Handbook of Solid Phase Microextraction

Handbook of Solid Phase Microextraction

Handbook of Solid Phase Microextraction

Janusz Pawliszyn University of Waterloo Waterloo, Ontario Canada

AMSTERDAM BOSTON HEIDELBERG LONDON NEW YORK OXFORD PARIS SAN DIEGO SAN FRANCISCO SINGAPORE SYDNEY TOKYO G

G

G

G

G

G

G

G

G

G

Elsevier 32 Jamestown Road London NW1 7BY 225 Wyman Street, Waltham, MA 02451, USA First edition 2012 Copyright r 2012 Elsevier Inc. All rights reserved No part of this publication may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopying, recording, or any information storage and retrieval system, without permission in writing from the publisher. Details on how to seek permission, further information about the Publisher’s permissions policies and our arrangement with organizations such as the Copyright Clearance Center and the Copyright Licensing Agency, can be found at our website: www.elsevier.com/permissions This book and the individual contributions contained in it are protected under copyright by the Publisher (other than as may be noted herein). Notices Knowledge and best practice in this field are constantly changing. As new research and experience broaden our understanding, changes in research methods, professional practices, or medical treatment may become necessary. Practitioners and researchers must always rely on their own experience and knowledge in evaluating and using any information, methods, compounds, or experiments described herein. In using such information or methods they should be mindful of their own safety and the safety of others, including parties for whom they have a professional responsibility. To the fullest extent of the law, neither the Publisher nor the authors, contributors, or editors, assume any liability for any injury and/or damage to persons or property as a matter of products liability, negligence or otherwise, or from any use or operation of any methods, products, instructions, or ideas contained in the material herein. British Library Cataloguing-in-Publication Data A catalogue record for this book is available from the British Library Library of Congress Cataloging-in-Publication Data A catalog record for this book is available from the Library of Congress ISBN: 978-0-12-416017-0 For information on all Elsevier publications visit our website at elsevierdirect.com This is an e-only product

To members of my research group at the University of Waterloo, my industrial and academic collaborators and the Department of Chemistry for help in the development of the SPME technology and this handbook.

Preface

The simplification of sample preparation and its integration with both sampling and the convenient introduction of extracted components to analytical instruments are a significant challenge and an opportunity for the contemporary analytical chemist. The results of current research will have a profound effect on future analytical technology. This monograph describes fundamentals and practical information about the solvent-free sampling/sample preparation/introduction approach: solid-phase microextraction (SPME). SPME techniques have been developed not only to address the need for a reduction in the size of the extraction instrumentation and solvent use but also to explore the ability of this approach to facilitate rapid and convenient sample preparation both in the laboratory and on site. There are many advantages of SPME, which can be realised to a higher or lesser degree depending on the geometric configuration of the instrument. Some designs of SPME better address issues associated with agitation, while others address the ease of implementing on-site analyses or sample introduction to the analytical instrument. For example, full automation of standard delivery, extraction and introduction performed sequentially is possible for gas chromatography (GC) using a coated fibre format and for liquid chromatography (LC) when an internally coated capillary is used. Conversely, the use of coated fibres or thin-films arranged to fit in a 96-well multi-well format facilitates parallel high-throughput sample processing. Small extraction devices facilitate on-site applications, including in vivo analyses, and allow for coupling to a variety of analytical micro-instrumentation, including capillary and microfluidics systems. Non-exhaustive microextraction techniques possess unique advantages because typically only a small portion of the target analyte is removed from the matrix. This feature allows the monitoring of chemical changes, partitioning equilibria and speciation in the investigated system because sampling causes minimum perturbation to the system. Therefore, the use of microextraction-based strategies results in better characterisation and more accurate information about the investigated system or process compared to exhaustive techniques. Non-exhaustive microextraction techniques provide signal magnitudes that are proportional to the free concentration of target analyte, defining the fraction of the analyte that is bioavailable. This unique feature of the non-exhaustive techniques allows the measurement of binding constants in complex matrixes, providing additional information about the investigated system. It also indicates the need for careful calibration and optimisation. Therefore, the development of robust quantitative analytical methods based on microextraction requires more time, but when the procedures are optimised, they

xvi

Preface

are more convenient and cost-effective compared to conventional exhaustive extraction approaches. This handbook has been developed to address this challenge by assisting the user in this task. It is the electronic edition of a print version published earlier and available from Supelco or our website: http://www.spme.uwaterloo.ca/. There are numerous untapped opportunities available for exploration, especially considering the unique features of SPME, making this research direction vital and scientifically interesting. I wish you a rewarding experience exploring SPME for your application. Janusz Pawliszyn Waterloo, Canada

List of Contributors

Barbara Bojko Department of Chemistry, University of Waterloo, Waterloo, Ontario, Canada Lucie Kudlejova Department of Chemistry, University of Waterloo, Waterloo, Ontario, Canada Heather Lord Department of Chemistry, University of Waterloo, Waterloo, Ontario, Canada Fatemeh Mirnaghi Department of Chemistry, University of Waterloo, Waterloo, Ontario, Canada Florin Marcel Musteata Department of Pharmaceutical Sciences, Albany College of Pharmacy and Health Sciences, Albany, NY, USA Gangfeng Ouyang School of Chemistry and Chemical Engineering, Sun Yat-sen University, Guangzhou, P. R. China Janusz Pawliszyn Department of Chemistry, University of Waterloo, Waterloo, Ontario, Canada Sanja Risticevic Department of Chemistry, University of Waterloo, Waterloo, Ontario, Canada Robert E. Shirey Supelco, Bellefonte, PA, USA Dajana Vuckovic Donnelly Centre for Cellular and Biomolecular Research, University of Toronto, Toronto, Ontario, Canada

1 Solid-Phase Microextraction in Perspective Janusz Pawliszyn Department of Chemistry, University of Waterloo, Waterloo, Ontario, Canada

1.1

Sample Preparation as Part of the Analytical Process

The analytical procedure for complex samples consists of several steps that typically include sampling, sample preparation, separation, quantitation, statistical evaluation and decision making (Figure 1.1). Each step is critical for obtaining correct and informative results. The sampling step includes deciding where to get samples that properly define the object or problem being characterised and then choosing a method to obtain samples in the right amounts. The objective of the sample preparation step is to isolate the components of interest from a sample matrix. This is because most analytical instruments cannot handle the matrix directly. Sample preparation involves extraction procedures and can also include ‘clean-up’ procedures for very complex ‘dirty’ samples. This step must also bring the analytes to a suitable concentration level for detection; therefore, sample preparation methods typically include enrichment. During the separation step of the analytical process, the isolated complex mixture containing target analytes is divided into its constituents, typically by means of chromatographic or electrophoretic techniques. Quantitation is the determination of amounts of the identified compounds. The identification can be based on a retention time combined with selective detection; more frequently, however, instruments providing more specific information (namely, mass spectrometers) are used to eliminate possible errors in quantification due to interferences. Statistical evaluation of the results provides an estimate of the target compound’s concentration in the sample being analysed. The data will then support appropriate decisions, which might include taking another sample for further investigation. It is important to note, as emphasised in Figure 1.1, that analytical steps follow one after the other, and a subsequent step cannot begin until the preceding one has been completed. Therefore, the slowest step determines the overall speed of the analytical process, and improving the speed of a single step may not result in a throughput increase. To increase the throughput of analysis, all steps need to be considered. Also, errors performed in any preceding step, including sampling, will result in the overall poor performance of the procedure. There have been major breakthroughs in the development of improved instrumentation, which involve miniaturisation of analytical devices and hyphenation of different steps into one system. It is recognised that an ideal instrument would Handbook of Solid Phase Microextraction. DOI: 10.1016/B978-0-12-416017-0.00001-2 © 2012 Elsevier Inc. All rights reserved.

2

Handbook of Solid Phase Microextraction

Action Decision Statistical evaluation

Separation and quantitation

Sample preparation

Sampling

Figure 1.1 Steps in the analytical process.

perform all the analytical steps without human intervention, preferably directly on the site where an investigated system is located rather than moving the sample to the laboratory, as is currently done. This approach would eliminate the errors and the time associated with sample transport and storage and result in more accurate, more precise and faster production of analytical data. Although such a total analysis system (TAS) is challenging to build, today’s sophisticated instruments, such as the gas chromatograph/mass spectrometer (GC MS) or liquid chromatograph/mass spectrometer (LC MS), can separate and quantify complex mixtures and automatically apply chemometric methods to evaluate results statistically. It is much more difficult to hyphenate sampling and sample preparation steps, primarily because the current state of the art in sample preparation techniques employs multistep procedures involving organic solvents. These characteristics make it difficult to develop a method that integrates sampling and sample preparation with separation methods for the purposes of automation. The result is that more than 80% of analysis time is currently spent on sampling and sample preparation steps. One of the reasons that progress in the area of sample preparation is so slow is that the fundamentals of extraction involving natural, frequently complex samples are much less developed and understood compared to physicochemically simpler systems used in separation and quantification steps, such as chromatography and mass spectrometry. This situation creates an impression that rational design and optimisation of extraction systems is not possible. Therefore, development of sample preparation procedures is frequently considered to be ‘art’, not ‘science’. This situation creates a tendency by practitioners and regulatory agencies to prefer exhaustive over non-exhaustive techniques (Figure 1.2), even though this choice

Solid-Phase Microextraction in Perspective

3

Extraction techniques Flow-through equilibrium and pre-equilibrium Exhaustive Purge and trap

Non-exhaustive In-tube SPME

Batch equilibrium and pre-equilibrium Exhaustive

Steady-state exhaustive and non-exhaustive

Non-exhaustive

LLE

Headspace

Sorbent trap

Soxhlet

LLME

SPE

Sorbents

SPME

SFE

MAE

Membrane

PFE

Figure 1.2 Classification of extraction techniques.

frequently results in labour-intensive and costly procedures. The main objective of the exhaustive techniques is to remove analytes completely from a sample matrix and transfer them to the extraction phase. The fundamental advantage of exhaustive methods is that, in principle, they do not require calibration because the vast majority of analytes are transferred to the extraction phase. There are alternative extraction techniques, however, with their own unique advantages, which have been developed to reduce solvent use and improve performance. A summary of extraction techniques is given below.

1.2

Classification of Extraction Techniques

Figure 1.2 provides a classification of extraction techniques and unifies the fundamental principles behind the different extraction approaches. In principle, exhaustive extraction approaches do not require calibration because most analytes are transferred to the extraction phase by employing overwhelming amounts of it. In practice, however, confirmation of satisfactory recoveries is implemented in the method by using surrogate standards. To reduce the amounts of solvents and time required to accomplish exhaustive removal, batch equilibrium techniques (e.g. liquid liquid extractions) are frequently replaced by flow-through techniques. For example, a sorbent bed can be packed with the extraction phase dispersed on a supporting material; when a sample is passed through, the analytes in the sample are retained on the bed. Large volumes of sample can be passed through a small cartridge, and the flow through the well-packed bed facilitates efficient mass transfer. The extraction procedure is followed by desorption of analytes into a small volume of solvent, resulting in substantial enrichment and concentration of the analytes. This strategy is used in sorbent-trap techniques and in solid-phase

4

Handbook of Solid Phase Microextraction

extraction (SPE).1 Alternatively, sample (typically a solid sample) can be packed in the bed and the extraction phase can be used to remove and transport the analytes to the collection point. In supercritical fluid extraction, compressed gas is used to wash analytes from the sample matrix; an inert gas at atmospheric pressure performs the same function in purge-and-trap methods. For example, in dynamic solvent extraction in a Soxhlet apparatus, the solvent continuously removes the analytes from the matrix at the boiling point of the solvent. In more recent pressurised fluid extraction techniques, smaller volumes of organic solvent (or even water) are used to achieve greater enrichment at the same time as extraction because of the solvent’s increased capacity and elution strength at high temperatures and pressures.2 Alternatively, non-exhaustive approaches can be designed on the basis of the principles of equilibrium, pre-equilibrium and permeation.3 Although equilibrium non-exhaustive techniques are fundamentally analogous to equilibrium-exhaustive techniques, the capacity of the extraction phase is smaller and is usually insufficient to remove most of the analytes from the sample matrix. This is because a small volume of the extracting phase is being used relative to the sample volume, such as is employed in microextraction [solvent microextraction4 or solid-phase microextraction (SPME5)] or a low sample matrix extraction phase distribution constant, as is typically encountered in gaseous headspace techniques.6 Pre-equilibrium conditions are accomplished by breaking the contact between the extraction phase and the sample matrix before equilibrium with the extracting phase has been reached. Although the devices used are frequently identical to those of microextraction systems, shorter extraction times are employed. The preequilibrium approach is conceptually similar to the flow-injection analysis (FIA) approach,7 in which quantification is performed in a dynamic system and system equilibrium is not required to obtain acceptable levels of sensitivity, reproducibility and accuracy. In permeation techniques, e.g. membrane extraction,8 continuous steady-state transport of analytes through the extraction phase is accomplished by simultaneous re-extraction of analytes. Membrane extraction can be made exhaustive by designing appropriate membrane modules and optimising the sample and stripping flow conditions,9 or it can be optimised for throughput and sensitivity in non-exhaustive, open-bed extraction.10 Because membrane extraction is a naturally diluting process resulting in lower concentration of analytes in the stripping phase compared to the sample matrix, incorporation of a sorbent interface to enrich the analyte in the stripping phase prior to detection is very useful, as discussed in Refs. 7 10. As the preceding discussion and Figure 1.2 indicate, there is a fundamental similarity among the extraction techniques used in the sample preparation process. In all techniques, the extraction phase is in contact with the sample matrix and analytes are transported between the phases. To ensure quantitative transfer of the analyte in an exhaustive technique, the phase ratio is higher and geometries are more restrictive than in non-exhaustive approaches. The thermodynamics of the process are defined by the extraction phase/sample matrix distribution constant. It would be instructive to consider in more detail the kinetics of processes occurring at the

Solid-Phase Microextraction in Perspective

5

extraction phase/sample matrix interface because this controls the time required by the analytical procedure. The analytes are often re-extracted from the extraction phase, but this step is not discussed here because this process is analogous and more basic in principle than removing analytes from a more-complex sample matrix. Fundamental understanding of extraction principles has advanced in parallel with the development of new technologies. This progress has been very important in the development of novel approaches that result in new trends in sample preparation, e.g. microextraction, miniaturisation and integration of the sampling and separation and/or quantification steps of the analytical process. The fundamentals of the sampling and sample preparation processes are substantially different from those related to chromatographic separations or other traditional disciplines of analytical chemistry. Fundamental principles of SPME techniques are outlined in Chapter 2 to aid selection of extraction mode, device geometry and operating conditions for a given application. Sampling and sample preparation techniques frequently resemble engineering approaches on a smaller scale. Some analytical scientists feel uncomfortable when working in the engineering discipline. Engineering progress, however, often drives development of new analytical technologies. For example, optical fibre manufacturing process is presently used to produce GC capillary columns. The availability of coated fused silica fibres originally developed for telecommunication applications was instrumental in the practical implementation of the SPME approach. Similarly, recent advancements in micro-machining and wireless communication are expected to have a profound impact on future analytical devices. These engineering developments do not preclude analytical scientists from making substantial contributions to the evolution and implementation of the new technologies; rather, these developments generate new opportunities for them. For example, over the last several decades, our engineering colleagues have developed micro-machining technologies including micro-electromechanical systems and components currently used to construct micro-total analysis systems (µTAS) devices. As analytical researchers investigate in detail these new technologies developed by engineers, they will continue to find new and unique opportunities and applications in analytical science.11 Currently there is increased interest in the incorporation of sample preparation into miniaturised devices to enable on-site deployment and/or automation. The key to rational choice, design and optimisation of sample preparation components to facilitate this objective is based on an understanding of fundamental principles governing the mass transfer of analytes in multiphase systems.

1.3

Perspective on Microextraction Techniques

Microextraction techniques have been developed not only to address the need for a reduction in solvent use and in the size of extraction instrumentation but also to explore the ability of this approach to facilitate rapid and convenient sample preparation, both in the laboratory and for on-site applications. Microextraction systems

6

Handbook of Solid Phase Microextraction

also minimise the impact on the sampled system. There are many advantages of microextraction, which can be realised to a higher or lesser degree depending on the geometric configuration of the instrument. Some designs better address issues associated with agitation, while others address the ease of implementing on-site analyses or sample introduction to the analytical instrument. For example, full automation of standard delivery, extraction and introduction is possible for GC using a coated fibre format and for LC when an internally coated capillary is utilised. Conversely, the use of either coated fibres (in SPME) or hollow fibres that contain small amounts of solvent Liquid Phase Microextraction (LPME) arranged in a 96-fibre configuration facilitate application with a multi-well format for high-throughput sample processing.12 Small extraction devices facilitate on-site applications, including in vivo analyses, and allow for coupling to a variety of analytical micro-instrumentation, including capillary and micro-fluidic systems. The non-exhaustive microextraction techniques possess unique advantages because typically only a small portion of the target analyte is removed from the matrix. This feature allows for the monitoring of chemical changes, partitioning equilibria, and speciation in the investigated system because sampling causes minimal perturbation to the system. Therefore, the use of microextraction-based strategies results in better characterisation and more accurate information about the investigated system or process compared to exhaustive techniques. Non-exhaustive microextraction techniques provide signal magnitudes that are proportional to the free concentration of target analyte, defining the fraction of the analyte that is bioavailable. This unique feature of the non-exhaustive techniques allows for the measurement of binding constants in complex matrices, providing additional information about the investigated system. It also indicates, as mentioned above, the need for careful calibration and optimisation. Therefore, the development of robust quantitative analytical methods based on microextraction requires more time, but when the procedures are optimised, they are more convenient and cost-effective compared to conventional exhaustive extraction approaches. One can draw several parallels between the development and applications of extraction techniques and that of electrochemical methods. For instance, the coulometric technique corresponds to total or exhaustive extraction methods. Although it is the most precise, this technique is not used frequently because of the time required to complete it. Equilibrium potentiometric techniques are more frequently used (pH electrode), particularly when the sample is a simple mixture and/or the selectivity of the membrane in an ion-selective electrode is sufficient to quantify the target analyte in complex matrices. The equilibrium microextraction approach has further advantages in selectivity, because the extraction is coupled with separation and/or specific detection (e.g. mass spectrometry), which enables identification and quantification of many components simultaneously. The advantage of electrochemical methods is a short response time because of the low capacities of electrodes. Design of micro-systems with cylindrical geometry facilitates rapid extraction, as in micro-electrodes.13 Some electrochemical methods, e.g. amperometry, are based on mass transport through the boundary layers, as in pre-equilibrium extraction techniques (e.g. Time Weighted Average (TWA) and diffusion-based).

Solid-Phase Microextraction in Perspective

7

Analogously, extraction calibration based on diffusion coefficients can be accomplished as long as the agitation conditions are constant, the extraction times are short and the coating has high affinity for the analytes.

1.4

Implementations of SPME

SPME was developed to address the need for rapid sample preparation, both in the laboratory and at the site of the investigated system. In this technique, a small amount of extracting phase dispersed on a solid support is exposed to the sample for a well-defined period of time. In one approach, a partitioning equilibrium between the sample matrix and extraction phase is reached. In this case, convection conditions do not affect the amount extracted. A second approach uses short preequilibrium extraction times, and the amount of analyte extracted is related to time if convection/agitation is constant. Quantification can then be performed based on timed accumulation of analytes in the coating. Figure 1.3 illustrates several implementations of SPME, which include mainly open-bed extraction concepts such as coated fibres, vessels and agitation mechanism disks, but also in-tube approaches. Some devices better address issues associated with agitation, while others focus on the ease of introducing the sample to the analytical instrument. The fibre technique remains, to this date, the most-used SPME approach. It should be noted that SPME was originally named after the first experiment using an SPME device, which involved extraction on solid, fused silica fibres. Subsequently, the name was retained as a reference to the appearance of the extracting phase (relative to a liquid or gaseous donor phase), even though it is recognised that the extraction phase is not always technically a solid.

Extraction phase Sample Sample flow Tube Vessel walls

Fibre Particle

Suspended particles

Figure 1.3 Configurations of SPME.

Stirrer

Disk/membrane

8

1.5

Handbook of Solid Phase Microextraction

Miniaturisation and Integration

Practical integration of sample preparation with the rest of the analytical process has been accomplished in several ways. The concept of FIA has facilitated the performance of sequential sample preparation processes and quantification in a single device with the help of a flowing stream.14 For example, these devices can be made very small by using capillary flow systems integrated with small semiconductor light emission and detection devices that use fibre optics.15 In addition, they can be implemented on-site in combination with single solvent drop detection.16 Further miniaturisation of FIA technology results in a whole sample preparation process being performed in the body of a single valve (‘lab in a valve’).17 The application of micro-machining technology to the construction of highly integrated analytical systems (µTAS or ‘lab on a chip’) has recently resulted in sample preparation being performed in machined micro-channels.18 µTAS enables the effective coupling of separation/detection processes with sample preparation similar to capillary-based devices but can potentially be mass-produced at much lower cost. The integration of sample preparation in the micro-devices with the other steps of the analytical process can be accomplished in two fundamentally different ways. First, analogous to FIA, sample preparation may be performed directly in the capillaries/micro-channels in the flowing systems. This approach would typically use flow-through sample preparation techniques (see Figure 1.2). Such devices are expected to be structurally complex and relatively large because they must incorporate valves to control flows, pumps or high voltage supply, samples, reagent ports and detection components.19 In addition, because of the high surface area/volume ratio, there is a possibility of sample losses and carryover in a complex channel network. Integration of the sampling step with this complex system might be a challenge. It can be addressed by using moisture-repellent sorbents, electromigrationfocusing mechanisms, membranes and solvent microextraction when the mobile phase in the separation technique is a solvent. For example, attempts are being made to integrate Capillary Electrophoresis with sampling/concentration20 and a micro-machined Gas Chromatography system with sampling via a micro-sorbent trap (R Sacks, personal communication). Membrane sampling interfaced to an investigated system could facilitate sampling of aqueous media, as is currently performed in microdialysis systems coupled to condensed phase separations21 or membrane extraction with sorbent interface (MESI) coupled to micro-gas chromatography.22 Recent developments in manufacturing micro-fluidic systems using polymers such as polydimethylsiloxane23 will facilitate these approaches because such materials are excellent extraction phases. Because the overall size of the fully integrated device is expected to be relatively large, there will always be some restriction on the dimensions of the object that can be investigated. The most significant limitation of this approach, however, is expected to be the cost because unique configurations are required for each specific application. This restriction would make the approach cost-competitive only for very popular applications, when mass production is justified.

Solid-Phase Microextraction in Perspective

9

The alternative approach involves the integration of sampling with sample preparation only, by performing extraction and sample processing directly in the sampling device, followed by on-site introduction to a micro-separation/quantification instrument. The extraction process can be made very selective for target analytes, limiting disturbance of the investigated system. If the sampling/sample preparation device is small enough, it can deliver the prepared samples directly into the separation channel/capillary of the separation/quantification micro-device. For example, in-micro-needle24 and on-fibre micro-sampling devices could enable such a method because processing reagents can be either drawn into the needle or delivered to the fibre by dipping or by using a spray.25 Prepared analytes can subsequently be introduced to the micro-device such as GC, LC or CE for separation and quantification. Because sample preparation is performed directly in the sampling system, external to the separation/quantification device, restrictions applied to one device will not have to be arbitrarily applied to the other. Low-cost generic micro-separation/detection devices can be used as long as they are designed to accommodate a specific configuration of sampler. The major limitation of this approach is in monitoring and parallel analysis applications for which separate miniaturised automated systems would be required to control the device to perform sample preparation and, occasionally, sampling as well. It is, however, sometimes possible to prepare an extraction phase that already contains the required reagents before sampling.26 In this approach to on-site analysis, optimisation of the design of the sampling/sample preparation systems is conducted independently. Much smaller and more flexible devices are expected, compared with the previously described fully integrated single micro-device. Several of the sample preparation technologies listed in Figure 1.2, including batch extraction techniques such as coated fibres, can be explored for this application.

1.6

In Vivo Analysis

The sampling procedures in the integrated on-site micro-devices described above are a significant departure from conventional ‘sampling’ techniques, in which a portion of the system under study is removed from its natural environment and the compounds of interest are extracted and analysed in a laboratory environment. There are two main motivations for exploring these types of configurations. The first is the desire to study chemical processes in association with the normal biochemical milieu of a living system; the second is the lack of availability, or the impracticality, frequently associated with size, of removing suitable samples for study from the living system. New approaches, such as an externally coated extraction phase on a micro-fibre mounted in a syringe-like device, packed micro-needles or online sampling from a membrane interface, seem to be logical targets for the development of such tools. As with any microextraction or membrane technique, compounds of interest are not exhaustively removed from the investigated system. In fact, conditions can be devised in which only a small proportion of the total

10

Handbook of Solid Phase Microextraction

compounds are removed and none of the matrix is removed, thus avoiding disturbance of the normal balance of chemical components. Second, because the analytical device is either a syringe-like one that can be physically removed from the laboratory environment for sampling or an integrated micro-membrane system, it is suitable for monitoring a living system in its natural environment, rather than trying to move the living system to an unnatural laboratory environment. Microdialysis systems are already used in animal studies27 and MESI has been used in breath monitoring.28 The coated micro-fibre approach has recently been used in drug pharmacokinetic studies the components of interest were extracted directly from a peripheral vein of an animal.29 Chapter 12 discusses additional in vivo SPME applications in more detail. To further improve the capability of SPME for in vivo sampling, biocompatible and specific coatings (e.g. affinity phases) should be developed for a range of important target analytes. The ultimate goal is to remove only those compounds required to characterise the system investigated, and none of the matrix, using molecular recognition approaches, as is frequently done in sensor arrays.30 This specific direct extraction approach is critical to minimising interference with the operation of the system being investigated. For example, the removal of neurotransmitters from the synaptic cleft results in elimination of the signal coming down the nerves and/or depletion of presynaptic stores of the transmitter. In addition, specific non-equilibrium direct extraction might facilitate sampling at the speed of biological processes. The extraction can be limited to a small number of molecules and combined with on-probe amplification approaches and/or single molecular detection schemes, facilitating investigation of analytes present in the system at a low copy number.

1.7

SPME Versus SPE

SPME is often mistakenly considered another form of SPE or micro-SPE. There are significant differences between the methods, however. SPE is essentially a three-step process. Initially, a sample is passed through the sorbent bed, and analytes present in the sample are exhaustively extracted from the sample matrix to the solid sorbent. In a second step, unwanted analytes are selectively desorbed from the solid sorbent by washing with a solution capable of desorbing unwanted analytes but leaving desired analytes retained on the sorbent. In the final step, the wash solution is changed for one able to desorb analytes of interest. The resulting eluent may then be concentrated by evaporation to the desired volume. SPME, however, takes advantage of equilibrium extraction and selective sorption from the matrix onto the coating. In the first step, the coating is exposed to the sample. Analytes with a high affinity for the sorbent are selectively extracted. In the second step, everything extracted by the fibre is desorbed into the analytical instrument. No intermediate clean-up step is normally implemented. Micro-SPE is more related to SPE, as it is a total extraction method, but it utilises a reduced sample and sorbent volume. A comparison with SPME, therefore, is inappropriate.

Solid-Phase Microextraction in Perspective

11

A degree of selectivity is required for any sample preparation method. It is impractical to introduce all compounds present in a sample to an analytical instrument. The method developed must eliminate compounds that are incompatible with the instrument, including matrix components. It is also desirable to remove as many of the unwanted compounds as possible, to make the ensuing data interpretation as clean and simple as it can be. Thus, with selective extraction, sample preparation is simplified and typically results in significant time savings and improvement in precision. Selectivity is, therefore, quite important when choosing a SPME coating. High capacity, even for a range of analytes, is more important for SPE, where prevention of breakthrough is a significant concern. Because breakthrough is not an issue that needs to be addressed in an equilibrium extraction method such as SPME, more emphasis may be placed on sorbent selectivity. SPME differs from SPE in another significant way: because of the large volume of sorbent required relative to SPME, SPE sorbent has the potential to retain nonadsorbed components in the void volume. It is difficult to design a wash regime that removes unwanted compounds completely, without affecting retention of the analyte(s) of interest. Therefore, unwanted compounds may remain, either adsorbed or present as non-adsorbed analytes in the bulk of the sorbent. Because of the geometry of the SPME device and the modes of extraction used, unwanted analytes are not normally present in the sorbent at the time of desorption. SPME devices have an open-bed structure relative to SPE devices, where the extraction medium is packed into a cartridge-like device. In SPME, the surface of the extraction phase is itself accessible for analysis. Although this is less true with in-tube SPME, limited surface characterisation has been performed with the capillaries as well. Therefore, with SPME, it is possible to perform convenient spectroscopic analysis of surface-adsorbed components not only extracted chemical species but also collected aerosols or particulates. This can represent an important advantage in the speciation and characterisation of natural systems. There are numerous untapped opportunities available for exploration, especially considering the unique features of microextractions that have been emphasised above, making this research direction vital and scientifically rewarding. The objective of this handbook is to constitute a practical guide to SPME for researchers and analytical chemists.

References 1. E Thurman & M Mills, Solid Phase Extraction (1998) John Wiley: New York, NY 2. J Dean, Extraction Methods for Environmental Analysis (1998) John Wiley: New York, NY 3. A Handley, Ed, Extraction Methods in Organic Analysis (1999) Sheffield Academic Press: Sheffield 4. F Cantwell & M Losier, Liquid Liquid Extraction, in J Pawliszyn, Ed, Sampling and Sample Preparation for Field and Laboratory (2002) Elsevier: Amsterdam

12

Handbook of Solid Phase Microextraction

5. J Pawliszyn, Solid Phase Microextraction (1997) Wiley VCH: New York, NY 6. B Ioffe & A Vitenberg, Headspace Analysis and Related Methods in Gas Chromatography (1984) John Wiley: New York, NY 7. J Ruzicka & E Hansen, Flow Injection Analysis, 1st ed (1981) Wiley: New York, NY 8. S Stern, Membrane Separation Technology (1995) Elsevier: Amsterdam 9. K Pratt & J Pawliszyn, Anal Chem 64 (1992) 2101 10. M Yang, M Adams & J Pawliszyn, Anal Chem 68 (1996) 2782 11. D Reyes, D Iossifidis, P-A Auroux & A Manz, Anal Chem 74 (2002) 2623 12. J O’Reilly, Q Wang, L Setkova, J Hutchinson, Y Chen, H Lord, C Linton & J Pawliszyn, J Sep Sci 28 (2005) 2010 13. J Heinze, Angew Chem Int Ed Engl 32 (1993) 1268 14. Z-L Fang, Flow Injection Separation and Preconcentration (1993) VCH: Weinheim 15. J Pawliszyn, Anal Chem 58 (1986) 3207 16. H Liu & P Dasgupta, Anal Chem 67 (1995) 4221 17. C-H Wu, L Scampavia & J Ruzicka, Analyst 127 (2002) 898, and references therein 18. P Greenwood & G Greenway, Trends Anal Chem 21 (2002) 726 19. Y Huang, E Mather & J Bell, Anal Bioanal Chem 372 (2001) 49 20. LY Zhu, CH Tu & HK Lee, Anal Chem 73 (2001) 5655 21. R Blakely, S Wages, J Justice Jr, J Herndon & D Neil, Brain Res 308 (1984) 1 22. A Segal, T Gorecki, P Mussche, J Lips & J Pawliszyn, J Chromatogr 873 (2002) 13 23. J Ng, A Stroock & G Whitesides, Electrophoresis 23 (2002) 3461 24. A-P Wang, F Fang & J Pawliszyn, J Chromatogr A 1071 (2005) 127 25. J Pawliszyn, Solid Phase Microextraction, in J Pawliszyn, Ed, Sampling and Sample Preparation for Field and Laboratory (2002) Elsevier: Amsterdam 26. J Koziel, J Noah & J Pawliszyn, Environ Sci Technol 35 (2001) 1481 27. Y Song & C Lunte, Anal Chim Acta 400 (1999) 143 28. H Lord, W Yu, A Segal & J Pawliszyn, Anal Chem 74 (2002) 5650 29. H Lord, R Grant, M Walles, B Incledon, B Fahie & J Pawliszyn, Anal Chem 75 (2003) 5103 30. K Michael, L Taylor, S Schultz & D Walt, Anal Chem 70 (1998) 1242

2 Theory of Solid-Phase Microextraction Janusz Pawliszyn Department of Chemistry, University of Waterloo, Waterloo, Ontario, Canada

2.1

Introduction

An understanding of solid-phase microextraction (SPME) theory provides insight and direction when developing methods and identifies parameters for rigorous control and optimisation. Effective use of the theory minimises the number of experiments that need to be performed. The theory has been developed to understand the principal processes involved in SPME by applying basic fundamentals of thermodynamics and mass transfer. To simplify mathematical relationships, discussion of the SPME theory assumes ideal conditions. The theory for ideal extraction conditions can be very accurate for trace concentrations in simple matrices like air or drinking water at ambient conditions, when secondary factors, such as thermal expansion of polymers, changes in diffusion coefficients due to the presence of solutes in polymers and heterogeneity of the matrix, can be neglected. When conditions are more complex, the theory for ideal cases still approximates well some of the parameters and general relationships between parameters and extraction times or amounts extracted. In this chapter, we describe both the thermodynamics and the kinetics of the extraction process. The amount of analyte extracted at equilibrium conditions can be calculated using thermodynamic principles, while the extraction time can be estimated by solving differential equations describing mass transfer conditions in the system. Three basic extraction modes can be performed using SPME: a direct extraction, a headspace extraction and an extraction involving membrane protection. Figure 2.1 illustrates the differences among these modes. In the direct extraction mode (Figure 2.1A), the coated fibre is inserted into the sample and the analytes are transported directly from the sample matrix to the extracting phase. To facilitate rapid extraction, some level of agitation is required to increase the transport of analytes from the bulk of the solution to the vicinity of the fibre. For gaseous samples, natural convection of air is sufficient to facilitate rapid equilibration. For aqueous matrices, more efficient agitation techniques such as fast sample flow, rapid fibre or vial movement and stirring or sonication are required.1,2 These conditions are necessary to reduce the effect caused by the ‘depletion zone’ produced close to the fibre as a result of fluid shielding and slow diffusion coefficients of analytes in liquid matrices. Handbook of Solid Phase Microextraction. DOI: 10.1016/B978-0-12-416017-0.00002-4 © 2012 Elsevier Inc. All rights reserved.

14

Handbook of Solid Phase Microextraction

Sample headspace

Coating

Fibre

Sample (A)

Membrane

Sample

Coating (B)

(C)

Figure 2.1 Modes of SPME operation: (A) direct extraction; (B) headspace SPME; and (C) membrane-protected SPME.

In headspace sampling, the fibre is inserted into the headspace above the aqueous matrix. Only relatively volatile analytes are extracted. Headspace sampling is advantageous for samples with high-molecular-weight interferences (Figure 2.1B). For samples containing both non-volatile target analytes and high-molecularweight interfering compounds, such as humic acids or proteins, the application of direct or headspace SPME may be challenging. In such cases, use of restrictedaccess materials (RAMs) or membrane-protected SPME results in better reproducibility and accuracy (Figure 2.1C).

