromatograph was modified in 1974 so that a complexing reagent such as PAR or Arsenazo could be added to the column effluent via a mixing tee [33]. This made it possible to detect virtually any metal ion that could form a highly-colored complex. A recorded chromatographic separation of all 30 rare earths was obtained in 1974 [30, 34], with this apparatus. This separation took 100 min. Five years later, Elchuk and Cassidy in Canada were able to obtain a better separation of earths in only 27 min using a similar but improved system [35]. 1.2.5 The Birth of Modern Ion Chromatography
Liquid column chromatography went ‘high performance’ around 1970 and is now commonly referred to as HPLC. Major improvements in speed and efficiency were obtained by using columns of relatively small bore packed with small spherical particles of uniform diameter, using a pump to provide constant eluent flow, and using automatic detection of the separated sample components. However, application of this technology for the separation of ions lagged. It was mainly the lack of satisfactory detectors that held up the development of high-performance ion-exchange chromatography. This situation changed dramatically with the publication of a landmark paper in 1975 by Small et al., working at the Dow Chemical Co. [36]. As the authors put it: ‘It would be desirable to employ some form of conductimetric detection as a means of monitoring ionic species in a column effluent since conductivity is a universal property of ionic species in solution and since conductance shows a simple dependence on species concentration. However, the conductivity from the species of interest is generally “swamped out” by that from
1.2 Historical Development
the much more abundant eluting electrolyte. We have solved this detection problem by using a combination of resins which strips out or neutralizes the ions of the background electrolyte leaving only the species of interest as the major conducting species in the effluent. This has enabled us to successfully apply a conductivity cell and meter as the detector system.’ This new system, which was given the name ‘ion chromatography,’ enabled the analyst to quickly separate and measure quantitatively the cations or anions at low concentrations in fairly complex samples. A diagram of the system for cation analysis is shown in Figure 1.3. The upper column, called the ‘separator column’, was packed with polystyrene–2% DVB particles, surface-sulfonated to obtain an exchange capacity of approximately 0.02 mmol g–1. The lower ‘suppressor column’ was packed with anion-exchange resin of high exchange capacity in the hydroxide form.
Figure 1.3 System for cation analysis by conductimetric chromatography. (From Ref. [36] with permission.)
A practical method for the separation of anions was described in the same paper [36]. This endeavor necessitated the development of a new low-capacity anion-exchange resin. It had been known for some time that cation- and anionexchange resins have a marked tendency to clump together. Using this principle, a satisfactory anion-exchange material of low capacity was prepared by coating surface-sulfonated cation exchanger In the original scheme for the separation of anions a mixture of sodium hydroxide and sodium phenate was used in the eluent. The suppressor column was packed with a cation-exchange resin of high capacity. The suppressor column con-
9
10
1 Introduction and Overview
verted the eluent ions to water plus phenol, while the sample anions A– were converted to the highly conducting pair H+A–. An instrument called the ‘ion chromatograph’ was offered commercially by the newly organized Dionex Co. and became an immediate success. The new technology made it possible to separate and determine both cations and most anions, but the ability to determine anions at low ppm concentrations had the greater impact. Many cations could already be determined by various spectral methods and by reasonably good chromatographic methods, but prior to the advent of ion chromatography there was no general analytical method for anions, especially at very low concentrations. Once the scientific world became aware that anions in fairly complex mixtures could be easily separated and quantified, even at low ppm concentrations, the use of ion chromatography exploded. A powerful new analytical technique had again facilitated scientific endeavors that were previously impractical. 1.2.6 Non-Suppressed-Ion Chromatography
A major disadvantage of the original Ion Chromatograph was that it required the use of a large suppressor column that contributed to peak broadening and required frequent regeneration. The eluent for anion separations had to be a base, and anions of very weak acids could not be detected because their acidic form after suppression was too weakly conducting. In 1979 a synthetic method was described for producing anion-exchange resins of very low exchange capacity [37]. A porous polymeric resin was chloromethylated under mild conditions and then alkylated with trimethylamine to form ionic quaternary ammonium groups. The exchange capacity could be varied from 0.2 to 1.5 meq g–1 by controlling the time and temperature of the chloromethylation. This drastically lower exchange capacity permitted the use of much lower eluent concentrations than had previously been possible. In 1979, Gjerde, Fritz and Schmuckler described a simple system for anion chromatography with eluents containing anions of very low conductivity such as benzoate or phthalate [38]. Anions were separated on a column containing macroporous anion-exchange particles of very low exchange capacity: 0.07, 0.04 or 0.007 mmol g–1. The eluent was an aqueous solution of the sodium or potassium salt of an organic anion that had a significantly lower equivalent conductance than the anions to be separated. In this method some of the eluent anion is replaced by a sample anion of significantly higher conductance as the sample ion is eluted from the column and passes through the conductivity detector. Because of the low resin capacity, an eluent containing only ca. 10–4 M of an organic acid salt, such as benzoate or phthalate, could be used. The eluent conductance was so low that no suppressor column was needed, and the separated sample ions could be detected with a simple conductivity detector. Numerous anion separations were demonstrated, and in some instances detection limits below 1 ppm were obtained. This method was initially called ‘single-column ion chromatography’ and later ‘non-suppressed-ion chromatography.’ An additional paper on anion chromatog-
1.2 Historical Development
raphy [39] was published in 1980. A chromatographic separation of halide ions is shown in Figure 1.4.
Figure 1.4 Separation of 4.8 ppm of fluoride, 5.1 ppm of chloride and 26.0 ppm of bromide on XAD-1, 0.04 mequiv g–1; eluent is 0.65 mM potassium benzoate, pH 4.6. (From Ref. [39] with permission.)
The non-suppressed method for anion chromatography was followed quickly in 1980 by a similar method for cations [40]. This method also introduced the concept of indirect conductivity detection. A 1 × 10–3 M solution of nitric acid was used as the eluent in conjunction with a sulfonated cation-exchange column of low exchange capacity. In this method, a baseline of relatively high conductance is established when the column is equilibrated with the acidic eluent. After introduction of a sample mixture, such as Na+ and K+, and continued elution with the eluent, the sample cations are gradually resolved into zones in which some of the highly conducting H+ (equiv. conductance = 350 S cm2 equiv–1) is exchanged for a sample cation of much lower conductance. A sharp ‘peak’ of lower conductance is obtained for each sample cation. A mixture of Li+, Na+, NH4+, K+, Rb+ and Cs+ was separated in less than 10 min with a blend of 0.17 meq g–1 and unfunctionalized cation-exchange resins with 1.25 × 10–4 nitric acid as the eluent (Figure 1.5). Although separation of divalent cations with nitric acid was not practical, a fast separation of magnesium and calcium in tap water was obtained with a 1 × 10–3 M ethylene diammonium nitrate eluent.
11
12
1 Introduction and Overview
Figure 1.5 Separation of alkali metal cations on ammonium cation exchange column with a conductivity detector. Column: 350 × 2.0 mm. Packed with 0.059 mequiv g–1 cation exchange resin.
Development of the ion-chromatographic methods that use a conductivity detector was accompanied by a significant increase in chromatographic efficiency. The ion-exchange materials were of much smaller and more uniform size and the packing efficiency of the column was also improved. The changes that occurred were not unlike those in partition chromatography when it went from ‘liquid chromatography’ to ‘high-performance liquid chromatography’ (HPLC).
1.3 Principles of Ion Chromatographic Separation and Detection
1.3 Principles of Ion Chromatographic Separation and Detection 1.3.1 Requirements for Separation
The ion-exchange resins used in modern chromatography are of smaller particle size but have a lower capacity than the older resins. Columns packed with these newer resins have more theoretical plates than the older columns. For this reason, successful separations can now be obtained even when there are only small differences in retention times of the sample ions. The major requirements of systems used in modern ion chromatography can be summarized as follows: 1. An efficient cation- or anion-exchange column with as many theoretical plates as possible. 2. An eluent that provides reasonable differences between the retention times of the sample ions. 3. A resin-eluent system that attains equilibrium quickly so that kinetic peak broadening is eliminated or minimized. 4. Elution conditions such that retention times are in a convenient range – not too short or too long. 5. An eluent and resin that are compatible with a suitable detector. 1.3.2 Experimental Setup
Anions in analytical samples are separated on a column packed with an anionexchange resin. Similarly, cations are separated on a column containing a cationexchange resin. The principles for separating anions and cations are very similar. The separation of anions will be used here to illustrate the basic concepts. A typical column used in ion chromatography might have the dimensions 150 × 4.6 mm, although columns as short as 50 mm in length or as long as 250 mm are also used. The column is carefully packed with a spherical anionexchange resin of rather low exchange capacity and with a particle diameter of 5 or 10 lm. Most anion-exchange resins are functionalized with quaternary ammonium groups, which serve as the sites for the exchange of one anion for another. The basic setup for IC is as follow. A pump is used to force the eluent through the system at a fixed rate, such as 1 mL min–1. In the FILL mode a small sample loop (typically 10–100 lL) is filled with the analytical sample. At the same time, the eluent is pumped through the rest of the system, by-passing the sample loop. In the INJECT mode a valve is turned so that the eluent sweeps the sample from the filled sample loop into the column. A detector cell is connected to a strip-chart recorder or a data-acquisition device so that a chromatogram of the separation
13
14
1 Introduction and Overview
(signal vs time) can be plotted automatically. A conductivity or UV-visible detector is most often used in ion chromatography. The eluent used in anion chromatography contains an eluent anion, E–. Usually Na+ or H+ will be the cation associated with E– The eluent anion must be compatible with the detection method used. For conductivity, the anion E– should have either a significantly lower conductivity than the sample ions or be capable of being converted to a non-ionic form by a chemical suppression system. When spectrophotometric detection is employed, E– will often be chosen for its ability to absorb strongly in the UV or visible spectral region. The concentration of E– in the eluent will depend on the properties of the ion exchanger used and on the types of anions to be separated. Factors involved in the selection of a suitable eluent are discussed later. 1.3.3 Performing a Separation
To perform a separation, the eluent is first pumped through the system until equilibrium is reached, as evidenced by a stable baseline. The time needed for this may vary from a couple of minutes to an hour or longer, depending on the type of resin and the eluent used. During this step the ion-exchange sites will be converted to the E– form: Resin–N+R3 E–. There may also be a second equilibrium in which some E– is adsorbed on the resin surface but not at specific ion-exchange sites. In such cases the adsorption is likely to occur as an ion pair, such as E–Na+ or E–H+. An analytical sample can be injected into the system as soon as a steady baseline has been obtained. A sample containing anions A1– , A2–, A3–,...., A1– undergoes ion exchange with the exchange sites near the top of the chromatography column. A1– (etc.) + Res E– > Res–A1– (etc.) + E– If the total anion concentration of the sample happens to be exactly the same as that of the eluent being pumped through the system, the total ion concentration in the solution at the top of the column will remain unchanged. However, if the total ion concentration of the sample is greater than that of the eluent, the concentration of E– will increase in the solution at the top of the column because of the exchange reaction shown above. This zone of higher E– concentration will create a ripple effect as the zone passes down the column and through the detector. This will show up as the first peak in the chromatogram, which is called the injection peak. A sample of lower total ionic concentration than that of the eluent will create a zone of lower E– concentration that will ultimately show up as a negative injection peak. The magnitude of the injection peak (either positive or negative) can be used to estimate the total ionic concentration of the sample compared with that of the eluent. Sometimes the total ionic concentration of the sample is adjusted to
1.3 Principles of Ion Chromatographic Separation and Detection
match that of the eluent in order to eliminate or reduce the size of the injection peak. Behind the zone in the column due to sample injection, the total anion concentration in the column solution again becomes constant and is equal to the E– concentration in the eluent. However, continuous ion exchange will occur as the various sample anions compete with E– for the exchange sites on the resin. As eluent containing E– continues to be pumped through the column, the sample anions will be pushed down the column. The separation is based on differences in the ion-exchange equilibrium of the various sample anions with the eluent anion, E–. Thus, if sample ion A1– has a lower affinity for the resin than ion A2–, then A1– will move at a faster rate through the column than A2–. 1.3.4 Migration of Sample Ions
The general principles for separation are perhaps best illustrated by a specific example. Suppose that chloride and bromide are to be separated on an anionexchange column. The sample contains 8 × 10–4 M sodium chloride and 8 × 10–4 M sodium bromide, and the mobile phase (eluent) contains 10 × 10–4 M sodium hydroxide. In the column equilibration step, the column packed with solid anion-exchange particles (designated as Res-Cl–) is washed continuously with the NaOH eluent to convert the ion exchanger completely to the –OH– form. Res-Cl– + OH– > Res-OH– + Cl– At the end of this equilibration step, the chloride has been entirely washed away and the liquid phase in the column contains 10 × 10–4 M Na+OH–. In the sample injection step a small volume of sample is injected into the ionexchange column. An ion-exchange equilibrium occurs in a fairly narrow zone near the top of the column. Res-OH– + Cl– > Res-Cl– + OH– Res-OH– + Br– > Res- Br– + OH– Within this zone, the solid phase consists of a mixture of Res-Cl– , Res-Br– and ResOH–. The liquid phase in this zone is a mixture of OH–, Cl– and Br– plus its accompanying Na+. The total anionic concentration is governed by that of the injected sample, which is 16 × 10–4 M (see Figure 1.6A).
15
16
1 Introduction and Overview
Figure 1.6 Anion exchange column: A, after sample injection; B, after some elution with 0.001 M NaOH.
In the elution step, pumping 10 × 10–4 M NaOH eluent through the column results in multiple ion-exchange equilibria along the column in which the sample ions (Cl– and Br–) and eluent ion (OH–) compete for ion-exchange sites next to the Q+ groups. The net result is that both Cl– and Br– move down the column (Figure 1.6B). Because bromide has a greater affinity for the Q+ sites than chloride has, the bromide moves at a slower rate. Because of their differences in rate of movement, bromide and chloride are gradually resolved into separate zones or bands. The solid phase in each of these zones contains some OH– as well as the sample ion, Cl– or Br–. Likewise, the liquid phase contains some OH– as well as Cl– or Br–. The total anionic concentration (Cl– + OH– or Br– + OH–) is equal to that of the eluent (0.0010 M) in each zone. Continued elution with Na+OH– causes the sample ions to leave the column and pass through a small detector cell. If a conductivity detector is used, the conductance of all of the anions, plus that of the cations (Na+ in this example) will contribute to the total conductance. If the total ionic concentration remains constant, how can a signal be obtained when a sample anion zone passes through the detector? The answer is that the equivalent conductance of chloride (76 ohm–1 cm2 equiv–1) and bromide (78) is much lower than that of OH– (198). The net result is a decrease in the conductance measured when the chloride and bromide zones pass through the detector. In this example, the total ionic concentration of the initial sample zone was higher than that of the eluent. This zone of higher ionic concentration will be
1.3 Principles of Ion Chromatographic Separation and Detection
displaced by continued pumping of eluent through the column until it passes through the detector. This will cause an increase in conductance and a peak in the recorded chromatogram called an injection peak. If the total ionic concentration of the injected sample is lower than that of the eluent, an injection peak of lower conductance will be observed. The injection peak can be eliminated by balancing the conduction of the injected sample with that of the eluent. Strasburg et al. studied injection peaks in some detail [41]. In suppressed-anion chromatography, the effluent from the ion-exchange column comes into contact with a cation-exchange device (Catex-H+) just before the liquid stream passes into the detector. This causes the following reactions to occur. Eluent: Chloride: Bromide:
Na+OH– + Catex-H+ → Na+Cl– + Catex-H+ → Na+Br– + Catex-H+ →
Catex-Na+ + H2O Catex-Na+ + H+Cl– Catex-Na+ + H+Br–
The background conductance of the eluent entering the detector is thus very low because virtually all ions have been removed by the suppressor unit. However, when a sample zone passes through the detector, the conductance is high due to the conductance of the chloride or bromide and the even higher conductance of the H+ associated with the anion. 1.3.5 Detection
This effect can be used to practical advantage for the indirect detection of sample anions. For example, anions with little or no absorbance in the UV spectral region can still be detected spectrophotometrically by choosing a strongly absorbing eluent anion, E–. An anion with a benzene ring (phthalate, p-hydroxybenzoate, etc.) would be a suitable choice. In this case, the baseline would be established at the high absorbance due to E–. Peaks of non-absorbing sample anions would be in the negative direction owing to a lower concentration of E– within the sample anion zones. Direct detection of anions is also possible, providing a detector is available that responds to some property of the sample ions. For example, anions that absorb in the UV spectral region can be detected spectrophotometrically. In this case, an eluent anion is selected that does not absorb (or absorbs very little). 1.3.6 Basis for Separation
The basis for separation in ion chromatography lies in differences in the exchange equilibrium between the various sample ions and the eluent ion. A more quantitative treatment of the effect of ion-exchange equilibrium on chromatographic separations is given later. Suppose the differences in the ion-exchange equilibrium are
17
18
1 Introduction and Overview
very small. This is the case for several of the transition metal cations (Fe2+, Co2+, Ni2+, Cu2+, Zn2+, etc.) and for the trivalent lanthanides. Separation of the individual ions within these groups is very difficult when it is based only on the small differences in affinities of the ions for the resin sites. Much better results are obtained by using an eluent that complexes the sample ions to different extents. An equilibrium is set up between the sample cations, C2+, and the complexing ligand, L–, in which species such as C2+, CL+, CL2 and CL3– are formed. The rate of movement through the cation-exchange column is inversely proportional to a, the fraction of the element that is present as the free cation, C2+.
References [1] F.W. E. Strelow and H. Sandrop, Distri-
[2]
[3]
[4]
[5]
[6]
[7]
[8]
[9]
bution coefficients and cation-exchange selectivities of elements with AG50WX8 resins in perchloric acid, Talanta, 19, 1113, 1972. J. S. Fritz and S. K. Karraker, Ion exchange separation of metal cations, Anal. Chem., 32, 957, 1960. F. W. E. Strelow, An ion exchange selectivity scale of cations based on equilibrium distribution coefficients, Anal. Chem., 32, 1185, 1960. F.W. E. Strelow, R. Rethemeyer, and C. J. C. Bothma, Ion exchange selectivity scales for cations in nitric acid and sulfuric acid with a sulfonated polystyrene resin, Anal. Chem., 37, 106, 1965. J. S. Fritz and B. B. Garralda, Cation exchange separation of metal ions with hydrobromic acid, Anal. Chem., 34, 102, 1962. J. S. Fritz, B. B. Garralda and S. K. Karraker, Cation exchange separation of metal ions by elution with hydrofluoric acid, Anal. Chem., 33, 882, 1961. J. S. Fritz and J. E. Abbink, Cation exchange separation of vanadium from metal ions, Anal. Chem., 34, 1080, 1962. F. W. E. Strelow, Separation of titanium from rare earths, beryllium, niobium, iron, aluminum, thorium, magnesium, manganese and other elements by cation exchange chromatography, Anal., Chem., 35, 1279, 1963. J. S. Fritz and L. H. Dahmer, Cation exchange separation of molybdenum, tungsten, niobium and tantalum from
[10]
[11]
[12]
[13]
[14]
[15] [16]
[17]
[18]
other metals, Anal. Chem., 37, 1272, 1965. F. W. E. Strelow and T. N. van der Walt, Separation of lead from tin, antimony, niobium, tantalum, molybdenum and tungsten by cation exchange chromatography in tartaric-nitric acid mixtures, Anal. Chem., 47, 2272, 1975. J. N. Story and J. S. Fritz, Forced-flow chromatography of the lanthanides employing continuous in-stream detection, Talanta 21, 894, 1974. J. S. Fritz and T. A. Rettig, Cation exchange in acetone-water-hydrochloric acid, Anal. Chem., 34, 1562, 1962. F. W. E. Strelow, A. H. Victor, C. R. van Zyl, and C. Eloff, Distribution coefficients and cation exchange behavior of elements in hydrochloric acid-acetone, Anal. Chem., 43, 870, 1971. J. Korkisch, Modern methods for the separation of rarer metal ions, Pergamon, Oxford 1969. Wescan Instruments, Inc., Santa Clara, CA, ‘Wescan Ion Analyzer #6’, (1983). K. A. Kraus and F. Nelson, Proc. First U. N. Int. Conf. on Peaceful Uses of Atomic Energy, 7, 113, 1956. K. A. Kraus. G. E. Moore and F. Nelson, Anion exchange studies. XXI. Th(IV) and U(IV) in hydrochloric acid. Separation of thorium, protoactinium and uranium., J. Am. Chem. Soc., 78, 2692, 1956. F. Nelson, R. M. Rush and K. A. Kraus, Anion exchange studies. XXVII. Adsorbability of a number of elements
References
[19]
[20]
[21]
[22]
[23]
[24]
[25]
[26]
[27]
[28]
[29]
[30]
in HCl-HF solutions, J. Am. Chem. Soc., 82, 339, 1960. J. B. Headridge and E. J. Dixon, The analysis of complex alloys with particular reference to niobium, tantalum and tungsten, Analyst, 87, 32, 1962. E. A. Huff, Anion exchange study of a number of elements in nitric-hydrofluoric acid mixtures, Anal. Chem., 36, 1921, 1964. L. Danielsson, Adsorption of a number of elements from HNO3HF and H2SO4HF solutions by cation and anion exchange, Acta Chem. Scand., 19, 1859, 1965. L. Danielsson, Adsorption of a number of elements from sulfuric acid solutions by anion exchange, Acta Chem. Scand., 19, 670, 1965. F. W. E. Strelow and C. J. C. Bothma, Anion exchange and a selectivity scale for elements in sulfuric acid media with a strongly basic resin, Anal. Chem., 39, 595, 1967. J. S. Fritz and B. B. Garralda, Anion exchange separation of thorium using nitric acid, Anal. Chem., 34, 1387, 1962. J. Korkisch and L. Hazan, Anion exchange behavior of uranium, thorium, the rare earths and various other elements in hydrochloric acid-organic solvent media, Talanta, 11, 1157, 1964. W. R. Heumann, Ion exchange in nonaqueous and mixed media, Crit. Rev. in Anal. Chem., 2, 425, 1971. J. S. Fritz and Marcia Lehoczky Gillette, Anion-exchange separation of metal ions in dimethylsulfoxide-methanolhydrochloric acid, Talanta, 15, 287, 1968. M. D. Seymour, J. P. Sickafoose and J. S. Fritz, Application of forced-flow liquid chromatography to the determination of iron, Anal. Chem., 43, 1734, 1971. M. D. Seymour and J. S. Fritz, Rapid, selective method for lead by forced-flow liquid chromatography, Anal. Chem., 45, 1632, 1973. J. N. Story and J. S. Fritz, Forced-flow chromatography of the lanthanides
[31]
[32]
[33]
[34]
[35]
[36]
[37]
[38]
[39]
[40]
[41]
employing continuous in-stream detection, Talanta, 21, 892, 1974. M. D. Seymour and J. S. Fritz, Determination of metals in mixed hydrochloric and perchloric acids by forced-flow anion exchange chromatography, Anal. Chem., 45, 1394, 1973. K. Kawazu and J. S. Fritz, Rapid and continuous determination of metal ions by cation exchange chromatography, J. Chromatogr., 77, 397, 1973. J. S. Fritz and J. N. Story, Chromatographic separation of metal ions on low capacity macroreticular resins, Anal. Chem., 46, 825, 1974. J. S. Fritz and J. N. Story, Selectivity behavior of low-capacity, partially sulfonated macroporous resin beads, J. Chromatogr., 90, 267, 1974. S. Elchuk and R. M. Cassidy, Separation of the lanthanides on high-efficiency bonded phases and conventional ion exchange resin, Anal. Chem., 51, 1434, 1979. H. Small, T. S. Stevens and W. S. Bauman, Novel ion exchange chromatographic method using conductometric detection, Anal. Chem., 47, 1801, 1975. D. T. Gjerde and J. S. Fritz, Effect of capacity on the behavior of anionexchange resins, J. Chromatogr., 176, 199, 1979. D. T. Gjerde, J. S. Fritz and G. Schmuckler, Anion chromatography with low conductivity eluents, J. Chromatogr., 186, 509, 1979. D. T. Gjerde, G. Schmuckler and J. S. Fritz, Anion chromatography with low-conductivity eluents II, J. Chromatogr., 187, 35, 1980. J. S. Fritz, D. T. Gjerde and R. M. Becker, Cation chromatography with a conductivity detector, Anal. Chem., 52, 1519, 1980. R. T. Strasburg, J. S. Fritz, J. Berkowitz and G. Schmuckler, Injection peaks in anion chromatography, J. Chromatogr., 482, 343, 1989.
19
21
2 Instrumentation 2.1 Components of an Ion Chromatograph (IC) Instrument
Most ion chromatographs are integrated instruments with the various components of the instrument designed so that the plumbing and wiring match and fit together perfectly. Ion chromatography is truly global. Commercial suppliers located in the United States, Europe and Asia include a number of vendors – Dionex, Metrohm, Schimadzu, Alltech, Lachat and others. These companies offer complete instruments with a wide variety of features and options. Integrated instrumentation is desirable; this type of instrument usually performs exceptionally well giving sharp and sensitive peaks and rapid separations. One example, the Personal Ion Analyzer 1000 available from Shimadzu (Kyoto, Japan), is a portable (15 kg) nonsuppression ion chromatograph. It is operated either by AC (power supply) or DC (battery), enabling analysis to be performed on site or where there is limited laboratory space. Ion chromatographs available from other vendors are integrated but also include an autosampler and sit on a laboratory bench. Conventional high-performance liquid chromatography (HPLC) hardware [1–3] has been undergoing an evolution in recent years to be able to operate at higher pressures and with smaller-diameter columns. In some cases, the newer hardware is called Ultra HPLC or UHPLC. Regardless of what it is called, HPLC hardware has many similarities to IC hardware, and having the option of using this instrumentation can be a valuable resource when determining the best hardware for a particular ion analysis problem, or at least assembling a workable system. In this chapter, we describe the various components of an ion chromatographic instrument, their function, how the instrument is built, and how to recognize parts of the instrument in the event that maintenance is needed. An understanding of some of the available instrumental options is also helpful in achieving better separations or using other detectors different from those included with an integrated IC instrument.
Ion Chromatography, 4th Ed. James S. Fritz and Douglas T. Gjerde Copyright © 2009 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim ISBN: 978-3-527-32052-3
22
2 Instrumentation
Figure 2.1 Block diagram of an ion chromatograph.
The block diagram in Figure 2.1 shows the general arrangement of components of an IC instrument. The essential hardware requirements are as follows: 1. Eluent reservoir. Electrical generation of eluent ions is used in some cases. 2. Pump and flow regulator. An optional gradient system permits programmed changes in eluent composition. 3. Sample injection unit. Purpose is to introduce a precise volume of sample for ion analysis. 4. Column and column oven to maintain a constant, pre-set temperature. 5. Suppressor. This useful component is only needed for systems that employ suppressed-conductivity detection. 6. Detector. Function is to detect sample ions as they are separated and pass through the detector. 7. Data processor. This unit is linked to the detector output so that a digital record of the chromatographic separation is obtained. The complete chromatogram can be viewed on a computer screen and printed out if desired. Each of these components is described in the following sections. The setup typically includes the IC instrument itself and a small computer with a display screen. The entire assembly occupies only a small amount of bench space.
2.2 General Considerations
2.2 General Considerations
Everything on the high-pressure side in the system, from the pump outlet to the end of the column, must be strong enough to withstand the pressures involved. The wetted parts are usually made of PEEK and other types of plastics, although other materials, such as sapphire, ruby, or even ceramics are used in the pump heads, check valves, and injector of the system. PEEK and other high-performance plastics are the materials of choice for ion chromatography because of their ability to withstand high fluid pressures. Stainless steel, usually found in HPLC equipment, can be used provided that the system is properly conditioned to remove internal corrosion and the eluents that are used do not promote further corrosion. Almost all IC eluents are not corrosive to stainless steel provided that this has been pretreated so that surface corrosion is not present. Most acids including sulfuric and nitric acid are not corrosive; however, hydrochloric acid is extremely corrosive to stainless steel. The reader is advised to consult the instrument manufacturer for care and upkeep instructions. It has been said that having a basic knowledge of plumbing can be valuable to the ion chromatograph user. It is interesting to note that, just as household plumbing requires some skill to make kitchens and bathrooms work properly, the plumbing in ion chromatography has to be done correctly for the same reason. One of the greatest points of concern is to configure the instrument so that system dead volume is kept to a minimum. System dead volume is any empty space or volume that the fluid occupies in an IC system. This includes all of the fluid path volume from the injection volume to the detector cell including the injection valve loop and fluid path, the tubing, tubing fittings and union, column end fitting, the space between the beads in the column and the detector fluid path. Too much dead volume or un-swept volumes will lead to peak broadening or peak tailing and consequent loss in separation efficiency. The system dead volume can be controlled by using small internal diameter tubing and keeping the tubing as short as is practical. It is important to use small bore tubing (0.007 inch, 0.18 mm) in short lengths when making connections between the injection and the column and between the column and the detector. Dead volume from the pump to the injector should also be kept small to help to make possible rapid changes in the eluent composition in gradient elution. The tubing fittings must be connected properly by bottoming out the end of the tubing into the fitting before final tightening. Although all regions in the flow path are important, the most important region where peak broadening can happen is in the tubing and connections from the exit end of the column to the detector cell.
23
24
2 Instrumentation
2.3 Eluent
The prepared eluents are contained in a reservoir and pumped at a constant or programmed rate into the chromatographic system. Eluents are usually prepared by dissolving buffers, acids or bases in aqueous solvent or in an aqueous–organic solvent mixture. Eluent entering the pump should not contain any dust or other particulate matter. Particulates can interfere with pumping action and damage the seals or valves. Material can also collect on the inlet frits or on the inlet of the column, causing pressure buildup. Eluents or the water and salt solutions used to prepare the eluents are normally filtered with a 0.2 or 0.45 lm nylon filter. Various eluents that can be used with ion chromatography are described throughout this book. In this chapter, hydroxide-based eluents are described because they are generated electrolytically by the instrumentation rather than simply mixing a base with water. NaOH, KOH or other hydroxide eluents are desirable because of their low suppressed background conductivity. This leads to the ability to form eluent gradients with small shifts in the baseline. Also, low background conductivity can improve the detection limit for ions. However, hydroxide is a weak driving anion, and high concentrations are needed to elute anions that are strongly held by the column. Also, unless the column is designed to operate with hydroxide eluents, the analyte peaks may not be symmetrical. For many workers, the advantages of gradients and low detection limits offset the disadvantages of long retention times and possible asymmetrical peaks. Pure hydroxide eluents are difficult to make, because of persistent contamination by carbon dioxide that is converted to carbonate by the high pH. Carbonate is a much stronger eluting anion than hydroxide, and its presence can shift sample retention times to much shorter (and inconsistent) retention times. Carbonate will also cause baseline shifts when gradients are generated. In fact, a baseline shift during a hydroxide gradient is a good diagnostic indication that one or more of the eluent reservoirs contain bicarbonate or carbonate anion. The electrolytic generation of hydroxide eluent was first described by Dasgupta and coworkers [4, 5]. Rather than mixing reagents, the hydroxide eluent is formed electrochemically as it is being used and is introduced directly into the elution column from the generator. The system permits direct electrical control of the eluent concentration, and gradient chromatography is accomplished without mechanical proportioning. The system contains an anode and a cathode across which a DC current is passed. The reduction reaction at the cathode produces the hydroxide anion. 2 H2O + 2 e– → 2 OH– + H2 ↑ (at cathode) A counterion to hydroxide is needed to conserve electric neutrality. Also, an oxidizing reaction occurs simultaneously at the anode. OH– is electrolytically neutralized and O2 is evolved.
2.3 Eluent
H2O – 2 e– → 2 H+ + 1⁄2 O2 ↑ (at anode without NaOH) However, the feed solution for the anode also contains NaOH. 2 Na+ + 2 OH– + H2O – 2 e– → 2 Na+ + 2 H2O + 1⁄2 O2 ↑ (at anode with NaOH) The Na+ and OH– are combined to form the eluent through the use of an ionexchange membrane. A cation-exchange membrane separating the anode from eluent flow allows the Na+ to join the OH– from the cathode.
Figure 2.2 Schematic representation of EG40 electrolytic production of potassium hydroxide (KOH) eluent (courtesy of Dionex Corp.)
A variation on the concept has been introduced by Dionex as the EG40 module [6]. In this case, KOH contained in a reservoir (labeled in Figure 2.2 as K+ Electrolyte Reservoir) is used rather than an NaOH feed solution. The process is the same; however, K+ is generated from the anode, because its counterion OH– is consumed in the production of H+. K+ migrates across the cation-exchange membrane to combine with OH– formed at the cathode. Carbon dioxide is removed from the eluent stream en route to the EG40, to prevent contamination by carbonate. The electrolyte reservoir must be changed when the K+Electrolyte is depleted. An analogous system can be used to generate methanesulfonic acid (MSA) eluent for the separation of cations (Figure 2.3). In this case, the anode generates H+ for eluent production. The cathode generates OH– anion that combines with the H+ in the MSA electrolyte reservoir. MSA– anion migrates across the anionexchange membrane to combine with the H+ eluent cation (maintaining electric neutrality). Dionex offers a ‘just add water’ system for electrolyte generation and purification of commonly used eluents such as KOH and methanesulfonic acid. Eluents are generated from deionized water using an Eluent Generator (EG) cartridge and then polished of contaminants using a continuous-regeneration trap column.
25
26
2 Instrumentation
Figure 2.3 Schematic representation of EG40 electrolytic production of methanesulfonic acid (MSA) eluent (courtesy of Dionex Corp.)
While very popular, the use of basic or acidic eluents with suppressed conductivity detection is certainly not the only useful form of IC. It is often advantageous to use another chemical type of eluent in conjunction with detection by nonsuppressed conductivity, UV–Vis spectroscopy, and electrochemical methods, among others. Eluents can be prepared at almost any desired pH and are generally quite stable except for eluents that are at very high pH. A freshly prepared eluent should be filtered before use to remove any sediment.
2.4 Pump
The IC system begins with the pump, which should be completely inert, robust, and capable of high precision and accuracy needed for reproducible chromatographic results. Ideally, the pump should support a range of flow rates consistent with using columns having diameters in the range 2–9 mm. Smaller column diameters usually require lower pump flow rates. The pump should also incorporate or have available vacuum degassing of the mobile phase and provide for gradient elution through a proportioning valve. Both single-piston (SP) and dual-piston (DP) pumps are available. IC pumps are designed around an eccentric cam that is connected to a piston (Figure 2.4). The rotation of the motor is transferred into the reciprocal movement of the piston. A pair of check valves controls the direction of flow through the pump head (discussed below). A pump seal surrounding the piston body keeps the eluent form leaking out of the pump head. The pump seal will wear and must be replaced periodically. In single-headed reciprocating pumps, the eluent is delivered to the column for only half of the pumping cycle. A pulse dampener is used to soften the spike of
2.4 Pump
Figure 2.4 IC pump head, piston, and cam.
pressure at the peak of the pumping cycle and to provide an eluent flow when the pump is refilling. Use of a dual-headed pump is better because heads are operated 180° out of phase with each other. One pump head pumps while the other is filling and vice versa. The eluent flow rate is usually controlled by the pump motor speed although there are a few pumps that control flow rate by control of the piston stroke distance. Figure 2.5 shows how the check valve works. On the intake stroke, the piston is withdrawn into the pump head, causing suction. The suction causes the outlet check valve to settle onto its seat while the inlet check valve rises from its seat, allowing eluent to fill the pump head. Then the piston travels back into the pump head on the delivery stroke. The pressure increase seals the inlet check valve and opens the outlet valve, forcing the eluent to flow out of the pump head to the injection valve and through the column. Failure of either of the check valves to seal properly will cause pump head failure, and eluent will not be pumped. In most cases, this is due to air trapped in the valve so that the ball cannot sit properly. Flushing or purging the head usually takes care of this problem. Using degassed eluents is also helpful. In a few cases, particulate material can prevent sealing of the valve, and in these cases the valve must be cleaned or replaced. The pump manufacturer has instructions on how to perform this operation.
27
28
2 Instrumentation
Figure 2.5 Check valve positions during intake and delivery strokes of the pump head pistons.
2.4 Pump
2.4.1 Gradient Formation
Isocratic separations are performed with an eluent at a constant concentration of eluent buffer or salt solution. Isocratic elution is desirable because it is simpler, but it is sometimes necessary or desirable to perform separations using eluent
Figure 2.6 High-pressure mixing systems use two or more independent pumps to generate the gradient. Low-pressure mixing systems use a single pump with a proportioning valve to control composition. The advantages of high-
pressure mixing are smaller dwell volumes and faster gradient formation. The advantages of low-pressure mixing are lower costs (single pump) and more versatile gradients (four solvents).
29
30
2 Instrumentation
gradients where the eluent strength is increased gradually over a chromatographic run. This allows the separation of anions that may have a wide range of affinities for the column. Weakly adhering anions elute first, and then, as the eluent concentration is increased, more strongly adhering anions can be eluted by the stronger eluent. Figure 2.6 shows the two most popular methods for forming gradients. In the first method, flow from two high-pressure pumps is directed into a high-pressure mixing chamber. One pump contains a weak eluent while the other contains the stronger eluent. After the mixing chamber, the flow is directed to the injector and then on to the column. Control of the relative pumping rates of each pump creates the gradient. The total flow from the two pumps is constant. The gradient starts with a high flow of the weak eluent pump and a low flow of the strong eluent pump. Then, over the course of the chromatographic run, the flow rate of the strong eluent pump is increased while the flow rate of the weak eluent pump is decreased, keeping the total flow rate constant. A more popular and less expensive method of forming gradients is by using a single pump and three or four micro-proportioning valves at the inlet of the pump. Each valve controls an eluent of a given concentration. Only one valve is open at a time. The concentration of the eluent that is delivered to the pump depends on the relative time that any particular valve is open. To start a gradient, the valve controlling the low concentration of the eluent is open for longer periods of time. As a gradient increases, this valve will close off while the valve controlling higher eluent concentrations is opened. Most gradients are formed with two of the valves although it is certainly possible to use more complex gradients with 3 or even 4 valves within one run. Normally, the other valves control different types of eluents or column cleaning solutions.
2.5 Sample Injector
The sample injection hardware is designed to introduce a small (1–100 lL) and reproducible volume of sample into the ion chromatograph. The injection system may be manual or automated. An automated system (often called an autosampler) permits the storage of multiple samples for unattended injection and analysis. The injection system may be manual or automated, but both rely on the injection valve. An injection valve is designed to introduce precise amounts of sample into the sample stream with variation usually less than 0.5% volume difference from injection to injection. Figure 2.7 schematically represents a 6-port and 2-position device valve. In one position the sample is loaded and the other it is injected. In the load position, the sample from the syringe or autosampler vial is pushed into the injection loop. The loop may be partially filled (partial loop injection) or completely filled (full loop injection). Partial loop injection depends on the precision filling of the loop with small known amounts of material. If partial loop injection is used, the loop must not be filled to more than 50% of the total
2.6 Columns
loop volume or the injection may not be precise. In full loop injection, the sample is pushed completely through the loop. Normally at least a two-fold amount of sample is used to fill the loop with excess sample from the loop going to waste. Typical loop sizes are 10–200 lL.
Figure 2.7 Schematic representation of partial- and full-loop injection methods.
At the same time that the sample loop is loaded with sample, the eluent travels in the by-pass channel of the injection valve and to the column. Injection of the sample is accomplished by turning the valve and placing the injection loop into the eluent stream. Usually the flow of the eluent is opposite to that of the loading of the sample into the loop. The injected sample travels to the head of the column as a slug of fluid. The ions in the sample interact with the column and the separation process is started with the eluent pushing the sample components down the column. Injection valves require periodic maintenance and usually have to be serviced after about 5000 injections. The instrument manual should be consulted for details on service.
2.6 Columns 2.6.1 Column Hardware
IC columns are usually made of PEEK or other polymer. Even the frits at the end of the column, which hold the column packing in place, are usually made of porous PEEK. The column lengths range from about 3 to 30 cm and the inside diam-
31
32
2 Instrumentation
eters from about 1 to 7.8 mm. Figure 2.8 shows the end of a column and the type of fitting used to connect the tubing to the column. Reusable PEEK fittings are used almost exclusively to connect tubing to columns and other instrument components. As stated earlier, the tubing should be bottomed out or pushed completely into the column end before the fitting is tightened, to ensure that there is no unnecessary dead volume in the connection.
Figure 2.8 Representation of a typical IC column and end fitting. The fittings, frits and body are normally PEEK. The PEEK fittings to attach the tubing to the column are normally tightened by hand and not with a wrench. Some columns will contain a replaceable disk or frit at the top of the column to protect the column from particulates and contaminants. The trend is toward smaller-diameter columns.
2.6.2 Column Protection
Not only does column protection extend the useful life of the separation column, but proper protection of the separation column can also result in more reliable analytical results over the lifetime of the column. Scavenger columns, located between the pump and the injector, are one means of protecting the column. The scavenger removes particulate material that may be present in the eluent, but can also contain a resin to ‘polish’ the eluent of any contaminant. An example is a chelating resin to remove metal contaminants. Besides protecting the separation column, scavenger columns may also improve detection of the analytes by reducing the background signal due to residual contaminants. Most IC users do not use scavenger columns but rather prefer the use of guard columns, located directly in front of the separation column. Guard columns generally contain the same material as the separation column. Therefore, material that would be trapped and would contaminate the separation column will instead get trapped by the guard column. Guard columns are changed when the separation of a standard is no longer acceptable and the column cannot be regenerated by the recommended procedures. Several guard columns may be used for protec-
2.7 Suppressor
tion over the lifetime of the separation column. Guard columns are generally smaller than the separation columns, but can add to the retention time of the separation. Some users may prefer to use an inline filter in place of or even in addition to the guard column. 2.6.3 Column Oven
There is about a 2% change in conductance per degree C change in temperature. If the eluent has a background conductance, then temperature control is important to reduce detector noise and improve detection limits. Good quality conductivity detectors have temperature control, temperature compensation, or both. An oven can keep the temperature of the fluid constant by the time it reaches the conductivity cell, and this also helps to improve detector noise and detection limits. Some column ovens are designed to preheat the eluent to a pre-set temperature as it enters the column. Use of a column oven permits IC separations to be run at temperatures ranging from sub-ambient to around 85 °C. 2.6.4 Two-dimensional IC
Applications that employ valve switching have been successfully used in ion chromatography. Systems are available that permit sample matrix diversion or matrix elimination prior to analysis of trace components. The strategy used is as follows: In the first stage, a large-volume sample loop is used to obtain a partial separation. The objective here is to focus the partially resolved peaks of interest onto a concentrator column in the second stage. This is accomplished by valve switching. The second column is of smaller diameter and is operated at a lower flow rate relative to the first column. Resolution of the peaks of interest is completed on the second column. Improved resolution is frequently possible by using eluents with different chemical compositions in the two columns.
2.7 Suppressor
A suppressor is only needed when analyte ions are to be detected by suppressed conductivity. Hydrogen ions pass through a cation-exchange membrane and convert a basic eluent to a low-conducting weak acid. An equivalent amount of the eluent cation, K+, passes through the membrane in the opposite direction to maintain electro-neutrality. An analyte anion, such as Cl–, now has the highly conductive H+ as its counterion, which can now be detected on top of the low-background eluent signal that has now been suppressed. An acidic eluent such as methanesulfonic acid (MSA), which is used in cation chromatography, works on essentially the same principle. The H+ of the eluent is
33
34
2 Instrumentation
neutralized by OH– from the basic regenerant which passes through an anionexchange membrane. Analyte cations are now paired up with OH–, which has a high conductance. The original micromembrane suppressor, introduced by Dionex in 1991, required a chemical regenerant with a flow three to ten times faster than the eluent flow. This problem was overcome by the Self Regenerating Suppressor (SRS) [7]. In this system, H+ or OH– for suppression is generated by electrolysis of water at a Pt anode (or cathode) placed in the regenerant chambers. The current required to start the electrolysis is proportional to the eluent concentration that has to be suppressed. A continuously electrolytically regenerated packed-bed suppressor (known as the Atlas) was made available by Dionex in 2001 [8]. This instrument contains monolithic cation-exchange disks attached to flow distributor disks with small holes to allow the liquid to flow. These are placed between the ion-exchange membranes. New models of membrane suppressors work with high eluent concentrations and have very low dead volume. Another suppression device uses a packed-bed suppressor in disposable cartridges. A device with solid-phase electrochemical cells called Electrically Regenerated Ion Suppression (ERIS) is marketed by Alltech [9]. A degassing unit for removing CO2 from suppression of carbonate is included in a more recent suppressor. Metrohm markets a column suppressor module that uses two suppressor columns. One is regenerated while the other is being used. When the first is exhausted then the other is available for eluent suppression.
2.8 Detector
The various types of detectors that can be used in ion chromatography are discussed in Chapter 4. Variable wavelength UV–Vis detectors are extremely useful for detection of sample ions with sufficient absorbance at the analytical wavelength. Various electrochemical detectors, including the pulsed amperometric detector (PAD), offer excellent selectivity and sensitivity. Detection by mass spectrometry is growing in popularity. This is undoubtedly a result of the high detection sensitivity and positive identification offered by MS in association with the fast development of interfaces between column and detector that have greatly simplified the IC/MS combination. Several combinations of IC and MS detection are available. These include ICP-MS (element-specific detection) and MS with atmosphere pressure ionization (API), operated with either electrospray ionization (ESI) or atmospheric pressure chemical ionization (APCI). The use of IC-MS where the column outlet is connected directly to a mass spectrometer may impose additional requirements on the IC system. Since a vacuum must be maintained, more volatile ammonium salts rather than sodium or potassium salts are used in the eluent. Inside the mass spectrometry interface evaporation of the IC liquid, ionization of neutral species to charged species, and removal
2.9 Data Acquisition and Calculation of Results
of a very large amount of vapors from the mobile phase, must take place to maintain the required vacuum conditions. But of all of the many choices for detector, the conductivity detector remains the most popular. Conductivity detection is universal, rugged and sensitive. Adding a suppressor (discussed in Chapter 4 and elsewhere) dramatically improves the sensitivity for the majority of anions. The only exception to this is weak base anions which become low conducting in a suppressor. But the gain in detection for most anions and cations is so impressive that suppressed conductivity is the most popular form of detection, which means that a suppressor is placed after the exit of the separation column and before the conductivity cell. Membrane suppressors are marketed by Dionex. As described above, several column suppressor patents have expired and there are several companies offering this type of suppressor. Nonsuppressed conductivity, where the exit end of the column is connected directly to the conductivity cell, is effective for measuring many types of cations and weak acid anions.
2.9 Data Acquisition and Calculation of Results
Chromatographic data acquisition has come a long way from earlier days when the output of the detector was connected to a pen-and-ink recorder moving at a fixed chart speed so that the resulting chromatogram could be recorded. Now, the results of the chromatographic separation are almost always stored and displayed on a computer. The computer uses an A/D (analog to digital) board to convert the analog signal from the detector to digital. The digital information is stored and manipulated to report the results to the user. The scale of both the vertical axis (detector signal) and the horizontal axis (elution time) can be adjusted to give a record of the separation. The most useful types of information are the peak retention times and the peak areas or peak heights. Retention times are used to confirm the identity of the various peaks, and peak area or peak height is a measure of concentration. Two general methods can be used to calculate the concentration of a given peak: use of a calibration curve or use of a standard-to-unknown ratio. Generally, it is usually better to prepare a calibration curve by a plot of peak area (or height) against the concentrations of known standards. Such a plot will normally be a straight line, but a perfectly valid calibration plot may deviate from linearity, especially at lower concentrations. If the calibration curve is linear and passes through zero, a fast method is to compare an analytical peak of interest to a single standard of known concentration. This calculation can be performed by use of a simple ratio: unknown concentration known concentration unknown peak area known peak area
35
36
2 Instrumentation
therefore: unknown concentration
known concentration × unknown peak area known peak area
If the calibration curve is not linear or does not pass through zero, then it best to use the calibration curve to calculate unknown concentrations.
References [1] C. F. Poole and S. K. Poole, Chromato-
[2]
[3]
[4]
[5]
graphy Today, Elsevier Science, Amsterdam, 1991. M. W. Dong, Modern HPLC for Practicing Scientists, Wiley Interscience, New York, 2006. S. Kromidas, Ed., HPLC Made to Measure: A Practical Handbook for Optimization, Wiley VCH, Weinheim, 2006. D. L. Strong and P. K. Dasgupta, K. Friedman and J. R. Stillian, Electrodialytic eluent production and gradient generation in ion chromatography, Anal. Chem., 63, 480, 1991. D. L. Strong, C. U. Joung, P. K. Dasgupta, Electrodialytic eluent generation and suppression: ultralow background conductance suppressed
[6] [7]
[8]
[9]
anion chromatography, J. Chromatogr., 546, 159, 1991. New products brochure EG40, Dionex Corp. Sunnyvale, CA , 1999. C. Pohl, R. Slingsby, V. Barreto, K. Friedman and M. Toofan, New membrane-based electrolytic suppressor device for suppressed conductivity detection in ion chromatography, J. Chromatogr., 640, 97, 1993. H. Small, Y. Liu, J. Riviello, N. Avdalovic, K. Srinivasan. Continuous Electrolytically Regenerated Packed Bed Suppressor for Ion Chromatograph, U.S. Patent 6 325 976, 2001. L.-M. Nair and R. Saari-Nordhaus, Recent developments in surfactant analysis by ion chromatography, J. Chromatogr. A, 804, 233, 1998.
37
3 Resins and Columns 3.1 Introduction
Most ion exchangers used for ion chromatography now have a polymeric base. This is quite the opposite to high-performance liquid chromatography (HPLC) where the use of silica-based materials still predominates. For convenience, we will frequently use the word ‘resins’ to refer to both silica-based and polymeric ion exchangers. Intensive research and development has resulted in ion-exchange materials with much greater selectivity and efficiency than was possible in the earlier days of IC. Some of the new columns are simply improvements of older columns having sharper peaks and shorter analysis times. One interesting side benefit to this is that as peaks become sharper the ability to detect anions at lower levels is also improved. Other column development has been driven by specific needs such as the need to analyze drinking water disinfectants. While it has become increasingly rare for practitioners to make their own resins and pack their own columns it is still important to understand how resins for the columns are made. In this way, both the power and the limitations of the commercial resins can be recognized and used. The chemistry and properties of ion-exchange resins and columns are discussed in this chapter. Specific commercial columns and the practice of anion chromatography are covered in Chapter 5 and cation chromatography in Chapter 7. A cation exchanger is a solid particulate material with negatively charged functional groups arranged to interact with ions in the surrounding liquid phase. For convenience, we will often refer to a cation exchanger as a ‘catex’. The most common type of catex contains sulfonic acid groups. Cross-linked polystyrene particles are converted to a catex by sulfonation with concentrated sulfuric acid. (3.1)
Ion Chromatography, 4th Ed. James S. Fritz and Douglas T. Gjerde Copyright © 2009 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim ISBN: 978-3-527-32052-3
38
3 Resins and Columns
In this equation, Res denotes resin or polymer. Silica-based cation exchangers are generally prepared by reacting silica particles with an appropriate chlorosilane or methoxysilane. A common type of silica catex has the structure:
In both of these materials, the sulfonate group is chemically bonded to the solid matrix. However, the H+ is attracted electrostatically to the –SO3– and can undergo exchange reactions with other ions in solution. For example: Solid–SO3– H+ + Na+ > Solid–SO3– Na+ + H+
(3.2)
2 Solid–SO3– Na+ + Mg2+ > (Solid–SO3–)2 Mg2+ + 2Na+
(3.3)
The physical form of the catex is such that ions from the surrounding solution can readily traverse through the solid to come into contact with the interior as well as the surface sulfonate groups. The exchange reactions (Eqs. 3.2 and 3.3) are reversible and are subject to the laws of chemical equilibrium. Most monovalent metal ions are more strongly held by the catex than H+. Cations of a higher charge are usually retained more strongly than those of lower charge. Ion-exchange equilibria are treated in more detail in Chapter 5.
3.2 Polymeric Resins 3.2.1 Substrate and Cross Linking
Historically, cation- and anion-exchange resins were intended for large-scale applications such as water softening and various areas of chemical processing. These materials were designed to be durable with repeated use and to provide a high ion-exchange capacity with good pH stability. Ion exchangers used in contemporary ion chromatography have a much lower exchange capacity and are designed to emphasize chromatographic separation ability with high efficiency. However, the IC resins of today gradually evolved from earlier resin technology and synthetic methods for introduction of various functional groups. A variety of polymeric substrates can be used in ion-exchange synthesis, including polymers of esters, amides and alkyl halides. But resins based on styrene-divinylbenzene copolymers are probably the most widely used ion exchangers. The polymer is schematically represented in Figure 3.1. The resin is made up primarily of polystyrene; however, a small amount of divinylbenzene is added during the polymerization to ‘cross-link’ the resin. This cross-linking confers mechanical sta-
3.2 Polymeric Resins
bility upon the polymer bead and also dramatically decreases the solubility of the polymer by increasing the molecular weight of the average polymer chain length. Typically, 12–15 wt% of the cross-linking compound is used for microporous resins and up to 55 wt% for macroporous resins. In many cases, the resin name will indicate the cross-linking of the material. For example, a Dowex 50 x 4 cation exchanger contains 4% divinylbenzene polymer.
Figure 3.1 Schematic representation of a styrene–divinylbenzene copolymer. The divinyl-benzene ‘cross-links’ the linear chain of the styrene polymer. A high percentage of divinylbenzene produces a more rigid polymer bead.
3.2.2 Microporous Resins
The starting material for cation- and anion-exchange ‘polymer’ resins can be classified either as microporous or macroporous. Most classical work has been done with microporous ion-exchange resins. Microporous substrates are produced by a suspension polymerization in which styrene and divinylbenzene are suspended in water as droplets. The monomers are kept in suspension in the reaction vessel through rapid, uniform stirring, and the use of a surfactant. Addition of a catalyst such as benzoyl peroxide initiates the polymerization. The resulting beads are uniform and solid but are said to be microporous. The size distribution of the beads is dependent on the stirring rate, that is, faster stirring produces smaller beads. The beads swell but do not dissolve when placed in common hydrocarbon solvents. After the resin is functionalized (the ion-exchange functional groups are attached to the polymer), the bead is considerably more polar. Depending on the relative number of functional groups, polar solvents such as water will now swell the ion-exchange resin. However, nonpolar solvents will tend to dehydrate the bead and cause it to shrink.
39
40
3 Resins and Columns
The extent of ion-exchange resin hydration will also depend on the ionic form of the resin. Ion-exchange resin beads with very little cross-linking are soft and tend to swell or shrink excessively when converted from one ionic form to another. However, the amount of cross-linking used in resin synthesis is still based on a compromise of resin performance. Microporous polystyrene resins usually contain about 8% divinylbenzene. Gel-type resins with a high cross-linking tend to exclude larger ions, and the diffusion of ions of ordinary dimensions within the gel may be slower than might be desired. Resins with cross-linking lower than about 2% are too soft for most column work. 3.2.3 Macroporous Resins
Macroporous resins (sometimes called macroreticular resins) are prepared by a special suspension polymerization process. Again, as with microporous resins, the polymerization is performed while the monomers are kept as a suspension of a polar solvent. However, the suspended monomer droplets also contain an inert diluent that is a good solvent for the monomers, but not for the material that is already polymerized. Thus, resin beads are formed that contain pools of diluent distributed throughout the bead matrix. After polymerization is complete, the diluent is washed out of the beads to form the macroporous structure. The result is rigid, spherical resin beads that have a high surface area. Rather than using an inert solvent to precipitate the copolymer and form the pores, the polymerization may be carried out in the presence of an inert solid agent such as finely divided calcium carbonate to create the voids within the bead. Later, the solid is extracted from the copolymer. Both of these polymerization processes create large (although probably different) inner pores. The average pore diameter can be varied within the range of 20 Å to 500 Å. The final resin bead structure of a macroreticular resin contains many hard microspheres interspersed with pores and channels. Because each resin bead is really made up of thousands of smaller beads (something like a popcorn ball), the surface area of macroporous resins is much higher than that of microporous resins. A gel resin has a (calculated) surface area of less than 1 m2 g–1. However, macroporous resin surface areas range from 25 to as much as 800 m2 g–1. Macroporous resins are remarkably rigid because of the large amounts of crosslinking agents normally used in the synthesis. Such resins are particularly advantageous for performing ion-exchange chromatography in organic solvents since changing solvent polarity does not swell or shrink the resin bed as it might for a gel-type resin. But the high cross-linking does not inhibit the ion-exchange process as it does in gel resins because the resins have pores and channels that are easily penetrated by the ions.
3.2 Polymeric Resins
3.2.4 Chemical Functionalization
Ion exchangers are created by chemically introducing suitable functional groups into the polymeric matrix. In a few instances, monomers are functionalized first and then polymerized into beads. An attractive feature of the aromatic copolymer used in many ion exchangers is that it can be modified easily by a wide variety of chemical reactions. More recently, some ion-exchange substrates have been polymers of esters (polymethacrylate) or amides. The reaction solvent is important in ion-exchange synthesis. In many cases, gel substrates must first be swollen in the reaction solvent to achieve complete functionalization of the resin. (We shall see that complete functionalization is usually not desired in resins used in ion chromatography.) The reaction solvent does not appear to be as critical in ion-exchange synthesis of macroporous substrates. Their surface is already ‘exposed’ and ready to be converted to an ion exchanger. Weak-acid cation-exchange resins that contain carboxylic acid or phosphonic acid ion-exchange groups are often used. A popular type of cation exchanger is made by introducing a sulfonic acid functional group. Resins with the sulfonic acid group are said to be strong-acid ion exchangers. Sulfonation reactions are performed by treating the polystyrene resin with concentrated sulfuric acid. Alternatively, the beads can be reacted with chlorosulfonic acid to produce a sulfonyl chloride group, and the sulfonyl chloride group is then hydrolyzed to the acid. Presumably, the latter reaction effects a more uniform placement of the ionogenic groups because the chlorosulfonic acid reagent is dissolved in an organic solvent. The solvent swells the bead, allowing free access of the chlorosulfonic acid to the aromatic rings. Concentrated sulfuric acid is more polar. Sulfonation with this reagent occurs first on the bead surface and then moves progressively toward the center of the bead. Even though this product is not as homogeneous, resins prepared with concentrated sulfuric acid are more popular for ion chromatography. The anionic group –SO3– that is produced is chemically bound to the resin and its movement is thus severely restricted. However, the H+ counterion is free to move about and can be exchanged for another cation. When a solution of sodium chloride is brought into contact with a cation-exchange resin in the hydrogen ion form, the following exchange reaction occurs: (3.4) If this reaction goes essentially to completion, the resin is said to be in the sodium (ion) form. Traditionally, anion-exchange resins have been made by a two-stage set of reactions (although other synthesis methods are now being used). The first step is a Friedel-Crafts reaction to attach the chloromethyl group to the benzene rings of styrene-divinylbenzene copolymer. Then the anion exchanger is formed by reaction of the chloromethylated resin with an amine. The most common type of
41
42
3 Resins and Columns
strong-base anion-exchange resin contains a quaternary ammonium functional group, which is obtained by alkylation with trimethylamine.
(3.5)
In these resins only the anion is mobile and can be exchanged for another anion. Another common strong-base anion exchanger is one that contains a hydroxyethyl group in place of a methyl group on the nitrogen. Weak-base anion exchangers are synthesized by reacting the chloromethylated resin with lower substituted amines or with ammonia. Weak-base anion-exchange resins cannot function as ion exchangers unless the functional group is protonated: Resin-CH2NH2 + H+ + NO3– → Resin-CH2NH3+ NO3–
(3.6)
The protonation, of course, depends on the basicity of the functional group and the pH of the solution in which the resin beads are immersed. 3.2.5 Resin Capacity
Resin capacity is an extremely important parameter in ion chromatography. Details of the effect of capacity on the behavior of resins for ion chromatography are found in Chapter 5. Generally, resins of lower capacity will lower the eluent concentration needed to elute sample ions from the column. The capacity of a resin is usually given in milliequivalents of exchangeable ion per gram of resin. In some cases it is expressed as milliequivalents per milliliter of resin. High-capacity commercial cation resins contain approximately one functional group per benzene ring and have an exchange capacity of around 4.5 mequiv g–1. Highcapacity anion-exchange resins are typically around 3.5 to 4.0 mequiv g–1 for a geltype resin and around 2.5 mequiv g–1 for a macroporous resin. In almost all cases, the bulk resins that are available commercially are high capacity. However, ion chromatography usually employs low-capacity ion exchangers. Capacities of these resins range from 0.01 to 0.2 mequiv g–1. Many low-capacity resins are pellicular, that is, the ion-exchange groups are on or near the surface of the bead. The density of the functional groups is high on the surface so that the local capacity is high, but the overall capacity of the ion exchanger is very low. However, macroporous resin substrates are porous and have a much higher surface area. Low-capacity ion exchangers made from these substrates have the functional groups distributed throughout the bead, although the exact relationship of the performance of the functional group to its position within the resin matrix is unknown.
3.3 Resins and Columns for Ion Chromatography
Anion-exchange resins of variable but low exchange capacities are produced under mild conditions and short reaction times in the chloromethylation reaction. Conditions for the amination are chosen to convert as much of the chloromethyl group as possible to the quaternary ammonium chloride, although experience indicates that some of the chloromethyl remains unreacted. A procedure devised by Barron and Fritz [1] uses concentrated hydrochloric acid and paraformaldehyde with a Lewis acid catalyst to chloromethylate the polymer. Exceptional control of the extent of chloromethylation is possible by adjusting the concentration of reagents, the reaction temperature, and the reaction time. Chloromethyl methyl ether is not generated in situ, except for possible traces, so the reaction is relatively safe to use. Following the chloromethylation step, amination is carried out by adding a large excess of 25% trimethylamine in methanol or water and allowing the reaction to proceed overnight. Depending on the conditions chosen, this procedure can produce anion exchangers with capacities from 0.005 to 0.16 mequiv g–1 when using XAD-1 as a substrate. This range includes the capacities most useful in ion chromatography.
3.3 Resins and Columns for Ion Chromatography
A large variety of columns for IC is available. Particle size of the resin column packing has steadily decreased from about 10 lm to 5 lm, and now to 3 lm. Resins are spherical and have a narrow particle size range. Column lengths are typically 5, 10, 15 or 25 cm. Column diameter is generally 4.0–4.6 mm, although columns 2.0 mm in diameter are also available. Both anion- and cation-exchange columns are available with a choice of resin matrix (polymeric or silica) and ionic functional group. Modern IC resins often contain imbedded polar groups in addition to the ion-exchange groups. The purpose is to reduce the hydrophobic nature of the solid ion exchanger and to modify the nonionic attraction of analyte ions for the solid phase. The effects of resin composition on anion selectivity will be discussed Section 3.4. 3.3.1 Monolith Columns
Enhanced chromatographic efficiency is obtained by using a column packed with very small ion-exchange particles. However, the pressure required to push eluent through a packed column increases rapidly with decreased particle size. A new and different type of chromatographic column called a ‘monolith’ offers significant advantages to traditional packed columns. A monolith support is a single polymeric rod that can be prepared by pouring a chemical mixture into a column and carrying out the polymerization in place. The monolith structure is designed to contain an uninterrupted, interconnected network of channels of a controlled
43
44
3 Resins and Columns
size range. The back pressure is much lower than with a packed column with particles of similar size. Svec [2] has reviewed the preparation of monoliths for chromatography. Monoliths are described as separation media in a format that can be compared with a single large particle that does not contain any interparticular voids. As a result, all of the mobile phase must flow through the stationary phase. This convective flow greatly accelerates the rate of mass transfer. In contrast to diffusion, which is the driving force for mass transfer within the pores of particulate phases used in packed columns, convective flow through the pores enables a substantial increase in the speed of large biomolecules. In packed-column chromatography, interstitial voids – empty spaces between the packing particles – take up space in the column but do not aid in the separation. In an ideally packed column with equal-sized spherical particles, some 30% of the column volume is lost to voids [3]. As the mobile phase is pumped through the column during the course of a chromatographic run, the fluid flows freely through the voids but meets resistance as it permeates the interior of the porous packing materials. When a sample solution is injected into the column, differences between analyte concentrations in the void spaces and interiors of the particles cause the analytes to be transferred back and forth between the two regions. The time required for equilibration reduces chromatographic efficiency and can result in slow separations, especially for larger analytes which tend to move sluggishly because of their small diffusion coefficients. Using chemical methods to polymerize liquid precursors into a continuous porous mass of coalesced particles, two sets of parameters can be controlled simultaneously. The nature of the material, porosity, and other properties that affect separations can be optimized. The size of channels and open spaces can be controlled independently. Monoliths for chromatographic use can be described as spongy with micrometer-sized channels winding through a mass of fused particles [3]. Monoliths are based on either silica or organic polymers. Silica materials can be prepared by a sol-gel synthesis. The reaction is based on hydrolysis and polycondensation of tetramethoxysilane in the presence of polyethleneglycol, urea, and other reagents. Heating the product yields an alkaline solution, which etches tiny holes into the silica walls. One of the main concerns is ensuring that the monolithic material remains in intimate contact with the walls of the surrounding tube. One strategy for ensuring the requisite contact is to bond the material covalently to bond the growing monolith to the walls of a quartz tube via SiO2 groups. Much of the development work on application of monoliths to IC separations has been performed on reversed-phase monolithic HPLC columns, which became available well before monoliths with ion-exchange sites. These HPLC columns can be adapted for use in ion chromatography by coating with an ionic surfactant. Monolith anion- and cation-exchange columns are particularly advantageous for separation of bio ions, although small anions and cations can also be separated. Effective anion columns with high efficiency can be prepared simply by coating a nonionic column with a cationic surfactant, such as cetylpyridinium chloride (CPC) [4].
3.4 Anion Exchangers
Paull and co-workers [5, 6] used a very short (10 mm × 4 mm) coated silica monolith for the separation of inorganic anions. The back pressure was sufficiently low that a peristaltic pump could be used to pass eluent through the system. Pelletier and Lucy [7] were able to achieve rapid low-pressure chromatographic separations of anions on short monolithic columns. Reversed-phase silica columns 0.5 and 1.0 cm long were coated with a long-chain cationic surfactant to convert the column to an anion exchanger. Excellent separations of seven inorganic anions were obtained in 1–2 min. The thickness of the surfactant coating, and thus the exchange capacity of the column, can be adjusted by varying the percentage of acetonitrile in the predominately aqueous coating solution from 1 to 5%.
3.4 Anion Exchangers
The most widely used anion-exchange resins may be divided into two general types. Porous polymeric materials have quaternary ammonium functional groups throughout the resin bead although the concentration of exchange sites deeper within the bead tends to diminish. Latex agglomerated anion exchangers are pellicular materials with the exchange sites on latex particles coated in a relatively thin layer at the outer perimeter of the bead. 3.4.1 Porous Anion Exchangers
The major anion exchangers for ion chromatography are based on two substrate types: macroporous and microporous (or gel type) materials. Microporous resins were formerly popular because of their superior ion-exchange kinetics. However, microporous substrates can be ‘spongy'. Columns packed with this material may eventually undergo bed compression leading to reduced column performance. Macroporous materials are rugged, and column beds made from this substrate are stable. However, some commercial materials show poor ion-exchange kinetics with certain eluents. Gjerde [8] described a macroporous anion-exchange resin that shows good IC separations with a variety of eluents including sodium carbonate/bicarbonate and sodium hydroxide. This resin, called the Transgenomic AN1, is highly cross-linked Polystyrene Divinylbenzene with 80 Å pores and 415 m2 g–1 surface area. The commercial material has an exchange capacity of 0.05 mequiv g–1 and an average particle size of 8 mm. The resin contains a quaternary ammonium functional group: dimethylethanolamine. The substrate polymer beads used in this work are quite hard and do not swell and shrink when the solvent is changed. A measure of bead hardness is given by the swelling propensity value. This value is obtained by measuring the backpressure of a column with a tetrahydrofuran mobile phase and then with an aqueous
45
46
3 Resins and Columns
phase. After correcting the eluent backpressure for viscosity, the swelling propensity, SP, is calculated by the following equation: SP
THFpressure H2 Opressure H2 Opressure
An SP value of zero indicates nonswelling material. The substrate used in this work had an SP of 0.8. Many PS-DVB resins have SP values well above 1.0. Columns packed with the AN1 resin have given excellent separations of common anions using a variety of eluents. Polymeric resins from Hamilton have been used extensively for anion chromatography. Their PS-DVB anex resins contain trimethylammonium groups. PRP X100 and PRP X110 have exchange capacities of 0.19 and 0.11 mequiv g–1 respectively. Their RCX10 has a somewhat higher capacity: 0.35 mequiv g–1. All of these have an average pore size of 100 Å. Good separations of anions have been obtained with a variety of eluents [9]. However, these resins are quite hydrophobic. For this reason, the eluents often contain ∼7.5% methanol or ∼0.1 mM sodium thiocyanate to give better peak shapes. A resin from Alltech, sold commercially as Durasep A1, is another example of a PS-DVB anion exchanger [10]. A highly cross-linked backbone makes this material chemically and mechanically stable. It withstands organic solvents and is stable over a wide range of pH, temperature and pressure. The resin is made of polydivinylbenzene with dimethylethanolamine functional groups and was designed for use with both nonsuppressed and suppressed conductivity detectors. The material is highly cross-linked and has both micro and macro pores for efficient mass transport. It was shown that resolution and peak shape are improved by adding 5–15% methanol to the eluent. In particular, methanol reduced or eliminated the hydrophobic interaction between nitrate and the resin, resulting in a symmetrical peak shape. However, addition of acetonitrile to the eluent seemed to increase the hydrophobicity of the column toward nitrate. A wide variety of resins based on polyacrylate polymers has been produced for use in chromatography. A type known as HEMA, a macroporous copolymer of 2-hydroxyethyl methylmethacrylate and ethylene dimethacrylate, has been used extensively in ion chromatography. It is extensively cross-linked to produce a polymeric matrix with high chemical and physical stability. The structure of HEMA is shown in Figure 3.2A. The tertiary carbonyl structure of pivalic acid is one of the most stable and least hydrolyzable esters known, which allows the HEMA stationary phase to be used with a variety of eluents in the pH range 2–12. The excess hydroxyl groups on the HEMA matrix also increase the hydrophilicity of this material, which will be shown later to result in improved peak shapes for polarizable anions. The strong-base anion exchanger of HEMA, shown in Figure 3.2B, is prepared by treating the HEMA precursor with an aqueous solution of trimethylamine. The preparation procedures and the influence of different functional groups on sorbent selectivity were discussed by Vlacil and Vins [11].
3.4 Anion Exchangers
Figure 3.2 Structures of (A) HEMA and (B) strong-base anion exchanger of HEMA.
The preparation and properties of polyacrylate anion-exchange resins by Alltech has been described as a universal stationary phase for the separation of a wide variety of anions [6, 7]. The Allsep anion column contains 7-lm particles packed into columns of various lengths. The resin is methacrylate based with quaternary ammonium functional groups The A2 anion column also has methacrylate, but has a quaternary amine with alkenol, rather than alkyl, groups. It is relatively hydrophilic and resolves acetate and formate from fluoride and chloride. Both columns have a broad pH range (pH 2–11) and can be used with 0–100% of an organic eluent modifier. 3.4.2 Effect of Functional Group on Selectivity
The polymeric matrix, the chemical type and structure of the ion-exchange groups, and solvation effects all have a significant effect on determining the selectivity of anion exchangers for competing ions [12]. Variations in the chemical structure of exchange sites have been particularly effective in producing resins with greater selectivity for certain ions.
47
48
3 Resins and Columns
Okada [13] compared the effects of –NH3+ and -NR3+ groups in anion exchangers with the same polymeric matrix and almost the same exchange capacity (0.4 mmol g–1). The –NH3+ resin had a more concentrated charge and a stronger electrostatic field than the resin with –NR3+, where the positive charge was more dispersed. Going from –NH3+ to –NR3+ resulted in decreased electrostatic and hydrogen bonding interactions with sample anions and increased ion-induced dipole and London dispersion interactions. The latter two effects resulted in preferable binding of larger ions by –NR3+. As an example, the ratio of retention factors of ClO4–:Cl– was 17.4 for the resins with –NR3+ but only 1.89 for those with –NH3+ groups. Anion-exchange resins containing a benzyltrimethylammonium functional group are a widely used type for anion chromatography. It was of interest to see how changing the chemical nature of the quaternary ammonium functional group might affect the selectivity of anion exchangers toward different anions. Barron and Fritz [14] prepared 13 different resins by reacting chloromethylated XAD-1 with different tertiary amines. To ascertain the effect of functional group structure on selectivity, the various resins should have a very similar exchange capacity so that identical elution conditions could be used for each resin. To accomplish this, the relative reactivities of various amines with chloromethylated XAD-1 had to be determined. Then the degree of chloromethylation of XAD-1 could be adjusted so that the aminated resins would have similar capacities. The retention times of 17 monovalent anions on resins with different functional groups but with almost identical exchange capacities (average: 0.027 mequiv g–1) were compared with the use of a solution of a monovalent anion (sodium benzoate) as the eluent [14]. Relative retention times were calculated by dividing the measured retention times by that of chloride. Data for resins with various trialkylammonium groups are presented in Table 3.1. The data show that the relative retentions of the weak-acid anions are almost independent of the size of the alkyl groups. However, as the size of the R groups increases, large changes occur with the more polarizable anions such as nitrate, iodide, chlorate and BF4– ions. The use of anion-exchange resins with different substituents offers a useful parameter for improving the separation of some anions. For example, Figure 3.3 compares the separation of five anions on columns containing resins of approximately the same capacity but with increasingly larger alkyl groups on the quaternary nitrogen. Identical elution conditions were used. The TMA resin column gave poor resolution of bromide and nitrate. The resolution was improved on the TPA column, and a baseline separation was obtained on the THA column. Most of the monovalent anions exhibit only small changes in their relative retention on resins containing one, two, or three hydroxyethyl groups, compared to the trimethylamine resins (Table 3.1). However, when a stronger eluent was used (phthalate, with two negative charges instead of benzoate, with a single negative charge) and several divalent anions were examined, the changes resulting from hydroxyethyl groups became more apparent [15].
3.4 Anion Exchangers Table 3.1 Relative retentions of anions on trialkylammonium
resins [24]. Anion
TMA
TEtA
TPA
TBA
THA
TOA
Cl–
1.0
1.0
1.0
1.0
1.0
1.0
0.66
0.70
0.69
0.71
0.68
0.69
Br
1.20
1.19
1.25
1.34
1.32
1.41
–
2.51
2.48
3.05
3.82
5.00
>5.0
H2PO4–
0.84
0.85
0.84
0.85
0.83
0.83
NO2–
0.82
0.82
0.86
0.90
0.89
0.98
–
1.30
1.32
1.38
1.54
1.63
1.72
Acetate
0.25
0.28
0.23
0.25
0.22
0.22
Formate
0.52
0.54
0.51
0.52
0.51
0.53
Lactate
0.44
0.49
0.46
0.47
0.45
0.49
Glycolate
0.45
0.48
0.44
0.45
0.43
0.45
Nicotinate
0.31
0.32
0.31
0.31
0.31
0.33
–
1.53
1.55
1.56
1.73
1.92
2.15
–
1.05
1.06
1.03
1.08
1.08
1.14
0.36
0.39
0.34
0.37
0.35
0.37
2.70
2.58
3.41
4.34
>7.5
–
1.00
1.06
1.01
1.06
1.01
1.07
8.3
8.0
9.5
8.9
–
F
–
I
NO3
ClO3
BrO3 N3– BF4
–
CH3SO3
–
tRCl(min)
10.6
9.4
TMA = trimethylamine, TEtA = triethylamine, TPA = tripropylamine, TBA = tributylamine, THA = trihexylamine, and TOA = trioctylamine.
The data in Table 3.2 show that methyldiethanolamine (MDEA) considerably lowers the relative retention of nitrate, chlorate, iodide and thiocyanate compared to the trimethylamine (TMA) resin. It also shows that with the phthalate eluent the tributyl resin (TBA) has longer retention times for nitrate and much longer for iodide, but shorter for sulfate and thiosulfate, all compared to the TMA material.
49
50
3 Resins and Columns
Figure 3.3 Separation of five anions on three different resins of similar capacity. The resins were packed in a 500 mm × 2.0 mm i.d. column and a solution of 0.0001 M benzoic acid was used as the eluent at a flow rate of 0.93 mL min–1 (From Ref. [14], with permission).
Virtually all anion-exchange separations have been carried out with resins containing a single nitrogen atom in each exchange group, either a quaternary ammonium group (– N+R3) or a protonated amine group (– N+HR2). A novel resin has been described, containing three nitrogen atoms in each functional group [16]. A chloromethylated PS-DVB resin was reacted with diethylenetriamine to give a functional group of structure a, or b, or a mixture of the two (see below).
3.4 Anion Exchangers Table 3.2 Relative retentions of ions on three anion exchangers
of differing polarity. Phthalate eluent (0.4 mM), pH 5.0. Ion
MDEA*
TMA**
TBA***
Cl–
1.0
1.0
1.0
NO3–
1.47
1.79
2.77
ClO3–
1.78
2.42
3.18
3.81
6.16
I
–
13.9
–
7.50
14.5
–
–
9.12
–
–
SO42–
7.31
S2O32–
15.75
C2O42–
7.41
7.18
6.30
9.50
9.92
8.72
SCN
ClO4
MoO4 WO4
2–
2–
tRCI- (min)
– 3.20
7.29 16.6
– 3.80
6.36 9.23
– 3.90
Values are expressed at tRanion/ tRCI* Capacity 0.090 mequiv g–1 ** Capacity 0.092 mequiv g–1 *** Capacity 0.096 mequiv g–1
By varying the pH at which the resin is used in an IC column, one, two or three of the N atoms can be protonated giving a net charge of 1+, 2+ or 3+ for each functional group. The retention times of sample anions become longer as the operating pH becomes more acidic and the net positive charge on the ion exchanger increases. Figure 3.4 plots the retention factor as a function of eluent pH for several sample anions. Thiocyanate and molybdate are very strongly retained, even at moderately acidic pH values. Several anions were separated chromatographically at pH 7.5–7.7 with different salts in the mobile phase and with direct UV detection at 200 nm. The results in Table 3.3 show that perchlorate is a significantly better eluting anion than chloride. However, sulfate and hydrogen phosphate both give even shorter retention times by virtue of their 2– charge.
51
52
3 Resins and Columns
Figure 3.4 Eluent pH vs. anion capacity factor, k′. Conditions: 100 mm × 4.6 mm (multicharge, weak base, anion exchanger) column, 15 mM sodium perchlorate eluent, UV detection at 200 nm (from Ref. [16], with permission). Table 3.3 Retention times (min) of several anions with different eluents.
Anion
NaClO4a
NaSO4a
NaCla
Na2HPO4a
Bromide
3.45
2.15
4.51
1.98
Nitrate
3.98
2.75
5.55
2.54
Iodide
6.25
7.01
12.4
5.95
Thiocyanate
13.7
NDb
25.6
17.4
a Each eluent: 5.0 mM at pH 7.5–7.7. b ND = not detected.
3.4.3 Effect of Spacer Arm Length
Polymeric anion exchangers are normally prepared by chloromethylation of the benzene ring followed by reaction with a tertiary amine to give a quaternary ammonium group. Thus the N+ is connected to the benzene ring by a single –
3.4 Anion Exchangers
CH2 group. Suppose the N+ was connected to the benzene ring by a longer series of –CH2 groups, sometimes called the spacer arm. What effect would a longer spacer arm have on ion-exchange selectivity? Warth and Fritz synthesized a series of resins with spacer arms of varying lengths [17]. The benzene ring of a macroporous resin (Rohm & Haas XAD-1) was reacted under controlled conditions with a bromoalkene with CF3SO3H as catalyst and was then quaternized by reaction with trimethylamine. This gave a resin of the following structure:
Conditions were adjusted so that all of the resins had almost identical exchange capacities. The spacer arm length varied from one to six methylene groups. Chromatographic retention times of a number of anions were compared, with 6.0 mM nicotinic acid and 2.0 mM phthalate (pH 6) as the eluents. In many cases, the length of the spacer arm had very little effect on the relative retention times (relative to Cl–). However, the selected data in Table 3.4 show that the relative retention times of bromide, nitrate, chlorate and iodide decreased, while that of sulfate increased slightly.
Table 3.4 Adjusted retention times as a function of spacer arm
length of anion exchange resins. Spacer arm length Anion
C1
C2
C3
C4
C6
Chloride
1.0
1.0
1.0
1.0
1.0
Nitrite
1.8
1.5
1.5
1.4
1.3
Bromide
2.5
2.1
2.0
1.8
1.9
Nitrate
3.6
3.1
2.5
2.4
2.6
Chlorate
7.1
4.5
3.9
3.7
4.1
Sulfate
7.4
8.3
8.5
8.5
8.5
Iodide
17.6
12.2
8.9
8.6
11.2
Longer spacer arms would of course reduce any influence the benzene ring might have on ion-exchange retention.
53
54
3 Resins and Columns
3.4.4 Latex Agglomerated Ion Exchangers
Pellicular materials in which the stationary phase is a layer on the outside perimeter of a spherical substrate have been used frequently in liquid chromatography. The relatively thin layer ensures a rapid equilibrium between the mobile and stationary phases even when the column packing has a relatively large particle size. In their original work on ion chromatography, Small et al. [18] used a surface-sulfonated material as a pellicular cation exchanger. Such materials are easy to prepare because sulfonation of a polymer containing benzene rings proceeds from the outside in. Sulfonation for a short period under mild conditions will insert sulfonic acid groups only on or near the outside surface of a spherical resin. It was soon discovered that efficient anion-exchange resins could be prepared by coating the outside of a surface-sulfonated polymer with a layer of latex particles functionalized with quaternary ammonium groups. The first commercial anion exchangers for ion chromatography (Dionex ASI) consisted of 0.15-lm latex particles coated onto a 25-lm sulfonated substrate. A two-dimensional diagram of this coating is shown in Figure 3.5A. The positively charged latex particles are firmly held by electrostatic attraction as shown in Figure 3.5B. Each latex particle has several quaternary N+ groups, so the coated substrate will have many quaternary groups available for ion exchange. Latex agglomerated resins are very stable chemically. Even 4 M sodium hydroxide is unable to cleave the ionic bond between the substrate and the latex bead.
Figure 3.5 Latex-coated anion exchanger (Courtesy Dionex Corp.)
Several advantages have been claimed for latex-coated anion exchangers [19]: . The substrate provides mechanical stability and gives a moderate back pressure. . The small size of the latex beads and their location on the outer surface of the substrate ensure fast exchange processes and thus a high chromatographic efficiency. . Swelling and shrinkage are minimal.
3.4 Anion Exchangers
The properties of latex resins can be varied by manipulation of several parameters. Hydrophobic attraction of the exchanger for some anions can be altered by varying the type and cross-linking of the polymeric substrate. The ion-exchange capacity is determined by the substrate particle size, the size of the latex beads, and the degree of latex coverage on the substrate surface. Selectivity for various anions is governed mainly by the type of functional groups attached to the latex bead and by the degree of latex cross-linking. Over the years, Dionex has developed a wide variety of latex-agglomerated resins to meet various needs in IC. A review of these developments is given in a book by Weiss [19]. A method has been described for preparation of latex-coated anion-exchange resins that does not involve sulfonation of the substrate. A suspension of quarternized latex beads in water containing 0.01 to 0.10 M sodium chloride is used to coat an unfunctionalized polymeric substrate. Once coated, the latex sticks tightly and is not washed off by aqueous solutions. The latex can be removed by washing with pure organic solvents and may then be recoated. The exchange capacity may be varied by changing the concentration of sodium chloride or latex in the coating solution. Various resin substrates were coated with exchange capacities ranging from 5 to 400 mequiv g–1. Columns packed with these latex exchangers gave unusually efficient separations of sample anions. Extensive research effort has gone into the development of resins and columns for IC. This is perhaps best illustrated by a specific example: a stationary phase for the determination of fluoride and oxyhalides such as chlorite, chlorate and bromate [20]. The determination of fluoride has been a problem owing to its low affinity for strongly basic anion exchangers. Although carbonate–hydrogen carbonate eluents are widely used, fluoride elutes very close to the system void and its detection by suppressed conductivity at lower concentration levels is very difficult owing to interference from the negative water dip. The water dip occurs when the injected aqueous sample passes through the conductivity cell, decreasing the background conductivity. In addressing this problem, an anion-exchange column with high exchange capacity gave good resolution of fluoride and early-eluting anions when used with a dilute eluent, but the elution times for bromide and nitrate exceeded 40 min. A commercially available polymeric quaternary ammonium (quat)-coated column gave excellent resolution of fluoride and other common inorganic anions, but a rinse with acetonitrile-water (90:10) destroyed the separation ability of the column. A periodic rinse with an organic solvent is often needed to remove humic acid or other organic matter that gradually builds up on a column. The goal of the research undertaken was to develop a new solvent-compatible ion-exchange material with which fluoride is resolved from the system void volume and oxyhalides are separated from other acid anions in the same run under isocratic conditions. Because of the high degree of hydration of the fluoride ion, it was necessary to generate a solid phase with an extremely high water content. This suggested the use of very hydrophilic ion-exchange sites. Because fluoride is so highly hydrated, the only viable method of obtaining a reasonable fluoride
55
56
3 Resins and Columns
retention was to produce a latex coating with extremely low cross-linking. A crosslinking level of well under 17 leads to the generation of polymers with very high water content. The use of a functionalized latex with extremely low cross-linking creates a potential problem with regard to the exchange capacity of the coated resin. The surface area of standard microporous particles is inadequate to provide sufficient capacity when used with these low-cross-linked latexes. A resin with an acceptable capacity was finally produced by coating a super-porous polymeric substrate. The structural and physical properties of the final product (with the commercial designation ion Pac AS 12A) are given in Table 3.5.
Table 3.5 Structural and physical properties of the Ion Pac
AS12A separator. Parameter
Value
Column dimensions
200 mm × 4 mm i.d.
Particle diameter
9 lm
Substrate material
Macroporous polyethylvinylbenzene cross-linked with 55% divinylbenzene
Pore size
200 nm
Column capacity
52 l equiv.
Latex polymer
Vinylbenzyl chloride
Latex cross-linking
Very low (0.15%)
Latex diameter
14 nm
Functional group
quaternary ammonium group
pH stability
0–14
Solvent compatibility
0–100%
Two important conclusions may be drawn from these tables. One is that the hydroxide ion is a much weaker eluent for anion chromatography than carbonate. The second is that the eluting power of sodium hydroxide is enhanced considerably by using latexes with one or two hydroxyethyl groups instead of those containing only alkyl groups.
3.4.4.1 Effect of Latex Functional Group on Selectivity Slingsby and Pohl [21] investigated the effect of varying the structure of the quaternary ammonium group on the latex while keeping the percentage of cross-linking and the polymeric backbone structure constant. They estimated that each
3.5 Cation Exchangers
5-mm spherical substrate particle was coated with approximately 28 000 quaternized latex beads. Retention factors (k), corrected for column capacity, were measured for four different columns. The latex functional group in column 1 was methyldiethanolamine (MDEA), column 2 was dimethylethanolamine (DMEA), column 3 was trimethylamine (TMA), and column 4 was triethylamine (TEA). The eluent in Table 3.6 was 5 mM sodium carbonate.