2.2

SPME Principle

The most widely used technique of sampling with SPME consists of exposing a small amount of extracting phase (coating), associated with a fibre, to the sample, for a predetermined amount of time (Figure 2.2).3 Typically, the microextraction process is considered complete when the analyte concentration reaches equilibrium in the sample matrix and the fibre coating. The equilibrium conditions can be described by Eq. (2.1) according to the law of mass conservation, if only two phases are considered (e.g. the sample matrix and the fibre coating): C0 Vs 5 CsN Vs 1 CfN Vf

ð2:1Þ

where CfN and CsN are equilibrium concentrations in the fibre coating and the sample, respectively. The distribution coefficient Kfs of the analyte between the fibre coating and sample matrix is defined as Kfs 5

CfN CsN

ð2:2Þ

Theory of Solid-Phase Microextraction

Fused silica fibre

15

Figure 2.2 Sample preparation with SPME: Vf, volume of fibre coating; Kfs, fibre/sample distribution coefficient; Vs, volume of sample; C0, initial concentration of the analyte in the sample.

Coating Vf Kfs Sample Vs Co

Equations (2.1) and (2.2) can be combined and rearranged into CfN 5 C0

Kfs Vs Kfs Vf 1 Vs

ð2:3Þ

Finally, the number of moles of analyte n extracted by the coating can be calculated from Eq. (2.4): n 5 CfN Vf 5 C0

Kfs Vs Vf Kfs Vf 1 Vs

ð2:4Þ

Equation (2.4) indicates that the amount of analyte extracted onto the coating (n) is linearly proportional to the analyte concentration in the sample (C0), which is the analytical basis for quantification using SPME. When the sample volume is very large, i.e. VscKfsVf, Eq. (2.4) can be simplified to n 5 Kfs Vf C0

ð2:5Þ

which points to the usefulness of the technique when the volume of the sample is unknown. In practice, this means that the fibre can be exposed directly to the flowing blood, ambient air, water and so on. The amount of extracted analyte will correspond directly to its concentration in the matrix without depending on the sample volume. Frequently, the extraction system is complex; for example, in a sample consisting of an aqueous phase with suspended solid particles having various adsorption interactions with analytes, plus a gaseous headspace. In some cases, specific factors have to be considered, such as analyte losses by biodegradation or adsorption on the walls of the sampling vessel. In the discussion below, we will consider only three phases: the fibre coating, the gas phase or headspace and a homogeneous matrix such as pure water or air. During extraction, analytes migrate between all three phases until equilibrium is reached.

16

Handbook of Solid Phase Microextraction

The mass of an analyte extracted by the liquid polymeric coating is related to the overall equilibrium of the analyte in the three-phase system. Because the total mass of an analyte should remain constant during the extraction, we have C0 Vs 5 CfN Vf 1 ChN Vh 1 CsN Vs

ð2:6Þ

where C0 is the initial concentration of the analyte in the matrix; CfN , ChN and CsN are the equilibrium concentrations of the analyte in the coating, the headspace and the matrix, respectively; and Vf, Vh and Vs are the volumes of the coating, the headspace and the matrix, respectively. If we define the coating/gas distribution constant as Kfh 5 CfN =ChN and the gas/sample matrix distribution constant as Khs 5 ChN =CsN , the mass of the analyte absorbed by the coating, n 5 CfN Vf , can be expressed as n5

Kfh Khs Vf C0 Vs Kfh Khs Vf 1 Khs Vh 1 Vs

ð2:7Þ

Also, Kfs 5 Kfh Khs 5 Kfg Kgs

ð2:8Þ

because the fibre/headspace distribution constant, Kfh, can be approximated by the fibre/gas distribution constant, Kfg, and the headspace/sample distribution constant, Khs, by the gas/sample distribution constant, Kgs, if the effect of moisture in the gaseous headspace can be neglected. Thus, Eq. (2.7) can be rewritten as n5

Kfs Vf C0 Vs Kfs Vf 1 Khs Vh 1 Vs

ð2:9Þ

The equation states, as expected from the equilibrium conditions, that the amount of analyte extracted is independent of the location of the fibre in the system. It may be placed in the headspace or directly in the sample as long as the volumes of the fibre coating, headspace and sample are kept constant.

2.3

Thermodynamics

2.3.1

Distribution Constant

The fundamental thermodynamic principle common to all chemical extraction techniques involves the distribution of the analyte between the sample matrix and the extraction phase. When a liquid is used as the extraction medium, the distribution constant, Kes, can be figured as follows: Kes 5

ae Ce 5 as Cs

ð2:10Þ

Theory of Solid-Phase Microextraction

17

This equation defines the equilibrium conditions and ultimate enrichment factors achievable by use of the technique; ae and as are the activities of analytes in the extraction phase and matrix, respectively, and can be approximated by the appropriate concentrations. This physicochemical constant, which reflects the chemical composition of the extraction phase, has been discussed in detail in fundamental chromatographic literature because it determines the retention and selectivity of a separation column. Although chromatography is frequently used to determine distribution constants, convenient sample-preparation techniques (e.g. SPME) can also be used to provide information about the thermodynamics of the partitioning process.4 Kes can be estimated using a variety of properties characteristic to matrix and analyte5 analogues, as is typically done when estimating octanolwater distribution constants (Kow).6 Chromatographic retention times obtained by using appropriate mobile and stationary phases corresponding to the extractant and the sample matrix, respectively, can be used occasionally to estimate the distribution constant.3 The extraction phase/samplematrix distribution constants are thermodynamic constants that depend on a variety of conditions, including temperature, pressure and sample matrix characteristics such as pH, salt and organic component concentration. For a solid extractant, adsorption equilibria can be explained using this equation: s 5 Kes

Se Cs

ð2:11Þ

where Se is the solid extraction phase surface concentration of adsorbed analytes. The above relationship is similar to Eq. (2.6) except for the replacement of the extraction phase concentration with the surface concentration. The Se term in the numerator indicates that the sorbent surface area available for adsorption also must be considered. This complicates calibration under equilibrium conditions because of displacement effects and the non-linear adsorption isotherm.1

2.3.2

Estimation of Distribution Constants

In addition to direct partitioning measurements, distribution constants can also be estimated from physicochemical data and chromatographic parameters. For example, distribution constants between a fibre coating and a gaseous matrix (e.g. air) can be estimated using isothermal gas chromatography (GC) retention times on a column with a stationary phase identical to the fibre coating material.4,7 The partitioning process in GC is analogous to the partitioning process in SPME, and there is a welldefined relationship correlating distribution constants and retention times. The nature of the gaseous phase does not affect the distribution constant unless components of the gas, such as moisture, swell the coating polymer, thus changing its properties. The formula that correlates the distribution constant and the retention time is Kfh 5 Kfg 5 ðtR 2 tA ÞF

T pm 2 pw 3 ðpi =po Þ2 2 1 1 Tm pm 2 ðpi =po Þ3 2 1 VL

ð2:12Þ

18

Handbook of Solid Phase Microextraction

where tR and tA are the retention times of the solute and a nonsorbed compound, respectively; F is the column flow measured by a soap-bubble flow meter; T and Tm are the temperatures of the column and flow meter, respectively; pm and pw are the flow meter pressure and the saturated water vapour pressure, respectively; pi and po are, respectively, the inlet and outlet pressures of the column; and VL is the column’s stationary phase volume. Usually, pm and po are equal to atmospheric pressure. The Kfg estimated by this method for the polydimethylsiloxane (PDMS) to gas partitioning of benzene agrees within a few percentage points with the value determined by SPME experimentation. A most useful method for determining the coating-to-gas distribution constants for PDMS extraction phases uses the linear temperature programmed retention index (LTPRI) system, which indexes compounds’ retention times relative to the retention times of n-alkanes. This system is applicable to retention times for temperature-programmed gasliquid chromatography. The logarithm of the coating-to-air distribution constants of n-alkanes can be expressed as a linear function of their LTPRI numbers (Figure 2.3). For PDMS, this relation is6 log Kfg 5 0:00415 LTPRI 2 0:188

ð2:13Þ

Thus, the LTPRI system permits interpolating the curve of Kfg versus retention time. LTPRI numbers are available from published tables, so this method can estimate Kfg values accurately without experimenting. If the LTPRI number for a 6.00

5.00

log Kfg

4.00

3.00

R2 = 0.99989

2.00

1.00

0.00 400

500

600

700

800

900

1000 1100 1200 1300 1400 1500

LTPRI

Figure 2.3 log Kfg (PDMS) as a function of LTPRI for n-alkanes at 25 C.

Theory of Solid-Phase Microextraction

19

compound is not available from published sources, it can be calculated from a GC run according to its definition:


tRðAÞ 2 tRðN Þ LTPRI 5 100 N 1 100 tRðN 1 1Þ 2 tRðN Þ

 ð2:14Þ

where N is the number of carbon atoms for tR(N), tR(A) is the analyte retention time, tR(N) is the n-alkane retention time less than tR(A) and tR(N 11) is the n-alkane retention time greater than tR(A). Note that the GC column used to determine LTPRI should be coated with the same material as the fibre coating. The fibre coating/water distribution constants can be calculated from the following equation: Kfw 5 Kfg Kgw

ð2:15Þ

where Kfg can be calculated from chromatographic data, as discussed above, by using LTPRI. Kgw is the gas/water distribution constant (Henry’s constant), which can be obtained from physicochemical tables or can be estimated by the structural unit contribution method.2 For example, the equation for a PDMS coating and aqueous matrix is log Kfw 5 0:00415 LTPRI 2 0:188 1 log10 Kgw

ð2:16Þ

This equation is obtained by substituting the expression for Kfg given in Eq. (2.13) to Eq. (2.15) (see Table 2.1). The results agree very well, considering that the errors in determination of Henry’s constants are typically above 10%. Therefore, by finding the relationship between Kfg and LTPRI for a given coating, the appropriate distribution constant can be conveniently calculated from chromatographic data and literature values of Henry’s constants. In addition, Henry’s constants are similar for closely related compounds, as illustrated in Figure 2.4, resulting in a single linear relationship between Kfw and LTPRI, characteristic for a group of different types of analytes. As expected, the slope of the curve has a value close to the slope for the curve in Figure 2.3, but the intercept varies and corresponds to a sum of the intercept value from Eq. (2.9), plus the average value of Henry’s constant for a given group of analytes. For example, as expected from their high Henry’s constant values, Kfw values Table 2.1 Comparison of Kfw for PDMS/Water Obtained by the log KfwLTPRI Relationship and the Experimentally Determined Kfw Values by Direct SPME at 22 C Hydrocarbon

LTPRI

Kfw by LTPRI

Kfw by SPME

Benzene Toluene o-Xylene

638.6 747.1 867.9

60 184 565

58 189 485

20

Handbook of Solid Phase Microextraction

logKfw = 0.0042 × LTPRI – [0.188 + logKaw]

5.50 5.00

Branched alkanes logKfw = 0.0041 × LTPRI + 1.33

logK (coating/water)

4.50 4.00 3.50 3.00

Benzene substituted aromatics Cyclopentane and cyclohexane derivatives

2.50

logKfw = 0.0043 × LTPRI – 0.88

logKfw = 0.0040 × LTPRI + 0.69

2.00

Kfw =

1.50 1.00

0

200

400

600

800

Cf Cw

=

1000

Cf

Ch

Ch

Cw

= KfhKhw

1200

1400

LTPRI Figure 2.4 log Kfw (PDMS) as a function of LTPRI for isoparaffins, substituted benzenes (aromatics) and cycloalkanes at 25 C.

for paraffins are larger than those for aromatic analytes. Because of this linear relationship, quantitation with minimum identification is possible, as long as a detector can selectively assign extracted analytes to appropriate groups of compounds.8 Some correlations can be used to anticipate trends in SPME coating/water distribution constants for analytes. For example, many investigators have reported the correlation between octanol/water distribution constant Kow and Kfw. This is expected because Kow is a very general measure of the affinity of compounds to the organic phase. It should be remembered, however, that the trends are valid only for compounds within the same group having similar structures, such as hydrocarbons, aromatics or phenols. Kow values cannot be used to compare different groups of compounds because of different analyte activity coefficients in the polymer. Another more universal approach is to use the common fragment constants applied for a given coating phase based on theories developed for and tested by liquid chromatography.911 In this method, contributions of various functional groups present in a given molecule are added together to estimate the appropriate constant.

2.3.3

Effect of Extraction Parameters on Distribution Constants

Thermodynamics predicts the effects of modifying certain extraction conditions on partitioning and indicates parameters to control for reproducibility. The theory can be used to optimise the extraction conditions using a minimum number of experiments and to correct for variations in extraction conditions without repeating calibration tests in the new conditions. For example, SPME analysis of outdoor air may be done at ambient temperatures that can vary significantly. The equation that

Theory of Solid-Phase Microextraction

21

predicts the effect of temperature on the amount extracted will allow calibration without the need for extensive experimentation. Extraction conditions that affect Kfs include temperature, salting, pH and organic solvent content in water.

2.3.3.1 Temperature If both sample and fibre temperature change from T0 to T, the distribution constant changes according to the following equation: 


 ΔH 1 1 2 Kfs 5 K0 exp 2 R T T0

ð2:17Þ

where K0 is the distribution constant when both fibre and sample are at temperature T0 (in degrees Kelvin), ΔH is the molar change in enthalpy of the analyte when it moves from sample to fibre coating and R is the gas constant.12 The enthalpy change, ΔH, is considered constant over temperature ranges typical for SPME experiments. It can be determined by measuring Kfs at two different temperatures. For coating/gas distribution constants, ΔH for a volatile compound is well approximated by the heat of vapourisation of the pure compound ΔHv for PDMS.3,13 Temperature effects must be considered when temperature variations occur while sampling outdoors and when heating is used to increase extraction rates, stop metabolic activity or enhance the release of analytes. When the Kfs value for an analyte is greater than 1, the analyte has a lower potential energy in the fibre coating than in the sample, so the analyte partitioning into the fibre must be an exothermic process (i.e. giving off heat), which means ΔH (the molar change in enthalpy of the analyte when it moves from sample to fibre coating) is greater than 0. Therefore, Eq. (2.17) shows that raising the temperature will decrease Kfs. This effect is illustrated in Figure 2.5. 5.00 n-Undecane

4.50

Mesitylene

log (K)

4.00

p-Xylene

3.50 3.00

Benzene

2.50 n-Pentane

2.00 1.50 3.20

3.25

3.30

3.35

3.40

1/Temperature (1/K) (× 10–3)

Figure 2.5 Effect of temperature on Kfg.

3.45

3.50

22

Handbook of Solid Phase Microextraction

Equation (2.17) applies only to partitioning between two homogeneous phases. The equation is not valid for partitioning between a fibre and a multicomponent sample but still can be used to estimate the effect. For example, for the extraction of benzene into PDMS from the headspace above an aqueous solution at 25 C, Kfw 5 125, Kfh 5 493, ΔHfw has been estimated as 13.9 kJ/K mol3 and ΔHfw can be estimated by the change in vapourisation enthalpy (42.8 kJ/K mol for benzene), where the subscript ‘w’ represents water and the subscript ‘h’ represents headspace. Therefore, a 100-μm PDMS fibre extracting benzene from 3 mL of headspace above a 2-mL, 100-ppb aqueous solution will extract 5.5 ng at 25 C according to Eq. (2.9). The amount extracted at 90 C can be calculated from Eqs (2.17) and (2.9): n5

n0 1 1ð1=Vf ÞVw Kfw exp½ðΔHfw =RÞð1=T 21=T0 Þ 1ð1=Vf ÞVh Kfh exp½ðΔHfh =RÞð1=T 21=T0 Þ

ð2:18Þ

which predicts 0.75 ng. Equation (2.17) does not apply directly to adsorption partitioning.

2.3.3.2 Salting Two common techniques used to enhance the extraction of organics from aqueous solutions are salting and pH adjustment. Salting can increase or decrease the amount extracted, depending on the compound and salt concentration, and the effect of salting on SPME has been determined to date only by experiment. In general, the salting effect increases with increasing compound polarity. Figure 2.6 illustrates the effect of salting on extraction of benzene and toluene from an aqueous matrix. A substantial increase of analyte extraction occurs at salt concentrations above 1% and leads to about an order of magnitude increase in sensitivity at the 30% level. Saturation with salt can be used not only to lower the detection limits of the analysis but also to normalise random salt concentration in natural matrices.

Log peak area

Figure 2.6 The effect of salt on extraction of toluene and benzene by SPME.

Toluene

6.0

5.8

5.6

Benzene

5.4 –4

–3

–2

–1

log (%NaCl)

0

1

2

Theory of Solid-Phase Microextraction

23

Note that salting can lower pH at high salt concentration levels because proton activity is increased with increased solution ionic strength. The effect of salting on SPME has not been examined theoretically, although theories have been developed for the effect of salting on liquidliquid extraction. Setchenow’s equation is   ðks Cs Þ2 ðks Cs Þ3 1 1? Kfs 5 K0 eks Cs 5 K0 1 1 ks Cs 1 2! 3!

ð2:19Þ

where K0 is the partition with no salt, ks is a constant and Cs is the salt concentration. This equation predicts salt effects, at least for low salt concentrations and sometimes for high concentrations as well. Long and McDevit14 have summarised the theory and also tabulated most published research on salting-out and salting-in up to 1951 in an extensive review article.

2.3.3.3 pH Assuming that only the undissociated form of the acid or base can be extracted by the fibre coating, adjusting the pH of an aqueous solution will change K for dissociable species, according to the following equation: K 5 K0

½H 1  Ka 1 ½H 1 

ð2:20Þ

where K0 is the distribution constant between the sample and the fibre of the undissociated form and Ka is the acidity constant of the dissociable analyte. This relation was confirmed by Yang and Peppard,15 whose results for the extraction of acids are illustrated in Figure 2.7. As pH decreases, more acid is present in neutral forms which partition into the coating, resulting in higher sensitivity. To obtain the highest sensitivity, pH needs to be two units lower than the pK value corresponding to the acid. Figure 2.7 The effect of pH on the SPME of acid compounds.

1.0 0.8

C7H15COO–

C7H15COOH

0.6 0.4 0.2 0 0

2

6

4

pH

8

10

24

Handbook of Solid Phase Microextraction

2.3.3.4 Polarity of Sample Matrix and Coating Material The presence of an organic solvent in water changes K according to the following equation:
 P1 2 P2 ð2:21Þ Kfs 5 2:303Kfw exp 2 where Kfw is the distribution constant for the analyte between fibre and pure water, P1 5 10.2 is the polarity parameter for water and P2 5 cPs 1 (1 2 c)P1 is the water/solvent mixture polarity parameter for a solvent of concentration c and polarity parameter Ps.2 This equation allows the prediction of the distribution constants for water heavily contaminated with miscible solvents, assuming that the solvent does not cause the coating to swell. This relationship indicates that the concentration of the solvent must be above 1% to change the properties of water and the distribution constant substantially. Figure 2.8 illustrates the decrease in extacted amount of benzene, toluene, ethylbenzene and xylenes (BTEX) into PDMS coating with the increase of methanol concentration in an aqueous matrix. Variation of organic composition in the matrix can be compensated for by using internal standard calibration techniques. Compounds have an affinity for a phase of similar polarity. Conzen et al.16 discussed the influence of polarity on fibre extraction of organics from water. The dielectric constant of PDMS varies from 2.6 to 2.8, depending on polymer molecular weight, and the dielectric constant of PA varies from 2.6 to 3.6.17 These polymer dielectric constants are similar to those of common organic compounds: for example, ε (toluene) 5 2.4 and ε (acetic acid) 5 4.1.18 The polymer and organic ε values are very low compared to that of water at room temperature (ε (H2O) 5 78).9 The trends in dielectric constants indicate high Kfs values for typical organics distributed between fibre coating and water.

2.3.4

Distribution Constants in the HeatingCooling Environment

High temperature allows the extraction of semi-volatile analytes and more efficient release of analytes from the matrix. However, a loss of sensitivity occurs because Figure 2.8 The effect of solvent on the SPME.

Log peak area

7 6 5

m,p-Xylene o-Xylene

4

Ethylbenzene Toluene Benzene

3 –1

0

1 log (%MeOH)

2

3

Theory of Solid-Phase Microextraction

25

of the corresponding decrease in the distribution constant. When heating of the sample is combined with simultaneous cooling of the fibre coating, a temperature gap is created between the hot headspace and the cooled fibre coating. This gap provides an additional advantage. In this heating/cooling environment, the coating/ headspace distribution constants of analytes increase dramatically.19 In headspace SPME, there are two processes involved: the release of analytes from their matrix and the absorption of vapourised analytes by the fibre coating. With the assumption that most analyte molecules can be released into the headspace during extraction, for the following discussion, we can simplify headspace SPME into two phases: the fibre coating and the headspace. Extraction at elevated temperatures enhances Henry’s law constants, increasing the concentrations of the analytes in the headspace. This results in rapid extraction by the extraction phase. The coating/sample distribution coefficient also decreases with increasing temperature, however, which results in reduction of the equilibrium amount of analyte extracted. To prevent this loss of sensitivity, the extraction phase can be cooled simultaneously with heating the sample. This ‘cold finger’ effect results in increased accumulation of the volatilised analytes in the extraction phase. This enhancement in the sample matrixextraction phase distribution constant associated with the temperature gap present in the system can be described by the following equation19: KT 5 K0


 Cp ΔT Ts Te exp 1 ln Te R Te Ts

ð2:22Þ

where KT 5 Ce(Te)/Cs(Ts) is the distribution constant of the analyte between the cold extraction phase on the fibre having temperature Te and the hot headspace at temperature Ts, Cp is the constant-pressure heat capacity of the analyte, ΔT 5 TsTe and K0 is the coating/headspace distribution constant of the analyte when both coating and headspace are at temperature Te. Due to the enhancement of the sample matrixextraction phase distribution constant, quantitative extraction of many analytes, including volatile compounds, is possible with this method. From Eq. (2.22), distribution constants (KT) of analytes between the two phases at different temperatures can be calculated. Table 2.2 lists KT values of BTEX compounds at selected headspace temperatures. Because KT values for the three xylene isomers are very similar, only o-xylene values are listed in the table. (Values for Cp are taken from the Thermodynamic Research Center’s tables.20) During the calculation, the average Cp value [listed as Cp(ave) in Table 2.2] of the compounds in the temperature range is used. The coating is PDMS and Tf 5 298 K (25 C). The distribution constants of BTEX at Ts 5 Tf(K0) are also listed in Table 2.2. Table 2.2 shows that if the coating temperature is maintained at 25 C as the headspace temperature goes up, the coating/headspace distribution constant of the analyte increases dramatically as well. For example, toluene has a distribution constant of 1,330 when both the coating and headspace are at 25 C. This value

26

Handbook of Solid Phase Microextraction

Table 2.2 KT Values (Calculated from Eq. (2.22), for Tf 5 25 C) and Other Parameters of Selected Compounds Ts (K)

Cp (J/K mol)

Benzenea 300 350 400 450 500 Tolueneb 300 350 400 450 500 Ethylbenzenec 300 350 400 450 500 o-Xylened 300 350 400 450 500

83.02 113.52 139.35 104.42 139.91 170.77 128.19 169.95 206.58 133.01 170.46 204.32

Cp (ave)

KT/K0

KT

82.73 90.36 97.98 104.44 110.9

1.01 1.36 2.36 5.15 14.18

496 670 1,161 2,539 6,993

104.08 112.96 121.83 129.54 137.26

1.01 1.41 2.7 6.92 23.57

1,330 1,867 3,571 9,144 31,153

127.8 138.24 148.68 157.83 166.99

1.01 1.47 3.15 9.65 41.77

3,286 4,804 10,294 31,500 136,418

132.66 142.02 151.38 159.85 168.32

1.01 1.48 3.2 9.88 42.85

4,444 6,537 14,139 43,622 189,260

At 298 K, K0 5 493; Cp 5 82.44 J/K mol. At 298 K, K0 5 1,322; Cp 5 103.75 J/K mol. At 298 K, K0 5 3,266; Cp 5 127.4 J/K mol. d At 298 K, K0 5 4,417; Cp 5 132.31 J/K mol. a

b c

increases to 31,153 when the headspace temperature is 227 C and the coating temperature is 25 C. With distribution constants known and the assumption that analytes can be completely released from the matrix, the mass of analytes extracted by the coating can be estimated in advance for the headspace extraction by the internally cooled SPME using Eq. (2.9). The coating/headspace distribution constant Kfs in this equation can be replaced by K0 if both the coating and headspace are at the same temperature and by KT if the coating and headspace are at different temperatures. This dramatic increase frequently leads to exhaustive extraction.

2.3.5

Multiphase Equilibria in SPME

SPME is a multiphase equilibration process. To simplify the system, however, only three phases will be considered initially: the fibre coating, the gas phase or

Theory of Solid-Phase Microextraction

27

headspace and a homogeneous matrix, such as pure water. During extraction, analytes migrate among the three phases until equilibrium is reached. The discussion below is limited to partitioning equilibrium involving liquid polymeric phases such as PDMS. Analysis for solid sorbent coatings is analogous because the total surface area available for adsorption is proportional to the coating volume, if constant porosity of the sorbent is assumed. The mass of an analyte extracted by the polymeric coating is related to the overall equilibrium of the analyte in the three-phase system as follows (see also Section 2.2): C0 Vs 5 CfN Vf 1 Ch VhN 1 CsN Vs

ð2:6Þ

where C0 is the initial concentration of the analyte in the matrix; CfN , ChN and CsN are the equilibrium concentrations of the analyte in the coating, the headspace and the matrix, respectively; and Vf, Vh and Vs are the volumes of the coating, the headspace and the matrix, respectively. If we define the coating/gas distribution constant as Kfh 5 CfN =ChN and the gas/sample matrix distribution constant as Khs 5 ChN =CsN , the mass of the analyte absorbed by the coating, n 5 CfN Vf , can be expressed as n5

Kfh Khs Vf C0 Vs Kfh Khs Vf 1 Khs Vh 1 Vs

ð2:7Þ

The proper expression for the above distribution constants would involve appropriate activities. However, concentration can serve as a good approximation of activity, considering the analytes are present in trace levels in the sample and assuming a pure matrix. The driving force in multiphase equilibrium is the difference between an analyte’s chemical potential in the three phases. SPME can be used to analyse compounds in various matrices, and the fibre coating can consist of immobilised liquid polymers or porous solid materials. In the following text, an analyte in a multiphase system including a liquid polymer, headspace and an aqueous matrix is used for the discussions of the equilibration theory. The general conclusions drawn from the discussion should be valid for other types of coatings and matrices. The chemical potential of an analyte in the headspace can be expressed as21
 ph μh 5 μ0 ðTÞ 1 RT ln 0 p

ð2:23Þ

where μh is the chemical potential of the analyte in the headspace, ph is the vapour pressure of the analyte in the headspace and μ0(T) is the chemical potential of the analyte at standard pressure p0 (usually p0 5 1 atm) and temperature T. Meanwhile, the chemical potentials of the analyte in the coating and the aqueous matrix can be expressed as
 pf μf 5 μ ðTÞ 1 RT ln 0 p 0

ð2:24Þ

28

Handbook of Solid Phase Microextraction

μs 5 μ0 ðTÞ 1 RT ln


 ps p0

ð2:25Þ

where μf and μs are the chemical potentials of the analyte in the coating and the aqueous matrix, respectively; and pf and ps are the vapour pressures of the analyte, in equilibrium with the analyte in the coating and the aqueous matrix, respectively. When the three-phase system is at equilibrium, the chemical potentials of the analyte in all three phases must be equal: μf 5 μh 5 μs

ð2:26Þ

From Eqs (2.23)(2.26), we can write pf 5 ph 5 p s

ð2:27Þ

According to Henry’s law,1 we have pf 5 KF CfN

ps 5 KH CsN

ð2:28Þ

where KF and KH are Henry’s constants of the analyte in the liquid polymer coating and the aqueous solution, respectively. Assuming that the ideal gas law, phVh 5 nhRT (where nh is the number of moles of the analyte in the headspace), is valid for the analyte vapour in the headspace, as follows: ph 5 ChN RT

ð2:29Þ

From Eqs (2.27)(2.29), we can easily connect the distribution constants with Henry’s constants as follows: Kfh 5

CfN RT 5 N Ch KF

ð2:30Þ

Khs 5

ChN KF 5 N Cs RT

ð2:31Þ

Khs 5

ChN KF 5 N Cs RT

ð2:32Þ

For direct SPME sampling from an aqueous solution, we have μf 5 μs or pf 5 ps at equilibrium. The distribution constant of the analyte, Kfs, between the coating and the aqueous solution can be expressed as Kfs 5 CfN =CsN 5 KH =KF because

Theory of Solid-Phase Microextraction

29

pf 5 KF CfN , ps 5 KH CsN and pf 5 ps when equilibrium is reached. It is intuitive based on Eqs (2.30) and (2.31) that Kfs 5

KH 5 Kfh Khs 5 Kfg Kgs KF

ð2:33Þ

because the fibre/headspace distribution constant, Kfh, can be approximated by the fibre/gas distribution constant Kfg and the headspace/sample distribution constant, Khs, by the gas/sample distribution constant, Kgs, if the effect of moisture in the gaseous headspace can be neglected. Then, Eq. (2.7) can be rewritten as n5

Kfs Vf C0 Vs Kfs Vf 1 Khs Vh 1 Vs

ð2:9Þ

The equation states that, as expected from the equilibrium conditions, the amount of analyte extracted is independent of the location of the fibre in the system. It may be placed in the headspace or directly in the sample, as long as the volume of the fibre coating, headspace and sample is kept constant. There are three terms in the denominator of Eq. (2.9), which describe the analyte capacity of each phase: fibre (KfsVf), headspace (KhsVh) and the sample itself (Vs). If we assume that the vial containing the sample is fully filled with the aqueous matrix (with no headspace), the term KhsVh in the denominator, which is related to the capacity ðChN Vh Þ of the headspace, can be eliminated, resulting in the following: n 5 CfN Vf 5 C0

Kfs Vs Vf Kfs Vf 1 Vs

ð2:4Þ

Both Eqs (2.9) and (2.4) describe the mass absorbed by the polymeric coating after equilibrium has been reached. For most analytes, Khs is relatively small (e.g. benzene has a Khs value of 0.26), and sampling from the headspace will not affect the mass absorbed by the coating, if the volume of the headspace is much lower than that of the aqueous solution (Vh{Vs). The detection limits of headspace SPME are, therefore, expected to be very similar to those of the direct SPME for these conditions. The relationship Kfs 5 KfhKhs in the headspace SPME can be generalised for multiphase equilibration systems (heterogeneous matrices). If there are n phases (e.g. different solids) present other than the coating and matrix during extraction, and, for convenience, they are numbered from 2 to n starting from the one closest to the coating and ending at the one next to the matrix, the distribution constant between the coating and the matrix (Kfs 5 CfN =CsN , where CfN and CsN are the analyte’s equilibrium concentrations in the coating and matrix, respectively) can be expressed as i 5 n 21

Kfs 5 Kf1 K12 K23 ?Kn 21;n Kns 5 Kf 1 Kns L Ki i51

ð2:34Þ

30

Handbook of Solid Phase Microextraction

N N where Kf1 5 CfN =C1N, Ki;i 1 1 5 CiN =CiN are the distribution 1 1 and Kns 5 Cn =Cs constants of coating/first phase, ith phase/(i 11)th phase and nth phase/matrix. The mass of an analyte extracted by the coating from that matrix is

n5

Kfs Vf C0 Vs Kfs Vf C0 Vs 5 P n Kfs Vf 1 K1s V1 1 K2s V2 1 ? 1 Kns Vn 1 Vs Kfs Vf 1 ii 5 5 1 Kis Vi 1 Vs ð2:35Þ

where Kis 5 CiN =CsN is the distribution constant of the analyte between the ith phase and the matrix of interest, which can be similarly determined through Eq. (2.4). Equation (2.35) turns into Eq. (2.4) when there is no intermediate phase during extraction and into Eq. (2.9) when there is a headspace present. Some real samples are heterogeneous and consist of many immiscible phases. The ability of the fibre coating to extract an analyte, as Eq. (2.35) suggests, is closely related to (i) the distribution constant (Kfs) of an analyte between that particular sample matrix and the coating, which is independent of the number of phases existing during extraction (refer to Eq. (2.34)), and (ii) the capacities of the other phases present in the sample for retaining the analyte. If the capacities of those in-between phases are small (such as for headspace), the mass of the analyte extracted by the fibre coating will not be affected significantly. However, the addition of an appropriate phase with high affinity towards interferences, but not towards target analytes, can be used to remove unwanted compounds from the fibre coating/sample matrix system. However, it should be emphasised that if those inbetween phases are liquid and the analyte has low diffusion coefficients in them, the mass transfer may be slow and the extraction process is kinetically limited. In addition, the rate of release of the analyte from solid matrices in the system can also control the overall mass transfer between the phases22,23: n5

Kes Ve C0 Vs P m Kes Ve 1 ii 5 5 1 Kis Vi 1 Vs

ð2:36Þ

where Kis 5 CiN =CsN is the distribution constant of the analyte between the ith phase and the matrix of interest, and Kes is the extraction phase/sample matrix distribution constant.24 Equation (2.36) simplifies to Eq. (2.4) if there are no competing phases in the sample matrix. Equation (2.36) clearly indicates the main challenge in SPME method optimisation: variations of matrix composition directly result in different amounts of analyte extracted by SPME. This means that careful calibration is necessary to compensate for the matrix effects when using SPME as the sample preparation method. However, this limitation represents a unique opportunity to study the partitioning properties of different phases present in the system. For example, SPME has been used to investigate the distribution of species in multiphase systems, both at equilibrium and prior to equlibrium, in order to understand the kinetics of the partitioning process. Thus, based on external calibration, SPME will give information about

Theory of Solid-Phase Microextraction

31

the concentration of chemical species in the phase of interest. Chapter 11 discusses the applications of SPME to investigate equilibria in biological matrices. The equations given above can be used to calculate the amount of analyte in the extraction phase, under equilibrium conditions. For equilibrium liquid microextraction techniques and large samples, including direct extraction from an entire investigated system, the appropriate expression is very simple25: n 5 Kes Ve Cs

ð2:37Þ

where Kes is the extraction phase/sample matrix distribution constant, Ve is the volume of the extraction phase and Cs is the concentration of the sample. This equation is valid when the amount of analytes extracted is insignificant compared with the amount of analytes present in a sample (large Vs and/or small Kes), resulting in negligible depletion of analyte concentration in the original sample. In Eq. (2.37), Kes and Ve determine the sensitivity of the microextraction method, whereas Kes determines its selectivity. The sample volume can be neglected, thus integrating sampling and extraction without the need for a separate sampling procedure, as discussed in more detail later. The non-depletion properties of the small dimensions typically associated with microextraction systems result in minimum disturbance of the investigated system, facilitating convenient speciation, investigation of multiphase distribution equilibria and repeated sampling from the same system in order to follow a process of interest. When significant depletion occurs, the sample volume, Vs, has some impact on the amount extracted and, therefore, on sensitivity.26 The amount extracted in this case can be calculated by using the equation: n5

Kes Ve C0 Vs Kes Ve 1 Vs

analogous to Eq. (2.4).

2.3.6

Matrix Effects

Two potential complications are typically observed when extracting analytes from complex matrices. One is associated with competition among different phases for the analyte and the other with the fouling of the extraction phase, due to the adsorption of macromolecules such as proteins and humic materials at the interface. The components of heterogeneous samples (including headspace, immiscible liquids and solids) partition in the multiphase system and are less available for extraction. This effect depends on analyte affinity and the volume of the competing phases and can be estimated, if appropriate volumes and distribution constants are known. The mass of an analyte extracted by an extraction phase in contact with a multiphase sample matrix can be calculated using Eq. (2.35), as previously discussed.

32

Handbook of Solid Phase Microextraction

Sample matrix Figure 2.9 Introduction of a barrier between the sample matrix and the Porous extraction phase to restrict transport of barrier

(A)

high-molecular-weight interferences. Extraction phase

(B)

Headspace Sample matrix . Analyte

Interference

The typical approach used to reduce fouling involves the introduction of a barrier between the sample matrix and the extraction phase to restrict transport of high-molecular-weight interferences (Figure 2.9). For example, the extraction phase can be surrounded by a porous membrane with pores smaller than the size of the interfering macromolecules (Figure 2.1C); e.g. the use of a dialysis membrane with the appropriate molecular cutoff. This approach is conceptually similar to membrane dialysis from complex matrixes, in which the porous membrane is used to prevent large molecules from entering the dialysed solution.27 Membrane separation has been used to protect SPME fibres from humic material.28 More recently, hollow fibre membranes have been used in solvent microextraction, both to support the small volume of solvent and to eliminate interferences when extracting biological fluids.29 This concept has been further explored by integrating a protective structure and the extraction phase in individual sorbent particles, resulting in RAM.30 The chemical nature of the small inner pore surface of the particles is hydrophobic, facilitating extraction of small target analytes, whereas the outer surface is hydrophilic, thus preventing the adsorption of excluded large proteins. In practice, fouling of the hydrophobic interface occurs to a great extent only when the interfering macromolecules are hydrophobic in nature. A gap made of gas is also a very effective separation barrier (Figure 2.9B). Analytes must be transported through the gaseous barrier to reach the coating, thus resulting in exclusion of non-volatile components of the matrix. This approach is that rates of extraction are low for poorly volatile or polar analytes because of their small Henry’s law constants. In addition, sensitivity for highly volatile compounds can suffer because these analytes have a high affinity for the gas phase, where they are concentrated. The effect of the headspace on the amount of analytes extracted and, therefore, on sensitivity can be calculated by using Eq. (2.9), which indicates that reducing its gaseous volume minimises the effect.

2.3.7

Characteristics of the Extraction Phases

The properties of the extraction phase should be carefully optimised because they determine the selectivity and reliability of the method. Properties include both bulk physicochemical properties (e.g. polarity) and physical properties (e.g. thermal

Theory of Solid-Phase Microextraction

33

stability and chemical inertness). Solvents and liquid polymeric phases (e.g. PDMS),31 are very popular because they have wide linear dynamic ranges associated with linear absorption isotherms. They also facilitate ‘gentle’ sample preparation because chemisorption and catalytic properties, frequently associated with solid surfaces, are absent. No loss or modification of the analyte occurs during extraction and/or desorption. Despite these attractive properties of liquid extraction media, solid phases are frequently used because of their superior selectivity and sensitivity for some groups of compounds. For example, carbon-based sorbents are effective for extraction of volatile analytes. The development of selective extraction materials often parallels that of the corresponding selective chemical sensors.32 Similar manufacturing approaches and structures similar to those sensor surfaces have been implemented as extraction phases. For example, phases with specific properties such as molecularly imprinted polymers33 and immobilised antibodies34 have recently been developed for extraction. An interesting concept, based on differences between bulk properties of the extraction phase and the highly specific molecular recognition centres dissolved in it, facilitates high-selectivity extraction with a minimum level of non-specific adsorption.35 Chemically tunable properties of the extraction phase have been controlled during the preparation procedure. For example, polypyrrole has been used successfully for a range of applications ranging from ion-exchange extraction to hydrophobic extraction based on selective interaction between the polymer and the target analytes.36 In addition, tunable properties of the polymer (e.g. the oxidation/reduction equilibrium in conductive polypyrrole) can be explored to control adsorption and desorption.37 Demands on the specificity of extraction phases are typically less stringent than for sensor surfaces because a powerful separation and quantification technique [e.g. gas chromatography/mass spectrometer (GCMS) or liquid chromatography/mass spectrometer (LCMS)] is typically used after extraction, facilitating accurate identification of the analyte. More demand, however, is placed on the thermal stability and chemical inertness of the extraction phase because the extraction materials are frequently exposed to high temperatures and different solvents during extraction and introduction to the analytical separation instruments. New coating chemistries (e.g. the solgel polymerisation approach) have recently been developed to address these needs.38 In order to optimise sensitivity, the choice of the extraction phase is frequently based on its affinity towards the target analyte. In practice, however, kinetic factors defined by dissociation constants, diffusion coefficients and agitation conditions frequently determine the amounts of analytes extracted from complex samples. Because overall extraction rates are slow, the amount of analytes extracted during experiments of limited duration do not reach equilibrium values.

2.4

Kinetics

Kinetic theory identifies extraction rate ‘bottlenecks’ of SPME and therefore indicates strategies to increase the speed of extraction. In the discussion below, zero

34

Handbook of Solid Phase Microextraction

interaction between analytes and vial surfaces or fibre core is assumed. Factors such as thermal expansion, swelling and analyte/analyte interactions are assumed to be negligible as well.

2.4.1

Direct Extraction

Let us first examine the direct extraction of the analytes from a homogeneous water sample into a fibre’s liquid polymer phase coating. No headspace is present in the system. Figure 2.10 shows the geometry of the system investigated, where b 2 a is the coating thickness.

2.4.1.1 Perfect Agitation Let us first consider the case where the water sample is perfectly agitated. In other words, the aqueous phase moves very rapidly with respect to the fibre so that all the analytes present in the sample have access to the fibre coating. The extraction process in this situation is shown in Figure 2.11. The curves in Figure 2.11 are concentration profiles obtained at different immersion times of the fibre into the aqueous solution. Both the concentration and z r

a Silica rod

θ

d Ds L

b Df

Liquid polymer Cf Kfs

Aqueous solution Cs Vial

Figure 2.10 Graphic representation of the SPME/sample system configuration, with dimensions and parameters labelled as follows: a, fibre coating inner radius; b, fibre coating outer radius; L, fibre coating length; d, vial inner radius; Cf, analyte concentration in the fibre coating; Df, analyte diffusion coefficient in the fibre coating; Cs, analyte concentration in the sample; Ds, analyte diffusion coefficient in the sample; Kfs, analyte distribution coefficient between fibre coating and sample; Kfs 5 Cf/Cs.

Theory of Solid-Phase Microextraction

35

f

1.0

e Coating/solution interface

0.8 Df t /(b – a)2 =

C/C∞

0.6

a: 0 b: 0.01 c: 0.05 d: 0.1 e: 0.5 f: ∞

0.4

0.2

d c b a

0 0

0.2

0.4 0.6 (r – a)/(b – a)

0.8

1.0

Figure 2.11 Absorption concentration versus radius profiles for different times after the fibre is exposed to a perfectly stirred sample, as calculated by the complete analytical solution given in Ref. 39. These profiles are valid for any analyte concentration in the sample and for any Kfs value. The curves have the following values for Dft/(b  a)2: a, 0; b, 0.01; c, 0.05; d, 0.1; e, 0.5; f, infinity.

position axes use appropriate dimensionless parameters to allow generalisation of this graph. The concentration of analyte in the aqueous phase is uniform and constant and is equal to the initial conditions. This constant concentration is ensured by the assumed infinite volume and perfect agitation concentration. To meet this condition in practice, the volume of the sample must be sufficiently large to ensure that the extracted amount does not change, within the limits of experimental error, with volume increase. Before the fibre is placed in the solution, no analyte is present in the coating (Figure 2.11, curve a). Immediately after immersion into the sample, only a thin layer close to the surface contains analyte (Figure 2.11, curve b). With time, analyte molecules diffuse progressively deeper into the coating (Figure 2.11, curves ce) and eventually reach equilibrium (Figure 2.11, curve f). Figure 2.11 indicates that under perfect agitation conditions, the speed of the absorption process is determined only by the diffusion of the analyte in the polymer coating. The area under a concentration profile curve from Figure 2.11 corresponds to the amount of analyte in the coating, expressed as a fraction of the mass extracted at equilibrium. This relationship can be represented as in Figure 2.12, and it is referred to as the extraction time profile. This is a universal graph because both the mass and time axes have dimensionless scales. The graph shows that immediately after the immersion of the fibre in solution, there is a rapid increase in the mass absorbed by the fibre. The rate of increase then slows and eventually reaches equilibrium.

36

Handbook of Solid Phase Microextraction

1.0

0.8

t95% =

2(b – a)2 Df

n/n∞

0.6

0.4

0.2

0 0

0.2

0.4 0.6 Df t /(b – a)

0.8

1.0

Figure 2.12 Mass absorbed versus time from a perfectly agitated solution of infinite volume. This profile is valid for any analyte concentration in the sample and for any Kfs value.

Figure 2.12 shows that the time required to reach equilibrium is infinite. In practice, a change in the mass extracted cannot be determined if it is smaller than the experimental error, which is typically about 5%. Therefore, the equilibration time is assumed to be achieved when 95% of the equilibrium amount of an analyte is extracted from the sample: te 5 t95% 5

ðb 2 aÞ2 2Df

ð2:38Þ

Using this equation, one can estimate the shortest equilibration time possible for the practical system by substituting appropriate data for the diffusion coefficient of an analyte in the coating (Df) and the fibre coating thickness (b 2 a). For example, the equilibration time for the extraction of benzene from a perfectly stirred solution with a 100-μm PDMS fibre is expected to be about 20 s. Equilibration times close to those predicted for perfectly agitated samples have been obtained experimentally for extraction of analytes from air samples (because of high diffusion coefficients in gas) or when very high sonication power was used to facilitate mass transfer in aqueous samples. Direct probe sonication, however, results in increased extraction temperature and poor precision. The energetic agitation approach is practical only for monitoring of flowing streams, which are self-cooled.1 Other approaches to agitation are available, which eliminate the need for additional stirring devices (such as vortex mixing) and approaches in which the extraction device itself performs agitation; e.g. updown movement,30 vibration and rotation40 of the fibre (SPME) or rotation of the magnetic stir bars coated with the extraction phase in stir-bar sorptive extraction (SBSE).41

Theory of Solid-Phase Microextraction

37

2.4.1.2 Practical Agitation Independent of the agitation level, fluid contacting a fibre’s surface is always stationary, and as the distance from the fibre surface increases, the fluid movement gradually increases until it corresponds to the bulk flow in the sample. To model mass transport, the gradation in fluid motion and convection of molecules in the space surrounding the fibre surface can be simplified as a zone of defined thickness in which no convection occurs and perfect agitation in the bulk of the fluid everywhere else. This static layer zone is called a Prandtl boundary layer (Figure 2.13).39 The thickness of the boundary layer is determined by the agitation conditions and the viscosity of the fluid. A precise understanding of the definition and thickness of the boundary layer is therefore useful. The thickness of the boundary layer (δ) is determined by both the rate of convection (agitation) in the sample and the diffusion coefficient of the analyte. Thus, the boundary layer thickness will be different for different analytes in the same extraction process. Strictly speaking, the boundary layer is a region where analyte flux is progressively more dependent on analyte diffusion and less on convection as the extraction phase is approached. For convenience, analyte flux in the bulk of the sample (outside the boundary layer) is assumed to be controlled by convection, whereas analyte flux within the boundary layer is assumed to be controlled by diffusion. Thus, δ is defined as the position where this transition occurs, or the point at which convection towards the extraction phase is equal to diffusion away from the extraction phase. At this point, analyte flux from δ towards the extraction phase (diffusion-controlled) is equal to analyte flux from the bulk of the sample towards δ, controlled by convection. An illustration of this model is presented in Figure 2.14 for the extraction of a hypothetical analyte from water, by a 10-μm coating with the following characteristics: Kfs 5 5, Ds 5 1.08 3 1025 cm2/s and Df 5 2.8 3 1026 cm2/s. The concentration profiles in Figure 2.14 were obtained at different immersion times of the fibre into the aqueous solution for a thin boundary layer (10 μm) and a thick boundary layer (100 μm), corresponding to well and poorly agitated samples. The concentration of analyte in the aqueous phase outside the boundary layer is

Extraction phase

Concentration

Boundary layer

Sample

Cs

δ

0 Position

Figure 2.13 Boundary layer model configuration. Regions and concentration versus radius profile, if the boundary layer determines the extraction rates.

38

Handbook of Solid Phase Microextraction

1

Equilibrium 25 s 20 s

0.8

C/C∞

0.6

10-µm static layer

15 s

Kfs = 5 a = 0.005 cm b = 0.015 cm

10 s

L = 1 cm Ds = 1.08 × 10–5 cm2/s Df = 2.80 × 10–6 cm2/s

5s 0.4

Bulk solution 0.2

0s 0 0.005

0.01

0.015

0.02

0.025

0.03

r (cm)

Figure 2.14 Concentration versus radius profiles in the coating and in the aqueous phase at different times after the fibre is exposed, in an SPME fibre/boundary layer system. The boundary layer is 10 μm thick. Parameters: a 5 0.005 cm, b 5 0.015 cm, L 5 1 cm, Ds 5 1.08 3 1025 cm2/s, Df 5 2.8 3 1026 cm2/s and Kfs 5 5.

uniform and constant and is equal to the initial concentration. This constant concentration is ensured by the assumed infinite volume and perfect agitation conditions in the bulk fluid. In both cases, before the fibre is placed in the solution (t 5 0), no analyte is present in the coating (Figure 2.14). Similar to the perfect agitation case (Figure 2.11), immediately after immersion into the sample (5 s), most of the extracted analyte is present in a thin layer in the coating close to the surface. However, now the concentration in the aqueous phase close to the fibre surface substantially decreases because a concentration profile is also produced in the boundary layer. This results in lower concentration gradients in the coating at the interface and slower mass transport in the system. With time, analyte molecules diffuse progressively deeper into the coating, and eventually the system reaches equilibrium. The area under a concentration profile curve in the coating, as shown in Figure 2.14, corresponds to the amount of analyte in the coating expressed as a fraction of the mass extracted at equilibrium. This relationship is shown in Figure 2.15 for the thin and thick boundary layers. For comparison, the equilibration time profile for the perfect agitation case is also included. In all cases, immediately after immersion of the fibre in the solution there is a rapid increase in mass absorbed by the fibre, which then evens out as the system

Theory of Solid-Phase Microextraction

39

Figure 2.15 Extraction versus time profiles: (a) perfect agitation conditions; (b) well agitated and (c) poorly agitated. Experimental parameters are the same as in Figure 2.14.

1 a

b

0.8

c

n/n∞

0.6

0.4

0.2

0

20

40

60

80

100

Time (s)

Figure 2.16 Dimensionless extraction versus time profile, corresponding to mass absorbed from an agitated solution of infinite volume, when the boundary layer controls the extraction rate.

100

n/n∞ (%)

80 60

t95% = 3

40

δ Kfs(b – a)

Ds

20 0 0

1

2 3 Dst / δKfs(b – a)

4

5

reaches equilibrium. The effect of the boundary layer size on the equilibration rate is very visible. The thin static film does not affect the extraction rate significantly. The equilibration time for a 10-μm-thick boundary layer is about 25 s (Figure 2.15, curve b), compared to 20 s for a perfectly agitated sample (Figure 2.15, curve a). In Figure 2.15, curve c the boundary layer is 100 μm, which is sufficiently thick for diffusion through this zone to determine the extraction rate. The equilibration time for a 100-μm-thick boundary layer was calculated to be 95 s. Figure 2.16 illustrates the relationship for the case when the extraction rate is determined by the presence of a boundary layer. This is a universal graph because both the mass and time axes have dimensionless scales. Note that an analyte with a

40

Handbook of Solid Phase Microextraction

high Kfs value will have a long equilibration time, even with a very thin boundary layer, which is characteristic of rapid agitation. Similar to the case of Eq. (2.38), the time required to reach equilibrium can be estimated from Figure 2.16: te 5 t95% 5 3

δKfs ðb 2 aÞ Ds

ð2:39Þ

where (b 2 a) is the fibre coating’s thickness, Ds is the analyte’s diffusion coefficient in the sample fluid and Kfs is the analyte’s distribution constant between fibre and sample. This equation can be used to predict equilibration times when the extraction rate is controlled by the diffusion in the boundary layer. In other words, the extraction time calculated by using Eq. (2.39) must be longer than the corresponding time predicted by Eq. (2.38), which leads to the following condition: δ.

ðb 2 aÞ Ds 6Kfs Df

ð2:40Þ

Diffusion coefficients for the most common coating, PDMS, are smaller than the corresponding coefficients in water by a factor of 56. Therefore, the above condition for water extraction and PDMS coating can be simplified to the ratio of the coating thickness and the appropriate distribution constant: δ.

b2a Kfs

ð2:41Þ

When the calculated value of b  a/Kfs is small compared to the thickness of the boundary layer, Eq. (2.39) should be used to estimate the extraction time. It should be emphasised that both equations substantially underestimate equilibration times for the situation when their values are close together because the diffusion through both phases has a cumulative effect on the equilibration times. The effective thickness of a boundary layer can be estimated from empirical formulae of fluid mechanics. If the fibre is placed in the centre of a vial stirred by a magnetic bar, liquid flow will be axis-symmetrical around the circumference of the fibre. The equation for a flat-plate boundary layer will apply, provided the layer thickness, δ, is small compared to the coated fibre radius, b.19 From the formula for heat transfer,42 the effective boundary layer thickness is δ 5 2:64

b pffiffiffiffiffiffi Pr 0:43 Re

ð2:42Þ

where b is the radius of the fibre; Re is the Reynolds number, Re 5 2ub/ν, where u and v are the fluid’s linear speed and kinematic viscosity (for water at 25 C, kinematic viscosity is 0.009 cm2/s); b is the coated fibre radius and Pr is the Prandtl number of the liquid, equal to v/D, where D is the diffusion coefficient of the analyte in the liquid.43 This formula applies for laminar flows: Re , 10.

Theory of Solid-Phase Microextraction

41

Extraction time can be shortened by optimising the position of the fibre in the sample vial. If the fibre is placed off-centre in the vial, the fluid flows past the fibre and, perpendicular to the fibre axis, the boundary layer thickness can be estimated by δ 5 9:52

b Re0:62 Pr 0:38

ð2:43Þ

The tangential velocity in water agitated by a stir bar in a cylindrical container is predicted by    r 2 for 0 , r # 0:74R ð2:44Þ uðrÞ 5 1:05πNr 2 2 0:74R uðrÞ 5 0:575πNR2

1 r

for r . 0:74R

ð2:45Þ

where R is the radius of the stir bar, and N is the revolutions per second.44 The point of maximum velocity is r 5 0:56R, where the velocity is u(r) 5 2.64NR. For example, when extracting benzene from water (Kfs 5 125) at 25 C using a 56-μm PDMS-coated fibre (b 5 125 μm) placed in the centre of the vial and magnetic stirring at 1,000 rpm, Eq. (2.39) predicts a boundary layer thickness of about 10 μm. This thickness satisfies the preceding condition (Eq. (2.38): 10 μm . 56 μm/125), which means that the extraction rate is controlled by the diffusion in the boundary layer. Therefore, we can use Eq. (2.36) to calculate equilibration time, which is found to be 180 s. If, however, the fibre is placed at the point of maximum velocity, r 5 0.56R, for a stir bar that is 2 cm long (R 5 1 cm), the same equations predict an equilibration time of 90 s. The experimental extraction times are close to these values. This calculation indicates that the position of the fibre in the vial should be kept constant, preferably close to the optimum position, about half the distance between the centre of the vial and the end of the stir bar. In practice, overestimation of the equilibration times is advisable, in order to eliminate variable extraction as a function of minute change in the fibre position. The extraction time profile from Figure 2.15 indicates that a 10-μm boundary layer (Figure 2.15, curve b) does not affect the equilibration time significantly, compared to the perfect agitation case, while, in the discussion above, the diffusion through the 10-μm layer controls the kinetics of extraction. The difference is associated with the distribution constant, which in the case of Figure 2.15 is only 5, while in the example above, it is 125. This comparison shows that it is not only the thickness of the boundary layer that is important but also the amount of analyte that needs to be transported through it to reach equilibrium. Therefore, for extraction from a small volume sample resulting in a drop of analyte concentration in the bulk of the sample, the equilibration time will be shorter because the amount extracted will be smaller. It should be emphasised that Eqs (2.39) and (2.40) can be used to calculate the boundary layer thicknesses for other agitation methods, as long as the relative velocity of sample phase versus coating can be estimated. For example, it is

42

Handbook of Solid Phase Microextraction

possible to estimate the boundary layer thickness for fibre movement agitation methods when the average velocity of the fibre in the solution is known.

2.4.2

Desorption of Extracted Analytes

After the extraction is complete, the coated fibre containing analytes is transferred to the injection port of a GC or high-performance liquid chromatography (HPLC) instrument. During the desorption process, the analyte diffuses from the coating into the stream of carrier fluid. Therefore, this process is the reverse of absorption from a well-agitated aqueous phase when the concentration of an analyte is zero at the coating/fluid interface. To ensure that this condition is fulfilled, a high linear flow rate must be generated. The high flow rate is required to ensure that the desorbed analyte is removed immediately from the vicinity of the coating, so as not to interact with the coating and slow the desorption process. Figure 2.17 shows a family of curves which describe changes in the concentration profile in the coating during the desorption process. At the start of desorption, analyte is removed from the layer of coating close to the interface (Figure 2.17, curve a) and then from the deeper parts of the coating (Figure 2.17, curves be). The desorption time profile is presented in Figure 2.18. As in the perfectly agitated case (Figure 2.11), the desorption process is completed at the time corresponding to ðb 2 aÞ2 =ð2Df Þ. The values for the distribution constant and diffusion coefficient will be different compared to extraction from a perfectly stirred sample because of the different temperatures. The above relationship indicates that the desorption time is independent of the distribution constant 1.0 a

b a: 0 b: 0.01 c: 0.05 d: 0.1 e: 0.5

C/C0

0.6

0.4

c d

Coating/gas interface

Df t /(b – a)2 =

0.8

0.2 e

0 0

0.2

0.4 0.6 (r – a)/(b – a)

0.8

1.0

Figure 2.17 Desorption concentration versus radius profile in a fibre coating, corresponding to different times after a fibre is exposed in an instrument injection port. The concentration at the coating/carriergas interface is assumed to be 0 at all times. The curves have the following values for Dft/(b  a)2: a, 0; b, 0.01; c, 0.05; d, 0.1 and e, 0.5.

Theory of Solid-Phase Microextraction

43

1.0

0.8

n/n0

0.6

0.4

0.2

0 0

0.2

0.4

0.6

0.8

1.0

Df t /(b – a)2

Figure 2.18 Mass in fibre versus time, showing the mass desorbed from a fibre by a fastmoving stripping phase after a fibre is exposed in an instrument injection port.

Kfg and independent of the initial concentration in the fibre. It can be calculated that the desorption time is about 1 s for a 100-μm coating, when 200 C temperature is used in the injector, for low-molecular-mass analytes. In practice, the flow of the mobile phase has finite value and, therefore, the desorption time might be longer for large molecule analytes, characterised by large distribution constants and high molecular weights.

2.4.3

Headspace Extraction

Equations (2.8) and (2.9) indicate that use of the headspace above the sample as an intermediate phase might be an interesting means of accelerating extraction for analytes characterised by high Henry’s law constants. When a thin extraction phase is used, the initial rate of extraction, and hence the extraction time, is controlled by diffusion of analytes present in the sample matrix through the boundary layer. Addition of a gaseous headspace facilitates enhanced transport into the extraction phase because of the high diffusion coefficients of the analytes into the gas phase. In order to increase transport from the sample matrix into the headspace, the system can be designed to produce a well-agitated, large sample/headspace interface. This can be accomplished by using large-diameter vials with good agitation, by purging, or even by the use of spray systems. At room temperature, only volatile analytes are transported through the headspace. For low-volatility compounds, heating of the sample is a good approach, if loss in magnitude of the distribution constant can be accepted. The ultimate approach is to heat the sample and cool the extraction phase at the same time. Heating the sample not only increases the Henry’s law constant but also induces convection in the headspace because density

44

Handbook of Solid Phase Microextraction

gradients associated with temperature gradients present in the system result in higher mass transport rates. The cooling of the sorbent increases its adsorption capacity. Collection of the analytes can be performed in the same vial, as discussed previously,45 or can be separated in space, similarly to the purge and trap technique. In the heatingcooling experiments, both kinetic and thermodynamic factors are addressed simultaneously. Headspace approaches are also interesting because, as discussed above, adverse effects associated with the presence of solids, oily or high-molecular-weight interferences, which can cause fouling of the extraction phase, are eliminated (see Section 2.3.6). In the headspace mode, the analytes need to be transported through the barrier of air before they can reach the coating. This modification serves primarily to protect the fibre coating from damage by high molecular mass and other non-volatile interferences present in the sample matrix, such as humic materials or proteins. The headspace mode also allows for modification of the matrix, such as a change of the pH, without damaging the fibre. Amounts of analyte extracted into the coating from the same vial at equilibrium using direct and headspace sampling are identical, as long as the volumes of the sample and gaseous headspace are the same. This is because the equilibrium concentration is independent of the fibre location in the sample/headspace system. If the above conditon is not satisfied, a significant sensitivity difference between the direct and headspace approaches exists only for very volatile analytes. The choice of sampling mode has a very significant impact on extraction kinetics, however. When the fibre coating is in the headspace, the analytes are removed from the headspace first, followed by indirect extraction from the matrix, as shown in Figure 2.19. Overall mass transfer to the fibre is typically limited by mass transfer rates from the sample to the headspace. Therefore, volatile analytes are extracted faster than semi-volatiles because they are at a higher concentration in the headspace, which contributes to faster mass transport rates through the headspace. Temperature has a significant effect on the kinetics of the process by determining the vapour pressure of analytes. In fact, the equilibration times for volatiles are shorter in the headspace SPME mode than for direct extraction, under similar agitation conditions. This outcome is the result of two factors: a substantial portion of the analyte is in the headspace prior to extraction, and diffusion coefficients in the gaseous phase are (A) High capacity KhsVh>> Kfh

(B) Low capacity KhsVh>> Kfh

Figure 2.19 Analyte distribution in headspace extraction, as a function of mass transfer rates.

Theory of Solid-Phase Microextraction

45

typically four orders of magnitude larger than in liquid media. However, because concentrations of semi-volatiles in the gaseous phase at room temperature are typically small, overall mass transfer rates are substantially lower and result in longer extraction times. The extraction times can be improved by using very efficient agitation or by increasing the extraction temperature.46 Figure 2.20A and B illustrates the effect of agitation on the extraction time profile obtained for polynuclear aromatic hydrocarbons (PAHs). As the rotational speed of the magnetic stirrer increases, the equilibration time of naphthalene and acenaphthene decreases from 8 min to 3 min and from 25 min to 10 min, respectively. For less volatile analytes such as phenanthrene and chrysene, the equilibration is not reached during the experimental period in either case, but the amount of analyte extracted after 70 min extraction at low agitation (Figure 2.20A) is about the same as after 45 min of more efficient stirring (Figure 2.20B). The use of even more efficient agitation techniques, such as direct sonication, further decreases extraction time. Elevated sampling temperature lowers the fibre/sample partition coefficients. This decreases the amount extracted at equilibrium, but it may be acceptable if target limits of detection can still be reached. This effect is demonstrated dramatically with the analysis of amphetamines.47 As shown in Figure 2.21, room temperature extraction produces a very long equilibrium extraction time but ultimately the highest amount extracted. Conversely, the highest temperature tested (73 C) produces a very fast equilibrium time (B5 min), but a significantly lower equilibrium amount is extracted. To enable the simultaneous analysis of very volatile substances (gases) and less volatile analytes, the headspace SPME technique can be combined with static headspace sampling by using a gastight SPME device.28 Upon injection, the volume of headspace is injected at the same time as analytes are desorbed.

20

20

Mass absorbed (ng)

(A)

(B) b

b

15

15

c

c

10

10 a

a

5

5 d

d 0 0

10

20

30

40

50

60

Extraction time (min)

70

80

0

0

10

20

30

40

Extraction time (min)

Figure 2.20 Effect of (A) low agitation speed and (B) high agitation speed on the extraction time profile for PAHs: a, naphthalene; b, acenaphthene; c, phenanthrene and d, chrysene.

50

46

Handbook of Solid Phase Microextraction

Methamphetamine extracted (ng)

500 450 400 350 300 250 200 150 100 50 0 0

20

40

60

80

100

Time (min)

Figure 2.21 Effect of temperature increase on the extraction time profile for methamphetamine. Key: (V) 22 C, (¢) 40 C, (’) 60 C and (K) 73 C.

2.4.4

Indirect SPME Extraction Through a Membrane

The main purpose of the membrane barrier is to protect the fibre against damage, similar to the use of headspace SPME, when very dirty samples are analysed. Membrane protection is advantageous for determining analytes with volatilities that are too low for the headspace approach. In addition, a membrane made from appropriate material can add a certain degree of selectivity to the extraction process. The kinetics of membrane extraction are substantially slower than for direct extraction, though, because the analytes must diffuse through the membrane before they can reach the coating. The thicker membranes can be used to slow down the mass transfer through the membrane, resulting in the time-weighted-average (TWA) measurement discussed later.

2.5

Extraction with Derivatisation

The selectivity and capacity of the extraction phase for analytes such as polar or ionic species, which are difficult to extract, can be frequently enhanced by introducing a derivatisation step.48 The objective of derivatisation is not only to convert the native analytes into less-polar derivatives that are extracted more efficiently but also to label them for better detection and/or chromatography. The most interesting implementation of this approach is simultaneous extraction/derivatisation. In this technique, the derivatisation reagent is present in the extraction phase during the extraction. The main advantage of this approach is that two steps are combined. Two limiting cases describe the combination of extraction and derivatisation.

Theory of Solid-Phase Microextraction

47

The first occurs when mass transfer to the fibre is slow compared with the reaction rate. Under these conditions, as discussed above, Eq. (2.38) describes the rate of accumulation of the analytes, assuming that the derivative is trapped in the extraction phase. In the second limiting case, the situation is reversed in that the reaction rate is slow compared with the transport of analytes to the extraction phase. In other words, at any time during the extraction procedure, the extraction phase is at equilibrium with the analyte in a well-agitated sample, resulting in a uniform reaction rate throughout the coating. This is typical for thinly dispersed extraction phases because the equilibration time for well-agitated conditions is very short compared with a typical reaction rate constant. The accumulation rate of the product in the extraction phase n/t can then be defined by ð n 5 Ve kr Kes Cs ðtÞdt

ð2:46Þ

where Cs is the initial concentration of analyte in the sample and kr is the chemical reaction rate constant. In short, when the sample is of large volume (e.g. direct sampling in the field), the reaction and accumulation of analyte in the extraction phase proceed with the same rate as long as reagent is present in excess. It is worth noting that the rate is also proportional to the extraction phase/sample matrix distribution constant. If the analyte concentration varies during accumulation, the amount collected corresponds to the integral over concentration and time, as will be discussed later for TWA sampling. For limited sample volume, however, the concentration of analyte in the sample phase decreases with time because it becomes partitioned into the coating and is converted to trapped product, resulting in a gradual decrease of the rate. The time required to extract analytes exhaustively from a limited volume can be estimated from the experimental conditions.

2.6

Extraction of Sample Matrices Containing Solids

The most challenging extractions occur when a solid is present as a part of the sample matrix. This situation will be considered as the most general example of extraction because it involves several fundamental processes occurring during the extraction procedure. If we assume that a matrix particle consists of an organic layer on an impermeable but porous core, and the analyte is adsorbed on the pore surface, the extraction process can be modelled by considering several basic steps, as shown in Figure 2.22. To remove the analyte from the matrix, the compound must first be desorbed from the surface (A(M,S)); it must then diffuse through the organic part of the matrix (A(M,L)) to reach the matrix/fluid interface (A(M,I)). At this point, the analyte must be solvated by the extraction phase (A(EP,P)), and it must then diffuse through the static phase present inside the pore to reach the portion of the extraction phase that is affected by convection. The analyte is then transported through

48

Handbook of Solid Phase Microextraction

Flow convection

kd

Ds

Kes

De A(EP,B)

A(EP,P) A(M,S) A(M,L) A(M,I)

Particle core Organic material

Figure 2.22 Kinetic of extraction from a solid matrix (organic layer on a porous core).

the interstitial pores of the matrix, eventually reaching the bulk of the extraction phase (A(EP,B)). The simplest way to design a kinetic model for this problem is to adopt equations developed by engineers49,50 to investigate mass transport through porous media.51 The leaching approach can be performed directly in a vessel (e.g. Soxhlet, sonication or microwave extraction) or can be combined with elution from the packed tube [via supercritical fluid extraction (SFE) or pressurised fluid extraction (PFE)]. For the purpose of further discussion, we will consider the efficient and frequently applied experimental arrangement for removing solid-bound semi-volatile analytes, which involves the use of an SPME fibre, such as cold-fibre SPME.