Table 3.6 Retention factors (k) for different latex functional groups.
Eluent: 5 mM sodium carbonate. Column
F–
Cl–
MDEA
0.06
0.24
DMEA
0.14
TMA TEA
Br–
NO3–
ClO3–
SO42–
HPO42–
0.92
1.1
1.0
0.20
0.31
1.1
4.5
5.0
4.9
3.0
6.7
0.30
4.4
19.2
22.5
21.2
51.4
>100
0.30
5.8
26.1
55.8
24.0
24.9
>100
3.5 Cation Exchangers
The science and technology of ion chromatography is continuously evolving. Certain trends become apparent which are often pushed aside after a few years in favor of a different trend. This is particularly true in the case of resins for cation chromatography. Some of the milestones in the development of cation exchangers for IC will be traced in this section. 3.5.1 Sulfonated Resins
Most of the cation exchangers used in IC fall into two major categories: sulfonated resins, sometimes called ‘strong-acid’ exchangers, and resins with carboxylic acid groups, sometimes called ‘weak-acid’ exchangers. The ion exchangers used in IC have a much lower exchange capacity than those intended for commercial applications such as the removal of calcium and magnesium ions from hard water. Low-capacity cation-exchange resins are obtained by superficial sulfonation of styrene-divinylbenzene copolymer beads. The resin beads are treated with concentrated sulfuric acid and a thin layer of sulfonic acid groups is formed on the surface. The final capacity of the resin is related to the thickness of the layer and is dependent on the type of resin, the bead diameter, and the temperature and time of contact with the sulfuric acid. Typical capacities range from 0.005 to 0.1 mequiv g–1.
57
58
3 Resins and Columns
It can be easily appreciated that, compared to a conventional cation-exchange resin, the diffusion path length is reduced because the unreacted, hydrophobic resin core restricts analyte cations to the resin surface. This results in faster mass transfer of the cations and consequently in improved separations. Also, because of the rigidity of the resin core, there is less tendency for the bead to compress. This means that higher flow rates (at relatively low back pressures) can be used than would be possible with conventional resins. Superficially functionalized resins are stable over the pH range of 1 to 14 and swelling problems are minimal. The selectivity of the superficial cation-exchange resins for ions is similar to that observed for conventional resins. For chromatographic purposes it is necessary to control the conditions for sulfonation so that the polymer beads are sulfonated evenly and the exchange capacity is very reproducible. It is also desirable to be able to increase or decrease the resin capacity in a predictable fashion. Pepper [22] proposed that the sulfonation of neutral poly(styrene-divinylbenzene) copolymer beads proceeded in a ‘layer-by-layer’ fashion and could be stopped at any particular depth. From this concept, Pepper [23] prepared the first superficially sulfonated polymer beads. Parrish [24] also prepared surface-sulfonated beads and gave a brief example of their utility. Fritz and Story [25, 26] prepared several low-capacity sulfonated resins for the chromatographic separation of various metal cations. Macroporous resins were used and their selectivity was somewhat different from that of conventional gel resins. The surface sulfonation of poly(styrene–divinylbenzene) beads with cross-linking ranging from 0.5 to 8% divinylbenzene was studied by Small [27, 28]. Examination of the sulfonation depth of a resin bead in terms of optimum separation of several inorganic cations was subsequently studied by Stevens and Small [29]. These resins were used for the chromatographic separation of simple cations. Low-capacity resins have been prepared specifically for use in ion chromatography by Fritz et al. [30]. The resin used in the separations was a 3:2 blend of neutral resin with low-capacity sulfonated resin giving a final capacity of about 8 lequiv g–1. Very good efficiency was demonstrated for separation of the alkali cations and alkaline earths. Papanu et al. [31] fabricated a cation-exchange resin by agglomerating a sulfonated latex onto larger beads of a low-capacity anion-exchange resin. Battaerd [32, 33] prepared a superficially sulfonated bead by graft polymerization of a sulfonated olefin onto a polyolefin core. Kirkland [34] impregnated a porous fluoropolymer bead with a sulfonated fluoropolymer, giving a pellicular type of ion exchanger. Horvath et al. [35] polymerized a coating of polystyrene-DVB onto glass beads and formed a cation exchanger by sulfonation. Several authors have prepared cation exchangers by introduction of sulfonated organic group onto the surface of silica supports [36–38]. Sevenich and Fritz [39] examined the sulfonation of resins for use in IC in some detail. Spherical microporous resin beads of 4%, 6.5% and 12% cross-linking were selected for their study. Initially, sulfonation of 4% cross-linked resin
3.5 Cation Exchangers
beads under identical conditions gave poor reproducibility. This was partly due to agglomeration of the small resin beads. Dispersion of the resin beads and even initial wetting of the surface by sulfuric acid seemed to be major problems in achieving reproducible sulfonation. The procedure finally developed involved mechanical and ultrasonic dispersion of the particles in methanol, passing the resin slurry through a small sieve, then removal of as much methanol as possible by suction filtration. The resin was then sulfonated. The small amount of methanol present when the resin is added to the hot sulfuric acid immediately volatilizes and is swept out in the inert gas stream. This is confirmed by a short burst of vapor issuing from the reaction vessel when the resin is added. The resin capacity is approximately linear with reaction time and the reproducibility was ±5%. Sulfonation of gel beads of 4%, 6.5% and 12% cross-linking was compared. Data for sulfonation of each for 30, 60 and 90 min are given in Table 3.7. These results show that resins of low capacity can be obtained in all cases but that the reaction is better controlled with resin beads of higher cross-linking.
Table 3.7 Retention factors (k) for different latex functional groups.
Eluent: 5 mM sodium carbonate. Column
F–
Cl–
MDEA
0.06
0.24
DMEA
0.14
TMA TEA
Br–
NO3–
ClO3–
O42–
HPO42–
0.92
1.1
1.0
0.20
0.31
1.1
4.5
5.0
4.9
3.0
6.7
0.30
4.4
19.2
22.5
21.2
51.4
>100
0.30
5.8
26.1
55.8
24.0
24.9
>100
The location of sulfonate groups in the resin bead was visualized by completely replacing H+ with UO22+ as the counterion and obtaining transmission electron micrographs on thin slices of the resin bead. Because uranium is a very heavy metal, the uranyl ions have higher stopping power for electrons and appear as a darker area on the micrograph. A dark outline around the resin slice indicated that the uranyl ions (and hence the sulfonate groups) are located in a thin zone (approximately 200 Å) at the outer perimeter of the resin bead. Knowing the resin capacity and the estimated thickness of the sulfonated layer, a simple calculation shows nearly complete sulfonation within the sulfonated layer. By ‘complete’ we mean that there is approximately one sulfonate group for each benzene ring of the polymer. The density of sulfonate groups in this layer is similar to that in the entire bead of a typical high-capacity cation-exchange resin. Columns packed with the 12% cross-linked resin (6.1 mequiv g–1 exchange capacity) gave good separations of metal cations. Using 8 different concentrations of perchloric acid eluents, from 0.10 M to 1.00 M, retention times for 36 metal ions
59
60
3 Resins and Columns
were measured [40]. The selectivity data obtained are given in Tables 5.3 and 5.4 in Chapter 5. The Dionex Co. has developed a number of latex-coated cation-exchange materials. Their anion-exchange resins have a surface layer of quaternary ammonium latex on a surface-sulfonated substrate. Addition of a second coating layer of sulfonated latex beads provides an outer layer of resin consisting of latex beads with exposed sulfonate groups. These groups undergo cation exchange with sample and eluent cations in IC. A schematic representation for these cation exchangers is: Resin–SO3– N+R3 latex N+R3 SO3–-latex- SO3–
(3.7)
A diagram of such a resin (Ion Pac CS3) indicates that the sulfonated latex beads are of larger diameter than the quaternized beads (Figure 3.6). This configuration of the column packing permitted the use of smaller size substrate beads, provided a shorter mean free path for the analytes and greatly improved peak efficiencies for cationic analytes. Although Group I and Group II cations could be separated simultaneously, the elution time was long (about 28 min) unless some form of gradient elution was employed.
Figure 3.6 Schematic representation of IonPac CS3, a latexcoated strong-acid cation exchanger (Courtesy Dionex Corp.)
A few years later (1990), a sulfonated cation-exchange column (Ion Pac CS10) with decreased cation-exchange site density was developed. This lower exchange capacity was obtained by adding a monomer that is deactivated by sulfonation to the latex polymerization mixture. The substrate of the column packing was 50% cross-linked, instead of 4% as with previous materials, in order to obtain better compatibility with organic solvents. Six common inorganic cations could be separated isocratically in about 15 min [41].
3.5 Cation Exchangers
Several improved versions of sulfonated latex columns were developed over the next few years [41], but this type of cation-exchange column was abandoned in favor of materials with a weak-acid function. Catex columns with sulfonic acid functionalities have a relatively low selectivity for hydronium ions. A divalent cation component must be added to the eluent to efficiently elute divalent cations such as magnesium and calcium. 3.5.2 Weak-acid Cation Exchangers
Owing to their differences in selectivity, it is often difficult to find conditions for separation of cations of different positive charge on a sulfonated resin column. Eluents that provide good separation of monovalent cations are too weak to elute divalent cations in a reasonable time. There is now a trend to use weak-acid cation-exchange columns. These materials contain carboxylic acid functional groups, or in some cases mixed carboxylic acid and phosphonic acid groups. At more acidic pH values these groups are gradually converted from the ionic to the molecular form, and thus their ability to retain sample cations is diminished. By adjusting the operating pH to an appropriate value it becomes possible to separate a wider variety of cations in a single run. One company advertises their chromatographic columns of this general type as a Universal Cation Column. The properties and performance of a commercial weak-acid resin column (Dionex CS12) have been described [42]. The substrate is a highly cross-linked, macroporous ethylvinylbenzene–divinylbenzene polymer with a bead diameter of 8 lm, a pore size of 6 nm, and a specific surface area of 300 m2 g–1. In a second step, this substrate was grafted with another polymer containing carboxylate groups. The exchange capacity is listed as 2.8 mequiv per column for a 250 mm × 4 mm i.d. column. With this column, simple eluents such as hydrochloric or methanesulfonic acid can be used to separate mono- and divalent cations rapidly and efficiently under isocratic conditions. Morris and Fritz [42] described the preparation and chromatographic applications of two weak-acid resins that are easily synthesized and carry the exchange group on the cross-linking benzene ring of the resin or on a short spacer arm from the ring. The first resin (resin I) was prepared by reaction of a cross-linked polystyrene resin with succinic anhydride in a Friedel–Crafts reaction with aluminum chloride as the catalyst. The carboxyl groups are connected to the resin benzene rings by a three-carbon atom spacer arm, thus: –COCH2CH2CO2H. The second cation exchanger (resin II) was prepared by reaction of the resin with phenylchloroformate to give a phenyl ester attached to the resin benzene rings, thus: – CO–OC6H5. The ester groups were then hydrolyzed by refluxing for 1 h in a sodium hydroxide–ethanol solution to give the sodium salts of the carboxylate. The exchange capacity of resin I was 0.60 mequiv g–1 and that of resin II was 0.39 mequiv g–1. Resin II in particular gave excellent separations of divalent metal cations with a complexing eluent.
61
62
3 Resins and Columns
Most of the cation-exchange resins used today have a polymeric matrix and contain carboxylate, or both carboxylate and phosphonate functional groups. Carboxylic acid phases are used in a weakly ionized form with only a fraction of the ion-exchange sites actually available for retention of cations. Because of this, the overall functional capacity must be increased in order to get sufficient retention of inorganic cations and amines with commonly used eluent systems. In producing the Dionex CS 12A cation exchanger [41], a higher capacity was obtained by using a highly crosslinked, macroporous, spherical substrate with a high surface area. In this particular case, the substrate particle consisted of a 55% crosslinked ethylvinylbenzene–divinylbenzene core with a surface area in the region of 450 m2 g–1. A polymeric film containing a mixture of carboxylic acid and phosphonic groups was then applied to a thickness of 5–10 nm over the entire bead surface as illustrated in Figure 3.6. Because of the porous nature of the substrate, the applied coating also penetrates the bead so that exchange groups are distributed throughout the resin particle. The average particle size of the final ion exchanger was 8 lm and the pore size was 150 Å. Selectivity in cation chromatography involves more than just an electrostatic attraction of analyte cations for a negative charge at exchange sites. The entire carboxylate functional group (–CO2–) or phosphonate group (–PO32– or –PO3H–) affects cation selectivity. Selectivity can therefore be altered by introduction of phosphonic acid groups into carboxylic acid ion exchangers in varying ratios. An example is shown in Figure 3.7. The resin contained a high ratio of phosphonic acid to carboxylic acid groups. The manganese peak has moved out beyond the calcium peak in going from a 1:5 ratio to a 1:1 ratio of phosphonate to carboxylate groups. Although not shown, calcium and strontium co-elute with the 1:1 phase, which is impractical because significant amounts of strontium are commonly found in groundwaters. The final product in the CS 12A exchanger actually contains a slightly lower amount of phosphonate than the upper chromatogram in Figure 3.7.
Figure 3.7 Schematic of the Ion Pac CS12A, a weak-acid cation exchanger (Courtesy Dionex Corp.)
3.6 Other Resins
3.5.3 Other Types
Considerable interest has been shown in a novel cation exchanger first developed by Schomburg et al. [43]. The material consists of a silica substrate of very uniform particle size coated with a poly(butadiene–maleic acid) copolymer which serves as the cation-exchange moiety. Maleic acid has two acidic dissociation constants (pK1 = 2.0, pK2 = 6.3), and this retains cation-exchange properties down to a fairly low pH. Analyte cations may be eluted with acidic or complexing eluents using commercially available columns [44, 45]. As an example, lithiuim, sodium, ammonium, potassium, magnesium and calcium are readily separated in a single run with 3 mM methanesulfonic acid as the mobile phase. Eluents containing a high concentration of common organic solvents can be tolerated [44]. Macrocyclic groups have been incorporated into resins to impart a different selectivity for metal cations than is possible with ordinary cation-exchange resins. A chemical structure known as 18-crown-6 has often been used. This a doughnutlike ring with six oxygen atoms connected together in via ethylene groups to form a ring with a cavity in the center. These ligands provide novel IC separations owing to the unusual specificity with which they bind cations of various ionic radii to form charged complexes. For example, the potassium ion fits into the cavity to form a stronger complex with the macrocycle oxygen atoms than does the sodium or ammonium ion. Potassium(I) elutes later from the cation-exchange column and is easily separated from sodium or ammonium ions. Application of macrocyclic ligands to ion chromatography has been discussed [46]. A practical column was created by absorbing tetradecyl-18-crown-6, which is a macrocycle with a long hydrophobic tail, onto a polystyrene-divinylbenzene substrate. When packed into a column, the macrocycle remains adsorbed to the resin [46]. The log K binding constants to 18-crown-6 in water for monovalent inorganic cations are as follows: Li+ = 0, Na+ = 0.8, Cs+ = 0.99, NH4+ = 1.23, Rb+ = 1.56, K+ = 2.03. Exactly the same elution order was observed for these same ions with the TD18C6 column [47].
3.6 Other Resins 3.6.1 Chelating Ion-exchange Resins
The selectivity of ordinary cation-exchange resins for various metal ions is somewhat limited. However, if a suitable chelating functional group is built into a polymeric resin, it often is possible to take up only a small group of metal ions. Other chelating resins may complex a larger group of metal ions, but selectivity is attained through pH control. Chelating resins also are valuable in sorbing a
63
64
3 Resins and Columns
desired metal ion (or small group of metal ions) from solutions containing a very high concentration of a noncomplexed metal salt. Frequently the selectivity of a chelating resin is so great that a very short column can be used to retain the desired metals. Although a great many chelating resins have been described, only a few have sufficiently fast rates of metal ion equilibrium and chromatographic efficiency for practical separations. Several chelating functional groups for resins are listed in Table 7.8. Chromatographic separations using chelating resins are discussed in Section 7.5. 3.6.2 Metal Oxides
Pietrzyk and coworkers [48] showed that hydrated alumina can function as a lowcapacity anion exchanger. The isoelectric pH for alumina is about 7.5 and its ionexchange capacity (about 2 mequiv g–1 maximum) is dependent on the sample ion, the eluent pH, and the pretreatment of the alumina. The alumina must be first hydrated and then treated with either an acid or base. Anion-exchange capacity increases as the eluent pH is made more acidic. The changeover from an anion exchanger to a cation exchanger is gradual and occurs in the vicinity of eluent pH 7.5. In anion chromatography the elution order of common anions is almost the reverse of conventional ion exchangers. Thus, using a Spherisorb A54 column with a 0.10 N acetate buffer at pH 6.50, the elution order was: perchlorate, iodide, bromide, chloride and bromate. Fluoride and phosphate were so strongly held by the alumina that elution was virtually impossible. The modification of silica gel with various metals is a simple and effective way to prepare ion exchangers that often have unique selectivities for analyte ions. Ohta et al. [49] described the preparation of a cation exchanger in which silica gel was first immersed in zirconium butoxide, Zr (OC4H9)4. Then the material was calcined (heated) at temperatures up to 1000 °C to form a silica–zirconia product. An excellent separation of all the alkali metal ions plus ammonium was obtained with 10 mM tartaric acid as the eluent. Divalent metal cations were strongly retained. 3.6.3 Multi-purpose Resins
Quite logically, new columns are developed to meet a real analytical need. Proliferation in the use of surfactants is a case in point. Surfactants have both hydrophilic and hydrophobic centers and are widely used in many industries because of their ability to reduce surface tension. Anionic surfactants, classified as alkanesulfonates, alkyl sulfates and alkylbenzenesulfonates, are commonly used in detergents, cleansing agents, cosmetics and hygienic products. Cationic surfactants are quaternary ammonium compounds, which are used in cosmetics, disinfectants, foam depressants, and textile softeners. Nonionic surfactants are also present in a
3.6 Other Resins
host of commercial products. Very large quantities of surfactants are discharged into the environment. Therefore, the determination of surfactants is important for product control and for environmental monitoring [50]. Chromatographic determination of surfactants can be performed by utilizing different separation modes, such as reversed-phase, ion-exchange and size-exclusion techniques. However, a mixed-mode stationary phase has been developed that is suitable for analyzing anionic, nonanionic and cationic surfactants in a single run [51]. The packing material of the commercially available column consists of 5-lm particles with a pore size of 120 Å and a specific surface area of 300 m2 g–1. The chemical surface of the particles consists of hydrophobic alkyl chains, tertiary amino groups and polar amide functional groups. Surfactant analytes are separated by mixed-mode and dipole–dipole interactions. Retention of neutral, cationic and anionic surfactants can be controlled independently by changing ionic strength, pH and organic solvent content of the mobile phase. A higher ionic strength in the mobile phase reduces the retention time (t) of anionic surfactants but gives somewhat longer t values for cationic analytes owing to stronger ion pairing with the eluent anion. The pH is an important parameter in optimizing a separation. A higher pH reduces the net positive charge on tertiary amine groups and gives lower t values for anionic surfactants. Cationic analytes have somewhat higher t values because there is less repulsion by protonated sites as the pH is increased. The eluent pH has relatively little effect on the retention of nonionic surfactants. Each of the three types of surfactant molecule has a large alkyl chain, and there is a strong hydrophobic component in its retention by the stationary phase. A higher percentage of organic solvent in the mobile phase reduces this attraction. Amide groups embedded in the stationary phase introduce a possible dipole– dipole contribution to the retention mechanism. In addition to these mechanisms, anionic surfactants are retained by ion exchange with the protonated tertiary amine groups. Cationic analytes are repulsed to some extent by the protonated amine sites and generally elute before anionic surfactants. 3.6.4 Ion-exchange Disks
How long does an IC column need to be for an effective separation? Separation of anions on monoliths 5–10 mm in length was described in Section 3.3.1. In some cases an anion-exchange disk can function as an effective ‘column’ even though the diameter of the disk is much greater than that of the mobile phase path through the disk. The growing use of short monolithic beds in the form of one or more disks placed in a tube has been reviewed [52]. Cellulose membranes similar to filter paper can be used to concentrate ions from aqueous samples without retaining most neutral solutes. The –CH2OH groups of the cellulose are functionalized to introduce either diethylaminoethyl groups (to prepare an anion exchanger), or sulfopropyl groups to prepare a cation exchanger.
65
66
3 Resins and Columns
The 3M Co. produces membrane disks containing embedded ion-exchange particles. Each disk is < 1 mm in thickness and a small disk of any convenient diameter can be cut from a larger sheet with a cork borer. The practical use of the small disks is illustrated by a 15-s separation of chromium(III), which is a 3+ cation, and anionic chromium(VI), which exists primarily as HCrO4– [53]. Two small extraction disks are placed one on top of the other in a plastic holder. Then a syringe containing the aqueous sample is attached to the holder and the sample is pushed through the disk at a rate of about 4–5 mL min–1. All of the chromium(VI) is retained on the top anion-exchange disk and the chromium(III) is extracted by the second cation-exchange disk. The concentrations on each disk are several hundredfold higher than they were in the original sample. The amounts of chromium(III) and (VI) extracted were measured directly on the surface of the respective disks by diffuse reflectance spectroscopy (DRS).
References [1] R. E. Barron and J. S. Fritz, Reproduc-
[2]
[3]
[4]
[5]
[6]
[7]
[8]
ible preparation of low-capacity anion exchange resins, Reactive Polymers, 1, 215, 1983. F. Svec, Porous monoliths: emerging stationary phases for HPLC and emerging methods, LC-GC electronic ed., Vol. 1, No. 1, June 23, 2004. M. Jacoby, Monolithic chromatography, Chemical and Engineering News, Dec. 11, 2006, p. 14. J. Li, Yan Zhu and Y. Guo, Fast determination of anions on a short coated column, J. Chromatogr. A, 118, 46, 2006. D. Victory, P. Nesterenko and B. Paull, Low-pressure gradient micro-ion chromatography with ultra-short monolithic anion exchange columns, Analyst, 129, 700, 2004. D. Connoly, D. Victory and B. Paull, Rapid, low pressure and simultaneous ion chromatography of common inorganic anions and cations or short permanently coated monolithic columns, J. Sep. Sci., 27, 912, 2004. S. Pelletier and C. A. Lucy, Achieving rapid low-pressure ion chromatography separations on short silica-based monolithic columns, J. Chromatogr. A, 118, 12, 2006. D. T. Gjerde, New macroporous stationary phase for the separation of anions, Advances in Chromatography, vol. 2,
p 169, Century International, Medfield, MA 1990. [9] R. E. Barron and J. S. Fritz, Effect of functional group structure and exchange capacity on the selectivity of anion exchangers for divalent anions, J. Chromatogr., 316, 201, 1984. [10] L. M. Nair, B. R. Kildew and R. SaariNordhaus, Enhancing the anion separations on a polydivinylbenzene-based anion stationary phase, J. Chromatogr. A, 739, 99, 1996. [11] F. Vlacil and I. Vins, Modified hydroxyethyl methacrylate copolymers as sorbents for ion chromatography, J. Chromatogr., 391, 133. 1987 [12] J. S. Fritz, Factors affecting selectivity in ion chromatography, J. Chromatogr. A, 1085, 8, 2005. [13] T. Okada, Nonaqueous anion-exchange chromatography. I. Role of solvation in anion-exchange resin, J. Chromatogr. A, 758, 19, 1997. [14] R. E. Barron and J. S. Fritz, Effect of functional group structure on the selectivity of low-capacity anion-exchangers for monovalent anions, J. Chromatogr., 284, 13, 1984. [15] R. E. Barron and J. S. Fritz, Effect of functional group structure and exchange capacity on the selectivity of anion exchangers for divalent anions, J. Chromatogr., 316, 201, 1984.
References [16] L. Li and J. S. Fritz, Novel polymeric res-
[30] J. S. Fritz, D. T. Gjerde and R. M. Becker,
ins for anion-exchange chromatography, J. Chromatogr., 793, 231, 1998. [17] L. M. Warth and J. S. Fritz, Effect of length of alkyl linkage on selectivity of anion exchange resins, J. Chromatogr. Sci., 26, 630, 1988. [18] H. Small, T. S. Stevens and W. G. Bauman, Novel ion-exchange chromatographic method using conductimetric detection, Anal. Chem., 47, 1801, 1975 [19] J. Weiss, Ion Chromatography, 2nd Ed., p 43, VCH, Weinheim, Germany, 1995 [20] J. Weiss, S. Reinhard, C. Pohl, C. Saini and L. Narayaran, Stationary phase for the determination of fluoride and other inorganic anions, J. Chromatogr. A, 706, 81, 1995. [21] R. W. Slingsby and C. A. Pohl, Anionexchange selectivity in latex-based columns for ion chromatography, J. Chromatogr., 458, 241, 1988. [22] R. W. Pepper, Chemistry Research, 1952, p. 77, Her Majesty’s Stationary Office, London, England, 1953. [23] K. W. Pepper, Sulphonated cross-linked polystyrene: A monofunctional cationexchange resin, J. Appl. Chem., 1, 124, 1951. [24] J. R. Parrish, Superficial ion-exchange chromatography, Nature, 204, 402, 1965. [25] J. S. Fritz and J. N. Story, Selectivity behavior of low-capacity, partially sulfonated macroporous beads, J. Chromatogr., 90, 267, 1974. [26] J. S. Fritz and J. N. Story, Chromatographic separation of metal ions on lowcapacity macroreticular resins, Anal. Chem., 46, 825, 1974. [27] H. Small, Solvent extraction process for the recovery of uranium and rare earth metals from aqueous solutions, U.S. Patent 3 102 782, 1962. [28] H. Small, Gel liquid extraction. The extraction and separation of some metal salts using tri-n-butylphosphate gels, J. Inorg. Nucl. Chem., 18, 232, 1961. [29] T. S. Stevens and H. Small, Surface sulfonated styrene divinyl benzene – optimization of performance in ion chromatography, J. Liq. Chromatogr., 1, 123, 1978.
Cation chromatography with a conductivity detector, Anal. Chem., 52, 1519, 1980. S. Papanu, C. Pohl and A. Woodruff, New high speed cation exchange columns for ion chromatography, Paper presented at the Pittsburgh Conference and Exposition on Analytical Chemistry and Applied Spectroscopy, Atlantic City, NJ, U.S.A.1984. H. A. Battaerd, Core-dual shell graft copolymers with ion exchange resin shells, U.S. Patent 3 565 833, February 23, 1971. H. A. Battaerd and R. J. Siudak, Synthesis and ion-exchange properties of surface grafts, J. Macromol. Sci. Chem., A4, 1259, 1970. J. J. Kirkland, Superficially porous chromatographic packing with sulfonated fluoropolymer coating and chromatographic packing with chemically bonded organic stationary phases, U.S. Patents 3 577 266, May 4, 1971, and 3 722 181, March 27, 1973. C. G. Horvath, B. A. Preiss and S. R. Lipsky, Fast liquid chromatography: An investigation of operating parameters and the separation of nucleotides on pellicular ion exchangers, Anal. Chem., 39, 1422, 1967. C. Horvath and S. R. Lipsky, Column design in high pressure liquid chromatography, J. Chromatogr. Sci., 7, 109, 1969. D. C. Locke, J. T. Schmermund and B. Banner, Bonded stationary phases for chromatography, Anal. Chem., 44, 90, 1972. D. H. Saunders, R. A. Barford, P. Magidman, L. T. Olszewski and L. T. Rothbart, Preparation and properties of a sulfobenzylsilica cation exchanger for liquid chromatography, Anal. Chem., 46, 834, 1974. G. J. Sevenich and J. S. Fritz, Preparation of sulfonated gel resins for use in ion chromatography, Reactive Polymers, 4, 195, 1986. G. J. Sevenich and J. S. Fritz, Metal ion selectivity on sulfonated cationexchange resins of low capacity, J. Chromatogr., 371, 361, 1986.
[31]
[32]
[33]
[34]
[35]
[36]
[37]
[38]
[39]
[40]
67
68
3 Resins and Columns [41] D. Jensen, J. Weiss, M. A. Rey and
[42]
[43]
[44]
[45]
[46]
[47]
C. A. Pohl, Novel weak-acid cationexchange column, J. Chromatogr., 640, 65, 1993. J. Morris and J. S. Fritz, Ion chromatography of metal cations on carboxylic acid resins, J. Chromatogr., 602, 111, 1992. P. Kolla, J. Köhler and G. Schomburg, Polymer-coated cation-exchange stationary phases on the basis of silica, Chromatographia, 23, No. 7, 465, 1987. L. M. Nair, R. Saari-Nordhaus and J. M. Anderson, Jr., Simultaneous separation of alkali and alkaline-earth cations on polybutadiene-maleic acid-coated stationary phase by mineral acid eluents, J. Chromatogr., 640, 41, 1993. L. M. Nair, R. Saari-Nordhaus and J. M. Anderson, Jr., Ion chromatographic separation of transition metals on a polybutadiene maleic-acid-coated stationary phase, J. Chromatogr. A, 671, 43, 1994. R. Saari-Nordhaus, H. Pham and J. M. Anderson, Jr., A new cation stationary phase for challenging ion chromatography applications. Poster 2292, Pittcon 699, Orlando, FL. B. R. Edwards, A. P. Giauque and J. D. Lamb, Macrocycle-based column
[48]
[49]
[50]
[51]
[52]
[53]
for the separation of inorganic cations by ion chromatography, J. Chromatogr. A, 706, 69, 1995. G. L. Schmitt and D. J. Pietrzyk, Liquid chromatographic separation of inorganic anions on an alumina column, Anal. Chem., 57, 2247, 1985. K. Ohta, M. Morikawa, K. Tanaka, Y. Uwamino, M. Furikawa and M. Sando, Ion-chromatographic behavior of alkali metal cations and ammonium ion on zirconium-adsorbing silica gel, J. Chromatogr. A, 884, 123, 2000. L. M. Nair and R. Saari-Nordhaus, Recent developments in surfactants analysis by ion chromatography, J. Chromatogr. A, 804 233, 1998. X. Liu, C. A. Pohl, and J. Weiss, New polar-embedded stationary phase for surfactant analysis, J. Chromatogr. A, 1118, 29, 2006. O. W. Reif, V. Nier, U. Bahr, and R. Freitag, Use of short monolithic beds for isolation and separation of biomolecules, J. Chromatogr. A, 664, 13, 1994. A. Steiner, M. D. Porter and J. S. Fritz, Ultrafast concentration and speciation of chromium(III) and (VI), J. Chromatogr., A. 1118, 62, 2006.
69
4 Detectors 4.1 Introduction
This chapter describes several different detectors that may be used in ion chromatography. But the reader may ask why so many detectors? Why not just use the conductivity detector that came with my instrument? Too many times the methodology that is devised for a particular analytical problem is just ‘good enough’. The analysis can be performed but the method is barely adequate in terms of resolution of peaks or in terms of sensitivity. And because it is just barely adequate, the method is not rugged. Often these issues can be solved by understanding how detectors operate, how ions are detected and choosing a better detector. For example, detection of iodide or nitrate in the presence of a salt (sodium chloride) matrix can be accomplished with conductivity detection. But UV detection would be much better because iodide will absorb UV light and chloride will not. Because the detection is selective for iodide, the separation conditions can be optimized for rapid interference-free elution. In IC, the detector must be able to ‘pick out’ and measure sample ions in the presence of a background of eluent ions. There are several methods that can be employed to make this possible. One is to choose a detector that will respond only to the sample ions of interest, but not to the eluent ions. Another method is to use indirect detection (sometimes called replacement detection). This is where the eluent has a background signal and the presence of sample ions causes a decrease in eluent ions through a replacement process. The detector measures the decrease in eluent ions when the sample ion peak elutes and a decreasing signal is detected. The most widely used method for ion chromatography detection is to treat or choose the eluent prior to detection to make the eluent ions less detectable and/or make the sample ions more detectable. The most common example of this is chemical suppression used in conductometric detection. The suppressor is really a chemical reactor that reacts with the post-column eluent stream and changes the ionic counterion for the eluent and for the sample peaks. In its most common form, sodium or potassium ions are removed from the stream and hydronium ions are added in an exchange process. This makes the background signal less Ion Chromatography, 4th Ed. James S. Fritz and Douglas T. Gjerde Copyright © 2009 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim ISBN: 978-3-527-32052-3
70
4 Detectors
conducting and the sample signal more conducting. Another example of treatment prior to detection is post-column reaction with a color-forming reagent. PAR, a color-forming chelator, can be added, post-column, to a separation of metal ions to make them detectable by visible spectrophotometric detection. The eluent ions do not react with the color-forming reagent. Detectors can be classified either as general or selective. A general detector will respond to all or most of the ions that pass through the detector cell. A conductometric detector is classified as a general detector because all ions will conduct electricity (although to different degrees). UV–Vis spectrophotometric, atomic emission, atomic absorption, and electrochemical detectors can be considered to be selective detectors because they respond only to certain ions. However, any of these can be made into a general-type detector by a post-column reaction of sample ions with an appropriate reagent. General and selective detectors both have their place in ion chromatography. If only one detector were available, a general detector would be most desirable because it would be of the most general use. However, a selective detector can be extremely effective in measuring a single ion of interest (or a small group of ions) from a high eluent background or a high sample matrix ion. Another key advantage to selective detection is the ability to achieve lower detection limits. A selective detector may or may not have a higher sensitivity (signal per unit concentration) for a particular ion. However, the lower background signals produced by selective detection will translate into lower detection limits because the signal-to-noise ratio is improved.
4.2 Conductivity Detectors
Conductance is the property of a solution containing a salt to conduct electricity across two electrodes. Conductance should not be confused with conduction, which is the mechanism by which electrons flow, or with conductivity, the property of a material. The ability of a solution to conduct is directly proportional to the salt concentration and the mobility of the individual anions and cations. As the ionic character of a molecule is increased, the conductivity increases. Small, mobile ions conduct quite readily and to a much greater extent than large bulky ions. For example, the hydroxide anion is small and mobile and will conduct much better than the propionate anion, which larger and bulkier. The relative conductances of ions are shown in Table 4.1. Molecular substances such as solvents (water and methanol) and solutions of nonionized organic acids and carbohydrates do not conduct electricity and are not detected by conductivity. The portion of a weak acid that does ionize will contribute to the conductivity signal. The ionic form of the weak acid depends on pH. A weak acid with a pKa larger than about 6 cannot be detected with suppressed conductivity detection because all anions are converted to the acid form by the suppressor. So while it is possible to separate both arsenite and arsenate on the
4.2 Conductivity Detectors
anion-exchange column (with a high pH eluent), it is only possible to detect the stronger acid, arsenate, by suppressed conductivity, and is not possible to detect the weak acid, arsenite. Other anions that cannot be detected by suppressed conductivity detection include borate, silicate and cyanide. Nonsuppressed, single-column methods, selective detection, or post-column reaction methods are required for the detection of salts of weak acids.
Table 4.1 Limiting equivalent ionic conductances in aqueous
solution at 25 °C. (Units: S cm2 equiv–1) Anions
k–
Cations
k+
Hydroxide
198
H3O+
350
+
Fluoride
54
Li
39
Chloride
76
Na+
50
Bromide
78
K+
74
Iodide
77
Nitrate
71 2–
Sulfate
80
78
+
77
Rb Cs
2+
Mg
53
69
2+
Ca
60
Chlorate
72
Sr2+
59
Perchlorate
67
Ba2+
64
66
2+
71
2+
53
2+
55
2+
53
2+
Phosphate
3–
+
Thiocyanate Hydrogen carbonate Carbonate
2–
Chromate
45 72 85
Pb
Zn Cu Co
Cyanide
82
Cd
54
Formate
55
Fe2+
54
41
3+
68
3+
70
Acetate Propionate 2–
Oxalate
3–
Citrate
36 74 56
Fe La
3+
Ce
NH4
70 +
Benzoate
32
CH3NH3
Phthalate2–
38
Et4N+
Borate/gluconate
26
73 +
58 33
71
72
4 Detectors
There are some IC users who employ direct or indirect nonsuppressed conductivity detection where the separation column is connected directly to the conductivity cell. The advantage of this detection method is simplicity in instrument design and operation and the ability to detect salts of weak acids. Cation detection can be more sensitive. But because of high sensitivity and reduced background noise, the most common form of conductivity detection is with the use of suppressors (see Section 6.2). Dasgupta and coworkers have proposed a novel approach to simultaneously practicing both suppressed and nonsuppressed ion chromatography [1–3]. Several renditions were proposed (Figure 4.1). The most effective and practical (Figure 4.1c) is based on first using a conventional NaOH eluent-suppressed IC system. The effluent from detector 1 is combined with a constant concentration of NaOH. The NaOH is introduced by an in-line electrolytic NaOH generator. Then the flow is directed to a second conductivity detector. The second detection background conductance is typically maintained at a level of 20–30 lS cm–1 corresponding to about 0.1 mM NaOH. The second detector output is the same as what would be observed in a single-column mode with a low-concentration NaOH eluent. Together, the two detector outputs provide detection at the microgram per liter level across the pKa range, from fully dissociated to very weak acids up to pKa 9.5.
Figure 4.1 Three potential approaches to perform simultaneous, suppressed, and nonsuppressed detection. (a) Split stream approach. DD, dummy dispersion device; Ds, suppressed detector; Dn, nonsuppressed detector. Restrictors R1 and R2 are adjusted to provide the
same residence time and flow rate in each of the branches. (b) Single column approach. Dn precedes the suppressor. (c) NaOH introduction approach. A small constant quantity of NaOH is introduced after Ds; no restriction is placed on eluent NaOH concentration [3].
4.2 Conductivity Detectors
The two-detector system also provides qualitative information about peak purity and sample pKa. This information is gained by plotting the two detector outputs against each other. In addition to the usual ratio plots, it is suggested that multidimensional detection in any separation system is best served by plotting the raw detector outputs against each other. This type of implementation of single-column and suppressed IC provides information beyond the sum total of that obtained with either approach alone. 4.2.1 Conductivity Definitions and Equations
Electrolytic conductivity is the ability of an electrolytic solution to conduct electricity between two electrodes across which an electric field is applied. Ohm’s law, V = I R, is obeyed, and the magnitude of the current depends, in part, on the magnitude of the applied potential. The conductance, G, of a solution is expressed in terms of the solution electrolytic resistance. It is measured in reciprocal ohms (mhos) or in the Sl unit siemens (S). G = 1/R
(4.1)
Specific conductance (k) takes into account the area of the electrodes (A) in cm2 and the distance (l) between electrodes, in cm. Conductance increases with the area of the electrodes but decreases as the distance between the electrodes is increased. k = G (l/A)
(4.2)
Thus, k has the units S cm–1. The cell constant (K) is equal to l/A in Eq. (4.2) and has the units cm–1. K = l/A
(4.3)
k= G K
(4.4)
Equivalent conductance takes into account the concentration of the chemical solution and is defined by the following equation: K = 1000 k/C
(4.5)
where C is the concentration in equivalents per 1000 cm3 and K has the units S cm2 equiv–1. Combining Eqs. (4.4) and (4.5) gives the following equation, which relates equivalent conductance to measured conductance, G: G = K C/1000 K
(4.6)
73
74
4 Detectors
A conductivity detector consists of a detection cell, a readout meter, and the electronics required for measuring the conductance and varying the sensitivity setting. The readout for conductance G is given in Siemens (S), or, actually, microSiemens (lS) for solutions that are comparatively dilute. The specific conductance for a solution can be calculated from the solution conductance (G) if the cell dimensions are known [see Eq. (4.2)]. However, the usual practice is to measure the conductance of a dilute solution of known specific conductance (such as 0.00100 N KCI) and calculate the cell constant from Eq. (4.4). Once the cell constant is known, the specific conductances of other solutions can be calculated from the measurement of G. The most common cell constant for detectors is 1 or 10; the smaller value is more sensitive. With a conductivity detector with a known cell constant, the conductances of various solutions of known concentration can be calculated from a table of equivalent conductances using Eq. (4.6). The limiting equivalent conductances of some common ions are given in Table 4.1. The equivalent conductances of ions generally decrease with increasing concentration because of interionic effects. For dilute solutions (10–5 to 10–2 N) the equivalent conductances are not greatly different from the values listed in the table. Example: Calculate the expected conductance of a 1.0 × 10–4 N solution of sodium benzoate in a conductivity cell with a cell constant of 10 cm–1. From Table 4.1, the equivalent conductance of sodium benzoate will be kNa+ + kBz-. Therefore: kNa+Bz- = 50 + 32 = 82 Substituting this into Eq. (4.6): G = (82 × 1.0 × 10–4)/(1000 × 10) = 8.2 × 10–7 mhos or 0.82 lS A more typical ion to detect might be chloride. The equivalent conductance is higher so a higher signal would result from this ion, assuming the same peak width and the same sodium counterion. If hydronium ion rather than sodium is the counterion to chloride, then the signal will be multiplied by another factor of 3.4. 4.2.2 Principles of Cell Operation
When an electric field is applied to two electrodes in an electrolytic solution, anions in the solution move toward the anode electrode and cations toward the cathode electrode. The number and the velocities of the ions in the bulk electrolyte determine the resistance of the solution. The ionic mobilities, or the velocity of the ion per unit electric field, depend on the charge and size of the ion, the temperature and type of solution medium, and the ionic concentration. As the potential that is applied across the electrodes is increased, the ionic velocities increase. Thus, the detector signal is proportional to the applied potential.
4.2 Conductivity Detectors
This potential can be held to a constant value or it can oscillate to a sinusoidal or pulsed (square) wave. Cell current is easily measured; however, the cell conductance (or reciprocal resistance) is determined by knowing the potential to which the ions are reacting. This is not a trivial task. Ionic behavior can cause the effective potential that is applied to a cell to decrease as the potential is applied. Besides electrolytic resistance that is to be measured, Faradaic electrolysis impedance may occur at the cell electrodes resulting in a double layer capacitance. Formation of the double layer capacitance lowers the effective potential applied to the bulk electrolyte. 4.2.3 Conductance Measurement
Techniques involving the use of alternating electrode potentials eliminate the effects of the processes associated with the electrodes. Reversing the polarity of the applied electrode potential reverses the direction of the ion motion, changes the type of electrolysis and changes the type of capacitance formation. The relaxation time (or the ability to recovery) is different for each type of process. As the frequency is increased, effects due to electrolysis are reduced or eliminated and the bulk of current flow is through capacitance formation. An upper frequency limit for detector operation is approximately 1 MHz. At this point the ions cease to move in response to the electric field, although dipole reorientation of the ion electron structure will still occur. Capacitance effects are controlled by matching the cell capacitance in the electronic circuitry or by measuring the instantaneous current. The instantaneous current is the current that is obtained when the potential is first applied and the double layer has not formed. Some detectors apply a sinusoidal wave potential across the cell electrodes at 100 to 10 000 Hz. A typical detector of this type operates at a frequency of 1 kHz and at a potential of up to 20 V with no electrolysis occurring. This detection method is called synchronous detection. This is when the only current component measured is that current which is in phase with the applied potential frequency. In effect, the measured current flow is always due to an ‘instantaneous’ potential. Other detectors use a bipolar pulse conductance technique [4, 5]. The technique consists of the sequential application of two, short (about 100 ls) voltage pulses to the cell. The pulses are of equal magnitude and duration and opposite polarity. At exactly the end of the second pulse, the cell current is measured and the cell resistance is determined by applying Ohm’s law. Because an instantaneous cell current is measured in the bipolar pulse technique, capacitance does not affect the measurement and an accurate cell resistance measurement is made. 4.2.4 Conductivity Hardware and Detector Operation
Conventional conductivity detector cells where the electrolyte is in contact with the electrodes are likely to use electrodes made from 316 stainless steel. A new
75
76
4 Detectors
cell should be treated with 1N nitric acid for about 60 min to ‘deactivate’ or ‘passivate’ the cell and stabilize the signal. In fact, such nitric acid treatment is a good idea for all parts of a stainless steel IC system. The mobility of ions in solution varies with the solution temperature. Ionic solutions will increase in conductivity about 2% for every degree increase in temperature. Conductivity detectors usually compensate automatically for temperature change by employing a thermistor monitor and compensation circuitry in which resistance changes linearly with solution temperature. Still, the detector cell (and even the column and tubing) should be placed in an oven for the best detector performance. It is helpful to insulate other components of the ion chromatograph. If the laboratory has large daily temperature swings, temperature control of the column and cell becomes more important. Generally, the instrument temperature is set to at least 5 °C above the maximum temperature that the laboratory is likely to reach in a given day. Control of the instrument oven and detector temperature and simply keeping the instrument out of drafts are probably the most important two things that a user can do to contribute to better detector stability. The flow path may also include a compartment containing the thermistor probe for electronic feedback. Even though conductance cells tend to be low dead volume, the total dead volume in a conductance detector can be quite large, and considerable mixing of the eluent stream may take place after the peak has been measured. If two detectors are used, it may be best to place the conductivity detector cell last to avoid peak broadening. 4.2.5 Contactless Conductivity Detection
Zemann and coworkers have developed a novel contactless conductivity detector [6–12]. Contactless conductivity detection offers the advantage of avoiding detection dead volumes. This is especially important for miniaturized chromatographic and electrophoresis systems. The detector works without direct contact of the electrode with the eluent or sample. The sensor is based on two metal tubes that are placed around a fused silica capillary with a detection gap of approximately 1.5 mm (Figure 4.2). The conductivity sensor is based on two metal tubes that act as cylindrical capacitors. The electrodes may be placed around any nonconducting tubing such as fused silica, PEEK, or Teflon. Dead volume of the connecting tubing is minimized and an extremely low dead volume cell can be manufactured. A high oscillating frequency of 40–100 kHz is applied to one of the electrodes. A signal is produced on the other electrode as soon as an analyte zone with a different conductivity compared to the background passes through the detection gap. An amplifier and rectifier are connected to the second electrode to measure resistance between the two electrodes. To isolate the two capacitors associated with each electrode, a thin piece of copper is placed between the electrodes and grounded.
4.3 Ultraviolet-Visible (UV–Vis) Detectors
Figure 4.2 Schematic drawing of the contactless capacitively coupled conductivity detector [6].
4.3 Ultraviolet-Visible (UV–Vis) Detectors 4.3.1 UV–Vis measurement
A spectrophotometric UV–Vis detector is selective, yet its selectivity can be changed simply by changing the wavelength monitored by the detector. Versatility of the detector can be increased by adding a color-forming reagent to the eluent or the column effluent. The fundamental law under which ultraviolet-visible (UV– VIS) detectors operate is the Lambert–Beer law. It can be stated in the following form: A= e b C
(4.7)
A is the absorbance of a species of concentration C, and with an absorptivity e, in a cell of length b. Concentration is usually in molar concentration units and the path length is measured in cm. The term (molar) absorptivity has units that are the inverse of the C and e units. This leaves A dimensionless; it is usually described in terms of absorbance units. A detector set to a certain sensitivity, for example, 0.16, is said to be at 0.16 Absorbance Units Full Scale sensitivity (0.16 AUFS sensitivity). The Lambert–Beer equation is useful for choosing conditions for the separation and detection of ions. The eluent ions should have a low absorptivity and the sample ions should have reasonably high absorptivity. In the special case of indirect detection this should be reversed. In this case, the eluent has an absorption signal and the sample is detected by a decrease in the background signal. It is important to note that when discussing the properties of the eluent and sample ions, it makes a difference whether one is separating anions or cations.
77
78
4 Detectors
For example, if a separation of anions is being discussed, then the absorptivities of the eluent anion and sample anions are considered. But if a low background signal is needed, then of course the cation that is counter to the eluent anions must have a low absorptivity as well. Some absorbance data are given in this chapter and in some of the applications described in this book. If an ion absorptivity for a particular wavelength is unknown, it can be measured with a spectrophotometer and ion solution of known concentration. The discussion of UV–Vis detectors for use in ion chromatography is divided into two parts: (a) the direct monitoring of column effluents and (b) post-column derivatization with subsequent spectrophotometric measurement. 4.3.2 Direct Spectrophotometric Measurement
Alkali metal ions are not detected by UV. However, many anions do absorb at lower wavelengths. A list of anions and their detection wavelengths is shown in Table 4.2.
Table 4.2 Solutes for direct spectrophotometric detection in IC
after ion-exchange separation. Solute AsO43–, AsO33– Au(CN)2
–
Wavelength (nm)
Ref.
200
[13]
214
[14]
–
Br
195–214
[14]
BrO3–
195–210
[13, 16, 17]
C2O42–
205
[18]
Citrate
205
[18]
–
Cl
190
[19, 20]
CIO2–
195
[13]
CIO3–
195
[13]
CN–
200
[21]
200
[21]
365
[22]
350, 220
[23]
225
[24]
CNO– CrO4 III
2– VI
Cr , Cr -EDTA –
Cr, Pt, Au-Cl complexes
4.3 Ultraviolet-Visible (UV–Vis) Detectors Table 4.2 (Continue) Solutes for direct spectrophotometric
detection in IC after ion-exchange separation. Solute
Wavelength (nm)
Ref.
HCOO–, CH3COO–
190
[19]
210–235
[25, 26]
195–210
[13, 17]
Metal Cl complexes
210–225
[27-30]
Metal CN– complexes
210–214
[31, 32]
Metal EDTA complexes
210
[33]
200
[21]
195
[13]
200–214
[20, 34, 35]
210
[36]
190
[19]
215
–
205
–
I
–
IO3
– –
MoO4
2–
N3– –
NO2 , NO3
–
Organoarsenic acids PO4
3–
S2– 2–
2–
S2O3 , S3O6 , –
SCN
S4O62–
195–205
[37, 38]
–
SeCN
195
[13]
SeO32–
195
[13]
SeO42–
195
[13]
SO32–
200
[38]
SO42–
190
[19, 20]
The references listed show the different applications of the detection method. The 190 to 210 nm range can be used for the detection of azide, chloride, bromide, bromate, iodide, iodate, nitrite, nitrate, sulfite, sulfide, and selenite. Other work has shown that the 210 to 220 nm range is useful for detecting trithionate, tetrathionate, and pentathionate down to the low nanogram levels. UV detection is particularly useful for anions such as nitrate and iodide, which absorb at the longer wavelengths. There have been a number of methods reported for the UV detection of nitrite and nitrate in drinking water and cured meat [39]. Aromatic acids absorb well and methods for detecting these anions are powerful. The usefulness of direct UV detection can be considered to be limited because sulfate is not detected by UV, and chloride, phosphate and others are difficult to detect. Sulfate is probably the most widely analyzed anion by ion chromatography
79
80
4 Detectors
so this is a serious limitation. On the other hand, anions that are difficult to detect make ideal eluent anions. The absorbance of metal chloride complexes in the ultraviolet spectral region has been used extensively to automatically detect metal ions in liquid chromatography [27–30]. The absorption wavelength maxima of the metal chloride complexes are shown in Table 4.3. Metal–EDTA complexes also absorb quite well. Another technique is indirect detection. In this method, the eluent absorbs strongly in the visible or ultraviolet spectral region. A wavelength is selected where the (usually aromatic) eluent absorbs but the sample ions do not [41–43]. Briefly, because an ion-exchange process is involved, a sample ion can only be eluted by displacement of the eluent ion. This results in a decrease in the signal when a sample ion peak is eluted. While several authors and users have been successful with indirect UV detection, it can be difficult to get the right conditions for separation and detection. Indirect UV detection is generally used only in cases where the separation and detection conditions have been carefully worked out and where a high quality UV detector is available. Temperature control of the column is recommended to control the baseline noise. 4.3.3 Post-column Derivatization
The post-column method of derivatization of column fractions has been well established from older ion-exchange separations. An appropriate reaction is performed on each fraction to determine the metal ion concentration in that fraction. The automatic addition of a color-forming reagent to an ion-exchange column effluent and analysis by flow-through cell detection is more recent. However, many of the color-forming reagents and buffers used in ion chromatography are the same as those used in the classical fraction method determinations. The ideal color-forming reagent reacts with a large number of metal ions and has low background absorption. Sickafoose [44] studied the reagents alizarin red S, arsenazo III, chlorophosphonazo III, chrome azurol S, quinalizarin, 4-(2-pyridylazo)resorcinol (PAR), 4-(2-thiazolylazo)resorcinol (TAR), and xylenol orange. Chrome azurol S and quinalizarin are of limited value, but the other reagents each react with 20 or more metals. PAR is the most general, reacting with 34 metals. Table 4.4 shows a list compiled by Fritz and Story [45] of metals and their reaction with color-forming agents. The 0.0125% PAR solution was made in 5 M ammonium hydroxide, the 0.00375% arsenazo III solution was in 2 M ammonium hydroxide and 1 M ammonium acetate, and the arsenazo I solution was in 3 M ammonium hydroxide. Other work [46, 47] was done with more dilute PAR solutions (4 × 10–4 M PAR with 3 M ammonium hydroxide and 1 M acetic acid). For lower detection limits (because of lower background signal), the PAR concentration can be reduced even more. But care should be taken not to ‘overload’ the reagent with too high concentrations of sample ions. Imanari et al. have reported a spectrophotometric detection of many inorganic anions using a post-column reactor [48]. A stream of ferric perchlorate, which is
4.3 Ultraviolet-Visible (UV–Vis) Detectors Table 4.3 Complex formation of inorganic anions with ferric
perchlorate color-forming reagent[a] [48]. Anion
kmax(nm)
Anion
kmax(nm)
CrO42–
305, 344
SO32–
308
–
3–
310
PO3
305
H2PO2–
<300
305
IO3
–
<300
SO42–
306
CO32–
<300
CI–
335
Br–
<300
SCN
Fe(CN)6
4–
Fe(CN)6
3–
P2O7
4–
I– P3O105– S
2–
310
B4O7
306, 350
BrO3–
310 <300
S2O3
2–
2–
–
CN
<300
<300 <300 <300
2–
<300
–
[b]
SiO3
308
NO3
NO2–
372, 360
F–
[b]
PO43–
310
ClO3–
[b]
[a] [b]
0.8 M HClO4, 0.05 M Fe(ClO4)3. Does not react.
essentially colorless, is mixed with the column effluent. The ferric perchlorate is colorless because perchlorate is a poor complexing anion, but most anions will complex the iron and form colored species that can be detected at 330–340 nm (Table 4.3). A similar detection method works for ions such as orthophosphate, pyrophosphate, nitrilotriacetic acid (NTA), and ethylenediaminetetraacetic acid (EDTA) [49].
81
82
4 Detectors Table 4.4 Reactions of color-forming reagents and metals.
X denotes a positive reaction [44]. Metal ion
Reagent Arsenazo I
Arsenazo III
PAR
Thorium(IV)
x
x
Zirconium(IV)
x
x
Hafnium(I/)
x
x
Aluminum(III)
x
Chromium(III)
x
Lanthanides(III)
x
Bismuth(III)
x
Iron(III)
x
Iron(II)
x
Vanadium(IV)
x
Maganese(II)
x
Cobalt(II)
x
Nickel(II)
x
Copper(II)
x
x
Zinc(II)
x
x
Cadmium(II)
x
Mercury(II)
x
Lead(II)
x
Magnesium(II)
x
Calcium(II)
x
x
x
Barium(II)
x
Strontium(II)
x
4.3.4 UV–Vis Hardware and Detector Operation
Detectors should have variable wavelength capability. The UV range 190 to 370 nm is used most often for direct detection. The visible wavelength region is used quite often with the post-column reagents. Metal separations with PAR detection are
4.4 Fluorescence Detector
operated at 520–535 nm and arsenazo III detection at 653 nm. Many of the eluents and other solutions can be corrosive. With highly corrosive solvents, glass or plastic tubing and connectors are required. Several companies now produce commercial post-column reactors specifically designed for IC separations. In addition to this, reactors designed for flow injection analysis (FIA) instrumentation often work. The most important features of any commercial post-column reactor are uniform mixing of the column effluent and the color-forming reagent and the delivery of the mixture to the detector without broadening the peak. In order to accomplish this, the mixing chamber must be small and efficient. The mixing must be thorough and adapt to any fluctuations in flow from either stream. Sometimes this is not so easy, considering that delivery of the column effluent can pulse with the pump strokes of the pumps. The color-forming reagent concentration should be in large enough excess to keep favorable reaction kinetics and a linear calibration curve. The buffer should be at a high concentration so that reactions are reproducible, rapid, and complete. High buffer concentrations are sometimes necessary to control the pH of a variety of eluents at high concentrations. Because of the high concentrations, reagents may precipitate, tubing and valves may plug, and cell walls may cloud. All solutions should be filtered before use. The instrument may have to be cleaned daily.
4.4 Fluorescence Detector
Fluorescence detection is closely related to UV–Vis detection, but in this case the emission rather than the absorption of light is measured. Some molecules that are planar and rigid have the ability to fluoresce. Fluorescence is the emission of light from a molecule at a lower energy after it has absorbed higher energy light. With the absorption of the appropriate radiant energy, a molecule is raised from a ground state vibrational level in to one of many possible vibrational levels in one of the excited electronic levels. The fluorescence emission of light results from a relaxation to a lower electronic energy state from its excited higher electronic state giving a wavelength of light corresponding to the decrease in energy states. The light wavelength is the inverse to the energy of the light so that longer wavelengths have lower energy and shorter wavelengths have higher energy. The absorption and emission wavelength of light are specific for that particular fluorescent dye molecule. The emission spectrum is measured by keeping the excitation wavelength constant and measuring the emission signal intensity at the various wavelengths. The absorption maximum is always at a shorter wavelength (higher energy) than the emission maximum (lower energy). After molecules have absorbed radiant energy and become excited to a higher energy state, they must lose their excess energy in order to return to the normal ground electronic state. Most molecules will simply heat up and lose the excess energy through heat radiation. Molecules that are rather planar and rigid and have conjugated double bonds have a greater tendency to fluoresce light. These
83
84
4 Detectors
molecules are most often referred to as fluorescent dyes. Common dyes used in molecular biology are listed in Table 4.5 and include FAM, TET, HEX, NED, and many others.
Table 4.5 Fluorescent dyes for tagging nucleic acid fragments
and the conditions for excitation and detection. Data assembled from web sites of www.SyntheticGenetics.com, www.MolecularProbes.com and www.idtdna.com. Dye
Excitation Maximum (nm)
Emission Maximum (nm)
6-FAM (6-carboxy-fluorescein)
494
518
JOE
520
548
TAMRA (Carboxytetramethyl rhodamine)
565
580
ROX (Rhodamine X)
585
605
Pacific Blue
416
451
Fluorescein
492
520
HEX(Hexachlorofluorescein)
535
556
TET Tetrachlorofluorescein)
521
536
Texas Red
596
615
Cy3
550
570
Cy5
649
670
Cascade Blue
396
410
Marina Blue
362
459
Pacific Blue
416
451
Oregon Green 500
499
519
Oregon Green 514
506
526
Oregon Green 488
495
521
Oregon Green 488-X
494
517
Rhodamine Green
504
532
Rodol Green
496
523
Rhodamine Green-X
503
528
Rhodamine Red-X
560
580
4.5 Electrochemical Detectors Table 4.5 (Continue) Fluorescent dyes for tagging nucleic acid
fragments and the conditions for excitation and detection. Data assembled from web sites of www.SyntheticGenetics.com, www.MolecularProbes.com and www.idtdna.com. Dye
Excitation Maximum (nm)
Emission Maximum (nm)
BODIPY FL
502
510
BODIPY 530/550
534
551
BODIPY 493/503
500
509
BODIPY 558/569
559
568
BODIPY 564/570
563
569
BODIPY 576/589
575
588
BODIPY 581/591
581
591
BODIPY FL-X
504
510
BODIPY TR-X
588
616
BODIPY TMR
544
570
BODIPY R6G
528
547
BODIPY R6G-X
529
547
BODIPY 630/650-X
625
640
One of the great strengths of DNA and RNA Chromatography (Chapter 12) is that UV detection can be used for the direct detection of nucleic acids. But many researchers also use fluorescence detection because of its extreme sensitivity. Fluorescence detection lowers the amount of DNA or RNA that can be detected by a factor of 10 to 100 (or even 1000 in some reported cases).
4.5 Electrochemical Detectors
Electrochemical detectors can be classified according to the three fundamental parameters of voltage or potential (V), resistance (R), or current (I). These terms are related via Ohm’s Law, which is V = I R. Electrochemical detectors are considered to be conductometric, potentiometric, amperometric or coulometric detectors. Conductometric detection has been discussed earlier in this chapter and there is only limited discussion in this section. Coulometric detection is not commonly used and is discussed only briefly.
85
86
4 Detectors
Potentiometric detectors measure potential in volts (V) under conditions where I essentially equals zero; and amperomteric and coulometric detectors measure current in amperes (A) as a function of applied potential. Conductometric detectors can be considered universal whereas potentiometric amperometric and coulometric detectors are selective detectors. Electrochemical detectors offer many advantages such as high sensitivity, high selectivity, and wide linear range. They are easily adapted to microchromatographic and electrophoric separation systems because the response is dependent on electrode area and not path length as in optical absorbance methods. The detectors are often simple, rugged, and relatively inexpensive. Unfortunately, electrochemical detectors are also sensitive to flow rate and eluent constituents including dissolved oxygen, which is difficult to eliminate or control. Conductometric detection is least affected, especially when using eluent suppression technology, which has facilitated its popularity. Potentiometric and amperometric techniques often suffer from slow response times, and they are sometimes difficult to use because of their heterogeneous detection process. In other words, analytes must diffuse to an electrode surface in potentiometric and amperometric techniques, which can lead to ‘poisoning’ of the electrode surface. Many amperometric techniques require daily polishing of the electrode surface. Pulsed electrochemical detection (PED) is designed to mitigate these types of problems. 4.5.1 Potentiometric Detection
Potentiometric detectors typically measure the potential difference (DE) across a membrane, which originates from the difference in analyte concentration in the eluent versus an internal reference solution. The most common potentiometric measuring device is a pH electrode, in which a glass membrane responds to hydronium ion concentration in the test solution. Other ion-selective, or indicator, electrodes are also available commercially. The attribute of an indicator electrode to be highly selective for a particular species is also its drawback, in that a different electrode is needed for each type of ion. Halides and sulfates can be monitored using silver/silver salt and lead/lead salt electrodes, respectively [50]. 4.5.2 Conductometric Detectors
Conductivity detectors are used primarily to detect ions in conjunction with ion chromatography (IC); a detailed review is outlined earlier in this chapter. This type of detector monitors the ability of the eluent to conduct electricity. In nonsuppressed ion chromatography, the conductivity of the eluent is usually minimized by the careful selection of reagents and control of their concentration. Under these conditions, charged analyte ions are more conductive than the eluent and a signal is generated as they are pumped through the detector. In suppressed ion chromatography, the response of the background is neutralized using a post-col-
4.5 Electrochemical Detectors
umn suppressor (for example, eluent hydroxide ions in anion-exchange chromatography and eluent hydrogen ions in cation-exchange chromatography are neutralized to water). Under these conditions, only the charged analyte ions are detected. Suppressed IC inherently and in practice has superior detection limits as compared to nonsuppressed IC, despite its having a larger dead volume. Postcolumn suppression offers the following benefits: . Increased signal-to-noise. Increased analyte conductivity increases the signal, and decreasing background eluent conductivity lowers noise. As a result, suppressed conductivity often provides lower detection limits and a wider dynamic range, and dirty samples can be diluted more, thereby extending column life. . Improved chromatograms. The elimination of interference from eluent and sample counterions leads to fewer system peaks and baseline artifacts. . Improved system performance. In addition to improved gradient compatibility, suppressed conductivity allows for the use of more concentrated eluents, which provides a greater range of elution control and the ability to use larger sample volumes. Suppressor-ion chromatography is widely accepted across the scientific community. It has been used routinely in thousands of laboratories over the past twenty years, and methods using suppressed-ion anion chromatography have been accepted and approved by governmental regulatory agencies. In fact, the US EPA has incorporated this technique in many of its own methods of analysis including EPA 300.1 for anions in drinking water. Nonsuppressed conductometric detection can be used for all anions but is normally is used for weak acid anions and cations. 4.5.3 Amperometric/Coulometric Detection
Amperometric and coulometric detection can exploit the property of a compound to undergo either oxidation or reduction at an electrode to which a potential has been applied. The rate of the electrochemical reaction is observed as current, and, hence, these techniques fall under the heading of amperometry, or amperometric detection. The output from this type of electrochemical detector may be measured in either amperes or coulombs if the signal is integrated over time. The conversion efficiency, or the percent of analyte converted to product, is typically less than 5% [51]. If 100% of the analyte is oxidized or reduced to product, then the technique is referred to as coulometry. The quantity of analyte can be determined via Faraday’s Law: Q=nFN
87
88
4 Detectors
where Q is the number of coulombs, N is the number of moles converted to product, n is the number of electrons per reaction, and F is the Faraday constant. Since 100% of the analyte is consumed, the need for a standard can theoretically be avoided. Following separation by ion chromatography, the eluent flows into the electrochemical cell. As the eluent band of an analyte passes over the working electrode, it acts as an electron sink to either accept or donate electrons to the analyte should it be oxidized or reduced, respectively. The electroactivity of a compound is dependent on several factors, including molecular structure, accessibility of filled and unfilled orbitals, and functional groups present. Since the eluent is in motion over the electrode surface, diffusional mass transfer is assisted by forced convection to bring the analyte to the electrode surface. As a hydrodynamically controlled system, the limiting current (ilim) of the analyte can be described by the following equation; C ilim nFAD b d where A is the area of the electrode, D is the diffusion coefficient of the analyte, Cb is the concentration of the analyte, d is the thickness of the diffusion layer, and n and F are the same as for Faraday’s Law described above. In addition to sample concentration, the limiting current, or reaction rate, is dependent on applied potential and the physical and chemical properties of both the eluent and the electrode material. Table 4.6 shows the useful potential limits for some common electrodes under various pH conditions [52].
Table 4.6 Useful potential limits for common electrodes under
various pH conditions. Electrode material
Supporting electrolyte
Potential Limit (V) Negative
Positive
DV
1 M H2SO4
–0.3
+1.2
1.5
pH 7 buffer
–0.7
+1.0
1.7
1 M NaOH
–1.0
+0.6
1.6
Au
1 M NaOH
–0.9
+0.8
1.7
Hg
1 M H2SO4
–1.1
+0.3
1.4
1 M KCl
–1.9
+0.1
2.0
1 M NaOH
–2.0
–0.1
2.1
1 M HClO4
–0.2
+1.5
1.7
0.1 M KCl
–1.3
+1.0
2.3
Pt
C
4.5 Electrochemical Detectors
The simplest potential–time waveform that is applied to an electrode is that of a constant potential, which is known as dc amperometry. The high sensitivity and selectivity of ED is ideally suited for complex samples, as evinced by its application to the determination of neurotransmitters in complex biological samples (e.g., brain extracts). Neurotransmitters are typically aromatic compounds (e.g., phenols, aminophenols, catecholamines, and other metabolic amines), which are detected easily by anodic reactions at a constant (dc) applied potential at inert electrodes [53, 54]. Hypochlorite, ascorbate, hydrazine, arsenite, thiosulfate, nitrite, nitrate, cobalt and iron are a partial list of the ions that have been detected using amperometric detection [55]. The most common ions determined by amperometric detection in IC include inorganic anions forming complexes with cyanide, sulfide, and iodide at a silver electrode [56]. The relevant chemical reaction is shown below; Ag0 (the electrode) + X– → AgX(s) + ewhere, Ag is the electrode and X– is the halide ion. Silver is used as a sacrificial electrode. A sacrificial electrode is one where the electrode undergoes the redox reaction, and the electrode is consumed in the process. Other ions detected by this electrode include fluoride, chloride, and bromide. 4.5.4 Pulsed Electrochemical Detection (PED)
The most common electrode materials in PED electrochemical detection are Au, Pt, and carbon. Electronic resonance in aromatic molecules stabilizes freeradical intermediate products of anodic oxidations, and, as a consequence, the activation barrier for the electrochemical reaction is lowered significantly. Even for reversible redox couples that are considered to be well-behaved, dc amperometry is often accompanied by the practice of disassembling the electrochemical cell and mechanically polishing the working electrode ‘daily’. In this manner, fouling from nonspecific adsorption processes and/or mechanistic consequences is physically removed from the electrode surface. In contrast to aromatic moieties, absence of p-resonance for aliphatic compounds results in very low oxidation rates even though the reactions may be favored thermodynamically. Stabilization of free-radical products from aliphatic compounds can be achieved alternatively via their adsorption to the surface of noble metal electrodes. Unfortunately, adsorption of organic molecules and free radicals also has the consequence of fouling of the electrode and loss of its activity [57]. The historical perspective of nonreactivity for aliphatic compounds at noble metal electrodes can be attributed to surface fouling as a result of high, but transient, catalytic activity. An alternative approach is to combine electrochemical detection with ‘on-line’ cleaning. Hence, in order to maintain uniform and reproducible electrode activity at noble metal electrodes for polar aliphatic compounds,
89
90
4 Detectors
pulsed electrochemical detection (PED) was developed [58]. A detailed review of all aspects of PED has been published [59]. Although increased sensitivity and reproducibility has also been reported for pulsed potential cleaning at carbon electrodes by several researchers [60, 61], these electrodes have not generally been successful for the detection of polar aliphatic compounds. This effect is attributable to the absence of appropriate electrocatalytic properties of carbon surfaces to support the anodic oxygen-transfer reaction mechanisms of polar aliphatic compounds. Most frequently, PED is applied at Au electrodes under alkaline conditions (pH > 12). Figure 4.3 shows a cyclic voltammogram of (- - - - -) glucose at an Au electrode in (– – – –) deaerated 0.1 M NaOH using a Ag/AgCl reference electrode. All aldehydes, including ‘reducing sugars’ like glucose, are anodically detected during the positive potential excursion at the oxidefree surface in the region of ca. –0.6 to +0.2 V. Large anodic signals are obtained for alcohols, polyalcohols, and nonreducing sugars in the region of ca. –0.3 to +0.2 V (Mode I: oxide-free detections) with an attenuated signal for most compounds from ca. +0.2 to +0.6 V. Nitrogen- and sulfurcontaining compounds, for which a nonbonded electron pair is present, are adsorbed at oxidefree Au surfaces for E < ca. +0.1 V and can be
Figure 4.3 Cyclic voltammogram of glucose at gold electrode in 0.1 M NaOH. Conditions: 900 rpm rotation speed, 200 mV s–1 scan rate. Solutions (.....) aerated 0.1 M NaOH, (——) deaerated 0.1 M NaOH, and (- - - - -) 0.4 mM
glucose. (A) oxide formation; (B) O2 evolution; (C) oxide dissolution; (D) dissolved O2 reduction; (E) aldehyde oxidation; (F) hydroxyl oxidation.
4.5 Electrochemical Detectors
anodically detected by oxidecatalyzed (Mode II) reactions during the positive scan or E > ca. +0.1 V. Detections at E > ca. +0.8 V are not recommended because of the deleterious effects of evolution of O2. Although there is little or no individual compound selectivity, functional group selectivity is clearly evident. Electro inactive surface-adsorbing species can be detected by suppression of the oxide formation process (Mode III) at potentials > ca. +0.2 V. The presence of dissolved O2 can be a problem for detection in PED. Figure 4.3 also shows the signal (marked ‘D') for dissolved O2 in the supporting electrolyte. Gold electrodes are commonly used in PED because of the presence of a ‘background-free’ region from –0.1 V to +0.2 V in 0.1 M NaOH. Voltammetric resolution of complex mixtures is futile, since electrocatalyticbased detection of various members within a class of compounds is controlled primarily by the dependence of the catalytic surface state on the electrode potential rather than by the redox potentials (Eo) of the reactants. Therefore, general selectivity is achieved via chromatographic separation prior to electrocatalytic detection. This does not preclude some limited selectivity from control of detection parameters.
4.5.4.1 Pulsed Amperometric Detection (PAD) Oxidefree detections are often implemented with a quadruplepulse potential–time waveform (Table 4.7) at a frequency of ca. 2.0–0.5 Hz, which is appropriate for most IC applications to maintain chromatographic peak integrity. The detection potential (Edet) is chosen to be appropriate for the desired functional group, and the faradaic signal can be sampled during a short time (e.g., 16.7 ms) after a delay of t del. Typical values of tdel are in the range of 100 to 600 ms. Following the detection step, the electrode surface is subjected to three additional potential excursions: 1. Reductive cleaning by a large negative potential pulse, or Ered (–1.6 to –2.0 V) for a period of tred. In addition, any Au ions in the diffusion layer are reduced to elemental gold, which greatly extends the life of the electrode 2. Oxidative cleaning by a positive potential pulse, or Eoxd (+0.6 to +0.8 V) for a period of toxd 3. Reductive activation and adsorption by a large negative step to Eads (–0.8 to +0.1 V) for tads prior to the next detection cycle. This last step can also be used to increase adsorption, or preconcentration, of the analyte to the electrode surface.
Amperometric detection under the control of a simple potential-time waveform is known as pulsed amperometric detection (PAD), which is a subset of PED. Table 4.7 lists the typical quadruplepulse potential-time waveform used in PAD for carbohydrates in 0.1. M NaOH. Figure 4.4 shows the waveform of the applied potential as a function of time for PAD detection.
91
92
4 Detectors Table 4.7 Typical waveforms for (A) PAD and (B) IPAD. The applied potential vs. time waveforms corresponding to this table are shown in Figure 4.4.
IPAD
PAD Time (min)
Potential (V) vs. Ag/AgCl reference
Time (min)
Potential (V) vs. pH reference
0.00
+0.10
0.00
+0.33
0.20
+0.10
0.16
+0.33
0.17
+0.55
0.41
+0.55
0.42
+0.33
Parameter
Edet
0.40
+0.10
0.51
+0.33
0.41
–2.0
0.52
–1.67
0.42
–2.0
0.53
–1.67
Eoxd
0.43
+0.6
0.54
+0.93
Eads
0.44
–0.1
0.55
+0.13
0.50
–0.1
0.60
+0.13
Ered
Integration
Begin
End
Figure 4.4 The potential vs. time waveforms for PAD and IPAD. The times and potentials for the waveforms are listed in Table 4.7.
In addition to carbohydrates and alcohols, compounds which also have amine and/or sulfur moieties (e.g., aminoalcohols, aminosugars, thiosugars) rely on the detection of the –OH groups to take advantage of the simplicities of PAD waveforms for Mode I detections. The amine and/or sulfur moieties are exploited to increase the adsorption of the analyte to the electrode surface. Strong adsorption of reacting molecules is considered very beneficial because the residence time of
4.5 Electrochemical Detectors
the molecule on the electrode surface is increased substantially, thereby increasing the probability of a successful detection reaction. In the case of compounds containing only amine and/or sulfur groups, only Mode II detections are often available. Numerous organic and inorganic sulfur compounds are adsorbed at the oxidefree surfaces of Au and Pt electrodes and can be detected by Mode II [62]. These compounds include thioalcohols, thioethers, thiophenes, thiocarbamates, organic thiophosphates, and numerous inorganic compounds. Adsorption is a prerequisite to detection and therefore at least one nonbonded electron pair must reside on the S-atom. The kinetics for detection of adsorbed S-compounds are quite favorable at pHs from 0 to 14. Since alcohol and amine groups are detected only under highly acidic and/or alkaline conditions, the detection of sulfur compounds under mildly acidic conditions is highly selective.
4.5.4.2 Integrated Pulsed Amperometric Detection (IPAD) Anodic detection of amine and sulfurcontaining compounds occurs in a potential region where there is a significant signal for the concurrent formation of surface oxide. As a result, a large baseline signal is often encountered for amino acids and sulfur compounds (Mode II). Furthermore, the large baseline current is frequently observed to drift to large anodic values, especially for new or freshly polished electrodes. This drift is the consequence of a slow growth in the true electrode surface area as a result of surface reconstruction caused by the oxide on-off cycles in the applied multistep waveforms. Mode II detections performed with PAD are subject to a number of disadvantages due to the formation of surface oxide, which is required and concomitant with the detection of amine- and sulfurbased compounds. These disadvantages are as follows: . Baseline Sensitivities Any changes or gradients in pH, organic modifiers, ionic strength, or temperature may lead to baseline drifting. The baseline drift is attributable to variations in the extent of surface oxide formation. . Poor Signal-to-Noise Ratio – The sample current is only a fraction of the total signal. The majority of the signal for Mode II detections is derived from surface oxide formation. Oxideformation signal tends to be noisy. . Post-Peak ‘Dips’ The presence of the analyte at the electrode surface interferes with surface oxide formation. Hence, the background differs in the presence and absence of the analyte, which often results in a dip after the chromatographic peak.
The disadvantages listed above are either alleviated or greatly diminished by the use of Integrated Pulsed Electrochemical Detection (IPAD), which is also a form of PED. Figure 4.4 shows the IPAD waveform and Table 4.7 describes the times and potentials that are generally used for amino acid detection. Here, the electrode current is integrated electronically throughout a rapid cyclic or square wave
93
94
4 Detectors
scan of the detection potential (Edet) within the pulsed waveform. The potential excursion proceeds into oxide formation (positive scan or step) and back out of the oxide formation (negative scan or step) region of the oxidecatalyzed reaction for detection by Mode II. The anodic charge for oxide formed on the positive sweep tends to be compensated by the corresponding cathodic charge (opposite polarity) for dissolution of the oxide on the negative sweep. Hence, the ‘background’ signal on the electronic integration at the end of the detection period can be virtually zero and is relatively unaffected by the gradual change of electrode area. IPAD combines cyclic voltammetry with potential pulse cleaning to maintain uniform electrode activity.
4.5.4.3 IC–PED IC separations are controllable via changes in pH, ionic strength, and organic modifiers. For example, carbohydrates are readily separated based upon anion formation in alkaline media on anion exchange columns [63] using pH and salt gradients. The separation of amino acids using AAA-Direct™ technology from Dionex (Sunnyvale, CA) uses a complex quaternary gradient system (i.e., pH, salt, and organic modifier) on an anion exchange column to achieve baseline resolution of all the natural amino acids (see above). For a particular column type as described above, the following generalities about the mobile phase can be made to assist in the development of IC methods when using PED. . pH The effect of pH on the oxide formation process is attributable to the pHdependent nature of Au oxide formation, i.e., Au(H2O)ads → Au–OH + H+ + e–, and the potential for onset of oxide formation at Au electrodes shifts to more negative values with increases in pH at a rate of ca. –60 mV per pH unit. The negative shift in oxide formation with increasing pH can be reflected by a large baseline change in IC– PED under pHgradient elution when Edet remains constant throughout the gradient [64]. In isocratic separations, the effect of pH can be mitigated by changing the waveform potentials proper for a particular pH. This effect can also be alleviated to a great extent by substitution of a pHsensitive glassmembrane electrode for the Ag/AgCl reference electrode in the PED cell. Because the response of the glassmembrane electrode is ca. –60 mV per pH unit, the value of Edet is automatically adjusted during execution of pH gradients. . Ionic strength Under ionic strength conditions suitable for electrochemical detection (that is, l > 50 mM), the effect of changing ionic strength is reflected as minor perturbations in the background signal from oxide formation. This effect is not noticeable under isocratic IC conditions. Under gradient conditions (e.g., increasing acetate concentration), both positive and negative baseline drifts have been observed.
4.5 Electrochemical Detectors .
Organic modifier concentration In comparison to ionic strength effects, changes in the concentration of organic modifiers can have a much greater effect on the baseline signal in IC–PED. This can occur, even for electroinactive organic additives, because the modifiers are frequently adsorbed at the electrode surface with a resulting suppression of the oxide formation process (Mode III). In addition to alteration of the IC–PED baseline, adsorbed organic modifiers can severely attenuate the analytical signal for carbohydrates by interfering with access to specific adsorption sites on the electrode needed for the reaction to occur.
The current produced by analyte detection based on an oxide-catalyzed detection is accompanied by current from surface oxide formation. Consequently, variables (e.g., pH, organic modifier, and ionic strength) which affect the rates of oxide formation and dissolution will be reflected as drift in the baseline in IC–PED. For these, IPAD was designed to apply a waveform which coulometrically rejects the oxide background by summing the charges due to oxide formation and oxide dissolution which are expected to be of equivalent magnitude but opposite polarity. It is important to note that IPAD can virtually eliminate drift and changes associated with variations in composition (e.g., ionic strength and organic modifier) of the mobile phase, as well as changes in the total surface area of the noblemetal electrode surface. Changes in pH may require the simultaneous use of a pH reference electrode. 4.5.5 Post-column Derivatization
Sometimes called secondary electrochemical detection, this technique is useful when the sample species of interest is not electroactive. A reagent is mixed with the column effluent to transform the sample, M, into an electroactive species. An example is shown by Eq. (4.8): [Hg - DPTA]3– + Mn+ + 2e- → [M - DPTA](5–n)– + Hg
(4.8)
The DTPA (diethylenetriaminepentoacetate) metal complex is added and the sample metal ‘displaces’ the mercury ion. The measured current is from the reduction of mercury. There are many other schemes that can be applied to a variety of ions [65–67]. 4.5.6 Electrochemical Hardware and Detector Operation
Electrochemical detectors are constructed with three electrodes. The electrolysis of interest takes place at the working (marked W in Figure 4.5) electrode at a potential measured by the reference electrode. The auxiliary (or counter) electrode
95
96
4 Detectors
potential is controlled to maintain the reference potential. The reference electrode is usually saturated calomel or silver/silver chloride. Platinum or glassy carbon is generally used for the auxiliary electrode [68, 69]. Choice of the working electrode depends on the sample ions to be determined (Table 4.6). In general, silver is used for halides and pseudohalides. Glassy carbon is useful for oxides and organics. Platinum is probably the most general electrode. It is useful for the oxidation detection of inorganics and organics. Mercury, with its large negative over-potential, is useful for species that can be reduced. Gold is the least susceptible to interference from dissolved oxygen. Cells are classified according to how the working electrode is positioned relative to the flow stream. There are three major configurations: tubular, thin layer, and wall jet. The tubular cell (open or packed) with its greater working electrode surface area is used for coulometric detection (see Figure 4.5). The thin layer and wall jet designs are used for amperometric detector cells. In thin layer cells, the eluent flow is in the same plane as the working electrode surface. The wall jet cells are made exactly the same way except that the working electrode is positioned opposite the eluent inlet port. Thus, the eluent ‘jet’ flow is perpendicular to the wall of the working electrode surface. Generally, the wall-jet cell is more sensitive and self-cleaning compared to thin layer cells, but, it is also more flow-sensitive. Detector response depends on the eluent flow rate, pH, ionic strength, and temperature. In practice, the operating potential of the instrument should be set at the smallest possible value on the diffusion current plateau. Higher potentials do not increase detector response but do increase background noise resulting from the presence of trace contaminants in the eluent. Lower potentials sacrifice some response, but may improve selectivity. Interference may be a problem. Trace substances in the eluent or samples may change detector response. It should be remembered that oxygen can diffuse into the tubing and cause a signal.
Figure 4.5 The most common flow-through electrochemical cell configurations are as follows: (A) thin-layer, (B) wall-jet, (C) tubular, (D) porous. The direction of fluid flow is
denoted by the arrows. In addition to the working electrode, all electrochemical cells will typically incorporate a counter or auxiliary electrode and a reference electrode.
4.7 Evaporative Light Scattering Detector (ELSD)
4.6 Refractive Index Detector
Extensive work has been performed with refractive index (RI) detection for ion chromatography [70–72]. The refractive index detector can be considered to be a universal detector because any salt (or acid or base) added to water will cause a change in refractive index of the solvent. The differences in refractive index can be measured as the sample ions pass through the detector window replacing some of the eluent ions and changing the refractive index. Depending on the relative change, the peak may be increasing or decreasing. The detector has been used for both anion and cation separations with good results. The refractive index of solution increases with the molecular weight (and concentration) of the solute dissolved in the solution. Organic ions usually cause a greater change than inorganic ions in the refractive index of a solution. Refractometers can be used for indirect detection if the eluent ion is organic, for example, phthalate. Sample peaks of most sample ions would be seen as decreasing refractive index peaks. Direct RI detection may be possible when the sample ions are organic. The eluent in this case would probably be inorganic. Refractive index detection allows an extremely wide latitude in the selection of the eluent type, eluent pH and the ionic strength. In principle, refractive index detection can be substituted for conductance or UV absorption detection in many separations. However, in early work, refractive index detection was found to be only moderately sensitive and was considered to be somewhat interference-prone [73]. Minimum detectable quantities for common anions such as chloride, nitrate, or sulfate were reported to be in the 20 ng to 50 ng range (compared with 1 to 5 ng for direct conductance detection). The stability of RI detection has improved dramatically, and it should be considered seriously as an option for rugged, sensitive detection. This is due, at least partly, to control of the cell temperature, improved electronics and improved optical transducers. The new detectors are much less sensitive to variations in room temperature. The highest sensitivity levels can be reached only with careful control of the chromatographic temperature. Therefore, column ovens should be considered when using RI.
4.7 Evaporative Light Scattering Detector (ELSD)
The evaporative light scattering detector works by measuring the light scattered from the particles remaining after the eluent solvent has been nebulized and evaporated. The eluent solvent enters the detector and is evaporated in a heated device. As chromatographic peaks are eluted from the column and enter the detector the solvent is evaporated away. Conditions are selected so that the sample peak molecules become particulate. These particles enter a chamber into which a
97
98
4 Detectors
laser is directed, and when the peaks are present light is scattered and detected. ELSD can be quite sensitive with detection limits in the nanogram range. ELSD detectors are available from several companies including Eurosep, Varian, Shimaszu, Alltech, Sedere and Waters. The three components of the ELSD detection apparatus are the nebulizer, evaporation chamber and detection cell (Figure 4.6.)
Figure 4.6 Schematic of an evaporative light-scattering detector. Eluent and sample peaks enter the top of the cell and are nebulized. Solvent is evaporated away leaving particles of analyte sample peaks. A light is directed at the particles and scattered light is detected, amplified and measured.
4.7.1 Nebulizer
This first step transforms the liquid phase flowing from the IC column into fine droplets. There is an optimum size of the droplets. Small droplets of solvent are easier to evaporate in the next step. However, small droplets also form small particles when the peaks are eluted. It is desirable to have larger particles because they scatter more light and can be detected more easily. Gas flows are optimized to produce the optimum droplet size depending the properties on the eluent solvent and the analyte being detected. Usually the nebulization process cannot be performed at 100% efficiency; large droplets are formed and drain away. In general, the nebulization chamber incorporates a Venturi-type flow of gas around the eluent inflow. In some detectors, the nebulization chamber has a shape that is used to remove the biggest droplets of mist. In others chambers, all of the mist droplets are passed into the next part of the detection – the evaporation chamber.