2.6.1

Convolution Model of Extraction

The discussion above applies only when the analytes are initially present in a fluid phase. If dynamic extraction is performed from the beginning of extraction, in most practical circumstances the system is not expected to achieve the initial equilibrium conditions. This is due to the slow mass transport between the matrix and the fluid (e.g. slow desorption kinetics or slow diffusion in the matrix). The expected relationship between the amount of analyte removed from the matrix and elution time can be obtained in this instance by convoluting the function describing the rate of mass transfer between the phases, F(t), with the extraction time profile, m/mo(t)52: ðτ 5 t τ50

mðt 2 τÞ FðτÞdτ mo

ð2:47Þ

Theory of Solid-Phase Microextraction

49

The resulting function describes a process in which elution and the mass transfer between the phases occur simultaneously. In this discussion, we will refer to this function as the ‘extraction time profile’ to emphasise that, for most extractions, these two processes are expected to be combined. F(t) describes the kinetics of the process, which defines the rate of release of analyte from the sample matrix and can include, for example, the matrixanalyte complex dissociation rate constant (assuming a linear adsorption isotherm), the diffusion coefficient, the time constant that describes swelling of the matrix that will facilitate removal of analyte or a combination of the above. A detailed discussion, graphical representations and applications of this model to describe and/or investigate processes in SFE have been described elsewhere.53 The conclusion reached above can be stated more generally: convolution among functions describing individual processes occurring during extraction describes the overall extraction process and is a unified way of describing the kinetics of these complex processes. The exact mathematical solution of the convolution integral is frequently difficult to obtain, but the solution can be represented graphically by use of the Fourier Transform or by numerical approaches. It is frequently possible to incorporate mathematical functions that describe a combination of the unit processes. It should be emphasised that the convolution approach considers all processes equivalently. In practice, however, a small number (in fact, frequently just one unit process) controls the overall rate of extraction, enabling simplification of the equation. Determination of the limiting step is not possible exclusively by qualitative agreement with the mathematical model because the effect on recovery of most of the unit processes has an exponential yield curve. For proper recognition of all unit processes, quantitative agreement and/or the effect of extraction conditions must be examined. Identification of the limiting process provides valuable insight into the most effective approach to the optimisation of the extraction. Fundamental understanding of the extraction process leads to better strategies for optimisation of performance. In heterogeneous samples, for example, release of solid-bound analytes from the sample matrix, by reversal of chemisorption or inclusion, frequently controls the rate of extraction. Recognition of this fact enables extraction conditions to be changed to increase the rates of extraction. For example, dissociation of the chemisorbed analytes can be accomplished either by the use of high temperatures or by the application of additives, facilitating desorption. This led to the development of high-temperature SFE,54 then to evolution of both the PFE approach55 and microwave extraction, with more selective energy focusing at the sample matrix/extraction phase interface.56 There is also an indication that milder conditions can be applied by taking advantage of the catalytic properties of the extraction phase or additives.23 To realise this opportunity, however, more research must be performed to gain insight into the nature of interactions between analytes and matrices. Deconvolution of experimental data can be used to investigate the matrix effect associated with the slow release of analytes from the matrix. Benefits include not only improved speed but also selectivity resulting from the application of appropriate conditions. This strategy of simultaneous extraction and

50

Handbook of Solid Phase Microextraction

cleanup has been successfully applied to the very difficult extraction of polychlorinated dibenzo-p-dioxins from fly ash.57 If the rate of extraction is controlled by mass transport of analytes in the pores of the matrix, the process can be successfully enhanced by application of sonic and microwave energy, which induce convection even in the small dimensions of the pore. Diffusion through all or some of a sample matrix containing natural or synthetic polymeric material frequently controls the rate of extraction.58 In such circumstances, swelling of the matrix and increasing the temperature result in increased diffusion coefficients and, therefore, increased extraction rates.

2.7

Solid Versus Liquid Sorbents

There is a substantial difference between the performance of liquid and solid coatings (Figure 2.23). With liquid coatings, the analytes partition into the extraction phase, in which the molecules are solvated by the coating molecules. The diffusion coefficient in the liquid coating enables the molecules to penetrate the whole volume of the coating within a reasonable extraction time, if the coating is thin (Figure 2.23A). With solid sorbents (Figure 2.23B), the coating has a glassy or a well-defined crystalline structure, which, if dense, substantially reduces diffusion coefficients within the structure. Within the time span of the experiment, therefore, sorption occurs only on the porous surface of the coating (Figure 2.23B). During extraction by using a solid phase and high analyte/interference concentration, after long extraction times, compounds with poor affinity towards the phase are frequently displaced by analytes characterised by stronger binding, or those present in the Figure 2.23 Sorption mechanism for (A) liquid and (B) solid sorbents.

t=0

Adsorption

Absorption

t = te

(A)

(B)

Theory of Solid-Phase Microextraction

51

sample at high concentrations. This is because only a limited surface area is available for adsorption. If this area is substantially occupied, competition occurs, and the equilibrium amount extracted can vary with the concentrations of both the target and other analytes.59 In extraction, with liquid phases, however, partitioning between the sample matrix and extraction phase occurs. Under these conditions, equilibrium extraction amounts vary only if the bulk coating properties are modified by the extracted components. This occurs only when the amount extracted is a substantial portion (a few percent) of the extraction phase, resulting in a possible source of nonlinearity. This is rarely observed because extraction/enrichment techniques are typically used for analysis of trace contaminants. One way to overcome the fundamental limitation of porous coatings in a microextraction application is to use an extraction time much less than the equilibration time, so that the total amount of analytes accumulated by the porous coating is substantially below the saturation value. For such pre-equilibrium SPME methods, diffusion-based and kinetic calibration methods can be used in order to determine the initial analyte concentration in the sample. Chapter 6 discusses the theoretical basis and proper implementation of various SPME calibration methods.

2.8

Passive TWA Sampling

Consideration of different arrangements of the extraction phase, including the protective barriers discussed earlier, is beneficial. For example, extension of the boundary layer by a protective shield that restricts convection would result in a TWA measurement of analyte concentration (see Eq. (2.39)). A variety of diffusive samplers have been developed based on this principle. One system consists of an externally coated fibre with the extraction phase withdrawn into the needle (Figure 2.24). When the extracting phase in an SPME device is not exposed directly to the sample but is contained within protective tubing (a needle), without any flow of sample through it, diffusive transfer of analytes occurs via the static sample (a gas phase or other matrix) trapped in the needle. This geometric arrangement is a very simple method, capable of generating a response proportional to the integral of the analyte concentration over time and space (when the needle is moved through space).60 Under these conditions, the only mechanism of analyte transport to the Figure 2.24 Principle of fibre TWA sampler. Z

c(t) 0

Z

z

52

Handbook of Solid Phase Microextraction

extracting phase is diffusion through the matrix contained in the needle. During this process, a linear concentration profile (shown in Figure 2.24) is established in the tubing between the small needle opening, characterised by a surface area A and the distance, Z, between the needle opening and the position of the extracting phase. The amount of analyte extracted, dn, during the time interval, dt, can be calculated by considering Fick’s first law of diffusion6: dn 5 ADm

dc ΔCðtÞ dt 5 ADm dt dz Z

ð2:48Þ

where ΔC(t)/Z is an expression of the gradient established in the needle between the needle opening and the position of the extracting phase, Z; ΔC(t) 5 C(t)  CZ, where C(t) is the time-dependent concentration of analyte in the sample in the vicinity of the needle opening and CZ is the concentration of the analyte in the vicinity of the coating. CZ is close to zero for a high extraction phase/matrix distribution constant; therefore, ΔC(t) 5 C(t). The concentration of analyte, CZ, at the coating position in the needle will increase with the integration time, but it will remain low compared with the sample concentration because of the presence of the extraction phase. The amount of analyte accumulated over time can therefore be calculated as n 5 Dm

ð A Cs ðtÞdx Z

ð2:49Þ

As expected, the amount of analyte extracted is proportional to the integral of sample concentration over time, the diffusion coefficient of the analyte in the matrix filling the needle, Dm, and the area of the needle opening, A, and inversely proportional to the distance, Z, of the coating from the needle opening. It should be emphasised that Eqs (2.48) and (2.49) are valid only when the amount of analyte extracted onto the sorbent is a small fraction (below the standard deviation of the measurement, typically 5%) of the equilibrium amount for the lowest concentration in the sample. In order to extend integration times, the coating can be placed further into the needle (larger Z), the opening of the needle can be reduced by placing an additional orifice over the needle (smaller A), or a higher capacity sorbent can be used. The first two solutions will result in low measurement sensitivity. Increasing the sorbent capacity is a more attractive proposition. It can be achieved either by increasing the volume of the coating or by changing its affinity for the analyte. Because increasing the coating volume would require an increase in the size of the device, the optimum approach to increasing the integration time is to use sorbents characterised by large coating/gas distribution constants. If the matrix filling the needle is something other than the sample matrix, an appropriate diffusion coefficient should be used in Eq. (2.49). SPME-based TWA applications are disscussed in Chapter 8. In the system described, the length of the diffusion channel can be adjusted to ensure that mass transfer in the narrow channel of the needle controls overall mass transfer to the extraction phase, irrespective of convection conditions.61 This is a

Theory of Solid-Phase Microextraction

53

very desirable feature of TWA sampling because the performance of this device is independent of the flow conditions in the system investigated. This is difficult to ensure for high surface area membrane permeation-based TWA devices, e.g. passive diffusive badges,62 semi-permeable membrane devices (SPMDs)63 or thin-film SPME membranes discussed in Chapter 3. For analytes characterised by moderate to high distribution constants, mass transport is controlled by diffusive transport in the boundary layer. The performance of these devices, therefore, depends on the convection conditions in the investigated system.64

2.9

In-Tube SPME

The discussion above is limited to systems involving extraction phases dispersed on the surface of a rod which is exposed from the protective tubing (needle). In addition to this approach, in some SPME systems discussed in Chapter 3, the extraction phase remains in the tubing during the extraction. These approaches involve either coated rods retracted into the needles (Figure 2.24) or an extracting phase dispersed onto the inner surface of the tubing. Below, we briefly discuss the theoretical aspects of the extraction process that use this in-tube geometric arrangement. Assume that the system consists of a piece of fused silica capillary, internally coated with a thin film of extracting phase (a piece of open tubular capillary GC column), or that the capillary is packed with extracting phase dispersed on an inert supporting material (a piece of micro-LC capillary column). In these geometric arrangements, the concentration profile along axis x of the tubing containing the extracting phase as a function of time, t, can be described by adopting the expression for dispersion of the concentration front65,66: Cðx; tÞ 5

  x 2ðut=1 1 kp Þ 1 pffiffiffi C0 1 2 erf 2 σ 2

where u is the linear velocity of the fluid through the tube and k is the partition ratio defined as kp 5 Kfs

Vf Vv

ð2:50Þ

where Kfs is a sample/coating distribution constant, Vf is the volume of the extracting phase and Vv is a void volume of the tubing containing the extracting phase. σ is the mean square root dispersion of the front defined as σ5

rffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi u Ht 1 1 kp

ð2:51Þ

54

Handbook of Solid Phase Microextraction

where H is equivalent to the height equivalent to theoretical plate (HETP) in chromatographic systems. This can be calculated as a sum of individual contributions to the front dispersion. These contributions depend on the particular geometry of the extracting system.67 Equation (2.51) indicates that the front of the analyte migrates through the capillary with a speed proportional to the linear velocity of the sample and inversely related to the partition ratio. For short capillaries with a small dispersion, the extraction time can be assumed to be similar to the time required for the centre of the band to reach the end of the capillary: te 5

Lð1 1 Kfs ðVf =Vv ÞÞ u

ð2:52Þ

where L is the length of the capillary holding the extraction phase. As expected, the extraction time is proportional to the length of the capillary and inversely proportional to the linear flow rate of the fluid. Extraction time also increases with an increase in the coating/sample distribution constant and with the volume of the extracting phase but decreases with an increase in the void volume of the capillary. It should be emphasised that the above equation can be used only for direct extraction when the sample matrix passes through the capillary. This approach is limited to particulate-free gas and clean water samples. The headspace SPME approach can broaden the application of the in-tube SPME. In that case, careful consideration to the mass transfer between sample and headspace should be given in order to describe the process properly (analogous to the discussion in Section 2.4.3). Also, if the flow rate is very rapid, producing turbulent behaviour, and the coating/sample distribution constant is not very high, then the perfect agitation conditions are met, and Eq. (2.39) can be used to estimate equilibration times. Removal of analytes from a tube is an elution problem analogous to frontal chromatography and has been discussed in detail in Ref. 67 In general, if the desorption temperature of a GC is high and thin coatings are used, then all the analytes are in the gas phase as soon as the coating is placed in the injector and the desorption time corresponds to the elution of two void volumes of the capillary. For liquid desorption (for example in an LC system, see Chapter 3), the desorption volume can be even smaller because the analytes can be focused at the front of the desorption solvent.

2.10

Experimental Verification

Some of the useful conclusions given in this chapter on SPME theory have been verified experimentally and are discussed below: 1. The concentration of the sample has no impact on the concentration time profile and the equilibration time. Figure 2.25 illustrates this fact based on experimental data for the extraction of benzene from water. In other words, if the extraction is optimised for a

Theory of Solid-Phase Microextraction

100 a

Cs = 10 ppm

b

Cs = 1 ppm

c

Cs = 0.1 ppm

10 Mass (ng)

55

10

1

0.

10

0

20

30

40

50

60

Figure 2.25 Effect of analyte concentration on the absorption versus time profile of 2,500-rpm stirred benzene in water extracted with a 56-μm-thick coating on a 1-cm-long fibre, for Kfs 5 125, a 5 0.007 cm, b 5 0.0126 cm, L 5 1 cm, Ds 5 1.08 3 1025 cm2/s and Df 5 2.8 3 1026 cm2/s. (a) Cs 5 10 ppm, (b) Cs 5 1 ppm and (c) Cs 5 0.1 ppm.

Time (s)

40

Figure 2.26 Equilibration time profiles for magnetic stirring conditions at various rotational speeds.

a b

30

Mass (ng)

c 20

d a: Perfect agitation b: 2,500 rpm c: 1,800 rpm d: 400 rpm e: No stirring

e

10

0 0

200

400

600

Time (s)

given concentration, the equilibration time will be the same for other concentrations as well. This condition is valid as long as the system behaves linearly, in the other words, distribution constants between the various components in the SPME/sample system remain constant with a concentration change. 2. Agitation conditions determine the extraction rate and equilibration time for extraction from aqueous samples. Figure 2.26 shows experimental data comparing equilibration time profiles obtained for magnetic stirring conditions at various rotational speeds (Figures 2.27 and 2.28). 3. Increased mass transfer conditions can be obtained by moving the fibre with respect to the solution and by vibrating the vial in a sonicator bath. 4. Extraction time is affected by the coating thickness. Figure 2.29 shows experimental results demonstrating that the coating thickness changes not only the amount of analyte extracted but also the equilibration time. It is important to use the thinnest coating that gives acceptable sensitivity.

56

Handbook of Solid Phase Microextraction

40 a

Mass (ng)

30

20

b c a: 2,500 rpm magnetic stirring

10

d

b: 100 W sonication c: Repeated fibre insertion/retraction d: No stirring

0 200

0

400

600

Time (s)

Figure 2.27 Extraction profile using various stirring techniques.

Mass absorbed (ng)

d

50

c

40

(a) Benzene in a static aqueous phase (b) Benzene in a well-agitated aqueous phase (c) o -Xylene in a static aqueous phase (d) o -Xylene in a well-agitated aqueous phase

30 20

b a

10 0

Figure 2.28 Extraction time profiles for benzene and o-xylene under static and well-agitated conditions.

0

2

4

6

8

10

Extraction time (min)

Figure 2.29 Extraction time profiles of benzene as a function of the coating thickness.

80 a

Mass (ng)

60

Sample: 0.1 ppm benzene in water

40

b a: 100 µm b: 56 µm c: 15 µm

20 c 0 0

200

400 Time (s)

600

Theory of Solid-Phase Microextraction

57

Figure 2.30 Extraction time profiles for BTEX as a function of their distribution constants.

30

a Mass (ng)

20 a: p-Xylene, Kfs = 831 b: Toluene, Kfs = 294 c: Benzene, Kfs = 125

10

b c

0 0

1,000

2,000

3,000

Time (s)

5. The distribution constant affects the equilibration time. Figure 2.30 shows experimental data for BTEX and differences in the corresponding extraction time profiles. The amount of analyte extracted increases with Kfs, but the equilibration time becomes longer as well. The compound with higher affinity for the coating reaches equilibrium later. Using a thin selective coating will not help the extraction times for aqueous and other heterogeneous matrices because in these cases, the transport to the coating controls the rate of extraction.

References 1. S Motlagh & J Pawliszyn, Anal Chim Acta 284 (1993) 265 2. T Gorecki, X Yu & J Pawliszyn, Analyst 124 (1999) 643 3. FM Musteata, Solid Phase Microextraction in Drug Discovery, Ph.D. Thesis, University of Waterloo, Canada (2005) p. 2 4. Z Zhang & J Pawliszyn, J Phys Chem 100 (1999) 17648 5. S Poole & C Poole, Analyst 120 (1995) 1733 6. R Schwarzenbach, P Gschwend & D Imboden, Eds, Environmental Organic Chemistry (1993) John Wiley & Sons: New York, NY 7. P Martos, A Saraullo & J Pawliszyn, Anal Chem 69 (1997) 402 8. A Saraullo, P Martos & J Pawliszyn, Anal Chem 69 (1997) 1992 9. J Li & PW Carr, J Chromatogr A 659 (1994) 367 10. JH Park, JE Lee, MD Jang, J Li & PW Carr, J Chromatogr 586 (1991) 1 11. AJ Dallas & PW Carr, J Phys Chem 98 (1994) 4927 12. CL Arthur, LM Killam, KD Buchholz & J Pawliszyn, Anal Chem 64 (1992) 1960 13. P Martos & J Pawliszyn, Anal Chem 69 (1997) 206 14. FA Long & WF McDevit, Chem Rev 51 (1952) 119 15. X Yang & T Peppard, LCGC 13 (1995) 882 16. JP Conzen, J Burck & HJ Ache, Appl Spectrosc 47 (1993) 753 17. J Brandrup & EH Immergut, Eds, Polymer Handbook 3rd ed (1989) John Wiley & Sons: Toronto, ON 18. Handbook of Chemistry & Physics, 70th ed (1989) CRC Press: Boca Raton, FL

58

Handbook of Solid Phase Microextraction

19. ZY Zhang & J Pawliszyn, Anal Chem 67 (1995) 34 20. The Thermodynamic Research Center, TRC Thermodynamic Tables (1992) The Texas A&M University System: College Station, TX 21. PW Atkins, Physical Chemistry, 4th ed (1978) Freeman: New York, NY 22. J Pawliszyn, J Chromatogr Sci 31 (1993) 31 23. N Alexandrou & J Pawliszyn, Anal Chem 61 (1989) 2770 24. J Pawliszyn, Solid Phase Microextraction, Theory & Practice (1997) John Wiley & Sons: New York, NY, pp. 4447 25. H Lord & J Pawliszyn, J Chromatogr 885 (2000) 153 26. T Gorecki & J Pawliszyn, Analyst 122 (1997) 1079 27. M Mulder, Basic Principles of Membrane Technology 122 (1991) Kluwer: Dordrecht, p. 1079 28. Z Zhang, J Poerschmann & J Pawliszyn, Anal Commun 33 (1996) 129 29. K Rasmussen, S Pedersen-Bjergaard, H Krogh, H Ugland & J Gronhaug, J Chromatogr 873 (2000) 3 30. K Boos & C-H Grimm, Trends Anal Chem 18 (1999) 175 31. D Louch, S Motlagh & J Pawliszyn, Anal Chem 64 (1992) 1187 32. B Eggins, Chemical Sensors & Biosensors, 2nd ed (2002) Wiley-VCH: New York, NY 33. B Sellegren, Ed, Molecularly Imprinted Polymers  Man-Made Mimics of Antibodies & Their Applications in Analytical Chemistry (2001) Elsevier: Amsterdam 34. V Pichon, M Bouzige, C Miege & M-C Hennion, Trends Anal Chem 18 (1999) 219 35. S Li & S Weber, Anal Chem 69 (1997) 1217 36. J Wu & J Pawliszyn, J Chromatogr 909 (2001) 37 37. J Wu, W Mullett & J Pawliszyn, Anal Chem 74 (2002) 4855 38. S-L Chong, D-X Wang, J Hayes, B Wilhite & A Malik, Anal Chem 69 (1997) 4566 39. AD Young, Boundary Layers (1989) BSP Professional Books: Oxford 40. H Geppert, Anal Chem 70 (1998) 3981 41. E Baltussen, P Sandra, F David & C Cramers, J Microcol Sep 11 (1999) 737 42. The mathematics of diffusion and heat transfer are equivalent because both processes are described by Laplace’s equation. Formulae for heat transfer can be used for mass transfer by substituting diffusion coefficients for thermal conductivity and diffusivity constants (see Ref. [20], Section 113I). The term D/δ in equation (339) corresponds to the heat transfer coefficient in Newton’s formula for heat transfer at interfaces. Therefore the formula for heat transfer at interfaces can be used to estimate the thickness of an effective diffusion boundary layer. 43. A Mikheyev, Fundamentals of Heat Transfer (1966) Mir Publishers: Moscow 44. A Ogawa, Vortex Flow (1993) CRC Press: Boca Raton, FL 45. Z Zhang & J Pawliszyn, Anal Chem 65 (1993) 1843 46. Z Zhang & J Pawliszyn, J High Resolut Chromatogr 19 (1996) 155 47. HL Lord & J Pawliszyn, Anal Chem 69 (1997) 3899 48. J Rosenfeld, Recent Developments in the Chemistry & Application of Analytical Derivatisations, in J Pawliszyn, Ed, Sampling & Sample Preparation for Field & Laboratory (2002) Elsevier: Amsterdam 49. P Wankat, Rate-Controlled Separations (1990) Elsevier Applied Science: New York, NY 50. F Dullien, Porous Media (1992) Academic Press: San Diego, CA 51. C Horvath & H Lin, J Chromatogr 149 (1978) 43 52. J Cadzow & F van Landingham, Signals, Systems, & Transforms (1985) Prince Hall: Englewoods Cliffs, NJ 53. J Langenfeld, S Hawthorne, D Miller & J Pawliszyn, Anal Chem 67 (1995) 1727

Theory of Solid-Phase Microextraction

59

54. J Langenfeld, S Hawthorne, D Miller & J Pawliszyn, Anal Chem 65 (1993) 338 55. BE Richter, BA Jones, JL Ezzell, NL Porter, N Avdalovic & C Pohl, Anal Chem 68 (1996) 1033 56. J Pare, J Belanger, K Li & S Stafford, J Microcolumn Sep 7 (1995) 37 57. Z Miao, Z Zhang & Pawliszyn, J Microcolumn Sep 6 (1994) 459 58. A Clifford & N Cotton, Anal Chem 63 (1991) 2371 59. D Ruthven, Principles of Adsorption & Adsorption Processes (1984) John Wiley & Sons: New York, NY 60. M Chai & J Pawliszyn, Environ Sci Technol 39 (1995) 693 61. Y Chen & J Pawliszyn, Anal Chem 75 (2003) 2004 62. J Koziel, Sampling & Sample Preparation for Indoor Air Analysis, in J Pawliszyn , Ed, Sampling & Sample Preparation for Field & Laboratory (2002) Elsevier: Amsterdam 63. J Petty, C Orazio, J Huckins, R Gale, J Lebo, K Echols & W Cranor, J Chromatogr 879 (2000) 83 64. B Vrana & G Schuurmann, Environ Sci Technol 36 (2002) 290 65. J Crank, Mathematics of Diffusion (1989) Clarendon Press: Oxford, p. 14 66. R Eisert & J Pawliszyn, Anal Chem 69 (1997) 3140 67. J Pawliszyn, J Chromatogr Sci 31 (1993) 31

3 Development of SPME Devices and Coatings Janusz Pawliszyn Department of Chemistry, University of Waterloo, Waterloo, Ontario, Canada

3.1

Historical Perspective

The early work on laser desorption/fast gas chromatography (Fast GC) conducted in our laboratory resulted in rapid separation times, even for very high molecular mass species.1 However, the preparation of samples took hours, which was more than an order of magnitude longer than the separation times. Optical fibres were used in this experiment to transmit laser light energy to the gas chromatography (GC) instrument. The sample preparation process used was analogous to standard solvent extraction procedures. The fibre tip was coated with the sample by dipping one end of the optical fibre in the solvent extract, coating the fibre and then removing the volatile solvents through evaporation. The fibre tip, prepared in such a way, was inserted into the injector of a gas chromatograph, and analytes were volatilised onto the front of the GC column by means of a laser pulse. During this work, a need for rapid sample preparation techniques was recognised in order to retain the time efficiency advantages made possible by using the laser pulse and a high-speed separation instrument. The challenge was addressed using fibres because optical fibres could be coated with several types of polymeric films. The original purpose of the coatings was simply to protect the fibres from breakage. Because of the thin films used (10 100 µm), the expected extraction times for these systems were very short. In addition, novel films could be prepared based on the knowledge gained from fused silica capillary column manufacturing experience. In the initial work on SPME, sections of fused silica optical fibres, both uncoated and coated with liquid and solid polymeric phases, were dipped into an aqueous sample containing test analytes and then placed in a GC injector.2 The process of introducing and removing the fibres required the opening of the injector, which resulted in loss of head pressure at the column. Despite their basic nature, the early experiments provided very important preliminary data that confirmed the usefulness of this simple approach because both polar and non-polar chemical species were extracted rapidly and reproducibly from aqueous samples. The development of the technique accelerated rapidly with the implementation of coated fibres incorporated into a microsyringe, resulting in the first SPME device.3 Figure 3.1 shows an example of an SPME device based on the HamiltonTM 7000 series microsyringe. The metal rod, which serves as the piston in a microsyringe, Handbook of Solid Phase Microextraction. DOI: 10.1016/B978-0-12-416017-0.00003-6 © 2012 Elsevier Inc. All rights reserved.

62

Handbook of Solid Phase Microextraction

Syringe barrel Syringe needle

Plunger cap

Silica fibre Coating

Epoxy glue Stainless steel microtubing

Figure 3.1 The custom-made SPME device based on the Hamilton 7000 series syringe.

is replaced with stainless steel microtubing with an inside diameter (i.d.) slightly larger than the outside diameter (o.d.) of the fused silica rod. Typically, the first 5 mm of the coating is removed from a 1.5-cm fibre, which is then inserted into the microtubing. High-temperature epoxy glue is used to mount the fibre permanently. Sample injection is performed in a manner similar to standard syringe injection. Movement of the plunger allows exposure of the fibre during extraction and desorption, and fibre protection in the needle during storage and penetration of the septum. In fact, SPME devices do not need expensive syringes like the Hamilton syringes. As Figure 3.2A illustrates, a useful device can be built from a short piece of stainless steel microtubing (to hold the fibre), another piece of larger tubing (to work as a ‘needle’) and a septum (to seal the connection between the microtubing and the ‘needle’). The design in Figure 3.2A is the basic building block of a commercial SPME device, described in Chapter 4, in Section 4.2, and illustrated in Figures 4.1 and 4.2. An additional improvement in the commercial device is the ability to adjust the depth of the fibre with respect to the end of the needle, which allows control of the exposure depth in the injector and extraction vessel. The commercial device also incorporates such useful features as colour marking of the fibre assemblies to distinguish among various coating types. Another simple SPME construction is based on a piece of internally coated tubing.4 This tubing can be mounted inside a needle or it can constitute the ‘needle’ of a syringe itself.5 Elimination of mechanical movement of a plunger of a syringe can be accomplished by sealing the tubing at one end and installing a microheater, as illustrated in Figure 3.2B. The expansion of air caused by a temperature increase allows the removal of desorbed analytes from the extracting phase located inside the tubing. A coated tubing approach is useful in the design of passive sampling devices discussed later because in this case, the extraction rate is limited by the diffusion of analytes into the needle.6 In addition, active sampling is possible by

Development of SPME Devices and Coatings

(A)

63

(B)

Attachment hub

Sealing septum

Cooling coils

Septum piercing needle

Coating Microtubing Coated fused silica fibre

Figure 3.2 Simple versions of SPME devices using (A) coated fibre or (B) internally coated tubing.

heating and cooling of air contained in the upper part of the tubing, which causes movement of liquid or gaseous samples into and out of the tubing, facilitating mass transport of analytes from the sample to the coating. The in-tube concept was also expanded to facilitate the automation of sample preparation for high-performance liquid chromatography (HPLC). In this in-tube solid-phase microextraction (SPME) approach, the sample components are extracted by the coating located on the inner surface of the hollow tubing and, after the extraction is completed, the analytes are washed into the HPLC column using the mobile phase or solvent. Everything is easily automated using a conventional autosampler (see Chapter 5). This concept is very similar to solid-phase extraction (SPE); some researchers used packed tubes.7 However, the fundamental difference between the two techniques is that SPE relies on exhaustive extraction, while in-tube SPME relies on equilibrium. Therefore, compared to SPE, in-tube SPME has different selectivity and fewer flow restrictions, and takes full advantage of phase capacity. The extraction coating has also been used on other elements of the analytical system. In addition to widely used fibre and in-tube geometry, SPME devices can take different configurations, such as the coated interior of vessels,8 the coated exterior of magnetic stirring bars9 or even pieces of polydimethylsiloxane (PDMS) tubes and thin membranes.10,11 Several of these implementations are shown in Chapter 1, in Figure 1.3. Thin-film devices are discussed in Section 3.2.4.

64

Handbook of Solid Phase Microextraction

The main reason for developing these alternative approaches is to enhance sensitivity by using a larger volume of the extraction phase and to improve the kinetics of the mass transfer between sample and the extraction phase by increasing the surface-to-volume ratio of the extraction phase. The main disadvantage of these approaches, however, is the loss of the convenience associated with a syringe configuration, in particular for the introduction of the sample into the analytical instrument. For the alternative-approach devices, sample introduction requires the use of high-volume desorption devices and creates difficulties in the automation of the extraction process. Other difficulties relate to handling volatile compounds, which may be lost during transfer of the extraction phase from the sample to the injection system. In view of the high sensitivities obtained for hydrophobic high-molecularweight compounds with fibre SPME, the advantages of using larger volume phases are limited, especially for small sample volumes.12 Although to date, SPME devices have been used mainly in laboratory applications, recent research has been directed towards remote monitoring, particularly for clinical, field environmental and industrial hygiene applications. The operating principles of such devices are analogous to those of the devices described above, but modifications are made for greater convenience in the given application. An example is shown in Figure 3.3: adding a tube with a small opening to cover the needle of the SPME syringe results in a useful device for breath analysis in a non-invasive clinical application.13 This design can be improved further by adding two one-way valves, mounted at the mouthpiece and on the exit aperture (Figure 3.4), but the concept remains the same.14 Figure 3.3 SPME device modified for breath analysis.

Exposed fibre Aperture

Inert tubing

SPME sampling port

Membrane module

Carrier gas in Mini fan

Breath sample out

Breath sample in

One-way valves

Figure 3.4 Breath analysis apparatus based on SPME.14 (Source: Reprinted with permission of American Chemical Society. r 2002.)

Development of SPME Devices and Coatings

65

The two valves allow the entrapment of alveolar air and prevent its accidental inhalation. The degree of dilution of alveolar air with environmental air can be measured by monitoring carbon dioxide concentration.15

3.2

Rational Design of SPME Devices

A better understanding of SPME theory allows more rational optimisation of extraction conditions. It can also be used to design devices that would offer improved extraction efficiency.16

3.2.1

Agitation for Air Sampling

When performing extraction from gaseous samples using short sampling times (pre-equilibrium), it has been observed that increasing wind speed to about 5 cm/s enhances extraction. The data indicate that the wind speed or air bulk movement significantly affects the volatile organic contaminant (VOC) mass transfer process from bulk air to the fibre. The VOC mass loading on the fibre increases as the wind velocity increases from 0 to 5 cm/s. No further change was observed as the wind speed was increased from 5 to 20 cm/s. This indicates that the thickness of the boundary layer between the fibre and air diminishes as the wind speed increases, which results in an increase of the mass transfer rate and mass loading on the fibre. After the thickness of the boundary layer has been decreased to a certain degree, uptake becomes limited by diffusion within the pores of the polymer, and no further increase in extraction is seen with increasing wind speed. In practice, when using SPME to sample in the field, it would be helpful to use a fan to move air samples across the fibre during sampling to eliminate possible imprecision in extraction due to variations in wind speed. Figure 3.5 shows an example of an agitation device for field air sampling, consisting of a modified hairdryer fan with a mounting for the SPME device.17,18 Appropriate devices for water sampling can be designed as well.

3.2.2

Agitation for Aqueous Sampling

Efficient agitation for aqueous sampling can be achieved using a bench drill and attaching the SPME fibre or PDMS thin-film to the drill like a drill bit (Figure 3.6). The drill chuck can be tightened onto the coloured hub of the SPME fibre to keep it secure. PDMS thin-films can also be used with this setup. The PDMS thin-films are cut into a particular house-like shape and secured using a metal wire. The wire can then be attached to the drill like a drill bit. Glass sampling vials were held immobile using ring stands and clamps.18,20 A portable, handheld device was developed from a cappuccino stirrer that was able to hold a commercial SPME fibre. Two Teflon disks were attached to the rotating shaft of the stirrer using small screws. The commercially available SPME fibre can screw into the top disk. The bottom disk has a small hole to secure the

66

Handbook of Solid Phase Microextraction

SPME holder O-ring SPME insert

Holding brace

Air sample Modified hairdryer (with reversed airflow)

SPME fibre Aluminium hollow cylinder

15 DC power supply

Figure 3.5 Constant agitation device for field air sampling by SPME. Figure 3.6 SPME field water sampling devices using a rotated fibre and a rotated membrane.19 (Source: Reprinted with permission from Elsevier.) Drill

Disk

PDMS membrane

SPME fibre

position of the fibre. This disk is movable and can be adjusted to expose the fibre during sampling or withdraw the fibre after use. Once the handheld stirrer is turned on, the fibre rotates and the water is agitated. The Teflon disks prevent the fibre from spinning outward during the sampling process. This designed shaft can also be used in the field with a portable drill.

Development of SPME Devices and Coatings

Inner capillary CO2 in

CO2 out

67

Figure 3.7 Design of internally cooled SPME device.

Outer capillary

Syringe barrel

Ferrule Needle Headspace

Coating Vial

Sample Heater

The challenge with these devices is ensuring that they provide constant agitation. Both the bench drill and the portable cordless constant speed drill have wellcontrolled speeds, but the portable cordless variable speed drill requires aid in fully depressing the trigger for the entire sampling period. The handheld stirrers are good for spot sampling, but only if the extraction time is short, because otherwise, the strength of the batteries decreases and slows the agitation rate. Also, if extraction times greater than 15 min are used, the motor of the handheld stirrer will burn out.

3.2.3

Cold-Fibre SPME

Extraction temperature is another important parameter to consider during SPME. At elevated temperatures, native analytes can effectively dissociate from the matrix and move into the headspace for rapid extraction by the fibre coatings. However, the coating/sample distribution coefficient also decreases with an increase in temperature, resulting in a diminution of the equilibrium amount of analyte extracted, as described previously (see Section 2.3.4). To prevent loss of sensitivity, the coating can be cooled simultaneously with sample heating. This idea was implemented in the design of the device shown in Figure 3.7. In this device, a fused silica tube is sealed and coated at one end (the outer surface of the capillary). Liquid carbon dioxide is delivered via the inner capillary to the coated end of the outer capillary, resulting in a coating temperature lower than that of the sample. This ‘cold finger’ effect results in accumulation of the volatilised analytes at the tip of the fibre. The internally cooled fibre approach can be used effectively to increase the sensitivity and extraction speed of SPME. Exhaustive extraction of many analytes, including volatiles, is possible with this method.21 Figure 3.8 shows the example of data corresponding to the extraction of a range of polynuclear aromatic hydrocarbons (PAHs) from the solid matrix as a function of the matrix temperature.

68

Handbook of Solid Phase Microextraction

120.0

Extracted amount (ng)

100.0 80.0 60.0 40.0 20.0

en e

Ac en ap ht hy l

N

ap ht ha le ne

0.0

Compounds 100°C

120°C

140°C

160°C

180°C

200°C

Figure 3.8 Extraction temperature profile of cold fibre for solid matrix sampling.

Coating temperature was kept constant at 40 C. The data indicates that with the increase in the temperature gap between fibre coating and matrix, the amount of analyte increases, as the theory would predict. However, the increase is more dramatic for larger PAHs since additionally the desorption rate from the matrix is enhanced with increase of the temperature similarly as it was previously observed in high-temperature supercritical fluid extraction (SFE) and accelerated solvent extraction (ASE). The advantage of the cold-fibre SPME approach compared to SFE and ASE is that the non-volatile interferences are eliminated because the analytes need to have some vapour pressure at the experimental conditions to reach the fibre, as it is a headspace arrangement. The results reported in Figure 3.8 are obtained on a fully automated system, as discussed in Chapter 5 in Section 5.1.6.2. An alternative approach to cooling the fibre uses a Peltier cooler.22

3.2.4

High-Surface-Area Samplers (Thin-Film Microextraction)

The extraction rate after exposure of the SPME device to the sample is proportional to the contact surface area between the sorbent and the sample (see Chapter 2). In order to increase the mass uptake rates (and, therefore, the sensitivities), large surface area sorbent geometries can be used. For example, the PDMS extraction phase can be a thin membrane, as shown in Figure 3.9.

Development of SPME Devices and Coatings

69

Figure 3.9 High-surface-area SPME samplers. 1 cm

Thin-film

Stainless steel wire

2 cm

2 cm

Step 1. Rotate thin-film as it is inserted into the liner

Stainless steel wire

Step 2. Place cap onto the liner

Desorption liner

Thin-film

Step 3. Introduce liner for thermal desorption

Coiled thin-film

Figure 3.10 Introduction of a high surface area sampler into a GC injector.

In this case, a high surface area-to-volume ratio is obtained, resulting in very high accumulation rates. This approach is particularly beneficial for hydrophobic, semi-volatile components characterised by very high distribution constants. To facilitate convenient introduction to the analytical instrument, the membrane can be attached to the holding rod (Figure 3.9), and, after extraction, the membrane can be rolled around the rod and introduced to the injection system for the desorption of extracted components (Figure 3.10).

70

Handbook of Solid Phase Microextraction

Sliding syringe

Figure 3.11 ‘Leaf’ SPME field sampling device.

Leaf

3.3 3.3.1

On-Site Samplers Field Samplers

An important feature of a field sampling device is the ability to preserve extracted analytes in the coating. A simple, practical way to accomplish this goal is to seal the end of the needle with a piece of septum. In addition, cooling extends the storage time. Polymeric septum material, however, may cause losses of analytes from the fibre. Therefore, a more appropriate approach is to use metal-to-metal (or solid polymer) seals. Figure 3.11 illustrates an example of a device construction based on a ‘leaf’ closure.19 It is anticipated that future devices designed for field applications will be more rugged than the current laboratory versions and will look more like sticks or pens than syringes. One of the possible designs is shown in Figure 3.12.23 Muller24 described several equilibrium-based SPME field sampling devices and discussed the typical losses of volatiles and contamination of the fibre during storage and transport. The percentage of compound retained in the coating was evaluated for the four devices, at different storage times and temperatures, and for different fibre coatings. The PDMS fibre demonstrated the lowest ability to retain these compounds. Carboxen PDMS had the highest ability, and PDMS DVB had intermediate retaining properties. The devices employed various sealing methods in order to preserve the samples on the fibre and to protect from external contamination. The Supelco field sampler seals by retracting the outer needle behind a silicone septum. Alternatively, a conventional manual sampler may be sealed by pressing the end of the needle into a piece of septum. Two prototype devices constructed were sealed by either a leaf system (Figure 3.12), which opens automatically when the outer needle is exposed, or by capping the outer needle after sampling with a Teflon cap. Another promising technique uses a valve syringe, where the fibre is withdrawn into the barrel of the syringe, along with a sample of the air being analysed, by

Development of SPME Devices and Coatings

71

(A)

(B)

(C-1)

(C-2)

Figure 3.12 SPME field air sampling device. (A) Configuration for standby, storage or transportation. (B) Configuration when the protecting shield is pulled outwards and locked in the sampling position. (C-1 and C-2) Configurations for TWA sampling and grab sampling, respectively.

means of retracting the syringe plunger. The valve is then sealed until analysis. Initial data describing the new approaches clearly demonstrate the advantages of the new designs to seal the fibre in the needle, compared to the commercial devices, by eliminating the losses of volatile components to the septa material as well as contamination.

3.3.2

NT Devices

As an alternative to using the coated fibre in the needle SPME device, sorbent can fill the whole diameter of the needle, forming a needle trap (NT) device.25 Convenient NT field sampling devices have been designed in the form of badges and pens for both spot sampling and time-weighted-average (TWA) sampling.26 There are two ways to immobilise the sorbent. One technique is to use a sidehole needle, and the other is to use an open tubular needle with the sorbent particles glued together on the surface. The side-hole needle device is easier to prepare, but it is more costly because the side-hole needle is more expensive. Two types of sorbents were used to build NT devices: a single layer and a segmented sorbent with progression of the sorbent strength from the tip to the inside of the needle. The segmented sorbent allows convenient desorption because in needle operation, the flow of the desorption system is naturally reversed compared to the extraction. Different desorption gas delivery techniques have been applied with NT, such as (i) the syringe piston delivery of a well-defined volume of the desorption gas; (ii) divergence of the gas flow after needle placement in the injector, the technique used primarily with the automated system available from PAS Technologies; and finally (iii) the design of the narrow neck insert in combination with the

72

Handbook of Solid Phase Microextraction

Sorbent

Side hole O

DVB

CAR Stainless spring

Needle

Figure 3.13 Schematic of an NT device with the side hole to facilitate convenient on-site desorption. MCounts 6

10 mm DVB NTD: 18 mL sampling volume

5 4 3 2

Allethrin

Benzene Toluene

Phenol

1 0 MCounts 6

65 µm DVB/PDMS fibre: 10 min sampling time

5 4 3 2 1 0 5

10

15

20

Figure 3.14 GC/MS chromatograms for SPME fibre and NT devices after sampling from mosquito-coil smoke.27

NT-containing hole above the sorbent position (Figure 3.13) to facilitate convenient desorption. In contrast to SPME, using NT devices is an exhaustive technique, but it can work very effectively in combination with needle SPME devices to better characterise the sample directly on-site. Figure 3.14 illustrates GC chromatograms obtained for SPME and NT of mosquito-coil smoke.27 This figure indicates that SPME does not extract volatile components very well because of low K values, but it performs very well for mid-range compounds, which exist as free analytes in air. Heavy (less volatile) components are particle bound and are not available for extraction by the fibre. However, they are extracted well using NT devices. Therefore, the combination of two needle-based devices (NT and SPME) provides a good characterisation of the complex air samples. SPME can be used to concentrate free analytes, which are available and can cause immediate impact, while NT measures the total amount of analytes in the sample. The combination of the two approaches can, therefore, provide useful information about the extent that the

Development of SPME Devices and Coatings

73

Table 3.1 Summary of Data for Analysis of Styrene, Obtained from a Field Study Using (A) TWA Charcoal Tubes and (B) TWA by SPME with 100 µm PDMS Fibre (retracted 0.3 cm) Sample Type

SPME 100-µm PDMS Charcoal Tubes Passive Badge PID

5 min sample (µg/L) 130 30 min TWA (µg/L) 56

97 54

90 72

50 250 N/A

analytes are bound to the particulate matter in air. This measurement can be performed directly and conveniently on-site using a single portable analytical instrument.

3.3.3

TWA Devices

The commercially available SPME device (with the fibre retracted in the needle) was used as a TWA diffusive sampler by Martos and Pawliszyn.28 Airborne (industrial) formaldehyde concentrations were measured with the device using on-fibre derivatisation and the results were compared with those produced by active air sampling using the NIOSH-2541 method.29 The device was also used for TWA air sampling of normal alkanes from C5 to C15.30,31 Field sampling trials were performed in a house, an apartment and a school. The results were compared with those of active sampling through charcoal tubes using the NIOSH-1550 method. The results obtained from the field study for the determination of TWA concentration of airborne styrene using the SPME device with PDMS coating and charcoal tubes are summarised in Table 3.1. Later, Chen and Pawliszyn used the fibre-retracted SPME device to determine the TWA concentrations of VOCs in air (Figure 3.15)32,33 and demonstrated that the face velocity of air across the needle opening does not affect sampling, due to the extremely small inner diameter of the fibre needle. Ouyang et al.34 extended the applications of this type of SPME device to TWA passive water sampling (Figure 3.16). The Hamilton 500 µL gastight syringe was modified as a TWA water sampling device to ensure that all the air in the SPME needle could be replaced with water. In addition, a removable needle was designed to avoid the effect of the adsorption of the target analytes on the outside wall of the needle. This field TWA water sampling device was used to monitor PAHs in Hamilton Harbour and Laurel Creek, Canada.34,35

3.3.4

In Vivo Samplers

In vivo research is better suited than in vitro research for observing an overall effect. SPME is one of the most promising techniques for in vivo sampling and subsequent analysis. According to SPME theory, above a certain sample size, sample volume does not affect the results (see Chapter 2); therefore, it is not necessary to define a specific sample size for the analysis, which is very desirable for on-site

74

Handbook of Solid Phase Microextraction

Figure 3.15 SPME TWA passive air sampling device: additional groves allow retraction of the fibre inside the needle for the desired length.32 (Source: Reprinted with permission of American Chemical Society. r 2003.)

Diffusion path

ID 0.76 mm silicon steel tube Teflon septum

Teflon septum

SPME fibre

Copper mesh

Figure 3.16 Schematic diagram and graphic of the SPME TWA field water sampler.34 (Source: Reprinted with permission from Elsevier.)

sampling. In addition, SPME directly extracts a small fraction of free analyte so that a negligible depletion of the free fraction is achieved after extraction. This depletion will be compensated by dynamic systems containing both bound and unbound analyte and the system will not be disturbed significantly. Finally, the technology is easy to miniaturise; thus, it is suitable for both small living systems and microanalytical instruments. Experimental errors and the time associated with sample transport and storage can be reduced, resulting in the collection of more accurate, more precise and faster analytical data. These advantages make SPME a promising tool for the direct assay of in vivo chemical concentrations.

Development of SPME Devices and Coatings

75

Highly or moderately volatile compounds are usually determined in the headspace using existing commercial SPME devices (Chapter 4). For headspace sampling, the investigated individuals, plant parts or insects are enclosed in special chambers that are usually designed to minimise interferences.36 45 Studies involving human exudates have also been conducted in a similar manner.46 48 When using headspace sampling, any extraction phase is suitable, as long as it is robust and innocuous; however, the use of PDMS and PDMS/DVB coatings predominates in most applications. In the case of polar, non-volatile analytes, only direct extraction is feasible. The fibre coating is either carefully rubbed against the sample36 38 or inserted with a special in vivo device49,50 (Figures 3.17 and 3.18). The development of biocompatible extraction phases for SPME has led to significant advances in bioanalysis: all sample preparation steps can be combined into a single one, even for complex biological samples, such as whole blood or plasma. Furthermore, biocompatible devices permit direct extraction of target analytes from the flowing blood of living organisms. As shown in Figure 3.18, an in vivo device that will soon become commercially available from Supelco, Inc., consists of an SPME fibre that is housed inside a Sampler for in vivo microextraction (based on a hypodermic needle) Tissue with blood vessel

Figure 3.17 Direct in vivo sampling of flowing blood.50 (Source: Reprinted with permission from Elsevier.) Figure 3.18 Photo of commercial in vivo assembly. (Source: Courtesy of Supelco.)

76

Handbook of Solid Phase Microextraction

hypodermic needle. This needle is used to protect SPME coating during storage and transportation as well as to pierce the tissue/blood vessel/sampling interface. The main features of SPME coating used in this device are that it is (i) biocompatible in order to prevent adverse reactions during sampling and to minimise fouling of the SPME coating surface by the adhesion of proteins present in biological fluid, (ii) immobilised on flexible metal alloy to render it sufficiently robust for in vivo use and (iii) compatible with most commonly used solvents in order to facilitate desorption. Using the above device, a fast in vivo microextraction technique that has the potential to replace current methods of analysis based on drawing blood was recently reported.49,51,52 The devices were used for direct extraction of drugs from the flowing blood of beagles over a period of 8 h, in order to construct pharmacokinetic profiles for diazepam and its metabolites. The drugs extracted on the probes were subsequently quantified by LC MS/MS. Two calibration strategies external and standard on the fibre were employed to correlate the amount extracted with the in vivo concentration. The microextraction technique was validated by comparison with conventional plasma analysis, with a correlation factor of 0.99 (Figure 3.19A). In addition to total concentrations, the method was very useful for determining free drug concentrations (Figure 3.19B). These results demonstrate the unique advantages of in vivo SPME and highlight its potential as a valuable new tool for fast clinical analysis. More data points may be obtained for each animal, and sampling can be performed simultaneously at multiple sites in one animal without the risk of exsanguination. Such an approach can be used not only for drugs but also for any other endogenous or exogenous compounds. Chapter 12 discusses in vivo applications of SPME in detail. In vivo sampling device has also been developed for tissue sampling, including fish sampling (Figure 3.20). A sterile hypodermic needle containing the SPME sampler is attached to the device. The user punctures the tissue by the needle to the desired depth, determined by the depth guide. Depressing the plunger allows the hypodermic needle to be removed while keeping the SPME probe embedded at the desired location. Retracting the entire device allows the SPME probe to remain in the tissue without the external sampler present (Figure 3.21). At the end of sampling, the SPME probe is simply pulled out.

3.4

Development of New SPME Coatings

The efficiency of the extraction process is dependent on the distribution constant Kfs (see Chapter 2). Kfs is a characteristic parameter that describes properties of a coating and its selectivity towards the analyte in contrast to other matrix components. Specific coatings can be developed for a range of applications. The coating volume also limits method sensitivity, but thicker coatings result in longer extraction times. It is important to use the appropriate coating for a given application. This is clearly demonstrated in Figure 3.22, which compares the performance of different coatings

Development of SPME Devices and Coatings

(A)

77

Concentration (ng/mL)

1000

100 Blood Plasma SOF-PPY SOF-PEG

10

1 0

1

2

3

4

5

6

7

8

9

Time (h)

Free concentration (ng/mL)

(B) 100

10 Diazepam Nordiazepam Oxazepam

1

0.1 0

1

2

3

4

5

6

7

8

9

Time (h)

Figure 3.19 (A) Diazepam pharmacokinetic profile, from three studies on three dogs (n 5 9). ‘Blood’: in vivo SPME from whole blood,50 ‘Plasma’: conventional analysis, ‘SOF-PPY’: in vivo standard on the fibre with PPY probes, ‘SOF-PEG’: in vivo standard on the fibre with PEG probes. (B) Free concentration profiles of diazepam, nordiazepam and oxazepam from the same study. (Source: Reprinted with permission from Elsevier.)

Figure 3.20 Picture of in vivo tissue sampler holder.

for analysis of polar and non-polar compounds from an aqueous matrix. The distribution constant and the sensitivity of the method drop over two orders of magnitude for o-xylene, and increase by an order of magnitude for 2,4-dichlorophenol when the film is changed from non-polar PDMS (Figure 3.22A) to polar polyacrylate polymer

78

Handbook of Solid Phase Microextraction

Figure 3.21 Juvenile rainbow trout in a tank with SPME probe embedded in back.

(A)

Total ion current

o-xylene K = 794

2,4-dichlorophenol K = 4.6

Total ion current

(B)

2,4-dichlorophenol K = 47 o-xylene K = 5.4

10

15 Retention time (min)

Figure 3.22 Analysis of compounds with different polarity from water using (A) PDMS and (B) PA coating.53 (Source: Reprinted with permission of American Chemical Society. r 1994.)

(PA) (Figure 3.22B).53 Chapter 4 discusses in detail all of the commercially available coatings and the type of applications they are most suitable for.

3.4.1

Coating Preparation Methods

In addition to liquid polymeric coatings, such as PDMS, for general applications, other more specialised materials have been developed for use as SPME coatings. There are several methods of depositing coatings onto fibres.

Development of SPME Devices and Coatings

79

3.4.1.1 Dipping Technique The dipping technique typically consists of placing a fibre in a concentrated organic solvent solution of the material to be deposited for a short time. After removal of the fibre from the solution, the solvent is evaporated by drying and the deposited material can be crosslinked.2

3.4.1.2 Electrodeposition An extension of dipping method is electrodeposition, which can be used to deposit selective coatings on the surface of metallic rods. The main limitation of this approach is coating thickness variance from fibre to fibre. Therefore, the preparation of films for commercial devices is carried out simultaneously with the drawing of the fused silica rod using dedicated equipment, in order to obtain very reproducible coating thickness. This process is identical to the preparation of optical fibres,54 so the required equipment is commercially available.

3.4.1.3 Hollow Fibre Membranes/Adhesive Tape The simplest way to prepare a coating is to use a piece of hollow fibre membrane (small i.d. tubing, commercially available), made from the desired extraction material.55 Preparation consists of swelling the membrane by means of an appropriate volatile solvent, placing the enlarged membrane onto the tip of the fibre and evaporating the solvent. Membrane thickness determines the thickness of the coating. Therefore, the volume of the coating can be large, reaching up to 3 µL for a 300-µm-thick PDMS hollow-fibre membrane. A porous hollow-fibre membrane can also be used for adsorption of target analytes, or its pores can be filled with organic solvent to allow for solvent microextraction.56 More recently, carbon tape (commercially available, traditionally used to immobilise samples prior to microscopy) was found to have good extraction properties for some non-volatile analytes amenable to LC.57 This type of coating is adhesive, so it is easily immobilised on a piece of stainless steel wire of desired diameter.

3.4.1.4 Adhesion of Coatings The use of strong adhesive to immobilise sorbent particles was originally proposed in 1997 to prepare SPME fibres suitable for GC use.58,59 The main idea in this approach is to immobilise a thin layer of sorbent particles (e.g. porous coated silica particles) on a metal wire using an appropriate adhesive. Depending on the choice of adhesive, coatings suitable for LC can also be prepared.60,61 The main advantages in this approach are (i) flexibility in the choice of glue, (ii) wide availability of different commercial sorbents so that the coating can be tailor-made for the particular application and (iii) low cost. For GC applications, the adhesive used should have good thermal stability, while for LC applications, good chemical stability is crucial so that the desorption step does not result in loss of coating. Depending on the diameter of metal support used as well as proper optimisation of coating

80

Handbook of Solid Phase Microextraction

procedure, a coating with excellent inter-fibre reproducibility can be made using this approach.61 Depending on the size and properties of the sorbent employed, these coatings tend to have reasonable extraction kinetics for drugs and other non-volatile compounds. For example, equilibrium for benzodiazepines could be reached in 30 min with agitation, while thick PDMS and carbon tape coatings require 10 h, or even longer, to reach equilibrium.57,61

3.4.1.5 Conducting Polymers Conducting polymers are versatile materials in which molecular/analyte recognition can be achieved in different ways, including (i) the incorporation of counter ions that introduce selective interactions, (ii) the use of inherent and unusual multifunctionality (hydrophobic, acid base interactions, polar functional groups, ion-exchange, hydrogen bonding, electroactivity, etc.) of the polymers, (iii) the introduction of functional groups to the monomers, (iv) the co-deposition of metals or other monomers within the polymer and (v) the application of appropriate electrochemical potentials. Based on these properties, conducting polymers have found wide application in many fields, including separation science, chemical sensors62 and electrochemical analysis.63 So far, the most widely used conducting polymers are based on polypyrrole (PPY), polythiophene and polyaniline. Of these three classes of materials, PPY and its derivatives have been intensively used and studied in recent years, due to additional advantages, to wit: (i) they can be polymerised easily from organic or aqueous media at a neutral pH by electrochemical or chemical methods, (ii) they are relatively stable in air and solution, and (iii) pyrrole monomer and some of its derivatives are commercially available. One of the main difficulties limiting the wide application of SPME LC is the absence of a suitable extraction phase that not only has high extraction ability for the polar analytes but is also stable in solutions of various matrices. PPY coatings have been investigated to fill this gap. PPY coatings can extract polar or even ionic analytes.64,65 They also open the possibility of exploring conductive properties of the polymer coating by applying a charge to the polymer during extraction in order to extract analytes of interest selectively, and then reversing the charge to facilitate desorption.66 Due to its biocompatibility, PPY-coated fibres were also developed as SPME probes for in vivo pharmacokinetic studies.49 Another advantage of PPY coatings is that they can be prepared as very thin films (,10 µm), which means that they require short extraction times in order to reach equilibrium. This facilitates their use in vivo under equilibrium conditions. As PPY is a porous coating, it extracts analytes mainly by adsorption processes. Consequently, the linear range of the probe is low and depends on the concentration of other compounds. This problem is significant in complicated matrices, such as whole blood or plasma, where many endogenous compounds exist. Another disadvantage of PPY coatings is relatively poor inter-fibre reproducibility (B30% standard deviation).67

Development of SPME Devices and Coatings

81

3.4.1.6 Sol Gel Coatings Another method of preparation of SPME fibres is using the sol gel chemistry.68 75 The main advantages of this type of the coating are the low cost and strong adhesion of the coating to the substrate, which can translate into improved thermal and chemical stability. For example, sol gel PDMS coating proposed by Chong et al.69 was able to withstand GC injector temperatures of 320 C in contrast to commercial PDMS coating, where bleeding may be observed for injector temperatures above 200 C. In addition, sol gel procedures can be used to prepare very thin coatings (as thin as 1 µm, but typically around 5 10 µm), which improves the extraction kinetics and results in shorter extraction times. The main steps of sol gel coating procedure typically include (i) pre-treatment of substrate surface, (ii) preparation of sol gel solution, (iii) coating of the substrate with sol gel solution using dipping method and (iv) conditioning of the coating.68 Most commonly employed sol gel precursors are methyltrimethoxysilane and tetraethoxysilane, while the extractive properties of the coating can be adjusted by incorporation of various modifiers such as crown ether,71 PDMS,69 polyethylene glycol (PEG)73 and so on. For example, to prepare sol gel coating suitable for LC use, Gbatu et al.68 incorporated n-octyltriethoxysilane in order to increase hydrophobicity of the coating.

3.4.2

Affinity-Based Coatings

3.4.2.1 Molecularly Imprinted Polymer Coatings A very pronounced difference in selectivity towards target analytes and interferences can be achieved by using surfaces common to affinity chromatography. Using the method of polymer imprinting, antibody mimics can be generated with specificity to a desired analyte.76 Briefly, the desired affinity can be introduced by adding an amount of the compound of interest to the polymerisation reaction. This ‘pattern’ chemical may be removed after polymerisation, leaving vacant sites of a specific size and shape that are suitable for binding the same chemical again from an unknown sample. Non-specific binding should be controlled for, but enhancements in sensitivity are obtained, particularly at low analyte concentrations. Molecularly imprinted polymers (MIPs) have been successfully immobilised on SPME fibres or used as coatings for in-tube SPME.77 81 For example, Hu et al.80 recently developed an MIP SPME method for the determination of tetracyclines in milk, chicken feed and chicken muscle. The main advantages of MIP coating were found to be high selectivity, excellent chemical stability, easy preparation, short equilibrium extraction times (30 min) and reusability (.100 times).

3.4.2.2 Immunoaffinity Coatings Antibodies show high affinity and extraordinarily specific recognition ability towards their complementary antigen in biological systems. Immunoassays, an analytical method based on the antibody antigen specific reactions, has become an essential routine and research tool throughout the biological sciences. It is

82

Handbook of Solid Phase Microextraction

particularly useful in clinical analysis because of its unique selectivity, extremely low limits of detection and applicability to a variety of compounds. The selectivity of antibodies allows for the quantification of compounds at trace levels in the presence of other structurally or chemically similar compounds. The use of antibodies enables a selective preconcentration that is essential for the analysis of complex samples, such as biological matrices. Therefore, affinity sorbents, such as those based on immobilised antibodies, play an ever-increasing role in the development of new biological separation and analysis methods. Recently, the extraordinary selectivity of antibodies was used in combination with SPME by immobilising antibodies on glass fibres using covalent glutaraldehyde bonding.82,83 The specificity of immobilised antibodies, the maximum antigen binding capacity and the competitive binding of different antigens to the immobilised antibodies were studied. The coatings were successfully applied for the trace analysis of 7-aminoflunitrazepam in urine, with the limits of detection and quantitation of 16 and 34 pg/mL, respectively.82

3.4.3

Biocompatible Coatings

The successful preparation of biocompatible and hemocompatible SPME coatings represents an important step towards developing powerful biomedical, pharmaceutical and forensic applications, as the advantages of SPME would be directly applicable to the analysis of biological samples. In addition, the main deterrent of in vivo application of SPME has been the lack of suitable extractive phases. It is expected that the number of such applications will increase significantly with the introduction of new biocompatible coatings. SPME fibres for in vivo applications should be robust, as thin as possible, unbreakable and highly flexible. The creation of non-fouling surfaces is one of the major prerequisites for microdevices (such as those shown in Figure 3.19) for biomedical and analytical applications. The term ‘biocompatibility’ refers to nonrejection of biological products or artificial devices that are in contact with living tissue.84,85 Incompatibility can lead to toxic reactions or immunological rejection. A material could be considered biocompatible if the sum of adverse humoral and cellular reactions occurring during exposure is lower than for a reference material. Furthermore, all the materials should be sterilisable, preferably by autoclaving (a widely available and accepted sterilisation procedure). Body fluids contain B1% NaCl, which constitutes a corrosive environment for the metallic part of the SPME fibres. The interactions of the coatings with body cells, products of corrosion and metal ions released from the wear debris can have adverse effects on the body and on the coatings. These effects can include cellular damage, infections, blood coagulation (potentially leading to thrombosis) and failure of the probe. Ti and its alloys are among the most biocompatible metals, but their wear resistance is relatively low. Co Cr alloys have better wear resistance but are less biocompatible. For the usual SPME applications, with in vivo extraction times of no more than 30 min, all the above-mentioned alloys are suitable. However, when highly flexible fibre cores are needed, Ti alloys should be used.

Development of SPME Devices and Coatings

83

Coating the fibres with biocompatible protective films, which can reduce corrosion and wear, may prevent or alleviate the problems described above. There are several approaches that are currently applied for effective direct extraction of lowmolecular-weight compounds from complex liquid matrices: (i) biocompatible polymer films which minimise protein adhesion, (ii) restricted access materials (RAMs) and (iii) use of a hollow membrane to form a concentric sheath around a coated SPME fibre in order to block the access of large particles or proteins to the coating surface, while target analytes with low molecular weight diffuse through the membrane and reach the extraction phase.

3.4.3.1 Biocompatible Polymer Coatings A useful strategy that can overcome the problem of biofouling is to passivate conventional extraction phases by creating a thin biocompatible interface (or film) through the coupling of certain neutral and hydrophilic macromolecules, such as polyhydroxyethyl methacrylate, polyacrylamide, poly(N,N-dimethyl acrylamide), dextran, polyacrylonitrile (PAN) and polyethylene glycol.86 These protective layers repel proteins and allow extraction of small molecules of target analytes. The modified SPME fibres can be easily prepared by physical adsorption of the biocompatible polymer onto the surface of regular commercial extraction phases. For example, when existing commercial coatings such as carbowax-templated resin (CW/TPR) were covered with a layer of PAN, the equilibration time remained essentially the same as illustrated in Figure 3.23 for verapamil. The extraction capacity of the fibres was reduced slightly with the addition of a thin biocompatible layer (Figure 3.23). Furthermore, the mechanical stability of the fibres coated with 18 16

Amount (ng)

14 12 10 8 6 4 2 0

0

5

10

15

20

25

30

35

Time (min)

Figure 3.23 Comparative extraction time profiles (n 5 3 fibres) for verapamil using CW/TPR commercial fibre and CW/TPR commercial fibre rendered biocompatible with a thin film of PAN.

84

Handbook of Solid Phase Microextraction

PAN was significantly improved, as indicated by the improved reusability (durability increased from B20 reuses to 50 reuses for CW/TPR fibres coated with PAN). A different strategy consists of incorporating high-capacity extraction phases into the bulk of a biocompatible polymer such as PAN or PEG.51,86,87 Useful extraction phases may consist of C18-silica, CN-silica, HS-F5-silica, RP-amidesilica or any other material used as stationary phase in HPLC or as sorbent in SPE. Although PDMS has been used in medical devices such as implants, catheters, pacemaker encapsulants, ocular lenses and, more recently, microfluidic chips due to its useful properties such as flexibility and low toxicity, the use of PDMS in this strategy is not recommended. The main reason for this is that PDMS is a relatively low biocompatibility material because of possible serious surface instability characterised by hydrophobicity recovery even when the surface is initially made hydrophilic.88 Better biocompatible polymers to use in this application are PEG and PAN. For example, the combination of C18-bonded silica/PEG coating was recently evaluated in vivo for monitoring diazepam and its two main metabolites in beagles.51 As probes were prepared manually, the between-probe variability ranged from 15% to 25% standard deviation. Further improvements in coating technique are expected to reduce this variability. In another study, the combination of PAN and C18-bonded silica was used to prepare biocompatible SPME coatings.86 PAN was selected in this study because of its excellent biocompatibility89 and better elasticity and mechanical stability in comparison to PEG.

3.4.3.2 Restricted Access Materials RAMs, such as alkyl-diol-silica (ADS) and ion exchange diol silica (XDS), constitute a class of promising biocompatible sample preparation materials. These materials consist of silica particles with a diameter of 5, 10 or 25 µm and with pores of about 3 nm in radius; the small pores yield a molecular mass cut-off of B15 kDa that allows direct fractionation of a sample into the protein matrix and the analyte fraction. In addition to a defined pore size, the specific feature of diol silica particles is the topochemically bifunctional surface of the particles: the outer particle surface is modified with hydrophilic diol groups, whereas the inner pore surface is covered with hydrophobic alkyl chains and/or ion exchange groups. RAMs were successfully used as the extraction phase in SPME for the determination of angiotensin-I in whole blood90 and benzodiazepines in urine.60

3.5

Interfaces to Analytical Instrumentation

Because of its solvent-free nature and the small size of the fibre or capillary, SPME can be interfaced conveniently to analytical instruments of various types. The sensitivity of determinations using the SPME technique is very high, facilitating trace analysis. Although the entire complement of analytes is not extracted from the sample (non-exhaustive extraction) in most cases, all material that is extracted is transferred to the analytical instrument, resulting in good performance.

Development of SPME Devices and Coatings

85

Also, the solvent-free process results in narrow bands reaching the instrument, giving taller, narrower peaks and better quantification.

3.5.1

SPME GC Interface

The analytical instrument used most frequently with SPME is the gas chromatograph. Only extracted analytes are introduced into the instrument because the extracting phase is non-volatile. Thus, there is no need for complex injectors designed to deal with large amounts of solvent vapour, and injector design can be simplified for use with SPME. Standard GC injectors, such as split/splitless, can be applied to SPME as long as a narrow insert with an i.d. close to the o.d. of the needle is used. The narrow inserts are required to increase the linear flow around the fibre, resulting in efficient removal of desorbed analytes. The split should be turned off during SPME injection. Under these conditions, the desorption of analytes from the fibre is very rapid, not only because the coatings are thin but because the high injector temperatures produce a dramatic decrease in the coating/gas distribution constant and an increase in the diffusion coefficients. In many cases, the speed of desorption is limited by the time required to introduce the fibre into the heated zone. One way to obtain sharper injection zones and faster separation times is to use rapid injection autosampling devices. An alternative solution is to use a dedicated injector, which should be cold during needle introduction but which heats up very rapidly after exposure of the fibre to the carrier gas stream. A schematic diagram of such an injector is presented in Figure 3.24.91 During desorption, the fibre is located inside the heated part of the fused silica capillary, its end being close to the bottom of the heated zone. The distance between the fibre and the capillary wall is B0.15 mm. A close match between the inner diameter of the capillary and the outer diameter of the fibre assures effective heat transfer from the heater to the fibre and a high linear flow rate of the carrier gas along the fibre. The injector is rapidly heated via a capacitive discharge. Heating rates of 1,000 C/s have been determined experimentally. The injector just described has achieved separation of benzene, toluene, ethylbenzene and xylenes (BTEX) in 8.2 s (Figure 3.25).92 Separation of the 28 volatile organic compounds listed in U.S. EPA method 624 has been accomplished in 150 s with a reproducibility of better than 5% standard deviation for most analytes (Figure 3.26).92 The fibre can also be designed to contain the heating element, as shown in Figure 3.27.92 In this case, no injector is necessary. The modified fibre can be introduced directly into the front of the column, and analytes can be desorbed rapidly by heating with a capacitive discharge current after the fibre has been exposed from the needle. Alternatively, flash desorption injectors can be designed by passing a current directly through the fibre. This is possible if the rod is made of conductive material, as it is in the case of the electrochemical SPME devices already mentioned. Figure 3.28 illustrates such an interface.93 When the electrical connection is made at the bottom of the interface, the fibre is rapidly heated by the discharging current.

86

Handbook of Solid Phase Microextraction

4

1 2

3

Septum purge

5 7

Carrier

8 6 9

DC supply

11

AC supply

12

13 10

Figure 3.24 Schematic diagram of the flash SPME injector: 1, injector body; 2, washer; 3, septum; 4, nut; 5, needle guide; 6, 0.53 mm i.d. fused silica capillary; 7, nut; 8, ferrule; 9, heater; 10, butt connector; 11, relay; 12, capacitor; 13, switch.

Figure 3.25 Rapid analysis of BTEX in water by SPME-flash injector-GC: 1, benzene; 2, toluene; 3, ethylbenzene; 4, m,p-xylenes; 5, o-xylene.

4

5 3

2 1

0

6

12 Time (s)

18

Development of SPME Devices and Coatings

87 20

25 26

23

15 12

11 13 6

8

5

9 1234

28 19

10

16 17 14

24

27

18

7

30

22 21

60

90

120

Time (s)

Figure 3.26 Fast separation of EPA-624 volatiles. Conditions: 0 30 /min 70 , 2.1 atm, capacitor voltage 24 V, MS detector, mass range 45 200. 1: chloromethane; 2: vinyl chloride; 3: bromomethane; 4: chloroethane; 5: trichlorofluoromethane; 6: 1,1-dichloroethene; 7: dichloromethane; 8: 1,2-dichloroethene; 9: 1,1-dichloroethane; 10: trichloromethane; 11: 1,1,1-trichloroethane; 12: tetrachloromethane; 13: benzene; 14: 1,2-dichloroethane; 15: trichloroethene; 16: 1,2-dichloropropane; 17: bromodichloromethane; 18: 2-chloroethyl vinyl ether; 19: cis-1,3-dichloropropene; 20: toluene; 21: trans-1,3-dichloropropene; 22: 1,1, 2-trichloroethane; 23: tetrachloroethylene; 24: dibromochloromethane; 25: chlorobenzene; 26: ethylbenzene; 27: tribromomethane; 28: 1,1,2,2-tetrachloroethane.