4.7 Evaporative Light Scattering Detector (ELSD)
4.7.2 Evaporation Chamber
The droplets from the nebulizer are carried by the gas flow into the heated area located before the detection cell. The efficiency of evaporation depends on the shape of the tube and the temperature required by the eluent. Gas and temperature conditions are selected to completely remove the eluent mist while retaining the particles due to the sample peak. This can be an issue if the sample components are low boiling. The eluent solvent is completely removed to produce particles of sample solutes that contain no solvent. In practice, a temperature in the range 40–60 °C is sufficient to evaporate solvents used in HPLC of carbohydrates where high percentages of water or polar solvents are frequently used. 4.7.3 Detection Cell
The sample particles are carried by the gas to pass through a flow cell. The particles are hit with an incident light beam. A detection transducer is located an angle so that the incident light is not measured but can capture the scattered light. The amount of light scattered is proportional to the amount of sample present in the peak. In some instruments, a secondary gas inlet is used to direct the particles to the center of the detection chamber. In ion chromatography, the ELSD has been most commonly used in the detection of carbohydrates [74]. One example reported is where seven sugars and oligomers in a beer standard (fructose, glucose, sucrose, maltose, maltotriose, maltotetraose and maltopentose) were separated in 18 minutes by gradient elution with acetonitrile/water and with acetonitrile/acetone and water on a Prevail Carbohydrate ES, 5lm, 250 mm × 4.6 mm column and detected by ELSD detection [75]. This acetonitrile, acetone, and water eluent is ideal for this detector because it is completely volatile. Ammonium acetate may also be a useful eluent because of its high volatility. This eluent can provide both anions and cations as eluent ions. Sodium (or potassium or lithium) carbonate bicarbonate eluents can be suppressed and then sent through an ELSD. This eluent operates at a high pH so can be used to separate on an anion exchanger weak acid anions such as borate or several organic acids. While these anions cannot be detected by suppressed conductivity detection (because of low conductance) it may be possible to adjust detector conditions to detect many of these weak acid anions by ELSD. The quality of the eluent solvent is critical to achieve a low background signal. The amount of residue after evaporation must be less than 1 mg L–1. Solvents must be filtered through compatible sub-micron filters (0.4 or 0.2 lm). Also, if possible, samples must also be filtered with special filters before injection. Obviously, the salt or buffer in the eluent is also important because this can produce a background signal. The ideal eluent solvent for this detector contains no salts or buffers. Another possibility is to remove the eluent buffer with a suppressor device before detection. Of course this may limit the types of chromatography
99
100
4 Detectors
where the detector can be used. But the detector can be very useful, and very sensitive, in many cases where the eluent can be made to contain no particulates.
4.8 Other Detectors
Many other detectors have been used to monitor ion chromatography separations. Most of these detectors have been used only in special cases. Flame photometric detection [76, 77] has been used to detect alkali, alkaline earth, and some rare earth metals. Atomic absorption (AA) detectors [78–80] have been used for arsenite, arsenate, monomethyl arsenate, dimethyl arsinate, and p-aminophenoarsenate separations. Detectors of this type can be extremely sensitive, detecting arsenic down to 10 ng mL–1. Research into the development of environmental methods might consider the use of AA detection. Perhaps some of the more interesting detectors are those that use inductively coupled plasma (ICP) as an energy source and either atomic emission (AE) or mass spectrometry (MS) as the detector. ICP-AE and ICP-MS are well-developed analytical tools. One of the major advantages of these techniques is that a mixture of metals can be analyzed without the need for separation. Thus, workers who use these instruments normally do not think about their use as detectors. However, ICP-AE and ICP-MS cannot determine the oxidation or chemical state of a particular metal ion (Chapter 13). Some samples are quite important from a toxicological and environmental standpoint since the toxicity of a metal may depend on its oxidation state. For example, Cr(III) is not toxic (and even considered an essential element), but Cr(VI) is extremely toxic. An inductively coupled plasma atomic emission [81] detector was used to detect rare earth metals. References [1] I. Berglund and P. K. Dasgupta, Two-
[4] D. E. Johnson and C. G. Enke, Bipolar
dimensional conductometric detection in ion chromatography: post suppressor conversion of elute acids to a salt, Anal. Chem., 64, 3007, 1991. [2] I. Berglund and P. K. Dasgupta, Twodimensional conductometric detection in ion chromatography: postsuppressor conversion of elute acids to a base, Anal. Chem., 63, 2175, 1991. [3] I. Berglund, P. K. Dasgupta, J. Lopez, and O. Nara, Two-dimensional conductometric detection in ion chromatography: sequential suppressed and single column detection, Anal. Chem., 65, 1192, 1993.
pulse technique for fast conductance measurements, Anal. Chem., 42, 329, 1970. [5] K. J. Caserta, F. J. Holler, S. R. Crouch, and C. G. Enke, Computer controlled bipolar pulse conductivity system for applications in chemical rate determinations, Anal. Chem., 50, 1534, 1978. [6] K. Mayrhofer, A. J. Zemann, E. Schnell, and G. K. Bonn, Capillary electrophoresis and contactless conductivity detection of ions in narrow inner diameter capillaries, Anal. Chem., 71, 3828, 1999. [7] A. J. Zemann, E. Schnell, D. Volgger, and G. K. Bonn, Contactless conductiv-
References
[8]
[9]
[10]
[11]
[12]
[13]
[14]
[15] [16]
[17] [18]
[19]
ity detection for capillary electrophoresis, Anal. Chem. 70, 563, 1998. A. J. Zemann, K. Mayrhofer, E. Schnell, and G. K. Bonn. Contactless conductivity detector for capillary electrophoresis, Book of Abstracts, 216th ACS National Meeting, Boston, August 23-27, ANYL075, 1998. M. Macka, J. Hutchinson, A. Zemann, S. Zhang, P. R. Haddad, Miniaturized movable contactless conductivity detection cell for capillary electrophoresis, Electrophoresis 24, 2144, 2003. A. J. Zemann, Capacitively coupled contactless conductivity detection in capillary electrophoresis, Electrophoresis 24, 2125, 2003. A. Zemann, Conductivity detection with capillary electrophoresis and capillary chromatography, G.I.T. Laboratory Journal, Europe 7, 60, 2003. P. A. Gallagher, N. D. Danielson, Capillary electrophoresis of cationic and anionic surfactants with indirect conductivity detection, J. Chromatogr., 781, 533, 1997. R. J. Williams, Determination of inorganic anions by ion chromatography with ultraviolet absorbance detection, Anal. Chem., 55, 851, 1983. P. R. Haddad and N. E. Rochester, Ioninteraction reversed-phase chromatographic method for the determination of gold (I) cyanide in mine process liquors using automated sample preconcentration, J. Chromatogr., 439, 23, 1988. Waters Ion Brief No. 88112. P. R. Haddad and P. E. Jackson, The determination of ascorbate, bromate and metabisulfite in bread improvers using high performance ion-exchange chromatography, Food Tech. Aust., 37, 305, 1985. Wescan Application #164. D. R. Jenke, Quantitation of oxalate and citrate by ion chromatography with a buffered, strong acid eluent, J. Chromatogr., 437, 231, 1988. G. P. Ayers and R. W. Gillett, Sensitive detection of anions in ion chromatography using UV detection at wavelengths less than 200 nm, J. Chromatogr., 284, 510, 1984.
[20] L. Eek and N. Ferrer, Sensitive determi-
nation of nitrite and nitrate by ionexchange chromatography, J. Chromatogr., 322, 491, 1985. [21] T. E. Boothe, A. M. Emran, R. D. Hnn, P.J. Kothari and M.M.Vora, Chromatography of radiolabeled anions using reversed-phase liquid chromatographic columns, J. Chromatogr., 333, 269, 1985. [22] Dionex Application Note 51. [23] A.F. Geddes and J.G. Tarter, The ion chromatographic determination of chromium(III)-chromium(VI) using an EDTA eluant, Anal. Lett., 21, 857, 1988. [24] D.T. Gjerde and J.S. Fritz, Chromatographic separation of metal ions on macroreticular anion-exchange resins of a low capacity, J. Chromatogr., 188, 391, 1980. [25] R.L. Smith, Z Iskandarani and D. J. Pietrzyk, Comparison of reversed stationary phases for the chromatographic separation of inorganic analytes using hydrophobic ion mobile phase additives, J. Liq. Chromatogr., 7, 1935, 1984. [26] I. Molnar, H. Knauer and D. Wilk, High-performance liquid chromatography of ions, J. Chromatogr., 201, 225, 1980. [27] M. D. Seymour, J. P. Sickafoose, and J. S. Fritz, Application of forced-flow liquid chromatography to the determination of iron, Anal. Chem., 43, 1734, 1971. [28] M. D. Seymour and J. S. Fritz, Rapid, selective method for lead by forced-flow liquid chromatography, Anal. Chem., 45, 1632, 1973. [29] J. S. Fritz and L. Goodkin, Separation and determination of tin by liquid-solid chromatography, Anal. Chem., 46, 959, 1974. [30] L. Goodkin, M. D. Seymour, and J. S. Fritz, Ultraviolet spectra of metal ions in 6M hydrochloric acid, Talanta, 22, 245, 1975. [31] B. Grigorova, S.A. Wright and M. Josephson, Separation and determination of stable metallo-cyanide complexes in metallurgical plant solutions and effluents by reversed-phase ion-pair chromatography, J. Chromatogr., 410, 19, 1987.
101
102
4 Detectors [32] D.F. Hilton and P.R. Haddad, Determi-
[43] R. A. Cochrane and D. E Hillman, Anal-
nation of metal-cyano complexes by reversed-phase ion-interaction high-performance liquid chromatography and its application to the analysis of precious metals in gold processing solutions, J. Chromatogr., 361, 141, 1986. [33] S. Matsushita, Simultaneous determination of anions and metal cations by single-column ion chromatography with ethylenediaminetetraacetate as eluent and conductivity and ultraviolet detection, J. Chromatogr., 312, 327, 1984. [34] J. Osterloh and D. Goldfield, Determination of nitrate and nitrite ions in human plasma by ion-exchange highperformance liquid chromatography, J. Liq. Chromatogr., 7, 753, 1984. [35] Waters Ion Brief No. 88111. [36] M. Marno, N. Hirayama, H. Wada and T. Kuwamoto, Separation and determination of organoarsenic compounds with a microbore column and ultraviolet detection, J. Chromatogr., 466, 379, 1989. [37] Y. Michigami, T. Takahashi, F. He, Y. Yamamoto and K. Ueda, Determination of thiocyanate in human serum by ion chromatography, Analyst (London), 113, 389, 1988. [38] A. Mangia and M.T. Lugari, Separation and determination of inorganic anions by means of ion-pair chromatography, Anal. Chim. Acta, 159, 349, 1984. [39] P.E. Jackson, P.R. Haddad and S. Dilli, Determination of nitrate and nitrite in cured meats using high-performance liquid chromatography, J. Chromatogr., 295, 471, 1984. [40] R. L. Cunico and T. Schalbach, Comparison of ninhydrin and o-phthalaldehyde post-column detection techniques for high-performance liquid chromatography of free amino acids, J. Chromatogr., 266, 461, 1983. [41] M. Denkert, L. HacEzell, G. Schill, and E. Sjogren, Reversed-phase ion-pair chromatography with UV-absorbing ions in the mobile phase, J. Chromatogr., 218, 31, 1981. [42] H. Small and T. E. Miller, Jr., Indirect photometric chromatography, Anal. Chem., 54, 462, 1982.
ysis of anions by ion chromatography using ultraviolet detection, J. Chromatogr., 241, 392, 1982. [44] J. P. Sickafoose, Ph. D. Dissertation, lowa State University Ames, lowa (1971). [45] J. S. Fritz and J. N. Story, Chromatographic separation of metal ions on low capacity, macroreticular resins, Anal. Chem., 46, 825, 1974. [46] S. Elchak and R. M. Cassidy, Separation of the lanthanides on high-efficiency bonded phases and conventional ionexchange resins, Anal. Chem., 51, 1434, 1979. [47] C. H. Knight, R. M. Cassidy, B. M. Recoskie, and L. W. Green, Dynamic ion exchange chromatography for determination of number of fissions in thorium-uranium dioxide fuels, Anal. Chem., 56, 474, 1984. [48] T. Imanari, S. Tanabe, T. Toida, and T. Kawanishi, High-performance liquid chromatography of inorganic anions using iron(3+) as a detection reagent, J. Chromatogr., 250, 55, 1982. [49] A. W. Fitchett and A. Woodruff, Determination of polyvalent anions by ion chromatography, L. C., 1, 48, 1983. [50] Hershcovitz, H., Yarnitsky, Ch., and Schmuckler, G., Ion chromatography with potentiometric detection, J. Chromatogr., 252, 113, 1982. [51] LaCourse, W. R., Pulsed Electrochemical Detection in High Performance Liquid Chromatography, John Wiley & Sons, New York, 1997, 61. [52] LaCourse, W. R., Pulsed Electrochemical Detection in High Performance Liquid Chromatography, John Wiley & Sons, New York, 1997. [53] Adams, R. N., Electrochemistry at Solid Electrodes. Marcel Dekker, New York, 1969. [54] Kissinger, P. T. Laboratory Techniques in Electroanalytical Chemistry, Kissinger, P. T. and Heineman, W. R. (editors), Marcel Dekker, New York, 1984. [55] Fritz, J.S. and Gjerde, D.T., Ion Chromatography, 3rd ed., Wiley-VCH, New York, 2000, 72.
References [56] LaCourse, W.R., Pulsed Electrochemical
[72] P. R. Haddad and A. L. Heckenberg,
Detection in High Performance Liquid Chromatography, John Wiley & Sons, New York, 1997, 70. [57] Gilman, S., Electroanalytical Chemistry, vol. 2, Bard, A. J. (editor), Marcel Dekker, New York, 1967. [58] Hughes, S., Meschi, P. L. and Johnson, D. C. 1981, Anal. Chim. Acta 132, 11. [59] LaCourse, W. R., Pulsed Electrochemical Detection in High Performance Liquid Chromatography, John Wiley & Sons, New York, 1997. [60] Ewing, A. G, Dayton, M. A. and Wightman, R. M., Anal. Chem. 53, 1842, 1981. [61] Tengyl, J. 1984. In: Ryan, T. H. (Editor), Electrochemical Detectors, Plenum Press, New York and London. [62] Vandeberg, P. G, Kowagoe, J. L., and Johnson, D. C., Anal. Chim. Acta. 260(1), 1, 1992. [63] Paskach, T. J, Lieker, P. J. and Thielecke, K., Carbohydr. Res. 215, 1, 1991. [64] LaCourse, W. R, Jackson, W. A. and Johnson D.C., Anal. Chem., 61:22, 2466, 1989. [65] Y. Takata and G. Muto, Flow coulometric detector for liquid chromatography, Anal. Chem., 45, 1864, 1973. [66] J. E. Girard, Ion chromatography with coulometric detection for the determination of inorganic ions, Anal. Chem., 51, 836, 1979. [67] I. S. Krull and W. R. LaCourse, PostColumn Photochemical Derivatizations, in: PCR Detectors for HPLC, vol. 37, Marcel Dekker: New York, 1986. [68] R. J. Rucki, Electrochemical detectors for flowing liquid systems, Talanta, 27, 147, 1980. [69] Dionex Corporation, Electrochemical Detector, Sunnyvale, CA, February, 1981. [70] Chrompack, Inc., Bridgewater, New Jersey, Chrompack Topics, Vol. 8 (1981). [71] F. A. Buytenhuys, Ion chromatography of inorganic and organic ionic species using refractive index detection, J. Chromatogr., 218, 57, 1981.
High-performance liquid chromatography of inorganic and organic ions using low-capacity ion-exchange columns with indirect refractive index detection, J. Chromatogr., 252, 177, 1982. [73] T. Jupille, UV-visible absorption derivatization in liquid chromatography, J. Chromatogr. Sci., 17, 160, 1979. [74] M. Lafosse, M. Dreux, L. Morin-Allory, J. M. Colin, Some applications of a commercial light scattering detector for liquid chromatography, J. High Resolution Chromatogr., 8, 39, 2005. [75] Alltech Application Note 0054E The use of acetone in the mobile phase to speed up run times for carbohydrates in beer by reversed phase HPLC and evaporative light scattering detection is described, March 29, 2004. [76] D. J. Freed, Flame photometric detector for liquid chromatography, Anal. Chem., 47, 186, 1975. [77] S. W. Downey and G. M. Hieftje, Replacement ion chromatography with flame photometric detection, Anal. Chim. Acta, 153, 1, 1983. [78] E. A. Woolson and N. Aharonson, Separation and detection of arsenical pesticide residues and some of their metabolites by high pressure liquid chromatography-graphite furnace atomic absorption spectrometry, J. Assoc. Off. Anal. Chem., 63, 523, 1980. [79] G. R. Ricci, L. S. Shepard, G. Colovos, and N. E. Hester, Ion chromatography with atomic absorption spectrometric detection for determination of organic and inorganic arsenic species, Anal. Chem., 53, 610, 1981. [80] A. A. Grabinski, Determination of arsenic(III), arsenic(V), monomethylarsonate, and dimethylarsinate by ionexchange chromatography with flameless atomic absorption spectrometric detection, Anal. Chem., 53, 966, 1981. [81] K. Yoshida and H. Haragachi, Determination of rare earth elements by liquid chromatography/inductively coupled plasma atomic emission, Anal. Chem., 56, 1984.
103
105
5 Principles of Ion Chromatographic Separations 5.1 General Considerations
The first step in performing an ion-chromatographic separation is to select an appropriate column, eluent and detector. The eluent and detector must of course be compatible. For example, a UV–Vis detector would be appropriate for separation of anions that absorb in the UV spectral region (Br–, I–, SCN–, aromatic anions, etc.) but a nonsorbing eluting ion (OH–, SO42–, ClO4–, etc.) must be selected. The expected difficulty of a separation also needs to be considered. For an easy separation, a shorter column may be appropriate to obtain a faster analysis. Sample ions that are strongly retained by the column may require a more concentrated eluent in order to avoid overly long retention times. A complex sample may require a programmed increase in eluent concentration (gradient elution) to adequately resolve the sample peaks. Information that accompanies commercial IC instruments will often provide instructions for appropriate conditions to start with. Modern instruments are, quite properly we think, designed to be as user-friendly as possible. By following recommended protocols, it is relatively easy to perform a simple separation. However, it is much better to have a real understanding of the factors that affect chromatographic efficiency and selectivity. It is also personally satisfying to devise a method to solve a challenging analytical problem. The goal of this chapter is to provide information on fundamental aspects of IC.
5.2 Chromatographic Terms
Figure 5.1 shows a generalized chromatogram that might be obtained for the separation of two sample ions. The following terms are commonly used: to, called the dead time, is the time in minutes or seconds for a nonsorbed marker to pass through the column and detector.
Ion Chromatography, 4th Ed. James S. Fritz and Douglas T. Gjerde Copyright © 2009 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim ISBN: 978-3-527-32052-3
106
5 Principles of Ion Chromatographic Separations
t1 and t2 are the retention times for the sample ion peaks (t = to + t in the figure). t′1 and t′2 are the adjusted retention times. w1 and w2 are the peak widths at the base of the peaks as shown in the figure. k′ (or simply k) is the retention factor, formerly called the capacity factor. k
t to to
(5.1)
The separation factor, a is a
k2 t2 k1 t1
to to
(5.2)
The magnitude of a is an indication of the difficulty of a given separation. t2 t1 Peak resolution, Rs. Rs w . Here w is the average peak width at the base. For Gaussian peak, w = 4r, where r is the average standard deviation. At Rs = 1.0 there is some overlap of two peaks at their base (2.3% of the area overlaps). At Rs = 1.25, only 6% of peak areas overlap. At Rs = 1.5, overlap is only 0.13% of peak areas and ‘baseline resolution’ is achieved.
Figure 5.1 Some common terms used in chromatography.
5.2.1 Retention Factor
In modern chromatography it is more convenient to use the capacity factor (k or k′) instead of the distribution coefficient. The name retention factor has been suggested to replace the name capacity factor. The definition for capacity factor or retention factor is the same:
5.2 Chromatographic Terms
k
amount of analyte in column stationary phase amount of analyte in column mobile phase
(5.3)
The most convenient way to calculate k is from the retention time (tR) of a sample ion and from the dead time (to or tM), measured from a chromatogram. The value of to (also called tM) is the time for a nonretained substance to pass through the chromatographic column and detector. The substance chosen can be either a detectable ion or molecule, but care must be taken to choose a substance that passes through the column with absolutely no retention. Once tR and to have been measured, the k for that ion can be calculated from the equation: k
tR
to =to
(5.4)
Alternatively, the retention volume (VR) and the dead volume (Vo) can be used. k VR
Vo =Vo
(5.5)
The retention factor is a useful constant for any given analyte and chromatographic system, because it does not vary with column length or flow rate. For example, as column length is increased, to and tR will both increase, but k′ does not change. The retention time for any analyte follows the simple relationship: t = to (1 + k)
(5.6)
The actual basis of any chromatographic separation is that the various analytes have differing distribution constants between the stationary and mobile phases. This causes the analytes to move through the column at different rates and to appear as time-separated peaks on the final chromatogram provided there is a reasonable difference in migration rates and the column is of sufficient length. In ion chromatography the distribution constant of a sample ion and the eluent ion represents a competition by these two ions for exchange sites on the solid ion exchanger. The migration rate of S will depend on the magnitude of the distribution constant. If we represent the eluent ion by E, a sample ion by S, and R represents the solid anion or cation resin, the equilibrium becomes R-E + S > R-S + E
(5.7)
R-SE SR-E
(5.8)
K
The retention factor is defined as the ratio of the sample ion in the resin phase to that in the liquid phase. k = [R-S]/[S]
(5.9)
107
108
5 Principles of Ion Chromatographic Separations
The amount of E in the resin phase will greatly exceed that in the liquid phase because so little sample is used. Thus, we can represent [R-E] as C, the exchange capacity of the resin in millimoles g–1. Making these substitutions in Eq. (5.8), K
kE C
(5.10)
k
KC E
(5.11)
In Eq. (5.11), C and E both have the concentration units mequiv g–1. Converting to logarithmic terms log k = –log E + log k + log C
(5.12)
When the exchange capacity, C, of the ion exchanger is fixed, the last two terms of Eq. (5.12) can be combined into a new constant which we will call the intercept. log k = –log E + Intercept
(5.13)
This last equation is very fundamental for ion chromatography. It predicts a linear decrease in log k as log E is increased. At fixed eluent concentration, the effect of resin capacity on log k may also be predicted. The preceding equations are derived for equilibria where the two ions S and E both have the same charge. A plot of log k against log [E] will have a slope of –1. If S is divalent, such as sulfate, and E has a 1– charge, the slope of the plot will be –2. The predicted slope will be –1/2 if S is 1– and E is 2–. The retention time of a sample ion, t, is linked to its retention factor, k, by the simple relationship, t = to(1+k)
(5.14)
where to is the dead time. Conditions need to be adjusted so that the sample ions are separated in a reasonable time period. Resolution of the sample components is unlikely if the retention factors are too small but an overly long separation will result if they are too large. If possible, conditions should be selected such that the k values fall within a range of about 2–12. The best way to do this is to adjust the eluent concentration either up or down using Eq. (5.13). Further adjustment of the actual retention times is possible by manipulating the value of to. Thus a larger to is obtained by using a longer column or by using a slower flow rate. The process of adjusting eluent concentration is perhaps best illustrated by an example. Suppose we have three sample ions with kC values of 40, 60 and 100. The retention factors, calculated by substitution into Eq. (5.13), are plotted as a function of eluent concentration in Figure 5.2. If to = 0.8 min, the retention times of the three ions are calculated to be as follows:
5.3 Selectivity
[E] = 5 mM : 7.2, 8.8, 16.8 min [E] = 15 mM : 2.2, 4.0, 6.1 min [E] = 30 mM : 1.8, 2.4, 3.4 min At [E] = 5.5 mM, the peaks will be well resolved, but the separation is slow. Peak retention times at [E] = 30 mM are very fast but all of the peaks may not be resolved, especially if one of the sample ions is present in a much higher concentration and therefore will give a broad peak. An intermediate eluent concentration of 15 mM gives a complete separation in just over 6 min and each of the peaks is separated from the others by approximately 2 min.
Figure 5.2 Retention factor, k, as a function of eluent concentration.
5.3 Selectivity
Selectivity in ion-exchange chromatography may be defined as the relative ability of sample ions to form an ion pair with sites of the opposite charge within the stationary phase. In cation exchange, sample cations pair up with negative sulfonate or carboxylate groups within the resin. Sample anions are generally considered to be excluded from the resin by a wall of negative sulfonate, carboxylate, or phosphonate groups that act as an electrostatic barrier. Similar considerations apply to anion-exchange resins where sample anions pair up with positively charged quaternary ammonium sites within the resin. A considerable amount of information on ion-exchange selectivity is available in the older literature. This is often in the form of elution orders of various ions or in terms of selectivity coefficients.
109
110
5 Principles of Ion Chromatographic Separations
5.3.1 Selectivity Coefficients
A selectivity coefficient measures the relative affinity of two ions, A and B, for the ion-exchange resin. For monovalent ions, the following equilibrium takes place: As + Br > Ar + Bs
(5.15)
Here s denotes the solution phase and r the resin phase. The equilibrium constant, which is called the selectivity coefficient is as follows: KBA
Ar Bs Br Ss
(5.16)
The reference ion, B, is usually H+ for cations and Cl– for anions. Unlike the equilibrium constant k in Eq. (5), which is measured by column experiments, the selectivity coefficient is generally measured by static equilibrium of a fixed amount of resin in the B form with a fixed volume of A in a solution of known concentration. The brackets indicate the ion concentration in mmol mL–1 for the solution phase and in mmol g–1 for the resin phase. For accurate calculations of equilibrium constants, we should really use the activities (to calculate the effective concentration) of the ions instead of their concentrations. However, activity coefficients are sometimes difficult to measure, especially within the ion-exchange resin matrix. In ion chromatography, ionic concentrations are often low and ion activities approach unity. For the sake of simplicity, only ion concentrations are used in this discussion. For the selectivity coefficient for a 2+ ion, such as Cu2+, compared to a 1+ ion, + H , the equilibrium will be 2 H+r + Cu2+ > Cu2+r + 2 H+s
(5.17)
The selectivity coefficient now has two squared terms: 2
KBA
Cu2 r H s 2 H r Cu2 s
(5.18)
Tables 5.1 and 5.2 show selectivity coefficients for several cations on AG 50WX8 and for some anions on Dowex 1X8, respectively.
5.3 Selectivity Table 5.1 Selectivity coefficients, KAB, for cations on AG 50WX8
cation-exchange resin. Here, B represents H+.
Cation
Selectivity coefficient
H+
1.0
+
0.9
Li
Na
+ 4+
1.5
NH
2.0
K+
2.5
Rb+
2.6
Cs+
2.7
+
5.3
+
7.6
Cu Ag
2+
Mn
2.4
Mg2+
2.5
Fe2+
2.6
2
2.7
2+
2.8
2+
2.9
2+
3.0
2+
Ni
3.0
Ca2+
3.9
Sr2+
5.0
Zn Co
Cu Cd
2+
Hg
7.2
2+
7.5
2+
8.7
Pb Ba
These data are for classical resins of high exchange capacity. Selectivity coefficients for the resins used in modern ion chromatography are apt to be significantly different. Nevertheless, the values in Tables 5.1 and 5.2 do give some indication of the relative affinity of the resins for various ions. Selectivity coefficients can be used to estimate the effectiveness of different ions as eluents in ion chromatography. Ions with high selectivity coefficients usually make the most efficient eluents. This means that a relatively dilute solution of the eluting ion can be used. Sometimes an eluent will elute the sample
111
112
5 Principles of Ion Chromatographic Separations
ions too quickly, even after further dilution of the eluent. In some cases it may be necessary to choose an eluent ion that has a lower selectivity coefficient. As a general rule, divalent eluent ions are best for separating divalent sample ions and monovalent eluents are for separating monovalent sample ions.
Table 5.2 Selectivity coefficients, KAB, for anions on Dowex 1×8.
Here, B represents Cl–.
Anion
Selectivity coefficient
Salicylate
32.2
Iodide
8.7
Phenoxide
5.2
Bisulfate
4.1
Nitrate
3.8
Bromide
2.8
Cyanide
1.6
Bisulfite
1.3
Nitrite
1.2
Chloride
1.00
Bicarbonate
0.32
Dihydrogen phosphate
0.25
Formate
0.22
Acetate
0.17
Aminoacetate
0.10
Fluoride
0.09
Hydroxide
0.09
5.3.2 Other Ion-exchange Interactions
In addition to true ion exchange, other interactions can take place between the sample solutes and the resin. Adsorption is one of the commonest of these interactions. For example, the benzoate anion appears to be adsorbed somewhat by the polystyrene–divinylbenzene matrix of organic ion exchangers. This may be due to an attraction of the p electrons of the aromatic polymer for the benzoate. Benzoic
5.3 Selectivity
acid, which exists mostly in the molecular form, is absorbed to a much greater degree than benzoate salts. Adsorption of eluent and sample ions by an ion-exchange resin appears to be quite common, although the degree to which adsorption occurs varies tremendously. Large organic ions and some of the larger inorganic ions seem to be appreciably adsorbed, while adsorption of small, polar ions may be negligible. Much of the selectivity in separating mixtures of organic cations and anions has been shown to come from differences in adsorption rather than from differences in ion-exchange selectivity [1, 2]. However, adsorption and subsequent desorption is apt to be a slower process than ion exchange and is therefore to be avoided as much as possible in ion chromatography. 5.3.3 Selectivity of Sulfonated Cation-exchange Resin for Metal Cations
The selectivity of sulfonated ion-exchange resins for metal cations is often expressed qualitatively in terms of elution orders. Numerical selectivity data for cations are limited [3–5]. Strelow and coworkers [6–8] published comprehensive lists of distribution coefficients for metal ions with perchloric acid and other mineral acid eluents. However, their data are for sulfonated gel resins of high exchange capacity. Sevenich and Fritz [9] published a comprehensive study of metal cation selectivity with resins of a more modern type. The studies were made on a column packed with a 12% cross-linked polystyrene divinylbenzene resin 12–15 mm in diameter and with an exchange capacity of 6.1 lequiv g–1. The resins were prepared by rapid sulfonation so that the sulfonic acid groups were concentrated on the outer perimeter of the resin beads [10]. The approach taken was to measure retention times of metal cations on columns containing low-capacity resins with eluents containing perchloric acid or various perchlorate salts. The perchlorate anion was used to eliminate any possible complexing of a metal ion by the eluent anion. Retention factors (k) were calculated from the retention data. Eluents containing hydrogen ions or sodium ions were used at concentrations ranging from 0.10 to 1.0 M. In general, narrow chromatographic peaks were obtained, although the rare earth peaks were broad with some tailing. The retention factors were calculated from the retention times and are given in Tables 5.3 and 5.4. The data show that, as expected, eluents containing sodium(I) are more efficient than those containing hydrogen(I). The data in these tables are arranged in order of increasing capacity factors. In this way it is possible to compare the relative affinities of the various divalent and trivalent metal ions for resin sites. Note that there are crossovers in the retention factors of lead(II) and several trivalent metal ions as the eluent concentration is increased.
113
114
5 Principles of Ion Chromatographic Separations Table 5.3 Retention factors (k) for various cations when
perchloric acid eluents are used. Cation
Perchloric acid concentration (M) 1.00
0.85
0.75
0.60
0.50
0.40
0.25
0.10
V(IV)
0.08
0.10
0.12
0.16
0.28
0.40
0.91
5.23
Be(II)
0.10
0.18
0.24
0.30
0.38
–
1.86
4.59
Zr(IV)
0.14
0.20
0.24
0.38
0.65
0.95
2.00
13.31
Mg(II)
0.16
0.18
0.26
0.36
0.51
0.67
1.39
7.49
Fe(II)
0.16
0.18
0.26
0.38
0.53
0.85
1.94
9.19
Mn(II)
0.16
0.18
0.26
0.42
0.55
0.95
2.02
12.14
Hg(II)
0.16
0.18
0.30
0.42
0.69
0.99
2.99
11.70
Ni(II)
0.16
0.22
0.28
0.38
0.59
0.89
2.00
12.75
Zn(II)
0.12
0.18
0.30
0.46
0.71
0.97
2.55
12.67
Cu(II)
0.16
0.22
0.32
0.52
0.61
0.97
3.19
12.83
U(VI)
0.16
0.22
0.28
0.40
0.67
0.99
3.52
12.83
Co(II)
0.20
0.22
0.28
0.38
0.48
1.05
2.02
13.56
Cd(II)
0.22
0.28
0.32
0.59
0.77
1.39
3.56
25.25
Ca(II)
0.32
0.40
0.57
0.97
1.47
2.26
5.98
21.41
Sr(II)
0.44
0.66
0.81
1.39
2.06
3.09
8.55
35.35
Ba(II)
1.01
1.29
1.52
2.73
3.94
6.63
15.76
–
Pb(II)
1.21
1.64
1.90
2.99
4.75
7.37
28.48
100.0
Al(III)
0.68
1.01
1.60
2.95
5.27
9.90
43.0
–
Fe(III)
1.07
1.68
2.89
5.54
9.98
18.42
80.0
–
V(III)
1.13
2.00
2.89
4.67
10.10
20.08
–
–
In(III)
1.15
2.12
3.05
5.84
10.59
19.84
63.8
–
Lu(III)
2.67
4.71
6.57
13.07
27.1
50.5
–
–
Yb(III)
2.93
5.15
7.17
14.14
28.7
52.5
–
–
Tm(III)
2.97
5.25
7.54
14.75
28.9
57.0
–
–
Y(III)
3.13
5.94
8.18
16.38
30.1
59.6
–
–
Er(III)
3.13
5.31
7.92
15.72
30.5
59.8
–
–
Ho(III)
3.37
5.62
8.59
16.71
32.9
63.4
–
–
Dy(III)
3.78
6.10
9.25
18.61
35.8
70.9
–
–
5.3 Selectivity Table 5.3 (Continue) Retention factors (k) for various cations
when perchloric acid eluents are used. Cation
Perchloric acid concentration (M)
Gd(III)
4.89
8.40
12.48
24.65
47.7
91.3
–
–
Eu(III)
5.45
9.25
13.84
27.47
52.9
105.0
–
–
Sm((III)
6.04
11.37
15.43
30.30
59.4
116.0
–
–
Nd(III)
7.13
13.64
17.76
35.35
69.3
137.8
–
–
Pr(III)
7.44
14.22
18.57
37.58
71.3
142.2
–
–
Ce(III)
8.69
14.44
21.2
43.2
80.8
167.7
–
–
La(III)
9.68
15.64
23.8
47.9
93.7
187.7
–
–
Bi(III)
8.75
13.1
21.2
31.9
53.7
135.0
–
–
Table 5.4 Retention factors (k) of various cations when sodium
perchlorate eluents are used. Cation
Sodium perchlorate concentration (M) 1.00
0.75
0.50
0.25
0.20
Be(II)
0.06
0.06
0.12
–
1.05
Mg(II)
0.06
0.10
0.18
0.73
1.23
Ni(II)
0.06
0.14
0.26
1.09
2.42
Fe(II)
0.08
0.14
0.32
1.15
3.58
Co(II)
0.08
0.16
0.20
1.25
2.79
Zn(II)
0.10
0.12
0.30
1.03
2.20
Hg(II)
0.10
0.18
0.36
1.11
2.95
Zr(IV)
0.12
–
0.30
1.29
2.71
V(IV)
0.14
0.16
1.37
0.95
2.40
Mn(II)
0.16
0.20
6.26
1.21
3.45
Cd(II)
0.16
0.22
0.67
1.49
3.23
Ca(II)
0.18
0.30
0.48
2.06
4.36
Cu(II)
0.20
0.30
0.46
1.31
2.79
Sr(II)
0.28
0.51
0.97
3.62
5.92
U(VI)
0.61
0.87
1.27
3.54
7.25
115
116
5 Principles of Ion Chromatographic Separations Table 5.4 (Continue) Retention factors (k) of various cations
when sodium perchlorate eluents are used. Cation
Sodium perchlorate concentration (M)
Ba(II)
0.75
1.21
2.26
4.46
–
Al(III)
–
0.75
2.32
–
–
Pb(II)
1.15
1.51
3.03
–
27.7
Lu(III)
1.31
2.63
6.91
44.24
–
Yb(III)
1.37
2.87
7.29
62.6
–
Tm(III)
1.56
3.03
7.92
64.8
–
Y(III)
1.54
2.95
8.28
65.0
–
Er(III)
1.62
3.03
8.34
67.1
–
Ho(III)
1.66
3.31
8.95
67.9
–
Dy(III)
1.94
3.98
10.85
86.1
–
Tb(III)
2.19
4.12
11.33
88.9
–
Gd(III)
2.51
4.89
13.19
104.0
–
Eu(III)
2.79
5.49
14.71
116.2
–
Sm(III)
3.29
6.00
16.55
123.6
–
Nd(III)
3.13
6.24
16.85
131.7
–
Pr(III)
3.21
6.20
17.92
138.6
–
Ce(III)
3.33
6.83
20.40
147.7
–
La(III)
3.80
7.86
21.42
173.5
–
In(III)
3.33
5.01
–
–
–
The elution order of metal ions reported in Table 5.4 is similar to that reported earlier by Strelow and Sondrop [8] with perchloric acid eluent and gel resins of high exchange capacity. Strelow and coworkers noted anion-complexing effects on elution orders in several cases when eluent acids other than perchloric acid were used. Eq. (5.13) predicts that a plot of k against the activity of the eluent ion should be a straight line. If the eluent ion is monovalent, the slope should be the same as the charge on the metal ion, M. The data in Table 5.3 (perchloric acid eluent) and Table 5.4 (sodium perchlorate eluent) were plotted according to Eq. (5.13). Linear plots were obtained in all cases. The slopes, obtained by linear regression, are given in Table 5.5 for perchloric acid, and similar results were obtained for sodium perchlorate. In most cases, the negative slope was very close to the charge on the metal ion. The somewhat
5.3 Selectivity
low values obtained for Be(II) and Bi(III) in Table 5.5 may be due to partial hydrolysis of the metal cation. The slopes obtained for vanadium(IV) and zirconium(IV) indicate that the metal ions are present as VO2+ and ZrO2+, respectively.
Table 5.5 Linear regression data for log k plotted against log
[HClO4]. Cation
Slope (this work)
c (Ref.*)
Be(II)
–1.64
–1.02
–0.990
Mg(II)
–1.66
–1.25
–0.998
Ca(II)
–1.87
–1.54
–0.996
Sr(II)
–1.91
–1.66
–0.998
Ba(II)
–1.97
–1.92
–0.997
Mn(II)
–1.89
–1.36
–0.998
Zn(II)
–1.98
–1.32
–0.998
Ni(II)
–1.87
–1.35
–0.998
U(VI)
–1.96
–1.32
–0.996
Cu(II)
–1.92
–1.33
–0.997
Fe(II)
–1.79
–1.36
–0.999
Co(II)
–1.86
–1.34
–0.992
Cd(II)
–2.08
–1.40
–0.997
V(IV)
–1.83
–1.11
–0.997
Pb(II)
–2.00
–1.88
–0.994
Hg(II)
–1.92
–1.76
–0.997
Zr(IV)
–1.95
–
–0.999
Al(III)
–2.88
–1.80
–0.999
Bi(III)
–2.67
–2.24
–0.992
Fe(III)
–2.99
–1.81
–1.000
In(III)
–2.78
–l.87
–0.999
V(III)
–2.87
–
–0.996
Lu(III)
–3.00
–
–0.998
Yb(III)
–2.95
–
–0.998
Tm(III)
–2.99
–
–0.999
117
118
5 Principles of Ion Chromatographic Separations Table 5.5 (Continue) Linear regression data for log k plotted
against log [HClO4]. Cation
Slope (this work)
c (Ref.*)
Y(III)
–2.95
–2.20
–0.999
Er(III)
–3.00
–
–0.999
Ho(III)
–3.01
–
–0.999
Dy(III)
–3.01
–2.23
–0.999
Tb(III)
–3.04
–
–0.998
Gd(III)
–2.97
–
–0.999
Eu(III)
–3.01
–
–0.999
Sm(III)
–2.97
–
–0.999
Nd(III)
–2.97
–
–0.998
Pr(III)
–2.95
–
–0.998
Ce(III)
–3.01
–2.48
–0.999
La(III)
–3.04
–2.49
–0.999
* Calculated from data of Strelow and Sondorp [8].
The necessity for using the activity of the eluent should be emphasized. When the concentration of H+ (in HClO4) was used, the slopes for the rare earths were approximately –3.25. However, the slopes obtained using the activity were very close to the theoretical value of –3.0. In view of the very low capacity of the resin used, it may seem remarkable that slopes so close to the theoretical values were obtained. One might expect the ionexchange sites to be so scattered that it would be impossible for a 3+ cation to exchange with three sites. However, other work has demonstrated that the exchange sites are in a thin layer on the outer perimeter of the ion-exchange bed. The results obtained here indicate that the concentration of sites in this outer layer is sufficiently dense that essentially theoretical exchange is obtained for polyvalent metal ions. The capacity factors for metal cations were measured with 0.75 M perchloric acid eluent on resins of 6.1, 13.8, and 24.9 lequiv g–1 exchange capacity. The results, shown in Table 5.6, show a substantial increase in capacity factors with increased resin capacity. According to Eq. (5.12), the logarithm of k should vary linearly with the logarithm of resin capacity, C. Linear plots are indeed obtained but the slopes are lower than the theoretical values and they vary from one metal ion to another.
5.3 Selectivity Table 5.6 Retention factors (k) for cations with resins of different
bulk densities (eluent: 0.75 M perchloric acid).
Cation
Resin capacity (mequiv/g) 6.1
13.8
24.9
V(IV)
0.12
0.40
0.81
Be(II)
0.24
0.32
0.38
Zr(IV)
0.24
0.40
1.03
Mg(II)
0.26
0.32
0.65
Fe(II)
0.28
0.40
0.93
NI(II)
0.28
0.51
0.87
U(VI)
0.28
0.59
1.21
Mn(II)
0.28
0.67
1.27
Zn(II)
0.30
0.42
0.85
Cu(II)
0.32
0.46
0.95
Cd(II)
0.32
0.51
1.11
Ca(II)
0.57
1.05
2.46
Ba(II)
1.52
5.05
–
Pb(II)
1.90
4.97
11.05
Fe(III)
2.89
5.68
11.0
In(III)
3.05
6.12
–
Lu(III)
6.6
15.9
35.4
Yb(III)
7.2
16.4
36.8
Tm(III)
7.5
17.8
38.4
Er(III)
7.9
18.8
42.4
Y(III)
8.2
19.3
42.8
Ho(III)
8.6
20.4
46.9
Dy(III)
9.3
24.4
55.4
Tb(III)
10.6
26.7
65.0
Gd(III)
12.5
31.5
73.1
Eu(III)
13.8
35.4
84.2
Sm(III)
15.4
39.4
95.8
Nd(III)
17.8
45.0
106.3
Pr(III)
18.6
48.9
115.0
Ce(III)
21.2
56.6
124.6
La(III)
23.8
68.1
172.1
Bi(III)
–
42.4
122.6
119
120
5 Principles of Ion Chromatographic Separations
An explanation of these observations might be that sulfonation of the resins proceeds inwardly from the outside of the resin beads. Thus, resins of higher capacity are apt to have thicker sulfonation layers. The microscopic selectivity coefficient may well change in various parts of the sulfonation layer. The selectivity coefficients measured would be an average of different microscopic values and might not follow the behavior predicted. 5.3.4 Factors Affecting Selectivity
The data listed in Tables 5.3–5.6 are simply observations concerning the effect of eluent concentration and resin exchange capacity on the retention factors of metal cations. A more fundamental approach is to examine the effect of physical and chemical variations in both the mobile and stationary phases on chromatographic behavior of ions. The factors affecting selectivity of ion chromatography have been reviewed in a recent publication [11]. It is often assumed that the strength of electrostatic attraction between a sample ion and oppositely charged sites within the ion exchanger is the overriding effect that determines selectivity. However, in the IC separation of organic cations, it has long been suspected that the retention mechanism involves more than electrostatic attraction. Hoffman and coworkers [1, 2] suggested that two mechanisms occur in such cases: ion exchange and hydrophobic interaction between the sample cations and the resin matrix. Dumont et al. [12] also concluded that much of the selectivity for protonated amine cations comes from hydrophobic interaction between the carbon chain of the analyte ions and the polymer matrix of the ion exchanger. By operating in a nonaqueous mobile phase (methanol, ethanol, 2-propanol) the hydrophobic interactions were greatly reduced and aliphatic amine cations from one to ten carbon atoms had almost the same retention factor. Selectivity in ion-exchange chromatography may be defined as the relative tendency of sample ions to form an ion pair with the sites of opposite charge within the stationary phase. It is helpful to divide the factors that affect selectivity into two general classes. 1. Electrostatic attraction (ES). This attraction is stronger when the sample ion has a higher charge and when the paired ions are very close to one another. 2. Enforced pairing effects (EP). Stronger ion pairing occurs within the stationary phase than in the aqueous solution because of a combination of effects that may include hydrophobic attraction, hydrogen bonding, lower dielectric constant and what has been termed ‘water-structure induced ion pairing’. The latter theory was developed by Diamond and coworkers to explain elution orders observed in anion-exchange chromatography [13]. Large, poorly hydrated univalent ions in aqueous solution intrude into the surrounding water structure
5.3 Selectivity
without being able to strongly orient the water molecules around themselves into coordinate hydration shells. This contributes to a tightening of the water structure around the ions. Large ions tend to be rejected by the water phase and are more easily extracted by nonstructured organic solvents than are smaller ions. The strong retention of polarizable inorganic anions, such as thiocyanate and perchlorate, and organic anions is likely the result of EP effects. Hydrophobic attraction between these anions and the ion exchanger certainly plays an important role. The relative attraction of smaller, hydrated inorganic ions and inorganic ions of higher charge, such as sulfate, is governed more strongly by ES effects.
5.3.4.1 Polymeric Matrix Effect Selectivity in anion IC is affected by the polymeric matrix of the ion exchanger. In one study [14] anion-exchange columns were prepared by applying a permanent coating of a cationic surfactant such as cetylpyridinium chloride to a cross-linked polystyrene resin (XAD-1) or to a polyacrylate resin (XAD-8). Polarizable anions, such as nitrate and iodides, gave significantly higher retention factors (relative to chloride) on a coated polyacrylate. Conversely, sulfate and thiosulfate gave considerably higher retention factors on the coated polystyrene resin. Polarizable anions are often found to give relative long retention times and tailed peaks when chromatographed on anion exchangers with a polydivinylbenzene polymeric matrix [2, 3]. This is most likely the result of hydrophobic interaction of the sample anions with the resin. Incorporation of about 10% methanol into the mobile phase has been used to alleviate this difficulty. In preparing anion-exchange columns by permanently coating a cationic surfactant onto a reversed-phase silica column, the underlying surface can be made less hydrophobic by applying an initial coating of a nonionic surfactant. Peak shape and chromatographic efficiency were markedly better when a C18 silica column was coated first with Triton X-100, a nonionic surfactant, and then with the cationic surfactant cetylpyridinium chloride [15]. Ion-exchange behavior is determined by both the lower and upper coating layers acting in concert and not by the charged upper surfactant layer alone. The physical nature of the polymeric matrix can also affect selectivity. Two basic types of polymeric ion exchangers have been used. Early gel-type resins were basically polystyrene materials with a limited degree of cross-linking. These were often designated as X-4, X-8, etc., to indicate the percentage of styrene-divinylbenzene (DVB) added to the polymerization mixture as the cross-linking tended to restrict the access of larger ions to the interior of the resin where the bulk of the actual ion exchange took place. The polymeric ion exchangers used in modern ion chromatography often are macroporous materials. The structure of each spherical bead consists of a network of micro particles with numerous pores and channels. The bulk of the ion exchange takes place at exchange sites on the solid micro particles with the resin bead. The pore size and chemical nature of the polymer will affect the ease with which competing ions can enter the resin and undergo ion exchange.
121
122
5 Principles of Ion Chromatographic Separations
Although ion-exchange beads have some porosity throughout, many of the materials currently used for ion chromatography are ‘pellicular’, that is, ion exchange only occurs at a thin layer on the outer surface of the spheres. Sulfonation of polymer beads proceeds from the outside to the interior, hence most of the sulfonate groups will be located near the surface when a short sulfonation time is used. Anion exchangers can be prepared by coating the surface-sulfonated beads with positively charged latex particles. However, the selectivity of these pellicular materials is still affected by the porosity and degree of cross-linking of the polymer interior as well as by the nature of the latex particles.
5.3.4.2 Resin Functional Group Variations in the chemical nature of the ion-exchange functional group have a major effect on selectivity in IC. This is particularly true for anion exchangers. Some of these variations in functional group structure were discussed in Chapter 3. Okada [16] compared the effects –NH3+ and –NEt3+ groups in anion exchangers of the same polymeric matrix and almost the same exchange capacity (0.4 mmole g–1). The –NH3+ resin had a more concentrated charge and a stronger electrostatic field than the resin with –NEt3+, where the + charge was more dispersed. Going from – NH3+ to –NEt3+ resulted in decreased electrostatic and hydrogen bonding interactions with sample anions and increased ion-induced dipole and London dispersion interactions. The latter two effects resulted in preferable binding of larger ions by –NEt3+. As an example, the ratio of retention factors of ClO4– : Cl– was 17.4 for the resins with –NEt3+ but only 1.89 for those with –NH3+ groups. Barron and Fritz investigated the effect of functional group structure on the selectivity of low-capacity anion exchangers for monovalent [17] and divalent [18] anions. A number of macroreticular anion-exchange resins of low capacity were prepared by chloromethylation of XAD-1 under mild conditions, followed by reaction with the appropriate tertiary amine. These resins were then evaluated to determine their relative selectivities for 17 monovalent anions. Increasing the size of the R group in the –NR3+ exchange sites in a series from R = methyl to R = octyl substantially increased the retention (relative to chloride) of larger ions like nitrate, chlorate, iodide and thiocyanate. However, the relative retention of divalent anions (sulfate, thiosulfate, oxalate) decreased as R became larger. A novel anion-exchange resin was prepared by reacting chloromethylated polystyrene particles with diethylenetriamine to produce a functional group with three nitrogen atoms [19]. By adjusting the pH of the mobile phase, this group would have a net charge ranging from 1+ to 3+. With an eluent (sodium perchlorate) that was completely ionized, the retention times of sample anions decreased steadily going from pH 2.2 to 8.2. An interesting feature of this resin was the unusually long retention time of sulfate compared to chloride and other monovalent anions. This may have been the result of very strong electrostatic attraction between the 2– sulfate and the high net charge on the resin cation.
5.3 Selectivity
Virtually all anion-exchange resins contain a quaternary ammonium group as the cation. By using a resin with the bulkier tetraalkylphosphonium group, the retention of nitrate is increased substantially. An ion exchanger of this type selectively removes nitrate from contaminated water supplies [20]. The chemical structure of the ion-exchange functional group will of course also affect the tenacity with which the eluent is held. An eluent containing NaOH or KOH is ideal for suppressed anion chromatography because the hydroxide is converted to water when it passes through the suppressor unit. However, the hydroxide ion is only weakly retained by the commonly used anion exchangers with the –N R3 functional group. By replacing one or two of the alkyl R groups with a hydroxyl ethyl group, –CH2CH2OH, a dilute solution of sodium or potassium hydroxide became an attractive eluent.
5.3.4.3 Solvation Effects Incorporation of some organic solvent into the aqueous eluent is an intriguing way to modify selectivity, but the ion exchanger must be compatible with a partially organic solution. Even a small percentage of organic solvent in the eluent caused early Dionex materials to swell badly, but higher crosslinking of the polymer was found to solve the problem. In discussing solvation effects, Rubin and Stillian [21] stated that solvent molecules are oriented around ions in an ordered manner. Use of a mixed organic– aqueous mobile phase allows intrusion of organic molecules into the aqueous solvation sphere. For an ion to adsorb onto an ion-exchange site, it must first rearrange and eventually partially shed its solvation sphere to allow close approach to the ion-exchange site; the closer it can approach the more tightly bound it becomes. Likewise, an ion-exchange site must reorient its solvation sphere to permit the ion to approach. Combining these concepts leads to the notion of selective mediation. The use of a mixed aqueous-organic mobile phase often has the practical advantage of providing faster separations. In one example, a mixture of nine anions was separated on a Dionex AS-11 column with 45 mM NaOH in 40% methanol as the stationary phase [11]. Early work in nonsuppressed cation chromatography illustrates the vital role of solvation within the ion-exchange phase [22]. Attempts to separate mixtures of alkali metal and alkaline earth cations with a column containing a lightly sulfonated macroporous polystyrene–DVB resin all failed. However, a sulfonated microporous polystyrene–4% DVB resin gave excellent separations of 1+ and 2+ cations. Later research has reaffirmed these early results [23]. A separation of alkali metal ions was first attempted in water alone using the lightly sulfonated macroporous cation exchanger with aqueous 3 mM methanesulfonic acid as the eluent. Under these conditions the sample cations exhibited very similar retention times. The selectivity of the macroporous resin for alkali metal ions was improved con-
123
124
5 Principles of Ion Chromatographic Separations
siderably by chemically introducing hydroxymethyl groups [24] prior to sulfonation. These results indicate that solvation of the resin plays a role in imparting selectivity for the various sample ions. Microporous cation-exchange resins form a gel and are highly hydrated within. With sulfonated macroporous resins the hydrated alkali metal ions may be repelled somewhat by the hydrophobic resin matrix. The presence of hydroxymethyl groups on the macroporous resin makes it less hydrophobic and improves selectivity for the hydrated alkali metal cations. When the sulfonated macroporous resin was used with the same acidic eluent in 100% methanol instead of water, a very good chromatographic separation was obtained. The alkali metal ions are solvated more with methanol than with water and the resin matrix is probably coated with a thin layer of methanol, which make the ions and resin surface more compatible with one another. Okada and Haroda [25] studied the local structures of chloride and bromide within an anion-exchange resin by means of X-ray adsorption fine structure (XAFS). Water molecules are coordinated with the paired anion within the resin, up to an average hydration number of 3. However, the anions are not as highly hydrated within the resin as they are in the external solution. An average of 2.1 water molecules are stripped off in transfer of chloride from the bulk solution to the resin. An average of 2.6 water molecules are stripped off bromide in the same transfer.
5.4 Chromatographic Efficiency
The basic concept of ion chromatography is actually quite simple. When a sample is injected into a packed column through which an ionic mobile phase (the eluent) is flowing, the sample ions A, B, C, ... are slowed to varying degrees by interaction with the ion-exchange particles (the stationary phase). As they move along the column, A, B, C, ... are gradually separated into distinct zones. As the ions emerge from the column and pass through the detectors, each zone appears as a Gaussian peak on the recorded chromatogram. Excellent chromatographic efficiency is exemplified by the completeness of peak separation from one another and by obtaining peaks that are as narrow and symmetric as possible. The separation ability of a chromatographic column is often measured by the number of theoretical plates, N. This concept comes originally from distillation theory in which the ability to separate volatile compounds by fractional distillation was related to the number of actual plates in the packed distillation column. In chromatography the number of theoretical plates in a column is calculated from the retention time, t, and the average peak width, w. N 16
t2 w2
(5:19
5.4 Chromatographic Efficiency
Since the shape of chromatographic peaks is essentially the same as a standard error plot in statistics, the peak width at the base may be written in terms of standard deviation. N 16
t2 t2 2 2 r
4r
(5:20
A convenient way to measure r from a chromatogram is to apply the relationship that the peak width at one-half the peak height is equal to 2.35 r. A drawback in using theoretical plates as a measure of chromatographic separation power is that N is not the same for all peaks and varies with the magnitude of the retention time, t. This is especially true for early peaks when the column dead time, to, is rather long and only a little smaller than t of the early peaks. This can lead to excessively high calculated N values even though chromatographic resolution of the early peaks may be rather poor. This problem was first recognized in the early days of capillary column gas chromatography. The concept of an effective theoretical plate number was introduced: Neff
t-to r2
2
(5:21
Neff tends to be very low when the difference between t and to is slight and is a better indication of separation power than N. In classical theory, N was used as a measure of separation power because N increased with greater column length. Height equivalent of a theoretical plate, H, served as a measure of chromatographic efficiency. H in mm or cm is calculated by dividing the column length, L, by the plate number H
L N
(5:22
A low value for H is an indication of high efficiency. Nowadays, many authors refer to N as the ‘efficiency’ and do not use H at all. This is somewhat misleading because N is actually a function of the column length as well as the true efficiency. A more fundamental way to look at efficiency is to examine the factors that contribute to peak broadening. Although the amount of sample used in IC is typically quite small, the number of ions involved in the separation is very large indeed. If only 10 lL of a dilute sample containing 10–5 M of a sample ion is injected, calculation using Avogadro’s number shows that the order of 1016 sample ions are present. These ions do not all behave identically in their passage through the column, but instead they follow a statistical distribution. As an example, Cl– and Br– are gradually resolved into separate zones as they are move through the column. Separation peaks for Cl– and Br– are recorded as these ions pass through the detector. Each peak is Gaussian; that is, it has the same shape as that of a standard error plot in statis-
125
126
5 Principles of Ion Chromatographic Separations
tics. The goal in IC is to use columns and systems where the peaks are as sharp and narrow as possible. Put another way, we are striving for conditions that will minimize peak broadening on the recorded chromatogram. The first opportunity for zone broadening occurs when the analytical sample is injected. A sample is introduced into a valve and then injected smoothly into the eluent stream as a small plug. Very little zone broadening should occur during this step unless an unusually large sample volume is used. Care should also be taken that tubing connecting the injector with the column is of small bore with no pockets or change in diameter where mixing could occur. The length of connecting tubing should also be as short as possible. With modern equipment, most of the zone broadening occurs within the column. These effects may be grouped into two major categories. The first is axial diffusion. As the sample ions are resolved into discrete zones, the zones tend to expand because of diffusion of the ions. Since diffusion across the column is limited by the small column diameter, the net zone diffusion takes place along the column and hence is called ‘axial diffusion’. Axial diffusion may cause substantial zone broadening for late-eluting sample ions or when a very slow eluent flow rate is employed. The magnitude of zone spreading by axial diffusion is a function of the time spent in the column and the diffusion coefficient of the particular ion. The major sources of zone spreading within the column are often lumped together in a term called ‘resistance to mass transfer’. This term includes several factors that contribute to peak broadening. 1. Differences in migration paths through the column, or ‘multipaths.’ This source of zone broadening is minimized by using ion-exchange particles that are spherical and of small and uniform diameter. As the science of chromatography has improved over time, the average diameter has become ever smaller: > 10 lm, 10 lm, 5 lm and often to 3 lm. Efficiency, as expressed by plate height H, has been shown to vary inversely with particle size. A very small particle size leads to higher resistance to liquid flow through a column, but the total back pressure (resistance to flow) can be reduced by using a shorter column. 2. Differences in diffusion of ions within the stationary phase. In many types of ion-exchange particles, exchange sites are located deeper within the particles and not just in the outer surface. Also, diffusion within a porous solid is slower than it is in the mobile liquid phase. Thus some ions of each sample component spend more time within the particle than others. 3. Equilibrium kinetics. Attaining chemical equilibrium is never instantaneous, and the time required for equilibrium between two distinct phases is always slower than it is within a single phase. Very slow equilibration or even nonequili-
5.4 Chromatographic Efficiency
brium distribution of ions between phases may result when the polymeric ion-exchanger surface is very hydrophobic. Sometimes a certain percentage of an organic solvent is added to the aqueous mobile phase to relieve this problem. 4. Mobile phase flow rate. Chromatographic efficiency, as measured by plate height (H), almost always decreases (higher H value) as a faster flow rate is used. A slower flow rate allows more time for equilibrium within the column to be achieved and decreases the amount of peak broadening. But an overly slow flow rate results in long retention times and increases peak broadening resulting from axial diffusion. The effect of eluent flow rate on efficiency, (H), is best illustrated by a classical Van Deemter plot (Figure 5.3) The linear flow rate (u) is used instead of the volume flow rate; u = column length (L) divided by the dead time, to. The very best efficiency is obtained at the flow rate where H is lowest. But a somewhat higher flow rate is generally used in practice in order to obtain a faster separation. Variations in temperature throughout the column can also contribute to zone spreading. For this reason the column is contained within a constant temperature oven in most contemporary IC instruments. Elevated column temperature is being recognized as a useful variable in selecting optimal conditions for a separation.
Figure 5.3 Van Deemter plot showing the effect of eluent flow rate on the theoretical plate height, H.
127
128
5 Principles of Ion Chromatographic Separations
As we trace the path through the IC column, the final incremental contributions to the ultimate peak broadening take place as the ion zones leave the column and pass through the suppressor unit, if there is one, and the detector. A certain amount of mixing occurs during passage through these units. It is therefore important to minimize the volume contained in each of these units. As an example, a detector designed for use with ordinary columns (typically about 4 mm i.d.) may result in excessively broad peaks if used for capillary IC columns. How much does each of the peak broadening mechanisms discussed above contribute to the final peak width? This question can be answered if we are able to make an estimate of the magnitude of each of the contributing factors. Recall that chromatographic peaks are Gaussian and that the width of a peak at its base is approximately four standard deviations: w = 4r. The factors that contribute to peak broadening are additive provided the variance (r2) is used instead of the standard deviation. The total variance (r2tot) is the sum of variances due to multipaths (r2mp), axial diffusion (r2dif ), resistance to mass transfer (r2mt) and extra-column (r2ec). (r2tot) = (r2mp) + (r2dif) + (r2mt) + (r2ec).
(5.23)
The differences in tortuous paths through the packed column (i.e., multipaths) is minimized by using spherical solid particles of a very narrow size range, packed very evenly into the column. The value of r2mp will increase with the average diameter of the packing material. The second term, r2dif , represents the zone spreading that each sample component exhibits due to diffusion along the column axis. Diffusion coefficients in aqueous solution are generally low, so the contribution of this term is relatively small unless the retention time is quite long because of a very slow flow rate or a high retention factor. Resistance to mass transfer, r2mt, is by far the major contributor to sample zone spreading within the column. This term is minimized by a column packing that attains equilibrium of the analyte between the mobile and stationary phases as quickly as possible. A moderate linear flow rate, u, should be employed because r2mt increases with flow rate. The last term, r2ec, points out that substantial peak broadening may occur outside the column. To avoid this, the transfer lines from sample injection to column and column should be as short as possible. Stagnant areas in the system must be avoided. These can occur, for example, if the connection between two pieces of tubing bows out. The detector cell should have a low dead volume. With carefully designed IC systems, resistance to mass transfer is usually the major source of peak broadening. The interface between the hydrophilic mobile phase and the more hydrophobic stationary phase is a critical factor with regard to rapid mass transfer. Solute ions may get stuck at the interface for different time periods. This results in peak broadening and peak distortion in more severe cases. So we try to make the phase interface more amenable to rapid transfer of solutes
References
across the interface. This transfer must take place in both directions across the interface to satisfy equilibrium requirements. Addition of a certain amount of a common organic solvent (methanol or acetonitrile) helps to correct this situation. However, organic anions or cations may still have long retention times and broad peaks. A better answer to this problem may be to use a more hydrophilic ion exchanger or to use an organic additive of higher molecular weight to the mobile phase. Li and Fritz [26] demonstrated that an aqueous phase containing only 0.25% 1,2-octanediol or an aqueous – acetonitrile phase containing 1.5% 1-hexanol gave shorter retentions and sharper peaks for carboxylate and amine analytes. The organic additive concentrates at the phase interface and facilitates mass transfer of solute ions. Rapid equilibration of solute ions within the ion-exchange material is another critical factor. This is frequently achieved by using resins with accessible exchange sites, often concentrated near the outer surface, and a resin with sufficient porosity. The liquid within the resin is also likely to influence solute mobility. The effect of solute transport across the interface and within the solid phase will be a recurring theme throughout this book, with respect to design of new resins and the mass transfer mechanism with its ion-exchange, ion exclusion and zwitterion stationary phases. What if we could eliminate a liquid–solid phase interface entirely? A significant improvement in chromatographic efficiency would be expected, with unusually sharp analyte peaks. This is indeed the case for ion chromatography in unpacked capillaries where ions are separated by differences in their electrophoretic migration (Chapter 11).
References [1] A. Rahman and N. E. Hoffman, Reten-
[2]
[3]
[4] [5]
[6]
tion of organic cations in ion exchange chromatography, J. Chromatogr. Sci., 28, 157, 1990. K. Lee and N. E. Hoffman, Retention of some simple organic anions in ion exchange HPLC, J. Chromatogr. Sci., 30, 98, 1992. O. D. Bonner and L. L. Smith, A selectivity scale for some bivalent cations on Doxex 50, J. Phys. Chem., 61, 326, 1957. J. F. Helfferich, Ion Exchange, McGrawHill, New York, 1962, p169. W. Rieman III and H. F. Walton, Ion Exchange in Analytical Chemistry, Pergamon, Oxford, 1970, p. 45. F. W. E. Strelow, An ion exchange selectivity scale of cations based on equilibrium distribution coeficients, Anal. Chem., 32, 1185, 1960.
[7] F. W. E. Strelow, R. Rethemeyer and
C. J. C. Bothma, Ion exchange selectivity scales for cations in nitric acid and sulfuric acid media with a sulfonated polystyrene resin, Anal. Chem., 37, 106, 1965. [8] F. W. E. Strelow and H. Sondorp, Distribution coefficients and cation-exchange selectivities of elements with AG50WX8 resins in perchloric acid, Talanta, 19, 1113, 1972. [9] G. J. Sevenich and J. S. Fritz, Metal ion selectivity on sulfonated cationexchange resins of low capacity, J. Chromatogr., 371, 361, 1986. [10] G. J. Sevenich and J. S. Fritz, Preparation of sulfonated gel resins for use in ion chromatography, Reactive Polymers, 4, 195, 1986.
129
130
5 Principles of Ion Chromatographic Separations [11] J. S. Fritz, Factors affecting selectivity in
ion chromatography, J. Chromatogr. A, 1085, 8, 2005. [12] P. J. Dumont, J. S. Fritz and L. W. Schmidt, Cation-exchange chromatography in nonaqueous solvents, J. Chromatogr. A, 706, 109, 1995. [13] R. M. Diamond, The aqueous solution behavior of large univalent ions: a new type of ion pairings, J. Phys. Chem., 67, 2513, 1963. [14] D. L. Duval and J. S. Fritz, Coated anion-exchange resins for ion chromatography, J. Chromatogr., 295, 89, 1984. [15] J. S. Fritz, Zhu Yan and P. R. Haddad, Modification of ion chromatographic separations by ionic and nonionic surfactants, J. Chromatogr. A, 997, 21, 2003. [16] T. Okada, Nonaqueous anion-exchange chromatography I, role of solvation in anion-exchange resin, J. Chromatogr. A, 758, 19, 1997. [17] R. E. Barron and J. S. Fritz, Effect of functional group structure on the selectivity of low-capacity anion-exchangers for monovalent anions, J. Chromatogr., 284, 13, 1984. [18] R. E. Barron and J. S. Fritz, Effect of functional group structure and exchange capacity on the selectivity of low-capacity anion-exchangers for diva-
lent anions, J. Chromatogr., 316, 201, 1984. [19] J. Li and J. S. Fritz, Novel polymeric resins for anion-exchange chromatography, J. Chromatogr., 793, 231, 1998. [20] J. E. Lockridge and J. S. Fritz, Determination of water using nitrate selective ion-exchange resin, U. S. Patent 4 944 878, July 31, 1990. [21] S. Rubin and J. Stillian, Practical aspects on the use of organic solvents in ion chromatography, J. Chromatogr. A, 671, 63, 1994. [22] J. S. Fritz, D. T. Gjerde and Rose M. Becker, Cation chromatography with a conductivity detector, Anal. Chem., 52, 1519, 1980. [23] P. J. Dumont, J. S. Fritz and L. W. Schmidt, Cation-exchange chromatography in nonaqueous solvents, J. Chromatogr. A, 706, 109, 1995. [24] J. J. Sun and J. S. Fritz, Chemically modified resin for HPLC, J. Chromatogr., 522, 95, 1990. [25] T. Okada and M. Harada, Anal. Chem. (submitted). [26] S. Li and J. S. Fritz, Organic modifiers for the separation of organic acids and bases by liquid chromatography, J. Chromatogr. A, 964, 91, 2002.
131
6 Anion Chromatography 6.1 Scope and Conditions for Separation
A very large number of anions have been separated successfully by anion chromatography. Table 6.1 lists many of these anions, grouped according to their general type. Particular attention has been paid to the so-called common anions. These can be separated in a single run with suppressed conductivity detection (Section 6.2), nonsuppressed conductivity (Section 6.3), or by one of the other detection methods discussed in later sections. If necessary, the sample is diluted so that the ions to be determined are in the low to medium concentration range. However, sample anions down into the lower parts-per-billion (ppb) range can be successfully separated and quantified.
Table 6.1 Anions determined by ion chromatography (grouped
according to general type). 1.
Common anions: Fluoride, chloride, bromide, nitrite, nitrate, phosphate, sulfate.
2.
Polarizable: Iodide, thiocyanate, thiosulfate, perchlorate, chromate, molybdate, tungstate.
3.
Inorganic anions of weak acids: Borate, bicarbonate, carbonate, cyanide, silicate, sulfide.
4.
Other inorganic anions: Arsenite, arsenate, azide (N3–), bromate, chlorite, cyanate, chlorate, perchlorate, iodate, periodate, sulfamate (NH2SO3–) sulfite, selenite, selenate.
5.
Smaller organic anions: Amino acids, alkane carboxylic acids (formate, acetate, propionate, butyrate), chloro carboxylic acids (chloroacetate, dichloroacetate), hydroxy acids (hydroxyacetate, lactate, tartrate, citrate), glycolate, gluconate, pyruvate, dicarboxylic acids (oxalate, malonate, succinate, glutarate, fumarate, maleate), alkanesulfonic acids (methanesulfonate, ethanesulfonate).
6.
Larger organic anions: aromatic carboxylic acids, aromatic sulfonic acids, carbohydrates (aldoses, ketoses), nucleotides, nucleic acids, proteins, surfactants.Polarizable anions include those listed in Group 2 of the table. These anions have a relatively high affinity for the ionexchange stationary phase and therefore require a stronger eluent for their separation.
Ion Chromatography, 4th Ed. James S. Fritz and Douglas T. Gjerde Copyright © 2009 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim ISBN: 978-3-527-32052-3
132
6 Anion Chromatography
The anions listed in Group 3 require an alkaline pH for their separation so that they will be in the anionic rather than the molecular form. Some of the ions listed in Group 4 will also require an alkaline eluent for separation. Smaller organic anions such as those listed in Group 5 are usually sufficiently hydrophilic to be separated in aqueous solution. If necessary, some methanol or acetonitrile may be included in the eluent to increase their solubility and also to avoid their retention by the stationary phase due to a non-ion-exchange partitioning mechanism. The elution of ions in a homologous series (formate, acetate, propionate...) is usually in order of increasing molecular weight. One or more hydroxyl groups in the molecule tends to decrease retention time. Separation of anions by ion chromatography was initially concerned primarily with the separation of inorganic and small organic anions, but this is changing. Ion chromatography is being used more and more to separate larger organic ions and bio ions (see Chapter 12). Anions can also be separated on columns with no ion-exchange groups using ion-pairing techniques (Chapter 9), on zwitterion stationary phases (Chapter 10), or on open-tubular capillaries with electrical, rather than mechanical pumping (Chapter 11). 6.1.1 Columns
A large number of excellent columns for anion chromatography are now available from commercial vendors. In fact, the variety of columns is so large that intelligent selection of an appropriate column for particular application may be confusing. While the columns available commercially are constantly changing, the major ones from leading suppliers are listed in Tables 6.2 – 6.6. Notes from the manufacturer concerning recommended use are included in the tables. The columns in Tables 6.2 and 6.3 all consist of a polymeric spherical material coated with an outside layer of a functionalized latex. The latex is held firmly in place by electrostatic attraction between –SO3– groups on the outside of the substrate and positively charged groups from the coating latex. The anion-exchange of the column material comes from the quaternary ammonium sites on the latex. Since the latex particles are very small and the coating layer is thin, these materials provide very efficient anion-exchange separations. Variations in the type of latex, its particle size, and total anion-exchange capacity impart different properties to these materials. However, the chemical nature of the substrate polymer and its porosity also affect the affinity of sample anions for the solid phase. Although the anion-exchange resins in Tables 6.2 and 6.3 were designed specifically for use with suppressed conductivity detection (Section 6.2), they are also generally useful for IC involving other modes of detection. The anion-exchangers described in Tables 6.4, 6.5 and 6.6 are not latex-coated but their exchange groups are mostly in a narrow band near the outside perimeter of the spherical ion-exchange particle. In some instances the ion-exchange part of the particle consists of a functionalized polymer grafted onto the outer surfaces of the substrate. It should be emphasized that these columns work well for all forms
6.1 Scope and Conditions for Separation
of detection including suppressed conductivity. But the properties of these columns vary, so again it is important to select a column that is compatible with the intended IC separation.
Table 6.2 Dionex Ion Pac hydroxide-selective, anion-exchange columns.
Designation
Capacity
Recommended use
AS 5
Low
Polyphosphates, oxanions, EDTA complexes, metal cyanide complexes, iodide, thiosulfate, thiocyanate
AS 10
High
High-ionic-strength samples, low-molecular-weight-aliphatic acids
AS 11
Low
Fast profiling of inorganic and organic anions in food, beverages, process solutions, waste-water brines
AS 11-HC
High
Similar to AS 11 but allows for injection of more concentrated samples
AS 15
High
Trace levels of inorganic anions and organic acids in highpurity water samples.
AS 16
High
Hydrophobic anions (iodide, thiocyanate, thiosulfate, perchlorate), polyphosphates, polycarboxylates
AS 17
Low
Fluoride, acetate, formate, chloride, nitrite, bromide, nitrate, sulfate, phosphate in simple matrices
AS 17-C
Low
Common anions in high-purity water. Provides low sulfate blanks and fast equilibration time
AS 18
–
Inorganic anions in a variety of sample matrices
AS 19
High
Oxyhalides and inorganic anions. Key application is trace bromate in ozonated drinking water
AS 20
High
Trace perchlorate in drinking water and ground water
AS 21
–
Fast analysis of trace perchlorate in drinking water prior to MS/MS detection
Fast Anion III A
–
Phosphoric and citric acids in cola soft drinks
AS 24
–
Haloacids prior to MS/MS detection. Analysis in drinking water at low lg L–1 concentrations
133
134
6 Anion Chromatography Table 6.3 Dionex carbonate-based, anion-exchange columns.
Designation
Capacity
Recommended use
AS 4 A-SC
Low
Solvent-compatible column for fast isocratic separation of inorganic anions
AS 9-SC
Low
Fast analysis of inorganic anions and oxyhalides
AS 9-HC
High
Inorganic anions and oxyhalides
AS 12 A
Medium
Fast isocratic separation of inorganic anions and oxyhalides
AS 14
Medium
Fast isocratic separation of common anions
AS 14-A
High
Fast isocratic separation of common anions
AS 22
V. high
Fast isocratic separation of common anions
AS 23
V. high
Trace bromate in drinking water
Table 6.4 Alltech anion-exchange columns.
Designation
Material
Recommended use
Allsep Anion
Methacrylate Quat-ammonium
Hydrophilic; symmetrical peaks for all anions
Allsep A-2
Methacrylate; Alkanol Quat-ammonium
Strongly retained anions; weak and strong acid anions separated in one run; resolves acetate and formate away from fluoride and chloride
Anion/S 2.0-5.5
Silica, Quat-ammonium
Rapid mobile phase equilibration, low adsorption. Sharp peaks for polarizable anions; pH range
Anion/R
Polystyrene, Quat-ammonium
Separations at high pH; common anions, silicate, borate, cyanide, formate, acetate
6.1 Scope and Conditions for Separation Table 6.5 Hamilton anion-exchange columns.
Designation
Material
Recommended use
PRP-X 100
DVB
Separation of common inorganic anions and organic anions; 10–500 ppm
PRP-X 110
DVB
Inorganic and organic anions; 20 ppb–20 ppm Good separation of fluoride from water dip
PRP-X 500
Acrylate
Fast separation of large proteins; limited porosity
PRP-X 600
Acrylate, weak base
Good separation of proteins
RCX-10
DVB
Separation of carbohydrates (anionic at high pH). Pulsed amperometric detector recommended
RCX-30
DVB
Similar to RCX-10 but longer retention times
These columns generally contain 7 lm particles, are stable at pH 1–13, and tolerate common organic solvents 0–100%.
Table 6.6 Metrohm-peak anion-exchange columns.
Designation
Material
Particle Size, m
Col. Capacity lmoles
Recommended use
Supersep
Acrylate
12
24
Common anion
Metrosep A
Styrene-DVB
6
700
NO2–, Br–, NO2– in high salt
Dual 3
Acrylate
6
34
F–, Ac–, Cl– in bio samples
Dual 4-25
Monolith SiO2
2
23
High flow rate, fast sep. of NO2– / NO3–
Dual 4-50
″
″
2
45
Common anions < 5 min
Dual 4-100
″
″
2
90
Nitrate, sulfate, perchlorate in high Cl–
6.1.2 Separation Conditions
The goals of the ion chromatographer are first, to achieve a satisfactory separation of the sample components of interest and second, to perform the separation as quickly as possible. Several parameters can be manipulated to fulfill these objectives. Of these, the choice of eluent composition and selection of the detector are undoubtedly the most important. Detectors fall into either of two classifications. Conductivity detectors may be classed as general detectors because all ions (cations as well as anions) contribute
135
136
6 Anion Chromatography
to the signal, although to varying degrees. Spectral detectors (UV–Vis) are selective because ions that absorb at the specific wavelength used contribute to the signal. Both types of detectors can be used in IC in either the direct or indirect detection mode. When direct detection is used, the eluent anion has a significantly lower equivalent conductance than the sample ions to be detected. As an example, a benzoate salt (limiting equivalent conductance = 32 S cm2 equiv–1) can be used for direct detection of ions such as chloride, bromide, iodide, nitrate and sulfate, which have a limiting equivalent conductance between 71 and 80 S cm2 equiv–1. A significant increase in overall conductivity is obtained as each of the sample ions passes through the detector, provided that the eluent concentration is not too high. A highly conductive anion such as hydroxide (limiting equivalent conductance = 198 S cm2 eq–1) would normally be used for detection of anions by indirect conductivity. This would establish a chromatographic baseline at a high conductance level. When a sample anion that has a much lower equivalent conductance passes through the detector, a ‘peak’ of much lower conductance is observed. This occurs because the total ionic concentration (sample anion plus OH– plus cations) remains constant but the zone containing the sample anion has a lower conductance that that of the background eluent. With direct spectral detection, the detector is set at a wavelength where the sample anions absorb but the eluent cation and anion do not. A number of inorganic anions including nitrate, bromide, iodide, thiocyanate, bromate and thiosulfate, as well as aromatic organic anions, can be detected. Sodium chloride, sodium perchlorate, or a number of other anions could be used in the eluent. Indirect spectral detection requires the use of a strongly absorbing anion in the eluent. Eluents containing molybdate, p-hydroxybenzoate or phthalate have been used successfully. In another form of direct detection, the baseline, or background, signal is suppressed by a chemical reaction while preserving or even enhancing the detector signal resulting from the sample anions. This reaction takes place within a device called a ’suppressor’ that is placed between the outflow end of the column and the detector cell. Chemical suppression provides a simple yet elegant way to reduce the background conductance of the eluent and at the same time to enhance the conductance of sample ions. In its original form, a second ion-exchange column was placed between the separator column and the conductivity cell. For anion analysis a basic anion was used in the eluent and a large, H+-form cation-exchange column was used as the second ’stripper’ column [1]. Modern methods of chemical conductivity suppression utilize much smaller and more efficient devices than the original stripper columns, but the basic principles are largely unchanged. The reactions for suppressed anion chromatography are as follows:
6.1 Scope and Conditions for Separation
Eluent NaOH + Eluent (Highly Conductive)
RSO3H Suppressor
> H2O + Suppressed Eluent (Weakly Conductive)
RSO3Na Suppressor
Analyte NaX + Analyte Salt (Conductive)
RSO3H Suppressor
> HX + Acid form of Analyte (Highly Conductive)
RSO3Na Suppressor
The basic eluent (OH–) is neutralized by H+ from the cation exchanger of the suppressor to form water. A sample zone passing through the suppressor is converted from the sodium as the counterion (NA+OH–) to the more highly conducting hydrogen counterion (H+X–). The major devices for suppressed conductivity detection in ion chromatography have been reviewed [2]. These are described in chronological order in the following sections. The goals of the ion chromatographer are first, to achieve a satisfactory separation of the sample components of interest and second, to perform the separation as quickly and conveniently as possible. Several parameters can be manipulated to fulfill these objectives. Resin structure. The chemical structure of the ion exchanger can have a major effect on its selectivity for various ions. As an example, anion-exchangers with quaternary ammonium functional groups have a lower affinity for hydroxide anions than for most of the other common inorganic anions. Replacement of one or more of the methyl groups with –CH2CH2OH resulted in an anion exchanger with a much greater affinity for OH–. This development made it possible to replace eluents containing bicarbonate and carbonate with hydroxide eluents, which were much more compatible with suppressed conductivity. Resin capacity. A lower exchange capacity will result in faster elution of analyte ions or will permit elution with an eluent of lower anion concentration. Divalent anions are affected more relative to monovalent anions. Column diameter. Use of a column with an i.d. of 2 mm instead of 4.0 or 4.6 mm will result in a faster elution at the same volume flow rate. Eluent. Eluent type and concentration. Since eluent anions and sample anions compete for ion-exchange sites on the stationary phase, it follows that more strongly retained eluent anions will give faster elution of sample anions. Ionexchange equilibria were discussed in Chapter 5. For example, an organic 2– anion such as phthalate, C6 H4
CO2 22 , is a stronger eluent than benzoate, C6H5CO2–, or p-hydroxybenzoate, HOC6H4CO2–. A solution of a polarizable inorganic anion, such as perchlorate, ClO4–, is a more powerful eluent than is a solution containing the same concentration of chloride. Column length. A relatively short column length, 50–100 mm, may be adequate for an easy separation. Indeed, some studies have indicated that a short column is capable of better chromatographic efficiency than longer columns. However, a
137
138
6 Anion Chromatography
longer column, 150–250 mm, may be needed for samples with several analytes or when two or more peaks are difficult to resolve. Flow rate. A ‘normal’ flow rate is often considered to be 1.0–1.5 mL min–1. A somewhat faster eluent flow rate may be used to obtain a faster separation, but at a cost of diminished chromatographic efficiency. Very low flow rates are best avoided as they increase peak broadening due to axial diffusions. Organic solvents. A low percentage, 2–5% of methanol, acetonitrile or other organic solvent is often added to the eluent, especially when the ion-exchanger matrix is hydrophobic. The organic additive tends to be adsorbed at the interface between the hydrophilic mobile phase and the more hydrophobic stationary phase to provide a more effective transfer of ions back and forth between the phases. A somewhat higher percentage of eluent additive is sometimes used to reduce the retention times of organic sample ions. Isocratic and gradient elution. Use of an eluent of fixed composition (isocratic elution) is the simplest, most straightforward method for ion chromatography. But when there is a large difference between the retention times of early- and lateeluting ions, or when a very complex sample is to be analyzed, gradient elution may be desirable or even essential. In the simplest form of gradient elution two eluents (A and B) are used, with eluent B being a more concentrated and thus a stronger eluent than A. Starting with 100% A, increasing percentage of the stronger eluent B is added to A. The elution is programmed so that B is mixed with A at set time intervals or B is continuously mixed with A at a set rate. The net result is that later peaks are eluted more rapidly than would be possible with isocratic elution. It should be kept in mind that, after a gradient elution, re-equilibration with eluent of the lower strength is necessary before the next run can begin. Quantification with gradient elution is also more complex. It is best to calibrate peak height or area under the exact conditions of the gradient elution.
6.2 Suppressed Anion Chromatography
In the original suppressed-conductivity system the suppressor column contained considerably more ion-exchange resin than the separation column in order to obtain enough exchange capacity for continuous operation. This bulky column had a large dead volume that caused considerable peak dispersion and broadening. Regeneration of the suppressor column was another serious problem. After several hours of operation, the ion-exchange bed would become exhausted and had to be regenerated. This was done offline by passing a sulfuric acid solution through the column, and then flushing with water. The fiber suppressor [3, 4] was the first device based on the use of an ionexchange membrane. It consisted of a long, hollow fiber made of a semi-permeable ion-exchange material. Column effluent containing zones of separated sample ions passed through the hollow center of the fiber. Here the sodium counterion was exchanged for H+ from the membrane. The outside of the hollow fiber
6.2 Suppressed Anion Chromatography
was bathed in an acidic solution, allowing for continual replacement of the H+ as the effluent passed through. The main advantage of this design was that it permitted continuous operation of the IC system. Band broadening in this suppressor was less than with the large packed-bed devices but was still significant. Fiber suppressors were also limited in their ability to suppress flow rates above 2 mL min–1 or eluents above 5 mM concentration. A flat membrane suppressor from Dionex, known as the Micro-Membrane Suppressor (MMS), had a much higher capacity and lower dead volume than previous devices and was able to operate around the clock with minimal attention. The internal design of the MMS is shown in Figure 6.1. Two semi-permeable ionexchange membranes are sandwiched between three sets of ion-exchange screens. The eluent screen is of fine mesh to promote the suppression reaction while occupying a very low volume. The ion-exchange membranes on either side of this screen define the eluent chamber. There are two ion-exchange regenerant screens that permit tortuous flow of the regenerant solution toward the membranes. These screens provide a reservoir for suppressing ions without having a counterion present.
Figure 6.1 Internal design of the MicroMembrane Suppressor (from Ref. [2] with permission).
The flow pattern for the anion suppressor is shown in Figure 6.2. Column effluent flowing through the suppressor exchanges Na+ for H+ from the cationexchange membranes, as shown in the middle part of the figure. Since the suppressor is actually a sandwich configuration with fairly broad cation-exchange membranes placed very close together, the exchange reaction proceeds rapidly and there is adequate exchange capacity to handle eluents of higher concentrations. A mineral acid such as dilute sulfuric acid flows through outer parts of the
139
140
6 Anion Chromatography
Figure 6.2 Suppression mechanism for the Anion MicroMembrane Suppressor (from Ref. [2] with permission).
suppressor to provide continuous regeneration. Regenerant flow is counter-current to the column effluent and at approximately three to ten times the chromatographic flow rate. A major drawback of the membrane suppressors was that they required a constant flow of regenerant for continuous suppression. This could consume up to 10 mL min–1 of regenerant solution. To reduce the large volumes of regenerant needed, an accessory for continuous regenerant recycling was introduced by Dionex in 1987. A large ion-exchange cartridge was used to remove the comparatively low concentrations of waste products (Na+, etc.) and replace them with fresh regenerant ions (H+). A pump recirculates the regenerant through the suppressor cartridge. The net effect is that only a small reservoir of regenerant solution is required for effective operation. 6.2.1 Electrolytic Suppressors
The ideal way to regenerate a suppressor for IC is to electrolyze water to produce the H+ or OH– needed. Strong and Dasgupta invented a practical suppressor of this type in 1989 [5]. In this device, a platinum wire-filled tube made of a Nafion cation-exchange membrane is inserted into another, larger Nafion tube and coiled into a helix. The helical assembly is inserted within an outer jacket packed with granular conductive carbon. An alkaline eluent, for example, NaOH or Na2CO3, flows in the annular channel between the two membranes, and pure water flows through the inner membrane and the outer jacket countercurrent to the direction of eluent flow. A DC voltage (3–8 V) is applied across the carbon bed and the platinum wire.
6.2 Suppressed Anion Chromatography
Sodium ions in the eluent migrate to the cathode compartment, resulting in water as the suppressed eluent. Up to 500 min–1 of sodium hydroxide could be suppressed effectively with a membrane 50 cm in length. The band dispersion was 106 lL for a 20lL sample. In 1992 Dionex introduced a commercial electrochemical suppressor called a Self Regeneration Suppressor, or SRS [6]. The internal design is similar to the membrane suppressor, but the regenerating ion (H+ for anion chromatography) is produced by electrolysis of water. This allows the use of very low flow rates for regenerant water and avoids the use of the independent chemical feed needed for earlier suppression devices. The mechanism for the Anion SRS is described in Figure 6.3. Hydrogen ions generated at the anode traverse the cation-exchange membrane to neutralize the basic eluent. Sodium counterions are attracted to the negatively charged cathode, where they permeate the membrane in the cathode chamber and pair off with electrogenerated hydroxide ions to maintain electric neutrality. Waste gases, hydrogen from the cathode and oxygen from the anode, are vented with a liquid waste of aqueous sodium hydroxide. As with other suppressors, analyte ion signals are enhanced by exchange of their counterions for hydrogen ions.
Figure 6.3 Mechanism of suppression for the Anion Selfregenerating Suppressor (from Ref. [2] with permission).
6.2.2 Solid-Phase Reagents, 1990 [7]
Gjerde and Benson discovered that post-column addition of a suspension of sulfonated polystyrene particles may be used to reduce the background conductance of basic eluents used in anion chromatography [7]. The eluent cation (typically Na+) is also replaced in the analyte ion bands by the more highly conducting H+ as the
141
142
6 Anion Chromatography
Figure 6.4 Detailed diagram of hardware configuration for post-column addition of SPR. (1 = Conductivity detector: Waters 431 detector, four electrode cell design; 2 = waste line: 4 × 0.009 in. stainless connected to 431 + 24 × 1/16 × 0.060 in. PTFE tubing; 3 = tee to 431 15 × 1/16 × 0.010 in. PTFE to 431 inlet; 4 = column to tee: shortest 1/16 × 0.010 in. PTFE from column to tee; 5 = tee: Unmount tee from check valve block for shortest path
length; 6 = analytical column: Waters IC PAK A or IC PAK A HR; 7 = check valve to tee: 2 × 1/8 in o.d.: PTFE; 8 = check valve; 9= polisher column to check valve: 3 × 1/8 in. o.d. PTFE; 10 = polisher column: 8 × 25 mm containing AGI × 8200 mesh; 11 = reservoir to polisher column: 12 × 18 in. o.d. PTFE; 12 = air supply: minimum of 90 p.s.i. compressed air supply; 13 = reservoir for SPR: reconfigure with outlet on left side (from Ref. [9] with permission).
counterion to a sample anion. Since the added reagent is a solid, it is ‘invisible’ to detectors that respond only to the liquid phase, for example, conductivity and potentiometric detectors. A typical solid-phase reagent (SPR) is a colloidal polystyrene sulfonic acid material, 2.3 mequiv g–1 exchange capacity, used as a 0.5–1.0% suspension at a flow rate of 1.0 mL min–1. This material stays in suspension and undergoes rapid ion exchange with ions in the column effluent. A detailed diagram of the hardware required is shown in Figure 6.4. This unique system works well for gradient elution [8] and for the determination of analytes of extreme concentration ratios [9]. 6.2.3 Typical Separations
The mobile phase used for anion chromatography with suppressed conductivity detection must (i) have adequate strength for elution of sample anions and (ii) produce products with very low conductivity after passing through the suppressor unit. A mixture of sodium bicarbonate and sodium carbonate has been widely used as an eluent for many years. Carbonate with a 2– charge is a stronger eluent than bicarbonate. By using a mixture of the two, the eluent strength can be adjusted as
6.2 Suppressed Anion Chromatography
desired. The suppressed products of this eluent are carbonic acid in equilibrium with carbon dioxide, which gives a reasonably low conductance. The background can be lowered by use of a suppressor that removes carbon dioxide gas by aeration. The ultimate eluent in terms of suppressed conductivity is the hydroxide ion, which gives water as the suppressor product. With the advent of anion-exchangers with an increased affinity for hydroxide and suppressors that tolerate a higher eluent concentration, the use of sodium or potassium hydroxide has become more popular. However, it is difficult to remove all of the carbonate from chemical solutions of sodium hydroxide. Electrolytic generation is now the preferred way to produce hydroxide eluents for IC. The product is almost entirely free of carbonate and the electrolytic generation provides excellent control of the concentration. Electrolytic generators are described in Chapter 1. Particular attention has been paid to the determination of common inorganic anions in a broad variety of aqueous samples. Excellent resolution of sample anions can usually be obtained by isocratic elution with either carbonate–bicarbonate or hydroxide eluents. Figure 6.5 shows a separation of ten anions in about 5 min with a carbonate–bicarbonate eluent. A similar separation on a different type of column is illustrated in Figure 6.6. Polyvalent anions, phosphate and sulfate, have longer retention times due to a lower carbonate concentration in the eluent.
Figure 6.5 Rapid separation of inorganic anions on an Ion Pac AS 12A column. Eluent: 0.5 mM sodium hydrogen carbonate/10.5 mM sodium carbonate. Detection: suppressed conductivity. Injection volume: 10 lL.
Peaks: (1) fluoride, (2) chlorite, (3) bromate, (4) chloride, (5) nitrite, (6) orthophosphate, (7) sulfate, (8) bromide, (9) chlorate, (10) nitrate (courtesy Dionex).
143
144
6 Anion Chromatography
Figure 6.6 Common anions, suppressed conductivity detection. Conditions: Hamilton PPRP-X110S, 150 × 4.1 mm, 1.7 mM sodium bicarbonate/1.8 mM sodium carbonate/0.1 mM sodium thiocyante, 2.0 mL min–1 (courtesy Hamilton Co.).
6.2.3.1 Isocratic and Gradient Elution Chromatographic separation of anions with a broad range of affinity for the stationary phase can present a difficult situation if an eluent of fixed concentration (isocratic elution) is to be used. An eluent of relatively low concentration is needed to resolve early-eluting anions such as fluoride, acetate, formate and chlorite, but more strongly retained anions such as bromide, nitrate and sulfate will have long retention times. A more concentrated eluent will elute the latter anions more quickly but may fail to resolve the early peaks. In this event, gradient elution should be considered. In gradient elution, two or more eluents are mixed in different proportions over a set time period so that the eluent strength is gradually increased. Chromatographic separations by isocratic elution and gradient elution are compared in Figure 6.7. With an isocratic sodium borate–sodium hydroxide eluent the first six peaks are nicely separated within 5 min but the later peaks become increasingly broad with some tailing. In the gradient separation eluent A is pure water and eluent B is a fairly concentrated solution of borate and hydroxide. A liner gradient from 22% to 73% B was run over 16 min. Sometimes a dilute eluent of fixed concentration is run for a few minutes before starting the gradient. This is illustrated in Figure 6.8. A potassium hydroxide eluent was used for the first 6 min, then a gradient of increasing KOH concentration was applied A large sample volume may be used to achieve very low limits of detection. The chromatogram in Figure 6.9 was obtained with a column and gradient similar to
6.2 Suppressed Anion Chromatography
that in Figure 6.8, but a 1000lL sample injection was used to detect sample anions in the very low ppb range.
Figure 6.7 Comparison of isocratic and gradient separation on an IonPac AS 12A column. Isocratic eluent. 20 mM sodium borate/ 18 mM sodium hydroxide. Gradient separation: Eluent A, water. Eluent B, 50 mM sodium borate/3.75 mM sodium hydroxide.
Gradient: linear from 22% to 73% B in 16 min. Peaks: (1) fluoride, (2) acetate, (3) formate, (4) chlorite, (5) bromate, (6) chloride, (7) carbonate, (8) nitrite, (9) bromide, (10) chlorate, (11) nitrate, (12) orthophosphate, (13) sulfate (from Ref. [10] with permission).
Figure 6.8 Separation of organic and inorganic anions on a 2mm AS15 column; a potassium hydroxide gradient is used: 9 mM from 0 to 6 min, 9 mM to 47 mM from 6 to 15 min (courtesy of Dionex Corp.).
145
146
6 Anion Chromatography
Figure 6.9 Ion separation at very low concentrations with a 1000 lL injection and a potassium hydroxide gradient: 8 mM from 0 to 6 min, 8 mM to 45 mM from 6 to 17 min (courtesy of Dionex Corp.).
6.2.3.2 Influence of Organic Solvents Even a cursory look at publications on ion chromatography will show that separation of inorganic anions has been given much greater prominence than that accorded to organic anions. One reason for this is that larger organic anions tend to have long retention times and broad peaks when conventional aqueous eluents are used. This is largely a consequence of the larger size and greater attraction of organic ions by the stationary phase. However, anion-exchangers currently used have an excellent tolerance for organic solvents. A higher percentage of an organic additive in the mobile phase reduces the hydrophobic attraction of the organic anions and results in faster elution and sharper peaks. Alternatively, a gradient can be run in which the percentage of organic solvent is increased through the chromatographic run. An example is shown in Figure 6.10 in which several nucleotide anions are separated. Another reason for the apparent neglect of organic sample ions in conventional ion chromatography is the success of ion-pair methods for determination of organic ions. Ion-pair methods are discussed in Chapter 9.
6.3 Nonsuppressed Ion Chromatography
Figure 6.10 Separation of nucleotides. Conditions: Hamilton PRP-X100 column 150 × 4.1 mm; A: 25 mM citric acid pH 5.4; B: 17.3 24 mM citric acid: acetonitrile, linear gradient 0 to 50% B from 0 to 10 min (courtesy Hamilton Co.).
6.3 Nonsuppressed Ion Chromatography 6.3.1 Principles
Ion chromatography with suppressed conductivity detection has been extremely successful in filling a large gap that previously existed in inorganic analysis. However, the necessity for a suppressor device does add to the complexity of the instrumentation. It also restricts the type of eluent that can be used and to some extent limits the separating ability of the method. In the middle to late 1970s several workers experimented with changing the capacity of ion-exchange resins [10]. It was found that a significant reduction in resin ion-exchange capacity will permit a substantial reduction in eluent ion concentration while maintaining retention times of analyte ions in a desirable range. This led to the development of a very simple form of ion-exchange chromatography with conductivity detection [11, 12]. At the time it was called single-column
147
148
6 Anion Chromatography
ion chromatography (SCIC) and is now referred to as nonsuppressed ion chromatography (NSIC). Nonsuppressed ion chromatography employs a conventional liquid chromatographic system with a conductivity detector cell connected directly to the outlet end of an ion-exchange separation column. No suppressor unit is required. The successful development of this method was made possible by three principal innovations: (i) the use of an anion- or cation-exchange resin of very low capacity (initially 0.007 to 0.04 mequiv g–1), (ii) an eluent with a low ionic concentration and hence a low conductivity, and (iii) an eluting ion in the eluent that has a significantly lower equivalent conductance than that of the analyte ions. For separation of anions, the ionic concentration of the eluent was typically 0.5 mM in the original work, although a somewhat higher concentration is now used. A solution containing the alkali metal salt of benzoic or phthalic acid is a suitable eluent. The benzoate anion has a limiting equivalent conductance of 32 (S cm2 equiv–1). By contrast, the equivalent conductance of common anions such as chloride, bromide, nitrate and sulfate is 70 to 80 (S cm2 equiv–1). Thus when one of these analyte anions passes through the detector cell, there is a significant increase in conductance over the background conductance. Figure 6.11 shows an early separation of chloride, nitrate and sulfate when 0.5 mM potassium phthalate at pH 6.2 is the eluent [10]; the sulfate peaks are a composite of five runs with sulfate concentrations ranging from 2.75 to 13.75 ppm. A plot of sulfate peak height or area vs. concentration is linear.
Figure 6.11 Separation of standard solutions of sulfate (2.75–13.75 ppm) from chloride and nitrate. Resin: XAD-1, 44-57 lm 0.04 mequiv g–1; eluent, 5.0 × 10-4 M potassium phthalate, pH 6.2.
6.3 Nonsuppressed Ion Chromatography
Figure 6.12 Separation of oxalate, thiosulfate and other anions. Conditions: Allsep anion column 100 × 4.6 mM, 4 mM phthalic acid, pH 4.2, 1.0 mL min–1 at 35 °C, conductivity detector (courtesy Alltech).
Figure 6.13 Speciation of arsenic anions. Allsep anion column, 100 × 4.6 mm. Mobile phase: 4 mM lithium p-hydroxybenzoate, pH 8.5. Conductivity detection. Peaks: (1) arsenite, (2) arsenate (courtesy Alltech).
149
150
6 Anion Chromatography
A separation of eight anions with a more contemporary IC system is shown in Figure 6.12. Quantitative results are possible in the low-to-medium ppm concentration range for each of the anions. A very broad variety of anions can be separated and quantitated by direct conductivity detection. The speciation of arsenic (III) and (IV) is illustrated in Figure 6.13 where arsenic and arsenate anions are easily separated in a single run. 6.3.2 Explanation of Chromatographic Peaks
When a sample containing salts of various anions, M+A–, M+B–, M+C–, etc. is injected, the sample anions will be taken up by the resin and exchanged for an equivalent amount of eluent anion, E–. The sample volume is rather small (usually 10–100 lL). The sample zone travels down the separator column at a rate equal to the eluent flow rate. This zone contains the cations present in the original sample and an eluent anion concentration equivalent to that of the sample anions. If the conductance of the cations and anions in this zone is greater than that of the eluent, a positive pseudo peak will be observed when this zone passes through the conductivity detector. However, if the conductance of ions in this zone is lower than that of the eluent, a negative pseudo peak will result. The pseudo peak in Figure 6.11 is negative while that in Figure 6.12 is first negative and then positive. This matrix peak causes no difficulty, provided elution of the sample anions is delayed until after the matrix peak has passed. After the sample plug has passed, the baseline is quickly restored to that obtained with the eluent alone. The solute anions gradually move down the column with the eluent. The total ion concentration in solution in the column is fixed by the eluent anion concentration because solute anion can only enter the solution phase by uptake of an equivalent number of eluent anions. The change in conductance when a sample solute band passes through the detector results from replacement of some of the eluent anions by solute anions, although the total ion concentration remains constant. This change is directly proportional to the sample concentration and the difference in equivalent conductance of the eluent anion and the sample anion. Conductivity detector response is discussed in greater detail in Section 6.3.6. 6.3.3 Eluent 6.3.3.1 General Considerations
The eluent must be carefully chosen if a conductivity detector is to be directly coupled to an anion-exchange separation column. In general, a good eluent is an aromatic organic anion which has a high selectivity coefficient for the anionexchange resin.
6.3 Nonsuppressed Ion Chromatography
As the affinity of the anion exchanger for an eluent anion increases, the eluent concentration needed to move sample anions along the ion-exchange column becomes progressively lower. A low eluent concentration is desirable because of the correspondingly low background conductance. There is probably some limit as to how great the affinity of the ion-exchange resin for the eluent anion should be; if the affinity becomes too great, sorption of the eluent anion (as opposed to ion exchange) is likely to occur. Hence, the major general criterion for eluent selection is a different equivalent conductance compared to the sample ions and a high enough affinity for the resin to promote effective elution of the sample ions. So far we have emphasized that the eluent should have a low equivalent conductance. But an eluent such as sodium hydroxide, with a high equivalent conductance, can be used in nonsuppressed IC with indirect detection. Many different eluents have been used for NSIC. The major types are discussed below.
6.3.3.2 Salts of Carboxylic Acids Lithium, sodium, potassium, or other salts of benzoic acid, phthalic acid, sulfobenzoic acid, citric acid, and others are useful eluents for anions. These are rather large organic anions that are less mobile than most inorganic anions and therefore have lower equivalent conductances. For example, Table 4.1 shows that the benzoate anion has a limiting equivalent conductance of 32 S cm2 equiv–, while chloride, nitrate, sulfate, and other typical sample anions have higher equivalent conductances (approximately 70 S cm2 equiv–). If a sodium benzoate eluent is used, the equivalent conductance is the sum of sodium ion (50) and benzoate (32), or 85 Scm2equiv–. The equivalent conductance of an anion is the sum of equivalent conductances of the sodium ion (50) and the anion (70), or 120 S cm2 equiv–. On an equivalent basis, this amounts to almost a 50% increase in conductance. Benzoate or p-hydroxybenzoate, which is more soluble and therefore to be preferred, are two of the most useful carboxylate ions for eluents. Phalate or hydrogen phthalate is also widely used. Benzoate salts are useful for separation of acetate, bicarbonate, fluoride, chloride, nitrite, nitrate, and other early-eluting anions. Divalent anion and other late-eluting anions such as thiocyanate and perchlorate are not eluted effectively by benzoate. The concentration of a benzoate eluent that should be used depends on the type and capacity of anion-exchange resin used, but is typically 0.5–5.0 mM. Potassium phthalate eluents can be conveniently prepared by dissolving potassium acid phthalate in pure water and adjusting the pH to around 6.1 to 7.0. In this pH range, the 2- phthalate anion is the predominant species. Phthalate is a more powerful eluent than benzoate and is used for separation of divalent anions and other late-eluting anions such as iodide, thiocyanate and perchlorate. Earlyeluting anions such as bicarbonate, acetate and fluoride are usually indistinguishable from the pseudo peak when a phthalate eluent is used. The eluting power of phthalate can be modified by changing the pH of the mobile phase. At pH 8.2 the
151
152
6 Anion Chromatography
hydrogen phthalate (HPh) form is primarily present in solution. Around pH 8.5, the ionic form has been converted to the phthalate (Ph–2), which is a more powerful eluent by virtue of its 2– charge. Alkali metal salts of benzenesulfonic acid are similar to benzoate in their eluting power although benzenesulfonate retains its 1– charge at a lower pH. In some cases, eluents in mixed ionic forms can elute both weakly and strongly retained sample ions in a single run. Solutions of p-hydroxybenzoic acid are a good example of a mixed eluent. At pH 8.5, the carboxyl group is completely ionized. The phenolic group is a weaker acid (pKa = 9.3) and becomes increasingly ionized at more alkaline pH values. Thus, by adjusting the pH, a mixture of 1– and 2– driving ions can be obtained, making it possible to elute ions from fluoride through sulfate in a single run.
6.3.3.3 Basic Eluents Anions of very weak acids such as arsenite, borate, carbonate, cyanide and silicate exist as anions only in basic solution. It is therefore necessary to use a basic eluent to separate these anions. A solution of sodium hydroxide can be used. Detection with sodium hydroxide is different than that with the organic salt and acid eluents. Since the hydroxide ion is more mobile and has a higher equivalent conductance than most other anions, the peaks for the sample anions appear as negative peaks (decreased conductance). However, the peak height (or area) is still a function of the amount of sample anion and the sensitivity is even better than with the more acidic eluents where positive peaks are obtained. A practical example of a chromatographic separation with indirect conductivity detection is illustrated in Figure 6.14. The theory of negative peaks for sample anions is easy to understand. Suppose we use 1.0 × 10–3 M sodium hydroxide as the eluent. The background conductance will be the sum of the sodium and hydroxide conductances and will be relatively high. Injection of a sample will result in the uptake of the anions by the resin column with an equivalent amount of resin hydroxide ion passing into solution. Once the matrix peak is through, the anion concentration of the column effluent will be constant, as fixed by the 1.0 × 10–3 M eluent concentration. When a sample anion, A–, is eluted from the column and passed through the detector, the eluent hydroxide ion will be decreased because a constant anion concentration must be maintained ([A–] + [OH–] = 1.0 × 10–3 M). The equivalent conductances of most anions range from about 30 to 80 S cm2 equiv–1, while the hydroxide ion has an equivalent conductance of 199. Thus, the conductance will decrease when a sample anion is eluted and the height (or area) of this negative peak will be proportional to the concentration of the anion. Although sodium hydroxide can be used as an eluent for cyanide, acetate, arsenite, fluoride, and other easily eluted anions, the hydroxide ion is a rather weak eluent for many anions. A sodium hydroxide solution containing small amounts of sodium benzoate (in about 1:10 molar ratio) behaves similarly to hydroxide (the peaks are usually still of decreasing conductance), but is a more powerful eluent than sodium hydroxide alone.
6.3 Nonsuppressed Ion Chromatography
Figure 6.14 Separation of weak acid anions. Conditions: Alltech Anion/R column, sodium hydroxide/sodium benzoate eluent, 1.5 mL min–1, conductivity detection. Hydroxide is the primary eluent for this separation and the weak
acid anions are detected with indirect conductivity. A small amount of benzoate is added to speed up the elution of the anions (courtesy Alltech).
6.3.3.4 Carboxylic Acid Eluents The detection sensitivity in anion chromatography can be improved markedly if the acidic form rather than the sodium salt of a carboxylic acid eluent is used. Gjerde and Fritz [13] separated seven inorganic anions with a l.25 mM solution of benzoic as the eluent. The change in conductivity was calculated to be almost eight times greater than the same separation with sodium benzoate. The key to this higher detection signal is that benzoic acid at the eluent concentration used is approximately 80% in the molecular form and only 20% ionized (H+ + B7–). When a sample anion, A–, passes through the detector an equivalent concentration of benzoic acid plus benzoate is replaced by highly ionized H+A–. Since the exchange and ionization equilibria are dynamic, approximately 80% of the H+ needed to pair with the sample anion, A–, comes from molecular benzoic acid. Several other organic acids have been used as eluents in anion chromatography. These include succinic, nicotinic, salicylic, fumaric and citric acid [14]. One objective was to find an eluent that is not adsorbed by the resin matrix and that attains equilibrium faster than benzoic acid. The following organic acids, listed in order of their increasing eluting power for sample anions, attain equilibrium fairly rapidly: nicotinic, succinic, citric, fumaric, salicylic.
153
154
6 Anion Chromatography
6.3.4 System Peaks
Suppose that we have attained equilibrium between a mobile phase containing sodium benzoate and an anion-exchange column, and nonsuppressed conductivity detection is to be used for a chromatographic run. When a small volume of an aqueous sample in injected, a zone is formed in the column that has a different conductance from that of the mobile phase. When this zone has moved through the column and passes through the detector cell, a baseline dip or peak is observed in the chromatogram. Most frequently, the sample conductance is lower than that of the eluent and a dip is observed. Eluent dips of this type were described by Gjerde and Fritz [13]. Stevens et al. [3] described the effect of the system peak in suppressed ion chromatography. Called a carbonate dip, the system peak was said to be the absent peak (from the injection) of the carbonic acid that is retained by the unexhausted portion of suppressor column. Injection of a dilute sample may cause another baseline disturbance that will be observed sometime later in a recorded chromatogram. In our example, the benzoate ion of the eluent is in equilibrium with molecular benzoic acid. C6H5CO2– + H+ > C6H5CO2H However, a second equilibrium process can occur in which the molecular form of the eluent solute is adsorbed by the resin matrix. This adsorption process also shifts the acid–base equilibrium so that a greater proportion of the eluent is in the molecular form. When a sample is injected that contains no benzoate or benzoic acid, a new equilibrium is established in which some of the adsorbed benzoic acid passes from the stationary phase into the mobile phase. After the zone occupied by the sample has passed, the resin re-equilibrates with the mobile phase to replace the adsorbed benzoic acid it has lost. This creates a ‘vacancy’ in the mobile phase that contains a lower total concentration of benzoate plus benzoic acid. In the example described, a decrease in the detector signal is observed when the vacancy passes through the detector. This is commonly called the ’system peak’ even if it is in the negative rather than the positive direction. This baseline disturbance is more pronounced when a considerable amount of the molecular form of the eluent is adsorbed and when the sample has a more basic pH than the eluent. This pH difference ionizes the adsorbed molecular species and desorbs it into the mobile phase. This action then creates a larger vacancy. In general, system peaks can be minimized by using a more hydrophilic ion exchanger and a more polar eluent. A more basic eluent pH is also helpful.
6.3 Nonsuppressed Ion Chromatography
6.3.5 Scope of Anion Separations
Anions of strong or medium-strength acids may be separated at a moderately acidic pH (as in Figure 6.12) or at a basic pH. However, anions of very weak acids such as borate, silicate or cyanide require a basic pH to be sufficiently ionized for separation and detection. Detection of these anions by suppressed conductivity is not very feasible because the suppressor converts the anions into the weak acid. However, anions of both weak and strong acids are nicely separated by elution with an alkaline solution of sodium hydroxide or sodium hydroxide and sodium benzoate as the eluent. (Figure 6.14) Detection was by indirect conductivity and the peaks represent a decrease in conductivity. Indirect conductivity detection with a sodium hydroxide eluent offers excellent sensitivity. As an example, formate (limiting equivalent conductance = 55 S cm2 equiv–1) replaces an equivalent amount of the more conductive hydroxide anion limiting equivalent conductance = 198 and thus gives a lower signal when the formate passes through the detector. 6.3.6 Sensitivity
The term ’sensitivity’ will be used here in its technically correct context as the change in detector signal per unit concentration. It should not be confused with detection limits, which are dependent on baseline noise. (Baseline noise is dependent on the signal magnitude, temperature variations, the electronics, etc. and varies from instrument to instrument.) In this section we shall examine the factors that determine the sensitivity that can be attained in anion chromatography. The approach taken utilizes Table 4.1 (of limiting equivalent conductances of various anions and cations) to estimate the changes in conductance that will occur using various chromatographic techniques. The calculations are verified by comparing the experimental and calculated background conductances of several eluents. Expected changes in conductances are calculated for separation of selected anions using different chromatographic techniques. Experimental measurements of the conductance of various eluents were made by Gjerde and Fritz [13]. The eluent was pumped through a Wescan model 213 conductivity detector until a steady reading was obtained. Most of the measurements were made at 22.0–22.3 °C, although a few were at a slightly higher or lower temperature. The cell constant was measured at 22.3 °C with a 1.00 × 10–3 solution of potassium chloride. It was found to be 33.0 cm–1 as calculated from the equation: k = GK, where k is the known specific conductance of the potassium chloride in lS, G is the conductance in lS, and K is the cell constant. Next, the conductance G was measured for several different eluents. For comparison, the conductances of the same eluents were calculated from a table of limiting ionic equivalent conductances, with the equation:
155
156
6 Anion Chromatography
G
lS
k k CI106 103 K
(6:1
where k is the limiting equivalent conductance of the cation or anion, C is the normality, and I is the fraction of the eluent which is ionized. The factor 106 converts G into lS. The term is divided into 103 because there are (approximately) 103 cm3 in 1 liter. The following illustrates this calculation for 100% ionized, 2 × 10–4 M potassium benzoate:
7:35 32:4
2 × 10 106 0:64 lS 1000
33:0 4
G
(6:2
The calculated and experimental values in Table 6.7 show quite reasonable agreement. The accuracy of these comparisons is limited by use of limiting equivalent conductances instead of the equivalent conductances at the concentrations actually used, imprecise temperature control, and somewhat limited accuracy in measuring the cell constant. Nevertheless, calculations from a table of equivalent conductances provide a useful estimation of expected experimental results.
Table 6.7 Background conductances of eluents[a}.
Eluent, concentration
Temp. [°C]
Conductance [S] Measured
Calculated
NaBz, 2 × 10-4 M, pH 7.0
22.0
0.50
0.50
NaBz, through catex
21.3
0.92
0.98
KBz, 2 × 10-4 M, pH 7.0
22.0
0.65
0.64
HBz, 8.4 × 10 M
22.2
2.15
2.32
-4
NaHCO3, 2 × 10 M, pH 7.5
22.8
0.53
0.57
NaHCO3, through catex
22.7
0.15
Pure H2O
23.0
0.042
Pure H2O, through catex
23.0
0.050
NaHCO3, 0.003 M + Na2HCO3, 0.0024 M
22.8
20.9
NaHCO3 + Na2HCO3. through catex
22.8
0.63
-4
-4
KPh, 2 × 10 M, pH 6.7
22.1
1.36
KPh, through catex
22.0
1.93
K3Cit, 2 × 10-4 M, pH 7.6
22.0
2.37
K3Cit, through catex
22.0
1.96
a Bz = benzoate; Ph=phthalate; Cit=citrate; catex=cation-exchange column in H+ form.
6.3 Nonsuppressed Ion Chromatography
The results in Table 6.7 show that it is better to use a sodium salt than a potassium salt as the eluent, because of the lower equivalent conductance of the sodium ion. A lithium salt would, presumably, be even better. The background conductance of some of the benzoate or phthalate eluents used in single-column chromatography are not greatly different from the background conductance of the more concentrated carbonate–bicarbonate eluent used in suppressed anion chromatography. For example, the background conductance of 2.0 × 10–4 M sodium benzoate is actually lower than that with the carbonate–bicarbonate eluent that is widely used in suppressed ion chromatography. Thus, the baseline noise can be similar in the two chromatographic methods. However, higher eluent concentrations, up to 5 mM are more common in nonsuppressed work.
Table 6.8 Anions that can be detected by direct UV.
Inorganic Anions
Organic Anions
Azide
Acetate
Bromate
Aromatic carboxylates and sulfonates
Bromide
Citrate
Chromate
Formate
Dichromate
Glutamate
Iodide
Lactate
Iodate
Malate
Nitrate
Malonate
Nitrate
Oxalate
Nobel metal complexes
Phenolates
(AuC4–,
Succinate
PtC62–,
Perrhenate Thiosulfate Vanadate
2–
PdCl4 , etc.)
Tartrate
157
158
6 Anion Chromatography Table 6.9 Detection limits (ppm) with a 10-L injection volume
(data from Ref. [23]). Anion
IO3– BrO3
–
NO3
Eluent 15 mM NaCl Sample in H2O
Sample in 20 000 ppm Cl–
0.12
0.26
0.32
1.2
0.03
0.06
–
Br
0.16
0.29
NO3–
0.08
0.12
MoO42–
0.18
0.40
0.70
1.32
1.70
2.6
0.57
0.82
0.60
1.0
CrO4
2–
VO3– I
– –
SCN
6.3.6.1 Conductance of a Sample Peak The general equation for detector response in ion chromatography was derived in Chapter 5. The following is a more specific derivation that takes into account that the eluent and sample are not always completely ionized. In ion chromatography, once the pseudo peak has passed through the column and detector the total cation and anion concentration in solution is constant and equal to that of the eluent. Since we are dealing with an ion-exchange process, a sample anion moves down the column by exchange with an equivalent amount of eluent anion. When a sample anion passes through the detector, the eluent anion concentration is decreased by an amount equivalent to the concentration of the sample anion. Thus, the change in conductance will be determined by the relative equivalent conductances of the sample and eluent ions. In nonsuppressed IC the sample anion must have a higher equivalent conductance than that of the eluent anion in order to get a positive peak. The relative sensitivity of different methods of anion chromatography can be compared by calculating the change in conductance resulting from the replacement of the eluent anion with an equivalent concentration of a sample anion. The eluent cation also affects the conductance, but in single-column chromatography the cation contribution to conductance remains constant. When a suppressor is used and the eluent cation is exchanged for a new cation, the cation contribution will of course change. Therefore, we shall consider only the cations which are present in the detector cell so that the equations developed can apply to both single-column and suppressor ion chromatography.
6.3 Nonsuppressed Ion Chromatography
Equation (6.1) can be used to calculate the change in conductance when a sample anion is present. Let CE be the concentration of the eluent anion, and CS be the concentration of the sample anion. Taking all forms of eluent and sample into account: (6.3) Eluent: HE > H+ + E– (6.4) Sample: HS > H+ + S– And the ion-exchange reaction is: > Resin-E + H+ + S– Resin-S + H+ + E– ↑↓ ↑↓ HE HS
(6.5)
The background conductance [taken from Eq. (6.1)], where there is no sample present, is given by the equation: GB
kE kE CE IE 10 3 K
(6:6
where E+ and E– are the eluent cation and anion, respectively, and IE is the fraction of the eluent that is ionized: IE
E HE E
(6:7
and CE = [HE] + [E–]
(6.8)
The concentration of E– during a sample peak elution will be (CE – CS)IE . The concentration of S– during a sample peak elution is CSIS. The conductance of the eluent and sample during a peak elution, GS, is described with Eq. (6.9). The conductance is the sum of the contributions from the eluent ions and the sample ions. GS
kE kE
CE 10 3 K
CS IE
(6:9
Subtraction of Eq. (6.8), the background conductance, from Eq. (6.9), the peak conductance, gives the change in conductance when a sample peak is eluted: GS
GB DG
kE kS IS
kE kE IE
CS 10 3 K
(6:10
where DG is the detector response. The change in conductance of a sample peak is greater with suppressed conductivity than with nonsuppressed detection. This is largely the result of the very high equivalent conductance of H+ which is present in exactly the same concentration as the analyte anion. The baseline conductivity is also lower with sup-
159
160
6 Anion Chromatography
pressed conductivity than in the nonsuppressed mode. Detection limits are also affected by the goodness of temperature control, sample volume and the inherent sensitivity of the conductivity detector. With direct injection of a 50lL sample, the detection limits for chloride, nitrate and sulfate in drinking water have been estimated to be around 10 ppb (lg L–1) using suppressed conductivity detection [15]. Separation of common anions, each at the 1 ppm level, is now routine. Detection at low ppb range is possible when a hydroxide eluent and a large sample volume are used (Figure 6.14). However, detection of anions at the low ppb level has also been realized with nonsuppressed conductivity with a nicotinic acid eluent [14]. While very low detection limits are definitely needed for analysis of high-purity water samples, this aspect of ion chromatography should not be overly emphasized. Determination of anions at the low ppm (mg L–1) level is quite adequate for a great many samples. In fact, it is quite common to dilute a sample before injection in order to bring the analyte concentration into the desired range. This process is often called ‘dilute and shoot.’ Suppressed and nonsuppressed methods are equally valid for many sample types.
6.4 Coated Columns
Coating porous polymeric particles of uniform size with a thin layer of an ionic surfactant offers a convenient way to prepare exchangers suitable for IC. This may be accomplished either by coating loose particles and packing them in a column or by passing the coating solution through a standard HPLC column. Cassidy and Elchuk [16] demonstrated that the coating is stable when aqueous eluents are used. Coated columns can also be prepared with different exchange capacities by varying the concentration and aqueous-organic solvent composition of the coating solution. Duval and Fritz [17] observed a linear relationship between adjusted retention times of inorganic anions and exchange capacity in the range of 0.002– 0.012 mequiv g–1. Several investigators have obtained good separations of anions on coated columns. Connolly and Paull [18] obtained fast ion-chromatographic separation of common anions using a 3-cm silica C18 column permanently coated with DDAB. Short columns were found to have better chromatographic efficiency than longer columns. Fritz, Yan and Haddad [19] found that coating a packed HPLC column with a mixture of a nonionic surfactant and CPC resulted in a more efficient column than one coated with CPC alone. The most efficient column was obtained by coating in steps – first with Triton-X and then with CPC. Reversal of the coating sequence (i.e., using CPC first) gave a column with almost no measurable anionexchange capacity. Figure 6.15 shows an excellent separation of several anions with a low concentration (2.0 mM) of sodium perchlorate as the eluent. Faster separation of the late-eluting anions would require a higher eluent concentration.
6.4 Coated Columns
Figure 6.15 Separation of anions on HPLC column (150 × 4.6 mm) coated with Triton X-100 and then with cetylpyridinium chloride. Eluent: 2 mM sodium perchlorate. UV detection at 210 nm. Peaks: (1) acetate, (2) nitrite,
(3) bromide, (4) nitrate, (5) oxalate, (6) iodide, (7) 4-hydrobenzoate, (8) chromate, (9) thiocyanate, (10) phthalate and (11) benzoate (from Ref. [19] with permission).
Several research groups throughout the world have taken advantage of the low resistance to mobile-phase flow offered by short monolithic columns. Pelletier and Lucy [20] separated several anions on the 1-cm, coated monolith column using a simple syringe pump (Figure 6.16). Although the peak efficiency is reasonably good, it should be much better. Extra-column peak broadening was a major factor. One end of the column holder connector had five times larger tubing than that used in the rest of the system. However, broadening from the suppressor unit (~35 lL internal volume) was the major source of extra-column broadening. Another example of a fast separation with a surfactant-coated monolith is shown in Figure 6.17 [21] In this case a column only 2.5 cm long was eluted with 10 mM sodium sulfate using an unusually fast flow rate of 3 mL min–1. A monolith coated with cationic latex particles has been used to separate inorganic anions [22]. Although a reasonable separation of five anions was obtained, the retention time for sulfate was quite long (22.9 min). Plate heights of the anion peaks were higher than would be expected from a really efficient ion-chromatographic column. This was attributed to extra dead volume in the suppressor, detector cell and connecting tubing.
161
162
6 Anion Chromatography
Figure 6.16 Low-pressure separations with syringe pump on a 0.5 cm coated monolith column. Eluent: 6.0 mM 4-cyanophenol, pH 73. Suppressed conductivity detection (from Ref. [20] with permission).
Figure 6.17 Anion separation on short Chromolith Flash RP-18e column, 25 × 4.6 mm, coated with cetylpyridinium chloride. Eluent: 10 mM sodium sulfate, UV detection 210 nm. Peaks: (1) acetate, (2) nitrite, (3) bromide, (4) nitrate, (5) tungstate, (6) chromate (from Ref. [21] with permission).
6.5 Optical Absorbance Detection
6.5 Optical Absorbance Detection 6.5.1 Introduction
Ion chromatography was originated with conductivity detection and has grown up with suppressed and nonsuppressed conductivity as the leading forms of detection. However, conductivity detection does have some drawbacks. Temperature affects conductance, so a constant-temperature device is an essential part of the IC system. Suppressed conductivity requires an extra suppressor unit in the system, which inevitably contributes to peak broadening. Conductivity is in effect a ‘universal’ detection system because all ions passing through the detector contribute to the measured conductance. Although universal detection can be an advantage, it is often better to use a more selective detector. Several such detectors will be discussed in the following sections, but a UV–Vis spectrophotometric detector is one of the very best. It is probably best to use a variable-wavelength UV–Vis spectrophotometer and not a fixed wavelength instrument. These detectors are very stable and have excellent sensitivity. We have obtained an excellent baseline and chromatograms when the detector was set on 0.003 AU full scale. The variable-wavelength feature is useful because detectors of this type can be used to take advantage of rather small differences in absorbance spectra of various ions. Absorbance detection has been applied to ion analysis through two different approaches: direct detection of the sample ion and indirect detection. In some cases, a post-column, color-forming reagent can be added to the column eluate to detect sample ions. 6.5.2 Direct UV Absorption
Table 6.8 lists some common inorganic and organic anions that absorb sufficiently at wavelengths slightly above 210 nm to be detected by direct UV. Many anions (fluoride, chloride, perchlorate, sulfate and others) are UV-transparent at these wavelengths. All that is needed is to select a suitable nonabsorbing anion for use in the mobile phase and carry out the IC separation using direct UV detection of the analyte ions. For example, sulfate is a good eluting ion by virtue of its 2– charge. Perchlorate is an even stronger eluting ion because of its strong affinity for anion-exchange sites on the column packing. Direct UV detection is probably the easiest and one of the most sensitive ways to monitor the separation of UVabsorbing anions. Direct UV absorbance detection is often applicable for samples where the use of conductivity detection may be problematic. The determination of trace levels of anions in the presence of very high levels of sample matrix ions is a fairly common analytical problem. Seawater, for example, contains approximately 0.50 M
163
164
6 Anion Chromatography
sodium chloride, corresponding to a chloride concentration of approximately 18 000 ppm. Ion-chromatographic separation of trace anions becomes difficult in this situation because the large chloride peak may obscure peaks of the analyte anions. A good answer to this problem is to use an eluent containing the same ion as the sample matrix. Marheni, Haddad and McTaggart used eluents containing chloride or sulfate for samples containing a high concentration of chloride or sulfate, respectively [23]. With a 50 × 4.6 mm column packed with a polymethacrylate anion exchanger of 30 ± 3 l mL–1 capacity, the chromatographic behavior of several anions was studied by direct photometric detection at 210 nm. Linear plots were obtained for log k (log retention factor) vs. log NaCl concentration in the eluent. The eluent chosen for practical separations contained 15 mM sodium chloride and 5 mM phosphate buffer at pH 6.5. Separation of ten anions gave very similar chromatograms for samples in water alone and in 20 000 ppm chloride, although the retention times were slightly shorter in the latter case. The mechanism of this separation is as follows: Introduction of the sample containing 20 000 ppm chloride (560 mM) results in a solution at the top of the column that is much higher than the eluent (15 mM chloride). But because the anion exchanger is already in the chloride form, the sample zone of very high chloride can pass rapidly through the column at the same linear flow rate as the eluent. Although the retention factors of the analyte anions are lower in the presence of the high sample chloride concentration, their k values quickly return to normal as soon as the high-chloride zone has passed. Table 6.9 shows that direct UV detection gives low detection limits for many inorganic anions, even when only a small (10 lL) sample is injected. Low limits of detection were also obtained for samples containing 20 000 ppm chloride. 6.5.3 Indirect Absorbance
In this detection mode, an eluent anion is chosen that absorbs in the visible or UV spectral region. The elution of sample anions is monitored by measuring the decrease in absorbance at the detection wavelength as transparent sample ions replace a fraction of the absorbing eluent anions. Under conditions where the solute anion (S–) is fully dissociated, which is the usual case for IC, the change in absorbance (DA) is given by: DA = (eS- – eE- ) CS- – b
(6.11)
where eS- and eE- are the molar absorptivities of the solute and eluent anions, CSis the molar concentration of the solute anion, and b is the cell path length. Ordinarily, the eluent anion and detection wavelength will be selected so that the absorbance will decrease and the solute peaks will be in the negative direction.
6.5 Optical Absorbance Detection
The concentration of the eluent anion must be high enough to elute the sample ions within a reasonable time and to exceed the concentration of any sample anion at its peak maximum. However, if the concentration of E– greatly exceeds that of S–, the background absorbance will be relatively high and the noise of peak detection will be poor. So while a reasonable eluent concentration is needed to promote ion exchange with the column, the concentration should still be as low as possible to reduce noise. To achieve the best sensitivity, eE- should be as high as possible so that the difference in molar absorptivities in Eq. (6.14) will be large. In summary, the eluting anion selected for indirect UV detection should have a strong affinity for the ion exchanger so that a relatively low concentration can be used, and it should have a high molar absorptivity. Several anions have proven to be effective for indirect photometric detection. Figure 6.18 shows an excellent separation of common anions with 2.0 mM potassium phthalate, pH 6.0, with indirect detection at 280 nm. By switching to 4.0 mM p-hydroxybenzoic acid, pH 8.5 with 2.5% methanol, detection at 310 nm, and reducing the column length to 10 m, the same anions plus phosphate could be separated in only 2.3 min [24].
Figure 6.18 Separation of common anions with indirect UV detection at 280 nm. Eluent: 2.0 mM potassium hydrogen phthalate pH 6.0, 1.2 mL min–1 (courtesy Hamilton Co.).
165
166
6 Anion Chromatography
A sodium molybdate eluent with indirect detection at 250 nm was found to provide an excellent separation and a very sensitive detection of inorganic anions [25]. The separation in Figure 6.19 was performed on an experimental latex-coated anion exchanger. Miura and Fritz investigated the use of polycarboxylic acid salts as eluents in anion chromatography [26]. At alkaline pH values, 1,3,5-benzenetricarboxylic acid (BTA) exists as the 3– anion and pyromellitic acid (1,2,4,5-benzenetetracarboxylic acid) can exist as the 4– anion.
Figure 6.19 Separation of anions on a latex-coated anionexchange column of very low exchange capacity (27 leq g–1). Eluent: 4 mM sodium molybdate. Detection: Indirect UV at 250 nm. Peaks: 1=ethylsufonate, 2=propylsulfonate, 3=chloride, 4=nitrite, 5=bromide, 6=nitrate, 7=sulfate (from Ref. [25] with permission).
6.6 Detection
Potentiometric detection of anions is feasible when an electrode is available that responds quickly, reversibly and reproducibly to the concentration (or more precisely to the activity) of sample ions. It is often possible to detect a given ion or class of ions with excellent selectivity. For example, solid-state or crystalline ion selective electrodes have been used in IC to detect halide anions. The fluoride ionselective electrode is particularly selective [27]. A copper wire electrode has been used to detect anions such as iodate, bromide and oxalate [28].
6.6 Detection
A metallic silver electrode responds rapidly and reproducibly to the activity of free silver ions in solution. At 25 °C: E = E° + 0.05915 (log Ks – log aX–)
(6.12)
where E is the electrode potential (V) and E° is the standard reduction potential for Ag+ + e– = Ag°. If the silver metal is coated with a slightly soluble solid (AgX), aAg+ and hence E is determined by the concentration of X– in solution via the solubility product of AgX. KS = aAg+ × aX– ; aAg+ = KS/aX–
(6.13)
Combining Eqs. 6.12 and 6.13: E = E° + 0.05915 (log KS – log aX–)
(6.14)
Lockridge et al. [29] prepared a number of silver wire electrodes coated with one of the following insoluble silver salts: AgCl, AgBr, AgI, AgSCN, Ag2S and Ag2PO4. The coating process was performed by anodic oxidization of a silver wire for 3–5 min in a solution of the appropriate anion. The electrodes coated with AgCl or AgSCN were found to give the best response with either of these electrodes. A reproducible potentiometric response could be obtained for any of the halide or pseudohalide anions but only if the electrode was conditioned by dipping it into a solution of the analyte 3 or 4 times before use as a detector. Electron micrograph photos showed the surface to be a composite of several silver salts covering the underlying silver chloride precipitate on the electrode surface. The cell for potentiometric detection is shown in Figure 6.20. The eluate from the IC column flows past the coated silver indicator electrode and then out past a small silver/silver chloride reference electrode (not shown in the figure). The eluate itself serves as a salt bridge between the two electrodes. A silver wire electrode coated with AgCl has excellent selectivity. An ion chromatogram obtained with 4.5 mM sodium perchlorate as the eluent gave sharp peaks for 1 mM chloride, bromide, iodide, thiocyanate and thiosulfate but no
Figure 6.20 Flow cell for potentiometric detection with coated silver electrode (from Ref. [29] with permission).
167
168
6 Anion Chromatography
response to equimolar concentrations of nitrate, phosphate, carbonate, sulfate and acetate. The background response of this potentiometric system is virtually unaffected by changes in the concentration of sodium perchlorate or sodium sulfate eluent. This permits the use of gradient elution as shown in Figure 6.21, where a gradient of 3.5–10.0 mM sodium perchlorate was used to obtain a fast separation of five halide and pseudohalide ions.
Figure 6.21 Gradient elution with potentiometric detection. Eluent: 3.5–10.0 mM sodium perchlorate; flow rate, 1.6 mL min–1; injection volume, 20 lL; analyte concentration, 1.0 mM (from Ref. [29] with permission).
6.7 Pulsed Amperometric Detector (PAD)
The principles of the PAD were discussed in Chapter 4. Perhaps the major use of the PAD in anion chromatrography has been for the detection of carbohydrates. At pH around 11 or higher sugars become anionic and can be separated by anion chromatography. For many years a somewhat awkward post-column derivatization reaction was used for detection of carbohydrates after a chromatographic separation, but the use of a PAD now provides a simple and direct detection method. All carbohydrates (aldoses and ketoses) and polyalcohols produce a large anodic peak response at about +0.15 V. The peak for glucose corresponds to a reaction
6.7 Pulsed Amperometric Detector (PAD)
with n approaching 10 equiv mol–1 for fluid velocities typical of flow-through detection cells. This n value is consistent with an oxidative cleavage for the C1–C2 and C5–C6 bonds to form two moles of formate and one mole of dicarboxylate dianion [24]. Either a gold or platinum electrode may be used in a PAD, although gold electrodes are more popular. One reason is that dissolved oxygen contributes a cathodic response at a platinum electrode over the entire useful range for anodic detection. Serious oxygen interference at the gold electrode can be avoided by careful selection of detection potential. A simple but useful example of pulsed-amperometric detection is shown in Figure 6.22 where glucose, fructose and a trace of sucrose are determined in honey by anion chromatography. Much more complex samples can be resolved using gradient elution. This is demonstrated in Figure 6.23 where 18 carbohydrates were separated. Elution of the later peaks is speeded up by gradually reducing the eluent pH to inhibit ionization of the carbohydrates. However, post-column addition of 0.4 M sodium hydroxide was needed to restore the effluent to a pH sufficiently alkaline for effective pulsed amperometric detection.
Figure 6.22 Sugars in honey. Conditions: Hamilton RCX-10 column 250 × 4.1 mm, 30 mM sodium hydroxide, pulsed amperometric detection with dual gold electrode. E1 = 350 mV for 166 ms, E2 = 900 mV for 166 ms, E3 = –800 nV for 333 ms (courtesy of Hamilton Co.).
169
170
6 Anion Chromatography
Figure 6.23 Separation of carbohydrate mixture. Pulsed amperometric detection (model I) at Au; Ag/Ag/Cl reference; Edet = +0.10 V (tdet = 610 ms, tdel = 400 ms, tint = 200 ms); Eoxd = +80 V (toxd = 120 ms); Ered = –0.60 V (tred = 200 ms). Column: Dionex AS-6 Carbopac. Solvents: (A) 100 mM NaOH, (B) 50 mM NaOH +0.5 M NaOAc, (C) H2O. Elution: isocratic (0–6 min) with A–C (50:50); linear gradient (6–15 min) to A–B (50:50); isocratic
(15–21 min) with A–B (50:50). Post column addition: 0.4 M NaOH. Peaks: a = inositol; b = xylitol; c = sorbitol; d = mannitol; 3 = fucose; f = rhamnose; g = arabinose; h = glucose; I = xylose; j = fructose; k = sucrose; l = unknown; m = maltose; n = maltotriose; o = maltotetraose; p = maltopentaose; q = maltohexaose; r = maltoheptaose (from Ref [32] with permission).
6.8 Evaporative Light Scattering Detector (ELSD)
The ELSD is in some respects the perfect detector because actual mass of analytes is detected directly without converting a signal, such as optical absorbance, to concentration. The ELSD responds to all analytes as long as they are appreciably less volatile than solutes in the mobile phase. It is not necessary for the analytes to have a chromophore or other identifying chemical structure. The principles of the ELSD were described in Chapter 4. Basically, generation of a signal involves a three-step process: 1. Nebulization. Column effluent passes through a needle where it mixes with nitrogen gas and forms an aerosol. 2. Evaporation. In this step the aerosol evaporates, leaving a fine mist of dried analyte particles in solvent vapor. 3. Detection. The fine sample particles pass through an optical cell where they pass through a laser light beam. The light scattered by the fine particles generates an electrical signal.
6.8 Evaporative Light Scattering Detector (ELSD)
A mobile phase that contains only chemicals that are volatile, or can be converted to volatile substances, is a necessary requirement for the use of evaporative light scattering as a chromatographic detector. Nonionic organic analytes are typically separated on a reversed-phase column using a mobile phase containing acetonitrile and water. Organic anions can be separated on an anion-exchange column and detected by ELSD using a mobile phase containing an inorganic ammonium salt, which gives volatile products when heated in the detector. The ELSD may be used in conjunction with an eluent containing nonvolatile ions provided that an appropriate suppressor is inserted between the column exit and the detector. This technique is illustrated by the determination of carbohydrates in cola drinks [30]. The separation is depicted in Figure 6.24. Carbohydrates such as glucose and sucrose are anionic at high pH values and can be separated by anion chromatography with a dilute KOH eluent. As the mobile phase leaves the column and enters the suppressor, nonvolatile KOH is converted to water; hence a very low background signal is obtained. Only a mist of carbohydrate particles remain to be detected upon evaporation in the ELSD.
Figure 6.24 Determination of carbohydrates in a cola drink. Dionex PAI column, ELSD detector. Eluent: 32 mM KOH. Peaks: 1 = glucose; 2 = fructose; 3 = sucrose (from Ref. [30] with permission).
171
172
6 Anion Chromatography
6.9 Inductively Coupled Plasma Methods (ICP) 6.9.1 Atomic Emission Spectroscopy (AES)
ICP–AES is often used to determine the concentrations of various elements in a sample. However, an element may be present in a variety of chemical forms or species. By coupling an ICP–AES detector to an ion-chromatographic column, a more complete description of the sample species can be obtained. Such a coupling generally requires a nebulizer to introduce the column effluent into the ICP. Conventional pneumatic nebulizers operate at about 1 mL min–1 sample flow and may introduce as little as 1% of the sample into the plasma. A newer direct-injection nebulizer (DIN) operates at sample flow rates only 5 to 10% that of a conventional nebulizer [31]. The separation of various arsenic species is a good example of the application of ICP–AES detection to anion chromatography [32]. A microbore column 10 cm × 1.7 mm i.d. was used with a low flow rate (<100 mL min–1). The column was packed with a low-capacity anion-exchange material (0.05 mequiv g–1 and solution containing 5 mM ammonium carbonate and 5 mM ammonium bicarbonate at pH 8.6) served as the mobile phase. The column hardware was connected directly to the inlet of the DIN–ICP–AES via a short length of 0.3mm i.d. PEEK tubing. A separation of arsenite, arsenate and monomethylarsonate (MMA) is shown in Figure 6.25. The detection limit for arsenic was 10 lg L–1 and the minimum detectable quantity was 100 pg. It was also possible to separate and detect selenium(IV) and (VI) under similar conditions.
Figure 6.25 Anion-exchange separation of arsenic species. A 5.0 mM ammonium carbonate/5.0 mM ammonium bicarbonate eluent was used at 80 lL/min flow rate and a 10 lL sample volume. Peaks (in order): As(III), MMA, As(V) (from Ref. [27] with permission).
References
6.9.2 Inductively Coupled Plasma Mass Spectrometry (ICP-MS)
It is also possible to use mass spectroscopy (MS) detection for anionic species separated by IC. Limits of detection are very low and the use of MS permits the measurement of different isotopes of the same element. The separation of selenite (SeO32–) and selenate (SeO42–) by anion chromatography with ICP–MS detection is a typical example [33]. The connection of the anion-exchange column and direct injection nebulizer to the mass spectrometer and chromatographic conditions are similar to those employed with AES detection [32]. It is important to minimize the dead volume between the column and DIN, otherwise significant broadening of the separated peaks can occur. Retention time was 3.4 min for SeO32– and 9.0 min for SeO42–. Separation of 5 ng of each selenium species gave an RSD of 3.0%. The absolute detection limit was approximately 15 pg for each species. The relative amounts of 78Se and 74Se could also be measured. It is also possible to couple a small anion-exchange column to the MS with an electrospray for sample introduction. The conditions here are mild enough for the electrospray molecules to survive.
References [1] H. Small, T. S. Stevens and W. C. Bau-
man, Novel ion exchange chromatographic method using conductimetric detection, Anal. Chem., 47, 1801, 1975. [2] S. Rabian, J. Stillian, V. Barreto, K. Friedman and M. Toofan, New membrane-based electrolytic suppressor device for suppressed conductivity detection in ion chromatography, J. Chromatogr., 640, 97, 1993. [3] T. S. Stevens, J. C. Davis and H. Small, Hollow fiber ion exchange suppressor for ion chromatography, Anal. Chem., 53, 1488, 1981. [4] J. Stillian, An improved suppressor for ion chromatography, LC Mag., 3, 802, 1985. [5] D. L. Strong and P. K. Dasgupta, Electrodialytic membrane suppressor for ion chromatography, Anal. Chem., 61, 939, 1989. [6] A. Henshall, S. Rabin, J. Statler and J. Stillian, Recent development in ion chromatography detection: the selfregenerating suppressor, Am. Lab., 24, 20R, 1992.
[7] D. T. Gjerde and J. V. Benson, Suspen-
[8]
[9]
[10]
[11]
[12]
sion postcolumn reaction detector for liquid chromatography, Anal. Chem., 62, 612, 1990. P. E. Jackson, P. Jandik, J. Li, J. Krol, G. Bondoux and D. T. Gjerde, Practical applications of solid-phase reagent conductivity detection in ion chromatography, J. Chromatogr., 546, 189, 1991. D. T. Gjerde, D. J. Cox, P. Jandik and J. B. Li, Determination of analytes at extreme concentration ratios by gradient ion chromatography with solid-phase reaction detection, J. Chromatogr., 546, 151, 1991. D. T. Gjerde and J. S. Fritz, Effect of capacity on the behavior of anionexchange resins, J. Chromatogr, 176, 199, 1979. D. T. Gjerde, J. S. Fritz and G. Schmuckler, Anion chromatography with lowconductivity eluents, J. Chromatogr., 186, 509, 1979. D. T. Gjerde, G. Schmuckler and J. S. Fritz, Anion chromatography with low-
173
174
6 Anion Chromatography
[13]
[14]
[15]
[16]
[17]
[18] [19]
[20]
[21]
[22]
[23]
[24]
[25]
conductivity eluents. II, J. Chromatogr., 187, 35, 1980. D. T. Gjerde and J. S. Fritz, Behavior of various benzoate eluents for ion chromato-graphy, Anal. Chem., 53, 2324, 1981. J. S. Fritz, D. L. DuVal and R. E. Barron, Organic acids as eluents for single-column anion chromatography, Anal. Chem., 56, 1177, 1984. J. Weiss, Ion Chromatography, VCH, Weinheim, Germany, 2nd Ed. 1995, p348. R. M. Cassidy and S. Elchuk, Dynamically coated columns for the separation of metal ions and anions by ion chromatography, Anal. Chem., 54, 1631, 1982. D. L. Duval and J. S. Fritz, Coated anion-exchange resins for ion chromatography. J. Chromatogr., 295, 89, 1984. D. Connolly and Paull, J. Chromatogr. A, 953, 209, 2002. J. S. Fritz, Yan Zhu and P. R. Haddad, Modification of ion chromatographic separations by ionic and nonionic surfactants, J. Chromatogr. A, 997, 21, 2003. S. Pelletier and C. A. Lucy, Achieving rapid low-pressure ion chromatography separations on short silica-based monolithic columns, J. Chromatogr. A, 118, 12, 2006. J. Li, Y. Zhu and Y. Guo, Fast determination of anions on a short coated column, J. Chromatogr. A, 118, 46, 2006. K. M. Glenn, C. A. Lucy and P. R. Haddad, Ion chromatography on a latex-silica monolith column, J. Chromatogr. A, 1155, 8, 2007. Marheni, P. R. Haddad and A. R. McTaggart, On-column matrix elimination of high levels of chloride and sulfate in nonsuppressed ion chromatography, J. Chromatogr., 546, 221, 1991. Hamilton Co., Reno, NV, USA. Hamilton applications handbook, Application #94. L. M. Warth, J. S. Fritz, J. O. Naples, Preparation and use of latex-coated res-
ins for anion chromatography, J. Chromatogr., 462, 165, 1989. [26] Y. Miura and J. S. Fritz, Benzenepolycarboxylic acid salts as eluents in anion chromatography, J. Chromatogr., 482, 155, 1989. [27] M. P. Keuken, J. Slanina, P.A.C. Jongjan and F. P. Bakker, Optimization of ion chromatography using commercially available detection system and software, J. Chromatogr., 439, 13, 1988. [28] P. W. Alexander, P. R. Haddad and M. Trojanowicz, Potentiometric detection in ion chromatography using a metallic copper indicator electrode, Chromatographia, 20, 179, 1985. [29] J. E. Lockridge, N. E. Fortier, G. Schmuckler and J. S. Fritz, Potentiometric detection of halides and pseudohalides in anion chromatography, Anal. Chim. Acta, 192, 41, 1987. [30] J. Li, M. Chen and Y. Zhu, Separation and determination of carbohydrates in drinks by ion chromatography with a self-generating suppressor and an evaporative light-scattering detector, J. Chromatogr. A, 1155, 50–56, 2007. [31] D. R. Wiederin, R. E. Smyczek and R. S. Houk, On-line standard additions with direct nebulization for inductively coupled plasma mass spectrometry, Anal. Chem., 63, 1626, 1991. [32] D. T. Gjerde, D. R. Wiederin, F. G. Smith and B. W. Mattson, Metal speciation by means of microbore columns with direct-injection nebulization by inductively coupled plasma atomic emission spectroscopy, J. Chromatogr., 640, 73, 1993. [33] S. C. K. Shum and R. S. Houk, Elemental speciation by anion exchange and size exclusion chromatography with detection by inductively coupled plasma mass spectrometry with direct injection nebulization, Anal. Chem., 65, 2972, 1993.
175
7 Cation Chromatography 7.1 Introduction
The basic principles of cation chromatography are very simple. A suitable cationexchange column is used in conjunction with the auxiliary equipment typical of liquid chromatography: eluent reservoir, pump, injection loop, guard column, detector and detector cell, and a data acquisition device. After equilibration of the system so that a steady baseline is obtained, an eluent is pumped through the system, a sample is injected, and the cationic analytes are separated. A common separation mechanism is one in which the sample cations are pushed at different rates down the column by the cations in the mobile phase. However, a number of cations have very similar selectivity coefficients for the cation exchanger and cannot be separated by this method. A second general separation method uses a complexing reagent (or mixture of reagents) to move the sample cations down the column by partial complexation. This latter method has been used very successfully for separation of metal cations such as the divalent transition metals and the lanthanides. A number of experimental variables can be manipulated in the quest to obtain a good analysis for an analytical sample. Each of these can exert a major effect. 1. Column. Various functional groups affect selectivity. Exchange capacity and column length affect the time required for a separation. Compatibility with organic solvents and pH stability also need to be considered. 2. Eluent and Detector. When selecting the eluent and the detector for a particular separation, these need to be considered together and not separately. The chemical type and concentration of eluent cation must be able to separate the analyte cations within a reasonable time, but the eluent must also be compatible with the detector. Monovalent separation can be obtained. Gradient elution with a programmed change in eluent composition may be needed for more complex samples. 3. Column Temperature. This is often a neglected parameter. A higher column temperature may improve chromatographic Ion Chromatography, 4th Ed. James S. Fritz and Douglas T. Gjerde Copyright © 2009 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim ISBN: 978-3-527-32052-3
176
7 Cation Chromatography
efficiency and also allow a separation to be completed more quickly. 4. Organic solvents. A low percentage of an organic solvent such as methanol or acetonitrile is sometimes added to the mobile phase to facilitate mass transfer between the two phases. A somewhat higher percentage of organic solvent in the mobile phase sharpens the peaks and reduces the retention times of organic ions. Cation chromatography can even be performed in a nonaqueous solvent. These experimental variables are considered in greater detail in the following sections of this chapter. Illustrative examples are included in the discussion.
7.2 Columns
Several types of cation-exchange column have been used successfully for cation chromatography. Almost all of the cation exchangers are of low exchange capacity, so that an eluent of a fairly low ionic concentration can be used. Low-capacity
Table 7.1 Dionex Ion Pac cation-exchange columns.
Name
Type
Special features/applications
CS 10
Sulfonate
For inorganic and biogenic amine cations with HCl or diaminopropionic acid eluent
CS 11
Sulfonate
Moderate capacity
CS 12
Carboxylate
High capacity, methanesulfonic acid eluent
CS 12A
Carboxylate and phosphonate
High capacity, for fast isocratic separation of inorganic cations in diverse sample matrices
CS 14
Carboxylate
Medium capacity, for amine cations
CS 15
Carboxylate and Crown ether
High capacity, for very disparate conc. ratios of sodium and ammonium at elevated temperature.
CS 16
Carboxylate
Sodium and ammonium in power plant and environmental waters.
CS 17
Carboxylate
For gradient separation hydrophobic and biogenic amines with aqueous eluents at elevated temperature.
CS 18
Carboxylate
For separation of polar and biogenic amine cations at elevated temperatures
ProSwift Cation
Monolith
High resolving power with improved speed, low back pressure
7.2 Columns
cation-exchange resins are obtained by superficial sulfonation of styrene-divinylbenzene copolymer beads, as originally described by Small et al. [1] and by Fritz et al. [2]. The resin beads are treated with concentrated sulfuric acid, and a thin layer of sulfonic acid groups is formed on the surface. The final capacity of the resin is related to the thickness of the layer and is dependent on the type of resin, the bead diameter, and the temperature and time of contact with the sulfuric acid. Typical capacities range from 0.005 to 0.1 mequiv g–1 compared to 5 mequiv g–1 for conventional cation-exchange resins. Over the years, the performance of cation-exchange columns for ion chromatography has improved significantly. Columns are packed with spherical particles with a narrow range of particle sizes. The average size generally is 5–10 lm, although 3lm particles are becoming available. Most contemporary cation exchangers have a polymeric matrix or a silica core with a polymeric coating. These materials generally can be used over a broad pH range and are resistant to common organic solvents. Cation-exchange columns currently available from major commercial suppliers are listed in Tables 7.1 and 7.2.
Table 7.2 Additional cation-exchange columns.
Name
Type
Special features
Alltech Cation R
Sulfonic acid
10 lm particles, pH 1–12, separate mono-, di- and trivalent cations
Alltech Universal Cation
Polybutadiene/ maleic acid coated silica
3 lm particles, high resolution
Hamilton PRP-X200
Polymeric sulfonic
Stable pH 1–13, 100% organic solvent acid
Hamilton PRP-X400
Sulfonate/glycophosphate
7 lm particles, higher capacity
Hamilton PRP-X800
Polymeric
Excellent durability, any pH, 100% organic solvent
Metrohm Metrosep C1
Polybutadiene/maleic acid
For biogenic amines
Metrohm C2
–
Very rapid separation of monoand divalent metal cations
Metrohm C3
Polyvinyl alcohol with carboxyl groups
High efficiency, for mono- and divalent cations, transition metal ions
Metrohm Nucleosil 5 SA
Sulfonic acid
For divalent metal ions only
177
178
7 Cation Chromatography
7.2.1 Historical Development
The development of commercially available columns has been strongly influenced by Dionex columns. The historical development of the latter has been reviewed [3]. In the years following the introduction of suppressed-ion chromatography in 1975 [1], column packings for cation separations consisted of surface-sulfonated 25-lm beads of polystyrene crosslinked with 2% divinylbenzene. At the eluent concentrations commonly used in ion chromatography, these resins had a much higher selectivity for divalent than for monovalent cations. Although ammonium and alkali metal cations could be eluted with dilute hydrochloric acid, a 2+ cation (protonated m-phenylenediamine) was needed to elute the divalents. The organic eluent was found to slowly poison the column. The first latexed column (Ion Pac SC3) was introduced in 1985. A layer of micro latex particles, functionalized with a tertiary amine to generate positively charged sites, was attached to a surface-sulfonated polymeric bead. Then an additional layer of sulfonated latex particles was electrostatically attached to the outer surface so that the final bead had cation-exchange properties (See Figure 3.6). The smaller substrate size, together with a shorter mean free path for the analytes, produced greatly improved peak efficiencies. The ability to separate six common inorganic cations (Li+, Na+, K+, NH4+, Mg2+ and Ca2+) in a single isocratic run has always been the most important single application of cation chromatography. Although an isocratic separation of these cations with the CS3 stationary phase was possible, the retention time for calcium was excessive (about 28 min), and the monovalent peaks were not baseline resolved. In 1990 a sulfonated cation-exchange column (Ion Pac CS 10) with a lower cation-exchange site density was developed. This material was produced by adding a monomer that partially deactivates sulfonation to the latex polymerization mixture. The lower exchange capacity reduced the retention of both the monovalent and divalent cations. The six common cations could be separated isocratically in about 15 min by adding a 2,3-diaminopropionic acid hydrochloride (DAP.HCI) to the acidic eluent. The substrate of the CS 10 material was 50% crosslinked (instead of the 4% crosslinking of previous materials), giving the CS 10 much improved compatibility with organic solvents. This enabled the newer columns to be used with solvent-containing eluents and, more importantly, to be washed with an organic solvent to remove hydrophobic or water-insoluble contaminants that gradually build up when real samples are analyzed. At about this time, the advantages of using a cation exchanger with carboxylic acid instead of sulfonic acid functional groups were becoming apparent. A silicabased stationary phase coated with a maleic acid polymer was shown to separate both 1+ and 2+ inorganic cations within a reasonable time [4]. The eluent is acidic and also contains a mild chelating reagent that combines only with the divalent metal ions.
7.2 Columns
Dionex introduced carboxylic acid ion-exchangers in 1992 and 1993 (CS 12 and CS 14). The six common inorganic cations, including magnesium and calcium, could be separated in about 8 min without using a divalent ion in the eluent. It was necessary to abandon the latex-based system that was previously used in order to produce viable cation exchangers with carboxylate exchange groups. A much higher concentration of carboxylic acid groups is required to provide a sufficient number of ionic exchange sites for cations because most of the groups remain in the neutral -CO2H form. The latex-based system provides high efficiency in a very thin film design, but the exchange capacity is far too low when carboxylic sites are used. To achieve the required higher capacity, the substrate particles consisted of 55% crosslinked ethylvinylbenzene-divinylbenzene core with a surface area of about 450 m2 g–1. A polymer containing three carboxylic groups per unit was applied as a thin film of about 5–20 lm thickness over the entire substrate surface. Ion exchange occurs throughout the applied film and provides a sufficiently high density of exchange sites for practical use. 7.2.2 Phosphonate Columns
The CS 12 A cation exchanger was introduced in 1995 [3]. This column had the unique feature of containing a combination of carboxylic acid and phosphonic acid monomers. This combination provides somewhat different selectivities for inorganic cations and allows the separation of a wide variety of amines as well. The chromatographic effect of changing the ratio of phosphonate to carboxylate groups in the ion exchanger was studied. A phosphonate:carboxylate ratio of 1:5 gave a satisfactory resolution of both monovalent cations in about 10 min. The monovalent cations eluted much faster when material with a 1:1 ratio was used. A 5:1 ratio resulted in rapid elution of all cations within 3 min, but peak resolution was very poor. The final product in the CS 12A material contained a slightly lower ratio than the 1:5 phosphonate:carboxylate ratio used in the development work. Adding phosphonate as a comonomer in the stationary phase for the CS 12A resulted in longer retention times for magnesium(II) and calcium(II) compared to a similar resin that contained only carboxylic acid groups (Figure 7.1). The chemical type as well as the eluent concentration are important parameters in optimizing any given separation. Relatively minor differences in the separation of inorganic cations were observed when sulfuric acid and methanesulfonic acid (MSA) eluents of almost identical concentration were compared. Increasing the concentration of sulfuric acid eluent from 11 mM to 15.5 mM reduced the time for separation of common inorganic cations from about 11 min to 7 min (Figure 7.2). Column temperature is an important parameter that is too often ignored. Column heaters that provide good temperature control are relatively readily available. The chromatograms in Figure 7.3 were performed under identical conditions except for temperature. Efficiency for all cations, as indicated by peak sharpness
179
180
7 Cation Chromatography
Figure 7.1 Comparison of Ion Pac CS12A and CS12. Eluent: 11 mM sulfuric acid. Detection: suppressed conductivity. Peaks: 1 = lithium (0.5 mg L–1); 2 = sodium (2 mg L–1); 3 = ammonium (2.5 mg L–1); 4 = potassium (5 mg L–1); 5 = magnesium (2.5 mg L–1); 6 = calcium (5 mg L–1). (From Ref. [3] with permission.)
and greater peak height, is significantly increased by raising the temperature to 50 °C. This is most likely the result of improved mass transport at the higher temperature. Retention times of the analyte cations are also significantly shorter especially for the divalent cations. Organic amines can also be separated by ion chromatography as protonated cations. Morpholine is used in the power industry as a corrosion inhibitor, and it is important to monitor its concentration as well as that of inorganic cations. In many applications involving amines, the hydrophobic part of the amine cation interacts with the ion exchanger matrix, giving broad and tailed peaks. The severity of this effect is greater for amines with larger organic groups. Figure 7.4 shows that a reasonable separation of organic amines can be obtained with an entirely aqueous system when an elevated column temperature is used.
Figure 7.2 Fast separation of alkali and alkaline earth metals and ammonium with CS12A. Eluent: 15.5 mM sulfuric acid. Detection: suppressed conductivity. Peaks: 1 = lithium (0.05 mg L–1); 2 = sodium (2 mg L–1);
3 = ammonium (2.5 mg L–1); 4 = potassium (5 mg L–1); 5 = magnesium (2.5 mg L–1); 6 = calcium (5 mg L–1). (From Ref. [3] with permission.)
7.2 Columns
Figure 7.3 Temperature effect on peak efficiencies of the inorganic cations and ammonium with the CS12A. Eluent: 18 mM methanesulfonic acid. Detection: suppressed conductivity. Peaks: 1=lithium (1 mg/l); 2=sodium (4 mg/l);
3=ammonium (5 mg/l); 4=potassium (10 mg/l); 5=rubidium (10 mg/l); 6=cesium (10 mg/l); 7=magnesium (5 mg/l); 8=calcium (10 mg/l); 9=strontium (10 mg/l); 10=barium (10/mg/l). (From Ref. [3] with permission.)
Figure 7.4 Solvent-free gradient separation of aliphatic amines on CS12A at elevated temperature. Eluent: 15.5 mM to 25 mM sulfuric acid in a 10-min gradient. Detection: suppressed conductivity. Column temperature: 60°C. Peaks: 1 = ethylamine (5 mg L–1); 2 = propylamine (7.5 mg L–1); 3 = tert-butylamine
(12.5 mg/l); 4 = sec-butylamine (12.5 mg L–1): 5 = isobutylamine (12.5 mg L–1); 6 = n-butylamine (20 mg L–1): 7 = 1,2-propane-diamine (20 mg L–1); 8 = 1,2-dimethylpropylamine (20 mg L–1); 9 = di-n-propylamine (40 mg L–1). (From Ref. [3] with permission.)
7.2.3 Macrocycle Columns
A resin containing a macrocyclic function, tetradecyl-18-crown-6 (TDI8C6) was described in Section 3.5.3. Use of this column adds a new dimension in selectivity for the separation of inorganic cations [5]. Selectivity is determined primarily by
181
182
7 Cation Chromatography
the extent that cations interact to fit into the macrocyclic cavity and are bound by the multiple oxygen atoms in the ring. Table 7.3 gives the binding constants for metal cations. Among the monovalent ions, Li+ is not bound at all, but K+ is very strongly bound.
Table 7.3 Binding constants, log K, for cations to 18-crown-6 in
water at 25 °C. Cation
log K
Cation
log K
Li+
0
K+
2.03
0.8
2+
Ca
2.26
Cs+
0.99
Sr2+
2.72
NH4+
1.23
Ba2+
3.87
Na
+
+
Rb
1.56
The order of elution of cations follows the same order as that indicated in the table. A unique chromatographic feature is the wide separation of NH4+ and the later-eluting K+. Small concentrations of potassium can also be detected in samples with an inordinately high concentration of sodium. The large differences in binding of cations with the macrocyclic resin make it difficult to separate Group I and Group II cations in a single chromatographic run because calcium(II) and potassium(I) elute so much later than the other monovalent cations. However, the retention time of K+ is greatly reduced by operating at a higher column temperature of up to 80 °C. 7.2.4 Surfactant Columns
The affinity of an ion-exchange stationary phase for various ions is determined by a combination of attractive forces and not by electrostatic attraction alone (see Section 3.4). Hydrophobic attractive forces assume a more important role as the organic bulk of an analyte becomes greater. A multi-mode column, the Dionex Acclaim Surfactant column, was developed for analysis of ionic analytes with long organic chains [6]. The surface chemistry of this packing material consists of hydrophobic alkyl chains, tertiary amino groups and polar amide groups. The resulting separation mechanism is mixed mode, including hydrophobic, anionexchange and dipole-dipole interactions. Retention of analytes can be controlled by changing ionic strength, pH and organic solvent content of the mobile phase. Surfactants generally consist of an aliphatic chain with a polar group at one end. Cationic surfactants have a quaternary ammonium group, while anionic surfactants contain a sulfonate or other negative group. A nonionic surfactant con-
7.2 Columns
Figure 7.5 Effect of mobile phase pH on the retention of anionic and non-ionic surfactants. Column: 150 mm × 4.6 mm I.D. Acclaim Surfactant, 5 lm; column temperature: 30 °C; mobile phase: acetonitrile/0.1 mol L–1 ammo-
nium acetate (70:30, v/v), flow rate: 1 mL min–1; injection volume: 10 lL; detection: ELSD. Peaks: (1) Triton X-100; (2) decyl sulfate; and (3) dodecyl sulfate. (From Ref. [6] with permission.)
tains hydroxyl or other polar groups. These three types of surfactants are affected differently by changes in mobile phase ionic strength and pH. The effect of raising eluent pH is illustrated in Figure 7.5. The tertiary amino groups are less positively charged as the pH is raised, resulting in less electrostatic attraction and shorter retention times for anionic surfactants. The retention time for the nonionic surfactant (Triton X100) is essentially unchanged. A higher eluent pH results in stronger retention of cationic surfactants, due to less electrostatic repulsion by the less positively charged amino groups. A simultaneous separation of cationic, anionic and nonionic surfactants was obtained by using gradient elution (Figure 7.6). Anionic surfactants have a stronger attraction for the positively charged nitrogen atoms and have longer retention times than the others. Peak 4 is a mixture of neutral species. Although xylene sulfonate (Peak 1) is anionic, it undergoes weaker hydrophobic attraction than the larger surfactants.
183
184
7 Cation Chromatography
Figure 7.6 Simultaneous separation of anionic, non-ionic and cationic surfactants. Column: 150 mm × 4.6 mm I.D. Acclaim Surfactant, 5 lm; column temperature: 30 °C; mobile phase: (A) acetonitrile and (B) 0.1 mol L–1 ammonium acetate, pH 5.4; gradient: 25% A to 85% A in 30 min; flow rate: 1 mL min–1;
injection volume: 25 lL; detection: ELSD. Peaks: (1) xylene sulfonate; (2) lauryldimethylbenzyl ammonium chloride; (4) Triton X-100; (5) decyl sulfate; (6) dodecyl sulfate; (7) C10-LAS; (8) C11-LAS; (9) C12-LAS; and (10) C13-LAS. (From Ref. [6] with permission.)
7.3 Separations 7.3.1 Suppressed-Conductivity Detection
With suppressed-conductivity detection, an acidic cationic eluent is used to separate the sample cations. The column effluent with zones of separated cations passes directly into the suppressor unit containing an anion-exchange membrane in the hydroxide form. The eluent cation is neutralized and the counteranions associated with the sample metal ions are exchanged for the more highly conducting hydroxide ion. For example, if a dilute nitric acid eluent is used and sodium and potassium sample ions are to be separated, the following reactions take place in the suppressor unit: Eluent: H+NO3– + Res-OH– → Res-NO3– + H2O
(7.1)
Sample: Na+NO3–, K+ NO3– + Res-OH– → Na+OH–. K+ OH– + Res-NO3–
(7.2)
7.3 Separations
The background conductivity is very low after the eluent passes through the suppressor unit; theoretically it is that of pure water. The equivalent conductance of sample ions is high; it is the sum of the conductances of the alkali metal cation and the hydroxide counter ion. Modern suppressors for cation chromatography are both efficient and selfregenerating. The principles are similar to those for the suppressors for anion chromatography, described in Chapter 6. The mechanism of suppression for a cation self-regenerating suppressor is illustrated in Figure 7.7 and described in some detail by Rabin et al. [7]. Suppressors for cation chromatography are limited to those cations that do not form precipitates with the hydroxide ions of the suppressor.
Figure 7.7 Mechanism of suppression for the cation selfregenerating suppressor. H+MSA– = methanesulfonic acid. (From Ref. [7] with permission.)
Excellent separations of all the alkali metal cations plus ammonium are obtained in 10 min or less with a strong acid (sulfonic acid) cation exchanger and a dilute solution of a strong acid as the eluent. However, divalent metal cations are more strongly retained by this column and require either an eluent containing a divalent cation or a more concentrated solution of the H+ eluent. By using an ion exchanger with carboxyl groups or with both carboxyl and phosphonate groups, it is possible to separate both monovalent alkali metal cations and certain divalent metal cations in a single run. A dilute solution of a strong acid such as methanesulfonic acid is generally used as the eluent. Particular emphasis has been placed on the separation of Li+, Na+, NH4+, K+, Mg2+ and
185
186
7 Cation Chromatography
Ca2+ because these ions are found in many types of samples. Figure 7.2 shows a separation of all six ions at concentrations of 0.5 to 5.0 ppm on a Dionex CS 12A bifunctional column [3]. Dilute sulfuric acid was used as the eluent with suppressed-conductivity detection. Gradient elution with suppressed-conductivity detection is also feasible. Several protonated aliphatic amine cations were separated using a gradient with increasing concentration of sulfuric acid (Figure 7.4). The increasing acidity served to reduce the effective exchange capacity of the ion exchanger and thereby speed up the elution of the larger amines and diamines. A different type of gradient was used to separate the metal ions and quaternary ammonium cations in Figure 7.8. In this case an eluent of fixed acidity was used (11 mM sulfuric acid) but the gradient increased the acetonitrile content of the mobile phase from 10% to 80% over 15 min. The purpose of the acetonitrile gradient was to decrease the hydrophobic affinity of the higher amine salts for the ion exchanger.
Figure 7.8 Gradient elution of hydrophobic quaternary ammonium ions. (Courtesy Dionex Corp.)
7.3 Separations
7.3.2 Non-Suppressed-Conductivity Detection
With modern columns and dilute solutions of a strong acid as the eluent, cations may be separated and detected with excellent sensitivity by direct conductivity as well as by suppressed conductivity [2]. The basis for direct-conductivity detection is that the highly conductive H+ (equivalent conductance = 350 S cm2 equiv–1) in the eluent is partially replaced by a cation of lower conductance when a sample zone passes through the detector. For example, the equivalent conductances of L+, Na+ and K+ are 39, 50 and 74 S cm2 equiv–1, respectively. The decrease in conductance on an equivalent basis can be calculated as follows: background: H+ + NO3–= 350 + 71 = 421 (S cm2 equiv–1). Sample peaks: L+ + NO3– = 39 + 71 = 110, a decrease of 311; Na++ NO3– = 50 + 71 = 121, a decrease of 300; K+ + NO3– = 74 + 71 = 145, a decrease of 276. Riviello et al. made a careful comparison of conductivity changes in cation chromatography using direct and suppressed-conductivity detection [8]. The calculation example is outlined in Figure 7.9. The change in conductivity, is actually slightly greater with non-suppressed-conductivity detection. However, the noise is much higher in the non-suppressed detection mode. Noise may be defined as the
Figure 7.9 Comparison of conductivity changes with direct and suppressed-conductivity detection. (Adapted from Ref. [7].)
187
188
7 Cation Chromatography
random signal (conductance in this case) that results from chemical background temperature fluctuations, hydraulics and electronics. It was found that noise is proportional to background conductivity. Pump noise and detector/electronic noise also increases with increased background conductivity. Temperature control is critical for direct detection and slightly improves suppressed-conductivity detection. A major difference between the two detection modes is that detection limits for alkali metal ions were 12–21 times lower with suppressed-conductivity detection [8]. Nevertheless, cations may be separated and quantified with direct detection down to fairly low concentrations. Another advantage is that the eluent remains acidic, and metal does not precipitate as easily as it might in the high pH environment of the suppressor. Figure 7.10 shows a separation of several metal cations including cesium and strontium at concentrations averaging only 1.0 ppm.
Figure 7.10 Separation of metal ions at low ppm concentrations. (Courtesy Alltech.)
7.3.3 Spectrophotometric Detection
Separations of metal cations with ionic eluents has been limited mostly to the alkali metals, ammonium, magnesium(II), calcium(II), strontium(II) and barium(II). Separations of other metal cations are usually performed with eluents that complex the sample cations to varying degrees (see Section 7.5). Some organic cations have also been separated with ionic eluents, although this appears to be an under-utilized area of cation chromatography. Either suppressed- or non-suppressed-conductivity detection is generally satisfactory for these separations. However, under carefully controlled conditions, indirect spectrophotometric detection may sometimes be used to advantage. Haddad
7.3 Separations
and Foley published a comprehensive study of protonated aromatic bases as eluents for cation chromatography [9]. A partial list of the bases studied is given in Table 7.4. The molar absorptivities are high enough to give good sensitivity for detection of sample cations by indirect UV. In this detection mode a wavelength near the kmax of the eluent component is used, giving a baseline of relatively high absorbance. When a zone of non-sorbing sample ions passes through the detector cell, the concentration of eluent component is reduced. This gives a negative peak for the sample ion that is proportional to the sample ion concentration.
Table 7.4 Properties of aromatic bases as prospective eluent
components. Eluent component
kmax , nm
e, L mol-1 cm-1
k+ , S cm2 equiv–2
pKa1
2,3-dimethylpyridine
66.4
6494
31.7
6.57
2,4-dimethylpyridine
258.5
5460
32.8
6.99
2,6-dimethylpyridine
269.2
7705
31.4
6.72
2-methylpyridine
262.3
6449
29.8
5.92
3-methylpyridine
262.3
7177
20.4
5.52
4-methylpyridine
253.4
6516
36.7
6.08
4-methylbenzylamine
261.6
276
36.3
9.36
2-phenylethylamine
256.7
219
39.5
9.84
benzylamine
255.9
249
31.6
9.33
2-aminopyridine
290.0
4188
33.4
6.71
7.6
4-aminopyridine
260.5
14175
24.4
9.11
6.44
pKa2
The eluent amines listed in Table 7.4 form 1+ cations with the exception of the aminopyridines, which can form 2+ cations. A typical separation of alkali metal cations employs approximately 0.1 mM protonated phenylethylamine or 4-methylbenzylamine. A higher concentration (approximately 10 mM) is used for separation of the divalent magnesium and alkaline earth cations. The equivalent conductances of the aromatic bases are low enough for non-suppressed-conductivity detection. However, the detection limits are generally lower with indirect UV detection, and the resolution is often better also. In Figure 7.11 chromatograms of the alkali metal cations are compared with direct conductivity and indirect UV absorption detection. With the aromatic base eluents, Li+ elutes between Na+ and K+ with an acidic eluent such as dilute nitric acid; Li+ always elutes before Na+.
189
190
7 Cation Chromatography
Figure 7.11 Chromatogram obtained with 0.2 mM 2,6-dimethylpyridine at pH 6.35 as eluent by use of (a) direct conductivity and (b) indirect UV absorption detection. Sample: 15 lL of a solution containing 2 × 10-5 M of each of the indicated ions. (From Ref. [9] with permission.)
The separations in Figure 7.11 were performed with 15 mL of a sample containing 2 × 10–5 M of each sample ion. This corresponds to an absolute amount of only 0.30 nmol of each ion. The detection limits of several inorganic ions have been calculated for aromatic bases as eluent components using a 100 mL sample. The results given in Table 7.5 show detection limits in the low ppb concentration range. Most of the detection limits were lower for indirect UV detection than for direct conductivity detection. There are certainly many other possibilities for direct spectrophotometric detection. A separation of several divalent metal ions is shown in Figure 7.12 with 3.5 mM cupric sulfate as the eluent and indirect detection at 220 nm.
7.4 Effect of Organic Solvents Table 7.5 Detection limits (ppb) for typical eluents, calculated
for a 100 lL injection. Indirect UV detection. Eluent
Li+
Na+
K+
NH4+
Mg2+
Ca2+
Sr2+
2-Phenylethylamine
1
2
2
3
1
4
12
Benzylamine
1
0.4
0.4
0.5
11
63
290
4-Methylbenzylamine
0.2
1
1
1
5
16
47
Figure 7.12 Separation of metal ions on a PRP-X200 column, 150 × 4.1 mm. Eluent 3.5 mM cupric sulfate, indirect UV detection at 220 nm. (Courtesy Hamilton Co.)