The other option is to use laser energy to desorb analytes from the surface of the fused silica optical fibre. Flash desorption injectors can be applied to interface SPME directly with a range of detection devices such as mass spectrometers and atomic emission devices. The sharp bands obtained during the desorption process result in very sensitive detection. For example, Figure 3.25 shows a sharp peak corresponding to the toluene band desorbed from the SPME fibre and directly detected by the mass spectrometer. The limit of detection is about two orders of magnitude lower than is obtainable with conventional GC MS techniques because of a much sharper band. To facilitate proper quantification, the extract needs to be very clean, which puts an additional demand on the coating selectivity. Tandem mass spectrometry (MS MS) may provide even lower detection limits. An important improvement of the SPME flash interface would be the addition of a cooling option, which would eliminate the broadening of the injection bands of volatile analytes because of desorption prior to the heating pulse. This would result in even faster desorption and separation.

3.5.2

Field Portable SPME Fast GC

SPME and high-speed gas-chromatography make a good combination for performing rapid, cost-effective investigations in the field, even for complex organic samples.

88

Handbook of Solid Phase Microextraction

Leads connected to capacitive discharge supply Holder body

Oblong opening in the plunger

Plunger with grooves

0.32 mm i.d. fused silica capillary

0.057 mm o.d. heating wire

Silicone rubber seal Nut

l

Needle

0.075 mm i.d., 0.15 mm o.d. fused silica capillary

Figure 3.27 Internally heated SPME device.

As discussed above, SPME is particularly suited for Fast GC, as it is solvent-free and the thin coatings can provide very fast desorption of analytes at high temperatures. Some instrumental modifications were performed recently in order to achieve successful fast separations.94 A portale system was optimised for SPME Fast GC field investigations and was commercialised by SRI Instruments (model 8610C, SRI Instruments, Torrance, CA). The instrument was tested in combination with a flame ionisation detector (FID), a photoionisation detector (PID) and a dry electrolytic conductivity detector (DELCD). A dedicated injector, presented in Figure 3.29, was mounted on the portable system in order to use SPME for high-speed separation. The injector guarantees very fast thermal desorption of the analytes from the SPME fibre.95 The injector for high-speed GC should produce as narrow an injection band as possible. Internal volumes of regular injector ports (e.g. split/splitless injector) are too large because they have been designed to accommodate large volumes of gaseous samples or vapours produced by solvent injection. Thermal focusing for separation improvement is not convenient for fast separations because temperature programming is impractical for high-speed GC. Hence, an injector port with a small internal volume was required for this application. Also, very fast thermal desorption from the SPME fibre was required to produce a narrow injection band and achieve effective separation. In the dedicated injector for SPME Fast GC, the injector port was kept cold during needle introduction and rapidly heated only

Development of SPME Devices and Coatings

89

2

3 DC supply 8

4

9

GC oven wall

AC supply 10

5

6 7 To ion-trap MS

Figure 3.28 Direct capacitive discharge desorption system: 1, SPME syringe; 2, electric connection I; 3, injector body; 4, steel wire; 5, gold coating; 6, electric connection II; 7, transfer line; 8, capacitor; 9, relay; 10, butt connector. 4

3

1

6

7

10

8

5

9

2

Figure 3.29 Dedicated injector for the field portable SPME Fast GC.

when the fibre was exposed to the carrier gas stream. The desorption area of the injector was heated by capacitive discharge that allowed heating rates as fast as 4,000 C/s, and very narrow injection bands were observed, as required by Fast GC.

3.5.3

SPME HPLC Interface

Research effort has also focused on designing interfaces for liquid-phase separation techniques to address the need for analysis of non-volatile and thermally labile analytes. The interface to HPLC can be a straightforward analogue of the traditional

90

Handbook of Solid Phase Microextraction

Detailed frame

Desorption chamber

Additional solvent

Metering pump

To waste 6-Port injection valve

SPME fibre

From pump To column

V/G ferrule, 0.4 mm i.d.

PEEK tubing 0.005” i.d.

Frame enlarged Interface SPME Device

Valco 3-way stainless steel tee

s.s. tubing 0.03” i.d.

Narrow bore HPLC column

HPLC pump

UV–VIS detector

Figure 3.30 Schematic diagram of the manual SPME HPLC interface.

loop injection system. A typical SPME HPLC interface consists of a custom-made desorption chamber and a six-port injection valve (Figure 3.30).94,95 The upper part of the polyether ether ketone (PEEK) tubing, fitted into a teeunion, is enlarged to fit the needle of the syringe. The internal tubing of the SPME device, which holds the fibre, can be sealed by the PEEK tubing and the tee-union tightly enough to withstand solvent pressures as high as 4,500 psi. The desorption chamber is placed in the position at which the injection loop normally resides on the injection valve. When the injection valve is in the ‘load’ position, it allows the fibre to be introduced into the desorption chamber under ambient pressure. It also allows for the introduction of a desorption solvent if it is different from the mobile phase. The valve is then switched to ‘inject’ to transfer the desorbed analytes to the column. A heater can be installed in the device to facilitate the desorption process. The interface performs well, and its desorption volume is similar to the volume of the typical injection loop. The use of small-volume desorption chambers results in very efficient supercritical fluid chromatography separation (Figure 3.31) with very narrow bore capillary columns.96 In addition to manual SPME interface described in this section, SPME can be coupled to HPLC using more automated methods, and these approaches are described in detail in Chapter 5.

3.5.4

SPME MALDI Interface

SPME can also be directly coupled to mass spectrometers. For example, SPME was recently coupled to a matrix-assisted laser desorption/ionisation (MALDI) for

Development of SPME Devices and Coatings

91

2 Phenanthrene

Fluoranthene

4 Acenaphthene 3 Benz[a]anthracene 5 Naphthalene 1

0

10

20

30

Time (min)

Figure 3.31 SFE separation of PAHs on a 50-µm i.d. capillary using the SPME device as a sample introduction technique.

the detection of large biomolecules.97 The tip of an optical fibre was silanised for extraction of analytes of interest from the sample. This treated optical fibre served as the sample extraction surface, the support for the sample plus matrix and the optical pipe to transfer laser energy from the laser to the sample. Both an ion mobility spectrometer and a quadrupole/time-of-flight (QqTOF) mass spectrometer were used for the detection of the SPME/MALDI signal. The QqTOF mass spectrum obtained for a solution of 5 pmol/µL of each peptide using the system from Figure 3.32 is shown in Figure 3.33. In these experiments, the sample tested was a mixture of enkephalin and substance P. Alpha-cyano-4-hydroxy cinnaminic acid was used as the matrix. Peaks at 726.4 and 1347.7 (shown in Figure 3.33) represent the protonated MH1 ions of the enkephalin and substance P, respectively. Increased selectivity and performance of SPME MALDI can be expected if general sorbent coating is substituted with a more specific affinity-based coating (described in Chapter 3, in Section 3.4.2). The combination of SPME/MALDI with a QqTOF system offers simple sample handling paired with the specificity and sensitivity of high-performance mass spectrometry. Even more important, it extends the usefulness of the SPME method to polar, high-molecular-weight biopolymers. The application of this technique holds promise, especially in biochemical analysis, pharmaceutical research, clinical diagnostics and screening. More recently, solid-phase microextraction/atmospheric pressure matrix-assisted laser desorption ionisation (SPME/AP MALDI) source configuration for a hybrid quadrupole linear ion trap instrument was developed in collaboration with Sciex (Figure 3.34).98 The multiplexed SPME plate is capable of simultaneous extraction from the wells of a 96-well plate, eliminating the need for extensive sample

92

Handbook of Solid Phase Microextraction

7

2 4 1

6

5 3

Figure 3.32 Schematic diagram of SPME/MALDI QqTOF system: (1) laser source, (2) focusing lens, (3) photodiode, (4) fibre holder, (5) SPME/MALDI fibre, (6) QqTOF and (7) computer.

726.4 100

Relative intensity (%)

5 pmol/µL

50

1347.7

0 200

400

600

800 1000 1200 1400 1600 1800 2000 2200 2400 2600 2800 3000 3200

m/z (amu)

Figure 3.33 Mass spectrum of enkephalin (m/z 5 726.4) and substance P (m/z 5 1347.7) obtained by SPME/MALDI.

preparation. Subfemtomole sensitivity was demonstrated for peptide standards and protein digests with run-to-run reproducibility ranging from B13% to 31% standard deviation.

3.5.5

SPME Nanospray MS Interface

SPME was also successfully interfaced to mass spectrometry via electronanospray tips.99 For the in situ extraction of peptides from the tryptic digests, trypsin was immobilised both on steel wires and on the inside wall of a vial. The devices were incubated together with the RAM SPME devices and a protein (casein) solution. After the protein digestion, the resulting peptides were analysed by SPME nanospray MS.

Development of SPME Devices and Coatings

93 ~2.5 Torr QJET: 5 cm

Laser beam Curtain plate

28°

2 mm

7 mTorr

Heated chamber

Plume

Curtain gas 1.2 L/min

760 Torr Orifice plate

High-throughput target plate (16 SPME fibres)

Figure 3.34 Schematic of the SPME/AP MALDI configuration used in these experiments. The target plate held an array of 16 SPME extraction fibres.

The vial approach provided the best results; up to eight peptides could be identified, which corresponds to a sequence coverage of 58%. The limit of detection of SPME nanospray MS for the extraction of peptides from an aqueous solution was about 50 fmol/mL. The results demonstrate that the direct coupling of SPME to nanospray can reduce analysis time and is an attractive alternative to conventional approaches like Zip-Tip purification.

3.5.6

Other Interfaces

Smaller injection volumes for applications of SPME to micro-HPLC and capillary electrophoresis can be attained by modifying microinjector designs; the sliding injector developed for capillary isotachophoresis is one example.100 An alternative approach is to design an appropriate sample introduction system based on guides, to introduce the fibre with the extracted components directly into the capillary (Figure 3.35). An on-column interface made of a Teflon block enables the direct insertion of an SPME fibre into the inlet end of a separation capillary.101 Very efficient separation was achieved using this method (Figure 3.36). Another approach combining the SPME and capillary electrophoresis (CE) techniques in protein analysis has recently been described and successfully demonstrated.102 Analytes desorbed from an SPME fibre were transferred by electrophoretic migration into a short piece of microdialysis hollow fibre located at the inlet of a CE system, in order to trap analytes with molecular weights greater than the molecular weight cut-off of the microdialysis material by dialysis. Then, an electric field with a different electrode polarity was applied, and the analytes trapped in the microdialysis hollow fibre migrated into the separation capillary.

94

Handbook of Solid Phase Microextraction

(A)

Capillary electropherograph Microscope SPME–CE coupler Recorder

Separation capillary

Fibre assembly +HV

(B)

Teflon rod

Polymer-coated silica fibre (~40 µm diameter)

Conical guide tube To detector Separation capillary (75 µm i.d. × 360 µm o.d.) SPME fibre assembly Buffer reservoir Pt electrode

Figure 3.35 Schematic of (A) the SPME CE system and (B) the interface.

Figure 3.36 Electropherogram of 10 phenolic compounds obtained by SPME CE using (A) a 40-µm o.d. PA-coated silica fibre and (B) a 40-µm o.d. bare silica fibre.101 Peak identities: 1: 2, 4-dimethylphenol; 2: phenol; 3: 4-chloro-3-methylphenol; 4: pentachlorophenol; 5: 2,4, 6-trichlorophenol; 6: 2-methyl-4, 6-dinitrophenol; 7: 2, 4-dichlorophenol; 8: 2-chlorophenol; 9: 4-nitrophenol; 10: 2-nitrophenol.

3

7 0.001 AU

4

5

1

A

6

2

8

9 10

B 0

5

10

15

20

Time (min)

SPME can be directly combined with optical detection based on reflectometric interference spectrometry.103 105 A light beam passing through an optically transparent fibre coated with transparent sorbing material interacts with absorbed substances through internal reflection. Therefore, if any of the extracted analytes strongly absorb the transmitted light, there is a loss in intensity that can be detected with a simple optical sensor.

Development of SPME Devices and Coatings

95

These devices demonstrate poor sensitivity, primarily because it is difficult to find light wavelengths that are specifically adsorbed by the analytes rather than by the coating or interferences. In an alternative design, the light can be passed directly through the absorbing polymer, which is then cooled to facilitate high sensitivity of determination.106 Fluorescence can be used to detect analytes in the coating. The selectivity of the extraction process and spectroscopy can be combined with selectivity of the electrochemical process, resulting in a spectroelectrochemical sensor (WR Heineman, University of Cincinnati, USA, personal communication).

References 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27.

J Pawliszyn & S Liu, Anal Chem 59 (1987) 1475 RG Belardi & J Pawliszyn, Water Qual Res J Can 24 (1989) 179 CL Arthur & J Pawliszyn, Anal Chem 62 (1990) 2145 J Pawliszyn, Method & Device for Solid Phase Microextraction & Desorption, PCT, International Patent Publication Number WO 91/15745 and national counterparts M McComb, E Giller & HD Gesser, 78th Canadian Society for Chemistry Conference & Exhibition, University of Guelph, Guelph, ON, 1995, Abs 528 M Chai & J Pawliszyn, Environ Sci Technol 29 (1995) 693 E Baltussen, F David, P Sandra, H-G Janssen & C Cramers, J High Resolut Chromatogr 21 (1998) 332 MA Nickerson, Sample Screening & Preparation within a Collection Vessel, US patent number 5,827,944 E Baltussen, P Sandra, F David, H-G Janssen & C Cramers, Anal Chem 71 (1999) 5213 I Bruheim, X Liu & J Pawliszyn, Anal Chem 75 (2003) 1002 W Zhao, G Ouyang, M Alaee & J Pawliszyn, J Chromatogr A 1124 (2006) 112 P Popp, M Moeder, ExTech’2001, Barcelona, Spain, September, 2001 C Grote & J Pawliszyn, Anal Chem 69 (1997) 587 H Lord, W Yu, A Segal & J Pawliszyn, Anal Chem 74 (2002) 5650 W Ma, X Liu & J Pawliszyn, Anal Bioanal Chem 385 (2006) 1398 J Pawliszyn, Solid Phase Microextraction: Theory & Practice (1997) Wiley: New York, NY F Augusto, J Koziel & J Pawliszyn, Anal Chem 73 (2001) 481 G Ouyang & J Pawliszyn, Anal Bioanal Chem 386 (2006) 1059 G Ouyang & J Pawliszyn, Trends Anal Chem 25 (2006) 692 Z Qin, L Bragg, G Ouyang & J Pawliszyn, J Chromatogr A 1196 1197 (2008) 89 Z Zhang & J Pawliszyn, Anal Chem 67 (1995) 34 Y Chen & J Pawliszyn, Anal Chem 78 (2006) 5222 G Ouyang & J Pawliszyn, Anal Bioanal Chem 386 (2006) 1059 L Muller, Applications of Solid Phase Microextraction, in J Pawliszyn, Ed, RSC Chromatography Monographs (1999) The Royal Society of Chemistry: Cambridge A-P Wang, F Fang & J Pawliszyn, J Chromatogr A 1072 (2005) 127 Y Gong, I-Y Eom, D-W Lou, D Hein & J Pawliszyn, Anal Chem 80 (2008) 7275 V Niri, I-Y Eom & J Pawliszyn, J Sep Sci 32 (2009) 1075

96

28. 29. 30. 31. 32. 33. 34. 35. 36. 37. 38. 39. 40. 41. 42. 43.

44. 45. 46. 47. 48. 49. 50. 51. 52. 53. 54. 55. 56. 57. 58. 59. 60. 61. 62. 63. 64. 65. 66.

Handbook of Solid Phase Microextraction

PA Martos & J Pawliszyn, Anal Chem 71 (1999) 1513 J Koziel, J Noah & J Pawliszyn, Environ Sci Technol 35 (2001) 1481 A Khaled & J Pawliszyn, J Chromatogr A 892 (2000) 455 J Koziel, M Jia, A Khaled, J Noah & J Pawliszyn, Anal Chim Acta 400 (1999) 153 Y Chen & J Pawliszyn, Anal Chem 75 (2003) 2004 Y Chen & J Pawliszyn, Anal Chem 76 (2004) 6823 G Ouyang, W Zhao, M Alaee & J Pawliszyn, J Chromatogr A 1138 (2007) 42 G Ouyang, W Zhao, L Bragg, Z Qin, M Alaee & J Pawliszyn, Environ Sci Technol 41 (2007) 4026 I Said, C Gaertner, M Renou & C Rivault, J Insect Physiol 51 (2005) 1384 J Tentschert, HJ Bestmann & J Heinze, Chemoecology 12 (2002) 15 J Tentschert, K Kolmer, B Holldobler, HJ Bestmann, JHC Delabie & J Heinze, Naturwissenschaften 88 (2001) 175 C Peeters, T Monnin & C Malosse, Proc R Soc Lond B Biol Sci 266 (1999) 1323 T Monnin, C Malosse & C Peeters, J Chem Ecol 24 (1998) 473 DC Gilley, G De Grandi-Hoffman & JE Hooper, J Insect Physiol 52 (2006) 520 D Djozan, T Baheri, R Farshbaf & S Azhari, Anal Chim Acta 554 (2005) 197 O Anderbrant, F Oestrand, G Bergstroem, A-B Wassgren, M-A Auger-Rozenberg, C Geri, E Hedenstroem, H-E Hoegberg, A Herz & W Heitland, Chemoecology 15 (2005) 147 D Rochat, P Ramirez-Lucas, C Malosse, R Aldana, T Kakul & JP Morin, J Chromatogr A 885 (2000) 433 X Chen, K Nakamuta, T Nakanishi, T Nakashima, M Tokoro, F Mochizuki & T Fukumoto, J Chem Ecol 32 (2006) 669 E Pionnier, C Chabanet, L Mioche, J-L Le Quere & C Salles, J Agric Food Chem 52 (2004) 557 E Pionnier, E Semon, C Chabanet & C Salles, Sci Aliment 25 (2005) 193 Z-M Zhang, J-J Cai, G-H Ruan & G-K Li, J Chromatogr B 822 (2005) 244 FM Musteata, ML Musteata & J Pawliszyn, Clin Chem 52 (2006) 708 FM Musteata & J Pawliszyn, J Biochem Biophys Methods 70 (2007) 181 A Es-haghi, X Zhang, FM Musteata, H Bagheri & J Pawliszyn, Analyst 132 (2007) 672 X Zhang, A Es-haghi, FM Musteata, G Ouyang & J Pawliszyn, Anal Chem 79 (2007) 4507 Z Zhang, MJ Yang & J Pawliszyn, Anal Chem 66 (1994) 844A P Cheo, Fibre Optics (1985) Prentice-Hall: Englewood, NJ, pp. 88 JP Hutchinson, L Setkova & J Pawliszyn, J Chromatogr A 1149 (2007) 127 C Jia, Y Luo & J Pawliszyn, J Microcolumn Sep 10 (1998) 167 R Vatinno, D Vuckovic, CG Zambonin & J Pawliszyn, J Chromatogr A 1201 (2008) 215 Y Liu, Y Shen & ML Lee, Anal Chem 69 (1997) 190 Y Liu, ML Lee, KJ Hageman, Y Yang & SB Hawthorne, Anal Chem 69 (1997) 5001 WM Mullett & J Pawliszyn, Anal Chem 74 (2002) 1081 D Vuckovic, E Cudjoe, D Hein & J Pawliszyn, Anal Chem 80 (2008) 6870 GG Wallace, M Smyth & H Zhao, Trends Anal Chem 18 (1999) 245 A Michalska, A Ivaska & A. Lewenstam, Anal Chem 69 (1997) 4060 J Wu & J Pawliszyn, Anal Chim Acta 520 (2004) 257 J Wu & J Pawliszyn, J Chromatogr A 909 (2001) 37 J Wu, W Mullett & J Pawliszyn, Anal Chem 74 (2002) 4855

Development of SPME Devices and Coatings

97

67. JK Schubert, W Miekisch, P Fuchs, N Scherzer, H Lord, J Pawliszyn & RG Mundkowski, Clin Chim Acta 386 (2007) 57 68. TP Gbatu, KL Sutton & JA Caruso, Anal Chim Acta 402 (1999) 67 69. SL Chong, D Wang, JD Hayes, BW Wilhite & A Malik, Anal Chem 69 (1997) 3889 70. A Fernandes de Oliveira, C Berto da Silveira, S Denofre de Campos, E Antonio de Capos & E Carasek, Chromatographia 61 (2005) 277 71. J Yu, C Wu & J Xing, J Chromatogr A 1036 (2004) 101 72. M Azenha, C Malheiro & AF Silva, J Chromatogr A 1069 (2005) 163 73. R Gomes da Costa Silva & F Augusto, J Chromatogr A 1072 (2005) 7 74. C Basheer, S Jegedesan, S Valiyaveettil & HK Lee, J Chromatogr A 1087 (2005) 252 75. H Bagheri, A Es-haghi & MR Rouini, J Chromatogr B 818 (2005) 147 76. V Pichon, J Chromatogr A 1152 (2007) 41 77. W Mullett, P Martin & J Pawliszyn, Anal Chem 73 (2001) 2383 78. EHM Koster, C Crescenzi, W den Hoedt, K Ensing & GJ de Jong, Anal Chem 73 (2001) 3140 79. E Turiel, JL Tadeo & A Martin-Esteban, Anal Chem 79 (2007) 3099 80. X Hu, J Pan, Y Hu, Y Huo & G Li, J Chromatogr A 1188 (2008) 97 81. X Hu, Y Hu & G Li, J Chromatogr A 1147 (2007) 1 82. H Lord, M Rajabi, S Safari & J Pawliszyn, J Pharm Biomed Anal 40 (2006) 769 83. H Lord, M Rajabi, S Safari & J Pawliszyn, J Pharm Biomed Anal 44 (2007) 506 84. TE Lipatova & YS Lipatov, Macromol Symp 152 (2000) 139 85. J Chanard, S Lavaud, C Randoux & R Philippe, Nephrol Dial Transplant 18 (2003) 252 86. ML Musteata, FM Musteata & J Pawliszyn, Anal Chem 79 (2007) 6903 87. D Vuckovic, B Shirey, Y Chen, L Sidisky, C Aurand, K Stenerson & J Pawliszyn, Anal Chim Acta 638 (2009) 175 88. H Makamba, Y-Y Hsieh, W-C Sung & S-H Chen, Anal Chem 77 (2005) 3971 89. MC Yang & WC Lin, J Polym Res 9 (2002) 201 90. FM Musteata, M Walles & J Pawliszyn, Anal Chim Acta 537 (2005) 231 91. T Gorecki & J Pawliszyn, J High Resolut Chromatogr 18 (1995) 161 92. T Gorecki & J Pawliszyn, Anal Chem 67 (1995) 3265 93. T Gorecki & J Pawliszyn, Field Anal Chem Technol 1 (1997) 277 94. J Chen & J Pawliszyn, Anal Chem 67 (1995) 2530 95. HL Lord, RP Grant, M Walles, B Incledon, B Fahie & JB Pawliszyn, Anal Chem 75 (2003) 5103 96. Y Hirata & J Pawliszyn, J Microcolumn Sep 6 (1994) 443 97. H Tong, N Sze, B Thomson, S Nacson & J Pawliszyn, Analyst 127 (2002) 1207 98. Y Wang, B Schneider, T Covey & J Pawliszyn, Anal Chem 77 (2005) 8095 99. M Walles, Y Gu, C Dartiguenave, FM Musteata, K Waldron, D Lubda & J Pawliszyn, J Chromatogr A 1067 (2005) 197 100. T McDonnell & J Pawliszyn, Anal Chem 63 (1991) 1884 101. C-W Whang & J Pawliszyn, Anal Commun 35 (1998) 353 102. Z Liu & J Pawliszyn, Analyst 131 (2006) 522 103. BL Wittkamp & DC Tilotta, Anal Chem 67 (1995) 600 104. GL Klunder & RE Russo, Anal Chem 49 (1995) 379 105. HM Yan, G Kraus & G Gauglitz, Anal Chim Acta 312 (1995) 1 106. J Pawliszyn, Device & Process for Increasing Analyte Concentration in a Sorbent, US Patent 5,496,741

4 SPME Commercial Devices and Fibre Coatings

Robert E. Shirey Supelco, Bellefonte, PA, USA

4.1

Introduction

There are several important criteria that are fundamental for the acceptance of solid-phase microextraction (SPME) as an extraction technique. These criteria include ease of use and solvent-free (or solventless) extraction capability. However, the performance of SPME is critically dependent on the availability and selection of appropriate coating. The SPME fibre coating is primarily responsible for the extraction of analytes. Because SPME is capable of extracting a wide range of analytes from volatile to non-volatile and from polar to non-polar, it is important to have fibre coatings that can extract this range of analytes. To accomplish this, different coating types have been developed. Different coating materials enable extraction of a variety of analytes with enhanced selectivity. This chapter will focus on the types and properties of the various commercially available SPME coatings and devices. It will also provide a guide on how to select the appropriate fibre for your application needs. For SPME to be widely accepted as a qualitative and quantitative extraction technique, the reproducibility between fibres and the durability of the coatings are very important. These two criteria allow the analyst to use one fibre for multiple extractions and to obtain similar results when multiple fibres are used. The repeatability of multiple extractions using a single fibre is greatly dependent upon the durability of the fibre coating. If the coating maintains its integrity, the repeated extractions of same samples should yield reproducible results. After many extractions, the fibre may slowly lose efficiency. When this occurs, the old fibre can be replaced with a new fibre of the same coating and thickness. If the preparation of the fibres is reproducible, the new fibre should produce similar results compared to the replaced fibre. This chapter will briefly describe how the fibres are manufactured and evaluated and how to select the appropriate conditions to extend the life of the fibre.

4.2

Description of SPME Fibre Assemblies and Holders

One of the important elements for making SPME attractive to users is the ease of use of the device. This was accomplished by designing an assembly that contains a Handbook of Solid Phase Microextraction. DOI: 10.1016/B978-0-12-416017-0.00004-8 © 2012 Elsevier Inc. All rights reserved.

100

Handbook of Solid Phase Microextraction

Figure 4.1 Schematic view of SPME manual fibre assembly.

Colour-coded screw hub Tensioning spring Sealing septum Ferrule

Septum-piercing needle

Fibre-attachment needle Coated SPME fused silica fibre

piercing needle and an inner needle or tubing that has a piece of coated fibre attached to it. Figure 4.1 shows a picture of a manual fibre assembly. The key to the fibre assembly is the sealing septum that seals the outer needle to keep it from leaking when inserted into a pressurised gas chromatograph (GC) injection port. The top of the plunger (tubing) is a coloured hub that indicates the type of coating on the fibre. The manual assembly contains a spring that helps to retract the fibre after exposure for extraction and desorption. The assembly used with autosamplers is identical to the manual style assembly, except that it does not contain a spring. The length of the coated fibre in the current SPME assembly is usually 1 cm, but the length can be as long as 2 cm. Lengths longer than 2 cm would not undergo efficient thermal desorption in a GC injector. The ‘hot zone’ (i.e. the point where the thermocouple is attached) in most injection ports is usually only 2 cm in length. The temperature decreases from the set temperature as the distance from the ‘hot zone’ increases, especially below the thermocouple because heat rises. The bottom tip of the 2-cm fibre may go below the positioning of the injection port in some GC instruments. A longer fibre would go even farther beyond this position. The assembly is attached to a holder that is shown in Figure 4.2. This figure shows both an external and internal view of the manual holder. On the left side, the fibre is in the exposed position, and on the right side, the fibre is retracted into the needle. The manual holder has a needle guide depth gauge that can be screwed up or down to determine how far the needle goes either in the injection port or into a vial. Also, it helps to support the needle and reduce breakage. The manual holder has a z-slot to lock the fibre in the exposed position. When the

SPME Commercial Devices and Fibre Coatings

101

(A)

(B)

External view

Internal view Plunger

Barrel

Plunger

z-slot Barrel Plunger retaining screw

Hub-viewing window

Adjustable needle guide/depth gauge

Retaining screw Slot

Colour-coded screw hub

Sealing septum

Fibre-attachment needle Retaining nut Needle ferrule

Fibre-attachment needle

Septum-piercing needle

Septum-piercing needle

Coated SPME fused silica fibre

Coated SPME fused silica fibre

Figure 4.2 Schematic view of SPME manual fibre assembly holder (external and internal view).

plunger is unlocked, the fibre will retract into the needle if it is a manual assembly. The holder designed for autosamplers does not contain the needle guide depth gauge or the z-slot. It is simply a straight slot. Fibre depth and plunger movement can be controlled by the autosampler. Please note that an autosampler assembly (without the spring) can be used in both the manual holder and autosampler holder, but the manual style spring can be used only in the manual holder. The z-slot is needed to expose the fibre with an assembly containing a spring. Assemblies with a 2-cm-length fibre do not contain a spring because they must go below the z-slot to fully expose the fibre. To adjust for the increased fibre length, the plunger tubing was shortened by 1 cm. Figure 4.3 shows a schematic of a typical SPME extraction (direct immersion) followed by thermal desorption in a hot GC injection port. Alternatively, for some fibres and applications, solvent desorption can be used instead of thermal desorption. For headspace extraction, the fibre would be exposed to the air (headspace) above the sample. In most cases, the vial would be heated to drive volatile analytes into the headspace. Minimising the headspace volume can enhance sensitivity.

4.2.1

Assemblies with 24- Versus 23-Gauge Needles

The original SPME fibre assemblies were introduced with 24-gauge needles that have an outside diameter of 565 μm. This gauge was selected because it is the smallest diameter that will enable the 100-μm-thick fibre coatings to be retracted

102

Handbook of Solid Phase Microextraction

Pierce sample septum

Retract fibre/remove Expose fibre/extract

Retract fibre/remove Pierce GC inlet septum

Expose fibre/desorb

Figure 4.3 Steps of typical SPME extraction (direct immersion) and thermal desorption in GC.

into the needle. A small diameter is desired to minimise septa coring. The 24-gauge needle is suitable for manual use, but it can be damaged more easily than 23-gauge needle when using an autosampler. Fibre assemblies with 23-gauge needles were introduced for use with Merlin Microsealt inlet seals. The Microseal ‘septum’ contains a duckbill opening that seals around the needle. For a proper seal, a 23-gauge needle is required. The 23-gauge needles have an outer diameter of 646 μm. Because of the larger diameter, the wall of the needles is thicker and is more durable compared to 24-gauge needles, which makes them more compatible with autosamplers. It is highly recommended that assemblies with 23-gauge needles be used with the CTC Analytics style autosamplers. Sealing sample vials with caps that contain thin septa can increase the life of the assembly further. The use of screw-cap vials containing 1.5-mm-thick septum will greatly increase the life of an assembly compared to 3-mm-thick vial septa, which are closed by crimping. The thinner septa are easier to pierce, which can minimise needle bending. The seals with the screw-cap vials have excellent sealing integrity as determined by Baltensperger and Shirey.1

4.2.2

GC Inlet Seals

The integrity of the inlet seal is critical for proper function of SPME apparatus. It is necessary to have a tight seal and to minimise coring. When septum fragments get into the heated zone of the inlet liner, extraneous siloxane peaks will appear on chromatograms. When a needle pierces an inlet, some oxygen always enters the

SPME Commercial Devices and Fibre Coatings

103

system. Septum particles exposed to oxygen in a hot inlet can form cyclic siloxanes that appear as chromatographic peaks. Simply ramping a GC oven without piercing the septum does not typically result in the appearance of these peaks. However, when a fibre is inserted, this allows oxygen into the system and the peaks appear. In most cases, the user will blame the fibre coating for the source of the siloxane peaks, but in reality, it is often the GC inlet septa. It is common for SPME fibres to introduce moisture into a GC inlet. The amount of water introduced can cause the appearance of extraneous peaks. Again, the fibre is frequently targeted as the source of the peaks, but this may not always be the case. Siloxane peaks have appeared when using fibres with coatings that do not contain any siloxanes. Also, the peaks appear when a small amount of water is injected into the GC inlet with a standard liquid syringe. Liners and inlets are coated with siloxane deactivants that can react with water and heat. Fibres do produce some extraneous peaks, but generally these peaks do not have a high intensity. In many cases, the peaks will reduce in size as the fibre is used. There are good alternatives to silicone septa for sealing GC inlets, including Merlin Microseals and Gerstel CIS inlet systems. Both have been evaluated and approved by the manufacturers for SPME use. These inlets do not core or produce extraneous peaks. The main limitation is that they are not available for all types of GC instruments. The other alternative is to use moulded silicone septa instead of cut septa. The moulded septa are made with a less rigid (lower durometer) silicone. It was determined that the moulded septa with a guide hole or injection hole resist coring. When the guide hole is present, the septum tends to tear and not core. Without the guide hole, coring will occur but will be less compared to a cut septum. Also, it is better to use moulded septum which does not contain iron oxide. These septa can be readily identified because of the red or orange colour from the iron oxide. Iron oxide is added to increase temperature stability, but it also increases the durometer of the septa, which makes it more susceptible to coring. Some vendors supply moulded septa with a guide hole or injection hole without iron oxide that are thermally stable to 350
C, which is adequate for SPME applications.

4.3 4.3.1

Description of Fibre Cores, Coatings and the Coating Process Types and Selection of Fibre Cores

4.3.1.1 Fused Silica Core The first SPME fibre cores investigated were fused silica or quartz. This highly inert glass has been used for the manufacturing of capillary columns, and it is the core material for optical fibre. Because coated optical fibre was readily available, it was the logical choice to use in SPME. The disadvantage of fused silica is that it can break, especially if the coating has been stripped off the fibre. With the

104

Handbook of Solid Phase Microextraction

exception of the 7-μm polydimethylsiloxane (PDMS) fibre, the phase coating must be stripped off the portion of the fused silica fibre that is inserted into the assembly plunger needle. The point where the fibre coating contacts the fibre plunger is a susceptible breaking point.

4.3.1.2 Stableflex Fibre Core The Stableflex fibre core was introduced in 1998. The Stableflex core consists of an 80-μm fused silica core coated with a 20-μm thermally stable ‘plastic-like’ polymer. This coating provides a protective barrier that reduces the chance of breaking the fibre. The other reason for the introduction of this fibre was to improve the coating process that will be discussed later in the chapter. Only adsorbent-type coatings are applied to the Stableflex core. Any extraction properties that the Stableflex coating has are minute with respect to the strength of the adsorbent coatings. Some of the adsorbent coatings bind more tightly to the Stableflex core compared to fused silica. The Stableflex core has a thermal limitation of 320
C, and it can produce some extraneous peaks depending upon the extraction conditions.