7.4 Effect of Organic Solvents
It is quite common to add a low percentage of methanol or acetonitrile to an aqueous mobile phase in order to obtain sharper peaks for organic sample ions, and it is entirely feasible to perform cation chromatographic separations in an aqueousorganic mixture with a much higher proportion of the organic solvent. Hoffman and coworkers [10, 11] have shown that two mechanisms occur in such cases: ion exchange and hydrophobic interaction between the sample and the resin when the eluent contains 70% acetonitrile. This is due to lower hydrophobic interaction in the 70% acetonitrile.
191
192
7 Cation Chromatography
Dumont, Fritz and Schmidt studied cation chromatography in organic solvents containing little if any water [12]. Under these conditions solvation of the lipophilic part of the cation should be sufficient to virtually eliminate the hydrophobic interaction between the sample cations and the ion-exchange resin. In this way the true ion-exchange selectivity could be measured. After trying several different inorganic acids, methanesulfonic acid was selected as the eluting acid for the separation of protonated amine cations. In IC of cations with H+ as the eluting cation, k should vary according to the following equation: log k = -m log [H+] + b where m is the slope of a linear plot and b is a constant. Linear plots were obtained for the C1-C10 n-alkylamines in methanol, ethanol, 2-propanol and acetonitrile. The slopes (m) were very close to the theoretical slope of –1 in the three alcohols and only a little less than –1 in acetonitrile. The effect of solvent was studied by measuring the retention factor, k, for a series of protonated alkylamine cations with 25 mM methanesulfonic acid in the appropriate solvent as the eluent [12]. Ordinarily a plot of log k vs the number of carbon atoms in such a homologous series would be linear. The slope of such a plot is at least in part an indication of the effect of the carbon chain on the retention factor. The retention factors were measured under identical conditions in each of four organic solvents. The k values of the alkylamines increased according
Figure 7.13 Separation of 12.5 ppm aniline (1), N-methylaniline and (2), N,N-dimethyl-aniline and (3), on a 5 cm sulfonated resin column (0.15 mmol L–1). The eluent was methanesulfonic acid in methanol at a flow-rate of 1 mL min–1. (From Ref. [13] with permission.)
7.4 Effect of Organic Solvents
to the solvent used in the following order: methanol, ethanol, 2-propanol, acetonitrile. However, in any given solvent the k values of the individual amines were almost constant from C1 to C10. Separation of the individual amines was not possible. Two other types of organic cations did show enough difference in their retention factors for practical separations. Separation of aniline, N-methylaniline and N,N-dimethylaniline in 100% methanol is shown in Figure 7.13. Significant differences in the k values of octylamine, dioctylamine and trioctylamine were observed in methanol, ethanol and 2-propanol. These results suggest that a successful separation of organic cations by IC depends on differences in hydrophobic attraction between the solute ions and the ion exchanger as well as on differences in electrostatic attraction. Incorporation of an organic solvent in the eluent will increase the solubility of samples containing organic solutes. However, it is usually better to work with a mixed organic-aqueous eluent rather than one that is entirely organic. Conductivity detection is feasible in organic-aqueous solutions or in 100% organic for the lower alcohols. 7.4.1 Separation of Alkali Metal Ions
Ion chromatographic separations of the alkali metal cations are normally performed with sulfonated microporous polymeric resins or with resins coated with a sulfonated latex. Dumont and Fritz selected a lightly sulfonated macroporous resin with a high degree of cross linking for a study on alkali metal ion separations in organic solvents [13]. Such a resin would be less likely to undergo volume changes due to swelling and should be more compatible with organic solvents. A separation of alkali metal ions was first attempted in water alone using a lightly sulfonated macroporous cation exchanger with aqueous 3 mM methanesulfonic acid as the eluent. Under these conditions the sample cations exhibited very similar retention times. When the macroporous resin column was used with the same acidic eluent in 100% methanol, the chromatographic separation was improved considerably, as the alkali metal ions were solvated with methanol and the resin matrix was probably coated with a thin layer of methanol, which made the ions and the resin surface more compatible with one another. Although several factors may influence the selectivity of cation-exchange resins for 1+ metal cations, electrostatic attraction of the sulfonate groups in the ionexchange resin for alkali metal cations suggests that cations with the smallest ionic radii would be the most strongly retained. The Pauling radii in Table 7.6 would predict a chromatographic elution order of Cs+, Rb+, K+, Na+, Li+, which is exactly the opposite of that observed in ion-exchange chromatography. However, hydrated ionic radii and approximate hydration number are in the opposite order to the Pauling radii, with Li+ being the most highly hydrated. When separations are carried out in a nonaqueous solvent, the solvation of the alkali metal cations is apt to change. This could result in changes in elution characteristics of the ions.
193
194
7 Cation Chromatography Table 7.6 Detection radii of alkali metal cations.
Li+
Na+
K+
Rb+
Cs+
Pauling radii (Å)
0.60
0.96
1.33
1.48
1.69
Hydrated radii (Å)
3.40
2.76
2.32
2.28
2.28
Approximate hydration number
25.3
16.6
10.5
10.0
9.90
Table 7.7 Retention factors (k) in organic and mixed solvents
with 0.5 mM methanesulfonic acid (MSA) as the eluent. Solvent
Li+
Water (100%) 2.27
Na+
K+
Rb+
Cs+
NH4+
2.27
2.71
2.74
2.96
3.04
Methanol 25%
2.66
2.47
2.77
2.80
2.98
3.13
50%
3.10
3.07
3.31
3.39
3.80
3.70
75%
4.43
5.07
6.32
7.30
8.33
5.82
100%
2.08
2.82
3.73
4.33
5.15
3.09
25%
2.75
2.57
2.78
2.78
2.96
3.13
50%
3.25
3.09
3.37
3.48
3.76
3.78
75%
4.61
5.14
6.90
7.62
8.76
5.98
100%
1.86
3.84
7.24
8.84
9.87
2.12
25%
2.20
1.98
2.11
2.05
2.17
2.45
50%
2.26
2.14
2.35
2.41
2.62
2.88
75%
3.41
3.52
4.42
4.81
5.66
4.69
100%
8.84
12.3
19.5
>20
>20
4.54
25%
2.10
2.10
2.37
2.38
2.54
2.51
50%
2.50
2.38
2.89
2.98
3.35
3.10
75%
3.00
3.12
3.79
3.95
4.45
3.98
100%
4.46
2.11
1.75
1.59
1.54
2.40
Ethanol
2-Propanol
Acetonitrile
7.5 Separation of Metal Ions with a Complexing Eluent
The results of these studies [13] are summarized in Table 7.7. The use of nonaqueous solvents with macroporous cation-exchange resin permits several separations that are very difficult with aqueous eluents. Methanol was found to be the most favorable solvent because it gives the best combination of resolution and peak shape. Acetonitrile and ethanol, although producing broader peaks, are useful for separating ions that usually elute close together, Li+/Na+ and K+/NH4+ respectively. Elution order in acetonitrile is the reverse of that found with aqueous eluents: Cs+ < Rb+ < K+ < Na+ < Li+. It was also discovered that addition of a crown ether, 18-crown-6, to the mobile phase improves both the peak shape and resolution of several ions.
7.5 Separation of Metal Ions with a Complexing Eluent 7.5.1 Principles
The cation separations discussed thus far are based on differing affinities of the sample ions for the cation-exchange resin. But divalent metal cations often have similar affinities and are difficult to separate from one another. Selective complexation with a mobile phase containing a ligand offers additional selectivity for separation of metal cations. The logarithm of the retention factor of metal ions is linearly proportional to the logarithm of the ligand concentration in the eluent (Figure 7.14).
Figure 7.14 Plot of log retention time against log aM for elution of rare earth cations with ethylenediammoniuim tartrate. (Courtesy of G. J. Sevenich.)
195
196
7 Cation Chromatography
7.5.2 Separations
An early paper by Sevenich and Fritz [14] described a separation of several divalent metal cations with an eluent containing 2.0 mM ethylenediammonium tartrate at pH 4.5. A sulfonic acid column was used with direct conductivity detection. Eluents containing ammonium tartrate and no ethylene-diammonium 2+ salt were ineffective for elution of the metal ions studied. The elution mechanism was described as an ion-exchange ‘pushing’ action of the 2+ cation and a weakly complexing ‘pulling’ effect of the tartrate anion.
Figure 7.15 Separation of zinc(II) (10.3 ppm), cobalt(II) (9.1 ppm), manganese(II) (160.0 ppm), cadmium(II) (16.1 ppm), calcium(II) (17.1 ppm), lead(II), and strontium II (20.3 pm). Eluent was 1.5 mM ethylenediammonium cation and 2.0 mM tartrate at pH 4.00. (Courtesy of G. J. Sevenich.)
A chromatographic separation of seven cations is shown in Figure 7.15. Lead(II) showed a dramatic decrease in retention time as the pH was increased and complexation by the tartrate became stronger.
7.5 Separation of Metal Ions with a Complexing Eluent
7.5.3 Use of Sample Masking Reagents
Ordinarily, a weak complexing reagent is added to the mobile phase to speed up a separation by partial complexation of the sample cations. But another complexation effect is produced by adding a second strong but selective chelating reagent only to the sample. Under these conditions, strongly complexed metal ions move very rapidly though the column while the other metal ions are only complexed weakly by the mobile phase reagent and can be separated from one another [15]. The practical usefulness of this technique is best explained by an example. Suppose we wish to determine small amounts of magnesium(II), calcium(II) and strontium(II) in a sample containing a much higher concentration of iron(III). In a chromatographic separation, the large iron peak would likely obscure the much smaller peaks of the divalent metal ions. But if an auxiliary chelating reagent could be used to selectively complex, and thereby mask, the iron(III), a good separation of the divalent metal ions would be possible. Simple calculations showed that EDTA does not complex metal ions such as magnesium(II) and calcium(II) at pH 4 but it does complex many other metal cations. Therefore, experiments were performed in which ETDA was added to the metal ion sample and the column was eluted with ethylenediammonium tartrate as before [14]. The amount of EDTA used was more than enough to complex the metal ions present, but an unduly high concentration of EDTA was avoided. The results obtained show that conditions can easily be established whereby magnesium and the alkaline earth cation peaks are hardly affected but metal ions that form stable EDTA complexes at about pH 4 are rapidly eluted. Because EDTA is added only to the sample and not to the eluent, it moves rapidly through the column and appears as part of the ‘pseudo peak’. Samples containing a large excess of iron(III) give extremely wide ‘pseudo peaks’ when the ethylenediammonium tartrate is used. This excess of iron(III) will totally obscure the magnesium peak while calcium and strontium appear on the tail of the ‘pseudo peak’. Figure 7.16 shows a chromatogram of the same sample in which EDTA is added to complex the iron(III). The additional peak is from an iron(II) impurity in the iron(III) solution used. Work thus far indicated that any metal ion that has an EDTA formation constant of about 1015 or higher should be masked effectively by adding EDTA to the sample. The very weak complexing of iron(II) by tartrate suggested that iron(II) might be determined quantitatively by cation chromatography. This was proved to be true by Fritz and Sevenich [14], who determined iron(II) in the presence of iron(III) and several other metal ions. Total iron in solution was determined after a preliminary reduction to iron(II) with ascorbic acid. Several other complexing reagents are effective masking agents when added to the sample but not to the mobile phase. Nitrilotriacetic acid (NTA) masks a 10- to 100-fold excess of aluminum(III), copper(II), nickel(II) or iron(III) when magnesium(II), calcium and manganese(II) are to be determined by IC. Analysis of rare-
197
198
7 Cation Chromatography
Figure 7.16 Separation of 0.10 mM each of Mg2+, Ca2+, Ca2+, Sr2+ in the presence of a 100-fold excess of Fe3+ with a no EDTA in the sample (pH 1.7), and (b) 0.010 M EDTA added to the sample (pH 1.7). Eluent was 2.00 mM ethylenediammonium tartrate at pH 4.50. (Courtesy of G. J. Sevenich.)
earth cations using sulfosalicylic acid to mask a large excess of aluminum also proved quite successful. 7.5.4 Weak-Acid Ion Exchangers
Separations of metal cations are best carried out with a carboxylic acid stationary phase. The mobile phase contains a chelating reagent, such as citric, oxalic, or 2,6-pyridinedicarboxylic acid (PDA). The concentration and pH of the mobile phase are adjusted so that metal ions in the sample will be in an equilibrium between the cationic form, M2+, and a neutral or anionic metal ligand complex form. The rate at which a metal analyte moves down the column will depend on the fraction that remains in the free cationic form. A greater degree of complexation results in a shorter retention time. Table 7.8 lists the logarithms of the formation constants of several metal cations with three common complexing reagents.
7.5 Separation of Metal Ions with a Complexing Eluent Table 7.8 Logarithms of formation constants of selected metal
complexes. Reagent
Log formation constant Cu2+
Ni2+
Co2+
Zn2+
Mn2+
Mg
Ca2+
Sr2+
Ba2+
Citric acid
5.60
5.11
4.83
4.70
3.70
3.25
3.18
2.81
2.55
PDA
8.80
6.60
6.35
6.43
4.70
2.02
4.30
3.50
3.13
Oxalic acid
4.53
3.70
3.25
3.43
2.60
2.10
1.66
1.25
1.02
Figure 7.17 Separation of transition metals with the use of a complexing eluent containing PDA. (Courtesy Dionex Corp.)
The elution order of metal cations should follow the order of decreasing complexation. This is indeed the case for the separation in Figure 7.17. Here, spectrophotometric detection at 530 nm was used after post-column addition of pyridylazoresorcinol (PAR) as a complexing reagent. At highly alkaline pH values the eluted metal ions form colored complexes with PAR that are more stable than the complexes with the mobile phase. However, the use of post-column detection requires somewhat more complicated equipment. A separation of eleven cations with indirect non-suppressed-conductivity detection is illustrated in Figure 7.18. The mobile phase contained both a strong acid (MSA) and oxalic acid as a complexing reagent [16]. The peaks actually denote decreasing conductivity.
199
200
7 Cation Chromatography
Figure 7.18 Chromatogram of standard cations. Peaks (mg L-1): 1 = copper (1.6); 2 = lithium (0.08); 3 = sodium (0.8); 4 = ammonium (0.32); 5 = potassium (0.8);
6 = magnesium (0.8); 7 = zinc (1.6); 8 = cobalt (1.6); 9 = nickel (1.6); 10 = calcium (1.6); 11 = strontium (1.6). (From J. Chromatogr. A, 1118, 68, 2006 with permission.)
Figure 7.19 Separation of divalent metal ions on a carboxylic acid column, 50 × 4.6 mm. Eluent: 1.0 mM ethylenediammonium, 0.1 mM PDA, pH 5.14. Peaks: 1 = Ca2+., 2 = Sr2+, 3 = Ba2+, 4 = Mg2+. (From Ref. [16] with permission.)
7.6 Chelating Ion-Exchange Resins and Chelation Ion Chromatography
A much faster separation of divalent metal cations is obtained by replacing H+ with a divalent cation such as ethylenediammonium (with a 2+ charge) as the predominant cation in the mobile phase [16]. Figure 7.19 shows excellent resolution of divalent cations with an acidic eluent containing ethylenediammonium PDA. The cations are eluted in the order of decreasing complexation of the sample ions: Ca2+, Sr2+, Ba2+, Mg2+. In the absence of a complexing mobile phase, magnesium would elute before the alkaline earth cations. The sample cations are detected with reasonable sensitivity by non-suppressed conductivity. Experiments designed to optimize separation conditions showed that the ethylenediammonium cation and a complexing anion were both needed in the eluent. Broad peaks and a few separations were obtained using the ethylenediammonium cation with only a non-complexing anion. Addition of any of the complexing anions tightened the chromatographic peaks considerably. However, very poor separations were obtained when the eluent contained a complexing anion in conjunction with a monovalent cation such as sodium. The presence of ethylenediammonium or some other divalent cation seemed to be necessary.
7.6 Chelating Ion-Exchange Resins and Chelation Ion Chromatography 7.6.1 Fundamentals
The selectivity of ordinary cation-exchange resins for various metal ions is somewhat limited. However, if a suitable chelating functional group is built into a polymeric resin, it is often possible to take up only a small group of metal ions. Other chelating resins may complex a larger group of metal ions, but additional selectivity is attained through pH control. Chelating resins are also valuable in sorbing a desired metal ion (or small group of metal ions) from solutions containing a very high concentration of a non-complexed metal salt. Frequently the selectivity of a chelating resin is so great that a very short column can be used to retain the desired metals. Preconcentration of selected metal ions is probably the main use of chelating resins in chemical analysis. Trace amounts of complexed metal ions may be concentrated from a large sample onto a very short column. Subsequent elution by acid breaks up the metal chelates and gives a much more concentrated solution of the metal ions for further analysis. However, a column packed with a chelating resin may also be used to separate sample metal ions based on differences in the strength of their chelates. A few of the many types of chelating resins that have been synthesized are listed in Table 7.9. Resins containing the iminodiacetic acid (IDA) functional group have received particular attention. For example, the material known as Chelex 100 has been available for many years but is not very efficient for chromatographic separation of metal ions. More modern IDA resins are more satisfactory.
201
202
7 Cation Chromatography
The IDA group forms chelates with a considerable number of metal ions and also provides good selectivity for metal ions that are not complexed. However, some problems can occur.
Table 7.9 Some typical chelating groups in resins.
Designation
Chelating group
Selective for:
IDA
N(CH2CO2H)2
Most M2+, M3+ , M4+
Amidoxime
-C(=NOH)NH2
Divalent transition metals
–
DTC
-NC(=S)S
Ag+, Hg2+, Cu2+, Cd2+, Zn2+, Co2+, Ni2+
Crown ether
-OCH2CH2)– (cyclic)
K+, some others
Hydroxamic acid
-C(=O)(NOH)CH3
Fe3+, Ti4+, Th4+, Zr4+, Mo(VI) [19]
A high concentration of IDA groups on the chelating resin results in more complete complexation of metal ions from solution. A high concentration of complexing groups may also cause the resin to retain metal ions from a more acidic sample. However, stronger complexation of metal ions means that a more concentrated acid solution must be used to break up the complex and thereby desorb metal ions from the resin. The presence of excess acid may complicate the determination of sample ions in subsequent analysis by ion-chromatography or capillary electrophoresis. An additional complication is that two kinds of metal ion uptake can occur with IDA resins. The desired kind of uptake involves chelation of metal ions with the nitrogen and carboxyl groups of the IDA as ligands. The other type is simple ion exchange of cations that are electrostatically attracted to the negatively charged carboxylate groups. This simple ion exchange can take up a significant amount of Na+ or other unwanted cations, particularly if the resin contains a high concentration of IDA groups. On balance, the best choice of an IDA resin might be one with a moderately low capacity (ca. 0.5 mequiv g–1, for example) for retaining metal ions by chelation. A resin particle size of ca. 10 lm should be used instead of the 40–50 lm size generally used in solid-phase extraction (SPE) cartridges. For chromatographic separations it is important to use an efficient resin in which the chelating groups are readily accessible, so that the chelation of metal ions is not sterically inhibited. The equilibrium between a divalent metal ion (M2+) and a chelating resin (R–H+) may be written: M2+ + 2 R–H+ > (RL)2M + 2 H+
(7.1)
7.6 Chelating Ion-Exchange Resins and Chelation Ion Chromatography
An acidic eluent is used to control this equilibrium so that the retention factor of the sample metal ion is in the desired range. Increasing H+ concentration in the eluent weakens the chelates and speeds the elution. Separations of different metal ions will occur because of differences in the equilibrium constants for Eq. (7.1). A second way to separate metal ions on a chelating resin column is to use a complexing eluent (E–), such as oxalate or tartrate, at a fixed pH. Here, a second equilibrium will come into play: (RL)2M + 2 E– > 2 RL– + ME2
(7.2)
Now the retention factor will be influenced by the type and concentration of L– in the chelating resin, the pH of the eluent, and the type and concentration of E– in the eluent. Because formation and breakup of metal chelates is slower than a simple ionexchange equilibrium, it is essential to select chelating resins with fast kinetics. When a complexing eluent is involved, the kinetic situation may become more difficult. Now we can envision a still slower equilibrium between the metal chelate, (RL)2M, and the eluent chelate, ME2. As we shall see, this does not necessarily prevent effective separations with a complexing eluent, but it may still be an inhibiting factor. Bonn, Reiffenstuhl and Jandik were able to separate a number of divalent metal ions effectively in a single run using an IDA resin [17]. A silica-based material (Nucleosil 300–7, with 7 lm diameter particles and 300 Å average pore size) was derivatized with 3-glycidoxy-propyltrimethoxy silane, and iminodiacetic acid was then covalently coupled to the epoxy-activated surface. The final material was a slurry and was packed into a 100 × 4.6 mm stainless steel column. A complexing eluent containing 10 mM citric acid plus 0.04 mM 2,6-pyridine dicarboxylic acid (PAD) gave a good separation of low ppm concentrations of Mg2+, Fe2+, Co2+, Cd2+ and Zn2+ using conductivity detection. Traces of Co2+, Zn2+ and Cd2+ were concentrated and separated with 10 mM tartaric acid at pH 2.54. In another example, a silica gel-based sorbent with chemically bonded amidoxime groups was used for chromatographic separation of transition and heavy metals [18]. The resin known as Amidoxim was used. Separation of five transition metals is shown in Figure 7.20 using 5 mM sodium at pH 3.6 as the eluent. Postcolumn detection was employed with 0.5 mM 4-(2-pyridylazo) resorcinol (PAR) in 3 M ammonia and 1 M acetic acid as the color-forming reagent.
203
204
7 Cation Chromatography
Figure 7.20 Separation of transition metals on the Amidoxim column with sodium oxalate as eluent. Eluent: 5 mM sodium oxalate (pH 3.6); other conditions as in Figure 7.1. Peaks: 1 = Cd2+; 2 = Co2+; 3 = Zn2+; 4 = Cu2+; 5 = Ni2+. (From Ref. [18] with permission.)
7.6.2 Examples of Metal-Ion Separation
Lanthanides and yttrium have been separated by chelation ion chromatography on a column packed with iminodiacetic acid bonded to silica [19]. The mobile phase contained 16 mM nitric acid and 500 mM potassium nitrate. The high nitrate concentration undoubtedly modified the ion-exchange and chelating action of the stationary phase by the partial formation of soluble nitrate complexes with the lanthanide cations. Post-column detection was used with Arsenazo III, which form intensely colored complexes with the lanthanides.
References
References [1] H. Small, T. S. Stevens and
[2]
[3]
[4]
[5]
[6]
[7]
[8]
[9]
W. C. Bauman, Novel ion exchange chromatographic method using conductimetric detection, Anal. Chem., 47, 1801, 1975. J. S. Fritz, D. T. Gjerde and R. M. Becker, Cation chromatography with a conductivity detector, Anal. Chem., 52, 1519, 1980. M. A. Ray and C. A. Pohl, Novel cationexchange stationary phase for the separation of amines and six common inorganic cations, J. Chromatogr. A, 739, 87, 1996. J. P. Kolla, J. Kohler and G. Schomburg, Polymer-coated cation-exchange stationary phases on the basis of silica, Chromatographia, 23, 465, 1987. B. R. Edwards, A. F. Giauque and J. D. Lamb, Macrocycle-based column for the separation of inorganic cations by ion chromatography, 1. J. Chromatogr. A, 706, 69, 1995. X Liu, C. A. Pohl and I. Weiss, New polar-embedded stationary phase for surfactant analysis, J. Chromatogr. A, 1118, 29, 2006. S. Rabin, J. Stillian, V. Barreto, K. Friedman and M. Toofan, New membrane-based electrolyte suppressor device for suppressed-conductivity detection in ion chromatography, J. Chromatogr., 640, 97, 1993. V. Barreto, L. Bao, A. Bordunov, C. Pohl and J. Riviello, Electrolytic suppression in ion chromatography. Paper No. 958, Pittcon 699, Orlando, Fl. P. R. Haddad and R. C. Foley, Aromatic bases as eluent components for conductivity and indirect ultraviolet detection of inorganic cations in non-suppressed ion chromatography, Anal. Chem., 61, 1435, 1989.
[10] A. Rahman and N. E. Hoffman, Reten-
[11]
[12]
[13]
[14]
[15]
[16]
[17]
[18]
[19]
tion of organic cations in ion exchange chromatography, J. Chromatogr. Sci., 28, 157, 1990. N. E. Hoffman and J. Liao, Ion exchange in reversed-phase chromatography of some simple organic cations, J. Chromatogr. Sci., 28, 428, 1990. P. J. Dumont, J. S. Fritz and L. W. Schmidt, Cation-exchange chromatography in nonaqueous solvents, J. Chromatogr. A., 708,109, 1995. P. J. Dumont and J. S. Fritz, Ion chromatographic separation of alkali metals in organic solvents, J. Chromatogr. A, 706, 149, 1995. G. J. Sevenich and J. S. Fritz, Addition of complexing agents in ion chromatography for separation of polyvalent metal ions, Anal. Chem., 55, 12, 1983. G. J. Sevenich and J. Fritz, Effect of complexing agents on the separation of polyvalent cations, J. Chromatogr., 347, 147, 1985. J. Morris and J. S. Fritz, Ion chromatography of metal cations on carboxylic acid resins, J. Chromatogr., 602, 111, 1992. G. Bonn, S. Reiffenstuhl and P. Jandik, Ion chromatography of transition metals on an iminodiacetic acid bonded stationary phase, J. Chromatogr., 499, 669, 1990. I. N. Boloschik, M. L. Litvina and B. A. Rudenko, Separation of transition and heavy metals on an amidoxime complexing sorbent, J. Chromatogr. A, 671, 51, 1994. P. N. Nesterenk and P. Jones, Isocratic separation of lanthanides and yttrium by high performance chelation ion chromatography on iminodiacetic acid bonded to silica, J. Chromatogr. A, 804, 223, 1998.
205
207
8 Ion-Exclusion Chromatography 8.1 Principles
Ion-exclusion chromatography (IEC) has developed into a very useful technique for separating relatively small weak acids (carbonic acid, carboxylic acids, hydrocarboxylic acids, etc.), weak bases (ammonia, amines) and hydrophilic molecular species such as carbohydrates and the lower alcohols. The analytical method actually involves the separation of molecular species rather than ions. Of course, ions can often be readily converted into molecular species as when anions of weak acids are acidified. The rationale for including IEC in a book on ion chromatography seems to be that a cation-exchange resin, or occasionally an anion-exchange resin, has generally been used for IEC separations. Also, it has become customary to include IEC in symposia and books devoted to ion chromatography. Ion-exclusion chromatography is a comparatively old technique, attributed primarily to Wheaton and Bauman [1]. Ionic material is rejected by cation- or anionexchange resin and passes through quickly, but non-ionic substances are held up and come through more slowly. Substances that can be separated include weak organic and inorganic acids, weak organic and inorganic bases, and hydrophilic neutral compounds such as sugars. The resin bed consists of three parts: 1. A solid resin network 2. Occluded liquid within the resin beads 3. The mobile liquid between the resin beads. The ion-exchange resin acts as a semipermeable membrane between the two aqueous phases, (2) and (3). Ionized sample solutes are excluded from the interior water (2) and pass quickly through the column. Nonionic materials are not excluded and they partition between the two water phases, (2) and (3). Thus, they pass more slowly through the column. Nonionic solutes differ in their degree of retardation by the resin phase because of: (i) differing polar attraction between the solute and resin functional groups, (ii) differing van der Waals forces between the solutes and the hydrocarbon portion of the resin.
Ion Chromatography, 4th Ed. James S. Fritz and Douglas T. Gjerde Copyright © 2009 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim ISBN: 978-3-527-32052-3
208
8 Ion-Exclusion Chromatography
Harlow and Morman [2] studied the behavior of a large number of organic and inorganic acids on a column containing the hydrogen form of Dowex 50 × 12. The eluent was distilled water. The strongest acids, such as sulfuric and hydrochloric, elute together with the column void volume because they are highly ionized and cannot enter the resin phase. Harlow and Morman reported several generalizations for predicting elution behavior: 1. Members of a homologous series emerge in order of increasing acid strength and decreasing 2. Water solubility. An example is the formic, acetic, and propionic acid series. 3. Dibasic acids elute sooner than nonbasic acids. Oxalic acid elutes before propionic acid. 4. An isoacid elutes before the corresponding normal acid. For example, isobutyric acid emerges before butyric acid. 5. A double bond tends to retard the elution of an acid. Acrylic acid elutes after propionic acid. 6. Acids with a benzene ring show a strong retention. According to the original theory, the retention volume of an analyte can be described by the equation: VR = V0 + KdVi
(8.1)
where: VR is the retention volume, Vo is the column void volume (volume not occupied by the column packing), Kd is the distribution coefficient for analytes between the mobile phase and the resin, and Vi is the ‘inner volume,’ which is the volume of liquid within the resin beads. It is assumed that anions are completely excluded from the resin by a wall of negative charges set up by the sulfonate sites on the resin particles. If this reasoning is correct, differences in the retention volume (or retention time) will depend on the value of Kd, which can vary only from zero to 1.0. It has been customary to use columns of larger than normal inner diameter (4–10 mm) in order to increase Vi and bring about a greater difference in retentions of the various analytes. Tanaka and Ishakuza [3] studied the separations of acids by IEC on a high-performance cation-exchange column with sulfonic acid functional groups, with water as the eluent. Values of the distribution coefficients, calculated from the retention volumes of the various acids may be summarized as follows: Kd = 0: HI, HBr, HClO4, HCl, H2SO4, HNO3, Oxalic Intermediate Kd: H3PO4 (0.09), H2SO3 (0.11), HF (0.36), formic acid (0.43), acetic acid (0.65), propionic acid (0.81), benzoic acid (0.98). Kd = 1.0: H2CO3, HCN, H3BO3. Kd > 1.0: butyric acid (1.10), H2S 1.40.
8.1 Principles
A linear plot was obtained for the retention of several weak acids as a function of pKa, indicating that retention is affected by the fraction that is in the molecular, non-ionized, form. Continuing research has demonstrated that the original ion-exclusion model does not adequately explain the results obtained when weak organic acids, amines or neutral, polar organics are to be separated. These analytes are also attracted to the stationary phase by hydrophobic effects and sometimes by hydrogen bonding, and not just by partitioning between the mobile phase and liquid within the stationary phase. The requirement that exclusion of analyte anions by negatively charged resin groups is necessary for IEC is no longer valid. The examples given in the following sections of this chapter indicate that what began as an ion-exclusion technique has become more of a special form of liquid chromatography that applies primarily to separation of smaller organic acids and amines, and very hydrophilic molecules such as carbohydrates. 8.1.1 Equipment
IEC systems consist of the same components as those used in ion chromatography. Either stainless steel or PEEK hardware is commonly used. Gel-type sulfonated cation exchangers or polyacrylate resins with carboxylic acid functional groups are employed for separation of weak-acid analytes. Weak bases are separated on gel exchangers with quaternary ammonium functional groups. The exchange capacities of the resins are generally higher than those used for ion chromatography. The inside diameter of IEC columns is often 6–8 mm, compared to an average i.d. of about 4.5 mm for IC. 8.1.2 Eluents
Acidic eluents are used in separating weak acids (such as carboxylic acids) to repress their ionization and give sharp chromatographic peaks. However, water alone is often a suitable eluent for very weak acids such as carbonic acid and boric acid. Organic amines require a basic eluent, such as dilute aqueous sodium hydroxide, to ensure that the amines are in the molecular form and are not ionized. Water alone can be used to elute weak molecular bases. Sometimes an organic solvent is added to the aqueous eluent to speed up the elution of sample compounds. The type and concentration of the eluent must always be chosen so that it will be compatible with the detector. Methanol can shrink the bed of a sulfonated polystyrene catex resin and cause irreversible damage to the column. However, up to 40% v/v acetonitrile can be used with this type of column. Ethanol and 2-propanol are acceptable in smaller amounts (about 15% and 10%, respectively) when they are used with the high-capacity cation exchanger.
209
210
8 Ion-Exclusion Chromatography
8.1.3 Detectors
Conductivity, direct absorbance or a differential refractometer are the most common forms of detection for IEC, PAD and ELSD. A pulsed amperometric detector (PAD) or, more recently, an evaporative light-scattering detector (ELSD) is appropriate for detection of carbohydrates. Both non-suppressed and suppressed conductivity have been used extensively. The need to incorporate a low concentration of a strong acid into the eluent has been an impediment to direct conductivity detection. Tanaka and Fritz [4] found that the conductivity detection of aliphatic carboxylic acids is much improved if a dilute solution of benzoic acid is used as the eluent. The background conductance is much lower than with aqueous solutions of strong mineral acids. A fairly low background conductance is obtained because benzoic acid is only partly ionized and a very dilute (5.0 mM) solution of benzoic acid is used. Nevertheless, this eluent is sufficiently acidic to give a good separation of aliphatic carboxylic acids with well-shaped peaks. Membrane suppressors that can be regenerated continuously have also been developed for the ion-exclusion chromatography of weak acids. In IEC, the primary contributor to the high eluent conductance is the hydrogen cation from the acid eluent. Membrane suppressors can reduce this background by exchanging the hydronium ion for tetrabutylammonium ion (TBA). This is shown in Table 8.1. Because of the TBA+/Cl– ion pair, the background conductivity is decreased significantly. It should be noted that the conductance of the sample species is reduced as well. The background conductivity may be reduced further by replacing the HCl acid eluent with a weaker acid. Two acids that have been used successfully for IEC are octylsulfonic acid (OSA) and tridecafluoroheptanoic acid (TDFA) [5].
Table 8.1 Relative conductance of 1 mM solutions in suppressed
ion-exclusion chromatography. Relative conductance, lS
Ion Pair H+/Cl– +
425 –
TBA /Cl +
100 –
TBA /OSA +
40 –
TBA /TDFHA
30
TBA = tributylamine, OSA = octylsulfonic acid, TDFHA = tridecafluoroheptanoic acid.
8.2 Separation of Organic Acids
8.2 Separation of Organic Acids
Gjerde and Mehra [6] compiled tables of retention factors for organic acids obtained on several commercial columns that are widely used for IEC. Some interesting questions concerning the mechanism of IEC have been posed [7]. If the mechanism for separation of these carboxylic acids is primarily ion exclusion, why does an aliphatic carboxylic acid with a larger alkyl (R) group elute later than one with a smaller alkyl group? We would expect the Kd value of the more bulky acid to be lower than that of the less bulky acid. Again, why should substitution of an aliphatic hydrogen by a hydroxyl group cause the hydroxyl compound to be eluted faster (for example, tartaric acid is eluted before malic acid, and both are eluted before succinic acid)? The reasonable answer seems to lie in the solute acid’s interaction with the resin matrix. An acid with a large alkyl group would have a greater hydrophobic attraction and thus a longer retention time. A hydroxy compound is more polar than its hydrogen analog and thus would interact less strongly with the resin matrix. In IEC, stronger acids (lower pKa) are eluted more rapidly than weaker acids, presumably because the stronger acids are incompletely converted to molecular form. If this were the case, coexistence of the ionic and molecular forms might produce broader peaks; however, these compounds produce very sharp peaks. A better explanation might be that stronger acids are more polar and therefore interact less strongly with the resin matrix. These considerations suggest that differences in partitioning between the mobile and resin phases is the main mechanism for chromatographic separation of carboxylic acids and that an ion-exclusion mechanism is not essential. The separation of carboxylic acids with a water–acetonitrile gradient (Figure 8.1) tends to support this conclusion. Other good separations were obtained [7] with an appropriately functionalized macroporous polymeric resin column. Because these resins are porous, they likely contain some stagnant mobile phase, but the amount of stagnant mobile phase is much less than that of gel resin columns and not nearly enough to account for retention of any of the carboxylic acid solutes. The separation mechanism seems to be a partitioning of solutes between the mobile phase and the resin. A nonfunctionalized polymeric resin can be used, but greater retention and better separations are obtained using a resin with polar substituents such as sulfonate or carboxylate. Good separations were obtained on a carboxylic acid resin column under conditions in which the resin’s carboxylic acid group was in molecular rather than ionized form. Separations are significantly faster on a macroporous resin column than on a gel resin column. Columns as short as 5 cm can be used with good results. Solvent gradients can be used with macroporous resins to obtain faster separations and sharper peaks for later-eluted compounds. A data acquisition system was used to correct a rising UV-absorbance baseline (see Figure 8.1).
211
212
8 Ion-Exclusion Chromatography
Figure 8.1 Gradient separation of carboxylic acids with background correction with a 15 cm × 4.1 mm Hamilton PRP-X300 column of 0.17 mequiv g–1 exchange capacity. Gradient: 1.0 mM sulfuric acid (pH 2.7) for 1.0 min, then a 0–20% acetonitrile linear gradient over 4.0 min, followed by a 5.0 min hold at 20%
acetonitrile. Other conditions were the same as in Figure 1. Peaks: 1 = oxalic acid, 2 = tartaric acid, 3 = maleic acid, 4 = citric acid, 5 = lactic acid, 6 = acetic acid, 7 = succinic acid, 8 = glutaric acid, 9 = propionic acid, 10 = butyric acid, 11 = valeric acid. From Ref. [7] with permission.)
It has long been held that a separation of carboxylic acids by IEC requires the use of an acidic eluent to repress ionization of the analytes and thereby give sharp peaks. However, equilibrium constant calculations indicate that alkane carboxylic acids are extensively ionized (60 – 97%) in predominately aqueous solution at the low concentrations generally used in liquid chromatography. It is questionable that added sulfuric acid is really effective in converting the solute acids to their molecular form. Morris and Fritz [8] studied the separation of low-molecular-weight carboxylic acids on a column (150 × 4.6 mm) packed with sulfonated PS-DVB resin spheres (0.25 mequiv g–1 exchange capacity). Elution with pure water gave fronted, poorly shaped peaks, as expected. However, incorporation of an organic solvent, but no
8.2 Separation of Organic Acids
acid, into the eluent gave an excellent separation of formic through valeric acid with sharp, well-shaped peaks. An optimal separation required 60% methanol (40% water v/v), 40% ethanol, 20% 1-propanol or 5% 1-butanol. Conductivity detection could be used with very good sensitivity because none of the eluents contained any added acid. The sensitivity improved markedly as the organic alcohol content of the eluent decreased. Similar peak heights were obtained at the same detector setting for analytes at the following concentrations: 60% methanol: 1 ppm formic acid, 25 ppm acetic acid, 30 ppm propionic acid, 50 ppm butyric acid, 75 ppm valeric acid. 40% ethanol: 0.25 ppm formic to 6 ppm valeric acid. 20% propanol: 0.25 ppm formic to 2 ppm valeric acid. 5% butanol: 0.25 ppm formic to 0.50 valeric acid. A separation in 5% butanol is shown in Figure 8.2.
Figure 8.2 Chromatographic separation on sulfonated PS-DVB resin column (150 × 4.6 mm) of 0.25 mequiv g–1 exchange capacity. Eluent conditions: 5% butanol/deionized water.
Detection is conductivity with an output range of 3 lS full scale. Flow rate is 1.0 mL min–1. Peaks are as identified. (From Ref. [8] with permission.)
213
214
8 Ion-Exclusion Chromatography
8.2.1 Effect of Alcohol Modifiers
The effect of methanol and ethanol on the separations could possibly be explained by stronger solvation of the sample solutes in the mobile phase. However, this explanation becomes unlikely for an eluent containing only 5% 1-butanol, the remainder being water. Scott and Simpson studied the adsorption of aliphatic alcohols, aldehydes and carboxylic acids in binary mixtures with water by ODS-2 silica [9]. They found that the distribution coefficient increases exponentially with the carbon number of the moderator. When using an aliphatic moderator having a chain length of four or five carbon atoms, the surface of a bonded phase could be completely covered with a monolayer. They stated further that the chromatographic characteristics of the surface could be changed by choosing appropriately active groups. Adsorption of a layer of alcohol on the polymeric resin surface is believed to explain the dramatic effects observed in our separations of carboxylic acids. Butanol has the highest distribution coefficient of the alcohol moderators studied, and only a low concentration in the aqueous eluent is needed to coat the resin surface. Partitioning of the various solute acids between the predominantly aqueous eluent and the coated resin surface is much different from that in the case of an uncoated polystyrene surface. In an effort to gain further insight into the retention mechanism, sodium salts of the carboxylic acids were injected rather than the acids themselves [8]. The chromatograms were almost identical, even with regard to peak height. At first it was assumed that the sulfonic acid groups on the resin converted the sodium salt to the molecular acid. But running the separations on unsulfonated PS-DVB resins still gave comparable chromatograms for the carboxylic acids and their sodium salts. In reviewing the complete chromatographic system, no source of acid other than carbonic acid (or CO2) could be found. Since the analytes are moderately weak acids, it is possible that carbonic acid could have an effect on the retention mechanism. It is known from previous work that carbonic acid has a substantial retention factor and thus spends a significant amount of time residing on the surface of the polymeric resin. It was demonstrated that carbon dioxide or carbonic acid in a 1-butanol/water eluent equilibrates with the resin to form an adsorbed layer that is responsible for conversion of carboxylate salts to the molecular form. CO2 (g) + H2O > H2CO3
(8.2)
Adsorbed H2CO3 + RCO2– → Adsorbed RCO2– + RCO2H
(8.3)
The adsorbed carbonic acid also affects the separations. Figure 8.3A shows the separation of four carboxylic acids after equilibration of the column with 5% 1-butanol eluent. Then the column was re-equilibrated with eluent that was freed of carbonic acid by placing an anion-exchange column in the hydroxide form in-
8.2 Separation of Organic Acids
line between the pump and injection valve. After 30 min, the background conductance dropped from approximately 900 nS to 180 nS, and the separation in Figure 8.3B was incomplete. After switching back to the original eluent the adsorbed carbonic acid layer gradually reformed, giving the chromatogram in Figure 8.3C. It was concluded that the presence of carbonic acid in the eluent appears beneficial for the separation of carboxylic acids so long as it is in moderate and controlled quantities. In the previous separation of aliphatic carboxylic acids, n-butanol in the mobile phase was believed to coat the surface of the polymeric resin and establish a dynamic equilibrium between the mobile and stationary phase. This increased the hydrophilicity of the resin surface and reduced the hydrophobic attraction of the analytes for the resin phase.
Figure 8.3 Chromatographic separation on sulfonated PS-DVB resin column (150 × 4.6 mm) of 0.25 mequiv g–1 exchange capacity. (A) initial separation, (B) separation after 30 min with anion-exchange column (hydroxide form) in-line between pump and injection valve, and (C) separation 30 min after removal of anion-
exchange column (initial configuration). Eluent conditions: 4% butanol/deionized water. Detection is conductivity with an output range of 3 lS full scale. Flow rate is 1.0 mL min–1. Peaks are as identified. (From Ref. [8] with permission.)
215
216
8 Ion-Exclusion Chromatography
Figure 8.4 Simultaneous ion-exclusion-CEC separation of anions and cations by elution with 5 mM tartaric acid/7.5% methanol-water at 1.2 mL min–1. Eluent conductivity: 536 lS
cm–1. Peaks: 1 = SO42–; 2 = Cl–; 3 = NO3–; 4 = eluent dip; 5 = Na+; 6 = NH4+; 7 = K+; 8 = Mg2+; 9 = Ca2+. (From Ref. [17] with permission.)
Several authors have reported that the organic solvent used in conventional aqueous–organic eluents used in HPLC undergoes an equilibrium to create a ‘third phase’ on the surface of the bonded silica. This effect is often quite small with commonly used solvents such as methanol or acetonitrile, but it can have a significant effect on chromatographic behavior when other mobile phase additives are used. McCormick and Karger [10] found that as little as 2% (v/v) of an organic modifier of higher molecular mass added to a methanol–water or acetonitrile– water mobile phase can have a major effect on HPLC separations. In a paper on the separation of ethanolamines with PS-DVB resin containing quaternary ammonium groups, Tanaka et al. [11] obtained a broad, very tailed peak for triethanolamine when water was used as the mobile phase. However, an aqueous mobile phase containing 0.2 M xylitol, fructose, glucose or sorbitol gave a sharp, well-resolved peak for each of the four analytes. The improved behavior was shown to be due to the increased hydrophilicity of the surface resulting from adsorption of the sugar. 8.2.2 Separation of Carboxylic Acids on Unfunctionalized Columns
Li and Fritz [12] successfully separated the first five alkylcarboxylic acids (formic – valeric) on a reversed-phase C18 column containing no functional groups. Water containing only a low concentration of an organic modifier served as the mobile
8.3 Simultaneous Determination of Anions and Cations
phase. It was not necessary to add any acid to the mobile phase to repress ionization of the sample acids. Sharp peaks were obtained using conductivity detection except when no organic additive was added to the mobile phase. Table 8.2 compares the retention times with different additives. These results demonstrate the dramatic effect of modifying the surface of a hydrophobic stationary phase by incorporating a larger polar additive into the eluent.
Table 8.2 Retention times for carboxylic acids on a Supercosil
LC-18 column (150 × 4.6 mm) with pure water containing organic additives. Conductivity detection was used. (From Ref. [12] with permission.) Acid
Retention time (min) Water only
2% BuOH
1% Hexanediol
0.25% Octanediol
Formic
2.0
0.9
1.4
1.1
Acetic
3.3
1.2
1.7
1.5
Propionic
6.3
1.7
2.2
.7
Butyric
20.1
2.9
3.7
2.6
Valeric
–
7.4 (fronted)
9.6
.4
to = 0.7 min
Ohta et al. [13] also separated mixtures of C1 and C7 aliphatic carboxylic acids on an unfunctionalized polyacrylate gel column. A 1.0 mM concentration of a C8 carboxylic acid (2methylheptanoic acid) was added to an acidic aqueous eluent to give lower retention times and sharper peaks for the analytes. Although a mixture of twelve carboxylic acids was well separated, a single run required about 40 min.
8.3 Simultaneous Determination of Anions and Cations
Stationary phases containing carboxylic acid functional groups are now frequently used for IEC instead of strong-acid sulfonated resins. The carboxylic acid particles are often a polymethacrylate gel. These materials have found extensive use for the analysis of low concentrations of both cations and anions in acid rain. Acid rain caused by SO2 and NOx in air is a major environmental pollution problem in many parts of the world. The major cationic components of acid rain are H+, Na+, NH4+, K+, Mg2+ and Ca2+, and the major anionic components are Cl–, NO3– and SO4–. The ionic balance between the total positive charge and negative charge of
217
218
8 Ion-Exclusion Chromatography
these ions is almost 100%, so the simultaneous determination of these ions is important. Tanaka et al. devised a method for the simultaneous determination of anions and cations in acid rain [14]. The column was packed with a weak-acid cation exchange resin (TOSAH TSK gel OA-PAK). Anions were separated by IEC and cations by cation ion-exchange chromatography. With water alone as the mobile phase, weak-acid anions and strong-acid anions were separated only as a group, and cations remained fixed on the column. When the aqueous mobile phase contained sulfuric acid, it was possible to separate the cations by cation exchange, but no separation of the anions was obtained. An eluent containing tartaric or citric acid (pK1 ≈ 2–3) made it possible to separate simultaneously both the cations and anions in acid rain. The optimized eluent contained 5 mM tartaric acid and 7.5% methanol. A separation of the ions in acid rain is shown in Figure 8.7 using conductivity detection. The peaks of the anions (sulfate, chloride and nitrate), which are highly ionized, are positive. The cation peaks are of lower conductivity than the tartaric acid eluent and hence are in the negative direction. The detection limits are low enough to handle most acid rain samples without any preconcentration (Table 8.3).
Table 8.3 Detection limits of major anions and cations related to
acid rain water determined by elution with 3 mM tartaric acid/ 7.5% methanol–water. Detection limita
Ion
lM
ppb
SO42–
0.16
15
–
0.10
3.6
0.14
9
Na+
0.20
4.6
NH4+
0.30
5.4
0.32
12.5
0.28
6.8
0.38
11.2
Cl
NO3
K
–
+ 2+
Mg
2+
Ca
a Signal-to-noise ratio = 3
8.3 Simultaneous Determination of Anions and Cations
In a later study [15], a highly sensitive method involving simultaneous ionexclusion and cation-exchange chromatography was accomplished by elution of the analyte ions with a strong acid (sulfosalicylic acid) having hydrophobic characteristics instead of the more hydrophilic acid (tartaric acid) used in the previous study. However, an unusually large column (two 15 × 7.8 mm columns connected in series) necessitated a long run time of about 30 min. The influence of acidic eluents on the retention behavior of anions and cations in ion-exclusion / cation-exchange chromatography was examined by a Japanese group [16]. Common inorganic anions and two weak-acid organic anions, as well as Na+, NH4+ and K+, were separated on a carboxylic acid gel column in < 10 min with a weakly acidic aqueous eluent (Figure 8.5).
Figure 8.5 Simultaneous separation of anions and cations on a TSK weak-acid gel column, 150 × 6.0 mm. Eluent: 20 mM succinic acid. Conductivity detection. Peaks: (1) SO42–,
(2) Cl–, (3) NO3–, (4) I–, (5) F–, (6) HCOO–, (7) CH3COO–, (8) elution dip, (9) Na+, (10) NH4+, (11) K+, (12) Mg2+, (13) Ca2+. (From Ref. [14] with permission.)
Several organic acid eluents were studied with regard to background conductivity (Table 8.4), ability to resolve sample peaks and retention time of the elution dip. Sulfosalicylic acid gave peak resolution but also high background conductivity for anions of weaker acids. Eluents containing a benzene ring (benzoic and salicylic acids) undergo considerable hydrophobic adsorption to the stationary phase and give eluent dips with long retention times. Tartaric acid and sulfosalicylic acid eluents, which are acids of low pKa, were suitable for rapid separation of strong-acid anions and of cations, but the high background conductivity level gave poor detector response for weak-acid anions. When propionic acid (pKa = 4.66) was used as the eluent, resolution of weak-acid anions was improved, but retention times for Ca2+ and Mg2+ were very long. With a succinic acid eluent (pKa = 4.00), all analyte ions were well resolved, as shown in
219
220
8 Ion-Exclusion Chromatography
Figure 8.5. Retention times of Ca2+ and Mg2+ were lower because of partial complexation by succinic acid.
Table 8.4 Background conductivities of eluents used for
separation of anions and cations. (Adapted from Ref. [16].) Acid
Conc’n
Conductivity, lS cm–1
Sulfosalicylic
1 mM
706
Tartaric
6 mM
540
Succinic
20 mM
397
6 mM
354
40 mM
282
Lactic Propionic
8.4 Conclusions
The original embodiment of IEC envisioned partition of polar molecular analytes between a primarily aqueous mobile phase and stagnant liquid within the stationary phase. Inorganic anions are excluded from the stationary phase by a barrier of negatively charged functional groups in the ion exchanger. A review of the examples cited in Section 8.2 shows that the original viewpoint of IEC was too limiting. Sample analytes are separated according to differences in partitioning between the mobile phase and the entire system within the stationary phase. This system includes the polymeric matrix, functional groups on the polymeric matrix, functional groups on the polymer and liquid within the gel. Several effects, in addition to liquid partition, may attract analytes to the stationary phase: 1. Hydrophobic attraction. Retention of RCO2H analytes becomes stronger as the size of the hydrophobic R group increases. 2. Hydrogen bonding. This seems to be a definite factor in the retention of carbohydrates, which have numerous –OH groups in their structure. 3. Complexation. Cation-exchange phases in the Ca2+ or Pb2+ form complex with sugar hydroxyl groups. Retention times for analytes, and often the selectivity, can be varied by changing the properties of the stationary phase. For example, research in solid-phase extraction (SPE) has demonstrated that introduction of polar groups, such as –CH2OH, –COCH3 or – CO2H, into polystyrene resins results in greater extraction of polar
8.4 Conclusions
organic solutes from aqueous solutions. Sulfonation of polystyrene resins markedly increases the retention of polar organics with a maximum of around 0.6 mmol g–1 sulfonate groups [17, 18]. Higher degrees of sulfonation gradually decrease analyte retention (Figure 8.6).
Figure 8.6 Effect of the degree of polystyrene resin sulfonation on the retention of phenol using an aqueous mobile phase. (Adapted from Ref. [17].).
Mobile phase additives also influence the chromatographic behavior of polar organic analytes. Addition of an organic solvent such as methanol or acetonitrile leads to shorter retention times and better peak shapes. Very low concentrations of higher alcohols or alkanediols may have an even greater effect. These additives appear to form a ‘third phase’ between the mobile and stationary phases that facilitates mass transfer of analytes across the interface. An acidic eluent has customarily been employed to inhibit the ionization of alkyl carboxylate analytes. Additional H+ shifts the following equilibrium by mass action: RCO2H > RCO2– + H+
(8.4)
However, the added H+ increases the background conductivity and reduces the detection sensitivity for weak acids that need to be in the ionic form. With a proper choice of a stationary phase, the molecular RCO2H will be retained more strongly than RCO2– and the equilibrium in Eq. (8.4) will be shifted accordingly.
221
222
8 Ion-Exclusion Chromatography
When the analyte passes through the detector after it leaves the column, a greater degree of ionization will enhance the conductivity detector signal. A mobile phase containing sulfosalicylic acid or succinic acid seems to work very well. The mobile phase is acidic enough to repress ionization of weakly acidic analytes but the background conductivity is reasonably low. Some of the organic acid in the eluent is adsorbed by the resin, and this leads to improved peak shape. The only downside of this eluent is that a ’system peak’ appears later in the chromatogram that may partially coincide with an analyte peak. The system peak stems from partial desorption of the sulfosalicylic or succinic acid when an aqueous sample is injected.
8.5 Determination of Carbon Dioxide and Bicarbonate
The determination of carbon dioxide is an important analytical problem, especially when low concentrations are to be measured. Ion-exclusion chromatography provides a convenient way to determine carbon dioxide or its form in solution, which is molecular carbonic acid. The separation column is packed with a cationexchange resin in the H+ form so that salts are converted to the corresponding acid. Ionized acids pass rapidly through the column while molecular acids are held up to varying degrees. A conductivity detector is commonly used. Unfortunately, carbonic acid is a very weak acid (pK1 = 6.4), and the conductance of the carbonic acid peak is consequently very low. In order to obtain more sensitive detection, Tanaka and Fritz inserted a cation-exchange column in the K+ form between the cation-exchange column and the detector [19]. Its purpose is to convert the carbonic acid to a more highly ionized form and thereby to increase the conductivity. This is called the first enhancement column. When it is in the K+ form, the exchange reaction is as follows: Rs–K+ + H2CO3 → Rs–H+ + K+ + HCO3–
(8.5)
With this column in place a standard sample of 1.0 mM bicarbonate gave a conductance of 0.504 lS, which was approximately 5.5 times greater than that obtained with no enhancement column. 8.5.1 Enhancement Column Reactions
A calculation of the relative detector signals shows that the conversion of carbonic acid to potassium bicarbonate and then to potassium hydroxide in the enhancement columns is essentially quantitative. In particular, it may seem surprising that an acid as weak as carbonic acid (k1a = 4.0 × 10–7) is able to exchange its H+ for K+ on the resin.
8.6 Separation of Bases
H2CO3 + Rs–K+ > Rs–H+ + K+ + HCO3–
(8.6)
However, two points should be kept in mind. One is that an increasing fraction of carbonic acid is ionized as the solution becomes more dilute (∼6.8% in 0.1 mM carbonic acid, for example). A second point is that the high concentration of K+ on the exchange column (∼4.2 M) pushes the ion-exchange equilibrium to the right. For 0.1 mM carbonic acid, it can be calculated that only a few theoretical plates would be needed for complete conversion of H2CO3 to K+ and HCO3–. Strong acids will also undergo ion exchange in the enhancement columns. Since the eluent is simply water, the enhancement columns will last a long time before regeneration or replacement is necessary unless the samples analyzed have a very large amount of strong acids or their salts. It is possible that ion-exchange membrane reactors, which can be continuously regenerated, could be used in place of the enhancement columns. A second enhancement column, placed just after the first enhancement column, provides still more sensitive detection. This is an anion-exchange column in the hydroxide form that converts K+HCO3– (equivalent conductance = 118) to K+OH– (equivalent conductance = 272). The chromatograms in Figure 8.7 show the effect of enhancement columns. Each one increases the carbonic acid peak height, but the baseline conductance is also increased somewhat. A pre-column packed with anion-exchange resin in the OH– form was then placed between the pump and loop injector to remove completely and continuously the carbon dioxide in the eluent. This arrangement resulted in a significant decrease in eluent background conductance, as shown in Figure 8.7D. An almost linear calibration plot was obtained from 0.05 to 5.0 mM bicarbonate. The detection limit was estimated to be 1.45 lM.
8.6 Separation of Bases
Haddad et al. measured retention volumes for a variety of bases on a quaternary ammonium functionalized PS–DVB stationary phase using dilute aqueous sodium hydroxide as the eluent [20]. Values for the retention volumes and distribution coefficients of selected bases are given in Table 8.5. Strong bases, which are fully ionized at the eluent pH, elute at the column void volume and have a Kd value of 1.0. Solutes intermediate between these two extremes are partly ionized and generally can be separated by an ion-exclusion mechanism.
223
224
8 Ion-Exclusion Chromatography
Figure 8.7 Comparison of ion-exclusion chromatograms HCO3– with and without first and second enhancement columns and precolumn: (A) separating column alone (H+ form) (no enhancement); (B) separating column (H+ form) + first enhancement column (K+ form); (C) separating column (H+ form) + first enhancement column (K+ form) + second enhancement column (OH– form);
(D) (C) with precolumn (OH– form) for removal of CO2 gas in water eluent. Conditions: first enhancement column is TSK SCX 5 lm, 4.6 × 50 mm. Separating column is TSK SCX (H+ form) 5 lm, 7.5 × 100 mm. Eluent is water (1 mL min–1). Sample mixture of 1 mM KCl and 1 mM NaHCO3 (0.1 mL). (From Ref. [19] with permission.)
8.6 Separation of Bases Table 8.5 Retention data for basic compounds. Bio-Rad
(Richmond, CA, USA). 300 × 7.8 mm column containing quaternary-ammonium PS-DVB resin. Column void volume 3.8 mL; sum of dead and the inner column volumes, 10.3 mL. (Data from Ref. [20].) Retention vol. (mL)
Kd
Solute
pK2
KOH
–10.00
3.90
0.02
NaOH
–5.00
3.90
0.02
Ca(OH)2
2.43
4.00
0.03
Ethylenediamine
4.07
5.30
0.23
Hydrazine
5.77
6.20
0.37
Methylamine
3.34
6.60
0.43
Ammonia
4.75
6.96
0.49
Triethanolamine
6.24
7.00
0.49
Ethylamine
3.30
7.15
0.52
Trimethylamine
4.19
7.36
0.55
Propylamine
3.40
8.00
0.65
Diethylamine
2.96
8.60
0.74
Methanol
15.00
10.30
1.00
Urea
13.82
10.32
1.01
Thiourea
14.26
10.40
1.02
Higher aliphatic amines (butylamine, pentylamine, diethylamine, etc.) had larger retention volumes and Kd values well above 1.0. A mixed retention mechanism involving hydrophobic adsorption and steric effects was observed for these compounds. Aromatic amines were found to be retained almost solely by a reversed-phase mechanism involving interaction of the solute with the unfunctionalized regions of the stationary phase. Retention of these solutes could be manipulated most easily by addition of acetonitrile to the eluent. Very weak bases can be determined by IEC, but the sensitivity of conductivity detection is not very good. Tanaka et. al. [21] separated ammonium from other common cations in biological nitrification–denitrification process water by IEC with ion-exchange enhancement of conductivity. More recently, Mori et al. [22] used a similar procedure to determine hydrazine (N2H4, or N2H5OH in aqueous solution). Trace levels of hydrazine in boiled water need to be monitored.
225
226
8 Ion-Exclusion Chromatography
Two small ion-exchange enhancement columns are connected in series with the outlet of the separation column. The following reaction of hydrazine takes place in the first anion exchange column, which is in the sulfate form: 2 N2 H5 OH Resin
SO4 → 2 Resin
OH
N2 H5 2 SO24
AQ6
8:7
In the second cation-exchange column, in the H+ form:
N2 H5 2 SO4 2 Resin H → 2 Resin
N2 H5 H2 SO4
(8:8
The final product, H2SO4, is highly ionized. A 1 ppm standard sample gave a conductivity response 27 times larger than that without the enhancement columns. A linear response was obtained between 0.001 and 100 ppm for either hydrazine or ammonium ions.
8.7 Determination of Water
Molecular organic compounds, including the lower aliphatic alcohols, can also be separated by IEC. The lack of a sensitive detector is a drawback. A very important analysis is the determination of small quantities of water in chemicals, pharmaceuticals and a host of other samples. The Karl Fischer method is widely used for this purpose, but it has some limitations. If an alcohol such as methanol can be separated by ion-exclusion chromatography using water as the eluent, why not do the reverse and separate water using a methanol eluent? Stevens et al. [23] did just this. They added a small amount of sulfuric acid to the eluent and detected the chromatographic water peak by a decrease in conductivity. The main drawback with this method was a non-linear calibration curve with very poor detection sensitivity in some concentration regions. Fortier and Fritz [24] separated water by IEC and devised a unique equilibrium system for in-line spectrophotometric detection. This method has been refined and its capabilities expanded by continuing research by Chen and Fritz [25–27]. Water is separated chromatographically from the other sample components on a short column packed with cation-exchange resin in the H+ form using dry methanol as the eluent. Detection of the water peak is made possible by addition of a low concentration of cinnamaldehyde to the methanol eluent. In the presence of an acid catalyst, such as an H+-cation exchanger, cinnamaldehyde reacts with methanol to form the dimethylacetal. (+ H+) C6H5CH = CHCHO + 2 CH3OH > C6H5CH = CHCH(OCH3)2 + H2O (8.9)
8.7 Determination of Water
Figure 8.8 Spectra of 0.0318 mM trans-cinnamaldehyde in methanol. (A) Spectrum immediately after the solution was prepared; (B) Spectrum after the solution had been shaken with Aminex Q-150S in the H+ form. (From Ref. [13] with permission.)
The UV spectra of cinnamaldehyde and its reaction product (an acetal) are quite different, as shown by Figure 8.8. The necessity for an acid catalyst must be emphasized. Cinnamaldehyde dissolved in methanol will retain its own spectrum (A in Figure 8.7) for some time, but the introduction of an acid catalyst results in a very fast equilibration to form the acetal. This is accomplished by the use of a sulfonic acid cation exchanger for the chromatographic separation. Since most of the cinnamaldehyde has been converted to the acetal, the background absorbance at 300 nm is low. However, a water zone passing through the column will shift the equilibrium toward the formation of more cinnamaldehyde, and the absorbance at 300 nm will increase. H+ H2O + acetal > aldehyde + 2 CH3OH
(8.10)
In methanol the equilibrium constant, K, has been measured: K = [aldehyde]/[acetal][H2O] = 5.3 × 10–4
(8.11)
The detector signal (Adet), which is the change in absorbance when water passes through the detector cell, is given by the following equation:
227
228
8 Ion-Exclusion Chromatography
Adet = k Cca(Csamp – Cblank)
(8.12)
where k is a proportionality constant related to the equilibrium constant K, Cca is the total concentration of cinnamaldehyde added to the eluent, Csamp is the water concentration of the sample, and Cblank is the water concentration in the eluent itself. As predicted by this equation, the detector signal has been shown experimentally to be a linear function of the total aldehyde and of the concentration of water present. Typically, a sharp water peak is obtained in approximately 2 min. The water peak is always well separated from an earlier injection peak that is due to the sample matrix. Under favorable conditions a very short column (length 2.5 cm) can be used, and a water peak is obtained in as little time as 20 s [25, 26]. Determination of water in samples containing aldehydes or ketones has always been a problem. These compounds can react with a solvent (methanol) to produce an acetal or ketal plus water. RCHO + 2 CH3OH > RCH(OCH3)2 + H2O
(8.13)
The key to this problem is that the above reaction will not take place unless H+ is present to catalyze the reaction. By using a cation-exchange column in the Li+ form, the water in the acetone can be separated from it chromatographically. An H+ form column placed in series then catalyzes the cinnamaldehyde–acetal equilibrium shift that is necessary for detection of the water. Reaction with methanol to form water is also catalyzed in this second column, but separation of the acetone and initial water has already taken place in the first column. The system described here is an example of a post-column reaction system which uses a solid-phase reactor. The set-up is simple and works very well. The reactants are already present in the mobile phase. The reaction simply does not occur until the catalyst column is reached. No additional reagents are mixed with the effluent stream. There is no need for the additional hardware (second pump, mixing tee or reaction chamber) commonly used in post-column reaction systems. Consequently, the problems inherent in a typical post-column reaction system are avoided. Chromatograms for determination of small amounts of water in several aldehydes are shown in Figure 8.9. This same ‘two-column’ method is also useful for other difficult samples. For example, peroxides are highly oxidizing and interfere with the Karl Fischer titrimetric determination of water. Water has been determined by the chromatographic method in several organic peroxides [25, 26].
8.7 Determination of Water
Figure 8.9 Determination of water in various aldehydes (A) 0.11% water in acetaldehyde, (B) water in propionaldehyde, (C) 0.81% water in heptaldehyde, and (D) 0.19% water in octylaldehyde. (From Ref. [25] with permission.)
8.7.1 Determination of Very Low Concentrations of Water by HPLC
The inability to obtain really dry methanol limits the ability of the liquid chromatographic (LC) method to determine very low concentrations of water. Attempts to remove water from methanol by treatment with molecular sieves or distillation from calcium hydride still gave a product with at least 50 to 150 ppm water. Virtually all the water can be removed from methanol by adding an ortho ester, trimethylorthoformate (TMOF) and a small amount of sulfuric acid to catalyze the reaction [25, 26]. H+ CH(OCH3)3 + H2O → HCO2CH3 + 2 CH3OH
(8.14)
The amount of TMOF needed to react with the water in the methanol eluent (containing cinnamaldehyde and a low concentration of sulfuric acid) is determined by a titration procedure. At first, each addition of TMOF solution to the eluent produces a decrease in the detector signal as the eluent is pumped through the chromatographic UV/Vis detector. As the water concentration in the methanol
229
230
8 Ion-Exclusion Chromatography
becomes progressively lower, the detector signal changes less and less. TMOF is added just until there is no further lowering of the detector signal. Removal of almost all the water from the methanol reduces the eluent baseline considerably and also increases the height of the water peak for a sample. With this treatment the detection limit is estimated to be <5 ppm water.
8.8 Separation of Saccharides and Alcohols 8.8.1 Introduction
IEC is widely used for the separation and determination of hydrophilic analytes, such as carbohydrates and alcohols, as well as for anions of weak acids. Analytical samples often contain carbohydrates (sugars) and organic acids, and it is desirable to perform the analysis in a single run. Wines containing, for example, organic acids, residual sugars and alcohols can be analyzed using ion-exclusion type columns. Mono-, di-, and polysaccharides, sugar alcohols and organic acids are often found together in food and drinks. Two general chromatographic methods are available for the determination of carbohydrates. Sugars are anionic around pH 10–12 and can be determined by anion-exchange chromatography. Hydrophilic carboxylate anions are often separated in the same run. The other approach is to separate carbohydrates and sugar alcohols at a slightly acidic pH by IEC. 8.8.2 Separation Mechanism and Control of Selectivity
The analytical columns used for separation of carbohydrates are similar to those employed for separation of weak organic and inorganic acids. The column contains fully sulfonated polystyrene polymer beads cross-linked with polydivinylbenzene. The polymers are fully hydrated and contain occluded water within the gel polymer matrix, just as in ion-exclusion polymer beads. Analytes partition between the occluded water within the bead matrix and the mobile phase. Water is most often used as the mobile phase. A major difference is that columns intended for carbohydrate separations may have either H+, NH4+, K+, Ca2+ or Pb2+ as the counter ion to the negative sulfonate groups within the resin. The hydroxyl groups on carbohydrate analytes form complexes of varying strength with the metal ions in the resin. This complex formation has a major effect on the retention of carbohydrates, and the separation mechanism is often described as ‘ligand exchange’. In chromatography it is always necessary to adjust conditions so that retention times of the analytes fall within a convenient time period. This is usually accomplished by adjusting the ionic concentration of the mobile phase, but water is commonly used as the mobile phase for carbohydrate separations, so other means
8.8 Separation of Saccharides and Alcohols
231
must be sought. Changing the ionic form of the column from hydrogen to sodium and then to calcium generally increases retention times and improves resolution for most carbohydrates. The hydroxyl groups on the carbohydrate form a more stable complex with the metal, resulting in an increase in retention. Changing to lead or silver has an even more dramatic effect, but may also result in unnecessarily long analysis times. Table 8.6 shows the effect of the ionic form of the polymer.
Table 8.6 Comparison of metal ionic form and cross-linking.
Data courtesy of Transgenomic. Retention Time (min) CHO611 CHO620 CHO682 COR87H COR87N COR87K H Na K Na Ca Pb 8 6 6 6 8 8
COR87C Ca 8
COR87Pb Pb 8
Arabinose
11.08
10.64
23.95
Digitoxose
10.18
10.26
21.95
Fructose
10.33
10.07
25.84
Fucose
10.96
10.57
Galactose
10.22
Glucose Mannose
Column Name ionic form % cross-linking Compound:
12.08
12.64
14.72
13.92
16.32
11.40
12.32
14.19
15.48
11.25
11.61
13.31
13.63
16.96
24.16
12.80
12.34
14.39
13.82
16.44
9.58
22.32
11.12
11.44
13.36
13.82
15.16
9.53
8.72
19.14
10.57
10.72
12.55
11.17
13.38
10.27
9.79
25.50
11.13
11.57
13.74
12.76
16.76
Rhamnose
9.88
9.64
22.56
11.94
11.08
12.83
12.86
15.26
Sorbose
9.33
9.50
22.38
10.08
11.08
12.66
12.86
15.24
Tagatose
10.29
11.53
11.15
11.36
12.82
16.46
20.80
Xylose
10.34
9.56
20.64
11.32
11.77
13.69
12.32
14.42
Cellobiose
7.17
6.65
15.58
8.43
7.90
9.26
8.94
10.98
Lactose
7.51
7.01
17.37
8.77
8.18
9.63
9.44
11.84
Lactulose
7.85
7.57
20.70
9.00
8.48
10.08
10.17
13.24
Melibiose
7.46
6.99
17.63
8.56
8.19
9.72
9.36
12.02
Trehalose
7.14
6.70
15.98
8.64
7.85
9.02
9.07
11.20
Sucrose
7.27
6.76
15.70
7.99
9.11
9.09
11.10
Maltose
7.37
6.89
16.61
8.57
8.08
9.48
9.17
11.54
Ribitol
10.13
10.94
30.72
12.44
11.26
11.84
15.55
20.44
232
8 Ion-Exclusion Chromatography Table 8.6 (Continue) Comparison of metal ionic form and cross-
linking. Data courtesy of Transgenomic. Retention Time (min) Column Name ionic form % cross-linking
CHO611 CHO620 CHO682 COR87H COR87N COR87K Ca6 Pb H Na K Na 8 8 6 6 8
COR87C Ca 8
COR87Pb Pb 8
Arabitol
10.52
12.32
39.82
12.65
11.64
12.10
18.36
25.24
Galactitol
10.23
13.05
52.43
11.80
11.15
11.61
20.46
31.60
Myo-Inositol
11.01
10.82
35.58
11.02
12.48
14.08
14.27
20.06
Lactitol
7.87
8.55
33.23
9.26
8.45
9.34
12.17
19.50
Maltitol
7.68
8.54
30.38
9.00
8.28
9.06
12.22
17.76
Mannitol
9.90
11.84
40.03
11.66
10.81
11.42
17.81
24.98
Sorbitol
10.38
13.64
56.56
11.77
11.32
11.86
21.34
33.40
Xylitol
11.01
13.93
51.15
12.82
12.16
12.64
21.30
31.10
Amiprylose
4.20
4.50
6.86
5.74
6.42
7.68
9.46
Melezitose
6.01
5.78
13.85
6.81
7.82
8.20
13.08
Maltotriose
6.22
5.91
15.17
6.98
8.16
8.28
10.54
Raffinose
6.10
5.86
14.40
6.88
7.92
8.24
10.22
Stachyose
5.39
5.28
13.41
6.33
7.28
7.77
9.58
Maltotetrose
5.54
5.37
14.07
7.30
6.42
7.46
7.80
9.84
Maltopentose
5.08
5.00
13.08
7.10
6.11
7.02
7.53
9.34
Maltohexose
4.87
4.78
12.24
7.00
5.94
6.74
7.38
8.80
Maltoheptose
4.60
4.66
11.74
6.96
5.84
6.61
7.28
8.52
Nitrate
4.20
4.50
10.30
6.85
5.70
6.40
7.30
8.40
7.72
Flow Rate: 0.5 mL/minute; Temperature: 90 °C
The calcium-form columns are the most popular and are the column of choice for most separations. Some separations require resin to be in different metal ionic forms. Arabinose and mannose, for example, are quite difficult to separate except with a lead-form column. Lead columns are used for mono-, di- and tetra-saccharides and for sugar alcohols. Hydrogen-form columns are also useful when mixtures of sugars, acids and alcohols are to be separated. Ion exclusion or ligand exchange are both appropriate names for this type of chromatography because both separation mechanisms apply. Because the column is in the hydrogen form, it is possible to use an eluent
8.8 Separation of Saccharides and Alcohols
containing an acid without fear of changing the ionic form. The pH of the eluent will control the selectivity of the organic acid analyte, but generally will not affect the sugars or the alcohols. Table 8.7 shows that lowering the eluent pH will retain organic acids more strongly, but will not affect sugars or alcohols. Generally, increasing column temperature will increase organic acid ionization and will decrease retention of those analytes.
Table 8.7 Retention times of acids, sugars and alcohols on a Transgenomic COR87H ion-exclusion column. Eluent: H2SO4, 0.01 N, 0.003 N, and 0.001 N. Flow Rate: 0.6 mL min–1. Temperature: 25° C. Detection: RI.
Substance
Retention Time (min)
Substance
0.01 N 0.003 N 0.001 N pH 2.1 pH 2.5 pH 3.0
Retention Time (min) 0.01 N 0.003 N 0.001 N pH 2.1 pH 2.5 pH 3.0
Oxalic acid
6.4
6.2
6.1
Ascorbic acid
9.9
9.8
9.5
Sucrose
6.8
6.9
6.8
Malonic acid
9.9
9.8
9.4
cis-Aconitic acid
7.1
9.8
8.4
Glyceric acid
10.5
10.5
9.9
11.2
10.3
9.0
11.2
11.2
11.0
Glucutonic acid
7.3
7.2
6.9
t-Aconitic acid
Oxaloacetic acid
7.5
7.5
6.7
Glucotonic acid c-Lactone
Citric acid
7.8
7.6
7.2
Glycerol
12.1
12.5
12.4
Isocitric acid
8.0
7.8
7.5
Lactic acid
12.1
12.1
11.8
a-Ketoglutaric acid
8.2
7.5
6.8
Succinic acid
12.1
12.1
11.8
Glucose
8.3
8.3
8.2
Shikimic acid
12.2
12.4
12.0
Ginconic acid
8.3
8.1
7.6
Formic acid
13.6
13.5
13.1
Galacturonic acid
8.3
8.1
7.6
Barbituric acid
13.9
13.5
13.5
Gluconic acid
8.4
8.4
7.9
Acetic acid
14.6
14.7
14.6
Maleic acid
8.4
7.2
7.2
Glutaric acid
15.0
14.9
14.7
Tartaric acid
8.5
8.3
7.7
Fumaric acid
16.8
15.4
13.6
Isocitric acid
8.7
7.4
7.2
Methanol
16.9
17.2
17.2
d-Lactone
233
234
8 Ion-Exclusion Chromatography Table 8.7 (Continue) Retention times of acids, sugars and
alcohols on a Transgenomic COR87H ion-exclusion column. Eluent: H2SO4, 0.01 N, 0.003 N, and 0.001 N. Flow Rate: 0.6 mL min–1. Temperature: 25° C. Detection: RI. Substance
Retention Time (min)
Substance
0.01 N 0.003 N 0.001 N pH 2.1 pH 2.5 pH 3.0
Retention Time (min) 0.01 N 0.003 N 0.001 N pH 2.1 pH 2.5 pH 3.0
c-Lactone Pyruvic acid
8.9
7.9
7.4
Propionic acid
17.3
17.3
17.2
Glyoxillic acid
9.1
8.9
8.6
Adipic acid
18.8
18.5
18.2
Fructose
9.2
9.3
9.1
Ethanol
18.8
18.7
18.8
Citramalic acid
9.4
9.3
8.7
Acetone
21.3
21.2
21.4
21.4
21.5
21.3
34.0
29.1
23.1
Malic acid
9.5
9.5
8.9
n-Butyric acid
Quinic acid
9.8
9.7
9.2
o-Phthalic acid
There are other subtle changes that can take place in column manufacturing that will provide differences in column performance. Changes in cross-linking density, microporosity, ion exchange capacity, and particle size all will affect the column performance. The beads are sulfonated to a capacity of 3–5 mequiv g–1 (1–2 sulfonic acid groups are present per polymer aromatic group). The polymeric beads used for these separations are low-cross-linked gel-like material. A more highly cross-linked bead is more mechanically durable. However, the selectivity may not be suitable for separation. In general, less highly cross-linked gels give the best separations, but they are the most fragile. The column bed can collapse through high pumping pressures. Once the bed starts to collapse, the feedback mechanism is positive, leading to higher eluent back-pressures and further collapsing of the column bed. By far the most fragile column is the Pb column, because this column is usually very low cross-linked. One should not attempt to pump eluent through this or any carbohydrate column without first heating the column. It may also be necessary to elevate the temperature of the column to eliminate separation of the various stereoisomers of monosaccharides. Besides keeping the backpressure of the eluent low, the higher eluent temperatures usually improve the separation. The optimum temperature is usually between 60 and 85 °C, and one should experiment to find the best conditions for each particular analysis. On the other hand, if chiral separation for sugars is desired, lowering column temperature to ambient and below slows mutarotation and allows resolution of many isomers.
8.8 Separation of Saccharides and Alcohols
8.8.3 Detection
Refractive index is the most common means of detection because the samples are not concentration limited (there is plenty of fructose in corn syrup, for example) and refractive index is a universal detector. UV detection is common when the analyte is UV absorbing. If high sensitivity is needed, pulsed amperometric detection (PAD) may be used to detect carbohydrates after separation by ion exchange. The base that is used to ionize the sugars also serves as a medium to allow oxidation of the sugars by PAD (see Chapter 4). Sugars will not oxidize and be detected so easily unless the pH is basic. Extra base is often added for PAD to work for ionexchange separations of carbohydrates (Dionex Technical Note 20). The same can be done for PAD detection of carbohydrates after separation by ion exclusion. The concentrations of sodium hydroxide needed for PAD detection are dependent on the detector setting, but are within the range 15–900 mM (Dionex Technical Note 21). An evaporative light-scattering detector (ELSD) can be used for IEC separations in which water is the eluent. Li, Chen and Zhu [28] also used this detector for the separation of sugars by anion chromatography. Glucose, fructose and sucrose in cola drinks were separated by anion-exchange chromatography with electricallygenerated KOH as the eluent. A suppressed-conductivity detector coupled to an evaporative light-scattering detector was used as the detector. The KOH and other salts were removed by the suppressor, leaving only the carbohydrates and water to pass through the ELSD [28]. 8.8.4 Contamination
Because of differences in swelling, columns are converted to the appropriate metal form in the bulk form before packing the column. Samples containing salts (cations) should be desalted to prevent displacement and precipitation of the metal from the column. If a metal gets displaced by other metals that may be in the sample or eluent, it is sometimes possible to regenerate the column with salt. For example, CaEDTA can sometimes be used to regenerate a Ca-form column. This regeneration should be done in the backflush mode. It is sometimes possible to remove organic contamination from a column by pumping a 40% acetonitrile solution in the backflush mode at a very low flow rate (0.1 mL min–1). This should be done only as a last resort to recover a column. The use of other types of organic solvents is not advised because bed shrinkage may result.
235
236
8 Ion-Exclusion Chromatography
References [1] R. M. Wheaton and W. C. Bauman, Ion
exclusion, Ind. Eng. Chem., 45, 228, 1953. [2] G. A. Harlow and D. H. Morman, Automatic ion exclusion-partition chromatography of acids, Anal. Chem., 36, 2438, 1964. [3] K. Tanaka and T. Ishisuka, Elution behavior of acids in ion exclusion chromatography using a cation-exchange resin, J. Chromatogr., 174, 153, 1979. [4] K. Tanaka and J. S. Fritz, Separation of aliphatic carboxylic acids by ion exclusion chromatography using a weak-acid eluent, J. Chromatogr., 361, 151, 1986. [5] Dionex Corp., Application note 25, Sunnyvale, CA, 1980. [6] D. T. Gjerde and H. Mehra, Advances in ion chromatography, P. Jandik and R. M. Cassidy, Eds, Century International, Medfield, MA, Vol. 1, p. 139, 1989. [7] J. Morris and J. S. Fritz, Separation of hydrophilic organic acids and small polar compounds on macroporous columns, LC-GC, 11, 513, 1993. [8] J. Morris and J. S. Fritz, Eluent modifiers for the liquid chromatographic separation of carboxylic acids using conductivity detection, Anal. Chem., 66, 2390, 1994. [9] R. P. W. Scott and C. F. Simpson, Solute-solvent interactions on the surface of reversed phase, Faraday Symp. Chem. Soc., 15, 69, 1980. [10] M. McCormick and B. L. Karger, Role of organic modifier sorption on retention phenomena in reversed-phase liquid chromatography, J. Chromatogr., 199, 259, 1980. [11] K. Tanaka and K. Ohta, Ion-exclusion chromatography of ethanolamines on an anion-exchange resin by elution with polyols and sugars, J. Chromatogr. A, 739, 317, 1996. [12] S. Li and J. S. Fritz, Organic modifiers for the separation of organic acids and bases by liquid chromatography, J. Chromatogr. A, 964, 91, 2002. [13] K. Ohta, A. Towata, M. Ohashi and T. Takeuchi, Application of polymethy-
[14]
[15]
[16]
[17]
[18]
[19]
[20]
[21]
crylate resin in liquid chromatography with UV detection for C1 – C7 aliphatic monoamines, J. Chromatogr. A, 1039, 161, 2004. K. Tanaka, K. Ohta, J. S. Fritz, S. Matsushita and A. Miyanaga, Simultaneous ion-exclusion chromatographycation exchange chromatography with conductimetric detection of anions and cations in acid rain waters, J. Chromatogr. A, 671, 239, 1994. K. Tanaka, K. Ohta, P. R. Haddad, J. S. Fritz, A. Miyanaga, W. Hu and K. Hasebe, Simultaneous ion-exclusion/ cation-exchange chromatography of anions and cations in acid rain waters on a weakly acidic cation-exchange resin by elution with sulfosalicylic acid, J. Chromatogr. A, 884, 167, 2000. M. Mori, K. Tanaka, T. Satori, M. Ikedo, W. Hu and H. Itabashi, Influence of acidic eluent for retention behaviors of common anions and cations by ionexclusion/cation-exchange chromatography on a weakly acidic cation-exchange resin in the H+-form, J. Chromatogr. A, 1118, 51, 2006. P. J. Dumont and J. S. Fritz, Effect of resin sulfonation on the retention of polar organic compounds in solid-phase extraction, J. Chromatogr. A, 691, 123, 1995. J. S. Fritz, P. J. Dumont and L. W. Schmidt, Methods and materials for solid-phase extraction, J. Chromatogr. A, 691, 133, 1995. K. Tanaka and J. S. Fritz, Determination of bicarbonate by ion-exclusion chromatography with ion-exchange enhancement of conductivity detection, Anal. Chem., 59, 708, 1987. P. R. Haddad, F. Hao and B. K. Glod, Factors affecting retention of basic solutes in ion-exclusion chromatography using an anion-exchange column, J. Chromatogr. A, 671, 3, 1994. K. Tanaka, R. Kutakowa, R. Nakashima and J. S. Fritz, Determination of ammonium ion in biological nitrification-denitrification process water with ion-
References exchange enhancement of conductivity detection, Bunseki Kagaku, 37, 99, 1988. [22] M. Mori, K. Tanaka, Q. Xu, M. Ikedo, H. Taoda and W. Hu, Highly sensitive determination of hydrazine ion by ionexclusion chromatography with ionexchange enhancement of conductivity detection, J. Chromatogr. A, 1039, 135, 2004. [22] M. Mori, K. Tanaka, Q. Xu, M. Ikedo, H. Taoda and W. Hu, Highly sensitive determination of hydrazine ion by ionexclusion chromatography with ionexchange enhancement of conductivity detection, J. Chromatogr. A, 1039, 135, 2004. [23] T. S. Stevens, K. M. Chritz and H. Small, Determination of water by liquid chromatography using conductometric detection, Anal. Chem., 59, 1716, 1987.
[24] N. Fortier and J. S. Fritz, Use of a post-
[25]
[26]
[27]
[28]
column reaction and a spectrophotometric detection, J. Chromatogr., 462, 325, 1989. J. Chen and J. S. Fritz, Chromatographic determination of water using spectrophotometric detection, J. Chromatogr., 482, 279, 1989. J. S. Fritz and J. Chen, New chromatographic methods for the determination of water, Am. Lab., 24J, July, 1991. J. S. Fritz, Principles and applications of ion-exclusion chromatography, J. Chromatogr., 546, 111, 1991. J. Li, M. Chen, and Y. Zhu, Separation and determination of carbohydrates in drinks by ion chromatography with a self-regenerating suppressor and an evaporative light-scattering detector, J. Chromatogr. A, 1155, 50, 2007.