4.3.1.3 Metal Fibre Core Because of the fragility of fused silica fibres and the thermal limitations of the Stableflex core, there was a need for a durable, thermally stable core material. The metal used as fibre core is a unique, non-ferrous alloy with shape memory properties. The shape memory process allows the fibre to be bent and flexed while retaining its straight properties or memory. This allows the fibre to be coiled and coated using optical fibre coating towers. The fibre is thermally stable up to 450
C; however, the fibre can lose the shape memory property if it is retained above 300
C for more than 16 continuous hours. As long as the heating process is not continuous, the fibre will retain its properties. In addition to improved durability and thermal stability, the metal fibre is highly inert and does not contain any ferrous material. It is known that amines are highly reactive with metal surfaces. Table 4.1 shows a comparison of the responses between the fibre cores for the extraction of amines. These responses are relative to the internal standard n-propanol. The concentration of all of the analytes in the sample was 1 μg/mL. The results indicate that there is virtually no difference between the cores. This shows that the metal fibre core is very inert and suitable as a replacement for fused silica fibre cores. Table 4.1 Comparison of Fibre Cores for the Extraction of Amines Fibre

Methylamine

Dimethylamine

Diethylamine

Metal Fused Stableflex

1.14 1.02 1.04

4.05 4.21 5.43

38.1 38.7 37.5

SPME Commercial Devices and Fibre Coatings

4.3.2

105

Types of Commercially Available SPME Coatings

SPME coatings can be classified primarily into four categories: by the type of coating, by the coating thickness, by polarity and by whether the coating is an absorbent or an adsorbent. A listing of commercially available fibres is shown in Table 4.2. The chemical structures of polymers commonly used as SPME coatings are shown in Figure 4.4. The type of phase applied determines the polarity of the coating. Polarity can provide selectivity by enhancing the affinity of the coating for polar analytes compared to a non-polar fibre coating. Essentially, all the SPME fibres are bipolar to some degree because they will extract both polar and non-polar analytes, but it is the overall properties of the coating that decide polarity. The PDMS surface does make the fibre less polar, but if polar analytes contact the pores, polar analytes can be extracted. Because porous adsorbents extract primarily by the size of the analyte, both polar and non-polar analytes can be extracted. This is why such fibres are classified primarily as bipolar in Table 4.2. Figure 4.5 illustrates how to select appropriate fibre coating depending on the analyte polarity and volatility. Table 4.2 Types of Commercially Available SPME Fibre Coatings Type of Coating

Extraction Mechanism

Polarity

7 μm PDMS 30 μm PDMS 100 μm PDMS 85 μm PA 60 μm PEG (Carbowax) 15 μm Carbopack ZPDMS 65 μm PDMSDVB, 55 μm/30 μm DVB/CarboxenPDMS 85 μm CarboxenPDMS

Absorbent Absorbent Absorbent Absorbent Absorbent Adsorbent Adsorbent Adsorbent Adsorbent

Non-polar Non-polar Non-polar Polar Polar Bipolar Bipolar Bipolar Bipolar

(A)

(B)

CH3 Si

O

Figure 4.4 Chemical OCH3 structures of common

polymers used as SPME coatings.

O

CH3

n

n Polyacrylate

PDMS

(C)

(D) CH2

CH n

HO

CH2 CH2

O

H n

Divinylbenzene

PEG (Polyethyleneglycol)

106

Handbook of Solid Phase Microextraction

Analyte properties

Low

Polarity

High

Carboxe

n

High

PDM

PD

MS

S–D VB

30

µm

G)

MS

E (P

PA

Volatility

PD

x wa rbo

µm

Ca

100

Low

PD

MS

7µ

m

Figure 4.5 Coating selection guide.

The mechanism for extraction is determined by whether a coating is an absorbent type or an adsorbent type (Section 2.7). The thickness of the coating determines the analyte capacity of the fibre. The thickness of the coating also determines the length of the extraction time required to reach equilibrium. It takes longer to reach equilibrium with a thicker coating compared to a thin coating. Volatile analytes require a thick coating to retain them, whereas thin coatings are preferred for the extraction of high-molecular-weight analytes.

4.3.2.1 Determination of Phase Volume For some applications, it is important to determine the distribution constant. The volume of phase on the fibre must be known to determine the distribution constant. To calculate the phase volume of a fibre, one must know the radius of the total fibre (the diameter of fibre core and coating divided by 2) and the radius of the fibre core. The following equation can be applied: V 5 πh r 2 total fibre 2 πh r 2 fibre core

ð4:1Þ

where V is the volume, r is the radius and h is the length of the fibre phase coating. Because the various fibre cores have different diameters, there is some variation in the coating volume, even though the thickness may be identical. Table 4.3 is a phase volume chart for commercially available fibres. Calculations are based on a 1-cm-length fibre. For a 2-cm-length fibre, double the phase volume.

SPME Commercial Devices and Fibre Coatings

107

Table 4.3 Physical Properties and Phase Volume of SPME Fibres Fibre Coating

Core Type

Core Core Diameter Volume (mm) (mm3 or µL)

Total Fibre Diameter (mm)

Total Volume (mm3 or µL)

Phase Volume (mm3 or µL)

100 μm PDMS

Fused silica Metal Fused silica Metal Fused silica Metal Fused silica Metal Metal

0.110

0.095

0.300

0.707

0.612

0.130 0.110

0.133 0.095

0.305 0.170

0.730 0.227

0.598 0.132

0.130 0.110

0.133 0.095

0.185 0.125

0.269 0.123

0.136 0.028

0.130 0.110

0.133 0.095

0.144 0.285

0.163 0.638

0.030 0.543

0.130 0.130

0.133 0.133

0.250 0.160

0.491 0.201

0.358 0.068

100 μm PDMS 30 μm PDMS 30 μm PDMS 7 μm PDMS 7 μm PDMS 85 μm PA 60 μm PEG 15 μm Carbopack Z/PDMS 65 μm PDMS/DVB 65 μm PDMS/DVB 65 μm PDMS/DVB 75 μm CarboxenPDMS 85 μm CarboxenPDMS 85 μm CarboxenPDMS 50/30 μm DVB/Carboxen Carboxen layer DVB layer 50/30 μm DVB/Carboxen Carboxen layer DVB layer 60 μm PDMSDVB HPLC

Fused silica Stableflex Metal Fused silica Stableflex

0.120

0.113

0.260

0.531

0.418

0.130 0.130 0.120

0.133 0.133 0.113

0.270 0.270 0.280

0.572 0.572 0.615

0.440 0.440 0.502

0.130

0.133

0.290

0.660

0.528

Metal

0.130

0.133

0.290

0.660

0.528

0.130 0.190

0.133 0.283

0.190 0.290

0.283 0.660

0.151 0.377

0.130 0.190 0.160

0.133 0.283 0.201

0.190 0.290 0.290

0.283 0.660 0.660

0.151 0.377 0.459

Stableflex

Metal

Special

4.3.2.2 Fibre Coating Temperature and pH Ranges The thermal limitations of the various fibre coatings and the acceptable pH range for the fibres are given in Table 4.4. Details on solvent compatibility are given in the detailed description of the various fibre coatings.

108

Handbook of Solid Phase Microextraction

Table 4.4 Temperature and pH Ranges of SPME Fibre Coatings Fibre Coating

Maximum Temperature (
C)

Recommended Operating Temperature (
C)

Recommended pH Range

100 μm PDMS 30 μm PDMS 7 μm PDMS Polyacrylate PEG Carbopack Z/PDMS PDMS/DVB CarboxenPDMS DVB/Carboxen/ PDMS

300 300 320 320 250 340 270 320 270

200300 200300 200320 220320 200240 200340 200270 250320 230270

210 211 211 211 29 211 211 211 211

4.3.3

Absorbent Coatings and Coating Process

The first SPME fibre coatings investigated and developed were coated with absorbenttype phases. Absorbent fibre coatings are composed primarily of ‘liquid-like’ polymers. The absorbent can be a gum or viscous oil that contains cross-linking agents. Heatand/or ultraviolet (UV)-free radical initiation can be used to cross link and bond the polymer to the fibre core. This produces a stabilised, high-molecular-weight polymer coating with fluid properties. The polymer can be applied in various thicknesses over the fibre. With absorbent-type fibre coatings, the analytes migrate in and out of the phase coating. The retention of the analytes is based primarily on the thickness of the fibre coating. The analytes are attracted to the phase coating primarily by polarity. As the analyte enters the coating, it can migrate deeper into the coating until it reaches the core. Small analytes will move more rapidly through the coating than larger analytes. Thus, it is difficult for absorbent phases to retain small analytes unless a thick coating is used.

4.3.3.1 Coating of Fused Silica Absorbent Fibres The advantage of fused silica fibre core is that it can be manufactured at a desired thickness in long lengths using a fibre optic tower. To produce such long lengths of optical fibre, the towers can be several stories tall. A piece of quartz rod is heated above 900
C and pulled through a form with a fixed diameter. The coating is performed immediately after the fibre is drawn prior to coiling, otherwise the fibre would break. The fused silica core diameter used for SPME fibres is 110 μm. The major advantage of coating using an optical fibre tower is that long lengths of fibre can be precisely coated. The diameter of the fibre core and the diameter of the coated fibre can be tightly controlled. This results in a very consistent product. Typically for SPME use, the fibre is coated to a length of 1,000 m. Because only

SPME Commercial Devices and Fibre Coatings

109

1.52 cm length of fibre is required to produce one (1-cm coated) fibre assembly, 50,00075,000 assemblies can be produced from a single (1,000 m) spool. All PDMS coatings and polyacrylate (PA) coatings on fused silica are produced using the optical fibre coating towers at 1,000 m lengths.

4.3.3.2 Coating of Metal Absorbent Fibres The 0.005v (128 μm) diameter metal fibre can be coated with the same fibre optic tower used to produce fused silica fibre. Instead of drawing the quartz fibre, a 1,200-m spool of the metal fibre is passed through the optical fibre coating tower. The same controls and guidelines used to coat the fused silica fibre are applied to the metal fibre. The consistent coating thickness helps to produce reproducible SPME fibre assemblies. All the absorbent PDMS coatings have been applied to the metal fibre as well as the polyethylene glycol (PEG) coating.

4.3.3.3 PDMS Coating The most common non-polar phase is PDMS. PDMS is commonly used in the manufacturing of GC columns. This polymer can be highly cross-linked, and it is thermally stable. PDMS is a good extraction material, though it does not have a high affinity for polar analytes. Because PDMS is easy to apply and is readily used in the fibre optic industry, coatings of various thicknesses can be manufactured reproducibly. Currently, there are three PDMS fibre coatings commercially available. The thickness is limited to about 100 μm due to the piercing needle opening. A significantly thicker coating would not retract into the needle without being damaged. All the PDMS coatings are cross-linked and are thermally stable. By the addition of coupling linkages in the coating polymer, the polymer covalently cross links with itself to form a higher molecular weight polymer as the fibre coating is thermally cured. The thicker 100- and 30-μm coatings have a lower maximum temperature limit of 300
C compared to the more highly cross-linked 7-μm PDMS fibre coating with a maximum temperature limit of 320
C (see Table 4.4). PDMS coatings are stable in water with a pH range from 2 to 11. Phosphate buffers are very good for obtaining this range. It is recommended to avoid the use of highly caustic bases such as KOH and NaOH, especially if direct-immersion SPME is used. PDMS coatings can swell in some organic solvents or if samples contain high concentration levels of these solvents. Chlorinated solvents, hydrocarbons and diethyl ether can swell the fibre coating. If the swelling is severe, the coating will be stripped off the fibre core by the outer needle when the fibre is retracted. In some cases, only a portion will be damaged. Swelling can also elongate the fibre coating. When this occurs, the fibre may retract cleanly into the needle, but a portion of the phase will stick out of the needle opening. Often this portion of the phase will get caught in the vial septa when the fibre is withdrawn from the vial. The phase coating will be stripped off the fibre core and will be hanging from the bottom of the vial septa.

110

Handbook of Solid Phase Microextraction

4.3.3.4 PA Coating PA is a moderately polar coating. This fibre is fairly universal for extraction of a wide range of analytes, both polar and non-polar. This polymer is more rigid than PDMS or PEG, so the migration in and out of this coating is slightly slower. The phase has a high affinity for aromatic compounds and oxygenated analytes. PA coatings have moderate thermal stability. The coating will darken over time and the coating thickness will decrease slightly as the fibre is used. Limiting the time that the fibre is desorbed above 280
C can minimise both of these problems. Also, it is very critical to have a high-purity carrier gas with a good purifier to remove trace amounts of oxygen. Good integrity of the inlet seals will increase the life of the fibre by reducing the amount of air that enters the inlet. The PA phase coating is stable over a pH range of 211. However, because the phase is an ester, it is recommended to limit the number of extractions in solutions with a pH above 9. Similar to PDMS, it is recommended not to expose this coating to sodium or potassium hydroxide solutions without neutralisation. PA fibres will swell slightly in water, but they are resistant to hydrocarbon solutions. A combination of a water-soluble solvent, such as acetonitrile, acetone or methanol, with water will cause the highest degree of swelling. It is good to test the degree of swelling by putting an exposed fibre in an uncapped vial containing the desired solvent/solvent mixture and see how much it swells. If swelling is severe or elongation occurs, simply remove the fibre from the vial and allow it to air-dry before retracting the fibre. The swelling will reduce as the solvent dissipates.

4.3.3.5 PEG Coating PEG, Carbowaxs, fibre coating is the most polar coating available on commercially produced SPME fibres. This phase tends to be more selective towards polar analytes. Also, it tends to extract fewer non-polar analytes relative to non-polar fibre coatings. The PEG coating is 60 μm thick and is coated on the metal fibre. Bonding of the phase to the metal fibre is better compared to the other cores. The coating tends to swell in water samples, but the swelling is reduced if the sample contains a high amount of salt. The reason for this is not known. The combination of some solvents with water can cause the most severe swelling. The PEG coating does not swell in hydrocarbon or aromatic solvents. The fibre has moderate thermal stability. It can be taken up to 250
C, but it is recommended that it not be used routinely above 240
C. The recommended pH range for this fibre is 29.

4.3.4

Adsorbent Coatings and Coating Process

Adsorbent coatings extract analytes by physical trapping. Instead of an analyte interacting with a liquid polymer, the interaction is with a solid particle. In adsorbent-type fibres, a solid material is suspended into a liquid polymer and coated on

SPME Commercial Devices and Fibre Coatings

111

a fibre. Usually the solid material is a porous polymer or a porous carbon or silica. With adsorbents, the analytes migrate into the pores of the adsorbent during the extraction process. The retention of the analyte depends upon the size of the analyte and the pore diameter. The surface of an adsorbent can interact with an analyte, such as ππ bonding, hydrogen bonding or van der Waals interactions. The ability of an adsorbent to retain analytes is dependent upon several factors. These factors include the total surface area, the amount of porosity and the size of the pores. Many adsorbents are characterised by the amount of surface area. This is suitable for non-porous materials, but it is only a partial measurement for determining the adsorbing capability of a porous adsorbent. In many cases, an adsorbent with a lower surface area may be much stronger than one with a higher surface area. The strength of an adsorbent is determined by the size of the analytes that it can retain. The strength is irreversibly proportional to the analyte size. Pores can be defined into three categories: macro, meso and micro. Macropores ˚. are primarily surface pores and have openings with diameters of .500 A ˚ Mesopores have openings in the range of 20500 A, while micropores are from ˚ . The amount of porosity is measured as the pore volume per gram of 2 to 20 A adsorbent in millilitres per gram. The average size of the micropore diameter is critical in determining the strength of the adsorbent. A pore can retain an analyte that is about half the size of the pore diameter. So ˚ can retain an analyte with an average molecular a pore with a diameter of 12 A ˚ diameter of 6 A. Therefore, both the degree of porosity and the average size of the pores are important in determining the extraction capability of an adsorbent. If a micropore volume is high and the average diameter is small, the adsorbent would have a very high amount of microporosity. Because the diameter determines volume, if the diameter is narrow, the length of the micropores would have to be greatly increased to obtain a large pore volume. Table 4.5 shows the physical properties of the two commonly used adsorbents with SPME fibres, a porous polymer called divinylbenzene (DVB) and a carbon molecular sieve called Carboxent 1006.

4.3.4.1 Preparation of Adsorbent Suspensions for Coating The typical particle size of the adsorbents used in SPME fibres is 35 μm. To prepare thicker coatings which have sufficient sample capacity, the particles must be Table 4.5 Physical Properties of Adsorbents Used in SPME Fibres Material

DVB Carboxen 1006

Surface Area (m2/g) 750 950

Porosity (mL/g) Macropore ˚ . 500 A

Mesopore ˚ 20500 A

Micropore Total ˚ 220 A

0.58 0.23

0.85 0.26

0.11 0.29

1.54 0.78

Average Micropore ˚) Diameter (A 16 12

112

Handbook of Solid Phase Microextraction

PDMS

Carboxen particles (3–5 µm)

Fibre core

Analytes can migrate between layers – increased capacity

Figure 4.6 Drawing of CarboxenPDMS SPME fibre.

suspended in a polymer and layered over the fibre until a desired thickness is obtained. The polymer must not block the internal pores of the adsorbent nor inhibit the migration of the analytes to the adsorbent surface and eventually into the pores. For the preparation of CarboxenPDMS fibres and DVBPDMS fibres, the particles are suspended in a high-molecular-weight proprietary PDMS. This material rapidly cross links and serves as an adhesive to retain the particles. Figure 4.6 shows a schematic of the CarboxenPDMS fibres. An actual 100 3 magnified picture of Carboxen particles suspended in PDMS shows that the Carboxen particles are spherical and that multiple layers of carbon particles can be observed. The layer of PDMS used to adhere the particles is very thin, so that it does not interfere with the uptake of analytes into the adsorbent.

4.3.4.2 Coating of Adsorbents on Fused Silica To coat adsorbents on the fibre, it takes multiple coats to build up a thickness. Putting multiple coats on a fibre using the optical fibre tower is very difficult, and this process has not been successful. Uncoated fused silica cannot be coiled because it is fragile and will break. Coating fused silica fibre not produced in an optical tower is limited to 1 m length. The fibre is pulled through a special applicator multiple times until the desired coating thickness is obtained. This is termed a ‘batch coating’ process. Because only 1 m pieces of fibre are coated one at a time, it is very difficult to manufacture each strand with exactly the same coating thickness. There are other variables such as heating the fibre after applying a layer of coating, controlling the speed with which the fibre is pulled through the applicator and potential problems with solvent evaporation. All these factors may affect the coating durability and reproducibility.

4.3.4.3 Coating Adsorbents on Stableflex and Metal Fibres Using the Continuous-Coating Process One of the primary reasons for introducing the Stableflex fibre was to be able to use a continuous-coating tower. Because of the thin ‘plastic-like’ coating, the

SPME Commercial Devices and Fibre Coatings

113

fibre can be coiled. Likewise, the metal fibre can be coiled onto a reel for continuous coating. A small-scale continuous-coating process was developed using a miniature coating tower. The device is capable of coating 2030 m of fibre at one time. With continuous-coating process, variables that were a problem with batchcoating fused silica have been resolved. The speed that the fibre is pulled through the applicator is controlled with a direct-drive motor. Each layer of coating is bonded when it passes through a temperature-controlled heater. A less volatile solvent blend was developed, along with a restricted opening above the applicator to reduce evaporation. With the former variables now being controlled, a more reproducible and durable fibre coating can be obtained. Table 4.6 shows a comparison between fibres coated using the batch-coating process and fibres coated using the continuous-coating process. As can be seen in Table 4.6, there is a marked reduction in the variability between the fibres when the fibres are coated with the continuous-coated tower process. There is a significant reduction in the variability of the analyte response and reduced variation of the coating thickness. These results show that the fibres produced with the continuous-coating process. All assemblies containing metal adsorbent fibres and Stableflex fibres have been manufactured with the continuous-coating process.

4.3.4.4 PDMSDVB Coating DVB is a commonly used porous polymer, especially in air monitoring. This material has a high degree of mesoporosity, but it also has some micropores, as indicated in Table 4.6 Comparison of Fibre Coating Procedures Using Analyte Response Two Continuous-Coated Lots  CarboxenPDMS Fibres (n 5 12)

Average area Standard deviation % Standard deviation

ETH PRO

BUT

PEN

192 27

3,300 390

10,679 15,860 18,182 753 740 1,428

14

12

7.1

4.7

HEX

7.9

Average of All Analytes

Coating (µm) 87 2

9.1

2.7

Two Batch-Coated Lots  CarboxenPDMS Fibres (n 5 10)

Average area Standard deviation % Standard deviation

ETH

PRO

BUT

PEN

HEX

132 27

2,088 696

7,512 1,848

13,089 2,053

16,275 2,744

21

33

25

16

17

ETH, ethane; PRO, propane; BUT, butane; PEN, pentane; HEX, hexane.

Average of All Analytes

Coating (µm) 80 4

22

5.4

114

Handbook of Solid Phase Microextraction

Table 4.5. Porosimetry has shown that the micropores of DVB are fairly uniform and large compared to Carboxen 1006. Because of these properties, this adsorbent is primarily used for the extraction of semivolatile analytes and larger volatile analytes. Because the micropores are uniform in diameter, there are some concerns about analyte displacement. Langmuir’s isotherm states that if the pores are uniform, the analytes with the highest affinity will displace those with less affinity if the total concentration of the analytes exceeds fibre capacity and by increasing the extraction time.2 The percentage of the surface filled or covered is represented by the term theta (Ø). Essentially, Ø equals the ratio of the number of filled adsorption sites to the number of adsorption sites available. Thus, Langmuir’s isotherm is represented by Eq. (4.2): Ø 5 KPA =ð1 1 KPA Þ

ð4:2Þ

where K 5 ka/kd, ka 5 rate of adsorption, kd 5 rate of desorption and PA 5 gas pressure. There are two factors that can help to prevent displacement of analytes. The first factor is to work within the linear concentration range of the fibre in order not to exceed the fibre capacity. The second consideration is to limit extraction time. By reducing the extraction time, the amount of analyte extracted by the fibre is reduced, especially for larger analytes, which tend to have higher affinities for the fibre. DVB has a temperature limit of 270
C. After extended time above this temperature, the pores could begin to collapse, which would change the properties of the adsorbent. The coating can be exposed to solutions with a pH of range 211. The range limits are based primarily upon the stability of the PDMS binder. During the development of the PDMSDVB fibres, it was determined that through an activation process, that highly polar, short-chained aliphatic amines could be extracted. Without this activation step, amines were not extracted. This proprietary activation process exposes more of the DVB surface that enables amines to be extracted. The degree of activation is proportional to the amount of the amines extracted. However, too much activation could weaken the bonding of the coating to the fibre core. Through study, a balance that provides good extraction of amines along with fibre durability has been obtained. Table 4.1 shows the affinity that the fibre has for amines. All the analytes are present at the same concentration, but the amount of amines extracted with respect to n-propanol is much greater. This mixture is used as a quality testing procedure for the evaluation of each fibre production lot.

4.3.4.5 Carboxen 1006PDMS Coating Carboxen 1006 is one of a family of carbon molecular sieves. SPME fibres with this adsorbent coating were developed to extract volatile and small analytes. Carboxen 1006 is an ideal adsorbent for SPME because of the variety of pore sizes, and the micropores are narrow enough to retain analytes in the C3 range. In terms of molecular weight, Carboxen 1006 will retain analytes with a molecular weight greater than 35 g/mol. Because the pores are tapered, larger analytes can also be

SPME Commercial Devices and Fibre Coatings

115

retained in the larger portion of the pores. Figure 4.7 shows a schematic of a tapered pore in a Carboxen 1006 particle. The pore has a large opening near the surface that tapers down as it enters deeper into the particle. The pore continues through the particle and widens as it gets close to the surface. This type of pore is called a ‘throughput pore’. Because the pore has two openings, analytes can be desorbed much more efficiently. The gas flow can enter a pore and push the analyte out the other opening. This is possible when the analyte is heated and moving more rapidly. The throughput pore greatly reduces peak tailing. Most carbon molecular sieves, such as coconut charcoal, have tapered sealed pores with one opening. This causes the analytes to condense in the pores. For desorption, the analyte must vapourise and transfer into the gas stream that is being hindered by the closed pore. This makes desorption very inefficient and increases peak tailing. There is a limit to the size of analytes that can be extracted and desorbed using Carboxen 1006. A larger planar molecule can have such strong interaction with the carbon surface that the analyte is released very slowly. In fact, the desorption can be so slow that the analytes are not detected. To maximise the upper molecular weight range and to improve desorption efficiency, it is best to desorb CarboxenPDMS fibres at a minimum temperature of 280
C. The fibre can be desorbed at 320
C without damaging the coating. Even at this temperature, a typical upper molecular weight range for Carboxen 1006PDMS fibres is about 150 g/mol. This limit can vary depending upon the shape and degree of branching of the analyte. This fibre works best in the headspace mode because this reduces the amount of non-volatile material from fouling the surface. Extraction is more rapid when analytes are in the gas phase. Because of the good retention properties of the CarboxenPDMS fibre coating, the sample can be heated to a higher temperature compared to a PDMS-coated fibre. This will drive more of the analytes into the headspace. With absorbent fibres, if the sample is too warm, the headspace temperature will increase and the analytes will begin to migrate out of the fibre coating more rapidly. But because of the adsorption mechanism of Carboxen, the maximum temperature for optimised extraction is usually 1020
C higher compared to Figure 4.7 Drawing of Carboxen 1006 particle highlighting throughput pore.

Macropore Mesopore Micropore

116

Handbook of Solid Phase Microextraction

absorbent coatings. The higher temperature can drive more analytes into the headspace and often can improve recovery. It is best to try to minimise the extraction time so that the extraction of large, less volatile analytes is minimised. If immersion is required when using the CarboxenPDMS fibres, the fibre can withstand a pH range between 2 and 11.

4.3.4.6 DVBCarboxenPDMS Coating Because the CarboxenPDMS fibres have difficulty desorbing higher molecularweight analytes and PDMSDVB have difficulty extracting analytes with low molecular weights, a new fibre was developed. This fibre contains both adsorbents that are layered to extend the molecular weight range of analytes extracted with one SPME fibre. Initially, the DVB and Carboxen adsorbents were mixed together in PDMS at the same ratio as the individual adsorbent containing fibres. When this fibre was evaluated, it behaved similarly to a Carboxen- and PDMS-coated fibres with a thinner coating. All the analytes migrated to the stronger adsorbent. The ratios were varied, but the results were the same. By applying the principle used in thermal desorption tubes and purge traps, the adsorbents were layered so that the larger analytes first contact the weaker adsorbent applied on the outer layer and the smaller analytes migrate through the DVBPDMS layer and into inner layer of CarboxenPDMS. Because the larger analytes would migrate slowly through the DVB layer, they should remain in the DVB coating or very slightly into the Carboxen coating. The smaller analytes should migrate relatively fast through the DVB layer and get trapped in the Carboxen 1006 layer. The total fibre coating capacity should be the same, but each coating layer will have less capacity compared to the fibres containing a single adsorbent. Figure 4.8 shows a schematic of this fibre. The key to manufacturing was determining the thickness of the layers. To do this, two analytes were selected. A sample containing a medium-sized polynuclear aromatic hydrocarbon (PAH) and a small hydrocarbon were placed in a vial and extracted with the PDMSDVB fibres and the CarboxenPDMS by heated headspace. The DVBPDMS strongly favoured the PAH over the hydrocarbon by 12:1, whereas the CarboxenPDMS fibres greatly favoured the hydrocarbon over the PAH by 15:1. The goal was to optimise the thickness of each layer produce a ratio that was less than 3:1 in either direction. In this range, both analytes are nicely extracted and easily detected and quantified. A properly working fibre will extract a wide molecular weight range of analytes. Table 4.7 shows a mixture of alkanes, from ethane to tetraeicosane, extracted by the various adsorbent fibres. Because several mixes were required to make a standard, the concentration of the alkanes is varied, but the standard was consistent among all the samples extracted in this study. The results show that the CarboxenPDMS coating is the best fibre for the extraction of C2C16 alkanes, but the C18C24 analytes were poorly desorbed from this coating, leaving very low responses and poor detection limits. However, the PDMSDVB fibres are the best

SPME Commercial Devices and Fibre Coatings

117

DVB–PDMS (55 µm) Front view

Carboxen 1006–PDMS (30 µm) Stableflex core (120 µm)

DVB–PDMS (55 µm) Lateral slice view

Carboxen 1006–PDMS (30 µm)

Fibre core Carboxen 1006–PDMS (30 µm) DVB–PDMS (55 µm)

Figure 4.8 Views of layered coatings on DVBCarboxenPDMS-coated fibres. Table 4.7 Comparison of Adsorbent Fibres for the Extraction of Alkanes as Measured by Area Counts Alkanes, Number of Carbon Atoms

Carboxen PDMS

DVBCarboxen PDMS

PDMS DVB

2 3 4 5 6 7 8 9 10 12 14 16 18 20 22 24

200 800 2,600 7,000 13,000 10,000 53,000 49,000 160,000 230,000 480,000 490,000 44,000 18,000 6,600 3,400

80 260 1,000 3,200 7,000 5,000 29,000 29,000 100,000 150,000 300,000 380,000 110,000 60,000 24,000 12,000

30 200 800 2,200 2,200 14,000 14,000 49,000 90,000 260,000 360,000 130,000 110,000 64,000 46,000

n 5 3, fibres per coating type; 15-min extraction headspace, 50
C. Bold values highlights fiber coating that yielded the highest response per analyte.

choice for the detection of alkanes C18C24, but they performed the worst for the extraction of the smaller alkanes. The amount of each analyte detected with the DVBCarboxenPDMS fibres was always between the amounts from the other two fibre types. It helps to even the responses over a broad molecular weight range, which enables the analysts to detect trace concentration levels of analytes in a complex mixture that might have been missed by either of the other adsorbent fibre coatings.

118

Handbook of Solid Phase Microextraction

4.3.4.7 Carbopack ZPDMS Fibres Carbopack Z is a porous graphatised carbon black with the pore size of approxi˚ . This fibre was initially developed to extract coplanar polychlorinated mately 100 A biphenyls (PCBs) and dioxins out of an organic solvent, but it may be suitable for other applications as well. The carbon-to-PDMS ratio is higher than usual in order to expose more carbon surface area. The coplanar congeners can interact more tightly with the carbon surface. The other congeners, which are not planar, do not have as much contact with the carbon surface. This results in the fibre being more selective towards the planar congeners that are more toxic. Because of the strong interaction of dioxins with the fibre, the coating is thin so that they can be desorbed. The fibre is used to extract a sample after a final cleanup procedure with a nonane rinse that is concentrated to about 100 μL. At this point, a 12 μL injection is made into the GCMS system. Instead of making the injections, the fibre is inserted into the sample and the analytes are extracted with the fibre. The work of Maeoka et al.3 showed that the fibre enhanced the concentration of coplanar dioxins and PCBs compared to a 2-μL liquid injection. It was determined that the fibre must be desorbed at a high temperature to remove the analytes and reduce carryover. A desorption temperature of 340
C is recommended when using the Carbopack ZPDMS fibres for this application. This fibre has been used for other applications. For example, Purcaro et al.4 used this fibre to extract PAHs out of cooking oil. The oil was diluted in hydrocarbon solvent, and the fibre was used to extract PAHs because of the strong ππ interactions between polyaromatic ring structures and the carbon surface.

4.3.5

HPLC and Biocompatible Fibres

The currently available SPME fibres have been designed primarily for use with GC and thermal desorption in order to remove the analytes. For high-performance liquid chromatography (HPLC) use, the analytes need to be removed by desorbing the SPME fibre in an organic solvent or an appropriate combination of organic solvent and water. However, the problem with many of the current SPME fibre coatings is that they tend to swell in various organic solvents. This can cause the fibre coating to be stripped off the core when the fibre is retracted into the needle. There are two fibre coatings that were developed for HPLC use, but these coatings contain GC phases. One fibre CarbowaxTPR has been deleted as a product due to the availability of materials. The other fibre has the same PDMSDVB coating, as described in Section 4.3.4.4, but it uses a thicker and less breakable core. Both of these fibre coatings have coating durability issues when exposed to certain solvents. A new series of fibres was developed that embeds coated HPLC silica particles into a biocompatible, non-swelling polymer. This fibre should be commercially available by the time of this printing. The bonded silica coating is applied to the metal fibre core to reduce breakage. The binding polymer is very resistant to a variety of solvents especially at room temperature. Table 4.8 compares the swelling properties of the new bonded silica fibres with CarbowaxTPR fibres.

SPME Commercial Devices and Fibre Coatings

119

Table 4.8 The Effects of Various Solvents on the Swelling of Two SPME Fibre Coatings

Solvent Water Acetonitrile (ACN) Methanol (MeOH) Dichloromethane Hexane Acetone Water:ACN (50:50) Water:MeOH (50:50)

Bonded Silica Fibre Coating Thickness (µm)

CarbowaxTPR Coating Thickness (µm)

Solvent Exposure

Difference in Thickness

Solvent Exposure

Difference in Thickness

None 44 44 44 44 44 44 44 44

0 0 0 0 0 0 0 0

None 50 50 50 50 50 50 50 50

10 1 11 2 0 0 20 18

15 min 44 44 44 44 44 44 44 44

15 min 60 51 61 52 50 50 70 68

Table 4.8 shows that no swelling is observed with the bonded silica fibres. However, the CarbowaxTPR fibres were swollen when exposed to certain solvents, particularly the aqueousorganic combinations commonly used in HPLC. The results show that the bonded silica fibres are the first fibres that are truly resistant to a wide variety of solvents and solvent mixtures. In addition to solvent resistance, the bonded silica fibres are also biocompatible thus enabling their use for in vivo SPME applications (see Section 3.3.4, in Chapter 3, and Chapter 12). A non-swelling proprietary binder has been developed that is used to hold the silica particles to the fibre. This biocompatible binder does not allow the deposit and build-up of large macromolecules such as proteins, complex carbohydrates and lipids on its surface. Some biocompatible materials are methacrylate, polyacrylonitrile, polysulfones and PEG. The binder tends to repel the macromolecules, but it allows the migration of the analytes of interest, usually drug metabolites, through the binder to interact with the silica surface. Because the binder is biocompatible, the fibre can be inserted directly into a biological fluid to extract the analytes of interest. With standard SPME fibres, the large macromolecules can adsorb to the fibre surface and prevent the extraction of the analytes or the bound macromolecules will not be removed during the desorption process. Figure 4.9 shows calibration curves obtained for diazepam in various biofluids obtained by Dajana Vuckovic at the University of Waterloo. The drug was extracted out of urine, plasma and phosphate buffer solution (PBS) (pH 7.4, containing 0.7% NaCl). Excellent linearity was obtained in all the matrices. The response obtained in plasma is lower due to the presence of proteins that bind some of the drug. The amount of drug extracted by SPME is directly proportional to free (unbound) concentration of the drug. In plasma, the free drug concentration is lower than in urine or PBS, thus resulting in a lower amount extracted by SPME (as shown by lower area ratio response in Figure 4.9).

120

Handbook of Solid Phase Microextraction

0.30 y = 0.007x – 0.0038 R2 = 0.996 y = 0.0010x – 0.0007 R2 = 0.9981

0.25

Area ratio

0.20

PBS Urine Plasma

y = 0.003x – 0.0017 R2 = 0.9996

0.15 0.10 0.05 0.00 0.00

100.00

200.00

300.00

400.00

500.00

600.00

700.00

Standard concentration (ng/mL)

Figure 4.9 Calibration curves for diazepam in various fluids extracted using bonded silica C18-coated fibre assemblies. Experimental conditions: diazepam spiked in fluids preincubated 60 min prior to extraction, 1.5 mL sample in 2 mL vial, 2 min direct immersion, static (PBS, urine) and 5 min direct immersion, static (plasma). Ninety minutes desorption in 100 μL 50:50 acetonitrile:water containing 100 ppb of internal standard, with agitation, detection using LC-Iontrap-MS; 20 μL injection volume; C18 column, 50 mm 3 2.0 mm, 5 μm; mobile phase: A: 10:90:0.1 acetonitrile:water:acetic acid; B: 90:10:0.1 acetonitrile: water:acetic acid; flow rate: 0.3 mL/min; detection: ESI 1 MS/MS; scan range: MS/MS transitions: 285.0/257.0 (diazepam), 247.0/204.0 (Internal Standard). (Source: Figure courtesy of Dajana Vuckovic and Janusz Pawliszyn, University of Waterloo.)