237
239
9 Ion Pair Chromatography 9.1 Principles
As an alternative to conventional ion chromatography, sample ions can be separated on a standard reversed-phase column of the type used for HPLC. This is accomplished by adding an ionic reagent to the mobile phase to pair with sample ions of opposite charge. Separations are based on differences in retention of the various ion pairs by the stationary phase. Lipophilic cations, such as tetrapropylammonium or tetrabutylammonium, are common pairing reagents for sample anions. Alkyl sulfonated, C6–C10, as well as certain inorganic anions such as perchlorate or PF6– are used as pairing reagents for sample cations. Equilibration is the first step in a separation. A mobile phase containing a pairing ion and a co-ion in an aqueous-organic solution is pumped through the column until equilibration is complete and a steady baseline is obtained. Suppose a tetrabutylammonium salt, Bu4N+Cl–, is to be used in the mobile phase for separation of the sample anion. The larger cation equilibrates between the two phases: Bu4N+ (mob) > Bu4N+ (sta) After reconditioning, the sample is introduced as the mobile phase continues to be pumped through the column. Analyte retention has been suggested to follow a double-layer model in which the organic pairing ion occupies a primary layer on the stationary phase and the other ions in the system complete for the secondary layer [1]. Sample anion or cations are separated by differences in their affinity for the pairing ion sites on the stationary phase. An alternative mechanism is possible when organic sample ions or polarizable inorganic ions, such as iodide or thiocyanate, are to be separated. In this case ion pairs are partially formed in the liquid mobile phase and the pairs can then equilibrate with the stationary phase. Retention factors of sample ions can be adjusted by changing the chemical nature and concentration of the pairing ion and by altering the proportion of organic solvent in the mobile phase. As an example, larger k′ values are obtained for anions with R4N+ as the pairing ion in the series: R = ethyl < propyl < butyl < penIon Chromatography, 4th Ed. James S. Fritz and Douglas T. Gjerde Copyright © 2009 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim ISBN: 978-3-527-32052-3
240
9 Ion Pair Chromatography
tyl. Retention factors are generally higher when the concentration of pairing ion is increased. However, k′ values decrease when more organic solvent is included in the mobile phase. Ion-pair chromatography is very versatile. Either anions or cations may be separated on a same column simply by changing the pairing ion in the mobile phase. The exchange capacity of conventional ion-exchange columns is fixed, but the effective capacity in ion-pair chromatography can be varied by changing the eluent composition. A variety of detectors can be employed. Ion-pair chromatography is actually quite broad in scope. Chromatographic behavior can be distinctly different when organic ions, rather than inorganic, are to be separated. The precise mechanism has been hotly debated (see Section 9.3), and several names have been applied to this type of separation: ion-interaction chromatography (IIC), mobile-phase ion chromatography (MPIC), as well as ionpair (IPC) chromatography. We find the last name to be simple and descriptive. The diverse nature of ion-pair chromatography is illustrated by specific examples [2]: 1. Separation of inorganic anions on a poly(styrenedivinylbenzene) adsorbent in the presence of tetraalkylammonium salts [3]. The first of the two major equilibria encountered involves retention of the R4N+ salt as a double layer on the stationary phase surface. The second equilibrium is between analyte anions and those occupying the secondary layer of the double layer. The mobile phase variables are the structure and concentration of the R4N+ salt. The retention factor, k′, for inorganic anions was adjusted to a desired range by varying the composition of the aqueous–organic mobile phase from 17.5% to 35% acetonitrile. The retention order of monovalent anions was similar to that in conventional anion-exchange chromatography: ClO4– > I– > NO3– > Br– > NO2– > Cl– > citrate > formate > OH–. Divalent anions were strongly retained. 2. A study on the separation of aromatic carboxylate and sulfonate anions with tetrapentylammonium (TPPA) as the pairing ion showed that the co-ion has a strong influence on partitioning between the mobile and the stationary phase [1]. Pure TPA salts were injected into a polymeric column and eluted with a mobile phase containing only 3:7 acetonitrile:water, and the retention factor, k′, was calculated for each salt (Table 9.1). The elution order of monovalent anions followed that generally observed in ion chromatography. However, sulfate (a divalent anion) had a much weaker affinity for the stationary phase than that observed in conventional anion chromatography.
9.1 Principles Table 9.1 Retention of tetrapentylammonium salts on PRP-1.
Conditions: 3:7 CH3CN:H2O mobile phase, conductivity or refractive index detector (adapted from Ref. [2]). Analyte
K′
R4N+OH–
0.27
R4N+F–
2.81
R4N+Form–
3.05
+
–
3.27
+
–
3.75
+
NO3–
4.44
R4N Cl
R4N Br R4 N
3. The chromatographic behavior of C2–C12 alkane sulfonates was studied on a polymeric column (PRP-1) or a C18 silica column with a metal cation (M+, M2+ or M3+) as the only counterion in the mobile phase [4]. Retention of the sulfonates was believed to be accompanied by double layer formation at the stationary phase surface, and diffuse secondary layer is made up of counter cations. The retention factors, k′, of the analytes increased substantially when the concentration of sodium acetate in the mobile phase was increased in steps from 10–4 to 10–2 M. When the mobile phase contained metal ion chlorides at constant ionic strength, the retention factors of C6–C10 alkanesulfonates followed the order: Al3+ > Ba2+ > Mg2+ > Na+ > Li+. These trends suggest that an appreciable association is taking place between the cation and the sulfonate analytes. The association reduces the charge at the anionic center and hence increases the hydrophobic interaction between the analyte and the stationary phase. 4. Roberts [5] found that the chromatographic behavior of protonated amines on a bonded-phase silica column is influenced markedly by the identity of the mobile phase anion. For example, the retention factor values of protonated nordoxepin, nortriptyline and amitriptyline all increased approximately 6-fold across the following series of anions employed as mobile-phase modifiers: H2PO4– < HCO2– < CH3SO3– < Cl – < NO3– < CF3CO2– < BF4– < ClO4– < PF6– In contrast, a neutral hydrophobic analyte, acenaphthene, showed no significant changes with respect to the mobile phase anion.
241
242
9 Ion Pair Chromatography
These effects were felt to be consistent with nonstoichiometric, double-layer ion-pairing models [6] and the rank of the anion in the Hofmeister series. Although the Hofmeister effect is not completely understood, it is well accepted that the rank of an ion in the series is a measure of its propensity to accumulate at or near interfacial regions and that the physical quantity responsible for the effect is intimately tied to the solvation properties of the ion.
9.2 Typical Separations
A variety of inorganic and organic ions (both anions and cations) have been separated by this type of chromatography. Haddad and Jackson give numerous examples [7]. The following experimental parameters can be adjusted to obtain satisfactory conditions for a separation: . Type of stationary phase. . Type of pairing reagent. A higher molecular weight shifts the equilibrium towards the surface. . Concentration of pairing reagent. A higher concentration also shifts the equilibrium towards the surface. . Type and concentration of organic solvent (organic modifier). Increasing concentrations shift the equilibrium away from the surface. The ionic pairing reagent necessarily introduces a counterion into the system. This ion preferably should be different from any of the sample ions to be determined. Some of the most useful separations are of organic ions or ionic inorganic complexes. The separation of eight metal cyanide complexes in Figure 9.1 would be difficult to accomplish by conventional ion chromatography. Detection in this case was by direct spectrophotometry at 214 nm. Suppressed conductivity detection may be used in some cases. A micromembrane suppressor is available that is solvent resistant and is permeable to quaternary ammonium ions. A tetrabutylammonium hydroxide eluent in aqueous– organic solution is often used in the separation of anions. The tetrabutylammonium ion is sufficiently large to act as a good ion-pairing reagent. In the suppressor unit the OH– counterion is converted to water while the counterion of sample anions is converted to the highly conducting H+. A separation of C5–C8 alkylsulfonates with suppressed conductivity detection is shown in Figure 9.2. The mobile phase consisted of 2.0 mM tetrabutylammonium hydroxide in 37% acetonitrile [8]. A sensitive method for the determination of thiosulfate, thiocyanate and polythionates in a mixture with photometric detection at 230 nm illustrates further the applicability of ion-pair chromatography [9]. Excellent resolution of a complex mixture was obtained on a C18 column with a mobile phase of 6 mM tetrapropylammonium acetate pH, 5.0, in 20–80% methanol–water (Figure 9.3).
9.2 Typical Separations
Figure 9.1 Ion-interaction separation of metal-cyano complexes. A Waters Nova Pak C18 column was used with 23:77 acetonitrile:water containing 5 mM Waters Low UV PIC A as eluent. Detection was by direct spectrophotometry at 214 nm (from Ref. [7] with permission).
Systematic studies compared the effect of eluent composition on the separation of six sulfur anions. Tetraethylammonium (TEA) salts did not resolve the peak; tetrabutylammonium (TBA) gave very long retention times, but tetrapropylammonium (TPA) acetate resolved all peaks in 22 min. The effect of varying TPA concentrations from 3–11 mM at a fixed 20% acetonitrile in water and pH 5.0 was studied. Increasing the TPA concentration slowed the elution rate of the sulfur anions. At TPA concentrations of 3 and 4 mM, all six anions eluted rapidly but thiosulfate was not separated from thiocyanate. TPA concentrations above 9 mM hexathionate required an elution time of >25 min. Good resolution of all peaks was obtained at 6 mM TPA. An increase in acetonitrile concentration accelerated elution of all anions. Adjusting the percentage of organic additive in the mobile phase is an excellent
243
244
9 Ion Pair Chromatography
Figure 9.2 Separation of alkylsulfonates by ion-pair chromatography. Eluent: 2.0 mM tetrabutylammonium hydroxide in 37% acetonitrile, suppressed conductivity detection (from Ref. [8]).
way to bring retention times of sample ions into the desired range. In this example a 15:85 v/v acetonitrile:water mobile phase resulted in a long retention time for hexathonate. A 20:80 mixture gave a good separation, but a 23:77 mobile phase failed to completely separate some of the anions. Ion-pair chromatography is particularly useful for separation of organic cations and anions. In conventional ion chromatography, organic ions frequently give long retention times and rather broad peaks. Many pharmaceuticals are large basic molecules that can be determined as protonated cations by ion-pair chromatography. Many organic ions can be detected conveniently by UV. Sarzanini et al. [10] studied the separation of sulfonates and carboxylates. A divalent pairing ion, hexamethonium bromide, was used. A reagent with manysided interaction points often provides enhanced selectivity for the separation of isomeric ions. A good separation of several organic anions was obtained on a C18 column with 7 mM hexamethonium bromide, 6:94 methanol:water at pH 7 as the eluent (Figure 9.4). Addition of 60 mM NaCl to the eluent gives peaks with much improved symmetry and somewhat shorter retention times. Two geometrically isomeric solutes were easily separated: maleate (cis form) and fumarate (trans form). The hydrophobic or hydrophilic nature of a soluble strongly affects the retention time in this system. Thus, the polar effects of the two carboxyl groups gave phthalate a shorter retention time than benzoate, which has a single carboxyl group. In conventional IC, the 2- phthalate anion would be
9.2 Typical Separations
expected to elute later than benzoate. The methyl group in 4- toluenesulfonic acid results in a longer retention time than that of benzenesulfonic acid. p-Hydroxybenzoate elutes much earlier than benzoate. The –OH group makes the analyte more hydrophilic and less retained.
Figure 9.3 Chromatogram of sulfur anions. Eluent: 6 mM tetrapropylammonium acetate, pH 5, in 20:80 v/v acetonitrile/water. Peaks: 1 S2 O23
1:5 lM; 2 SCN
1:5 lM; 3 S3 O26
5 lM; 4 S4 O26 1:5 lM; 5 S5 O26
1:0 lM; 6 S6 O26
1:0 lM: (from Ref. [9] with permission).
245
246
9 Ion Pair Chromatography
Figure 9.4 Effect of NaCl on the symmetry of peaks. Top: no NaCl; bottom: 60 mM NaCl. Stationary phase: LiChroSorb RP-18, 250 mm × 4 mm. Mobile phase: 7 mM hexamethonium bromide, buffered at pH 7.0; 6:94 v/v methanol/water (from Ref. [10] with permission).
9.3 Mechanism
The mechanism of what we will call ‘ion-pair chromatography’ has been the subject of a considerable amount of investigation. Horvath et al. demonstrated the practicality of this approach [11]. They proposed an ion-pair mechanism and gave a number of ion-pair formation constants.
9.3 Mechanism
Kraak, Jonker and Huber used anionic surfactants in conjunction with a bonded-phase silica column and an organic–aqueous mobile phase for the separation of amino acids [12]. A comprehensive study was made of the parameters, including the generation of gradients. Kissinger argued that an ion-pair mechanism is incorrect [13]. The pairing reagent partitions strongly onto the stationary phase, modifying its surface charge. This implies an ion-exchange mechanism. This interpretation would appear to be valid when the pairing reagent is very strongly adsorbed on the stationary phase surface. This situation is much like using a reversed-phase HPLC column packing with a permanent coating of a surfactant for ion chromatography. It is apparent from the four examples cited in Section 9.1 that ion-pair chromatography covers a broad range of chromatographic conditions. In the first example, the pairing ion is attached to the stationary phase so strongly that it may be considered almost as a permanent coating. Sulfate is strongly retained and the elution order of monovalent anions is virtually the same as that encountered in ion-exchange chromatography. The retention factors of analyte anions decrease with a higher ion concentration in the mobile phase, but the slope of a plot of log k′ versus log eluent concentration is less than the theoretical slope for an ionexchange mechanism. In examples 3 and 4, organic analyte ions are separated with inorganic pairing ions in the mobile phase. Retention factors of the analyte ions vary according to the chemical nature of both the analyte and pairing ions. Retention factors increase with a higher ionic concentration in the mobile phase. This is the opposite of that observed in the first example and in ion-exchange chromatography. The most likely mechanism is one in which the analyte and pairing ions partition between the mobile phase and the stationary phase where they are retained as ion pairs. C+ (mobile) + A– (mobile) > C+A– (stationary) Knox and Hartwick [14] studied the effect of alkyl sulfates with different numbers of carbon atoms on the retention of cations. The amount of mobile phase additive sorbed onto the stationary phase was measured. They concluded that presorption of the alkyl sulfate and subsequent dynamic ion exchange of sample cations was the dominant mechanism and that ion pair formation in the mobile phase was not important. Although the studies of Knox and coworkers are persuasive, their conclusions are based on hydrophobic additives that sorb strongly onto the stationary phase. It is hard to imagine that the same scenario applies to less hydrophobic organic and inorganic additives, especially when organic cations are to be separated. Dai and Carr published a detailed study of the role of ion pairing of anionic additives on the separation of cationic drugs on the reversed-phase column [15]. Two major retention mechanisms were examined by which anionic additives can influence the retention of cations: (i) ion-pair formation in the mobile phase with subsequent retention of the neutral ion pair, and (ii) pre-sorption of anionic addi-
247
248
9 Ion Pair Chromatography
tives on the stationary phase followed by dynamic ion exchange or electrostatic interaction with the analyte ions. Ion-pair formation constants were measured independently by capillary electrophoresis (CE) under mobile phase conditions identical to those used in packed-column chromatography. The authors concluded that under specific conditions the ion-pair mechanism is more important than dynamic ion exchange. The relative importance of the two mechanisms varies with experimental conditions, but ion pairing remains a significant contribution. We may conclude that the mechanism of ion-pair chromatography can vary between two extremes. In one case, a very hydrophobic counterion is strongly attracted to the solid phase during the equilibration step. Relatively hydrophilic sample ions then undergo ion exchange at the mobile–stationary phase interface in a process similar to that of conventional ion chromatography. At the other extreme, extensive ion-pair formation may take place within the mobile phase when the sample ions and pairing ions are both more hydrophobic or are poorly hydrated by water. The ion pairs may then partition between the mobile and the stationary phases. A combination of these two mechanisms seems likely for many specific separations.
References [1] Z. Ishkandarini and D. J. Pietrzyk, Ion
[2]
[3]
[4]
[5]
[6]
interaction chromatography of organic acids on a poly(styrene-divinyl benzene) adsorbent in the presence of tetraalkylammonium salts, Anal. Chem., 54, 1065, 1982. J. S. Fritz, Factors affecting selectivity in ion chromatography, J. Chromatogr. A, 1085, 8, 2005. Z. Ishkandarini and D. J. Pietrzyk, Ion interaction chromatography of inorganic anions on a poly(styrene-divinyl benzone) adsorbent in the presence of tetraalkylammonium salts, Anal. Chem., 54, 2427, 1982. D. Zhou and D. J. Pietrzyk, Liquid chromatographic separation of alkane sulfonates and alkyl sulfate surfactants: effect of ionic strength, Anal. Chem., 64, 1003, 1992. J. M. Roberts, A. R. Diaz, D. T. Fortin, J. M. Friedle and S. R. Piper, Influence of the Hofmeister series on the retention of amines in reversed-phase liquid chromatography, Anal. Chem., 74, 4927, 2002. S. G. Weber, L. L. Glarina, F. F. Cantwell and J. G. Chen, Electrical double-layer
[7]
[8]
[9]
[10]
[11]
[12]
models of ion-modified (ion-pair) reversed-phase liquid chromatography, J. Chromatogr. A, 656, 549, 1993. P. R. Haddad and P. E. Jackson, Ion Chromatography, Principles and Applications, p. 189, Elsevier, Amsterdam, 1990. J. Weiss, Ion Chromatography, 2nd ed., p 251, VCH, Weinheim, Germany, 1995. Y. Miura and A. Kawaoi, Determination of thiosulfate, thiocyanate and polythiorates in a mixture by ion-pair chromatography with ultraviolet absorbance detection, J. Chromatogr. A, 884, 81, 2000. M. C. Bruzzoniti, E. Mentasti and C. Sarzanini, Divalent pairing ion for ion-interaction chromatography of sulphonates and carboxylates, J. Chromatogr. A, 770, 51, 1997. C. Horvath, W. Melander, I. Molnar and P. Molnar, Enhancement of retention by ion-pair formation in liquid chromatography with nonpolar stationary phases, Anal. Chem., 49, 2295, 1977. J. C. Kraak, K. M. Jonker and J. F. K. Huber, Solvent generated ion-
References exchange systems with anionic surfactants for rapid separation of amino acids, J. Chromatogr., 142, 671, 1977. [13] P. T. Kissinger, Comments on reversephase ion-pair partition chromatography, Anal. Chem., 48, 883, 1977. [14] J. H. Knox and R. A. Hartwick, Mechanism of ion-pair liquid chromatography
of amines, neutrals, zwitterions and acids using anionic hetaerons, J. Chromatogr., 204, 3, 1981. [15] J. Dai and P. W. Carr, Role of ion pairing in anionic additive effects on the separation of cationic drugs in reversed-phase liquid chromatography, J. Chromatogr. A, 1072, 169, 2005.
249
251
10 Zwitterion Stationary Phases 10.1 Introduction
Zwitterion is a German word for an ion with both a positive and a negative charge. A suitable column for ion chromatography is easily prepared by coating a reversed-phase HPLC column with a zwitterionic surfactant. The most common type is a sulfobetaine stationary phase of the formula type:
R N
CH3 2 CH2 CH2 CH2 SO3 The R– group is usually an alkyl chain with 12-16 carbon atoms. A steroidal surfactant, such as CHAPS or CHAPSO2, is another type of stationary phase. A phosphocholine surfactant with the negative charge nearest to the R head group is used primarily for the separation of cations: O CH3—(CH2)15—O
P O—(CH2)2—N(CH3)3 + O–
A zwitterion phase has some unique properties. Sample anions are attracted to the N+ sites and cations are attracted to the negative sites. If the positive and negative sites are separated by a sufficient distance, these sites have little effect on one another. But in the examples shown, anions are attracted to the positive sites and are simultaneously repulsed by the nearby negative sites. Cations undergo a similar attractive–repulsive effect. Thus, the chromatographic properties of zwitterionic phases differ in several ways from the anion- and cation-exchange columns used in conventional chromatography: 1. Anions and cations are taken up simultaneously by zwitterion phases. The strength by which they are retained is affected by the properties of both the anion and the cation. This is different from conventional IC where the cation has
Ion Chromatography, 4th Ed. James S. Fritz and Douglas T. Gjerde Copyright © 2009 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim ISBN: 978-3-527-32052-3
252
10 Zwitterion Stationary Phases
little to do with the retention strength of sample anions by the anion exchanger. 2. With zwitterion phases, the charged group nearest to the hydrophobic chain determines whether the sample anions or cations will be retained more strongly. Thus, in a sulfobetaine (SB) phase, sample anions will be retained more strongly because of the close proximity of the N+ to the alkyl chain. 3. Sample ions can usually be eluted from a zwitterion column under milder conditions than are required with a regular ion-exchange column. A dilute eluent containing ions of low conductivity gives greater detection sensitivity for sample anions or cations. 4. Since sample anions and cations are retained in pairs by zwitterion phases, they can also be eluted in pairs with pure water as the eluent. This is of course a very favorable situation for conductivity detection of sample ions. However, multiple peaks may be obtained when the sample contains more than one anion and more than one cation. For example, cations C1 and C2 can combine with anions A1 and A2 to form four peaks: C1A1, C1A2, C2A1 and C2A2 The relative amounts of an ion (such as C1) is a function of the relative affinity of each ion pair (such as C1A1, C1A2) for the stationary phase. Multiple peaks for 2+ and 3+ cations can be avoided by passing the sample through an anion exchanger to give only one kind of anion, or more simply by adding an excess of salt such as NaI to the sample [1]. Na+ I– is so weakly retained that it elutes before any sample ions, and I– combines more strongly than other anions with the sample cations so that only one peak is obtained for each cation. 5. The ionic concentration of the mobile phase has much less effect on the retention times of sample ions on a zwitterion column than it does on a conventional IC column. At the eluent concentrations generally used with a zwitterion phase, an increase in eluent concentration results in only a slight decrease in the retention times. Sometimes longer retention times are observed as the eluent concentration increased. 6. The elution order for anions has been reported to be: F– < HPO42– < SO42– < Cl– < NO2– < Br – < NO3– < ClO3– < I– < ClO4– [2]. The early elution of hydrogen phosphate and sulfate is quite different from the elution order observed in ionexchange chromatography.
10.2 Simultaneous Separation of Anions and Cations
10.2 Simultaneous Separation of Anions and Cations
In 1981, Knox and Jurand [3] separated nucleotides, which have both positive and negative charges, on a reversed-phase column using an eluent containing a zwitterion, (1.25 mM 11-aminodecanoic acid) in addition to 75 mM ammonium phosphate. The retention mechanism was attributed to the formation of a quadripole between the zwitterion and two oppositely charged sites on the nucleotide. The enhanced retention of nucleotides was due to the 11-aminodecanoic acid that was adsorbed by the stationary phase. Yu and Hartwick [4] were apparently the first to use a zwitterion stationary phase in HPLC. The column was packed with silica particles to which the following chain was attached: Silica
OSiCH2 C6 H4 CH2 N
CH3 2 CH2 CH2 SO3 R N
CH3 2 CH2 CH2 CH2 SO3
A mixture of organic cations, anions and neutral solutes was separated with a mobile phase of a 10 mM ammonium phosphate buffer in 15–85% methanol– water. The results suggest that charged analytes were retained by a combination of electrostatic and hydrophobic attractive effects. In 1993, Hu, Takeuchi and Haraguchi [5] described a technique in which anions and cations were separated simultaneously on a zwitterion stationary phase with pure water as the eluent. An octadecyl silica column acquired a semi-permanent coating when a solution of a zwitterionic surfactant was passed through the column. It was demonstrated that analyte cations and anions were eluted together. The ability to separate ions by ion chromatography with a mobile phase containing no chemicals whatsoever is an ideal situation for detection of the ionic analytes. To explain their experimental results, it was suggested that the inorganic anions and their counter cations combine to make ′ion-pairing-like′ forms, as shown in Figure 10.1. During passage through the column, an analyte anion is attracted by the positive charge in the zwitterionic stationary phase. During passage through the column, an analyte anion is attracted by the positive charge in the zwitterionic stationary phase. But the anion is simultaneously repulsed by the positive charge, which is very close. The analyte cation will have the same electrostatic behavior. Therefore, neither the anion nor the cation can get close to the zwitterionic stationary phase. Nevertheless, there is sufficient attraction of an ion pair for the stationary phase to slow its migration through the column. As in any type of chromatography, separation of analytes is based on differences in their attraction to the stationary phase. When water is used as the eluent, peaks for all possible ion pairs may be obtained. However, some cation–anion pairs form more readily than others [6]. This is illustrated by the separation of NaSCN and CaCl2 at a 3:1 molar ratio (Figure 10.2). Most of the Cl– is paired with Na+, but the Ca2+ – 2Cl– peak elutes slightly later. More of the SCN– is paired with Ca2+ than with Na+, and the Ca2+ – 2SCN– peak has a significantly longer retention time. In more complex samples, some of the possible peaks have a low priority and are not seen at all.
253
254
10 Zwitterion Stationary Phases
Figure 10.1 Simultaneous electrostatic attraction and repulsion interactions between analyte ions and the zwitterionic charged stationary phase (from Ref. [5] with permission).
Figure 10.2 Chromatogram of an aqueous solution containing 3.333 mM each of Na2SO4, NaCl, NaBr, NaI, NaSCN, and Ca(SCN)2. Elution with pure water. Peaks: (1) 2Na+–SO42–, (2) Na+–Cl–, (3) Na+–Br–, (4) Ca2+–2Br–, (5) Na+–I–, (6) Na+– SCN– and (7) Ca2+–2SCN– (from Ref. [6] with permission).
10.3 Separation of Anions
Quantification of the analytical results can be a problem when an ion is eluted in more than one peak. If we determine all the ion pairs, the concentrations of the cations and anions in the original solution can be calculated. The results obtained in this manner for the analysis of tap water were in good agreement with conventional IC for chloride, sulfate and nitrate. Comparison with a separate analysis of metal cations by spectroscopy showed that the ′calcium′ peak was actually a mixture of Ca2+ and Mg2+ [7]. Macka and Haddad [8] separated the sodium salts of several inorganic anions on a column coated with a CHAPS zwitterion phase with only water as the eluent. Each of the peaks consisted of an Na+–anion ion pair. However, elution with an aqueous solution of histidine (an amino acid) at the pH of its isoelectric point produced separate peaks for Na+ and for each of the anions. It was concluded that small pH changes occurred which allowed a histidine anions to pair with Na+ and a histidine cation to pair with the sample anions. By comparison, a chromatographic run using a morpholinoethanesulfonic acid (MES) as anion and histidine as the eluent cation at pH 6.1 gave the same elution order as the histidine mobile phase, but the peaks were sharper and more symmetric.
10.3 Separation of Anions
In practice, it is often more convenient to determine anions on a sulfobetainecoated column with an aqueous eluent containing a low concentration of a weakly retained salt such as sodium tetraborate. This eluent has a low background conductivity and allows detection of common inorganic anions. Spectral detection is convenient for anions that absorb in the UV region. A separation of this type is illustrated in Figure 10.3 where inorganic anions are separated on a C18 column coated with CHAPS. The eluent was a 10 mM sodium phosphate buffer at pH 6.8 [9]. A sulfobetaine column is used primarily for the separation of anions because retention of mono- and divalent metal cations is weak. A column of this type has been used for the determination of very low concentrations of iodide in seawater [10]. The stationary phase shows almost no affinity for sulfate and very low affinity for chloride, but iodide is strongly retained. An eluent containing 0.2 mM sodium perchlorate was used to elute iodide as a sharp peak in about 6 min. UV detection of iodide at 210 nm was employed. Anions have been separated on a fluorinated carboxybetaine phase that was bound very tightly on a C18 column because of its high hydrophobicity [11]. This surfactant differs from the betaine-type surfactants used in previous studies in that the terminal functional group is a carboxylate group. The N+ and –CO2– sites are separated by only two –CH2 groups instead of three in the sulfobetaine phases. The carboxylic acid group is a weak acid, and its degree of ionization can be controlled by changing the eluent pH. When basic sodium hydrogen carbonate eluents were used, all of the test anions were unretained. This is probably the
255
256
10 Zwitterion Stationary Phases
Figure 10.3 Chromatogram of a mixture of inorganic solutes. Separation conditions: column, 250 mm × 4.6 mm i.d. packed with ODS and coated with CHAPS micelles; mobile phase, phosphate buffer; flow rate, 0.7 mL min–1;
detection at 230 nm. Analyte: aqueous solution of 1 mM each of thiosulphate, iodate, nitrite, nitrate, iodide and thiocyanate (from Ref. [9] with permission).
result of strong repulsion of the sample anions by the closely positioned positive and negative sites on the zwitterion phase. However, good resolution of inorganic anions was obtained when a more acidic pH was employed and the negative charge of the carboxyl group was diminished. The effects of pH control were demonstrated by separation of inorganic anions with an acetic acid/sodium acetate eluent. At pH 5.6, the seven anions were separated in only 12 min, but at pH 5.2 the anions were more strongly retained and the separation required 25 min.
10.4 Separation of Cations
The great majority of chromatographic methods have used a reversed-phase silica column coated with a zwitterionic surfactant of the sulfobetaine type. Anions are generally retained more strongly than cations on this type of column. It has been stated that EIC is a method for the separation of anions. The strong retention of anions by the sulfobetaine phase is probably due largely to the close proximity of the hydrophobic chain (often called the ‘head group’) to the N+ sites.
10.4 Separation of Cations
A column coated with n-hexadecylphosphocholine has the hydrophobic head group adjacent to the negative phosphate group. The positive and negative sites on phosphocholine (PC) phases are separated by two –CH2 groups. On the sulfobetaine phases, the oppositely charged sites were separated by three –CH2 groups. Again, sample cations and anions are taken up and eluted in pairs. The cation is attracted to the negative phosphate group and the anion pairs up with the positive nitrogen. Sample ions are retained more weakly by PC phases than with sulfobetaine columns. This may be because of greater electrostatic repulsion on the PC columns. Since sample ions are both taken up and eluted in pairs with water as the eluent, the chromatographic peak for a sample cation (C+) will actually be a cationanion pair (C+A–). The retention time for such a peak will be a function of how tightly A–, as well as C+, is retained by the phosphocholine. The retention time for C+ can be either increased or decreased by the choice of the accompanying anion, A –. When a sample contains more than one anion, multiple cation-anion peaks can result when water alone is the eluent. For example, cations C1 and C2 can combine with anions A1 and A2 to form four peaks: C1 A1, C1 A2, C2 A1 and C2 A2. In order to obtain only one peak for each sample cation, the sample must be made to contain only a single anion, or an excess of the sodium salt of a strongly retained anion must be added to the sample. Since all sodium salts are eluted very rapidly by water, this extra peak will not interfere with the later sample peaks. A significant advantage of the new system is its ability to separate ions using water as the mobile phase. This leads to excellent sensitivity using conductivity detection. Another very striking property of the HDPC-coated column is that hydrogen ions are eluted after mono- and divalent metal cations. Application of these principles has been demonstrated by Hu, Hasebe, Tanaka and Fritz [11]. Figure 10.4 shows the separation of several inorganic acids with only water as the mobile phase. The sample cations (H3O+) and anions always elute as ion pairs, but the retention time of each peak is determined both by the cation and anion. Thus, if we wish to determine acidity, the retention time for H3O+ will be divided among several peaks. The total acidity of a sample can be forced to elute in a single chromatographic peak by adding at least a five-fold excess of the sodium salt of a more strongly retained anion. The sodium salt is added only to the sample; water alone is the eluent. This principle is illustrated by chromatograms of 20 mM HCl alone, HCl plus sodium nitrate, and HCl plus sodium iodide. Peak height of the H3O+ remains constant, but the retention time is longer as the pairing anion is changed (Figure 10.5). Addition of an excess of a sodium salt such as NaI also permits quantitative measurement of total acidity when the sample contains weaker acids. Suppose we have a weak acid, HA, that is only partially ionized and we add an excess of NaI, which is completely ionized. The equilibrium is shifted so that ionization is complete. NaI + HA + H2O → H3O+ + I– + Na+ + A–
257
258
10 Zwitterion Stationary Phases
Figure 10.4 Separation of a 20-lL sample containing 10 mM concentration of each acid on a phosphocholine-coated column. Eluent: pure water (from Ref. [1] with permission).
Figure 10.5 Chromatograms of 2.0 mM HCl (left trace), 2.0 mM HCl with a 5-fold excess of sodium nitrate added to the sample (center trace) and 2.0 mM HCl with a 5-fold excess of sodium iodide. Eluent was water containing 10 lM HDPC; conductivity detection (from Ref. [1] with permission).
10.5 Mechanism
Figure 10.6 Simultaneous determination of total acidity, Ba2+, Mg2+ and Ca2+ using iodide (left trace) and thiocyanate (right trace) as the counterions. Sample: 1.0 mM each of Ba2+,
Mg2+ and Ca2+ ; 3.0 mM H+ (HCl). Peaks: 1, 1′ = H+ ; 2, 2′ = Ca2+ ; 3, 3′ = Mg2+ ; 4, 4′ = Ba2+ (from Ref. [1] with permission).
All of the hydrogen is paired with iodide in an H3O+I– pair that is eluted in about 28 min. A mixed peak of Na+I– and Na+A– is eluted in about 3 min. This principle was applied successfully for the determination of acidity of samples containing oxalic, tartaric, formic, acetic, pyruvic, malonic, propionic or phthalic acid. The protons in ascorbic acid were only about 39% converted [12]. Retention times of divalent metal cations can be manipulated by addition of the sodium salt of an anion with a strong pairing propensity. The retention times of the cations were all longer when iodide or thiocyanate was added to the sample (Figure 10.6). The elution order of the cations was entirely different from the order obtained in classical ion chromatography.
10.5 Mechanism
In describing a separation process which they called ‘electrostatic ion chromatography’, Hu and coworkers [5, 6] envisioned a mechanism in which sample ions were attracted to sites of a like charge in the stationary phase and concurrently repulsed by nearby sites of opposite charge. The analyte anions and their counter cations were forced into an ‘ion-pairing-like form‘ at some distance from the zwitterion charges. Separation of various ion pairs was achieved by differences in their migration rates along the column.
259
260
10 Zwitterion Stationary Phases
As additional experimental results have accumulated, it has become apparent that the strength with which sample ions are retained is governed by a combination of effects, and not solely by electrostatic attraction and repulsion. Electrophobic attraction, solvation, hydrogen bonding, and perhaps other effects also play a major part. In Section 5.3.4 the attractive forces relating to ion exchange were divided into two categories: electrostatic (ES) and enforced pairing (EP). Actually, there are a number of similarities between ion-pair chromatography and zwitterion ion chromatography. In both of these the uptake of anion is affected by the particular cation that is present, retention of anions is enhanced by increased ionic strength of the aqueous phase and, most importantly, the elution order of anions is almost identical. A very similar elution order is observed for anions in conventional ion-exchange chromatography except for small, highly charged ions such as sulfate and hydrogen phosphate. Irgum [13] concluded that the mechanism is different from that in conventional ion-exchange chromatography. Noting that ions elute according to increasing chaotropic properties in the Hofmeister series, they suggested it is reasonable to assume that ion-exclusion can be an effect, but partitioning with the polymeric substrate cannot be ruled out since the hydrophobic character of an ion is also connected to the level of hydration. Cook et al. [2] contributed several experimental observations that are helpful in understanding the separation process. Increasing the concentration of the eluent sodium salt resulted in a substantial increase or decrease in the retention factor, k′, between zero and 1–2 mM. Plots of k′ against concentration leveled out between about 2 mM and 15 mM. This initial change was attributed to disentangling of the surface morphology so that the positive and negative sites of the zwitterion phase were no longer paired with each other. Cook used capillary electrophoresis to measure the zeta potential of a C18-functionalized capillary that had been coated with a sulfobetaine surfactant. The unequal uptake of Na+ and Cl– by the zwitterion was confirmed by a zeta potential of –31 mV. The positive charge on the zwitterion nitrogen was partially neutralized by uptake of Cl–, but the stationary phase acquired a net negative charge because of much weaker attraction of Na+ to the negative sulfonate sites. Under similar conditions, sodium cyanate gave a zeta potential of –33 mV and sodium perchlorate a value of –53 mV because of the very strong retention of perchlorate. Attraction of inorganic cations becomes stronger as the charge increases, and this causes the zeta potential to become more positive. For chloride salts, the zeta potential of Na+ is –31 mV, Mg2+ –13 mV, and Ce3+ +40 mV. Cook et al. [2] also proposed a new mechanism whereby equilibrium of the bound zwitterions with a mobile phase containing a suitable electrolyte causes the establishment of a charged layer created by the terminal sulfonate groups of the zwitterion, which acts as a Donnan membrane. The magnitude and polarity of the charge on this membrane depends on the nature of the mobile-phase ions. The Donnan membrane exerts weak electrostatic repulsion or attraction effects on analyte anions. A second component of the retention mechanism is chaotropic interaction of the analyte anion with the quaternary ammonium functional group of
References
the zwitterion. This interaction exerts the major effect on the separation selectivity, such that analyte anions are eluted in order of increasing chaotropic interactions in accordance with the Hofmeister series. A subsequent paper proposed a mechanism for the chromatography of cations on a column coated with a phosphocholine zwitterion [14]. A review of the available facts on the separation of anions on a zwitterion stationary phase points strongly to a partitioning mechanism or to a mixed mode mechanism in which partitioning is dominant, with only a weak ion-exchange component. As in ion-pair chromatography, the driving force is the tendency of analyte ions and co-ions to pass from the mobile phase and pair up as C+A– in the hydrophobic stationary phase. The retention factors of sample ions are determined by the relative ability of pairs of ions to partition between the mobile and stationary phase. This is determined primarily by the hydrophobic properties of the ion pairs rather than by the relative concentrations of eluent and sample ions in the mobile phase. This is in contrast to ion-exchange chromatography where analyte and eluent ions compete for charged sites, and a higher concentration of eluent ions reduces the retention factor of analyte ions. The areas of electrostatic charge are much more diffuse in zwitterion phases than in conventional ion chromatography. As has been suggested [2], it might be appropriate to use the name ‘zwitterion ion chromatography’ (ZIC) to refer to ion chromatographic separations in which a zwitterionic stationary phase is employed.
References [1] W. Hu, K. Hasebe, K. Tanaka and
[2]
[3]
[4]
[5]
J. S. Fritz, Determination of total acidity and of divalent cations by ion chromatography with n-hexadecyl phosphocholine as the stationary phase, J. Chromatogr. A, 956, 139, 2002. H. A. Cook, W. Hu, J. S. Fritz and P. R. Haddad, A mechanism of separation in electrostatic ion chromatography, Anal. Chem., 73, 3022, 2001. J. H. Knox and J. Jurand, Zwitterionpair chromatography of nucleotides and related species, J. Chromatogr., 203, 85, 1981. L. W. Yu and R. A. Hartwick, Zwitterionic stationary phases in HPLC, J. Chromatogr. Sci., 27, 176, 1989. W. Hu, T. Takeuchi and H. Haraguchi, Electrostatic ion chromatography, Anal. Chem., 65, 2204, 1993.
[6] W. Hu, H. Tao and H. Haraguchi, Elec-
trostatic ion chromatography. 2. Partitioning behaviors of analyte cations and anions, Anal. Chem., 55, 2514, 1994. [7] W. Hu, H. Tao, M. Tominaga, A. Miyazaki and H. Haraguchi, A new approach for the simultaneous determination of inorganic cations and anions using ion chromatography, Anal. Chim. Acta, 299, 249, 1994. [8] M. Macka and P. R. Haddad, Elution mechanism in electrostatic ion chromatography with histidine as an isoelectric ampholyte mobile phase, J. Chromatogr. A, 884, 287, 2000. [9] W. Hu and H. Haraguchi, Simultaneous determination of organic and inorganic ultraviolet-absorbing compounds in human saliva by electrostatic ion chro-
261
262
10 Zwitterion Stationary Phases matography, Anal. Chim. Acta, 285, 335, 1994. [10] W Hu, P. J. Yang, K. Hasebe, P. R. Haddad and K.Tanaka, Rapid and direct determination of iodide in seawater by electrostatic ion chromatography, J. Chromatogr., 956, 103, 2002. [11] W. Hu, P. R. Haddad, K. Tanaka and K. Hasebe, Electrostatic ion chromatography using a carboxybetaine-type zwitterionic surfactant as the stationary phase, Anal. Bioanal. Chem., 375, 259, 2003. [12] W. Hu, K. Hasebe, K. Tanaka and J. S. Fritz, Determination of total acidity
and of divalent cations by ion chromatography with n-hexadecyl phosphocholine as the stationary phase, J. Chromatogr. A, 956, 139, 2002. [13] W. Jiang and K. Irgum, Covalently bonded zwitterionic stationary phase for simultaneous separation of inorganic cations and anions, Anal. Chem., 71, 333, 1999. [14] H. A. Cook, G. W. Dicinoski and P. R. Haddad, Mechanistic studies on the separation of cations in zwitterion ion chromatography, J. Chromatogr. A, 997, 13, 2003.
263
11 Capillary Electrophoresis 11.1 Introduction
Capillary electrophoresis (CE) is a micro method in which ions are separated by differences in their rates of migration through a silica capillary. The capillary is filled with an electrolyte solution, each end of the capillary dips into an electrolyte reservoir containing a platinum electrode. The electrodes at the two ends of the capillary are connected to a high-voltage power supply (0–30 kV). Ions in solution will flow through the capillary to complete an electric circuit. The sample ions to be determined migrate at different velocities toward the electrode of opposite charge (electrophoretic flow). The sample ions are detected spectrophotometrically, or by other means, as they pass through a cell near the end of the capillary. Separations by ion chromatography and capillary electrophoresis are both based on differences in the velocities at which ions move through a column or capillary. However in IC these differences in velocity are the result of differences in partitioning of sample ions between a stationary ion exchanger and the liquid mobile phase. In CE there is no partitioning between the two phases; differences in velocity are the result of differences in electrical mobility (electrophoretic mobility) through an open-tubular capillary. An additional separation factor is introduced by addition of a soluble ionic polymer or an alkylammonium salt to the capillary electrolyte. Now, separation of analyte ions is based on differences in ion association within the capillary electrolyte as well as by differences in electrophoretic mobility of the free ions. Capillary electrophoresis has several advantages. CE is truly a micromethod and requires only a very small sample. In terms of theoretical plates, CE has at least ten times the separation ability of a typical IC system. Separations are fast and a number of experimental conditions can be adjusted to obtain a good separation of difficult samples. A basic understanding of the fundamentals is very helpful in taking full advantage of the separation possibilities offered by CE. The purpose of this chapter is to present a compact treatment of the principles of CE and an idea of its scope as applied to the analysis of inorganic and small organic ions. Ion Chromatography, 4th Ed. James S. Fritz and Douglas T. Gjerde Copyright © 2009 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim ISBN: 978-3-527-32052-3
264
11 Capillary Electrophoresis
The essential parts of a CE instrument are shown in Figure 11.1. A background electrolyte (BGE) is placed in reservoirs (R) and then pumped through the capillary (C) A typical capillary is made of fused silica, 50 or 75 lm i.d. and approximately 60 cm long, although the length may vary. A platinum working electrode (E) is placed in each reservoir and connected to a high-voltage power supply (P) capable of generating up to 30 kV. A positive power supply makes the left-hand electrode the anode (positive charge) and the other electrode the cathode. With a negative power supply these polarities are reversed. Conditions must always be arranged so that sample anions will migrate from left to right and thus flow through the detector (D).
Figure 11.1 Schematic representation of a capillary electrophoresis instrument. R, electrolyte reservoirs; E, platinum electrodes; C, fused silica capillary; D, detector; Cell; P, high voltage power supply.
11.1.1 Steps in Analysis 11.1.1.1
Capillary Pretreatment
A separation with sharp peaks for sample ions and reproducible migration times requires a clean capillary surface. This is usually accomplished by frequent rinsing of the capillary with dilute aqueous sodium hydroxide. After a water rinse, the capillary is filled with the background electrolyte (BGE) solution. The BGE contains a pH buffer and a sufficient concentration of an electrolyte (frequently 20–50 mM) to maintain a steady current.
11.1.1.2 Sample Introduction To introduce the sample, the left-hand end of the capillary is dipped into a sample vial with several centimeters hydrostatic pressure for a fixed number of seconds. This will force a small volume of liquid sample into the end of the capillary.
11.2 Principles
Another method of sample introduction is to dip the end of the capillary into the sample vial and turn on the power for a few seconds. Sample ions thus flow into the system by electrical migration.
11.1.1.3 Sample Run When the sample has been added to the system the power is turned on. Most commonly a power of + or –20, 25 or 30 kV is applied, although sensitive samples may necessitate a lower power. Sample ions will now migrate by electrophoretic flow toward the electrode of opposite charge, provided that the electrophoretic flow is either in the same direction as or greater than the electroosmotic flow (EOF) (see p. 266). Separations are based on differences in electrophoretic mobility. Tables of limiting equivalent conductivity are a good guide to predicting the relative rates of migration of sample ions. To separate several cations, a positive power supply is used and the sample ions will migrate at different rates toward the detector end (negative electrode) of the system. Many separations of anions require that the direction of electroosmotic flow be reversed. This is accomplished by adding a flow modifier, such as a quaternary ammonium salt with a long hydrocarbon chain, to the BGE. A thin layer of the flow modifier is adsorbed on the capillary surface. This gives the surface a positive charge and causes electrolyte anions to give an electroosmotic flow toward the anode.
11.1.1.4 Detection Sample ions that absorb sufficiently in the UV or visible spectral region may be detected by direct spectrophotometry. Indirect spectrophotometric detection is commonly used for ions that do not absorb. An absorptive reagent is added to the BGE, and this gives a peak in the direction of reduced absorbance when a sample ion passes through the detector. The absorbing reagent, which is sometimes called a ‘visualization’ reagent, should have a mobility that matches those of the sample ions as closely as possible. Chromate is often used for the indirect detection of anions and a protonated amine cation, such as benzylamine, for detection of cations.
11.2 Principles 11.2.1 Terms and Relationships
Several terms that apply to capillary electrophoresis are listed in Table 11.1. Separations of anions are based on differences in electrophoretic flow. Inorganic ions are generally smaller and therefore more mobile than organic ions.
265
266
11 Capillary Electrophoresis Table 11.1 Some fundamental CE definitions.
Field strength
E = V/L
V = applied potential, V L = capillary length, cm
Electrophoretic mobility
lep = LdLt/tmV
Ld = length to detector, cm Lt = total length, cm tm = migration time, s V = applied potential, V
Electroosmotic mobility
leo = LdL/t
Ld, L, V as above t = neutral marker time, s
Observed or total mobility
lOB = lep + leo
The electrophoretic mobilities of inorganic ions are an inverse function of their hydrated ionic radii. Electrophoretic mobility is also affected by the charge on an ion and by the solvent medium. Tables of limiting ionic conductance are a convenient source for estimating electric mobilities of ions. In addition to electrophoretic flow, a second type of migration called electroosmotic flow (EOF) occurs in the capillary. A minute charge on the capillary surface (called the zeta potential) results from ionization of silanol groups, which have a pKa of 6 to 7. –SiOH > SiO– + H+ Cations from the bulk solution are attracted to the capillary surface. These cations are attracted toward the cathode. Since the cations are hydrated and the inner diameter of the capillary is quite small, their migration induces a bulk liquid flow (a plug flow). This means that all of the liquid and solutes in the capillary flow at the same rate toward the cathode (electroosmotic flow). The mobility of solute resulting from EOF is termed the electroosmotic mobility (leo). EOF in fused silica capillaries is pH-dependent. The EOF is very low, around pH 3–4, but rises rapidly to around pH 5–6 as the silanol groups become progressively more ionized. Finally, the EOF levels out at around pH 8–9. The magnitude and direction of electromigration can be indicated by vectors. For sample cations in neutral to alkaline solution, the electrophoretic and electroosmotic flow vectors are in the same direction. This condition is called co-migration. The electrophoretic vectors of anions are in the opposite direction to the cationic vectors. Usually the electroosmotic vector is larger than the electrophoretic vector, so the net movement of anions is still toward the cathode. This is important because the separated sample ions (anions in this case) must all pass through the fixed detector. A separation in which the electrophoretic and electroosmotic vectors are in the opposite direction is called counter-migration. A separation with counter-migration will take longer than one with co-migration, but the separation power of counter-migration is usually much better.
11.2 Principles
11.2.2 Zone Broadening
Capillary electrophoresis (CE) is sometimes called ‘capillary zone electrophoresis’ because the analytes are separated into discrete zones as they migrate along the capillary. The zones are translated into peaks as the zones pass through the detector. The recorded picture of analyte peaks as a function of migration time is called an ‘electropherogram.’ Analyte zones in CE are generally narrower than in conventional ion chromatography. Inefficient mass transfer of analytes between liquid and solid phases, together with a curved eluent flow profile due to pumping, are major sources of zone broadening in packed columns, but these are avoided in CE. Jorgenson and Lukacs [1] showed that the theoretical plate number, N, in CE can be expressed by the following equation: N = lepV/2D
(11.1)
where lep is the electrophoretic mobility, V is the applied potential and D is the diffusion coefficient of a sample ion. This equation suggests that axial diffusion is a major cause of peak broadening in CE. Lower values for lep and applied voltage results in a longer migration time and allows for a longer period for diffusion to take place. Resistance to mass transfer between phases is a major source of peak broadening in classical ion chromatography, but is not a factor in CE. Although Eq. (11.1) is a useful guideline, additional factors will affect separation efficiency in CE. Heat generated at the high voltages employed may be dissipated more rapidly near the capillary wall than nearer to the center of the capillary. This leads to unequal migration velocities within the zone and subsequent peak deformation. For this reason, separations are often at 15–20 kV rather than at the maximum of 30 kV applied potential. Interaction of analyte ions with the capillary wall is another source of peak broadening. Some ions may interact with silanol groups or be adsorbed onto the wall surface. In such cases, peaks may become tailed or overly wide. These effects can be avoided or minimized by careful cleaning of the capillary and by using a more concentrated electrolyte so that electrolyte ions, rather than sample ions, will be attracted to the wall surface. 11.2.3 Sample Injection
The introduction of samples into capillaries by means of differential pressure (hydrostatic injection) has become the most popular method in capillary electrophoresis. For sample introduction, the sample vial is raised to a defined level above the detection reservoir for a specified time; frequently about 5 s. To terminate the injection, the end of the capillary is removed from the sample and replaced into the vial containing the background electrolyte.
267
268
11 Capillary Electrophoresis
Electrokinetic injection of the sample may also be used. The capillary tip is dipped into the sample, potential is applied for a few seconds, and the capillary tip is replaced into the electrolyte reservoir. The sample injection step is another source of zone peak broadening in capillary electrophoresis. The plug of injected sample occupies a small but not insignificant volume at the injection end of the capillary. The sample ions are distributed throughout the volume of the sample plug. However, the sample ions can be concentrated into a very narrow zone by using a sample with a total ionic strength that is much lower than that of the background electrolyte in the capillary. When power is applied to begin a separation, a process called ‘electrostacking’ takes place. Electrostacking is a technique for injection that produces an unusually sharp, concentrated band of analyte ions in the front section of the capillary. The dependence of migration velocity (v) on the electric field (E) is: m = (li + leo)E
(11.2)
where: li = individual ion mobility; leo = electroosmotic mobility. According to the electrostacking condition, the sample zones in hydrostatic injections must have much lower ionic strength than the carrier electrolyte. When the separation voltage is applied, the low ionic strength of the sample zone creates a significantly higher resistance than that sustained by the electrolyte, which consequently produces a higher field strength. The increased field strength of the sample zone forces the analyte ions to migrate faster than the electrolyte ions, which are subjected to a lower field strength. The net effect of this differential migration rate is the accumulation of sample ions inside a very narrow zone at the sample–carrier electrolyte boundary. This precondition, or electrostacking, occurs before the migration of the analyte ion zone through the bulk of the carrier electrolyte solution.
11.3 Electroosmotic Flow (EOF)
The electrophoretic migration of all species within the capillary is called electroosmotic flow or EOF. The phenomenon of EOF is a fused silica capillary may be explained as follows [2]. 1. Silanol groups in the fused silica are partially ionized. (11.3) –SiOH + H2O > –SiO– + H3O+ This gives a negative zeta potential (Z) at the capillary wall and introduces an excess of hydronium cations into the solution. 2. Water dipoles and ions from the electrolyte solution form a fluid profile next to the solid–liquid interface, a relatively rigid and stagnant inner Helmholtz layer closer to the wall,
11.3 Electroosmotic Flow (EOF)
and a more dilute and less organized outer Helmholtz layer a little farther from the interface. 3. In the CE capillary both Helmholtz layers are situated in an electric field with an orientation parallel to the solid surface. A movement of the less tightly attached hydronium cations in the outer Helmholtz layer toward the cathode is induced. These hydronium ions are interlocked within the water structure so that their movement is possible only if the bulk of the water and ions is also moving along. Thus EOF causes all ions within the capillary to migrate at the same rate. 4. The magnitude of EOF, meof, in capillary electrophoresis is given by the Smoluchowski equation: (11.4) meof = –(EZ / g) E where E is the dielectric constant, Z is the zeta potential, g the viscosity, and E the applied electric field (V cm–1). The zeta potential, and hence the EOF, in fused silica capillaries is quite small, around pH 3–4, but increases rapidly at more alkaline pH values. 5. The magnitude and direction of electroosmotic flow can be changed by various treatments of the silica inner wall. For example, addition of a cationic surfactant to the background electrolyte results in a positive surface that gives a reversed EOF, i.e., toward the anode. The magnitude of EOF is easily determined by injecting a small, nonionic solute such as phenol or dimethylformamide (DMF) that migrates only by EOF. The electroosmotic mobility, leo, is then calculated from the migration time, tm, of the solute. leo
Lt Ld Vtm
(11:5
In this equation Lt is the total capillary length (cm), Ld is the length to the detector (cm), V is the applied potential (V), and tm is the migration time of the EOF marker. The observed, or apparent, mobility of an analyte ion, lOB, is calculated from an equation identical in form to Eq. (11.5) except that tm is the migration time(s) of the ion. The actual, or observed, migration of sample ions is the net result of two vectors, electrophoretic and electroosmotic migration. The observed mobility, lOB, is expressed simply as: lOB = lep + leo
(11.6)
269
270
11 Capillary Electrophoresis
Depending on the experimental conditions, lep and leo may be in the same direction (co-migration), or in opposite directions (counter-migration). The value of the electrophoretic mobility of an ion is calculated from Eq. 11.6 after separately determining lOB and leo. 11.3.1 Effect of EOF on Separations
Migration of sample ions through a capillary is affected by both electrophoretic and electroosmotic vectors. Separations can be carried out in any of the following modes: 1. Electrophoretic migration (ep) and negligible electroosmotic migration 2. Ep and eo migration to the detector (co-migration) 3. Ep migration to the detector and a weaker opposing eo 4. Strong eo migration to the detector and a weaker opposing ep. Both 3 and 4 above are forms of counter-migration. It is perhaps easier to illustrate these separation options by examples using linear velocity (v) of the vectors. For a column of fixed Ld v = Ld / tm, or
(11.7)
tm = Ld / v
(11.8)
As an example, ions that undergo only electrophoretic migration have the values v1 = 8.08 cm min–1 and v2 = 8.00 cm min–1. For a capillary with L = 40 cm, tm values calculated by Eq. (11.8) are t2 = 5.00 min and t1 = 4.95 min. The separation factor a = 5.00/4.95 = 1.01. Only partial resolution of peaks with such a small separation factor would be expected. But if there is an EOF in the opposite direction to the electrophoretic migration of the analyte ions, longer tm values are obtained and the separation factor for the two ions is more favorable. Better peak resolution is obtained at the price of a longer run time (see examples in Table 11.2). EOF in the same direction as the electrophoretic migration (co-migration) has the opposite effect. Migration times are shorter, but the separation factors are reduced. Nevertheless, co-migration is often used to speed up separations when migration times are rather long and analyte peaks are easily resolved. Although separations are usually based on differences in electrophoretic migration of sample ions toward the detector, a counter-migration system can be set up with a strong EOF to the detector. Conditions are adjusted so that the direction of electrophoretic migration is in the opposite direction. Electroosmotic flow is the driving force and electrophoretic migration is the opposing force. In this indirect separation, the order of migration times of sample ions is exactly the opposite of that obtained in direct-migration CE.
11.3 Electroosmotic Flow (EOF) Table 11.2 Effect of an opposing EOF on migration time and
separation factor. Calculated for fixed linear velocities on a capillary, L = 40 cm. Opposing veo
Net vep
Migration time
cm min–1
cm min–1
min t1
Separation factor
t2
1
2
None
8.08
8.00
4.95
4.0
4.08
4.00
9.80
10.0
1.02
6.0
2.08
2.00
19.23
20.0
1.04
7.0
1.08
1.00
37.0
40.0
1.08
5.00
1.01
11.3.2 Control of EOF
Anions are usually separated by CE at an alkaline pH, so that weak acids will be fully ionized. A negative power supply is used so that electrophoretic migration is toward the anode. However, capillary silanol groups are ionized at an alkaline pH, resulting in a negative capillary surface and a strong EOF to the cathode. Anions have long migration times and some may have a net migration away from the detector. This situation is corrected by adding a low concentration of a cationic surfactant to the electrolyte. The surfactant forms a micellar coating on the capillary wall and reverses the direction of EOF so that both EOF and electrophoretic vectors are toward the anode. Sometimes more subtle adjustment of the magnitude and direction of the EOF is needed. Yeung and Lucy [3] reported a monotonic alteration of EOF from fully reversed to near zero by coating the capillary with variable mixtures of cationic and zwitterionic surfactants. Addition of a low concentration of an aliphatic amine salt to the electrolyte is another way to vary the EOF. A commonly occurring problem in the separation of organic cations is that the electrophoretic mobilities are so high that cation migration times are very short and peaks are close together. Figure 11.2A depicts the separation of eight pharmaceutical cations at pH 3. Although two of the peaks are barely resolved, the efficiency is excellent with an average N = 275 000 plates. The ionic concentration of the electrolyte was high enough for electrostacking to occur and the organic BGE ions kept the current reasonably low. Addition of a low concentration of tributylammonium ethanesulfonate (TBA–ESA) to the previous electrolyte produced a dramatic change in the separation of the drug cations. (Figure 11.2B and C). The distances between the various peaks are now much greater.
271
272
11 Capillary Electrophoresis
Figure 11.2 Effect of protonated tributylamine (TBA) on the separation of protonated organic amine cations. Electrolyte contains 50 mM Tris-ethanesulfonate, pH 3, plus added TBA-ESA. A. None, B. 3 mM, C. 5 mM, D. 8 mM.
1 = 2-ethylaniline, 2 = iso-propylaniline, 3 = 4-tert-butylaniline, 4 = phenylpropylamine, 5 = pindolol, 6 = metoprolol, 7 = imipramine, 8 = laudanosine.
11.3 Electroosmotic Flow (EOF)
The increases in migration time observed in these experiments are the result of an increasing electroosmotic vector counter to the larger electrophoretic vector. This is illustrated in Figure 11.3, where the EOF is plotted against the concentration of alkylammonium in the electrolyte. These changes undoubtedly come about from equilibration of the amine additive between the electrolyte solution and the capillary wall. A higher concentration of an additive or a higher molecular weight imparts a greater positive charge to the wall and therefore a larger opposing EOF. This equilibration technique offers a myriad of possibilities for creating a desired EOF because of the large variety of amines that are available.
273
274
11 Capillary Electrophoresis
Figure 11.3 Effect of protonated tributylamine (upper plot) and protonated triethylamine (lower plot) on EOF. Electrolyte also contains 50 mM tris-ethanesulfonic acid (S. A. Steiner and J. S. Fritz, 2006).
11.4 Separation of Ions 11.4.1 Separation of Anions
Although both anions and cations can be separated, much of the developmental work in capillary electrophoresis has been directed toward the separation of anions. Inorganic and small organic anions are most often separated using a negative power supply because anions undergo electrophoretic migration toward the positive electrode at the detector end of the capillary. An alkaline buffer of approximately pH 8.5 is often selected so that anions of weak acids will be in the ionic rather than the molecular form. However, ionization of silica silanol groups is extensive at this pH, and electrophoretic migration of the anions is opposed by a strong EOF vector in the opposite direction. For a reasonably fast separation, the direction of electroosmotic flow must be reversed so that electrophoretic and electroosmotic flow will both be toward the anode. This is usually accomplished by adding a reagent to the BGE that will thinly coat the capillary surface, giving it a positive charge. Hydrated electrolyte anions will now move though the detector toward the anode, thus providing an EOF in the desired direction.
11.4 Separation of Ions
A typical flow modifier is a surfactant with a hydrocarbon chain of 14 or 16 carbon atoms and a quaternary ammonium group at one end of the chain. A 1–5 mM concentration of this surfactant in the electrolyte coats the capillary wall with surfactant micelles. These give the wall a positive charge and result in a strong EOF in the same direction as the electrophoretic migration of the analyte anions. UV–Vis spectroscopy is the most common detector technique used in CE. Some analyte anions absorb in the UV spectral region and can be detected by direct spectrophotometry. However, in many cases, indirect spectrophotometric detection is indicated. For indirect detection, a low concentration of an anion that absorbs strongly in the visible or UV regions is added to the BGE. This is called the ‘probe’ ion or sometimes the ‘visualization reagent’. A background signal is established by the probe ion passing through the detector at a fixed rate. Within a sample ion zone the concentration of the probe ion is reduced by an amount proportional to the sample ion concentration, thus resulting in a detection peak of reduced absorbance. The reason that the visualization ion concentration is lower within a sample zone is that the total ionic current in the capillary must remain constant. A requirement of an anion used for indirect detection is that its mobility must match that of the sample ions as closely as possible. If the mobilities do not match reasonably well, peaks may be fronted or tailed. Since the mobilities of sample ions will differ, the mobility of the visualization reagent is matched as closely as possible to the mobility of the middle sample ions. Chromate at a concentration of around 5 mM has been used very successfully for indirect detection of common inorganic anions at a wavelength of 254 nm. A now classic separation of some 30 anions in a single run is illustrated in Figure 11.4. Shamsi and Danielson proposed naphthalenedisulfonate (NDS) and naphthalenetrisulfonate (NTS) for indirect detection of inorganic and organic anions [5]. There is a good match in their migration times relative to those of common anions. Even though these are large anions, the migration times of NDS and NTS are short because of their 2– and 3– charges. Comparison of the electrophoretic mobilities of several probe anions and analyte anions (Figure 11.5) shows that chromate is a reasonably good match for fastmoving inorganic anions but is not so good for the other analytes listed. Lau showed convincingly that molybdate is a much better probe ion than chromate for indirect detection of common anions [4]. Sensitivity is better. With molybdate, molar absorptivity is 5650 at 230 nm compared to 3180 at 254 nm for chromate, molybdate solutions are more stable, and peak shapes are better. Systematic studies resulted in the following optimal conditions: 5 mM molybdate as the UVabsorbing ion, 0.15 mM cetyltrimethylammonium hydroxide (CTAH) as an electroosmotic flow modifier, 0.01% polyvinylalcohol as an additive to solve the comigration problem of fluoride and formate, and 5 mM tris(hydroxymethyl)aminomethane as a buffer to maintain a pH of 7.9. A separation of a standard anion mixture is shown in Figure 11.6.
275
276
11 Capillary Electrophoresis
Figure 11.4 Peak identity and concentrations (ppm) for a 30-anion electropherogram displayed in an 89-s electropherographic segment. Electromigration injection at 1 kV for 15 s. Peaks: 1 = thiosulfate (4); 2 = bromide (4); 3 = chloride (2); 4 = sulfate (4); 5 = nitrite (4); 6 = nitrate (4); 7 = molybdate (10); 8 = azide (4); 9 = tungstate (10); 10 = monofluorophosphate (4); 11 = chlorate (4); 12 = citrate (2); 13 = fluoride (1); 14 = formate (2); 15 = phosphate (4); 16 = phosphite (4);
17 = chlorite (4); 18 = galactarate (5); 19 = carbonate (4); 20 = acetate (4); 21 = ethanesulfonate (4); 22 = propionate (5); 23 = propanesulfonate (4); 24 = butyrate (5); 25 = butanesulfonate (4); 26 = valerate (5); 27 = benzoate (4); 28 = l-glutamate (5); 29 = pentanesulfonate (4); 30 = d-gluconate (5). The electrolyte is a 5 mM chromate and 0.5 mM EOF modifier adjusted to pH 8.0 (From W. R. Jones and P. Jandik, J. Chromatogr., 546, 445, 1991, with permission).
Figure 11.5 The electrophoretic mobilities of common UV-absorbing anions and analyte anions (from Ref. [3]).
11.4.1.1 Separation of Isotopes Chromatographic separation of isotopes is a very challenging analytical problem. Separation factors of isotopes are usually extremely small. McDonald and Lucy [6] demonstrated the extraordinary separation power of CE by obtaining a baseline
11.4 Separation of Ions
Figure 11.6 Separation of a standard anion mixture. BGE : 5 mM molybdate, 0.15 mM CTAH, 5 mM Tris buffer at pH = 7.9, 0.01% PVA; Capillary: 65 cm × 0.075 mm i.d. fused silica; Run: –20 kV; current: 12 lA; Injection:
8 cm for 20 s; Detection: 230 nm. Anions, 2 ppm each: 1 = chloride; 2 = sulfate; 3 = nitrate; 4 = fluoride; 5 = formate; 6 = phosphate; 7 = carbonate; 8 = acetate (from Ref. [4]).
resolution of two chloride isotopes, 35Cl– and 37Cl–. Counter-migration was used and conditions were adjusted so that the electroosmotic and electrophoretic flow vectors were very similar in magnitude but opposite in direction. Under these conditions the net migration velocity of the chloride ions is very slow, but the separation power is very great. The elegant isotope separation of Lucy and McDonald was successfully repeated somewhat later in the author’s (JF) laboratory. The following conditions were employed for the separation in Figure 11.7: –20 kV, pH 9.2 using a very dilute buffer containing 3 mM borate and 6 mM chromate for indirect spectrophotometric detection
Figure 11.7 Separation of chloride isotopes by CE using counter-migration (courtesy of Youchun Shi).
277
278
11 Capillary Electrophoresis
11.4.2 Separation of Cations
Free metal cations are best separated in acidic solution in order to avoid problems due to hydrolysis. A positive power supply is generally used so that electrophoretic migration will be toward the cathodic detector end of the system. The EOF vector is in the same direction as electrophoretic migration, although an acidic pH reduces the magnitude of the EOF somewhat. This co-migration condition often enables separations to be made in as little as 2 min. Electrophoretic migration of an ion is usually proportional to conductance; hence the feasibility of separating metal cations is readily predictable from a table of limiting equivalent conductances. Table 11.3 predicts an easy separation of Li+, Na+ and K+, but the similarity in the conductances in NH4+ and K+ makes their separation difficult to impossible. Magnesium(II) and the alkaline earths are not readily separated as the free ions except for Ca2+ and Sr2+. The divalent metal ions listed cannot generally be separated as the free ions except for Pb++, which has a higher conductance. The lanthanides all have very similar conductances and cannot be separated as the free ions.
Table 11.3 Limiting equivalent conductances k (S cm2 equiv–1)
of selected metal cations. Ion
k
Ion
k
Ion
k
Li+
39
Mg2+
53
Fe2+
54
60
2+
53
2+
Na
+
50
Ca
2+
Co
73
2+
Sr
59
Ni
54
K+
74
Ba2+
64
Cu2+
55
Rb+
78
Zn2+
53
77
2+
71
NH4
+
Cs
+
Pb
Indirect detection is commonly used for metal cations because these cations lack the high UV or visible absorptivity needed for direct photometric detection. Waters Associates introduced UV Cat 1 for indirect detection. This is an amine cation that absorbs strongly in the UV spectral region and has an electrophoretic mobility similar to 1+ and 2+ metal cations. Protonated phenylethylamine or 4-methylbenzylamine are suitable visualization reagents for indirect detection of metal cations at moderately acidic pH values. Lithium, sodium and potassium ions are readily separated by CE using indirect UV detection. However, NH4+ and K+ have almost identical conductances and their peaks are not resolved at pH 6.15 (Figure 11.8a). At pH 8.5 the acid-base
11.4 Separation of Ions
Figure 11.8 Effect of pH on separation of NH4+ ions from K+ ions. 1 = potassium, 2 = ammonium, 3 = sodium, 4 = lithium (from Ref. [8] with permission).
equilibrium has shifted so that a portion of the NH4+ has been converted to NH3. The average charge is now <1 and the ammonium peak has a longer migration time (Figure 11.8b). 11.4.3 Separations at Low pH
CE separations of inorganic anions are usually carried out at an alkaline pH to ensure that the analytes will be in the ionic rather than the molecular form. A flow modifier must generally be used to reverse the direction of EOF. However, it is advantageous to separate free metal cations, as well as complex metal anions such as AuCl4– and PtCl6–, at as acidic a pH as possible in order to eliminate multiple species due to hydrolysis. Thornton and Fritz [6, 7] found that metal cations and anionic metal complexes can be separated with a carrier electrolyte of either hydrochloric acid or perchloric acid at pH 2.0–2.4. A somewhat lower applied potential (+ or –10 kV) was used to reduce the high current obtained at 20 kV. An electrolyte containing around 10 mM of a strong acid has several advantages over more conventional capillary electrolytes: (i) Silica silanol groups are not ionized at the acidic pH used and EOF is therefore very low. (ii) No additional buffer is required because HCl or HClO4 acts as a kind of concentration buffer and resists pH changes. (iii) Metal cations resist hydrolysis in acidic solution, and metal chloro anions are stabilized by an HCl electrolyte.
279
280
11 Capillary Electrophoresis
Conditions for separation at pH 2.3 with direct UV detection were established 2+ and VO2+ [6]. for several cations including Cr3+, Cu2+, Fe3+, UO2 2 , VO Speciation of vanadium(IV) ions was achieved. In acidic solution, vanadium(V) exists as the vanadyl ion (VO2+), and vanadium(V) is present as VO2+ as well as the vanadate anion (VO3–). Cationic peaks were obtained. Linear calibration plots of vanadium(IV) and vanadium(V) were achieved (peak areas vs. the concentration of V(IV) and V(V). It is noteworthy that an essentially linear calibration could be obtained for the vanadium(V) cation because an anionic peak was also observed for vanadium(V) in acidic solution when the polarity was reversed. The following equilibrium is stated to exist: VO3– + 2H+ > VO2+ + H2O
(11.9)
Gold(III) and the platinum group of metals form stable chloro complexes. All of these absorb strongly in the UV region so that direct detection is possible. Excellent separation of a mixture of AuCl4–, platinum II (PtCl42–) and two platinum IV anions (PtCl62– and PtCl5 (H2O)–) was obtained in HCl–Cl– at pH 2.0–2.4 [7]. This CE technique is an excellent way to follow the course of slow hydrolytic reactions in which one or more of the chloride ligands is replaced by water. Figure 11.9 shows the separation of rhodium III by CE in HCl solution. The major peak for a freshly prepared rhodium(III) sample (Figure 11.9A) is RhCl5(H2O)2– with a very small later peak of RhCl4(H2O)2–. After standing for 24 h, extensive hydrolysis of the same sample had taken place and the electropherogram shown in Figure 11.9B was obtained. Almost all of the initial peak had been converted into RhCl4(H2O)2– and RhCl3OH(H2O)2–. 11.4.4 Capillary Electrophoresis at High Salt Concentration
It is commonly thought that even a moderately high ionic concentration in the BGE would lead to Joule heating and serious peak distortion. However, Ding, Thornton and Fritz found that very satisfactory separations of both inorganic and organic anions could be obtained in solutions as concentrated as 5 M sodium chloride using direct photometric detection [8]. The first experiments on the effect of high salt concentrations were run at pH 2.4 to almost eliminate EOF. A negative power supply (–10 kV) and a 75 lm i.d. fused silica capillary were used. Both the sample and the BGE contained 0.5 M sodium chloride. The results for several inorganic anions under these conditions with direct photometric detection were poor. The peaks were badly shaped and there was almost no resolution of individual peaks. However, peak shape and resolution improved dramatically with increasing sodium chloride concentration in the BGE. At 1.5 M sodium chloride, excellent separation was obtained for samples containing 0.5 M sodium chloride and low ppm concentrations of five inorganic anions. The salt content of the BGE needed to be at least three times that of the
11.4 Separation of Ions
Figure 11.9 Electropherograms of rhodium (III). Fresh solution in HCl (A), and same solution after 24 h (B). Peaks: (1) RhCl5(H2O)2–, (2) RhCl4(H2O)2–, (3) RhCl3OH(H2O)2– (from Ref. [7] with permission).
sample in order to provide sufficient peak focusing (electrostacking) during the sample introduction. These experiments demonstrated that the limits of salt concentration of both the sample and the BGE are much higher than had been expected. A short study was conducted on the pH effect over a broad range, from pH 3.0 to pH 12.0. The sodium chloride concentration was 1.5 M in all buffers. There were no observed differences in either migration times or peak shapes for I–, SCN–, NO3– and IO3– between pH 3.0, 7.0, and 12.0. This effect strongly indicates
281
282
11 Capillary Electrophoresis
that the electroosmotic flow is greatly suppressed, and the ionized silanol groups at the capillary surface are effectively shielded by the high concentration of cations, M+, in the buffer solution [8]. Sample solutions containing low concentrations of several inorganic anions were run at pH 8.5 with increasing concentrations of sodium chloride or lithium sulfate in the BGE. The plots in Figure 11.10 show several interesting effects. One is that the current increases rapidly with increasing salt concentration and levels out at 280 lA around 200 mM sodium chloride or lithium sulfate. This sharp
Figure 11.10 Plots of electrophoretic mobility and current against concentration of salts (A) sodium chloride and (B) lithium sulfate. Ions (top to bottom): Br–, CrO42–, NO2–, NO3–, MnO4–, CrO33–, ReO4– (from Ref. [6] with permission).
11.5 Capillary Electrophoretic Ion Chromatography
increase in current can be attributed to less electrical resistance. The maximum current that can be obtained in our instrument is set at 280 lA. In order to maintain this current, the voltage was automatically lowered as the salt concentration in the BGE continued to increase. The full power of the instrument’s power supply was then being used. The electrophoretic mobilities of the sample anions increased and the current also increased as the amount of sodium chloride or lithium sulfate increased from 0 to 200 mM (Figure 11.10). Actually, a decrease in electrophoretic mobility is predicted with increasing salt concentration. The initial increases can be explained by Joule heating. Under the conditions used a temperature of 49°C was calculated for the capillary at high salt concentrations [7]. The greatest differences between electrophoretic mobilities and electroosmotic mobility occur around 200 mM salt in the BGE. Figure 11.11 shows an excellent separation of inorganic anions at pH 8.5 in 220 mM sodium chloride. The high salt concentration suppresses the EOF to the extent that no flow modifier is required, even at pH 8.5.
Figure 11.11 CE separation of ten inorganic anions. The 200 mM NaCl was added in the carrier electrolyte. Peaks: 1 = Br– (10 ppm); 2 = NO2– (20 ppm); 3 = S2O3– (80 ppm);
4 = NO3– (2 ppm); 5 = N3– (40 ppm); 6 = Fe(CN)64–; 7 = MnO42– (40 ppm); 8 = WO42– (40 ppm); 9 = CrO33– (40 ppm); 10 = ReO4– (40 ppm) (from Ref. [6] with permission).
11.5 Capillary Electrophoretic Ion Chromatography
Separations in ion chromatography, and indeed in column liquid chromatography in general, are based on differences in partitioning of analytes between a liquid mobile phase and a solid stationary phase. Resistance to this mass transfer be-
283
284
11 Capillary Electrophoresis
tween the two phases is a major source of peak broadening. Peaks are much sharper in capillary electrophoresis because the separation process does not require any mass transfer between two phases. Separations are based on differences in electrophoretic mobility of the analytes. But CE has the serious limitation that analyte ions often have similar electrophoretic mobilities and therefore are difficult to separate. Several techniques are described in this section that can be used to introduce a chromatographic component into capillary electrophoresis. 11.5.1 Micellar Electrokinetic Chromatography (MEKC)
Capillary electrophoretic separation is based on differences in the electrophoretic mobilities of the analyte ions. Neutral organic analytes have no electrophoretic mobility and are not separated by CE. However, Terabe et al. described a new method called Micellar Electrokinetic Chromatography, or MEKC, in which an excellent separation of a complex mixture of uncharged phenols was obtained [9, 10]. The separation was accomplished by adding an anionic surfactant (sodium dodecylsulfate, or SDS) to the carrier electrolyte. At the concentration used (1.0 mM), spherical aggregates of SDS, called a micelle, are formed in which the hydrophobic tails constitute the inner region of the micelle and the ionic head groups are pointed outward. The micelles undergo electrophoretic migration because they have a charge. Neutral analytes partition into the micellar pseudo phase. Separation of the neutrals is based on differences in their partition coefficients. Terabe noted that electrokinetic chromatography is an electrophoretic separation technique based on the separation principle of chromatography. ‘The ionic micelle, which corresponds to the stationary phase in conventional liquid chromatography, incorporates the analyte and migrates with a different velocity from the surrounding medium by electrophoresis. Accordingly, differential partition and differential migration constitute the separation principle of EKC, similar to that of chromatography.' A low concentration of an ionic surfactant in the capillary electrolyte can also be used to modify the migration behavior of sample ions. A cationic surfactant such as cetyltrimethylammonium chloride (CTAC) is often used for anion separations, and SDS is often used when cations are to be separated. Kenata et al. [11] found that CTAC decreased the effective electrophoretic mobilities of several inorganic anions as the result of two processes: 1. Ion association equilibration with monomeric surfactant occurs at concentrations below the critical micelle concentration (CMC). 2. Partitioning into the micelle at concentrations occurs above the CMC.
11.5 Capillary Electrophoretic Ion Chromatography
A plot of the reciprocal of the effective electrophoretic mobility for iodide against surfactant concentration gave two linear segments. The slope of the segment above the CMC was greater because of partitioning of iodide into the micelle. 11.5.2 Partial Complexation
Our ability to separate free metal cations by CE is limited because many of the metal ions have similar electrophoretic mobilities. An excellent way to enhance the separation of metal ions is to add a relatively weak complexing ligand (L–) such as tartrate, lactate or a-hydroxyiso-butyric acid (HIBA) to the BGE. Now part of each metal ion will remain as the free ion (M2+, for example) and part will be converted to a complexed form (ML–, ML2, ML3–, for example). The total mobility (l) will be the sum of the products of the mole fractions of each species (a) multiplied by their respective mobilities. l = aMlM + aMLlML + aML2lML2 + ...+ leo
(11.10)
leo is the electroosmotic mobility. The free metal ion will make the greatest contribution to total mobility because of its higher positive charge. Different elements will in general be complexed to different degrees so that their net mobilities will vary even though the mobilities of uncomplexed cations may be almost the same. Jones et al. obtained excellent separation of 15 alkali, alkaline earth, and divalent transition metal ions with 6.5 mM HIBA at pH 4.4 to partially complex some of the cations [12]. A protonated amine cation containing a benzene ring (Waters UV Cat 1) was used for indirect UV detection. All of the 13 lanthanides were separated using HIBA under similar conditions (Figure 11.12). Lactate has the same a-hydroxycarboxylate complexing group as tartrate and HIBA, but it is a smaller molecule and forms somewhat weaker complexes than tartrate with most metal ions. Shi and Fritz [13] found that a lactate system gave excellent separations for divalent metal ions and for trivalent lanthanides. A brief optimization was first carried out to establish the best concentrations of lactate and UV probe ion and the best pH. Excellent separations were obtained for all thirteen lanthanides, alkali metal ions, magnesium and the alkaline earths, and several divalent transition metal ions. All of these except copper(II) eluted before the lanthanides. An excellent separation of 27 metal ions was obtained in a single run that required only 6 min (Figure 11.13). Lactic acid makes possible the separation of metal ions with almost identical mobilities by complexing the individual metal ions to varying degrees. However, NH4 andK cations also have virtually identical mobilities and are not complexed by lactic acid. Potassium ions can be separated by CZE if a suitable crown ether is incorporated into the electrolyte [15-18]. The K+ ion is selectivity complexed and its mobility is reduced just enough to permit a good separation.
285
286
11 Capillary Electrophoresis
Figure 11.12 Separation of 13 lanthanides using HIBA. Electrolyte: 4 mM HIBA, 5 mM UV Cat 1, pH 4.3. Applied voltage: 30 kV. Peaks: 1 = La3+; 2 = Ce3+; 3 = Pr3+; 4 = Nd3+; 5 = Sm3+; 6 = Gd3+; 7 = Tb3+; 8 = Dy3+; 9 = Ho3+; 10 = Er3+; 11 = Tm3+; 12 = Yb3+; 13 = Lu3+ (from Ref. [14] with permission).
An electrolyte solution containing both lactic acid (11 mM) and a crown ether (2.6 mM 18-crown-6) will permit an excellent electrophoretic separation of 16 metal ions, including [14] NH4 and K (See Figure 11.14). The electrolyte also contained 7.5 mM 4-methylbenzylamine as an indirect detection co-ion. Separations in which 12-crown-4 or 15-crown-5 was used in place of 18-crown-6 failed to separate NH4 and K . Incorporation of 18-crown-6 into the electrolyte containing lactic acid affects the migration of several metal ions other than K and NH4 . The crown ether increases the migration time of Sr2+ by 15%, Pb2+ by 18% and Ba2+ by 35%, apparently by complexation to form a bulkier, less mobile species.
11.5 Capillary Electrophoretic Ion Chromatography
Figure 11.13 Separation of 27 alkali, alkaline earth, transition, and rare earth metal ions in a single run with lactate. Electrolyte: 15 mM lactic acid, 8 mM 4-methylbenzylamine, 5% methanol, pH 4.25. Applied voltage: 30 kV. Peaks: 1 = K+; 2 = Ba3+; 3 = Sr3+; 4 = Na+; 5 = Ca2+; 6 = Mg2+; 7 = Mn2+; 8 = Cd2+; 9 = Li+;
10 = Co2+; 11 = Pb2+; 12 = Ni2+; 13 = Zn2+; 14 = La3+; 15 = Ho3+; 16 = Pr3+; 17 = Nd3+; 18 = Sm3+; 19 = Gd3+; 20 = Cu2+; 21 = Tb3+; 22 = Dy3+; 23 = Ho3+, 24 = Er3+; 25 = Tm3+; 26 = Yb3+; 27 = Lu3+ (from Ref [14] with permission).
11.5.3 Effect of Ionic Polymers
An easy way to add a chromatographic component to the separation of ions by CE is simply to add a water-soluble polymer with ionized sites to the BGE. As an example, migration of a sample anion through the capillary is slowed by interaction with a soluble polymer containing positive quaternary ammonium sites. Terabe and Isemura [9, 10] described a method of this type which they called ‘ionexchange EKC’. Of several polymers tried, poly(diallyldimethylammonium chloride), abbreviated as PDDAC, proved to be the most effective. The usual electrophoretic migration was modified by the analyte anions forming ion-pairs of varying stability with the ionic polymer. These authors developed the basic theory of this technique and demonstrated its effectiveness in separating various organic
287
288
11 Capillary Electrophoresis
Figure 11.14 Separation of 16 common metal ions and ammonium. Electrolyte, 11 mM lactic acid, 2.6 mM 18-crown-6, 7.5 mM 4-methylbenzylamine, 8% methanol, pH 4.3; applied voltage, 30 kV; injection time, 30 s, Peaks: 1 = NH4+ (5 ppm); 2 = K+ (5 ppm); 3 = Na (3 ppm); 4 = Ca2+ (3 ppm); 5 = Sr2+ (5 ppm);
6 = Mg2+ (1.5 ppm); 7 = Mn2+ (3.2 ppm); 8 = Ba2+ (5 ppm); 9 = Cd2+ (4 ppm); 10 = Fe2+ (3.2 ppm); 11 = Li+ (0.8 ppm); 12 = Co2+ (3.2 ppm); 13 = Ni2+ (3.2 ppm); 14 = Zn2+ (3.2 ppm) 15 = Pb2+ (5 ppm); 16 = Cu2+ (4 ppm).
anions. Cassidy and coworkers [15] applied a similar technique for the separation of several inorganic anions using indirect photometric detection. A comprehensive paper by Li, Ding and Fritz [16] used PDDAC as the soluble polymer in a BGE containing a relatively high concentration of sodium chloride or lithium sulfate for the separation of inorganic and organic anions. Anionexchange equilibrium was proposed, rather than a mechanism that involved only ion-pair formation. The ion-exchange equilibrium between a sample anion (A–) and the polymer + – (P Cl ) is given by the following equation: P+Cl– + A– > P+A– + Cl–
(11.11)
for which the equilibrium constant (K) is: K
P A Cl A P Cl
(11:12
11.5 Capillary Electrophoretic Ion Chromatography
At a fixed concentration of P+Cl–, a conditional constant, K′, may be written as follows: K′ = K [P+Cl–]
(11.13)
Combining Eqs.11.12 and 11.13, and rearranging gives: A Cl K′ P A
(11:14
The electrophoretic migration rate will depend primarily on the fraction of sample anion that is present as the free anion. This is true because the free anion (A–) will migrate rapidly toward the anode, while the fraction associated with the ion exchanger (P+A–) will move only very slowly in the opposite direction. The fraction present as A– will depend both on the total anion concentration ([Cl–] in Eq 11.11) and on the value of K′, which will be different for each sample anion. Incorporation of PDDAC into the capillary electrolyte also reverses the direction of EOF. Addition of 0.05% to 0.30% PDDAC to the BGE sets up a dynamic equilibrium in which PDDAC forms a thin coating on the inner walls of the capillary. This imparts a negative charge to the surface and sets up an electroosmotic flow toward the anode which is in the opposite direction to the usual cathodic EOF in uncoated capillaries. Under typical conditions the EOF in capillaries equilibrated with PDDAC was almost constant over a wide pH range (Table 11.4). A negative power supply (–10 kV) is used for anion separations. The BGE typically contains 0.05% to 0.30% PDDAC, up to 150 mM sodium chloride or lithium sulfate, and a 20 mM borate buffer. The EOF vector and the electrophoretic vectors of the sample anions are both in the anodic direction. The fraction of each sample anion that is attached to exchange sites in the PDDAC has a cathodic electrophoretic vector, although this is believed to be weak. The net electrophoretic vector for any given anion depends primarily on the fraction that exists as the free ion [Eq. (11.14)] and on the electrophoretic mobility of the free ion. Separations of common inorganic anions are quite fast, as indicated by the migration times in Table 11.4. It is also possible to obtain a baseline separation because of the greater ion-exchange affinity for iodide; the electrophoretic mobilities of bromide and iodide are almost identical. A separation of several inorganic anions is illustrated in Figure 11.15. The effect of experimental variables on the separation of anions can be summarized: 1. Increasing the BGE concentration in steps from 0.1 to 1.0% (approximately 6 to 60 mN) decreases the electrophoretic mobilities of sample anions to varying degrees. This results in longer migration times. 2. Increasing the concentration of LiSO4 in the BGE from 50 to 150 mM increases the electrophoretic mobilities of sample anions, probably by reducing the interaction between the anions and PDDAC [see (Eq. 11.11)].
289
290
11 Capillary Electrophoresis Table 11.4 Comparison of EOF and migration times at different
pH values for inorganic anions. Capillary: 40 cm × 50 mm; electrolyte: 150 mM Li2SO4, 0.05% PADDC, 20 mM borate for pH 8.5, or 20 mM sodium acetate for pH 5.0, or 20 mM HCl for pH 2.3; separation voltage: –10 kV; hydrostatic sampling: 40 s at 10 cm height; detection: UV: 214 nm; EOF marker: D.I. H2O. Migration time (min) Br– I–
PH
EOF (cm2 V–1 s–1)
2.3
–2.46 × 10–4
1.98
5.0
–2.65 × 10–4
8.5
–4
–2.74 × 10
NO3–
SCN–
2.06
2.17
2.34
1.95
2.03
2.13
2.30
1.93
2.00
2.10
2.26
Figure 11.15 Separation of organic anions. Capillary: 40 cm × 50 lm; electrolyte: 150 mM Li2SO4, 20 mM borate, 0.3% PADDC, pH 8.5; separation voltage: –10 kV; hydrostatic sampling: 40 s at 10 cm height; detection: UV, 214 nm. Peaks: 1 = benzenesulfonic acid;
2 = benzoic acid; 3 = p-toluenesulfonic acid; 4 = p-hydroxybenzoic acid; 5 = p-aminobenzoic acid; 6 = 2-naphthalenesulfonic acid; 7 = 1-naphthalenesulfonic acid; 8 = 3,5-dihydroxybenzoic acid; 9 = 2,4-dihydroxybenzoic acid (from Ref.[17] with permission).
The type of salt, as well as its concentration, can have a major effect on the migration of sample anions. In ion chromatography, sulfate is known to have a much stronger affinity for a solid quaternary ammonium anion exchanger than acetate, for example. Thus, acetate will have a much smaller inhibiting effect on the ion exchange of sample anions with PDDAC than the same concentration of sulfate. The migration times of bromide and iodide in 150 mM lithium sulfate are 5.74
11.5 Capillary Electrophoretic Ion Chromatography
and 6.88 min, respectively (a = 120). In 150 mM sodium acetate the migration times are 6.08 min for bromide and 8.77 min for iodide (a = 1.44). 11.5.4 Effect of Alkylammonium Salts
Alkylammonium cations, as well as quaternary ammonium polymers, may be used to modify the electrophoretic migration of anions. Steiner, Watson and Fritz [17] demonstrated that a 100-mM concentration of a quaternary ammonium salt provides an excellent medium for separation of anions. The electrophoretic migration of sample anions is slowed by ion association (or ion exchange) with the organic cation. The extent of anion–cation interact ion varies with the bulk and hydrophobicity of the cations as well as the cation concentration. This effect is illustrated by the electropherograms in Figure 11.16 where the BGE contained 100 mM R4N+Cl– and 5 mM buffer. The anion migration times increase in the order R = ethyl < propyl < butyl. Although excellent separations of inorganic anions and organic sulfonates were obtained, an opposing electroosmotic flow at alkaline pH values necessitated the use of an acidic pH for many of the separations. This drawback was avoided in later work [18] by precoating the capillary with PDDAC. Coating is accomplished simply by passing an aqueous solution of the soluble polymer through the capillary. The capillary acquires a semi-permanent layer of the polymer that gives a strong EOF in the desired direction (anodic) at any pH. Figure 11.17 shows the separation of eleven inorganic and organic anions with a PDDAC-coated capillary and a 100-mM aqueous solution of tetrabutylammonium acetate at pH 6.0. A negative power supply was used with co-migration of the EOF and electrophoretic vectors. In these separations, use of a moderately high concentration of an alkylammonium salt in the electrolyte again adds a chromatographic component to CE. Sample anions migrating electrophoretically toward the detector are slowed by transient association with alkylammonium cations moving in the opposite direction. This association is considered to result from a combination of ion–ion and dispersive interactions. The following equilibrium is established: R4N+ + A– > R4N+A–
(11.12)
with the association constant, KAS
R4 N A R4 N A
(11:13
This can be viewed as a dynamic equilibrium in which the anion spends a certain time fraction migrating as the free anion and another time fraction as the association complex, during which migration is much slower. The magnitude of this effect is a function of the type and concentration of the electrolyte cation, the counter-anion and of the properties of the analyte anion.
291
292
11 Capillary Electrophoresis
Figure 11.16 Effect of the alkylammonium salt on the separation of anions at pH 3. BGE contains 100 mM R4NCl + 5 mM buffer. Applied voltage, –15 kV. Column length, 50 cm, 42 cm to detector. Peak identification: 1 = bromide, 2 = iodide, 3 = thiocyanate, 4 = p-toluene-
sulfonate, 5 = 1-napththalenesulfonate, 6 = 2-naphthalenesulfonate. (A) tetraethylammonium chloride; (B) tetrapropylammonium chloride; (C) tetrabutylammonium chloride (from Ref [17] with permission).
11.5 Capillary Electrophoretic Ion Chromatography
Figure 11.17 Separation of inorganic and organic anions. Conditions: 100 mM tetrabutylammonium acetate electrolyte at pH 5.5; –15 kV. Peaks in order of increasing tm: nitrite, nitrate, iodide, 1,2-benzenedisulfonate,
thiocyanate, mandelate, p-toluenesulfonate, 1,2-naphthoquinone-4-sulfonate, salicylate, 2-naphthalenesulfonate, 1-naphthalenesulfonate (from Ref. [18]).
The association constants, KAS, were measured by the method of Keneta et al. [11]. The reciprocal of measured, or effective, electrophoretic mobility of the analyte anion (lef ) is plotted against the concentration (C) of R4N+ in the capillary electrolyte. 1 CK AS 1 lef lA lA
(11:14
The intercept of a linear plot gives the value of 1/lA, where lA is the electrophoretic mobility of the free, uncomplexed analyte anion. Knowing this, the value of the association constant, KAS, can be calculated. Values of KAS for analyte anions in several systems are presented in Table 11.5. As might be expected, KAS is larger for tetrabutyl than for tetraethyl quaternary ammonium salts. Examination of the values for tetraethylammonium salts, which generally gave r2 values of 0.99 or better, show that KAS for sulfate was lower than for acetate. This is an indication that the process we are observing involves ion exchange in solution and not merely ion association where the anion would play no role. R4N+E– + A– > R4N+A– + E–
(11.15)
In this equation, E– is the electrolyte anion and A– is the analyte anion. Of course, the concentration of E– is several orders of magnitude higher than that of A–.
293
294
11 Capillary Electrophoresis Table 11.5 Association constants, KAS , for several anions, based
on electrophoretic mobility measurements at 0.025, 0.05 and 0.10 M concentrations of quaternary ammonium salts. The values given are estimated to be ± 2–3%. KAS Anion
Et4Nac
Et4NSO4
Bu4Nac
BDSA
1.01
–
1.46
1.06
TSA
1.49
1.34
4.94
3.69
1-NSA
3.65
2.55
11.4
Bu4NSO4
–
11.5.4.1 Separation Mechanism In the experimental system, the capillary is filled with electrolyte with each end open to a separate reservoir of electrolyte. When a voltage is applied, an electric current is established in which electrolyte anions migrate toward the positive electrode (anode) and electrolyte cations move in the opposite direction. After sample injection, sample anions are separated by differences in their electrophoretic migration rates to the detector at the anodic end of the capillary. Although the sample anions migrate against a counter flow of electrolyte cations, interaction between the cations and anions is minimal in conventional CE where the background electrolyte typically contains a fairly low concentration of sodium or ammonium cations. The situation changes drastically when a higher concentration (about 100 mM) of a larger organic cation is used in the electrolyte. Conditions are now favorable for cation–anion interactions as the sample anions migrate through a veritable sea of the larger alkylammonium cations. These interactions slow the electrophoretic migration rates of sample anions to varying degrees, thus producing a strong chromatographic component to the separation. These interactions may be said to occur on a nano scale because the effective radii of the organic cations are of the order 1 nm or less. It is possible that the organic cations may undergo some attraction for one another and form a series of nano domains. Actual micelle formation can be ruled out by the relatively low molecular weight as well as by geometrical considerations for the quaternary ammonium salts used.
11.6 Summary
Capillary electrophoresis is a technique in which ions are separated by differences in their electrophoretic migration rates when a high voltage is applied. The ‘column’ is an open tubular capillary, similar to the capillary used in gas chromatography. Addition of an ionic surfactant, polymer, or a larger organic ion also modi-
11.6 Summary
fies the migration rates of sample ions through the capillary. These electrolyte additives slow the migration rates of oppositely charged analyte ions in much the same way as a solid ion exchanger does in conventional ion chromatography. The major difference is that in the capillary system, the cation–anion interactions between the electrolyte additives and analyte ions occur between the liquid carrier and a ‘pseudo phase’ when a surfactant micelle is involved. The cation–anion interactions take place within a single liquid phase when the electrolyte additive is a soluble polymer or simply a moderately large organic ion. These conditions result in much sharper sample-ion peaks than in conventional ion chromatography where mass transfer between two distinct phases is required. Terabe [10] has proposed the name ‘ion-exchange EKC’ for separations where the electrophoretic migration of analyte ions through an open-tubular capillary is slowed to varying degrees by ion exchange with soluble polymers in the capillary electrolyte. Capillary electrophoretic ion chromatography, or CE-IC, is also an appropriate name to describe separations of this kind. Although CE methods generally provide very good separations with sharp peaks, there can be some problems. Sometimes reproducibility of migration times is relatively poor and peaks are broad and unsymmetrical. These aberrations are most likely due to interactions between the analyte ions and the capillary wall. Most difficulties can be avoided by selecting experimental conditions that: (i) give a stable, reproducible interface between the capillary electrolyte and the wall, and (ii) avoid interactions between analyte ions and the wall. Reproducible migration times require a stable EOF. The magnitude of the EOF vector is influenced by the extent to which silica silanol groups are ionized to give a negative charge, and this varies considerably with pH. Thus, a stable EOF requires a thorough equilibration of the capillary surface with a pH buffer and finally with the buffered capillary electrolyte. All solutes within the capillary can equilibrate to varying degrees between the liquid and the capillary surface. But when the ionic concentration of the carrier electrolyte is much higher than that of the sample, analyte ions are kept from the capillary surface according to the law of mass action. There will be a stable equilibrium that will hardly be affected by analytical samples of different compositions. For this reason, a fairly high concentration (100–150 mM) usually gives more reproducible times and sharper peaks than the more dilute electrolytes that were formerly recommended. Suggested conditions for capillary electrophoretic separations are summarized in Table 11.6. A larger number of experimental options are available than with conventional ion chromatography. However, the separation power of CE is so great that an adequate separation can usually be obtained even with less than optimal experimental conditions.
295
296
11 Capillary Electrophoresis Table 11.6 Suggested conditions for capillary electrophoretic separations.
General Conditions Capillary Length:
About 40 – 60 mm.
Capillary i.d.:
50 lm.
Applied potential:
–15 – 20 kV for anions; +15 – 20 kV for cations.
Format:
Counter-migration with moderate opposing EOF if possible. Co-migration is faster but gives poorer peak resolution.
Detector:
Several options are available. Direct UV-vis is most common for ions with sufficient absorption. Indirect detection with an added visualization reagent may also be used.
pH:
Appropriate for sample chemistry. Acidic pH for protonated amine cations; alkaline pH for most anions.
EOF Control:
Vary pH. Use coating for capillary surface, if necessary.
Migration time:
2–10 min range if possible. Adjust by changes in pH, capillary surface, or applied potential.
Other Conditions
Effect
1. Capillary Surface a) Pretreatment: NaOH, wash, with Hydrolyzes SiOSi groups, prepares surface. BGE buffer b) Other treatment: ionic surfactant, Adjusts or reverses EOF to a desired value. ionic polymer or R4N+ 2. Background Electrolyte a) Use a higher conc. (100 mM)
Promotes electrostacking, inhibits analyte interactions with wall.
b) Organic ions, not inorganic
Gives a lower current, improves solubility of organic analytes.
c) 5–20 mM pH buffer
Maintains desired pH.
d) Consider use of an additive to give a chromatographic effect.
Improves peak resolution.
3. Sample a) More dilute than BGE
Increases electrostacking.
b) Short injection time (~ 5 s), Gives sharper peaks. unless a larger sample is needed. c) Keep total amount of analytes low Avoids sample overloading and poorly shaped peaks.
References
References [1] J. W. Jorgenson and K. D. Lukacs, Zone
[2]
[3]
[4] [5]
[6]
[7]
[8]
[9]
electrophoresis in open-tubular glass capillaries, Anal. Chem., 53, 1298, 1981. P. Jandik and G. Bonn, Capillary electrophoresis of small molecules and ions, p. 23, VCH, New York, 1993. K. K.-C. Yeung and C. A. Lucy, Improved resolution of inorganic anions in capillary electrophoresis by modification of the reversed electroosmotic flow and the anion mobility with mixed surfactants, J. Chromatogr. A, 804, 319, 1998. K. M. Lau, PhD Thesis, University of Hong Kong, December 1997. S. A. Shamsi and N. D. Danielson, Naphthalenesulfonates as electrolytes for capillary electrophoresis of inorganic anions, organic acids, and surfactants with indirect photometric detection, Anal. Chem., 66, 3757, 1994. C. A. Lucy and T. L. McDonald, Separation of chloride isotopes by capillary electrophoresis based on the isotope effect on ion mobility, Anal. Chem., 67, 1074, 1995. M. J. Thornton and J. S. Fritz, Separation of inorganic anions in acidic solution by capillary electrophoresis, J. Chromatogr. A, 770, 301, 1997. W. Ding, M. J. Thornton and J. S. Fritz, Capillary electrophoresis of anions at high salt concentrations, Electrophoresis, 19, 2133, 1998. S. Terabe, K. Otsuka, K. Ichikawa, A. Tsuchiya and T. Andg, Electrokinetic separations with micellar solutions and open-tubular capillaries, Anal. Chem., 56, 111, 1984.
[10] S. Terabe and T. Isemura, Effect of poly-
mer ion concentrations on migration velocities in ion-exchange electrokinetic chromatography, J. Chromatogr., 515, 667, 1990. [11] T. Kaneta, S. Tanaka, M. Taga and H. Yoshida, Migration behavior of inorganic anions in micellar electrokinetic capillary chromatography using a cationic surfactant, Anal. Chem., 64, 798, 1992. [12] W. R. Jones, P. Jandik and R. Pfeifer, Capillary ion analysis, an innovative technology, Am. Lab., 5, 40, 1991. [13] Y. Shi and J. S. Fritz, Separation of metal ions by capillary electrophoresis with a complexing electrolyte, J. Chromatogr., 640, 473, 1993. [14] Y. Shi and J. S. Fritz, New electrolyte systems for the determination of metal cations by capillary electrophoresis, J. Chromatogr. A, 671, 429, 1994. [15] C. Stathakis and R. M. Cassidy, Effect of electrolyte composition in the capillary electrophoretic separation of inorganic/ organic anions in the presence of cationic polymers, J. Chromator. A, 699, 353, 1995. [16] J. Li, W. Ding and J. S. Fritz, Separation of anions by ion chromatography-capillary electrophoresis, J. Chromatogr. A, 879, 245, 2000. [17] S. A. Steiner, D. M. Watson and J. S. Fritz, Ion association with alkylammonium cations for separation of anions by capillary electrophoresis, J. Chromatogr., 1085, 170, 2005. [18] S. A. Steiner and J. S. Fritz, Unpublished work, 2006.
297
299
12 DNA and RNA Chromatography 12.1 Introduction 12.1.1 Importance of DNA and RNA Chromatography
The reader may well ask what nucleic acids, deoxyribonucleic acid (DNA) and ribonucleic acid (RNA) have to do with ion chromatography. Nucleic acids are very large biological molecules that are more like linear polymers and bear little similarity to common anions such as chloride and sulfate. Nevertheless, DNA and RNA are anionic by virtue of the phosphate groups attached to each nucleotide sugar unit. Nucleic acids can be separated by chromatography, often by ion-pair methods similar to those described in Chapter 9. For these reasons, it seems appropriate to include a chapter on DNA and RNA chromatography in this book. Advances in separation technology over the past few years have shown that chromatography can perform remarkably well for separating, measuring and purifying these biological molecules. From a technical perspective, the instrumentation and many of the methods are very similar to what is used in traditional ion chromatography. In general, an ion chromatograph instrument can be used to perform DNA and RNA separations. Although the column is specifically designed for nucleic acid separations, the eluent buffers, methods of detection and so on are similar to what is used in many types of ion chromatography. However, DNA chromatography differs in several important ways from the ion chromatography of smaller ions. Specially designed columns are required to prevent strong adsorption of the bulky nucleic acids. Temperature of the eluent and sample is a major variable in DNA chromatography. Elevated column and eluent temperatures cause structure changes that can alter the elution profile of DNA fragments. Careful precautions must be taken to avoid contamination from traces of metal in the chromatographic system. While it is understood that analytical chemists are the primary audience for this 4th edition of Ion Chromatography, it is also true that the analytical chemist and ion chromatographer are being asked more and more to provide answers to biological problems. In a sense, many of the new research challenges facing the Ion Chromatography, 4th Ed. James S. Fritz and Douglas T. Gjerde Copyright © 2009 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim ISBN: 978-3-527-32052-3
300
12 DNA and RNA Chromatography
biologist are problems that are much more analytical in nature. Because biologists are not normally trained to use advanced analytical methodology, many of them seek out the analytical chemist to help them in their research. DNA provides the general ‘blueprint’ for the cells to make proteins. After the proteins are made, they go on to do the work of the cell in performing various cellular processes. RNA provides the mechanism for translating information held by DNA into these cellular proteins. Over the past several years, DNA and RNA have become of extremely high interest because of their potential to be used as drugs to treat disease. A huge amount of academic and industrial resources are now being spent on trying to understand, manipulate and use DNA and RNA. It has also been predicted that DNA-based drugs will someday cure disease with a single treatment (rather than our current treatment methods of a daily regimen of drugs) by incorporating new genes into cells. Silencing RNA has been demonstrated to be able to turn off or control specific biological functionality. Through research that is now being pursued, it has been predicted that RNA-based drugs will prove to be extremely powerful and will someday revolutionize treatment of disease. But in order to accomplish these ambitious goals, advanced analytical tools and methods will have to be used in biological research. And these advanced analytical tools are likely to include DNA and RNA chromatography. 12.1.2 Organization of this Chapter
We start this chapter by briefly describing nucleic acid composition and structure. Nucleic acids have several different forms or structures, depending on their biological function. These forms can be controlled (or preserved if desired) through chemical and temperature control. Several column types have been used to separate nucleic acids. A discussion of the historical development of DNA and RNA chromatography is followed by a brief description of the column development that has led to the modern use of ion-pairing chromatography. There are three modes of modern nucleic acid chromatography, which are mainly dependent upon the temperature of the eluent and column under which the separations are performed. A description of specific types of separations will be discussed, along with the hardware and columns needed to perform these separations. DNA and RNA chromatography require that the instrument, column and eluent be completely free of metal contamination. The effect, source, and control of metal contamination will be discussed. A high-quality oven must be used. The instrument normally includes the option of collecting the nucleic acid in a fragment collector for further research and processing. Finally, applications of DNA and RNA chromatography will be shown. The purpose of using this type of methodology is to provide the means to answer biological questions. The power of the methodology is illustrated through its various applications.
12.2 DNA and RNA Chemical Structure and Properties
The discussions in this chapter necessarily use biological terms, some of which may be unfamiliar to the analytical chemist. These terms will be defined at the start of each section.
12.2 DNA and RNA Chemical Structure and Properties
DNA and RNA are both nucleic acids. Nucleic acids are very large molecules that are made up with a backbone of alternating sugar and phosphate molecules, bonded together in a long polymeric chain. This backbone of the nucleic molecule is represented below: –
sugar –
phosphate
– sugar –
phosphate
– sugar –
phosphate
– sugar –
The polymer is of varying lengths that, depending on its biological function, range from just a few units up to tens of thousands of units. These varying sizes are called nucleic acid fragments. For DNA, the sugar molecule in the backbone is deoxyribose. For RNA, the sugar is ribose. To complete the nucleic acid molecule, nucleotide bases are bonded to each sugar molecule. This is shown below: nucleotide base
nucleotide base
nucleotide base
nucleotide base
|
|
|
|
– sugar –
phosphate
– sugar –
phosphate
– sugar –
phosphate
– sugar –
There are four different types of nucleotide bases. In DNA these are adenine (A), thymine (T), cytosine (C), and guanine (G). A DNA nucleic acid molecule is illustrated below in which the order of nucleic acids is A, T, C and G: A
T
C
G
|
|
|
|
– sugar – phosphate – sugar – phosphate – sugar – phosphate – sugar –
In 1953, Watson and Crick published a paper in the scientific journal Nature describing the structure of DNA. Watson and Crick showed that not only is the DNA molecule double-stranded, but the two strands wrap around each other forming a coil or helix, forming complementary strands. Each nucleotide in one strand is hydrogen bonded to another nucleotide base in a strand of DNA opposite to the original. This bonding is specific. Adenine always bonds to thymine (and vice versa) and guanine always bonds to cytosine (and vice versa). This bonding occurs across the molecule, leading to a double-stranded molecule as shown next:
301
302
12 DNA and RNA Chromatography – sugar – phosphate – sugar – phosphate – sugar – phosphate – sugar – |
|
|
|
T
A
G
C
|
|
|
|
A
T
C
G
|
|
|
|
– sugar – phosphate – sugar – phosphate – sugar – phosphate – sugar –
Double-stranded DNA has the unique ability to make exact copies of itself, i.e. self-replicate. When more DNA is required by a cell (such as during reproduction or cell growth), the hydrogen bonds between the nucleotide bases break and the two single strands of DNA separate. New complementary bases are brought in by the cell and paired up with each of the two separate strands, thus forming two new, identical, double-stranded DNA molecules. A similar process called polymerase chain reaction (PCR) can be performed in the laboratory under in vitro conditions to copy and amplify specific sequences of DNA. The order in which these nucleotide bases (also called the sequence of bases) appear in a particular nucleic acid molecule constitutes the blueprint for the information carried in the molecule. In a living cell, these sequences are contained in genes. The sequence of bases found in a particular gene provides the information for the synthesis of a particular protein. In the laboratory these sequences are called nucleic fragments. They may be called a fragment of x number of base pairs if the fragment is double-stranded, or a fragment of x number of bases if the fragment is single-stranded. Ribonucleic acid, or RNA, gets its name from the sugar group in the molecule’s backbone – ribose. Several important similarities and differences exist between RNA and DNA. Like DNA, RNA has a sugar-phosphate backbone with nucleotide bases attached to it. Like DNA, RNA contains the bases adenine (A), cytosine (C), and guanine (G). However, RNA does not contain thymine (T). Instead, RNA’s fourth nucleotide is the base uracil (U). Both RNA and DNA can exist in doublestranded and single-stranded forms, although RNA is normally single-stranded. RNA is the main genetic material contained in the organisms called viruses. RNA is also important for the production of proteins in other living organisms. RNA can transport within the cells of living organisms, and thus serves as a sort of genetic messenger. Information stored in the cell’s DNA is transported by RNA from the nucleus to other parts of the cell, where it is used to help make proteins.
12.3 DNA and RNA Chromatography
12.3 DNA and RNA Chromatography 12.3.1 Development of DNA and RNA Chromatography
While ion-pairing reverse-phase chromatography is the primary separation method described in this chapter, many different types of chromatographic methods have been used to separate nucleic acid. These include gel filtration, affinity and ion exchange [1–13]. In particular, ion-exchange chromatographic separations of nucleic acids performed on nonporous anion exchangers can be extremely rapid and useful [14–17]. J. A. Thompson and coworkers published many excellent papers in the form of both review articles [18–23] and research papers [24–26]. The review articles published in 1986 and 1987 are a series of six publications, each dealing with some aspect of nucleic acid separation. Taken together, the six review articles present a comprehensive description of the different chromatographic methods for nucleic acid separations. The fundamental technology leading to modern DNA chromatography and later RNA chromatography was first described by Guenther Bonn, Christian Huber and Peter Oefner in 1993 [27–29]. Using ion-pairing reverse-phase chromatography, they obtained rapid, high-resolution separations of both double-stranded and single-stranded DNA. The separations were performed usually in less than 10 min, and in many cases resolution of fragments differing in only a single base pair in length was achieved. This form of HPLC analysis is largely (though not entirely) based upon the unique separation properties of a nonporous polystyrene–divinylbenzene polymer bead that has been functionalized with C18 alkyl groups. An alkylammonium salt, usually triethyl ammonium acetate (TEAA), is added to the eluent to form neutral ion pairs when a DNA sample is introduced into the HPLC instrument. The fragments are adsorbed by the column, and then a gradient of water and acetonitrile is used to separate the DNA fragments. The smaller fragments come off the column first, and larger fragments are then eluted from the column and detected. Figure 12.1 shows a separation of double-stranded DNA obtained by this method. Baseline resolution of fragments up to 100 base pairs long can be achieved and fragments up to 2000 and larger base pairs can also be separated on the column. Bonn, Huber and Oefner showed that the DNA separations were performed according to the size of the fragment, just as they are in gel electrophoresis. Gel electrophoresis is the classical way of separating DNA and is based on using a high electrical potential to pull the DNA fragments through a gel slab sheet. Separations are based on the size of the DNA fragment, with smaller fragments traveling faster through the gel than larger fragments. For double-stranded DNA, Bonn and coworkers showed that the sequence does not contribute to the retention of the fragment on the column; retention is based only on DNA fragment size. Figure 12.2, taken from their work, demonstrates
303
304
12 DNA and RNA Chromatography
Figure 12.1 High-performance DNA chromatography separation of double-stranded DNA using a DNASep® column 50 mm × 4.6 mm. Sample was a mixture of BR322 Hae III restriction digest and UX174 Hinc II restriction
digest. Eluent and gradient: A: 0.1 M TEAA; B: 0.1 M TEAA, 25% acetonitrile, 35 to 45% B in 2 min, 45% to 57% in 10 min., 57% to 61% in 4 min, 1.0 mL min–1 flow rate, UV detection at 254 nm (from Ref. [23] with permission).
this with a plot of retention time vs. fragment size for a number of different fragments. Various plasmids (circular double-stranded DNA of known sequence) were digested with a number of different enzymes to cut the plasmid at specific, known sequences so that mixtures of DNA fragments with different, known lengths and sequences were generated. Since the base sequence of the original plasmids is known, the effect of DNA sequence can be shown. In the plot, all of the retention times of the various fragments fall on a line drawn through the data points. This shows that the retention of DNA is dependent on fragment size, and is independent of fragment sequence. It was shown that the technology could be used to separate single-stranded DNA; later, single-stranded RNA separations were developed. These separations are very rapid with base resolution possible for most short oligo mixtures. Depending on the type of eluent used, single-stranded separations are based on differences in size, polarity and shape of the molecule. By changing the ion-pairing reagent to be more nonpolar, the separation can become mostly size based. Also, discussed later, temperature is an important parameter for single-stranded separations, especially RNA separations.
12.3 DNA and RNA Chromatography
Figure 12.2 The retention times of various plasmid digest fragments are plotted according to size. The fragments have different sequences, but are separated according to their size using a DNASep® column 50 mm × 4.6 mm.
Eluent and gradient: A: 0.1 M TEAA; B: 0.1 M TEAA, 25% acetonitrile, 0% to 100% B in 30 min, 1.0 mL min–1 flow rate, UV detection at 254 nm (from Ref. [27] with permission).
12.3.2 Column Properties
A number of modern HPLC columns have been used for DNA and RNA separations. One example is a silica-based C18 material available from Varian Corporation (Walnut Creek, CA). There have been recent reports on monolith polymeric materials providing excellent separations [30–32]. The remarkable performance of the DNA and RNA separation columns is based on a number of properties including the porosity of the packing material, its polarity, the absence of metal contamination, and the small size and narrow size distribution of the packing material. Bead pores are small relative to the nucleic acid molecule so that column interactions of the nucleic acid with the bead are on the surface of the bead. Polarity of the bead is adjusted so that nucleic acid interactions with the surface are controllable with ion pairing reagents. Most column packings are about 2-lm particles functionalized with a hydrophobic, neutral C18 alkyl group. Polymer-based column materials are used because they are rugged and can withstand extremes in eluent pH and high-temperature operation. The most popular columns, DNASep® and OligoSep™ (Transgenomic, Inc. Omaha, NE), have been cited in more than 1200 journal articles (www.Transgenomic.com). They are 2-lm, C18 surface, nonporous polymeric columns, based on the original work published by Bonn et al. [27–29].
305
306
12 DNA and RNA Chromatography
12.3.3 Ion-pairing Reagent and Eluent
The ion-pairing reagent is an amine cation salt that forms a nonpolar ion pair with the phosphate anion group of the nucleic acid. Triethylammonium acetate (TEAA) is the most common amine cation salt used for this purpose. TEAA will pair with nucleic acid fragments to form a nonpolar ion pair and adsorb to the neutral nonpolar surface of the column. Acetonitrile is gradually added to the eluent, decreasing its polarity until TEAA/nucleic acid ion-pair fragments desorb from the column. In gradient elution, the concentration of acetonitrile that is pumped through the column is increased gradually as the separation proceeds. Smaller fragments elute first, followed by fragments increasing in size as the acetonitrile concentration is increased. DNA and RNA nucleic acid fragments are extremely large molecules relative to what can normally be separated by a chromatographic process. The average molecular weight of one base pair of DNA is approximately 660 g mol–1. As an example, a double-stranded 100 base pair DNA molecule has an approximate molecular weight of 66 000. As many as 1000 and even up to 2000 base pair fragments can be separated by DNA chromatography. It is remarkable that a chromatographic system can achieve base pair resolution for fragment sizes greater than 100 base pairs, with partial resolution up to 200 base pairs. It is possible to use a very rapid gradient program of perhaps 3–5 min, although such a rapid elution program is at the expense of lower resolution of peaks. In many cases, this is adequate because the mixture being separated is not complex. A slower gradient process of 20–30 min will produce the highest resolving conditions and the greatest separation of peaks although obviously this is at the expense of a longer analysis time.
Figure 12.3 Example chromatogram using the optimized siRNA reverse-phase denaturing HPLC purification protocol described here for the WAVE Oligo System (siRNA oligonucleotide: Luciferase Antisense. n-Hexylammonium
acetate ion pairing agent separates primarily based on length rather than base composition. The product purity was less than 50% but was enriched to greater than 95%.
12.4 Temperature Modes of DNA and RNA Chromatography
It was shown that TEAA produces size-based separations for double-stranded DNA. For single-stranded DNA and RNA, separations are not solely size-based when using the TEAA eluent because the polarity of the sequence fragment is not shielded by the ion pairing reagent. However, the use of a more hydrophobic reagent such as tetrabutylammonium bromide can provide a separation that is very close to size-based. Figure 12.3 shows a separation and purification of silencing RNA synthesized material separated on a Transgenomic OligoSep Prep HC cartridge using another hydrophobic reagent, n-hexylammonium acetate as the ion-pairing reagent. Since the separation is primarily size-based, synthesis failures can be removed and the product can be purified to more than 95% purity [33].
12.4 Temperature Modes of DNA and RNA Chromatography
DNA exists in both double-stranded and single-stranded forms. To a certain extent, RNA can exist in both forms, but it is more likely to be single-stranded with some double-stranded secondary structure. The temperature of the column and the fluid entering the column can be thought of as an additional reagent in the separation of nucleic acids. Temperature controls whether the nucleic acid is separated as a single molecule, double-stranded molecule or something in between. Double stranded nucleic acids are held together by hydrogen bonding of the two strands. As the temperature is increased, these bonds are broken, making two strands of single-stranded nucleic acid. Temperature is used to achieve the three modes of operation: nondenaturing mode, fully denaturing mode, and partially denaturing mode. 12.4.1 Nondenaturing Mode
The nondenaturing mode of DNA and RNA chromatography is used to separate double-stranded nucleic acids. This is most often used for DNA, but some short sequences of RNA can also be double-stranded. Obviously, it is important to keep the double strand intact while performing these separations. The breaking of hydrogen bonds of double-stranded DNA by increasing the temperature is called melting or denaturing. The temperature at which this occurs depends on the strength of the hydrogen bonding and the environment around the DNA. Higher salt and buffer content will raise the melting temperature. Also, DNA adsorbed onto a solid surface, such as column packing material, will require a higher temperature to melt the DNA. Conversely, the presence of an organic solvent such as acetonitrile will lower the melting temperature, and the effect is increased with increasing concentrations of the solvent. In the nondenaturing mode, a sufficiently high temperature is chosen to lower eluent viscosity (and therefore column back pressure), but not so high that denaturing occurs. The normal oven temperature of operation for nondenaturing mode is 50 °C.
307
308
12 DNA and RNA Chromatography
12.4.2 Fully Denaturing Mode
Single-stranded DNA and RNA may contain secondary structure due to complementary sequences within its own fragment. If portions of sequences are complementary, the fragments may fold back on themselves through the formation of intra-molecular hydrogen bonds, giving several different possible structures for the same fragment. The presence of these secondary structures is uncertain and nonreproducible and will lead to peak broadening if they occur. Increasing the temperature can break up the hydrogen bonding and therefore reduce some or perhaps most of the secondary structure [34]. Figures 12.4a and 12.4b show the separation of an RNA ladder at 40 and 75 °C. The peaks of the RNA become sharper and more uniform with increasing oven temperature. It is not certain that temperature will always reduce secondary structure. In these cases, using several different temperatures for the same sample can give clues as to the nature of stronger secondary structure.
Figure 12.4a Chromatogram of RNA ladder at 40 °C. RNA ladder (Cat. No. 15623010, Life Technologies) has nucleotide lengths of 155, 280, 400, 530, 780, 1280, 1520, and 1770 bases. Buffer A: 1 M TEAA, pH 7.0, buffer B: 1 M
TEAA, pH 7.0 with 25% v/v acetonitrile, gradient 0.0 min 38% B, 1.0 min 40%, 16 min 60%, 22 min 66%, 22.5 min 70%, and 23 min 100%. (from reference [34] with permission).
12.4 Temperature Modes of DNA and RNA Chromatography
Figure 12.4b Same conditions as Figure 12.4a except chromatogram is at 75 °C. It is not certain that temperature will always reduce secondary structure.
12.4.3 Partially Denaturing Mode
The use of partially denaturing mode can detect mutations in DNA through a heteroduplex detection process. Remember that the hydrogen bonding holding two complementary strands of DNA together is specific. The adenine base always bonds to thymine (and vice versa) and guanine always bonds to cytosine (and vice versa). If the two strands of a DNA fragment are perfectly matched, then they are hydrogen bonded at each and every nucleic acid site. They are said to be a homoduplex nucleic acid fragment because both strands of the DNA are completely complementary to the other. A heteroduplex nucleic acid fragment arises where a genetic mutation has occurred. One of the bases has mutated so that a doublestranded fragment is not completely complementary but now contains a base that cannot hydrogen-bond to the base located on the other fragment. Under a partially denaturing mode, differences in melting of double-stranded DNA are detected because increased melting decreases the retention time of the fragment. The operating temperature of the column and eluent is chosen to partially denature or melt the fragment to enhance this mismatch. Since singlestranded DNA elutes differently from double-stranded DNA, a mixture containing both homoduplex fragments and heteroduplex fragments can be separated. Typically, oven temperatures between 54 and 72 °C are used. The method can be called a ‘difference detecting engine’ because it detects the presence of a heteroduplex regardless of the sequence being studied. This type of chromatography is sometimes called DHPLC or denaturing HPLC. More details on how DHPLC works are described later in an example of the detection of DNA mutations.
309
310
12 DNA and RNA Chromatography
12.5 Instrumentation
The components of the HPLC system are quite similar to what is used for standard ion chromatography with some important refinements in the general flow path, oven, and detector. In many cases, the nucleic acid that is separated is collected in a fragment collector to be used for further research. These instrument refinements are described in this section. 12.5.1 Effect of Metal Contamination
In the course of commercializing the DNASep® column technology, a degradation effect was discovered after the column had experienced medium to longer term use [35–38]. The DNASep® column would work for a while and then would start giving either split peaks for each fragment eluting from the column or stop giving peaks at all. At first, the packing procedure was extensively studied because split peaks mean that the chromatographic bed has become cracked or disrupted. This was not unexpected since polymer beds are prone to this kind of behavior if they are not packed properly. The packing procedure was studied to rule out the possibility of unstable packed column beds. Then, the column was put into a standard HPLC system that destroyed the separation slowly and somewhat controllably. In one set of experiments, a new column showed excellent separation of a pUC 18 Hae III digest (Figure 12.5a). As the column was used, a degradation effect was observed as a loss of resolution for base pairs greater than 200 (Figure 12.5b). As the degradation continued, increasingly shorter fragments of DNA were affected. Many of the peaks were split or doublets (Figure 12.5c). Eventually, the DNA did not elute at all from the system. Thus, the degradation or decreasing resolution appeared to be a function of the length of the polynucleotide fragment being separated. The cause of the split peaks for the larger DNA fragments was discovered when attempts were made to clean the column. Clean-up with organic solvents did not improve performance. However, subsequent clean-up of the column with injections of tetrasodium ethylenediaminetetraacetic acid (EDTA), a metal-chelating agent, largely restored chromatographic resolution. It appeared that the chelating reagent passivated the metal in the system by removing surface oxidized metals so that they could not bind with the DNA. Adding small amounts (i.e., 0.1 mM) of tetrasodium EDTA to the mobile phase can be done without significant changes to the chromatography. The most significant sources of metal ions are HPLC components containing fritted filters made of stainless steel. Fritted filter components are used in mobile phase filters, check valve filters, helium spargers, mobile phase mixers, inline filters, column frits, and other parts of the HPLC.
12.5 Instrumentation
Figure 12.5a This is the first of a series of three chromatograms showing a separation, illustrating the degradation effect of running an HPLC releasing contamination over time to the column. The separation is of doublestranded DNA, pUC18 Hae III. The degradation occurs first on larger fragments.
Smaller fragments are eventually degraded as well. This first chromatogram uses a new DNASep® column, Buffer A: 1 M TEAA, pH 7.0, buffer B: 1 M TEAA, pH 7.0 with 25% v/v acetonitrile, gradient, 0 min 35% B, 3 min 55% B, 10 min 65% B, and 12 min 100 % B.
Figure 12.5b After several minutes of use, the column shows the initial degradation of the HPLC separation with fragments greater than about 200 bp affected.
311
312
12 DNA and RNA Chromatography
Figure 12.5c After several more minutes of use, the column shows the greater degradation of the HPLC separation. Now all fragments are affected (from Ref. [35] with permission).
Metal ion contamination, such as colloidal iron, can be released from frits, travel to other parts of the HPLC, and then be trapped. These types of contaminants will interfere with DNA in solution or after it has been released and trapped on a critical component of the HPLC such as the column, an inline filter in front of the detector, or at a back-pressure device located after the detector. Iron(III) and other metals can form insoluble complexes with phosphate anion. It is likely that surface metal ions and/or colloidal iron are combining with one or more phosphate groups on the nucleic acid fragments. If this happens, the ionpair process chromatographic separation can be interrupted, causing the peaks to broaden. In extreme cases, it is possible that this metal contamination is so severe that nucleic acid fragments are completely prevented from eluting from the system and no peaks are detected. Subsequent experiments showed that even if titanium or PEEK components are used in the fluid path, then some treatment was necessary before the components could be used. Although an improvement, the use of titanium frits did not initially give consistent results. However, treatment of the frits with dilute nitric acid and then with a chelating agent did improve the performance of the instrument. Similarly, as shown in the examples, PEEK frits were not consistently suitable for DNA chromatography, but acid treatment did improve their performance. Finally, degassing the fluid before it enters the liquid chromatography system in order to remove oxygen will inhibit oxidation and hence production of metal ions in stainless steel or titanium or other tubing containing iron. All of these improvements are incorporated into commercial instruments designed for nucleic acid chromatography. It is also important to follow preventive maintenance procedures and check filters for signs of colored deposits. Since double-stranded DNA is more susceptible to contamination, the use of precautions with respect to the method and the system is much more critical than when the
12.5 Instrumentation
system is used to separate single-stranded DNA. Failure to keep the column clean is probably the most common error that the user can make. More details on procedures and recipes for instrument and column maintenance are described in a book entitled, ‘DNA Chromatography’ by Gjerde et al. [39]. 12.5.2 The Column Oven
The oven controls the temperature of the fluid entering the separation column. This includes not only the eluent but the sample that is injected into the system. The fluid entering the oven compartment is normally cooler than the oven and there is a time lag before this new fluid comes to the oven temperature. In most HPLC ovens, the fluid never does reach the set point of the oven. In DNA and RNA separation instrumentation, a pre-heat tube is used to bring the fluid to column temperature before it reaches the column. The oven temperature should be accurate, should not drift and should be precise, i.e. it should come to the same temperature each time it is directed by the run method to go to a specific temperature. The oven remains one of the most critical and difficult parameters to control in the DNA chromatograph. The reader should consult the HPLC manufacturer for information on oven use, calibration and upkeep. Selerity Technologies (Salt Lake City, UT) has developed an active eluent preheater that uses a feedback technology to maintain a set temperature for the eluent prior to entering the column. This unit appears to be compatible with different HPLC instrumentation. 12.5.3 UV and Fluorescence Detection
Nucleic acids absorb light strongly in the UV, the absorption maximum being at a wavelength of 260 nm. Variable-wavelength detectors are set at 260 nm for detection; however, single-wavelength detectors work very well at 254 nm. The ion-pairing reagent does not absorb at these wavelengths, and the nucleic acid fragments may be detected directly at the sub-nanogram level by UV automatic detection. Fluorescence detection could also be used provided that fluorescence tags are added to the DNA. DNA is tagged by using special fluorescently tagged nucleotides when the fragments are being synthesized. Use of this detection method decreases the amount of DNA that can be detected by a factor of 10 to 100 (or even 1000 in some reported cases). Common dyes used in molecular biology include FAM, TET, HEX, TAMRA, NED, Pacific Blue, and many others. A table of fluorescent tags can be found in Chapter 4.
313
314
12 DNA and RNA Chromatography
12.5.4 Fragment Collection
One of the powerful features of DNA and RNA chromatography is that material can easily be purified by collecting directly from the detector effluent. Purification can be used in biological samples, where there may be several types of nucleic acids present and a particular type is desired for study. An example of this is shown later in this chapter where several types of RNA are detected in a cell extract and can be purified. The collection can be done by hand, but is normally accomplished with an automated fragment collector and controlling software. Collection can be into single vials or large 96-well plates. Care should be taken to ensure that the actual peak has been collected. Measurement of recovery is performed by taking a small portion of the recovered peak, reinjecting and measuring the area, multiplying this by the ratio of total collected volume to reinjected volume, and comparing this value to the area of the original peak. Normal recoveries are about 80% of injected material. Of course materials such as RNA may degrade (by enzymes) after they have been collected. Higher recoveries may not be possible owing to loss of material during the concentration process through precipitation or plating of material on surfaces. An important parameter in fragment collection is careful execution of the collection times. There is a lag from the time the peak is seen in the detector cell and when the fragment is deposited from the fragment collector probe. The most reliable collection method is the timed collection. But unless the timing is correct, it is easily possible to miss some of the peak or even the entire peak. It is also important that there is as little dead volume in the tubing from the detector cell outlet to the tip of the deposition probe. Too large a dead volume will destroy the resolution of the separation and could result in cross contamination of the peak of interest with neighboring peaks. It is important that the probe is cleaned between collection of peaks. This is normally done automatically by the fragment collector.
12.6 Applications of DNA Chromatography 12.6.1 DHPLC
The use of column temperature to control separations in DNA chromatography was described in 1996 through the insights of Oefner and Underhill [40–43]. They demonstrated that DNA chromatography possessed unique properties, enabling the separation of DNA based on its relative degree of helicity. Heteroduplex DNA has a lower melting point than homoduplex DNA. The retention of singlestranded DNA is lower on the column than that of double-stranded DNA. Heteroduplex DNA melts or denatures more easily, and it elutes earlier in the separation. This technique is called denaturing HPLC (DHPLC).
12.6 Applications of DNA Chromatography
All individuals have mutations in their DNA that have spontaneously occurred in the past (due to chemical or radiation exposure) and have been passed on from generation to generation. Some of these mutations are serious and may cause disease, but most are benign. Mutations in an individual can be found by comparing specific DNA fragments of that specific individual to the same fragment sequence in the general population. Usually, DNA fragments of about 600 base pairs are examined in genes that are known to be responsible for various types of diseases. If a mutation is present, a heteroduplex is formed and can be detected. If a mutation is not present only nonmutated homoduplexes are detected. Details of the fragment selection process are beyond the scope of this chapter and can be found in the Oefner and Underhill references. An excellent flash movie showing how homoduplex and heteroduplex species are formed and then detected by DHPLC can be viewed by clicking the DHPLC icon on the web page following the link www.transgenomic.com/ap/VariationApp4.asp. 12.6.2 Nucleic Acid Enzymology
Many chemical reactions occur within biological cells. A special class of protein molecules called enzymes speed up or control these reactions. Chemical reactions need a certain amount of activation energy to take place, and enzymes can increase reaction speed by allowing a different reaction path with a lower activation energy, making it easier for the reaction to occur. Enzymes are essential for the function of cells. They are very specific to the reactions they catalyze and the chemical substrates that are involved in the reactions. Enzymes are said to fit their substrate like a key fits its lock. The control of the expression of proteins by DNA genes relies on special proteins that bind to specific DNA sequences. The way in which these proteins recognize their binding sites in the genome is an important step in understanding how these processes occur at the molecular level. Nucleic acid enzymology is the investigation of the properties of enzymes that act on or with nucleic acids. These enzymes include DNA polymerases (for replicating or copying DNA), restriction endonucleases (for base cutting at a specific known base sequence) and reverse transcriptases (for transcribing the information contained in DNA). Chromatography, coupled with fluorescence detection, can be a generic analytical platform for nucleic acid enzymology.
12.6.2.1 Telomerase Assays Telomerase is a ribonucleoprotein complex that plays a critical role in cellular mortality [44]. The vertebrate enzyme catalyzes the addition of TTAGGG sequence repeats to the ends of chromosomes. In the absence of telomerase, human telomeres undergo progressive shortening with each round of cell division, an event that may contribute to cell aging and mortality. Telomerase is known to be asso-
315
316
12 DNA and RNA Chromatography
ciated with immortalized cancer cells but is absent in most normal tissues and has therefore become a focus for diagnostic investigation. The presence of telomorase can be measured by the chromatographic full-denaturing measurement of extension products. If telomerase is present, then the DNA fragments are extended to longer fragments. The amount of telomerase present is indicated by the extent to which the fragments have been extended.
12.6.2.2 Polynucleotide Kinase Assays Polynucleotide kinase catalyzes the addition of a phosphate group to DNA and RNA. The enzymatic phosphorylation of synthetic single-stranded DNA can be applied to experimental molecular biology [45]. Radioactive (32P) phosphate groups are used to make the single-stranded probes that can be incorporated into biological reactions. Measurement of the proportion of probe that has been successfully labeled is normally difficult to determine. However, the subtle ionic difference between phosphorylated and unphosphorylated DNA is readily resolved by full denaturing DNA chromatography in fragments up to 50 bases in length. Figure 12.6 shows the effects of phosphorylation of a single-stranded oligodeoxynucleotide upon retention time following DNA chromatography. Phosphorylation of the oligodeoxynucleotide reduces the retention time of the modified DNA, because the addition of the polar phosphate group diminishes the net hydrophobicity of the DNA. These data are readily corroborated by mass spectrometry [46].
Figure 12.6 The use of DNA chromatography in the resolution of differentially phosphorylated oligodeoxynucleotides under denaturing conditions. The difference in polarity between the two species is clearly sufficient to resolve the strands, which may be up to 50 to 100
nucleotides in length. Confirmation of the products was obtained by mass spectrometry [44] which is a powerful complementary technique to DNA chromatography since the eluted material is ideal for mass spectrometry immediately post column.
12.7 Applications of RNA Chromatography
12.6.2.3 Uracil DNA Glycosylase Assays Uracil DNA glycosylase (UDGase) is an enzyme that removes uracil from both singlestranded and double-stranded DNA [47]. The normal RNA base, uracil, can arise in DNA spontaneously. Uracil in DNA is pro-mutagenic and can disrupt the binding of sequence-specific gene regulatory proteins. All DNA-containing organisms contain a specific repair pathway which removes uracil from DNA with UDGase enzyme. A novel assay was developed using full denaturing DNA chromatography [45]. After treatment with UDGase when uracil is present in the DNA strand, there is a significant shift in retention time of the modified single-stranded DNA. Removal of the uracil base from the single-stranded DNA results in elution at an earlier retention time through loss of the uracil base and subsequent decrease in hydrophobicity of the DNA.
12.7 Applications of RNA Chromatography
RNA is found in a repertoire of intra- and inter-molecular hydrogen bonding interactions. One familiar structural element in an RNA molecule is the stem-loop, in which a noncomplementary segment separates two complementary stretches of nucleotides. This structural diversity is characteristic of biological RNA. The retention of RNA molecules on a chromatography column is a function of chain length, surface chemistry, morphology and localized single- and double-stranded character. Even with thorough thermal denaturation, it is unlikely that a given population of RNA molecules is completely free from secondary and tertiary interactions. Thus, the chromatography of RNA will be different and more unpredictable compared to the chromatography of single-stranded and double-stranded DNA. Still, RNA chromatography is best performed at elevated temperatures to denature as much as possible and reduce secondary of the RNA being separated. There are several types of RNA. Silencing RNA (siRNA) is a 20–25 nucleotide long double-stranded nuclear RNA studied in biological research and has high potential in new drugs. Silencing RNA gets its name because of its ability to ’silence’ or stop a protein from being produced. The sequence of the siRNA is chosen so that a strand of the RNA molecule can hydrogen-bond with the targeted sequence of a particular gene, which produces the desired effect. Research in RNA in gene control is increasing rapidly and it is certain that many methods will be developed. Other RNA species include tRNA (transfer RNA), mRNA (messenger RNA) and rRNA (ribosomal RNA). A typical RNA chromatography separation of a total RNA extraction obtained under full denaturing conditions is shown in Fig 12.7 [34]. The earliest eluting species include the population of tRNAs (and probably includes small nuclear RNAs), the middle section of the profile is dominated by the rRNA species, and finally, underlying the entire chromatogram, is a spectrum of mRNAs with many of the fragments centered on the later retention times.
317
318
12 DNA and RNA Chromatography
Figure 12.7 RNA chromatography of a total cellular extract of RNA from tobacco plants is displayed. The large, broad peak eluting between 12 and 15 min is primarily rRNA together with mRNA.
12.7.1 Separation of Messenger RNA from Ribosomal RNA
The cellular processes that are expressed in a living cell are the direct result of the proteins contained in the cell. This in turn is closely linked to the population of mRNA molecules produced by the genes contained in a given cell type. For this reason there has been considerable interest in evaluating qualitatively and quantitatively the various types of RNA contained in a cell. One of the key experimental procedures is the systematic removal of rRNA from the mRNA fraction. Typically, mRNA has a long poly A tail that can be captured on a poly T chromatography column or batch resin. The rRNA and tRNAs pass through the column and, after a suitable washing protocol, the mRNA fraction is concentrated and stored for subsequent experimentation. The effectiveness of the purification of a typical mRNA purification scheme was followed by RNA chromatography by Hornby and coworkers [34]. It was found that at least two rounds of enrichment are usually required in order to remove the bulk of the nonpoly A RNA. It is possible to apply RNA chromatography in a preparative mode in order to produce a series of RNA fractions which can be used for subsequent research. One of the major drawbacks of poly A mRNA isolation is that some mRNAs are not polyadenylated and therefore will be excluded from any subsequent analysis. The use of RNA chromatography in a preparative mode offers the potential for isolating at least a fraction of those mRNA species that do not co-elute with rRNA. This is clearly an important area for development, since the analysis of cell-specif-
12.7 Applications of RNA Chromatography
ic RNA populations in disease is becoming increasingly important in molecular medicine. 12.7.2 Analysis of Transfer RNA
Each cell contains a population (usually referred to as a pool) of tRNAs that meet the requirements of that particular cell’s (or in the case of bacteria, that organism’s) protein synthesis machinery. The range of sizes of tRNA molecules is particularly narrow compared with mRNA, between around 60 and 100 bases, and a given cell typically contains around 100 species. While this molecular weight range is ideal for RNA chromatography, the close similarity of molecular sizes represents a problem for resolving individual components in a typical cellular pool. This is readily seen in Figure 12.8 where the total pool of tRNAs from Escherichia coli has been separated by RNA chromatography [34, 48].
Figure 12.8 RNA Chromatogram of the entire pool of tRNAs purified from E. coli. By coupling separation with RTPCR it is possible to identify and quantify the individual tRNAs. A typical population of tRNAs will contain around 50 species.
319
320
12 DNA and RNA Chromatography
References [1] H. Ellegren and T. Laas, Size-exclusion
[2]
[3]
[4]
[5]
[6]
[7]
[8]
[9]
chromatography of DNA restriction fragment. Fragment length determinations and a comparison with the behavior of proteins in size-exclusion chromatography, J. Chromatogr. 467, 217, 1989. G. Bernardi, Chromatography of nucleic acids on hydroxyapatite column, in: L. Grossman and K. Moldave (eds.), Methods in Enzymology, Vol. 21, Academic Press: New York, pp 92–140, 1971. M. Soyab and Y. Sen, The isolation of extrachromosomal DNA by hydroxyapatite chromatography, in Methods in Enzymology (ed. R. Wu ), Vol. 68, Academic Press: New York, pp 199–206, 1979. A. D. Kelmers, G. D. Novelli and M. P. Stulberg, Separation of transfer ribonucleic acids by reverse phase chromatography, J. Biol. Chem., 240, 3979, 1965. A. D. Kelmers, H. O. Weeren, J. F. Weiss, P. S. Pearson, M. P. Stulber and G. D. Novelli, Reversed-phase chromatography systems for transfer ribonucliec acid – preparatory scale methods, in Methods in Enzymology (eds., K. Moldave and L. Grossman) Vol. 20, Part C, Academic Press: New York, pp 9–34, 1971. R. L. Pearson, J. F.Weiss and A. D. Kelmers, Improved separation of transfer RNAs on polychlorotrifluoroethylene-supported reversed-phase chromatography columns, Biochem. Biophys. Acta, 228, 770, 1971. R. D. Wells, S. C. Hardies, G. T. Horn, B. Klein, J. E. Larson, S. K. Neuendorf, N. Panayatatos, R. K. Patient, and E. Selsign, PRC-5 column for the isolation of DNA fragments, in Methods in Enzymology (eds. L. Grossman and K. Moldave.), Vol. 65, Academic Press, New York, pp 327–347, 1980. J. A. Thompson and R. W. Blakeley, Higher performance liquid chromatography purification of nucleic acids, Fed. Proc. 42, 1953, 1983. R. D. Wells, High performance liquid chromatography of DNA, J. Chromatogr., 336, 3, 1984.
[10] R. A. Hardwick and P. R. Brown, The
[11]
[12]
[13]
[14]
[15]
[16]
[17]
[18]
[19]
[20]
performance of micro particle chemically-bonded anion-exchange resins in the analysis of nucleotides, J. Chromatogr., 112, 650, 1975. E. Westman, S. Eriksson, T. Laas, P. Parental and S. E. Scold, Separation of DNA restriction fragments by ionexchange chromatography on FPLC columns Mono P and Mono Q, Anal. Biochem., 166, 158, 1987. F. Flack and V. Sarantoglou, Curved DNA fragments display retarded elution upon anion exchange HPLC, Nucleic Acids Res., 19, 4181, 1991. W. Bloch, Precision and accuracy of anion-exchange separation of nucleic acids, U. S. Pat. 5 856 192, 1999. E. D. Katz and M. W. Dong, Rapid analysis and purification of polymerase chain reaction products by high-performance liquid chromatography, Biotechniques, 8, 546, 1990. E. D. Katz, Protocol: Quantitation and purification of polymerase chain reaction products by high-performance liquid chromatography, Molec. Biotechnol, 6, 79, 1996. E. D. Katz, L. A. Haff and R. Eksteen, Rapid separation, quantitation and purification of products of polymerase chain reaction by liquid chromatography, J. Chromatogr. 512, 433, 1990. J. M. Wages, X. Zhao and E. D. Katz, High-performance liquid chromatography analysis of PCR products, PCR Strategies, Academic Press: San Diego, pp 140–153, 1995. J. A. Thompson, A review of high performance liquid chromatography in nucleic acids research I. Historical perspectives, Biochromatography, 1, 16, 1986. G. Zon and J. A. Thompson, A review of high performance liquid chromatography in nucleic acids research II. Isolation, purification, and analysis of oligodeoxyribonucleotides, Biochromatography, 1, 22, 1986. J. A. Thompson, A review of high performance liquid chromatography in
References
[21]
[22]
[23]
[24]
[25]
[26]
[27]
[28]
[29]
nucleic acids research III. Isolation purification, and analysis of supercoiled plasmid DNA, Biochromatography, 1, 68, 1986. J. A. Thompson, A review of high performance liquid chromatography in nucleic acids research IV. Isolation, purification, and analysis of DNA restriction fragments, Biochromatography, 2, 4, 1987. J. A. Thompson, A review of high performance liquid chromatography in nucleic acids research V. Nucleic acid affinity techniques in DNA and RNA research, Biochromatography, 2, 68, 1987. J. A. Thompson, S. Garfinkel, R. B. Cohen and B. Safer, A review of high performance liquid chromatography in nucleic acids research VI. Nucleic acid affinity techniques in DNA-binding Protein research, Biochromatography, 2, 166, 1987. J. A. Thompson, R. W. Blakesley, K. Doran, C. J. Hough and R. D. Wells, Purification of nucleic acids, by RPC-5 analog chromatography: peristaltic and gravity-flow application. Methods in Enzymology, Vol. 100 Academic Press: New York, pp 368-399, 1983. A. P. Green, J. Burzynski, N. M. Helveston, G. M. Prior, W. H. Wunner and J. A. Thompson, HPLC purification of synthetic oligodeoxyribonucleotides containing baseand backbone- modified sequences, Biotechniques, 19, 836, 1995. A. P. Green, G. M. Prior, N. M. Helveston, B. E. Taittinger, X. Liu and J. A. Thompson, Preparative purification of supercoiled plasmid DNA for therapeutic applications, Biopharm., pp 52-–62, May 1997. C. G. Huber, P. J. Oefner, and G. K. Bonn, Rapid analysis of biopolymers on modified non-porous polystryrene – divinylbenzene particles. Chromatographia, 37, 653, 1993. G. Bonn, C. Huber and P. Oefner, Nucleic acid separation on alkylated nonporous polymer beads, U.S. Pat. 5 585 236, 1996. C. G. Huber, P. J. Oefner, E. Preuss and G.K. Bonn, High-resolution liquid chro-
matography of DNA fragments on highly cross-linked poly(styrene-divinylbenzene) particles, Nucleic Acids Res. 21, 1061, 1993. [30] L. Trojer, S. H. Lubbad, C.P. Bisjak, W. Wieder and G. K. Bonn, Comparison between monolithic conventional size, microbore and capillary poly (p-methylstyrene-co-1,2-bis(p-vinylphenyl)ethane) high-performance liquid chromatography columns Synthesis, application, long-term stability and reproducibility, J. Chromatogr A. 1146(2), 216–24, 2007. [31] T. A. Jakschitz, C. W. Huck, S. Lubbad and G. K. Bonn, Monolithic poly[(trimethylsilyl-4-methylstyrene)-cobis(4-vinylbenzyl)dimethylsilane] stationary phases for the fast separation of proteins and oligonucleotides, J. Chromatogr A. 1147(1), 53–58, 2007. [32] W. Wieder, C. P. Bisjak, C. W. Huck, R. Bakry and G. K. Bonn, Monolithic poly(glycidyl methacrylate-co-divinylbenzene) capillary columns functionalized to strong anion exchangers for nucleotide and oligonucleotide separation, J. Sep. Sci. 16, 2478–2484, 2006. [33] Application note AN120, Optimized purification of siRNA oligonuclotides using the WAVE® Oligo System Transgenomic Inc. Omaha, NE. [34] D. T. Gjerde, D. P. Hornby, C. P. Hanna, A. I. Kuklin, R. M. Haefele and P. D. Taylor, Method and system for RNA analysis by matched ion polynucleotide chromatography, U. S. Pat. 6 576 133, 2003. [35] D. T. Gjerde, R. M. Haefele and D. W. Togami, Method for performing polynucleotide separations using liquid chromatography, U. S. Pat. 6 017 457, 2000. [36] D. T. Gjerde, R. M. Haefele, and D. W. Togami, System and method for performing polynucleotide separations using liquid chromatography, U. S. Pat. 5 772 889, 1998. [37] D. T. Gjerde, R. M. Haefele and D. W. Togami, Apparatus for performing polynucleotide separations using liquid chromatography, U. S. Pat. 6 030 527, 2000. [38] D. T. Gjerde, R. M. Haefele and D. W. Togami, Liquid chromatography
321
322
12 DNA and RNA Chromatography
[39]
[40]
[41]
[42]
[43]
systems for performing polynucleotide separations, U. S. Pat. 5 997 742, 1999. D. T. Gjerde, C. P. Hanna and D. Hornby, DNA Chromatography, Wiley-VCH, Weinheim, 2002. P. J. Oefner and P. A. Underhill, Detection of nucleic acid heteroduplex molecules by denaturing high-performance liquid chromatography and methods for comparative sequencing, U. S. Pat. 5 795 976, 1995. P. J. Oefner and P. A. Underhill, DNA Mutation detection using denaturing high-performance liquid chromatography, in Current Protocols in Human Genetics (eds N. C. Dracopoli, J. L. Haines, B. R. Korf, D. T. Moir, C. C. Morton, C. E. Seidman, J. G. Seidman, D. R. Smith), Wiley & Sons: New York, p 7.10.1 – 7.10.12, 1998. P. A. Underhill, L. Jin, A. A. Lin, S. Q. Mehdi, T. Jenkins, D. Vollrath, R. W. Davis, L. L. Cavalli-Sforza and P. J. Oefner, Detection of numerous Y chromosome biallelic polymorphisms by denaturing high performance liquid chromatography, Genome Res. 7, 996, 1997. P. A. Underhill, L. Jin, R. Zemans, P. J. Oefner and L. L. Cavalli-Sforza, A
pre-columbian Y chromosome-specific transition and its implications for human evolutionary history, Proc. Natl. Acad. Sci. 93, 196, 1996. [44] G. B. Morin, The human telomere terminal transferase enzyme is a ribonucleoprotein that synthesizes TTAGGG repeats, Cell, 59, 521, 1989. [45] J. Sambrook, E. F. Fritsch and T. Maniatis, in Molecular Cloning: A Laboratory Manual, 2nd edn, Cold Spring Harbor Laboratory Press, New York: 1989. [46] M. J. Dickman, M. M. Matin and D. P. Hornby, High throughput analysis of nucleic acid modification reactions using IP RP HPLC, Anal. Biochem. 301, 290, 2002. [47] G. Slupphaug, I. Eftedal, B. Kavli, S. Bharati, N. M. Helle, T. Haug, D. W. Levine and H. E. Krokan, Properties of a recombinant human uracilDNA glycosylase from the ung gene and evidence that ung encodes the major uracil-DNA glycosylase, Biochemistry, 34, 28, 1995. [48] M. J. Dickman, M. Conroy, J. Grasby and D. P. Hornby, RNA footprinting analysis using ion pair reverse phase liquid chromatography, RNA, 8, 247, 2002.
323
13 Sample Pretreatment 13.1 Dilute and Shoot or Pre-treat the Sample?
One of the strengths of ion chromatography is the simplicity and ruggedness of the analytical methods developed from its technology. The chromatograms generated can be extremely reliable in that exactly the same peak profile is generated from a given sample whether it is injected once or a dozen times. In many cases very little is done to prepare the sample prior to injection. One hears the phrase ‘just dilute and shoot’. Nothing else is needed other than to dilute the sample with de-ionized water, mix and inject the sample. But sometimes more may be needed, either to protect expensive columns or to make the sample detectable. It is possible that a single injection of the wrong sample can destroy a column through contamination or plugging. Or perhaps a sample component must be enhanced to increase the concentration to a detectable level. Sometimes more is needed than ‘just dilute and shoot’. This chapter describes the various sample pretreatment methods, why they are used and where they are used. Sample pretreatment methods can be classified depending on need or effect: 1. Remove particulate (that may plug the column) (a) Pre-injection removal (syringe filter) (b) In-line column filter (before the separation column). 2. Remove column contaminating material or particulate with a guard column. 3. Enhance or enrich the concentration to improve the detection limit. 4. Make the sample detectable by a process that collects the sample or converts the sample to a detectable form. 5. Remove an interfering ion (remove an ion that co-elutes with the sample ion). Whatever is done to pre-treat or prepare the sample for analysis, it must be done while preserving the integrity of the sample. New contaminants must not be added. Sample ions must not be lost or the chemical form changed. They must Ion Chromatography, 4th Ed. James S. Fritz and Douglas T. Gjerde Copyright © 2009 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim ISBN: 978-3-527-32052-3
324
13 Sample Pretreatment
not be oxidized or reduced to a different species. In that sense, just dilute and shoot is attractive because sample integrity is most likely held true in the injection process. But with proper method design, sample integrity is held with sample pretreatment as well.
13.2 Particulate and Column-contaminating Matter
The sample solutions injected into an ion chromatographic system must be free of particulate matter to avoid plugging of the capillary connecting tubing and the frits at the head of the analytical column. Even samples that appear to be clear may contain unsuspected fine particles. Disposable membrane syringe filters with a pore diameter of 0.45 lm are sufficient in most cases to clean a sample from harmful particulates. Samples with biological activity are filtered through aseptic filters with a pore diameter of 0.22 lm to avoid any change in sample composition due to bacterial oxidation or reduction. It is important to note that these filters may contain ionic contamination even if certified for use in ion chromatography applications. In general, membrane filters described below should be prerinsed with de-ionized water prior to use to avoid sample contamination. Then the sample is filtered, discarding the first portion that comes out of the filter that is diluted with the de-ionized water. Inline filters are very effective to protect columns from particulate material. These filters have low dead volume so that they can be placed anywhere after the injector and before the column. The filters are usually replaceable and cost effective to replace. The filter bodies and filter inserts can be metal or polymer. Metal filters are usually made of 316 stainless steel and can be used even if the rest of the chromatographic is plastic. It is a good idea to have an inline filter regardless of whether samples are pre-filtered or not. Solid or semi-solid samples may require extraction with an aqueous solution to isolate the ionic components in a form suitable for IC. The actual procedures vary widely, depending on the type of sample. For example, meat and sausage products to be analyzed for nitrate and nitrite are first homogenized mechanically, extracted with a 5% borax buffer solution in a hot water bath, and then subjected to a precipitation with strong solutions of potassium hexacyano ferrate and zinc sulfate. The aqueous extracts are diluted further with de-ionized water and filtered through a membrane prior to injection. Organic substances in the sample matrix may interfere with ion chromatographic separations. In some cases it is sufficient to add enough methanol or other organic solvent to completely dissolve the organic matter. But in samples that are soluble in water alone, the organic components may be adsorbed by the IC column packing and prevent reproducible results. This often occurs with dyes that are added to many commercial products. Dyes and many other types of soluble organic compounds can usually be removed by some form of solid-phase extraction (SPE) without altering the inor-
13.3 Preconcentration
ganic ion content of the sample. In simple cases an SPE cartridge or membrane disk is attached to the sample syringe. Then the liquid sample is injected through the SPE material into the ion chromatograph. SPE cartridges are attached to the sample syringe by means of a Luer tip. A solid disk can be cut from a larger SPE disk, such as the Empore disks produced by the 3M Co., and fitted tightly inside the sample syringe. A convenient semi-micro device for SPE with a syringe has been described [1]. Many methods are available for removal of organic material from aqueous samples by off-line SPE [2]. Hydrophobic organic material is best extracted by solid poly(styrene–DVB) polymers or reversed-phase silica extractants. Polyvinylpyrrolidone (PVP) is an appropriate choice for removal of humic acids, lignins and tannins from water samples. As stated before in connection with syringe filters, one should be careful of ion contamination with SPE columns. The columns should be pre-rinsed with de-ionized water. The last line of defense is the guard column, a short column that is configured directly in front of the separation columns. The guard column mimics the separation column so that whatever detrimental event would happen to the separation column happens instead to the less expensive, replaceable guard column first. In most cases, the guard column uses exactly the sample resin as the separation column (although in rare cases the particles may be a little larger). If one understands the sample and column chemistry well, it is possible to use a guard column that has a different chemistry. For example, a cation-exchange chemistry might be used in front of an anion-exchanger separation column to remove heavy metals from the sample. This might be useful to prevent certain metals from hydrolyzing and precipitating in the high-pH environment of the separation column.
13.3 Preconcentration 13.3.1 Collection of Ions from Air
In the manufacture of semiconductor integrated circuits, it is of the utmost importance to manufacture in a ‘contaminant-free’ environment. The wafers and chips containing integrated circuits will fail if a particulate or ion contaminant shorts out any one of the circuit connections. As more and more integrated circuits are placed on the silica wafer the distance between the circuit connections become less. This means contamination is becoming more of an issue so that it has been proposed to monitor the air for airborne ionic contaminants. Lue and coworkers have described a method of monitoring acidic airborne contaminants in clean rooms [3]. In this method, acidic contaminants were adsorbed on silica gel tubes by passing a known volume of air through the tubes. The adsorbed impurities were extracted by a solution of carbonate and hydrogen car-
325
326
13 Sample Pretreatment
bonate and determined by ion chromatography. The recovery of HF was 100% and that of HCl was 91–100%. 13.3.2 Preconcentration of Ions in Water
Ion chromatography is frequently used to determine anions and cations at very low concentration levels, often in the low lg L–1 (ppb) range. In the electric power industry the water used in steam generators must be almost free of Na+, Cl– and other ions to avoid stress corrosion cracking of turbines, pipes, etc. The ionic content of ultrapure water used in the electronics industry must be kept to extremely low levels. Semiconductor chip manufacturers require clean rooms with utility impurities of no more than 1 ppb for 0.35lm devices [3]. As advances in technology have been made, IC detection limits have been lowered through careful control of the column and detector temperature and improvements in pump and detector design. One way to extend ion chromatography to even lower limits of detection is to increase the injection volume of the sample. A sample loop of 100 lL or even 400 lL could be used instead of a more typical 10–50 lL loop. By injecting 10–20 times more sample, the limits of detection for sample ions should be correspondingly lowered. However, the zone occupied by the sample solution in the IC column will also be larger. Since the sample ionic strength is much lower than that of the eluent there may be a prolonged dip in the chromatographic baseline when this sample zone passes through the detector. A separate concentrator column is the method most commonly used to extend the working range of ion chromatography to significantly lower levels. A concentrator column is a short column (typically 35–50 mm in length) placed in a valve just before the analytical column. Sometimes a guard column containing identical or similar material to the separation column is used as the concentrator column. The function of a concentrator column is to strip ions from a relatively large volume of an aqueous sample of very low ionic content. A valve arrangement enables the sample to pass through the concentrator column directly to waste. Then a valve is switched and the ions taken up by the concentrator column are swept into the analytical column by the eluent stream where they are separated chromatographically. The advantage of this system is the ability to perform routine analyses for ion concentration levels at low lg L–1 (ppb). Although several valve arrangements may be used, the simplest configuration is illustrated in Figure 13.1. In the load mode the sample flows through the concentrator column and out to waste (7, 8, 4, 3 sequence). Simultaneously, the eluent bypasses the concentrator column and flows into the analytical column (1, 2, 6, 5 sequence). In the inject mode the valve is switched so that the eluent flows through the concentrator column in the opposite direction to sample loading and into the analytical column (1, 4, 8, 5 sequence). Simultaneously, the sample stream is directed to waste (7, 6, 2, 3 sequence). Sample introduction may be by a small pump or with a manual syringe.
13.3 Preconcentration
Figure 13.1 Configuration for a Dionex Low Pressure 4-Way Valve and a Concentrator Column (Courtesy Dionex Corp).
The sample breakthrough volume from the concentrator column needs to be measured in order to know how large a sample may be used. The sample must be of low ionic strength (<50 lS), otherwise the sample itself can act as an eluent for the sample ions. A good discussion of the use of concentrator columns in IC is available [4].
327
328
13 Sample Pretreatment
13.4 Sample Pretreatment 13.4.1 Anions in Acids
Samples that do not require preconcentration may contain substances that may adversely affect chromatographic performance. These substances may mask peaks of interest or be irreversibly retained, permanently damaging the column. In such cases sample pretreatment is necessary before an IC separation can be attempted. Determination of trace inorganic constituents in concentrated reagents is important in a variety of chemical and semiconductor processes. Direct determination of inorganic anions in acetic acid, hydrofluoric acid and formic acid by IC often gives high blanks and poor sensitivity. A method called ‘matrix elimination’ introduced by Siriraks and coworkers has been more successful [5]. A method was developed so that at low pH the weak matrix acids were protonated and passed unchanged through an anion-exchange column of moderate capacity. Then, a methanol–water wash was used to remove the last traces, and finally the retained inorganic anions were separated and determined by IC. Excellent recoveries were obtained for traces of sulfate, chloride, nitrate and bromide in the concentrated acids. 13.4.2 Neutralization of Strongly Acidic or Basic Samples
Determination of small concentrations of anions in samples containing 1 M sodium hydroxide and determination of trace cations in strong mineral acids are two examples where sample neutralization is needed. However, chemical neutralization with HCl or NaOH would introduce a high concentration of unwanted ions. A better way is to introduce the H+ or OH– needed for neutralization through an ion-exchange membrane. This process is called electrodialysis. Haddad, Laksana and Simons [6] described a device for off-line neutralization of strongly alkaline samples. The method uses an electrodialysis cell comprising three compartments separated from each other by cation-exchange membranes (Figure 13.2). Cations can pass through the membrane but anions cannot. The cathode compartment of the cell typically contained 20 mL of 0.1 M sodium hydroxide, the anode compartment 10 mL of 0.05 M sulfuric acid, and the sample compartment 2 mL of highly alkaline sample. Application of power (< 3 W) resulted in currents of 100–200 mA. As shown in the figure, water is reduced to H2 + OH– in the cathode and is oxidized to O2 + H+ in the anode compartment. To complete the electric circuit, cations must be carried from the anode through the sample solution to the cathode. Hydrogen ions entering the sample compartment will be neutralized by OH– to form water, and Na+ from the sample compartment will move across the membrane into the cathode compartment. The net
13.4 Sample Pretreatment
Figure 13.2 Schematic diagram of the electrodialysis process (from Ref. [6] with permission).
result is that both Na+ and OH– are removed from the sample without disturbing other sample anions. Devices are commercially available that maintain sample integrity and trace detection limits while eliminating interfering sample matrices. Dionex offers an Auto Neutralization Module that neutralizes concentrated acid with electrically generated hydroxide ions [7]. Workers at Alltech [8] have described two on-line methods for sample preparation. An ERIS Autosuppressor removes high concentrations of metal ions in samples to be analyzed for organic acids. It contains a cation exchanger that retains metal ions while allowing organic acids to pass through the cell. The ion exchanger is electrically regenerated. The device is constructed so that one cell is regenerated while the other cell is in use. Then the cells are reversed for the next run. Another device, the Alltech SCAN Sample Processor, is designed both for sample concentration and neutralization. The cell is packed with either anion- or cationexchange resin, depending on the application. 13.4.3 Dialysis Sample Preparation
In sample preparation by dialysis, sample ions transfer or migrate across a membrane from its original matrix into a receiving solution. Then, the receiving solution containing the sample ions is injected into the IC and the ion analysis is performed. An attractive feature of the dialysis process is that it isolates the sample matrix from the sample ions that are injected onto the analytical column. Some applications of dialysis sample pretreatment include the determination of ions in
329
330
13 Sample Pretreatment
milk and dairy products or canned goods after being homogenized. Fruit juices containing pulp can be analyzed for ionic content. Depending on the type of membrane used, dialysis may be either passive or active (Donnan type). Passive dialysis employs a neutral porous membrane with a fine pore structure. This type of dialysis is used primarily to remove sample matrix materials such as proteins and/or fiber or other solids [9, 10]. Donnan dialysis can do the same, but can also perform preconcentration of the ions in a sample. In Donnan dialysis, an ion-exchange membrane is used to exchange sample ions for receiving solution ions [11–14]. Each type of dialysis is described below.
13.4.3.1 Passive Dialysis This type of dialysis provides an attractive method for removing any undesired, solid or suspended material from an IC sample that would normally be difficult to filter. The technology is based on a hydrophilic membrane containing small pores through which ions and other small neutral material can migrate, but larger particles cannot. A typical membrane is a hydrophilic polysulfone ultrafilter designed to filter materials in the middle molecular weight range of 500–2000. The membrane may be of the hollow fiber or sheet type. In the first type, the sample is pumped into the hollow fiber and ions migrate out of the fiber into the receiving solution located on the outside. Generally, both the sample and the receiving solution are nonflowing or static. Older dialysis procedures based on static sample and receiving solutions were slow and required relatively large samples. These procedures have usually resulted in severe dilution of the sample. Newer dialysis methods are based on a flowing sample stream and a static receiving solvent (usually water) [10]. In this type of apparatus, a sheet membrane is used to separate sample and receiving channels. A typical static receiving channel volume is 240 lL. The flowing sample stream channel has a similar volume, but the sample typically flows at a rate of 0.8 mL min–1. The sample continues to flow until the ionic concentration comes to equilibrium. After dialysis, typically 20 lL of the receiving solution is injected into the IC. Using this method, 100% of the sample ionic concentration can be transferred to the acceptor solution. The sample solution will come to equilibrium with the receiving side so that concentrations of ions are on both sides of the membrane are equal. Calibration can be carried out easily with external standards or, if desired, internal standards. The sample must be a liquid or a homogeneous suspension before it is introduced into the dialyzer, heavy solids being first separated from the sample by centrifugation. The time required for the dialysis is about the same as the IC run time, so the dialysis time can be overlapped with elution of the previous sample. This type of system lends itself to fully automated sample handling.
13.4.3.2 Donnan (Active) Dialysis Donnan dialysis provides an alternative to column ion-exchange, filtration, or precipitation methods that are used to preconcentrate ions and/or remove interfering
13.4 Sample Pretreatment
ions. In Donnan dialysis, an (aqueous) sample is separated from a receiving electrolyte by an ion-exchange membrane. Only the desired ions and a small amount of neutral molecules go from the sample into the receiver. Depending on the charge of the membrane, either anions or cations can travel across it. An anionexchange membrane with fixed positive charges will allow the transport of anions across it, but will not allow the transport of cations. The reverse is true for a cation-exchange membrane, i.e. transport of cations is allowed and transport for anions is not. There are several processes that are occurring in Donnan dialysis sample preparation. One process involves the normal transport ions of like charge across the membrane. At the start of dialysis there is a strong tendency for the any particular ion to diffuse from a high concentration zone on one side of the membrane to the low concentration zone on the other side. As this process occurs, corresponding transfer of the same charge ion travels back across the membrane in order to preserve electroneutrality. The counter ion does not travel across so there can exist different concentrations of ions on each side. The example of 90% transport of a sample (Na+) Cl– at 0.0010 N and a receiving electrolyte (Na+) OH– at 0.050 N at 1 mL each is shown in Figure 13.3. Since NaOH is at the higher concentration, the driving force is to make the ionic concentrations equal on both sides. But Na+ is prevented from traveling across the membrane. Instead, most of the Cl– is replaced by OH–. The Cl– is never completely recovered on the receiver side, but most will travel across because the OH– is at a much higher concentration. Since OH– is at a much higher concentration, its dilution by transport to the sample is not greatly affected. The exact amount of sample ion transported across is affected by the contact time, the mixing of the solutions and probably other factors. That is why standards used to calibrate the instrument should undergo the same process as the
Figure 13.3 In this example, Cl– dialyzes into a 0.050 N OH– receiving solution where the sample and receiving volume are equal. 90% of the sample Cl– enters the receiving solution and a corresponding amount of OH– replaces it in
the original sample solution. The counterions, such as Na+ or K+, do not transport across the membrane. The exact amount of Cl– that transports across is time-dependent.
331
332
13 Sample Pretreatment
Figure 13.4 In this example, Cl– dialyzes into a 0.050 N OH– receiving solution where the sample volume is ten times greater than the receiving volume. As in the example of Figure 13.3, 90% of the sample Cl– also enters the receiving solution and a corresponding amount of OH–
replaces it in the original sample solution. But because of the volume differences, the concentration of Cl– is increased in the receiving solution by a factor of 10 compared with the example given in Figure 13.3.
sample. The use of internal standard spikes added to the sample is the best way to ensure proper calibration. Now, consider the case where the volume of the sample is 10 times greater than that of the receiver (Figure 13.4). There is ten times more Cl– available for transport, and in this example 90% of the chloride is transported. The receiver concentration of the OH is reduced to a greater extent, but the benefit is that the Cl– transported across is increased by a factor of 10. Finally, consider the case where the sample is flowing past the membrane and the receiver remains static. In this case, only the volume and concentration of the sample and the concentration of the receiver solution limit the concentration effect of the sample into the receiver solution. Obviously, sample anions can only be transported across the membrane if there are receiving ions available for transport back across. Electric neutrality must be preserved. It is of benefit to the Donnan dialysis process to have a high receiving ion concentration. But depending on the solution type and concentration, subsequent injection of this high ion concentration solution (that now contains the sample) can interfere with the ion chromatography analysis. Ideally, the receiving solution type should be the same as the eluent used for the IC analysis. A sodium carbonate/bicarbonate receiving solution is commonly used for suppressed IC. Also, it is possible to remove the receiving solution background ions prior to injection. This is accomplished by treating the receiving solution containing the sample anions with an H+-form column or membrane. This converts all of the anions to the acid form and effectively removes the carbonate/bicarbonate through conversion to carbon dioxide. The process gives high enrichment factors and wide linear dynamic range, and is free of many sample matrix effects.
13.4 Sample Pretreatment
Electrochemical dialysis is a refinement to the dialysis method. The transfer of ions through a membrane is driven by the application of an electric field. 13.4.4 Isolation of Organic Ions
Organic compounds that are ionic or that become ionic at certain pH values can be isolated selectively from neutral organic compounds by SPE with ion-exchange materials [2]. Amines become protonated cations in acidic solution and are retained by a short cation-exchange column in the H+ form. Actually, the cation exchanger or exchangers can convert a neutral amine to the protonated cation. RNH2 + catex-H+ → Catex-RNH3+ Neutral organic compounds that cannot exist as cations may be retained by physical adsorption but can be washed off the cation-exchange column by a brief rinse with an organic solvent. The amine cation can then be eluted from the column with a 1 M solution of trimethylamine in methanol. The trimethylamine converts the amine cation to the free amine, which is no longer retained by the cation exchanger. Because of its volatility, trimethylamine is easily removed from the eluate. After acidification, the sample amines can be separated by cation chromatography. The cation-exchange resin used in this operation should be a macroporous rather than a microporous polymer to minimize physical adsorption. An intermediate exchange capacity (∼0.6–1.2 mequiv g–1) is required and the column dimensions should be small to avoid use of an excessive amount of trimethylamine–methanol solution to neutralize the resin’s H+ capacity in the elution step. An Empore ion-exchange disk inserted into a tube is ideal for this use. In a similar manner, carboxylic acid anions, phenolates, etc. can be separated from other organic matter by retention on a small anion exchanger in the OH– form. RCO2H + Anex–OH– → Anex–RCO2– + H2O After a brief rinse the retained sample anions are eluted with a solution of 1 M HCl in methanol. Again, the ion-exchange column should be small and the resin should be a macroporous polymer of ∼1 mequiv g–1 exchange capacity.
333
334
13 Sample Pretreatment
References [1] D. L. Mayer and J. S. Fritz, Semi-micro
solid-phase extraction of organic compounds from aqueous and biological samples, J. Chromatogr. A, 773, 189, 1997. [2] J. S. Fritz, Analytical solid-phase extraction, Wiley–VCH, New York, 1999. [3] S. J. Lue, T. Wu, H. Hsu and C. Huang, Application of ion chromatography to the semiconductor industry I. Measure of acidic airborne contaminates in clean rooms, J. Chromatogr. A, 804, 273, 1998. [4] Dionex Corp. (Sunnyvale, CA U.S.A.), The use of concentrator columns in ion chromatography, 1994. [5] A. Siriraks, C.A. Pohl and M. Toofan, Determination of trace anions in concentrated acids by means of a moderate capacity anion-exchange column, J. Chromatography, 602, 89, 1992. [6] P. R. Haddad, S. Laksana and R. G. Simons, Electrodialysis for cleanup of strongly alkaline samples in ion chromatography. J. Chromatogr., 640, 135, 1993. [7] A. Siriraks and J. Stillian, Determination of anions and cations in concentrated bases and acids by ion chromatography, J. Chromatogr., 640, 151, 1993. [8] R. M. Montgomery, R. Saari-Nordhaus, L. M. Nair and J. W. Anderson, Jr.,
[9]
[10]
[11]
[12]
[13]
[14]
On-line sample preparation techniques for ion chromatography, J. Chromatogr. A, 804, 55, 1998. F. R. Nordmeyer and L. D. Hansen, Automatic dialyzing-injecting system for liquid chromatography of ions and small molecules, Anal. Chem., 54, 2605, 1982. Metrohm literature, On-line sample preparation in ion chromatography – no problem with the novel 754 dialysis unit, Metrohm Information, 74th ed., 26, 3, 1999. J. E. Dinunzio and M. Jubara, Donnan dialysis pre-concentration for ion chromatography, Anal. Chem., 55, 1013, 1983. J. A. Cox and J. Tanaka, Donnan dialysis preconcentrator for the ion chromatography of anions, Anal. Chem., 57, 2370, 1985. J. A. Cox and E. Dabekzlotorzynska, Determination of anions in poly-electrolyte solutions by ion chromatography after Donnan dialysis sampling, Anal. Chem., 59, 543, 1987. S. Laksana and P. R. Haddad, Dialytic clean-up of alkaline samples prior to ion chromatographic analysis, J. Chromatogr., 602, 57, 1992.
335
14 Method Development and Validation 14.1 Choosing a Method
When presented with an ion analysis problem, the worker may use a logical process of decision making to devise a working method for a new type of determination. The method may be based on one already published in the literature or on a standard method published by organizations such as the EPA, AOAC, or ASTM. Or the method may be entirely new, based on analyzing the problem of sample mixture and matrix. There are probably several different methods that could be developed to solve the same problem. The decision to use one method over the other is frequently based on the availability of a certain column or detector. The method chosen may not even be the best available in a perfect world, but it may be the best given financial, time or instrumental constraints. Certain sacrifices may have to be made on sensitivity, accuracy, and the ease of analysis, e.g., the number of different runs needed. It is even possible a decision may have to be made on whether the total goal can be accomplished. The chances of success are far greater if the literature is first searched. Good research papers or published methods are based on comprehensive testing of the proposed method. Gaining access to this experience can be quite valuable when faced with a new problem. Several papers may be published on a particular topic that, together, outline the various options available. Even a paper that shows limited success is useful because it can be compared to more successful methods. 14.1.1 Define the Problem Carefully
Exactly what is the analytical problem and what is the minimum analytical information needed to provide a reasonable answer? In this connection it is well to categorize the type of analysis desired: oxyhalides in drinking water, arsenic speciation in drinking water, speciation of chromium in plating baths, etc. The second step is to describe the expected sample composition as completely as possible. For oxyhalide analysis, it is good to know which oxyhalide species are Ion Chromatography, 4th Ed. James S. Fritz and Douglas T. Gjerde Copyright © 2009 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim ISBN: 978-3-527-32052-3
336
14 Method Development and Validation
present, their estimated concentrations and what accuracy will be needed. The expected concentration of all cations and anions should be noted even if only one or the other is to be determined. Frequently, the counterion can form a complex with the sample ion of interest, thus making it less available for analysis. Also, any nonionic materials such as alcohols, polymeric material, sugars or antioxidants should be listed as these may affect the analysis. The next step is to consult the literature for information. Useful sources include books and standard methods from sources such as EPA, AOAC and ASTM, company literature, and journal articles. It is likely that other workers have faced situations similar or even identical to the problem at hand. It is interesting to note that difficult problems frequently garner more publications than easier problems. At this point, it should be possible to select an analytical method. The method selected is likely to reflect the personnel and equipment available and the worker’s own preferences. The method selected need not be ideal; a method that simply works is sufficient. It is also necessary to consider any modifications needed to apply the selected method to your own analytical problem. The final step is to confirm that the modified analytical method works for your particular situation. Does the resolution of peaks, sensitivity and reproducibility meet your requirements? Correct identification of each analytical peak is essential. A standard for each unknown must be available and tested to determine if the method works. Normally, for method development, the sample is run without the standard and then a known amount of each standard is spiked into the sample. The peak that increases in height is identified as the ion of interest. The spike is normally obtained from concentrated standards so that the volume of the sample does not change with the addition of the standard. For example, a spike of 100 mL of a standard into a 10 mL sample volume changes the volume by only 1%. The concentration of the standard solution is chosen so that the increase in signal is measurable and appropriate. If a single concentration is spiked, then choosing a concentration that is expected to double the signal is appropriate. 14.1.2 Experimental Considerations
The different properties of anions and cations in the sample will affect the method development. These include whether the ion is organic or inorganic, multivalent or monovalent, and so on. The following discussion illustrates how these parameters can become important. Anions can be conjugate bases of either strong acids or weak acids. Strong-acid anions such as chloride or sulfate exist as the charged anion regardless of the eluent pH. Anions of weak bases such as acetate or formate are controlled by the eluent pH. If the separation mechanism is by ion exchange, then the pH must be high enough to ionize the anions and allow interaction with the column. Multivalent anions such as phosphate can change the extent of interaction by changing the pH. In many cases, a slight increase or decrease of eluent pH can result in improved resolution of a strong-acid anion and weak-acid anion pair. The easiest
14.1 Choosing a Method
way to increase the eluent pH is to add sodium or lithium hydroxide (carbonatefree) to the eluent. Lowering the eluent pH can be accomplished by increasing the concentration ratio of bicarbonate to carbonate. All other things being equal, changing the eluent concentration will affect divalent ions more than monovalent ions. Increasing the concentration will shift the retention of divalent ions to shorter retention times faster than monovalent ions. The reverse is also true. Monovalent eluents are best for separating monovalent sample ions and divalent eluents are most useful for divalent sample ions. Certain commercial columns might have superior selectivity for monovalent anions and others are selective for divalent anions. Consult the column manufacturer for details of column selectivities. Detection of weak-acid anions is best by indirect conductivity detection or postcolumn reaction detection because the suppressed conductivity detection will not perform. Indirect conductivity detection is often used because the high pH used to separate the anions will also facilitate indirect conductivity detection of these anions. Chapter 4 describes a method of combining suppressed conductivity detection and nonsuppressed detection. Conductivity detection is the most popular for ion chromatography. Although UV detection is often overlooked, it can be quite powerful. Amperometric detection, for example, offers selectivity and sensitivity, in many cases unsurpassed. The optimum eluent separation pH may not be the optimum pH for detection. An anion may be separated but not detected. This is especially true for some weak-acid anions and suppressed conductivity detection. Chapter 4 discusses the use of different detectors for IC. Anions that are polarizable will interact with the column more strongly than other anions. Nitrate will frequently tail slightly because the anion will interact both with the column ion-exchange site and the backbone of the ion-exchange matrix. Iodide, perchlorate, and many organic anions can be difficult to elute from the column in a sharp peak. Using a divalent eluent anion or addition of a small amount of organic solvent such as methanol may help. An eluent gradient also frequently helps elute a range of anion types. Depending on other materials also to be analyzed, polarizable ions may be easier to separate by ion-pairing chromatography. Ion-exchange separations are more resistant to changes in sample matrix. In ion-pairing separations, the sample matrix can cause the sample retention to shift to different times. This is less likely to happen in ion-exchange chromatography. Several charts of retention times for specific columns and eluents are listed in different chapters in this book and in company literature. While new columns are introduced, these charts can still be used as tools to determine the relationship between ions. Usually a combination of the table and a chromatogram will help predict what a chromatogram should look like. Keep in mind that the weak acids are affected most by eluent pH. Divalent ions are affected most by eluent concentration. Ion exclusion is useful for the separation of weak-acid anions. The decision to use ion exclusion rather than ion exchange frequently depends on the matrix of the sample. If a mixture of weak-acid anions and strong-acid anions is to be analyzed, then ion exchange is a separation tool that will be the most effective. How-
337
338
14 Method Development and Validation
ever, if weak-acid anions are to be analyzed in the presence of large concentrations of strong-acid anions, e.g., in the determination of acetate in hydrochloric acid, then ion exclusion will be the superior separation tool. Chloride, sulfate and similar anions are not retained by ion exclusion and therefore high matrix concentrations of these anions can be tolerated. Sample preparation (Chapter 13) includes simple procedures such as centrifugation or filtration. Other more complex sample preparation procedures include passive or active dialysis, preconcentration, combustion or precipitation of matrix ions. In some cases, choosing the correct eluent/column/detector system will negate the need for sample preparations. A selective detector will be able to detect a minor analyte in the presence of other ions. Ion exclusion allows the passage of strong-acid anions prior to separation of weak-acid anions. Methods that require the least sample preparation or minimal sample preparation directly prior to injection are the most desirable. 14.1.3 Example of Method Development
In order to illustrate the method development process, a description of a problem is presented. For example, one might want to determine sulfite in wine. Sulfite is a widely used food preservative and whitening agent. There have been many reported cases of allergic reaction to the ingestion of sulfite contained in foods or beverages. Since 1986, the FDA has required warning labels on any food or beverage containing more than 10 mg kg–1 or 10 mg L–1 of sulfite. The concentration of sulfite in wine is expected to be in the low ppm range; however, the sample matrix includes a wide variety of materials including several carboxylic acids, ethanol, sugars, and many other components. A search of the literature for sulfite determinations by ion chromatography yielded several references [1–22] describing a variety of separation methods, matrixes, and detection methods. Detection methods included inductively coupled plasma coupled to atomic emission (ICP–AE) [1], amperometric [3, 5, 9] refractive index [17], fluorometric [8, 14] UV [12] and other methods. The sample type ranged from vitamins [8, 14] to photographic fixers [13], animal feed [8, 14], and food [10, 18]. The search also found a paper on capillary electrophoresis [4]. Vendors including Dionex [19], Zellweger [20], Metrohm [21], and Sarasep [22] have published separations of sulfite. The best reference found was AOAC method 990.31 [18] that was developed by Kim and Kim. The method was based on extensive work with a variety of foods and beverages. The incentive for the work was to find an alternative to the modified Monier–Williams method [23], which is time-consuming and labor-intensive. The Kim and Kim method uses ion-exclusion chromatography with a dilute sulfuric acid eluent. Detection is based on amperometric detection with a Pt working electrode. The method is quite selective. Samples are only blended with a high pH solvent to extract both bound and unbound sulfite, and then filtered and injected into the ion chromatograph. Mannitol is added to the extraction solvent to slow the oxidation of sulfite to sulfate.
14.1 Choosing a Method
After examining the rest of the literature, several observations can be made. First, the most sensitive detection method is amperometric detection with a Pt working electrode. Other detectors may work, but may not be practical. ICP–AE or fluorescence work well but may not be available to the user. Ion exclusion is the preferred separation for sulfite determinations. However, considering all of the neutral materials that are contained in wines (sugars, organic acids and alcohols), ion exchange may be considered. In many anion-exchange columns, sulfite and sulfate will coelute, and this will be a factor if conductivity or another type of general detection method is used. Information is available from column vendors on which columns will resolve sulfite and sulfate. It should be noted that the eluent pH will determine whether sulfite is an anion or neutral. This is important to consider when choosing between ion exclusion and ion exchange. Also, the optimum eluent pH may not be the optimum detection pH. The Kim and Kim ion exclusion method works well. The amperometric detection procedures used are extremely selective and sensitive. However, this detection method is not without problems. Although good results have been reported, in some cases the sample matrix material can foul the electrode leading to nonreproducible peak areas and/or loss of sensitivity. Pulsed amperometric detection (see discussion in Chapter 4) has been suggested as an improved detector for this method [19] in order to prevent deactivation of the working electrode. Modern DC amperometric detectors are computer-controlled, and oxidation cleaning and reducing potentials can be applied to the working electrode after the run is completed. So a PAD detector may not be necessary as long as a routine cleaning operation is implemented as part of the process of analysis. Sample preparation can be quite important in sulfite determination since sulfite can easily be oxidized to sulfate. So in this case, stabilization or storage of the sample is also considered to be sample preparation. For two of the references discussed, aldehyde adducts have to be formed to stabilize the sample [2, 9]. Other work describes a sulfite–disulfite equilibrium process [7]. It is important that the eluent does not change the sample during the elution process. Eluents do not normally contain oxidizing agents except for, perhaps, dissolved oxygen. Degassing the eluent will remove dissolved oxygen. Optimum sample preparation appears to include keeping the sample tightly capped and out of sunlight until just before the injection. The sample should be quickly removed and analyzed. In general, it is better not to perform extensive sample preparation unless necessary. Of course, in many cases, sample preparation is needed. Although adding mannitol to the sample extraction solvent is not essential, it is probably important to ensure good results. This is just one example of how a method may be developed. There are others. The best method for the determination of trace anions in concentrated acids either requires removing the matrix anion, selective detection, or choosing a column that has sufficient capacity and selectivity to allow the matrix to travel quickly through the column while the analytes are retained and separated. A good method for determining chloride in a 1% solution of boric acid solution is to use nonsuppressed ion chromatography with conductivity detection. An eluent is chosen
339
340
14 Method Development and Validation
with a low pH (pH 4.5: phthalate, for example) so that the boric acid passes quickly through the column and chloride is retained and detected. Using a highpH eluent would swamp the column because the borate would be ionized and retained. Developing a method requires careful thought, an open mind, and just plain hard work evaluating the literature or trying out procedures. This book has tried to present fundamental principles of ion chromatography with the hope that using this information will make IC methods more productive.
14.2 Some Applications of Ion Chromatography
In selecting or modifying a chromatographic method, several questions need to be considered. What analytical information is truly needed? Which analyses will provide a definite answer? What accuracy is required? There is no need to perform a slow, careful determination that gives a result accurate to ± 0.3% if a faster method accurate to ± 10% will suffice. Is the method selected sufficiently rugged and dependable? There are countless situations where an ion-chromatographic determination will provide a fast and satisfactory answer to an analytical problem. Some of these examples may require the development of appropriate sample pretreatment or sampling techniques. For example, how does one sample fog? IC determination of trace inorganic ions in very pure water often requires some form of sample preconcentration to achieve the requisite limits for quantitative detection. The human and economic consequences of many IC analyses are very great indeed. Thus, extremely pure water is a necessity in the manufacture of semiconductors and in the electric power industry to avoid stress corrosion of turbine blades. Food and drugs need to be free of toxic impurities. Let us take a closer look at the analytical needs of the pharmaceutical industry that are apt to require chromatographic determinations. Drugs must be assayed for potency and purity, and impurities must be identified and quantified. The manufacturing process often necessitates control of certain impurities. In human and animal studies, determination of drug metabolites is required. Here are a few specific applications where ion chromatography has been the method of choice. 1. Analysis of counter ions in the development of small-molecule drugs which are protonated bases. These include chloride, bromide, phosphate, methanesulfonate and maleate. 2. Analysis of residual ions of concern in drugs. A new drug (filed but not yet approved) uses an IC method for residual cyanide analysis with control down to single ppm (parts per million) level. Another project has used IC for analysis of residual level of trifluoroacetic acid. 3. Analysis of residual ions for process optimization during drug development. One project has a synthetic step that gen-
14.3 Statistical Evaluation of Data
erates fluoride that subsequently forms HF during the process. An IC method was developed to analyze and control fluoride down to the low ppm level in the product, thereby avoiding damage by HF to expensive equipment. 4. Amino acid analysis for biopharmaceuticals such as proteins and peptides. This is a very common analysis. Commercial instruments such as the Hitachi L-8900 are useful for dedicated analysis of amine acids by IC.
14.3 Statistical Evaluation of Data 14.3.1 Common Statistical Terms
The object of any quantitative chemical analysis is to determine the amount of a particular constituent, X, in the analytical sample. In addition to getting a numerical answer for the amount or percentage of X, the analyst must be confident that the results are sufficiently accurate and can be repeated. Simple statistical concepts are very useful in this regard. Suppose a series of measurements of a give substance is repeated several times giving the values X1, X2, X3,....Xn. The following statistical terms are commonly used: Mean; X
RX n n
Variance; Var X
(14:1 R
X X n
2
(14:2
(When n is relatively small, the number of degrees of freedom, n–1, is used in the denominator) Standard deviation; S
X
p Var
Relative standard deviation; RSD RSD
%
Sx × 100 X
(14:3 Sx X
(14:4 (14:5
When applying statistical concepts to a chromatographic determination, X is considered to be a random variable. We assume that there is no systematic error (either positive or negative); X simply varies in a random manner. Repeated measurements of X are considered to follow a Gaussian, or normal distribution. The
341
342
14 Method Development and Validation
true mean of this underlying distribution is indicated by the Greek symbol, l, and the standard deviation by r. But a precise determination of the ‘true’ mean and standard deviation would involve a very large number of measurements and is therefore considered to be impractical. What we actually do is to take a smaller statistical sample. The mean of this sample, X, and the standard deviation, Sx, are estimates for l and r. There is often some confusion regarding the symbol used for standard deviation. The symbol r or rx is often used in place of S or Sx when the measured value is the average of 10 or more values. Under these conditions S is a very good estimator for r. The term S is used when the standard deviation is not already known. A normal distribution curve can be used to determine how good X and Sx are as estimates for the true mean and standard deviation. The density function, f(x) of the normal random variable, X, with mean l and variance r2, is given by the equation f
x
eExp
1=2
X l=r p r 2H
2
(14:6
The density function, f(x), can be considered as the probability that any given value of X will have a certain f(x). (Probability ranges from 0 to 1.) The resulting curve is bell-shaped. The curve flattens out as the standard deviation, r, becomes larger (see Figure 14.1). The location on the x-axis is determined by the mean, l. Instead of dealing with a family of distribution curves of differing shape and mean, a single distribution curve can be obtained by transforming any normal random variable X to a new normal random variable Z with mean 0 and variance 1: 1 f
Z p e 2p
z2 =2
(14:7
where Z
X l r
(14:8
A curved plotted in this manner is called a standard normal distribution curve (see Figure 14.2). Because the curve is symmetrical, there tends to be a negative error for each positive error of the same value. Integration of the area under the normal distribution curve within definite limits gives the probability that a measurement lies within that area. For example, 68% of the results lie between –1r and +1r from the mean. An abbreviated list of areas as a function of Z is given in Table 14.1. This shows that 90% of the results lie between –1.64r and +1.64r and that 95% lie between –1.96r and +1.96r. The relative frequency of measurements that have a large deviation from the mean is very small. Thus, only 0.3% of the measurements are expected to fall outside the limits of ±3r.
14.3 Statistical Evaluation of Data
Figure 14.1 Effect of standard deviation on a normal distribution curve.
Figure 14.2 A standard normal distribution curve.
343
344
14 Method Development and Validation Table 14.1 Areas of the standard normal distribution between –Z
and +Z. Z
Area
0.5
0.38
1.0
0.68
1.5
0.87
1.64
0.90
1.96
0.95
2.5
0.99
3.0
0.997
14.3.2 Distribution of Means
An analytical determination usually consists of two or three replicate analyses, and the mean is reported as the ‘best’ value. Thus, the mean is our most vital concern. We want to know how the number of replicate measurements affects the goodness of the mean and how much confidence can be placed in the value we report. A single analytical measurement, X, is a random variable. When several measurements of X are taken and the mean is calculated, and then several more are taken and a second mean is calculated, we usually find that the means are not exactly the same. If we follow this procedure a number of times, we can plot a frequency distribution of means in the same way as for the distribution of individual measurements. Thus, the mean, X, is also a random variable. Perhaps the most important question in chemical measurements is how accurate the mean of several repeated measurements is likely to be. This is not an easy question to answer with absolute certainty, but by using simple statistical principles, we can get a numerical answer that is very likely to be correct. It can be shown that the variance of the distribution of sample means is equal to the variance of the underlying distribution divided by the number of individual measurements making up the sample means, n. Var
X
Var
X n
(14:9
or
r rX px n
(14:10
14.3 Statistical Evaluation of Data
In Equation 14.10, r is called the ’standard deviation of the mean.’ This is a X very significant equation because it tells us that the standard deviation of the mean becomes smaller and smaller as the number of individual measurements, n, increases. For example, the standard deviation of the mean is cut in half by a fourfold increase in n. 14.3.3 Confidence Intervals
The standard deviation of the mean is a point estimate of l. However, a point estimate does not indicate the confidence that can be placed in such an estimate. When an objective measure of reliability is required, we report a range of values rather than a single value. These interval estimates are called confidence intervals (CI) or confidence limits. The confidence interval of a mean, X, is calculated by using the following equation: Zr CI X± px n
(14:11
The confidence limit gives the range that the mean should be in a certain percentage of the time. Thus, the confidence interval at the 90% level is calculated by inserting the Z value 1.64 from Table 14.1 into the equation. The use of these statistical contents in ion chromatography is perhaps best illustrated by an example. A standard sample was analyzed many times, giving X for chloride = 8.2 mg L–1 and a standard deviation, r = 0.12 mg L–1. Compare the expected confidence intervals at the 90% levels for a similar sample when only a single analysis is performed (n =1) and when the sample is analyzed in triplicate (n=3). CI
n 1 1:64 × 0:12 0:20mg=L CI
n 3
1:64 × 0:12 p 0:11mg=L 3
When r is not known from previous experience, it becomes necessary to calculate the standard deviation, S, for a small set of measurements. Thus, S is a less precise value than r, and the confidence limit for the mean will be larger when S is used. W. S. Gosset, writing under the pseudonym of ‘Student’, worked out a ‘t’ distribution that is used for calculating confidence limits of the mean when S, instead of r, is known. To formulate the confidence limits, the standard deviation S of the sample is calculated, and a value for t is found by consulting Table 14.2. The constant t depends on the probability level and on n, the number of measurements in the
345
346
14 Method Development and Validation
sample. (Many statisticians prefer n – 1, the number degrees of freedom.) The confidence interval is calculated from equation tS CI X± pX n
(14:12
By comparing the ratio of the confidence limit when S is used to that when r used, we can get an idea of how good S is as an estimator for r as n changes. This is done at the 90% confidence level in Table 14.3. The results show that S is a poor estimator for r when n = 2 or 3. Things improve as n increases, and S becomes a very good estimator for r at n = 15 or 20.
Table 14.2 Values of t for calculating confidence intervals.
Probability level Number of measurements
90%
95%
99%
2
6.31
12.71
63.66
3
2.92
4.30
9.93
4
2.35
3.18
5.84
5
2.13
2.78
4.60
6
2.02
2.57
4.03
8
1.89
2.37
3.50
10
1.83
2.26
3.25
15
1.76
2.15
2.98
21
1.73
2.09
2.85
41
1.68
2.02
2.70
14.4 Validation of Analytical Procedures Table 14.3 Ratio of confidence interval at the 90% level. The ratio given is of the confidence interval using S to that using Z.
Number of measurements, n
Ratio
2
3.84
3
1.78
4
1.43
5
1.30
6
1.22
8
1.15
10
1.11
15
1.07
21
1.05
41
1.02
14.4 Validation of Analytical Procedures
Validation of an analytical procedure is the process by which the performance characteristics of the procedure are shown to meet the requirements for the intended application. Validation is very important to meet the requirements of regulatory agencies, such as, the FDA (Food and Drug Administration). It is also necessary to ensure that the method can work reproducibly and reliably in QC (Quality Control) laboratories for support of manufacturing. The major steps in the evaluation process are listed below, although some may be changed or eliminated based on the particular situation. 1. Accuracy. The accuracy of an analytical procedure is the closeness of test results to the true value, as determined by analysis of a Reference Standard or by comparison with the results obtained by a second, well-characterized procedure. In the assay of a drug in a formulated product, a reference standard may be prepared in the form of a synthetic mixture of the drug product components to which known amounts of the analyte have been added. In the quantitative analysis of impurities accuracy can be assessed on samples of the drug spiked with known amounts of impurities. Accuracy is calculated as the difference between the mean and the accepted true value, together with confidence intervals. It is recommended that accuracy should be assessed
347
348
14 Method Development and Validation
using a minimum of nine determinations over three concentration levels, covering the specified range. 2. Precision. The precision of a method is the degree of agreement of individual test results when the procedure is applied repeatedly to multiple samplings of a homogeneous sample. The precision is usually expressed as the standard deviation or relative standard deviation. 3. Specificity (also called selectivity). This is the ability to measure the analyte specifically in the presence of other components of the sample matrix. Selectivity may often be expressed as the degree of bias between the analysis of a potent ingredient with and without other matrix components presents. 4. Limit of Detection (LOD). This is the lowest amount of analyte in a sample that can be detected, but not necessarily quantified. Limit tests merely substantiate that the amount of analyte is above or below a certain level. LOD is usually measured at three times the noise level. 5. Limit of Quantitation (LOQ). This is a characteristic of quantitative assays for low levels of compounds in a sample, such as impurities in bulk drugs or degradation products in finished pharmaceuticals. It is the lowest amount of analyte in a sample that can be determined with acceptable precision and accuracy. LOQ is expressed as a percentage or parts-perbillion and is usually measured at 10 times the noise level. 6. Linearity and Range. Linearity means that the test results are directly (or by a well-defined mathematical transformation) proportional to the concentration of analyte within a given range. The range is the interval between the upper and the lower levels of analyte that can be determined with suitable levels of precision, accuracy and linearity. The range is typically from the achievable sensitivity to 120% of the control limit for trace analysis, or 20% to 150% of the expected level for a major component, as in counterion analysis for a drug. For products ready for approval, validation also includes the following: 7. Robustness. This is to determine how reproducibly the method can be operated by intentionally varying the parameters within pre-defined ranges. 8. System Suitability or Ruggedness. This is to evaluate how reproducible the method can be when used on different instruments by different analysts from different laboratories. Statistical analysis is used to evaluate the validation data. Many laboratories use available statistical software, such as Excel or Statistica.
14.4 Validation of Analytical Procedures
14.4.1 Analytical Control
After the initial validation of a method has been completed, it is necessary to know that the method continues to give correct results over a period of time. The use of a control chart is perhaps the simplest and commonest way to ascertain that the quality of chemical analysis is being maintained. A control chart is a chronological plot of results from periodic analysis of a reference sample. Control charts can be used to indicate any significant trends and to point out any gradual deterioration in the analytical results. This is intended to enable a difficulty to be corrected before any serious consequences occur. To set up a control chart, several replicate analyses of a standard reference material are made periodically, such as once a week. If possible, the true value of the reference material and the standard deviation of the analytical method should be known. If not already known, the correct values can be estimated accurately by repeated analysis. The reference material must be very stable and should approximate as closely as possible to the composition of actual samples to be analyzed.
Figure 14.3 A control chart.
A control chart is illustrated in Figure 14.3. In this case each periodic determination was done in triplicate; hence the mean, X, of each subset is given by: . p SX rX 3
349
350
14 Method Development and Validation
There is no discernible trend for several time periods, and the process is considered to be under control. Then there is a sudden trend toward low results, and the warning limit of ± 2r is breached. At this point, ‘troubleshooting’ is required to correct the source of the difficulty and bring the analytical process under control.
References [1] D. R. Migneault, Enhanced detection of
sulfite by inductively coupled plasma atomic emission spectroscopy with high-performance liquid-chromatography, Anal. Chem., 61, 272, 1989. [2] T. Sunden, M. Lindgren, A. Cedergren and D. D. Siemer, Sulfite stabilizer in ion chromatography, J. Chromatogr., 663, 255, 1994. [3] R. Leubolt and H. Klien, Determination of sulfite and ascorbic acid by high-performance liquid chromatography with electrochemical detection. J. Chromatogr., 640, 271, 1993. [4] D. R. Salomon and J. Romano, Applications of capillary ion electrophoresis in the pulp and paper industry, J. Chromatogr., 602, 219, 1992. [5] H. R. Wagner and M. J. McGarrity, The use of pulsed amperometry combined with ion-exclusion chromatography for the simultaneous analysis of ascorbic acid and sulfite, J. Chromatogr., 546, 119, 1991. [6] J. E. Parkin, High-performance liquid chromatographic investigation of the interaction of phenylmercuric nitrate and sodium metabisulfite in eye drop formulations, J. Chromatogr., 511, 233, 1990. [7] B. J. Johnson, Sulfite-disulfite equilibrium on an ion chromatography column, J. Chromatogr., 508, 271, 1990. [8] S. M. Billedeau, Fluorimetric determination of vitamin K3 (menadione sodium bisulfite) in synthetic animal feed by high-performance liquid chromatography using a post-column zinc reducer, J. Chromatogr., 471, 371, 1989. [9] J. F. Lawrence and F. C. Charbonneau, Separation of sulfite adducts by ion chromatography with oxidative amperometric detection, J. Chromatogr., 403, 379, 1987.
[10] J. F. Lawrence and K. R. Chadha, Head-
[11]
[12]
[13]
[14]
[15]
[16]
[17]
space liquid chromatographic technique for the determination of sulfite in food., J. Chromatogr., 398, 355, 1987 N. Sadlej-Sosnowska, D. Blitek and I. Wilczynska-Wojtulewicz, Determination of menadione sodium hydrogen sulfite and nicotinamide in multivitamin formulations by high-performance liquid chromatography, J. Chromatogr., 357, 227, 1987. R. G. Gerritse and J. A. Adeney, Rapid determination in water of chloride, sulfate, sulfite, selenite, selenate and arsenate among other inorganic and organic solutes by ion chromatography with UV detection below 195 nm, J. Chromatogr., 347, 419, 1985. J. M. McCornick and L. M. Dixon, Determination of sulfite in fixers and photographic effluents by ion chromatography, J. Chromatogr., 322, 478, 1985. A. J. Speek, J. Schrijver and W. Schreurs, Fluorimetric determination of menadione sodium bisulfite (vitamin K3) in animal feed and premixes by high-performance liquid chromatography with post-column derivatization, J. Chromatogr., 301, 441, 1984. A. Gooijer, P. R. Markies, J. J. Donkerbroek, N. H. Welthorst and R. W. Frei, Quenched phosphorescence as a detection method in ion chromatography: the determination of nitrite and sulfite, J. Chromatogr., 289, 347, 1984. W. E. Barber and P. W. Carr, Ultraviolet visualization of inorganic ions by reversed-phase ion-interaction chromatography, J. Chromatogr., 260, 89, 1983. P. R. Haddad and A. L. Heckenberg, High-performance liquid chromatography of inorganic and organic ions using low-capacity ion-exchange columns with
References indirect refractive index detection, J. Chromatogr., 252, 177, 1982. [18] AOAC Official Method 990.31 in Official Methods of Analysis of AOAC International, 16th edn., Vol. II, P. Cunniff (ed.), 1995. [19] Dionex Application Note 54, Determination of sulfite in food and beverages by ion exclusion chromatography with pulsed amperometric detection, Dionex Corp., Sunnyvale, CA, 1999. [20] Lachat Ion Chromatography Data Pack, Zellweger Analytics, Milwaukee, WI, 1999.
[21] Metrohm Application Note S-12, Deter-
mination of lactate, chloride, nitrate, sulfite and phosphate in wine, Metrohm Ltd., Herisau, Switzerland, 1999. [22] Sarasep Application Note, Determination of sulfite by ion exchange and conductivity detection, San Jose, CA 1996. [23] AOAC Official Method 962.16 in Official Methods of Analysis of AOAC International, 16th edn., Vol II, P. Cunniff (ed.), 1995.
351
353
15 Chemical Speciation 15.1 Introduction
Elements often exist in a sample as multiple chemical species. The term chemical speciation refers to the determination and measurement of the various chemical forms of an element that are present in a particular sample. Several elements such as arsenic and selenium are converted to different species during weathering or as a result of biological processes. Drinking water may contain different oxyhalide species formed through various water disinfection processes. The concentrations of these species are now regulated in drinking water in the United States and Europe. Industrial samples such as plating baths contain different oxidation states of metals. The effectiveness of chromium plating baths depends on the amounts of Cr(III) and Cr(VI) that are present. Gold plating processes involve the use of Au(I) and Au(III) cyanide complexes. Vanadium speciation is important in the recovery of sulfur from geothermal water. From an environmental standpoint, there are two reasons to perform chemical speciation. One reason is that the toxicity of an element varies with its chemical form. For example, mercury and silver form an amalgam that is used to fill tooth cavities. Ninety per cent of dentists use amalgam to fill cavities, and mercury in this form can be in the body for years with no toxic effects. Elemental mercury has low toxicity and, although this is not recommended, can be touched and handled with little chance of ill effects. However, methyl mercury, caused by industrial pollution and sometimes found in sea food, is readily adsorbed into the body and once there is extremely toxic. Pollution incidents involving mercury have shown that total metal data are insufficient and often misleading in assessing the potential hazard of this metal. The other environmental reason for measuring the type and concentration of each chemical species in a sample is to help determine the technology needed for removal or cleanup. The chemical properties, and the technology needed to remove an element, will vary with chemical species formed from that element. A tragic example of why chemical speciation is needed is now coming out of northern India and Bangladesh. Possibly the largest mass poisoning in history is occurring because of arsenic in drinking water. In the 1970s, several international Ion Chromatography, 4th Ed. James S. Fritz and Douglas T. Gjerde Copyright © 2009 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim ISBN: 978-3-527-32052-3
354
15 Chemical Speciation
agencies headed by UNICEF began providing money to drill wells to provide ‘clean’ drinking water rather than relying on surface waters that contained parasites and other harmful materials. The wells, now numbering more than 10 million, have only been discovered in the last few years to contain dangerous amounts of arsenic. According to the World Health Organization, more than half the population of Bangladesh and 1 million Indians are now being exposed. Exposure to high levels of arsenic can cause cancers of the skin, bladder, kidney and lung, and diseases of the blood vessels of the legs and feet, as well as possibly diabetes, high blood pressure, and reproductive disorders. Monitoring the effectiveness of the cleanup technologies will require measurement of the various chemical forms of arsenic in the water. The general application of ion chromatography (IC) to environmental samples has been described in a review [1]. More laboratories are looking at speciation as methods are being developed. As well as the traditional ion chromatography companies, most of the inductively coupled plasma mass spectrometry (ICP–MS) companies (PE, Agilent, Thermo, Varian) have done applications or offer speciation columns and/or kits. A website on speciation called Evisa is sponsored by the EU and provides much information (www.speciation.net). There is a links tab that lists various elemental metals, with further lists of references and certified methods on several different species of each particular element. Environmental directives and legislation about elemental speciation have been limited. Therefore analytical laboratories have often lacked the incentive to invest in the necessary technologies to perform chemical speciation. However, this situation is changing, and methods in effect include EPA Method 6800 for speciation of chromium, EPA Method 314 for perchlorate, EPA Method 317 for oxyhalide anions, EPA Method 8323 to detect organotin compounds, and several others. In order to do elemental speciation analysis, the original distribution of chemical species must be carefully preserved until the analysis is done. It can be very difficult to keep the sample stable up to the point where it is injected into the instrument. Since the analytes can be oxidized or sometimes reduced to related species within the sample, just handling the sample may result in changes in the relative concentrations of the analytes. Preconcentration of the samples is not advised, and in many cases the samples must be treated to preserve the original state. For example, EDTA may be added to complex the metals in a sample, protecting them from oxidation or reduction reactions. Nitric acid may not be used for speciation because nitric acid is an oxidizing agent. Method development for speciation involves answering several questions. Is the sample photoreducible or photosensitive? Are any analyte species volatile, thermally unstable, or unstable when exposed to air? Could redox conditions change because of sampling, sample preservation, or sample preparation? The loss or conversion of nonanalyte sample components may also affect the balance of analyte species. The reader is advised to consider the chemistry closely and follow established speciation methods to prevent incorrect measurements of the various elemental species.
15.2 Detection
15.2 Detection
There have been several reports describing how general detection methods such as conductivity or post-column reaction photometry have been used for speciation detection. EPA Method 300 describes the use of conductivity detection for the speciation of chloride and bromide in drinking water. The chromatographic conditions are much more important when general detection is used. Major interferences must elute well away from the ions of interest and all of the species should elute in the same chromatogram. The ionic form of the metal species must be either all positive or negative in order to use either cation chromatography or anion chromatography. This is not necessarily easy to arrange for some metals. Chromium can exist as either a cation or an anion depending on its oxidation state. Detectors that are selective or can be ‘tuned’ to an element are quite useful. For example, it was shown in a review by Urasa [2] on DC plasma atomic emission spectrometry (DCPAE) detection that only one of the metal species need be retained by the chromatographic column in order to perform a speciation analysis. Cr(III) exists as a cation in aqueous solution and Cr(VI) exists as an anion (CrO42–). If one is using a cation-exchange column and injects a mixture of Cr(III) and Cr(VI), then the anion, Cr(VI), elutes with the void volume and Cr(III) is retained and elutes later. Selective detection allows quantification of the Cr(VI) peak in spite of the fact that many other materials are eluting at the same time. Cr(III) is eluted later as part of the cation-exchange process and detected. The eluent conditions needed to elute this peak are not very stringent. All that is needed is to elute Cr(III) quickly in a nice sharp peak after the Cr(VI). Since the eluent driving ions and counterions are not detected, the eluent type, pH, and/or concentration can be changed as needed. Of course, the reverse could be used if an anion-exchange column is used. Cr(III) would elute quickly and then Cr(VI) would follow. Strong chelating or complexing agents may form anionic complexes with metals. These anionic complexes can be separated by anion-exchange chromatography. Continuing with the chromium example, Cr(III) can be converted to an anionic complex by complexation with EDTA. Thus, a mixture of CrEDTA and CrO42– may be separated by anion-exchange chromatography and be detected by UV detection. Thus, methods for speciation can be developed quickly and performed easily provided there is access to selective detection. However, ICP–MS appears to be the most useful because it is very selective and sensitive. In many cases, the methods described for one form of detection can be translated into other selective detection methods, provided the sensitivity is high and interference from other analytes is low. Inductively coupled plasma coupled to atomic emission (ICP–AE) is also a powerful and sensitive detection method for elemental speciation.
355
356
15 Chemical Speciation
15.3 Chromatography
Although there has been significant work performed by ion-pair chromatography, the majority of speciation separations are performed by ion-exchange chromatography. The ion-exchange separation of metal species in a chromatographic system can be performed on the basis of either of two concepts. Sample species can be separated on a column on the basis of affinity differences between the species for the column. Sample species may also be separated by a complexing reagent in the eluent. The complexing reagent changes the form of the sample species, allowing it to move down the column more easily. These concepts are sometimes called the ‘push/pull’ mechanism of ionexchange chromatography [3]. The eluent that competes with sample species for ion-exchange sites and elutes the sample from the column operates under a ‘push’ mechanism. The eluent ‘pushes’ or elutes the sample species from the column. Adding a complexing reagent to the eluent and converting the sample species into a complex that is not as well retained by the ion-exchange column is called the ‘pull’ mechanism. The complexing agent ‘pulls’ the sample species down the column. The effect can be shown by equations. It has been shown that the ion-exchange reaction of the eluent and sample species competing for the ion exchangers can be written as an equilibrium equation. By chromatographic theory, the equilibrium equation can be rearranged to predict behavior in ion-exchange chromatography. The result is shown by the equation: log k = a/e log C – a/e log E + 1/e log Keq – a/e log e + D where k is the sample capacity factor, a is the charge of the sample ion, e is the charge of the eluent driving ion, C is the resin capacity, E is the eluent concentration, and Keq is the static solution ion-exchange equilibrium constant. The constant D includes the terms column void volume and resin density. Examination of the equation shows that controlling separations (changing log k or retention) by affinity differences is not very powerful. Options for controlling the separation are limited. The resin capacity, sample ion charge, and affinity of a sample (Keq) for a particular column are usually constant or not easily changed. The only real variables are the eluent type (affinity of eluent ion for the ion exchanger), eluent ionic charge, and eluent pH and concentration. Some retention crossovers can be achieved by the use of different eluent types and variations in the eluent concentration. However, the addition of a complexing agent to the eluent makes ion-chromatographic separations really powerful. The complexing agent controls the amount of ion species available to compete with the eluent for the ion exchanger. In effect, the Keq term is changed (usually lowered). In cation chromatography, Keq (and log k) is usually lowered because neutral or anionic complexes are formed. In anion
15.4 Valveless Injection IC
chromatography, Keq and retention are usually increased. The complexing agent is usually specific for a metal or group of metals. Thus, a particular separation can be achieved by choosing the type, concentration and pH of the complexing agent added to the eluent.
15.4 Valveless Injection IC
Normal IC separations rely on a high-pressure eluent pump, injection valve, column and detector. However, a novel method for speciation was developed by Gjerde and Wiederin [4], in which a low-pressure column method with valveless injection was employed. The system configuration is shown in Figure 15.1. An example of a separation of Cr(III) and Cr(IV)using this type of IC is shown in Figure 15.2. Nitric acid eluent is acceptable in this case because it does not oxidize Cr(III). The separation method is anion exchange, so Cr(III) is not retained and Cr(IV) is eluted by the nitrate driving anion. The sample probe is moved between eluent and standard or sample solutions using an autosampler. Some wells in the autosampler contain samples to be injected. Other wells contain standards of known concentration. Still others contain one or more eluents of different concentration. A peristaltic pump delivers the eluent and sample to the low-pressure column. Because a peristaltic pump is a flow-through device, the integrity of the sample is maintained as it is passed through the pump. However, because peristaltic pumps can only pump at a maximum of 100 psi, the columns must operate at a low backpressure. Development of a low-backpressure column depends on several factors. First, the ion-exchange resin must be small, uniform and efficient. The column hardware and fluid connections must be well designed and not contain flow paths
Figure 15.1 Schematic representation of a valveless, low pressure column, IC with IC–MS detection. The low backpressure column allows use of a peristaltic pump. The sample is loaded
through the autosampler and peristaltic pump. Then a gradient of eluents are introduced step by step as the autosampler probe takes up liquid from each eluent vial.
357
358
15 Chemical Speciation
Figure 15.2 Separation of Cr(III) and Cr(VI) by valveless IC and ICP–MS detection. The separation is affected by both the volume of sample introduced to the column and the speed of the peristaltic pump. The eluent was 0.35% (w/w) nitric acid adjusted to pH 1.6
with ammonium hydroxide. The column is a low-capacity anion exchanger ANX3202 with dimensions of 3.2 mm × 20 mm. Detection limits for Cr(III) and Cr(VI) are both <0.1 ppb (courtesy of Transgenomic, Inc., Omaha, NE).
that broaden the sample peaks. The column should be short so that the eluent backpressure is low. Finally, the detector must be able to operate at the eluent flow rates that are compatible with the column. Standard analytical separation columns probably operate at too high a backpressure to be used in this type of system. However, guard columns often contain the same separation material as an analytical column and have much lower backpressure. To a certain extent, the flow rate can be reduced to lower column backpressure. The injection process is as follows. During the process of injection, the probe is moved from the eluent well into the sample well for a fixed period of time and then back to the eluent well. Because a finite amount of time is needed to travel from the eluent to the sample and back again, an air segment is introduced on each end of the sample. This does not harm the separation as long as the time interval is not very long. A longer sampling time gives higher sensitivity, but worse chromatographic resolution. A shorter sampling time may be used for high sample concentrations. The same sampling time should be used for both samples and standards. One of the advantages of the valveless mode is the ability to perform step-gradient elution. In this case the autosampler vials contain eluents of increasing concentration. An eluent step gradient is formed as the autosampler ’samples’ in turn each eluent vial. Since ICP–MS detection is used, a nonanalyte element may be spiked into the eluent to provide a known, uniform signal for instrument optimization. Alternatively, the column may be removed from the sample introduction apparatus to allow direct pumping of optimization solutions to the detector. Another advantage of removing the column is that the total element content of a particular species can be determined and compared to the total amount of the various species. This could be done automatically with a six-port, two-position switching valve where
15.5 Speciation of some Elements
the column is positioned in one of the bypass loops. Each sample could be analyzed with and without the column inline. This, in fact, brings up an interesting point. Essentially, the only difference between this type of chromatograph and a normal ICP–MS instrument is the presence of a column and the types of solvents (eluents) used to introduce the sample. This point may help the valveless IC gain acceptance by the analytical spectroscopist.
15.5 Speciation of some Elements
In some instances it is only necessary to determine the metal by any appropriate analytical method. However, speciation by IC is often needed when the particular form of a metal needs to be known. The basic chemistry of several important metal elements is reviewed briefly in the following sections. 15.5.1 Chromium
Chromium exists primarily as Cr(III) and Cr(VI). Speciation is very important because Cr(VI) is far more toxic than the trivalent species. The actual form of hexavalent chromium is pH-dependent. Between pH 2 and 6 the yellow chromate, HCrO4–, and the more intensely colored orange-red dichromate, Cr2O72– , are in equilibrium. In very dilute aqueous solutions most of the chromium is present as the hydrogen chromate ion, but dichromate is the predominant species around 0.1 M. In basic solution, HCrO4– is the major species. Under strongly acidic conditions, only dichromate ion exists. Cr(III) species in acidic solution exist as Cr(H2O)63+ ions and in concentrated alkali have been identified as Cr(OH)63– and Cr(OH)5(H2O)2–. Chromium metal is produced by roasting the chromate ore. The largest single use of the metal is as a corrosion inhibitor in alloying steels. Chromium finds its way into the environment through industrial wastes from electroplating sludge, tannery wastes, the manufacture of corrosion inhibitors, and municipal sewage sludge. These sludges are of particular concern. Cr(III) is essential to human nutrition and is present in most soils, while Cr(VI) is only occasionally present. Cr(III) is less toxic and less mobile (in the environment) than Cr(VI). Cr(VI) compounds are very toxic to aquatic plants and animal life as evidenced by their widespread use as algicides. Cr(VI) toxicity may manifest itself in the form of skin ulceration, nasal perforation, and lung cancer. There have been several reports [5–9] that describe the speciation of chromium. Many of these reports described the use of DCPAE, ICP–AE and atomic absorption (AA) selective-type detection. UV–Vis detection has also been extensively used. Atomic emission detection has been found to be very promising because of its selectivity and low detection limits.
359
360
15 Chemical Speciation
The basis of many of the separations has been to convert Cr(III) to an anion by adding a complexing agent. Cr(VI) is already an anion (usually CrO4–); hence, anion chromatography can be used to separate mixtures. An example is separation carried out in the presence of KSCN, converting Cr3+ to Cr(SCN)4–. Pyridine dicarboxylic acid (PDCA) was used as a complexing agent for Cr3+ to form anionic Cr(PDCA)2–. The sample pH is critical to the separation. Optimum results were reported at pH 6.8. Both Cr(III) and Cr(VI) were detected by UV–Vis detection and diphenylcarbohydrazide as a post-column reagent. The method was found to be applicable for plating baths and waste-water analysis. Geddes and Tarter [8] reported the use of EDTA eluent and UV detection for Cr speciation. Two silica-based anion-exchange columns were connected in series. However, the chromatographic peaks were very broad and the method had poor detection limits. Much better chromatographic separations were shown by Powell et al. [10]. Spiking Cr(VI) into industrial waste-water samples showed no initial concentrations of Cr(III), but after 3 weeks, 30% of the Cr(VI) had converted to Cr(III). Other work [7] shows that nitric acid is an effective preservative for determination of chromium in surface water. At ambient pH, the unpreserved water exhibited over 50% loss of Cr(III) within 1.5 h due to precipitation. With nitric acid preservation, both chromium species were preserved in solution for a week. Care must be taken when spiking standards into any natural water sample. Organisms and suspended solids present in the water are easily oxidized by Cr(VI). All possible sample matrices cannot be anticipated, so acid (or other) preservation cannot be recommended as a general strategy. The best sample handling protocol for specific needs and specific matrices must be determined during method development. Prompt analysis of the sample is often the best approach for accurate determinations. 15.5.2 Iron
Iron is the second most abundant metal after aluminum. The highest oxidation state of iron is VI, although, of course, II and III oxidation states are the most common. Fe(II) forms a variety of complexes. In aqueous solution Fe(II) exists as Fe(H2O)6+, which is pale sea-green in color. Fe(II) is slowly oxidized in acid. Most Fe(II) complexes are octahedral. Perhaps the most important ferrous iron complex to humans is heme, which exists in hemoglobin. In aqueous solution at pH 5 or higher, Fe(III) has a strong tendency to hydrolyze and precipitate. In acidic solution Fe(III) is apt to be present as a mixture of hydrated and hydroxyl positively charged complexes. In the presence of complexing anions such as chloride, the hydrolysis of Fe(III) is even more diverse, and mixtures of chloro, hydrated, and hydroxy species are formed. Fe(III) complexes are mainly octahedral, like Fe(II) complexes. Fe(III) has a greater affinity for oxy ligands, whereas Fe(II) has a slight preference for ligands containing nitrogendonor atoms.
15.5 Speciation of some Elements
The simultaneous determination of Fe(II) and Fe(III) is important to understanding the environmental redox processes in biological systems. Iron activity affects several chemical processes in natural waters and its speciation concentration is a significant factor in the evaluation of water quality. Iron speciation of Fe(II) and Fe(III) is reported more often than any other speciation. The methods are based on cation exchange, anion exchange, and ion-pairing chromatography. Only a few of the methods are discussed here. Saitoh and Oikawa [11] simultaneously determined Fe(II) and Fe(III) by post-column reaction (PCR) detection. The PCR reagent was bathophenanthrolinedisulfonic acid and ascorbic acid. This procedure was found to be successful in spring water samples. Moses et al. [12] reported the determination of Fe(II) and Fe(III) in water samples. The detection system consisted of PCR reaction with PAR detection reagent. In this work, the Fe(II)/Fe(III) ratio increased at times – probably because Fe(III) was photochemically reduced. The presence of trace iron contaminants in gold plating baths can cause brittle deposits. Fe(CN)4– and Fe(CN)63– were reported to be separated by anion chromatography [13]. These complexes are multivalent and difficult to elute. Small amounts of Na2CO3 added to the mobile phase sharpened the peaks. 15.5.3 Arsenic
Arsenic is found in igneous and sedimentary rocks. The most common commercial source is as a by-product from the refining of copper, lead, cobalt, and gold ores. Although arsenic is actually a metalloid, it is grouped with metals for most environmental purposes. Arsenic chemistry is complex, involving a variety of oxidation states, both as anionic and cationic species, and both inorganic and organometallic compounds. Of these, III and V are the most common oxidation states. The oxidation states of arsenic change easily and reversibly. As(III) is commonly encountered as the arsenite ion, H2AsO3–. Arsenious acid is a weak acid, pKa1 = 9.2, pKa2 = 13. As(V) exists as H2AsO4– in aqueous solution. Arsenic acid is a weak tribasic acid. Its dissociation constants, pKa = 2.3, 6.8, and 11.5, are similar to those of phosphoric acid. Oxidation of As(III) to As(V) by dissolved oxygen is slow at neutral pH, but is much faster at either extreme [14]. In reducing environments, As(III) is produced, but As(V) is the more stable state in aerobic environments. As noted in the Bangladesh example in the introduction, arsenic is extremely poisonous. Most arsenic compounds are highly toxic, causing dermatitis, acute and chronic poisoning, and possibly cancer. Arsenic is found in virtually all soil and other environmental matrices [15]. Arsenic is present in coal, pesticides, preservatives, etc. Arsenite, a commercial form of arsenic, is one of the most toxic forms of arsenic. Arsenic speciation by IC can be simple and reproducible. A conductivity detection method of As(IV) was reported by McCrory-Joy [16]. The procedure showed no interference from ions such as nitrate. As(III) was separated but not detected
361
362
15 Chemical Speciation
Figure 15.3 Separation of dimethylarsinic acid, arsenic(III), methylarsonic acid, benzenearsonic acid and arsenic(V) using the AS7 anion column with an ammonium phosphate gradient and ICP–MS detection (courtesy of Dionex, Inc. Sunnyvale, CA).
because suppressed conductivity detection cannot detect the weak acid. An IC method for the separation of several species of arsenic (100 ppb in 1% NaCl) is shown in Figure 15.3. The method employs the Dionex IonPac AS7 anionexchange column (Dionex, Sunnyvale, CA) and ICP–MS detection. Arsenobetaine is considered relatively nontoxic but is a common form of arsenic in some foods such as seafood. Slingsby and coworkers developed a dual column selectivity method and electrospray mass spectrometry detection to determine several common arsenic species at high sensitivity [17]. Both anionic and cationic species of arsenic were determined. 15.5.4 Tellurium
Tellurium is used in the metallurgical industry as an alloy constituent. Tellurium improves the acid resistance of lead used in batteries. It is also used in the manufacture of heat- and abrasive-resistant rubber. Tellurium is obtained as a by-product in the electrolytic refining of copper. In the semiconductor industry, the ultratrace level determination of tellurium in tellurium-doped single crystals is often required. Tellurium species are highly toxic. Tellurium and selenium resemble each other chemically. The analytical chemistry of both elements is usually presented together. Tellurium and selenium are commonly found in oxidation states II, IV, and VI, as well as in the elemental forms. Oxidation state IV is the most stable. In aqueous chemistry, tellurium is mainly found as telluride (Te2–), tellurite (TeO32–), and tellurate (TeO42–). Zolotov et al. [18] developed an IC procedure for the separation of TeO32–, and TeO42– by a suppressed ion-chromatographic system. In the method, F– interfered with the determination of TeO32–, while SO42–
15.5 Speciation of some Elements
interfered with the determination of TeO42–. In a report by Chen et al. [19], a similar method was used to speciate tellurium. The detection limits for TeO42– were very poor in both methods. The studies were confined only to standard solutions. 15.5.5 Selenium
Selenium is widely dispersed. It is found, for example, in igneous rocks, volcanic sulfur deposits, hydrothermal deposits, and copper ores. Selenium is used in the electronics industry for the manufacture of rectifiers and photoconductivity cells. Selenium and its compounds are also used as additives in chromium-plating, glass, ceramics, pigment, rubber, photography, lubricants, pharmaceuticals, and organic substances. Selenium is both a toxic and an essential element. The toxicity depends greatly on the species. Selenium is a cumulative toxic substance and can be a serious health hazard when present in high concentrations in food and water. However, at very low levels, (lg kg–1) it is recognized as an essential trace element in animal nutrition. There has been increasing interest in the determination of selenium at trace levels in a wide variety of matrices. Selenium determinations of environmental samples have become quite important. Large-scale poisoning of waterfowl has occurred in several watershed areas of central California. Selenium is washed or leached into these areas, helped through the widespread irrigation of seleniumcontaining farm soil. Selenium is commonly found in oxidation states II, IV and VI, and as an element. Se(VI) is much more stable than Se(IV). The reaction chemistry of selenium is mainly that of selenide (Se2–), selenite (SeO32–) and selenate (SeO42–). Selenious acid is a weak acid with pKa values of 2.6 and 8.3. Selenious acid and selenite are much stronger oxidants than sulfurous acid and sulfite. Thus, many of the characteristic reactions are related to redox reactions, in which Se(IV) can be reduced to elemental selenium. The acid strength of selenic acid (H2SeO4) is similar to that of sulfuric acid. Selenium can exist in at least two different ionic forms in environmental samples: selenite (SeO32–) and selenate (SeO42–). The concentration and speciation of selenium in a given sample depend on the pH and redox conditions, the solubility of its salts, the biological interactions, and the reaction kinetics. For example, in seawater, SeO32– is the dominant species, but in river or tap water, selenium can be found as roughly equal amounts of Se(IV) and Se(VI) [20]. Zolotov et al. [18] developed a suppressed IC method for the simultaneous determination of SeO32– and SeO42– in the presence of F–, Cl–, NO3–, HPO42– and SO42–. The separation took about 30 min. The method was applied to river and tap water. Sensitivity and precision were good. Hydrogen peroxide was used in the sample pretreatment to decompose organoselenium compounds so that total selenium could be determined.
363
364
15 Chemical Speciation
The determination of SeO32– and SeO42– in water with graphite furnace atomic absorption detection was investigated by Chakraborti et al. [21]. Some interference by Cl– and F– was reported. DCPAE detection was used by Urasa and Ferede [22]. Results were 1000 times more sensitive than conductivity detection. One of the advantages of atomic emission detection was described in this work. Identical molar sensitivity was obtained for both species. Mehra et al. [23] developed a novel single-column IC method to determine selenium species in seleniferous soil samples. The separation took about 14 min and there were no reported interferences. 15.5.6 Vanadium
Vanadium has a relative abundance of about 0.02%. Oxidation states of V to I are known. Vanadium solutions generally contain several species in a complicated series of equilibria. V(V) and V(IV) are both stable, with the former mildly oxidizing and represented mainly by oxy species. Pervanadyl ion (VO2+) is a major species in strongly acidic solutions, while in strongly basic solutions the mononuclear vanadate, VO43–, exists. V(IV) ions are stable in acid, and give blue solutions of vanadyl ion (VO2+). A number of anions of the IV oxidation state are known, including VO32–, and V4O92–. These ions are stable under alkaline conditions. Much of the analytical chemistry of vanadium is concerned with its use in ferrous and nonferrous metallurgy. Vanadium also finds application in catalysis and in the paint and ceramic industries. Environmental concerns about vanadium arise primarily from air pollution problems. Vanadium can be released from fly ash and oil combustion products. There are only a few references on vanadium speciation. One reference reported the simultaneous determination of V(IV) and V(V) [24]. Postcolumn reaction with PAR resulted in detection limits of about 10 ppb, even in the presence of high concentrations of phosphate. Unfortunately, the studies were not carried out in samples. Urasa et al. [2] used DCPAE detection to speciate VO2+ and another vanadium species thought to be VOCl42–. 15.5.7 Tin
Tin forms two stable inorganic species of Sn(II) and Sn(IV). Sn(II) is added to tin/ lead alloy plating baths. The Sn(II)/Sn(IV) ratio is important to the plating bath performance. An IC separation was carried out with the use of 0.3 mM HCl eluent [25]. Neither Sn(II) nor Sn(IV) were strongly retained by the cation-exchange column used in this work. Inorganic tin speciation is quite difficult, because Sn(II) hydrolyzes easily at neutral and alkaline pH. Tin has a strong tendency to form organometallic complexes. Organotin compounds are used in marine antifoulant agents. Of course, there can be many tin species when combined with organics. MacCrehan [26] reported the separation of n-Bu3Sn+, Et3Sn+ and Me3Sn+ by cation chromatography and differential pulse amperometric detection. Organotin compounds tend to foul or poison the work-
15.5 Speciation of some Elements
ing electrode surface. The differential pulse technique used here eliminated this problem through the reoxidation of the reduction products on the electrode surface. The speciation of tin in natural water was performed by Ebdon et al. [27] by AA detection. Sn(II), Sn(IV) and Bu3Sn+ were separated. An extraction/preconcentration sample preparation procedure was used. Jewett and Brinckman [28] used graphite furnace AA (GFAA) to detect several diorganotin and triorganotin species. Cation-exchange and reverse-phase chromatography were used in the several samples listed. Chromatographic separation of both dialkyltin and trialkyltin species appear to follow cation-exchange separation mechanisms. 15.5.8 Mercury
The harmful effects of mercury and organomercury species are well known. Disastrous effects, both on personal levels and large population levels, have resulted from exposure to certain mercury species. A method for Hg(I) and Hg(II) was reported which used on-column derivatization with diethyldithiocarbamate complexing agent. This process of using a complexing reagent seems appropriate for this speciation determination. In this way, the mixture can be ‘locked in’ before the chromatography takes place. This is particularly important for metals that change oxidation states easily. A reverse-phase column and UV detection at 350 nm was employed [29]. Mercury compounds all appear to absorb UV light. The 254 nm detection has also been used with reasonable sensitivity and selectivity. ICP–MS detection was used by Bushee [30] to speciate Hg(II) and thimerosol, a mercury-containing preservative. Other work showed the separation of MeHg+, EtHg+ and Hg(II). The detection limits were extremely good, at about the 1 ppb level. 2-Mercaptoethanol (ME), when added to an eluent, complexes with mercury compounds to produce charge-neutral compounds that can be separated on reverse-phase columns. MacCrehan et al. described this procedure in a separation of Hg2+, MeHg+, EtHg+, and PhHg+ [31, 32]. 15.5.9 Other Metals
Aluminum species exist in oxidation state III. In aqueous solution, the simple ion exists as Al(H2O)63+ . This ion readily dissociates to give other ions such as Al(H2O)5OH2+, all of which are colorless. Over a wide pH range, under physiological conditions, in alkaline solution, the species appear to be Al(OH)2+, Al(OH)3, Al(OH)4–, Al3(OH)112–, Al6(OH)153+, and Al8(OH)222+. Study of the substitutions of aluminum aqua ion by ligands such as SO42–, citrate, and EDTA has been established by 27Al NMR spectroscopy [33]. A novel study on the speciation of aluminum in solution has been reported by Bertsch et al. [34]. Fluoro, oxalato, and citrate aluminum complexes were identi-
365
366
15 Chemical Speciation
fied as distinct peaks together with free Al(III). Post-column reaction/UV detection was used. These studies were used in kinetic, ion-exchange, and toxicological investigations. Gold cyanide complexes are important in gold-plating baths. As the Au(III) bath content increases, the plating efficiency is decreased. Mobile-phase ion chromatography can be used to determine total gold as well as Au(CN)2– and Au(CN)4– [34–36]. An anion-exchange method was also reported in which conductivity or UV detection could be used [37]. Lead speciation of Et3Pb+ and Me3Pb+ along with some other organometallic species was reported by MacCrehan et al. [32]. Lichrosorb NH2 column and an ion-pairing type eluent was used. Bushee, Krull and coworkers described an ionpairing reverse-phase method for the separation of Cu(I) and Cu(II) [38]. Detection was by ICPAE. The effect on the retention of Cu(I) of changing the ion-pairing reagent from pentanesulfonic acid to octanesulfonic acid was shown. Mn(II) and Mn(III), along with some other metals, were separated on a C18 column and detected by W [39]. An elegant approach to determine Pt, Pd and Au as their chloro complexes by IC was reported by Rocklin [40]. UV detection was used, resulting in detection limits of 0.03–1 ppm. In other work, IC was employed for the separation of chloro complexes of Pt, Pd, and Ir. UV detection was used with a 1 mM sulfosalicyclic acid, pH 4.2 eluent [41].EDTA anionic complexes of Cd, Ni, Cu. and Zn have been separated by IC [42]. The results agreed well with those obtained by AA and ICP methods. The separation and quantification of cyano complexes of various metals have been successfully carried out by Hilton and Haddad [43]. Cyano complexes of Cu(I), Ag(I), Fe(II), Fe(III), Co(II), Au(I), Au(III), Pd(II) and Pt(II) were analyzed by ion-pairing chromatography. The methods can be extended to many speciation problems.
References [1] W. T. Frankenberger, Jr., H. C. Mehra
and D.T. Gjerde, Environmental applications of ion chromatography, J. Chromatogr., 504, 211, 1990. [2] I. T. Urasa, S. H. Ram and V. D. Lewis, in Advances in Ion Chromatography, Vol. 2, P. Jandik and R. M. Cassidy (eds.), Century International Inc. Franklin, MA, 1990, p 93. [3] D.T. Gjerde, Eluent selection for the determination of cations in ion chromatography, J. Chromatogr., 439, 49, 1988. [4] D. T. Gjerde, D. Wiederin, F. G. Smith, and B. M. Mattson, Metal speciation by means of microbore columns with
direct-injection nebulization by inductively coupled plasma atomic emission spectroscopy, J. Chromatogr., 640, 73, 1993. [5] O. Shpigun and Yu.A. Zolotov, Ion Chromatography in Water Analysis, Wiley, New York, 1988. [6] Dionex (AU 144) - Determination of Hexavalent Chromium in Drinking Water Using Ion Chromatography. [7] US EPA Method 7199: Determination of hexavalent chromium in drinking water, groundwater and industrial wastewater effluents by ion chromatography and US
References
[8]
[9]
[10]
[11]
[12]
[13]
[14]
[15]
[16]
[17]
[18]
EPA Method 1636: Determination of hexavalent chromium by ion chromatography. A. F. Geddes and J. G. Tarter, The ion chromatographic determination of chromium(III)-chromium(VI) using an EDTA eluent, Anal. Lett., 21, 857, 1988. I. T. Urasa and S. H. Nam, Direct determination of chromium(III) and chromium(VI) with ion chromatography using direct current plasma emission as element-selective detector, J. Chrom. Sci., 127, 30, 1989. M. J. Powell, D. W. Boomer and D. R. Wiederin, Determination of chromium species in environmental samples using high-pressure liquid chromatography direct injection nebulization and inductively coupled plasma mass spectrometry, Anal. Chem., 67, 2474, 1995. H. Saitoh and K. Oikawa, Simultaneous determination of iron(II) and -(III) by ion chromatography with post-column reaction, J. Chromatogr., 329, 247, 1985. C. O. Moses, A. L. Herlihy, J. S. Herman and A. L. Mills, Ion chromatographic analysis of mixtures of ferrous and ferric iron, Talanta, 35, 15, 1988. J. Weiss Handbook of Ion Chromatography, 3rd Edition, 2 Volume Set, WileyVCH, 2004. R. R. Turner, Oxidation state of arsenic in coal ash leachate, Environ. Sci. Techn., 15, 1062, 1981. C. J. Craig, Organometallic Compounds in the Environment, Longman, London, 1986, p 198. C. McCrory-Joy, Single-column ion chromatography of weak inorganic acids for materials and process characterization, Anal. Chim. Acta, 181, 277, 1986. R. W. Slingsby, R. Al-Horr, C. A. Pohl and J. H. Lee, Use of dual-selectivity ICESI-MS for the separation and detection of anionic and cationic arsenic species., Amer. Lab., May, 2007 Y. A. Zolotov, O. A. Shpigun, L. A. Bubchikova and E. A. Sedelnikova, Ion chromatography as a method for the automatic determination of ions. Determination of selenium, Dokl. Akad. Nauk SSR, 263, 889, 1982, (Russian).
[19] S. G. Chen, K. L. Cheng and C. R. Vogt,
[20] [21]
[22]
[23]
[24]
[25]
[26]
[27]
[28]
[29]
Ion chromatographic separation of some aminopolycarboxylic acids and inorganic anions, Mikrochim. Acta, 1, 473, 1983. R. J. Shamberger, Biochemistry of Selenium, Plenum, New York, 1973, p 185. D. Chakraborti, D. C. J. Hillman, K. J. Irgolic and R. A. Zingaro, Hitachi Zeeman graphite furnace atomic absorption spectrometer as a seleniumspecific detector for ion chromatography. Separation and determination of selenite and selenate, J. Chromatogr., 249, 81, 1982. I. T. Urasa and F. Ferede, Use of directcurrent plasma as an element selective detector for simultaneous ion chromatographic determination of arsenic(iii) and arsenic(v) in the presence of other common anions, Anal. Chem., 59, 1563, 1987. H. C. Mehra and W. T. Frankenberger, Jr., Simultaneous determination of selenate and selenite by single-column ion chromatography, Chromatographia, 25, 585, 1988. R. E. Smith, Ion Chromatography Applications, CRC Press, Boca Raton, FL 1988. R.A. Cochrane in Trace Metal Removal from Aqueous Solution, R. Thompson, (ed.), The Royal Society of Chemistry, London, 1986, p 197. W. A. MacCrehan, Differential pulse detection in liquid-chromatography and its application to the measurement of organometal cations, Anal. Chem., 53, 74, 1981. L. Ebdon, S. J. Hill and P. Jones, Speciation of tin in natural waters using coupled high-performance liquid chromatography-flame atomic-absorption spectrometry, Analyst, 110, 515, 1985. F. E. Brinckman, W. R. Blair, K. S. Jewett and W. P. Iverson, Application of a liquid chromatograph coupled with a flameless atomic absorption detector for speciation of trace organometallic compounds, J. Chrom. Sci., 15, 493, 1977. R. M. Smith, A. M. Butt and A. Thakur, Determination of lead, mercury, and cadmium by liquid chromatography using on-column derivatization with
367
368
15 Chemical Speciation
[30]
[31]
[32]
[33]
[34]
[35]
[36] [37]
dithiocarbamates, Analyst, 110, 35, 1985. D. S. Bushee, Speciation of mercury using liquid chromatography with detection by inductively coupled plasma mass spectrometry, Analyst, 113, 1167, 1988. W. A. MacCrehan, R. A. Durst, Measurement of organomercury species in biological samples by liquid-chromatography with differential pulse electrochemical detection, Anal. Chem., 50, 2108, 1979. V. I. A. MacCrehan, R. A. Durst and J. M. Bellama, Electrochemical detection in liquid chromatography: application to organometallic speciation, Anal. Lett., 10, 1175, 1977. L. O. Ohman and S. Sjonberg, Equilibrium and structural studies of silicon(IV) and aluminum(III) in aqueous solution. Part 13. A potentiometric and aluminum-27 nuclear magnetic resonance study of speciation and equilibriums in the aluminum(III)-oxalic acid-hydroxide system, J. Chem. Soc. Dalton Trans., 12, 2665, 1985. P. M. Bertsch and M. A. Anderson, Determination of gold, palladium, and platinum at the parts-per-billion level by ion chromatography, Anal. Chem., 61, 535, 1989. E. Otu, C. Robinson and J. Byerley, Inorganic modifiers in ion-pair chromatographic separation of dicyanoaurate(I), Analyst, 118, 1277, 1993. J. Weiss, Ion Chromatography, VCH, New York, NY 1995. M. Nonomura, Ion chromatographic analysis for cyanide, Met. Finish., 85, 15, 1987.
[38] D. Bushee, I. S. Krull, R. N. Savage and
[39]
[40]
[41]
[42]
[43]
S. B. Smith, Jr., Metal cation/anion speciation via paired-ion, reversed phase HPLC with refractive index and/or inductively coupled plasma emission spectroscopic detection methods, J. Liq. Chromatogr., 5, 463, 1982. B. W. Hoffman and G. Schwedt, Application of HPLC to inorganic analysis. Part VII. Comparison between pre-column- and on-column derivatization for separation of different metal oxinates; quantitative determination of manganese(II) besides manganese(III) ions, J. HRC and CC, 5, 439, 1982. R. D. Rocklin, Determination of gold, palladium, and platinum at the partsper-billion level by ion chromatography, Anal. Chem., 56, 1959, 1984. O. A. Shpigun and Yu. E. Pazuktina, Ion chromatographic determination of platinum in the presence of other platinum-group metals and inorganic anions, Zh. Analit. Khim., 42, 1285, 1987. M. Yamamoto, H. Yamamoto, Y. Yamamoto, S. Matsushita, N. Baba and T. Ikushige, Simultaneous determination of inorganic anions and cations by ion chromatography with ethylenediaminetetraacetic acid as eluent, Anal. Chem., 56, 832, 1984. P. R. Haddad and N. E. Rochester, Ioninteraction reversed-phase chromatographic method for the determination of gold(I) cyanide in mine process liquors using automated sample preconcentration, J. Chromatogr., 439, 23, 1988.
369
Index a Acid rain 218 Adsorption 112, 154 Alcohols separation by IEC 230–235 Alkali metal ions elution order in acetone 195 retention factors 194 separation 193 Alkaline earths separation 191 Alkane sulfonates 241, 244 Alkylammonium salts effect on CE separations 291–294 Aluminum complexes speciation 365 Amidoxim 203, 204 Ammonia determination by IEC 225 Ammonium / potassium separation 182 Amperometric detection 87 Analytical control 349, 350 Anex see Anion exchange Anilines separation 192 Anion chromatography 10, –131 at ppb concentrations 218 at very low concentrations 146 common anions 165 detectors 137 eluents 137, 150 explanation of peaks 150 inorganic anions 143, 144 non-suppressed 147–154 organic anions 145 organic solvents 146 scope 131, 155 separation conditions –135
suppressed 17, 138 Anion exchange 3 arsenic species 172 Anion exchange columns latex coated 166 Anion exchangers 45, 56 Transgenomic AN 1 45 local structure 124 macroporous 45 pH effect 52 polyfunctional 50 porous 45 relative retention 51 selectivity 56, 57 spacer arm effect 52 Anion-exchange columns 132 Alltech 134 Dionex 133, 134 Hamilton 135 Metrohm 135 Anions detection limits 158 in concentrated acids 328 inorganic 157 relative retention 49 retention factors 59 retention times 52 selectivity 112 separation 6 Anions and cations simultaneous separation 216–219 Arsenazo 8 Arsenic anions 149 Arsenic poisoning 354, 361 Arsenic speciation 361, 362 Atomic absorption detection 100 arsenate 100 arsenite 100 dimethyl arsinate 100
Ion Chromatography, 4th Ed. James S. Fritz and Douglas T. Gjerde Copyright © 2009 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim ISBN: 978-3-527-32052-3
370
Index monomethyl arsenate 100 p-aminophenoarsenate 100 Atomic emission (AE) detection 100
b BGE 283, 288, 289 Background electrolyte 296 see BGE Baseline noise 155 Bases determination by IEC 223–226 Bicarbonate determination by IEC 222, 224 Bromide 292 Bromide complexes 4 Buffer tris(hydroxymethyl)aminomethane (TRIS) 275 By-pass loop 327
c CE see capillary electrophoresis CE-IC 295 CMC 284 CTAC 284 Calcium (II) 4 Calibration curve 35 Capacity factor 356 see retention factor Capillary electrophoresis 1, 2, 260, 263–297 at high salt concentrations 280–283 capillary pretreatment 264 conditions 296 current 282 detection 265 effect of ionic polymers 287–291 pH effect 281 power supply 265 principles –265 sample 291 sample injection 267 sample introduction 264 separation of ammonium and potassium 279 separation of anions 274–277 separation of cations –278 separation of chloride isotopes 277 separations at low pH 279 steps in analysis 264 summary 294–295 terms 266 vectors 266
Capillary surface 296 Capillary zone electrophoresis 267 Carbohydrate retention pH effect 233 Carbohydrate separation 170 Carbohydrates 171 detection 235 effect of metal ions 231, 232 separation mechanism 230 Carbon dioxide determination by IEC 222 Carboxylic acids separation on C18 column 216, 217 Catex 37 see Cation exchange Cation chromatography 175–205 Quaternary ammonium ions 186 at low concentrations 188 historical 178 inorganic ions 185 principles 175 temperature effect 179–181 variables 175 Cation exchange 3 elution order 5 selectivity 5 Cation exchange columns Alltech 177 Dionex 176, 177 Hamilton 177 Metrohm 177 Cation exchanger 37, 57 capacity 61 carboxylate 179 cross-linking effect 58, 59 18-crown-6 63 exchange capacity 57 exchange reactions 38 latex-agglomerated 58 location of exchange groups 59 phosponate and carboxylate 179 selectivity 4, 62 silica 38 spacer arm 61 sulfonated 57, 58, 59 weak acid 51, 198 Cation self-regenerating suppressor 185 Cation separations 3, 4 Cation-anion interaction in CE 294 Cations selectivity 111 Cell contant, conducitivity 73 Cetyl pyridinium chloride 44 Chaotropic interaction 260
Index Chelating resins 63 Chelex 100 201 Chemical speciation 353–368 Chloromethylation 43, 122 Choosing a method 335 Chromatographic principles 105 Chromatographic terms 105–109 Chromium in plating baths 353 speciation 355, 359 Chromium III, VI separation at 0.1 ppb 358 Cinnamaldehyde 226 equilibrium with water 227 spectra 227 Cis-trans separation 244 Citric acid 198 Co-migration 270 Coated columns 257 Maleic acid 178 Cobalt II 5, 5 Color-forming reagents 81, 82 Aresenazo I 82 Arsenazo III 82 PAR 82 alizarin red S 81 chlorophosphonazo III 81 chrom azurol S 81 4-(2-pyridylazo)resorcinol (PAR) 81 4-(2-thiazolylazo)resorcinol (TAR) 81 xylenol orange 81 Column TSK gel 219 Column hardware 31 Column metal contamination 310 Column oven 33, 127, 313 Column temperature 179, 180 Columns Alltech 46, 47 Hamilton 46 anion-exchange 44 cation exchange 44 coated 160, 165 enhancement 222–224, 226 monolith 43–45 monolithic 161, 162 surfactant 182 Complex metal chlorides 279 Complexing acids 4 Concentrator column 326, 327 Conducitivity hardware 75 Conductance 70, 73, 75, 155, 156 background 156, 159
ion pairs 210 metal cations 278 of a sample peak 158, 159 Conduction 70 Conductivity 73 background 210 Conductivity detector 70 Conductometric detector 86 Confidence intervals 345 Confidence limits 345 Contactless conductivity detection 76 Control chart 349 Copper II 5 Coulometric detection 87 Counter migration 270 Critical micelle concentration (CMC) 284 Crown ethers 285, 286 18-Crown-6 286 binding constants 182
d DHPLC, denaturing HPLC 309, 314 DNA 301 DNA chromatography 299 Data acquisition 35 Dead time 105, 108 Dead volume 23 Density function 342 Detection 323 ICP-MS 358 UV-Vis 8 chromate 275, 276 direct conductivity 215 in CE 275, 276 molybdate 275, 276 naphthalenedisulfonate 275 non-suppressed conductivity 187 on-line 8 optical absorbance 163 potentiometric 166, 167 spectrophotometric 188 suppressed conductivity 184–188 Detector 34 PAD 34 UV-Vis 34, 163 conductivity 9, 16, 35 mass spectronometry 34 Detectors 69 ICP-MS 173 atomic emission 172 direct UV 157 evaporative light scattering (ELS) 170 for IEC 210 indirect absorbance 164, 165
371
372
Index inductively coupled plasma 172 mass spectrometry 173 pulsed amperometric (PAD) 168 Dialysis Donnan (active) 330–333 passive 330 sample preparation 329 2, 3-Diamino propionic acid (DAP) 178 Dilute and shoot 1, 160, 323 Dionex AS12A 56 Dionex CS12A 179, 181 Dionex CS 3 60 Dionex CS 10 60 Dionex CS 12A 61, 62 Direct spectrophotometric detection 78, 79 Direct-injection nebulizer 172 Distribution coefficient (Kd) 208 Distribution coefficients anion exchange 7 Donnan membrane 260 Double layer capacitance 75
e EDTA 197 EOF 274, 275, 279, 295 EOF vector 278 Efficiency 124, 125 Electric field 268 Electrochemical cell 96, 97 porous 96 thin layer 96 tubular 96 wall jet 96 Electrochemical detector 85 Electrode material 88 Electrodialysis 328, 329 Electrokinetic injection 268 Electromigration 266 Electroosmotic flow 265, 268–274 control 271 effect on separations 270, 271 Electroosmotic mobility 266, 268, 269 Electropherogram 267 Electrophoretic flow 263, 265, 266 Electrophoretic mobilities 276 Electrophoretic mobility 263, 266, 282 Electrostacking 268, 268, 296 Eluent 164 H2SO4 / HF 6 aromatic bases 189 background conductance 156 background conductivities 220 basic 152
benzoate 11 carbonate 24 carboxylic acid 153 carboxylic acid salts 151 complexing 195, 196, 198, 199, 200, 203, 356 cupric sulfate 190 2,6-dimethylpyridine 190 dry methanol 226 electrolytic generation 24, 25, 26 ethylene diamine tartrate 195, 196, 200 hydroxide 24 kinetics 203 methane sulfonic acid 192 molybdate 166 nicotinic acid 53 perchlorate 167 phthalate 53 phthalates 151 sodium hydroxide 16 sodium perchlorate 160 succinic acid 219 sulfosalicylic acid 219 tartaric acid 219 Equilibration 2 ion exchange 15 Equilibrium ion exchange 110 Equivalent conductance 73, 74 bromide 16 chloride 16 Equivalent conductance, limiting 71 Evaporative light scattering detector (ELSD) 98, 99 detection cell 99 evaporation chamber 99 nebulizer 98 Exchange capacity 10, 108, 111 Extraction disks 325
f Faraday’s law 87 Field strength 268, 268 Filters 324 Flame photometric detection 100 Fluorescense detector 83 Fluorescent dyes 84–85 FAM 84 HEX 84 NED 84 TET 84 others 84, 85 Fluoride 55 Fluoride complexes 4, 6
Index Fragment collection, nucleic acid 314 Fructose 169, 171
g General detector 70 Geometric isomers 244 Glucose 169, 171 Gold cyanide complexes 353, 360 Gold III 280 Gradient 29, 30 Gradient elution 138, 144, 145, 186 Guard column 327 Guard columns 32
h HIBA 285 HPLC 8, 12, 21 Helmholtz layer 269 n-Hexylammonium acetate 306 Hydrazine determination by IEC 225 Hydrochloric acid 4, 6 Hydrogen peroxide 4 2-Hydroxy isobutyric acid (HIBA) 5
i IC see ion chromatography IC-PED 94 amino acids 94 ICP 100 atomic emission detection 100 mass spectrometry detection 100 ICP-AE 355 Iminodiacetic acid (IDA) 201, 202 Indirect detection 80 Inductivity coupled plasma (ICP) detection 100 Injection Peak 14 Injection loop 30, 31 Inner volume (Vi) 208 Inorganic cations detection limits 191 separation 178, 180, 181, 190 Instrumentation 21 Integrated pulsed amperometric detection (IPAD) 93 Iodide 292 Ion III masking 198 Ion association 263 Ion association constant 291, 293, 294
Ion chromatograph 10 block diagram 22 components 21 Ion chromatography 1, 2, 9 applications 340 basis for separation 17 birth of 8 capillary electrophoretic 283 experimental setup 13 forced-flow 8 historical development 2 migration of sample ions 15, 16 non-suppressed 10 principles 13 separations 105–109 single column 10 two dimensional 33 valveless injection 357 Ion exchange electrostatic attraction 120 enforced pairing 120, 121 hydrophobic attraction 120 matrix effect 121 mechanisms 120 selectivity 120 Ion exchange EKC 287 Ion exchange disks 65, 66 Ion exchange equilibrium with ionic polymers 288–289 Ion exchange sites 16 Ion exchanger hydrophilic 129 Ion exchangers Tetra alkyl phosphonium 123 cross-linking 55 exchange capacity 55 hydrophobic attraction 55 latex-agglomerated 54, 56 metal oxide 64 silica zirconia 64 solvation effects 123 Ion exclusion chromatography 207–209 alcohol modifiers 214 carbonic acid effect 214 carboxylic acids 212, 213, 215 conclusions 220–222 detectors 210 eluents 209 elution order 208 gradient elution 212 mechanism 211 n-butanol 215 principles 207
373
374
Index Ion pair chromatography 239–248 double layer 240 effect of salt 246 elution order 240, 241 mechanism 246–248 principles 239 scope –240 variables 242 Ion pair formation 247 Ion pairs 261 Ion-exchange EKC 295 Ion-pairing reagent 306 Ion-pairing reverse-phase chromatography 303 Ion-selective electrodes 166 Ions isolation from organic matter 333 Iron in plating baths 361 Iron III 5, 8 Iron speciation 360 Isocratic elution 29, 138, 144, 145
k Karl Fischer method 226, 228
l Lactic acid 187, 285 Lambert-Beer law 77 Lanthanides CE separation 285, 286 separation 204 Lead speciation 366, 366 Ligand 18, 285 Limiting current 88 Limiting equivalent conductance 71 Lithium I 189 Lithium sulfate 282, 289
Metal EDTA complexes 80 Metal cations CE separation 280 activity 118 elution orders 113 linear regression 117, 118 retention factors 114–116, 119 selectivity 113 Metal choride complexes 80 Metal complexes formative constants 199 Metal ion contamination 312 Metal ions CE separation 285, 287 separation 6 Metal-cyano complexes separation by IPC 243 Methane sulfonic acid 179, 185 generation 25 Methanol 5 Method Development 335–341 Method development AOAC method 338 defining the problem 335 example 338 experimental considerations 336 ion exclusion 338 Methylmercury 353 Micellar electrokinetic chromatography (MEKC) 284 Micelle 284 Migration rate 107 Migration time effect of acetate, sulfate, PDDAC 290 Mobility of ions 71, 76 Molar absorptivity 164 Molybdenum VI 4 Monolith columns 43, 44 Morpholine 180
m MEKC 284 Magnesium II 189 Manganese II 7 Mannitol 338 Masking reagents 197, 198 Mass spectrometry (MS) detection 100 Matrix elimination 328 Membrane cation exchange 25 hydrophilic 330 ion-exchange 331 Membrane suppressors 210 2-Mercaptoethanol 365 Mercury speciation 365
n Nano domains 294 Negative peaks 152 Nickel II 7 Nitric acid 7 Non suppressed conductivity detection 72 Nonporous polymer column 305 Normal distribution curve 342, 343 Normal random variable (Z) 342 Nucleic acid enzymology 315–317 polynucleotide assays 316 telomerase assays 315 uracil DNA glycosylase assays 317 Nucleotides 147
Index
o Observed mobility 269 Octylsulfonic acid 210 Ohm’s law 85 Organic acid retention pH effect 234 Organic amine cations separation by CE 272 Organic amines separation 180 Organic solvents 191, 192, 193, 194 effect on ion exchange 5, 7, 8 in IEC 213 Oxalic acid 198 Oxyhalides in drinking water 353
p PAR 8, 199, 203 PDA 199, 200 PDDAC 287–290 PED post-column derivatization 95 PEEK 23, 31, 32, 172 Pairing reagents 242, 243, 244, 247–248 Pauling ion radii 193, 194 Peak broadening 125, 126 axial diffusion 126 variance 128 Peak resolution 106 Perchloric acid 114 Peristaltic pump 357 Phosocholine 258 Phosphorolylated DNA 316 Plasmids 304 Plate height 161 Platinum electrode 264 Platinum group metals speciation 366 Platinum metals 280 Polymer macroporous 333 Post-column derivatization 81 Potassium I 189 Potentiometric detection 86 Power supply 278 Probability level 346 Protoactinium V 6 Pseudo phase 295 Pulsed amperometric detection (PAD) 91–93 S-compounds 93 alcohols 90 aminoalcohols 92 aminosugars 92
carbohydrates 91 organic thiophosphates 93 thioalcohols 93 thiocarbamates 93 thioethers 93 thiophenes 93 thiosugars 92 Pulsed amperometric detector 339 Pulsed electrochemical detection (PED) 89, 90 Pump 26, 30 check valves 27, 28 2,6-Pyridinedicarboxylic acid 203 3,6-Pyridinedicarboxylic acid 198
r RNA 302 RNA chromatography 299 mRNA, messenger RNA 316, 318 rRNA, ribosomal RNA 316 tRNA, transfer RNA 316, 319 Refractive index detector 97 Relative conductance 70 Resin anion exchange 9 cation exchange 9 Resin capacity 42 Resins HEMA 46, 47 adsorption of ions 113 anion exchange 48 chelating 201–204 cross-linking 38, 40, 121 effect of sulfonation 221 exchange sites 121 exchange sites location 118 for IEC 211 functional group 122 functional group effect 47, 48 gel type 40 gel-type 59 hydroxy ethyl groups 48 latex agglomerated 122 macro reticular 40 macrocycle function 181 macroporous 39, 211 methylmethacrylate 46 microphorous 39, 40 multi-purpose 64 polydivinyl benzene 46 polymeric 38, 179 polymeric matrix 121 pore diameter 40 retention factors 118, 119
375
376
Index selectivity 48 spacer arm effect 53 strong acid 41 substrate 38 sulfonated 41 weak acid 41 weak base 42 Resins and columns 37 Resistance to mass transfer 267 Resolution 2 Retention Volume (VR) 208 Retention factor 106–109 Retention factors with complexing eluents 195 Retention time 107 Retention volume 107 Rhodium III hydrolysis 280, 281
s SDS 284 Saccharides separation by IEC 230–235 Sample concentration 329 Sample contaminants 325 Sample homogenization 324 Sample injection volume 326 Sample injector 39 Sample integrity Sample ion pairs 257 Sample neutralization 328, 329 Sample particulate matter 324 Sample preconcentration 325, 326 Sample preparation 339 Sample pretreatment 323–334 Sample values 327 Scavenger columns 32 Selective detector 70 Selectivity 109 Selectivity coefficient 110, 111 Selenium speciation 363 Sensitivity 155 Separation factor 270, 271 Separation factor (a) 106, 113, 114 Separations alkali metal cations 12 group 4, 5 Separator column 9 Siemens, microSiemens 74 Silanol groups 271, 274 Silver electrode 167 Smoluchowski equation 269 Sodium I 189 Sodium hydroxide 152
Sodium iodide in zwitterion phases 258 Sodium perchlorate 115, 116, 122 Solid-phase extraction (SPE) 324 Solvation 192 Speciation EPA methods 354, 355 Specific conductance 73, 74 Standard deviation chromatographic peaks 125 Standard normal distribution 342, 343. 344 Statistical evaluation of data 341–347 Statistical terms 341 Step gradient 3 Student t distribution 345, 346 Sucrose 169, 171 Sulfur anions separation by IPC 245 Supporting electrolyte 88 Suppressed conductivity detection 72 Suppressor 33 electrolyte 140 fiber 138 membrane 138, 140 micromembrane 34 packed-bed 34 self-regenerating 141 solid-phase reagents 141, 142 Suppressor column 9 Surfactant phosphocholine 251 Surfactants 121 anionic 183 cationic 183 non-ionic 183 separation 65, 183, 184 Swelling propensity 46 System peak 222 System peaks 154
t Tartrate complexes 5 Tellurium speciation 362 Temperature Modes 307–309 fully denaturing mode 308 nondenaturing mode 307 partially denaturing mode 309 Tetrabutylammonium acetate 293 Tetrabutylammonium chloride 292 Tetrapentylammonium salts 241, 243 Theoretical plate 124, 267 effective 125 Thiocyante 292
Index Third phase 216 Thorium IV 6 Tin IV 5 Tin speciation 364 Titanium IV 4 Transition metal ions separation 204 Tributylamine 274 Tributylammonium ethanesulfonate 271 Triethyl ammonium acetate (TEAA) 303, 306 Triethylamine 274 Trimethylorthoformate 229 Triton-X 160
u UV and fluorescence detection 313 UV-VIS Hardware 82, 83 Ultraviolet-visible detector (UV-VIS) 77 Uranium VI 6
v Validation accuracy 347 limit of detection (LOD) 348 linearity and range 348 precision 348 quantitation (LOQ) 348 robustness 348 ruggedness 348 specificity 348 Validation of procedures 347–350 Valve switching 33 Van Deemter plot 127 Vanadium IV, V 4 Vanadium V equilibrium 280 Vanadium speciation 353, 364 Variance 342, 344 Visualization reagent 265, 278
Visualization reagents 278 Void volume (Vo) 208
w Water determination by IEC 226–230 Water determination at very low concentrations 229 effect of aldehydes and ketones 228 Weak acid anions 153
z Zeta potential 260, 268, 269 Zn II 5 Zone spreading at interface 128 diffusion 122, 128 equilibration kinetics 126 extra-column 128 flow rate 127 mass transfer 128 multipaths 126 Zwitterion ion chromatography (ZIC) 261 Zwitterion phases ion pairs 259 mechanism 259–261 separation of acids 258 Zwitterion stationary phases 251–262 CHAPS 255 Introduction 251 elution order 252 historical 253 ion pairing like forms 253, 254, 259 multiple peaks 252, 254 pH effect 256 properties 251, 252 separation of cations 256–259 separation of inorganic anions 256 water eluent 253, 254
377