4.4

A Guide for the Selection of the Appropriate SPME Fibre

There are four major criteria that are commonly used in selecting the proper fibre coating for a particular application. These are (i) the molecular weight and size of an analyte, (ii) analyte polarity, (iii) the analyte concentration level and range and (iv) the complexity of the sample. This section provides a guideline on how to select the appropriate fibre based upon the analyte and sample properties in order to help the analyst reduce method development time.

4.4.1

The Effect of Analyte Molecular Weight and Size

The molecular weight of an analyte determines how rapidly it can move in and out of the phase coating and through the sample. A smaller analyte will move faster and is not retained as well, whereas the larger analytes migrate through the coating and sample more slowly and take a much longer time to reach equilibrium.

SPME Commercial Devices and Fibre Coatings

121

Figure 4.10 shows an analyte molecular weight range typically extracted using a given SPME coating. The coating thickness plays a major role in determining the equilibrium time (see Chapter 2). The other factors that determine equilibrium time are the distribution constants and the thickness of the boundary layer.5 The boundary layer is the fluid that surrounds a solid surface that has less velocity than the surrounding fluid due to increased friction with the solid surface. The thickness of the boundary layer can affect the rate of diffusion of an analyte into the fibre coating. Desorption of analytes can be affected by the coating thickness. The rate of desorption will be slower with thicker fibre coatings. The effect of coating thickness in relationship to the size of the analytes and extraction time is shown in Figure 4.11. Five aromatic compounds ranging in molecular weight from 92 to 502 g/mol were extracted using three PDMS-coated fibres with a thickness of 7, 30 and 100 μm. A water solution containing 50 ng/mL of each of the analyte was extracted using the three fibres for 15 and 30 min. The figure shows the results adjusted to the same scale showing the area responses for each of the analytes. The responses for toluene and xylene for all of the fibre coatings were the same when comparing 15- and 30-min extraction times. This would indicate that these smaller analytes reached equilibrium on all the fibres in 15 min. The 7-μm fibre gave similar response for each analyte for the two extraction times tested. This indicates that with a thin coating, equilibrium can be reached quickly. The 30-μm fibre has a higher fibre capacity relative to the 7-μm fibre. The responses at 15 min for chrysene and decachlorobiphenyl (DCBP) were similar between the 7- and 30-μm fibres, but at 30 min, the responses for these analytes were greater on the thicker 30-μm fibre due to the increased capacity. The 100-μm fibre was the best fibre for retaining smaller analytes compared to the other fibre coatings. This would be expected because the thicker coating helps to retain the fast-moving smaller analytes. For the larger analytes, the advantage of the 100-μm fibre over the thinner coatings should get smaller. At 15 min, the analyte responses for DCBP and chrysene between the 30- and 100-μm fibre coatings were similar as expected. At 30 min, the response for the 100-μm fibre did not increase as rapidly as the 30-μm fibre. This could be due to two reasons. The first reason could involve the boundary layer. The boundary layer is affected by the

7 µm PDMS 30 µm PDMS 100 µm PDMS DVB DVB–Carboxen Carboxen 0

150

300

450

Analyte molecular weight range

Figure 4.10 Molecular weight extraction range for SPME fibre coatings.

122

Handbook of Solid Phase Microextraction

(A) 15-min extraction

10,000,000 9,000,000

7 µm

8,000,000

30 µm

7,000,000

100 µm

6,000,000 5,000,000 4,000,000 3,000,000 2,000,000 1,000,000 0 Toluene

Xylene

(B) 10,000,000 9,000,000 8,000,000 7,000,000 6,000,000 5,000,000 4,000,000 3,000,000 2,000,000 1,000,000 0

Acenapthene

Chrysene

DCBP

Chrysene

DCBP

30-min extraction 7 µm 30 µm 100 µm

Toluene

Xylene

Acenapthene

Figure 4.11 Comparison of PDMS SPME fibres of various thickness (7, 30 and 100 μm) for the extraction of aromatic compounds for (A) 15 min and (B) 30 min. Analytes in water at 50 ng/mL. Extraction by immersion for 15 or 30 min with agitation. Analysis by GCFID. Desorption temperature was 280
C for all fibre coatings. SLB-5 MS column, 30 m 3 0.25 mm ID, 0.5 μm; 50
C (hold 1 min) to 320
C at 20
C/min (hold 5 min).

stirring rate and can change the mass transfer rate.6 There is a possibility that the stirring rates were not identical, even though an effort was made to keep them similar. The second possibility is that there was incomplete desorption of the two larger analytes with the 100-μm PDMS fibre. The 30-μm fibre will reach equilibrium more quickly than the 100-μm PDMS, but the responses should be similar up to the time of equilibrium. If the extraction times were increased beyond the equilibrium of the 30-μm fibre, the increased capacity of the 100-μm fibre would allow more of the analytes to be extracted, but the equilibrium time for these analytes using the 100-μm fibre might be several hours. Lancaster Labs uses 30-μm PDMS fibres to screen their purge-and-trap samples by solid-phase microextractiongas chromatographyflame ionisation detector (SPMEGCFID) prior to analysis by purge-and-trap GCMS (T Schumacher, personal communication, Lancaster Laboratories, New Holland, PA, 1998). This allows them to dilute the samples prior to the GCMS analysis if any of the components are at a concentration greater than 200 μg/mL. This prevents the need to rerun samples and reduces the chances of contaminating the purge trap.

SPME Commercial Devices and Fibre Coatings

123

Figure 4.12 shows a chromatogram of the standard used by Lancaster Labs. Using the 30-μm PDMS fibre, the extraction time is only 12 s and the total cycle time is less than 5 min per sample. The 30-μm fibre was selected because of a good sample capacity coupled with a fast desorption rate. In addition to molecular weight consideration, the size and shape of a molecule also plays an important role in the selection of the coating. Essentially, a flat planar structure with no substitution groups can interact with the phase coating or adsorbent, usually though ππ interactions. Examples of these types of analytes are PAHs and long chained molecules with multiple double bonds that make them more rigid. With these analytes, you can effectively added 3050 amu to the molecular weight of the analyte. Conversely, if the analyte is highly branched or is an aromatic with substitution groups, such as chlorine or bromine, this reduces the interaction with adsorbents, and their effective size is actually smaller if the substitution group is highly electronegative. Because the size of the molecule is smaller than an all hydrocarbon structure of the same molecular weight, it has the extraction efficiency similar to a molecule that is approximately 30 amu less in molecular weight. This number will vary depending upon the degree of substitution and is just a guideline. Generally for smaller analytes (molecular weight , 125), an adsorbent fibre coating that can retain these fast-moving analytes is usually a better option than absorbent fibres (see Figure 4.10). The CarboxenPDMS fibres are the best choice for these analytes, especially at trace concentration levels. Other fibre coatings containing DVB and those with thicker phase coatings may be suitable, especially if the analytes are in higher concentration levels or the samples contain multiple analytes (Figure 4.13). In this figure, the area counts are labelled above the peaks. The CarboxenPDMS fibre coating extracted an average of 200 times more of these analytes than the 100-μm PDMS fibre. For analytes with higher molecular weights, absorbent fibres may be the better option, but DVB-containing fibres with larger pores are also suitable for many of these analytes. The comparison of the various fibre coatings for the analysis of larger molecular weight analytes is shown in Figure 4.14. The fibre that is least effective for the extraction of chrysene is CarboxenPDMS. This four-ringed PAH has strong interaction with the carbon surface and is not efficiently desorbed. The response is so low that the bar in Figure 4.14 representing the CarboxenPDMS fibre coating is not visible. Even though DCBP has a higher molecular weight than chrysene, it is desorbed more efficiently by the CarboxenPDMS fibres than chrysene. The multiple chlorine substitutions on DCBP shield the aromatic rings from the Carboxen surface, reducing the ππ interactions. Most of the other fibres extract these large non-polar analytes fairly well. The extraction time plays a major role. The extraction time for the analytes shown in Figure 4.14 was 30 min. The larger analytes are extracted more effectively by the 30-μm PDMS fibre than the 100-μm PDMS fibre. This is partly due to the migration rate of these large analytes into the fibre coating and partly due to the desorption rate. In general, the absorbent fibre coatings (those not containing DVB or Carboxen) are the better choice for these large analytes.

124

Handbook of Solid Phase Microextraction

1.8 1.6 mV (Span = 1.3)

17

1.4 1.2

9 11 1 7

1.0 0.8

4 2 5 3

15

16

8 10 13 14 12

6

0.6 0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0 2.2 2.4 2.6 2.8 Min

Figure 4.12 Screening of volatile compounds using SPME with 30-μm PDMS-coated fibre. Peak identities (1) methanol, (2) acetone, (3) methylene chloride, (4) MTBE, (5) cis-1,2dichloroethylene, (6) 1,1,1-trichloroethane, (7) benzene, (8) trichloroethane, (9) toluene, (10) tetrachloroethylene, (11) chlorobenzene, (12) ethylbenzene, (13) m-xylene, p-xylene, (14) o-xylene, (15) isopropylbenzene, (16) 1,4-dichlorobenzene and (17) naphthalene. Sample 0.7 mL of water spiked with 2600 μg/L of each analyte with 0.25 g NaCl. Headspace extraction for 12 s at ambient temperature. Desorption was for 3 min at 250
C; GC column  SPB-1, 10 m 3 0.20 mm ID, 1.2 μm coating; oven programme, 70
C (hold 0.2 min) to 180
C at 50
C/min; hydrogen carrier gas at 12 psi, FID detection.

The DVBCarboxen fibres were designed to extract a wider molecular weight range, as discussed in Section 4.3.4.6. This fibre coating consists of a layer of DVB suspended in PDMS coated over a layer of CarboxenPDMS. The larger analytes are retained by DVB layer, whereas the smaller analytes can migrate into the Carboxen 1006 adsorbent. This effectively increases the molecular weight range that the fibre can extract. As shown in Figures 4.13 and 4.14, this fibre was a good compromise over a very wide molecular weight range. The only limitation is that this fibre will have less sample capacity due to the thinner coatings of each layer.

4.4.2

The Effect of Analyte Polarity

The second criterion that is important for fibre selection is analyte polarity. For volatile, low-molecular-weight compounds, Carboxen- and PDMS-coated fibres are still the best choice. Once the molecular weight increases above 80 amu, the effect of fibre polarity becomes more evident for the extraction of polar analytes. There are only two polar fibre coatings: the 60-μm PEG phase and the 85-μm PA phase, as summarised in Table 4.2. However, the other fibres still have some affinity for polar analytes. Ideally, it is desirable for polar fibre coatings to be more

125

Propanal

Nitropropane

Acetone

311 528 2,119 1,671

208 121 758 499

5,930

14,420 857 1,062 7,229 3,209

316 192 1,306 837 7,829

PDMS PA PDMS–DVB CW–DVB DVB–CAR Carboxen

15,720

55,870

59,563

SPME Commercial Devices and Fibre Coatings

Propionitrile

Figure 4.13 Comparison of SPME fibre coatings for the extraction of low-molecular-weight analytes (area response). One microgram per millilitre of each analyte in water containing 25% NaCl, pH 7, headspace extraction at 50
C for 15 min using various SPME fibre coatings. Desorption temperatures varied per fibre coating. GC column  SPB-1 sulphur, amine deactivated 30 m 3 0.32 mm ID, 4.0 μm; oven programme, 45
C (hold 1.5 min) to 80
C at 8
C/min to 200
C at 20
C/min. Detector FID at 290
C.

MW = 502 MW = 154

7 µm 30 µm 100 µm PA PDMS–DVB DVB–CAR Carboxen

9.1E+06

5.6E+04

Acenaphthene

Decachlorobiphenyl

4.9E+03

MW = 228

Chrysene

Figure 4.14 Comparison of SPME fibre coatings for the extraction of higher molecular weight (MW) semivolatile analytes (area response). Concentration of analytes in water containing 25% NaCl, pH 11, was 50 ng/mL each. Extraction was by immersion for 30 min using various SPME fibre coatings with agitation. Analysis by GC/MS using SLB-5 MS column, 30 m 3 0.25 mm ID, 0.5 μm; 50
C (hold 1 min) to 320
C at 20
C/min (hold 5 min).

126

Handbook of Solid Phase Microextraction

selective  not only to extract polar analytes but also to reduce the extraction of non-polar analytes in comparison to non-polar fibre coatings. Figures 4.15 and 4.16 show the comparison of four SPME fibre coatings for the extraction of baseneutral analytes. Figure 4.15 shows the polar fraction of the analytes. The CarbowaxDVB fibre, which is no longer commercially available, is included in the evaluation, so that the PEG fibre can be compared to it. In all cases, the 60-μm PEG fibre was the best choice for these analytes. The PA fibre coating, which is slightly less polar than the PEG, was also suitable for the extraction of these analytes, while the 100-μm PDMS fibre extracted these polar analytes very poorly. Because counts were so low, the area response did not show on the graph in some cases, so the area counts are listed above the bars. Figure 4.16 shows that the analytes in neutral fractions are extracted efficiently by all the fibres, as indicated by the higher area counts. Interestingly, the moderately polar PA fibre was best for some of the neutral compounds. The PA fibre coating has an affinity for aromatic compounds. The PEG fibre, being the most polar fibre, extracted these analytes less efficiently than the other fibres. This indicates that the fibre does selectively extract polar analytes compared to the other SPME fibres. The selectivity of the PEG fibre was demonstrated further with the extraction of water-soluble solvents. Figure 4.17 shows a typical chromatogram of the extraction of these solvents out of water using the 60-μm PEG fibre. Figure 4.18 compares the area counts of the analytes obtained for SPME using various fibre coatings. The PEG fibre is the best for extraction of the polar alcohols. The less polar analytes, such as methyl tert-butyl ether (MTBE) and ethyl acetate, are best extracted by the PDMS-containing fibres. The polar PEG fibre is the least efficient for the extraction of these analytes. The fibre demonstrated selectivity by enhancing the extraction 6.E + 05

5.E + 05

85 µm Polyacrylate 60 µm PEG 65 µm CW–DVB 100 µm PDMS

0.E + 00 Aniline

405

4,217

1.E + 05

Benzyl alcohol

o-Toluidine

p-Chloroaniline

o-Nitroaniline

806

57,453

2.E + 05

1,202

198,216

3.E + 05

1,856

4.E + 05

m-Nitroaniline

p-Nitroaniline

Figure 4.15 Comparison of SPME fibres for the extraction of baseneutral analytes  polar fraction. Analyte concentration at 25 ng/mL in water with pH 9. Extraction by immersion for 30 min using CombiPal with agitation. Analysis by GCMS, SLB-5 MS column, 30 m 3 0.25 mm ID, 0.5 μm; 50
C (2.0 min) to 130
C at 12
C/min to 200
C at 20
C/min to 260
C at 15310
C at 20
C (4 min); ion trap MS, m/z 5 50230 at 0.65 μs per scan.

SPME Commercial Devices and Fibre Coatings

127

1.8E + 07 1.6E + 07 85 µm PA 60 µm PEG CW–DVB 100 µm PDMS

1.4E + 07 Area counts

1.2E + 07 1.0E + 07 8.0E + 06 6.0E + 06 4.0E + 06 2.0E + 06 0.0E + 00

Undecane

Benzoanthracene

Naphthalene

Acenaphthene

Figure 4.16 Comparison of SPME fibres for the extraction of baseneutral analytes  neutral fraction. All conditions are the same as in Figure 4.15. UV 6 8 Signal intensity

9

2–5

1,400

11 12

1,000

1

7

10

600 0

2

4

6 8 Retention time (min)

10

12

Min

Figure 4.17 Analysis of water-soluble solvents using PEG SPME fibre; analytes and concentration in μg/mL (1) methanol 20, (2) ethanol 20, (3) acetonitrile 20, (4) acetone 20, (5) isopropanol 20, (6) n-propanol 10, (7) MTBE 3, (8) ethyl acetate 5, (9) n-butanol 10, (10) 1,4-dioxane 20, (11) butyric acid 5 and (12) phenol 0.2, in water with 25% NaCl, pH 2, 0.05 M phosphate buffer; extraction: 10 min; direct immersion with 60-μm PEG fibre with agitation; desorption: 5 min at 240
C; injector: splitless/split, closed initial 0.75 min than opened 50:1 with 0.75-mm ID liner; column: SPB-1 sulphur, 30 m 3 0.32 mm, 4.0 μm; column oven: 5
C (1.5 min) to 80
C at 8
C/min to 230
C at 20
C/min (10 min); carrier gas: helium at 13 psi constant pressure, 40 cm/s at 40
C; detector: FID at 300
C.

of more polar analytes and decreasing the extraction of less polar analytes. The CarbowaxDVB fibres, though polar, was not as selective because the DVB portion also extracts the less polar analytes, as shown in the amount of MTBE and ethyl acetate extracted relative to the other fibres. Because the PEG fibre does not swell in hydrocarbon solutions, and because of its selectivity towards polar analytes, it can be used to extract oxygenated

128

Handbook of Solid Phase Microextraction

1.5E + 04 85 µm PA 60 µm PEG CW–DVB 100 µm PDMS

Area counts

1.2E + 04 9.0E + 03 6.0E + 03 3.0E + 03 0.0E + 00 ol

ol

an

h et

M

e

le

tri

an

h Et

i on

et

Ac

on

t ce

A IP

A

te

E

H

rO

P n-

TB

Et

e Ac

ne

H

ta

M

O

u -B

n

D

4-

1,

id

a iox

tr

Bu

yic

ac

l

no

e Ph

Figure 4.18 Comparison of analyte response of water-soluble solvents using various SPME fibre coatings.

Butane

20,000

SPME

Pentane

Ethanol n-Propanol

10,000

IPA

200,000

Butane Direct

Pentane

Ethanol

100,000

IPA

n-Propanol

0 0

1

2

3

4

Min

Figure 4.19 Analysis of oxygenates (400 ppm) in gasoline by SPME and direct injection. 3.5 mL in 4 mL vial; extraction: 15 min; direct immersion with 60-μm PEG fibre; 5-min desorption at 240
C; injector: split 50:1, 240
C; SPB-1 sulphur column, 30 m 3 0.32 mm, 4.0 μm; oven programme: 40
C (1.5 min) to 80
C at 8
C/min to 260
C at 20
C/min (10 min); carrier gas: helium at 13 psi constant pressure; FID detector at 300
C.

compounds used to boost the octane in gasoline. Figure 4.19 compares the SPME extraction of gasoline to a direct injection of gasoline. As can be seen in the figure, there is a 10-fold decreased response for hydrocarbons using SPME and an increased response for the oxygenated compounds relative to the direct injection. Due to differences in distribution constants, the PEG fibre showed a much lower response for hydrocarbons while demonstrating affinity for the oxygenated compounds. The gasoline sample was spiked with 400 ppm of each of the oxygenated compounds. Table 4.9 shows the enhancement achieved by SPME for the oxygenated compounds versus the direct injection method. The average response of the seven largest hydrocarbons that eluted near the oxygenated compounds was used as a hydrocarbon reference or internal standard. The actual hydrocarbons ranged from butane to heptane. The enhancement values shown in

SPME Commercial Devices and Fibre Coatings

129

Table 4.9 Comparison of SPME Extraction to Direct Injection for the Analysis of Oxygenated Compounds (400 ppm 5 0.4%) in Gasoline

Ethanol Isopropanol n-Propanol Average HC C4C7 IS

Area Counts

Relative to Butane

SPME

Injection

SPME

Injection

21,214 12,403 16,212 48,528

2,055 2,191 2,636 719,824

0.493 0.288 0.376

0.003 0.003 0.004

SPME Enhancement 153 84 91

Direct injection of gasoline, 0.2 μL split 50:1 SPME extraction (15 min); direct immersion with PEG fibre, split 50:1 during desorption.

Table 4.9 are lower than the actual values because all of the hydrocarbons used for reference saturated the FID with the direct injection. Because of the saturation, the average hydrocarbon value shown for direct injection is low.

4.4.3

Effect on Sample Concentration and Complexity on Fibre Selection

Sample concentration and the complexity of the sample also play a significant role during fibre selection. Fibres that extract by adsorption have a more limited capacity compared to absorbent fibres with a similar thickness. Because of this limited capacity, analytes tend to compete more for the available sites (see Chapter 2). To demonstrate the effect of concentration on fibre selection, a group of volatile analytes was extracted out of water over a large concentration range. By plotting the concentration versus the response of the analytes, one can determine the type of extraction mechanism. Carboxen 1006, used in the CarboxenPDMS fibres, has tapered throughput pores. Because the pores are not uniform, analytes with different size and shapes fit in different regions in the pores. Figure 4.20 shows the loglog plot of seven analytes across a wide concentration range. The loglog plots were required because four to five orders of magnitude were plotted. The results show that all the analytes were extracted at 5 ng/mL. The polarity of the analytes increases in order from top to bottom. The responses for all the analytes began to level off at 5 μg/mL, indicating that the maximum capacity for the analytes with this fibre coating. Between 10 and 100 μg/mL, the responses remained constant. Where lines cross over each other, this could be an indicator of displacement. When displacement occurs, generally the response for the displaced analyte will decrease and the analyte doing the displacement will increase. The decreased response for isopropanol may simply be a solubility issue as the concentration of the other analytes increased. The DVB fibres have larger micropores and a more uniform mesopore that could lead to displacement. Figure 4.21 shows a plot of the extraction of the same analytes under the same conditions using a PDMS- and DVB-coated fibres.

130

Handbook of Solid Phase Microextraction

6.5 6 5.5

Log response

5 4.5

15-min extraction time Pentane Nitropropane Methylene chloride Propionitrile Acetone Isopropanol Dioxane

4 3.5 3 2.5

1.5 0.5

100 µg/mL

100 ng/mL

2 1

1.5

2

2.5 3 3.5 Log of concentration

4

4.5

5

5.5

Figure 4.20 Analyte concentration versus analyte response using CarboxenPDMS fibres. Analytes in water with 25% NaCl, pH 2, 0.05 M phosphate buffer. Extraction: 10 min, direct immersion with agitation; desorption: 5 min at 300
C; injector: splitless/split, closed initial 0.75 min, then opened 50:1 with 0.75-mm ID liner; column: SPB-1 sulphur, 30 m 3 0.32 mm, 4.0 μm. Column oven: 45
C (1.5 min) to 80
C at 8
C/min to 230
C at 20
C/min (10 min); carrier gas: helium at 13 psi constant pressure, 40 cm/s at 40
C. Detector: FID at 300
C.

The results indicate that the response for the polar analytes is not detected until the concentration increases to 50100 ng/mL, while the non-polar analytes could be extracted at 510 ng/mL. The response begins to level off for some analytes between 10 and 50 μg/mL, and there appears to be some displacement. The response for methylene chloride continues to climb, while the slopes of the response lines for dioxane and acetone begin to decline. A longer extraction time would show this effect more dramatically. Figure 4.22 shows the same analysis using the absorbent 100-μm PDMS-coated fibre. The results show that the minimum detection limits are much higher for these smaller analytes, as expected with an absorbent coating, but the linearity is excellent up to 100 μg/mL, which was the highest concentration level evaluated. No displacement of the analytes was observed. The advantage of absorbent fibres is that each analyte appears to reach its point of saturation independently of the other analytes present. Even in the presence of samples with high concentration levels of a component, such as ethanol in alcoholic beverages, the trace flavour components still can be extracted and quantified.7 As samples become more complex, competition for space in the fibre can become more challenging. With adsorbent fibres, a component with a high

SPME Commercial Devices and Fibre Coatings

131

15-min extraction time 6

Pentane Nitropropane Methylene chloride Propionitrile Dioxane Acetone Isopropanol

5.5 5

Log response

4.5 4 3.5 3 2.5 2 1.5

10 ng/mL 1 0.5

1

100 ng/mL 1.5

2

1 µg/mL

10 µg/mL

2.5 3 3.5 Log concentration (ppb)

4

100 µg/mL 4.5

5

5.5

Figure 4.21 Analyte concentration versus analyte response using PDMSDVB fibres. Analytes in water with 25%NaCl, pH 2, 0.05 M phosphate buffer, extraction: 10 min, direct immersion with agitation; 5-min desorption at 270
C. All other conditions are the same as in Figure 4.20.

concentration level could displace a minor component, or a component with higher affinity for the adsorbent could displace a component with less affinity. In the work by Black and Fine,8 MTBE was being detected in gasoline that had leaked into waterways. In cases where the gasoline sample was at a high concentration level in the water sample, MTBE would often be displaced from the fibre. It was determined that shorter extraction times helped to reduce this problem. The use of the dual-layer fibre can help reduce problems with displacement because the DVB coating retains many of the analytes that would normally compete on the CarboxenPDMS surface if the DVBPDMS coating was not present. As long as the concentration levels are not high and the extraction times are relatively short, complex samples can be analysed using this fibre.

4.4.4

Summary of Selection Guidelines

The following summary can be made based upon the results of this section and can be used as a guideline for selecting the appropriate SPME fibre: G

G

Adsorbent fibres are better for analytes at low concentration levels. Adsorbent fibres have a limited capacity, so the linear range for each analyte needs to be determined.

132

Handbook of Solid Phase Microextraction

5.5

Log response

4.5

15-min extraction time Pentane Nitropropane Methylene chloride Propionitrile Acetone Dioxane Isopropanol Linear (Pentane)

3.5

2.5

1.5 100 µg/mL

100 ng/mL 0.5 1.5

2

2.5

3

3.5

4

4.5

5

5.5

Log concentration (ppb)

Figure 4.22 Analyte concentration versus analyte response using 100-μm PDMS fibre. Analytes in water with 25% NaCl, pH 2, 0.05 M phosphate buffer. Extraction: 10 min, direct immersion with agitation; 5-min desorption at 280
C. All other conditions are the same as in Figure 4.20. G

G

G

G

G

G

Absorbent fibres are better for complex samples with varied concentration ranges. DVBCarboxenPDMS fibres are good for complex samples at low concentration levels due to the two adsorbent beds. The 30-μm PDMS-coated fibre is good for screening samples at high concentration levels over a broad molecular weight range. Absorbent fibres are a better option for dirty samples that may contain multiple unknown compounds. PEG-coated fibres with polar selectivity are suitable for a wide range of analytes. PA fibres are suitable for the extraction of substituted aromatic analytes.

References 1. B Baltensperger & R Shirey, Supelco Reporter 22.2 (2004) 6 2. T Gorecki, Z Yu & J Pawliszyn, Analyst 124 (1999) 643 3. T Maeoka, M Miyzaki, T Kaneko & R Shirey, Dioxin 2003  23rd Symposium on Halogenated Environmental Pollutants and POPs (2003) RPJ Associates: Boston, MA, vol. 60, p. 57

SPME Commercial Devices and Fibre Coatings

133

4. G Purcaro, P Morrison, S Moret, LS Conte & PJ Marriot, J Chromatogr A 1161 (2007) 284 5. J Pawliszyn, Solid Phase Microextraction, Theory and Practice (1997) Wiley-VCH: New York, NY, p. 122 6. J Semenov, J Koziel & J Pawliszyn, J Chromatogr A 873 (2000) 39 7. S Selli, M Kurkcouglu, E Karfkas, T Cabaraglu, B Deminci, KHC Baser & A Canbas, J Inst Brewing 107 (2004) 837 8. L Black & D Fine, Environ Sci Technol 35 (2001) 3190

5 Automated SPME Systems Janusz Pawliszyna, Dajana Vuckovicb, Fatemeh Mirnaghia and Sanja Risticevica a

Department of Chemistry, University of Waterloo, Waterloo, Ontario, Canada Donnelly Centre for Cellular and Biomolecular Research, University of Toronto, Toronto, Ontario, Canada

b

5.1

Automated Solid-Phase Microextraction Gas Chromatography

The fibre configuration of solid-phase microextraction (SPME) is best suited for automation with gas chromatography (GC), due to its similarity to the traditional GC syringe for liquid injection. In principle, any autosampler that is able to perform syringe injection can be modified to be capable of automated SPME GC. The first commercial version of the laboratory SPME device was introduced by Supelco in 1993 (see Figure 4.3 in Chapter 4). In addition to standard PDMS coatings of various thicknesses and polyacrylate, Supelco developed new mixed phases based on solid/liquid sorption, such as Carbowaxs, CAR PDMS, polydimethylsiloxane divinylbenzene (PDMS DVB) and DVB CAR PDMS. New coatings and devices are expected to follow as interest in SPME grows, along with the unprecedented number of new applications appearing in the literature.

5.1.1

Automation of Fibre SPME GC

The first reference in the literature to automated SPME analysis was by Arthur et al.,1 published even before the first commercial SPME fibres were released. This work involved adapting a Varian model 8100 syringe autosampler so that it could accept an SPME device. Magnetic stirring was used for agitation. This was achieved by setting up a microstirrer so that it would be in close proximity to the vial being sampled. The first commercial GC autosampler with a capability for SPME was the model 8200, released by Varian in 1993 (Z Penton, personal communication, 2004; JR Berg, personal communication, 2004). Initially, this autosampler could perform only static sampling and was not temperature controlled. However, it was able to start the extraction of the next sample, whilst the previous one underwent GC analysis (Z Penton, personal communication, 2004).2 During 1996, a modified device was launched that allowed vibration of the fibre to agitate the sample, while in 1998, temperature control in the form of a thermostated sample carousel was added (Z Penton, personal communication, 2004; JR Berg, personal communication, 2004). One limitation of the system was the vial sizes that it could accept.3,4 Handbook of Solid Phase Microextraction. DOI: 10.1016/B978-0-12-416017-0.00005-X © 2012 Elsevier Inc. All rights reserved.

136

Handbook of Solid Phase Microextraction

CTC Analytics (Zwingen, Switzerland) introduced an SPME option on its CombiPALt autosampler at the beginning of 1999 (Figure 5.1; B Baltensperger, personal communication, 2004). This autosampler allows additional capabilities, such as full temperature control of individual samples (sample trays that can be heated or cooled), stirring provided by rotation of an agitator tray and a fibre conditioning station, which allows ‘bake-out’ of the SPME fibre outside the injection port. The CombiPAL is also sold by a number of other instrument distributors and, in some cases, under different names. A different robotic arm6 for analysing samples with an awkward shape or size, such as living plants,7 was described. However, this device is able to perform only static sampling. More recently, Gerstel GmbH (Mulheim an der Ruhr, Germany) introduced the MultiPurpose Sampler (MPS 2) as a new generation autosampler capable of SPME GC automation. This autosampler offers both orbital agitation and a stirring mechanism during incubation and extraction processes. It also offers simplified, user-friendly software for the specification of SPME method parameters. Next, Thermo Fisher Scientific (Milan, Italy) launched its version of an autosampler compatible with SPME GC, namely, the TriPlus autosampler. The latter autosampler version is significantly modified compared to the previous two due to the implementation of a rocking agitation mechanism and the possibility of closing the agitator tray during extraction. PAS Technology (Magdala, Germany) also introduced its version of the CONCEPT autosampler for SPME GC applications. This autosampler incorporates the use of both the orbital and magnetic stirring agitation mechanisms.

Figure 5.1 Commercial SPME GC autosampler (CTC Analytics CombiPAL): A sample preparation/injection arm; B sample trays; C needle heater/fibre conditioning station; D sample preparation station/agitator tray. (Source: Reprinted with permission from Ref. 5. r Wiley-VCH Verlag GmBH & Co. KGaA.)

Automated SPME Systems

137

5.1.1.1 Septum Coring In any fibre SPME GC application, septum coring can be an issue when traditional GC septa are used, due to the wider 23- or 24-gauge needle of an SPME fibre assembly, compared to the traditional 26-gauge needles used for standard liquid GC injection. Septum coring problems become more significant when using automated methods, due to the higher throughput and the fact that the system is likely to be unattended for significant periods of time. This problem has traditionally been resolved by frequent septa replacements and subsequent liner cleaning procedures. However, this solution is not acceptable in the case of automated high-throughput methods and minimally supervised systems. Septum coring can also be minimised by using septa that are predrilled with an SPME needle prior to installation in the injection port or solved by using a septumless injection device.3,8 The Merlin Microsealt septum replacement is an example of a septum-free injector. It is equipped with two seals to provide (i) sealing of the syringe/SPME needle during the injection and (ii) closure of the inlet while the separation procedure takes place. Septum replacement kits are commercially available for most of the commonly used types of injectors. The main advantages of the septum replacement kits are the non-coring nature, durability and wide range of applicable operation temperatures and pressures. The lifetime of the Merlin Microseal can range from 1,000 to over 10,000 injections, depending on the sample type and operating conditions during the analysis.9 Gerstel, Inc., also offers a septumless head (SLH), which is used in conjunction with the CIS 4 programmable temperature vapourisation (PTV) inlet (Figure 5.2). The SLH offers the same advantages of the Merlin Microseal system in terms of the non-coring nature of injection, durability and wide range of applicable temperatures and pressures. The SLH/CIS 4 combination is also compatible with both standard fused silica and metal SPME fibres. CIS also allows the fibre to be inserted into a cooled inlet, which is subsequently heated rapidly, while a special narrow bore liner provides a higher linear velocity gas flow around the fibre coating to allow more efficient analyte desorption. Fixing cap Teflon needle guide Kalrez O-ring

Plunger

O-ring Spring

Figure 5.2 Septumless head used in conjunction with CIS 4 from Gerstel, Inc.

138

Handbook of Solid Phase Microextraction

5.1.1.2 Sample Agitation The four most frequently used methods for sample agitation in automated fibre SPME are fibre vibration (Varian autosampler), tray rotation (CombiPAL, MPS 2 and CONCEPT) and tray rocking motion (TriPlus). By contrast, in manual fibre SPME operations, magnetic stirring is the most common technique. The advantage of vibration, rotation and rocking shaking techniques over magnetic stirring lies in the fact that a foreign object is not added to the sample. This saves sample preparation time because the stirrers do not have to be added manually to the sample vials and also prevents any chance of sample contamination, if old stirrer bars are cleaned and reused. The fibre vibration technique, while more efficient than magnetic stirring in small (2 mL) vials, was not particularly effective in larger sample volumes.10 The method also puts stress on the fibre needle.11 Similarly, the rotating tray agitation mechanism causes stress on the fibre needle through the rotating action, and agitation is therefore restricted to speeds ranging from 250 to 750 rpm with the CombiPAL, MPS 2 and CONCEPT systems. If 24-gauge needles are used, the stirring speed often must be less than 750 rpm to avoid breakage of the needle due to stress. This problem appears to have been overcome by the use of the 23-gauge needles available for automated operations. Stirring restrictions can result in slower equilibration times and lower amounts of analytes extracted in pre-equilibrium conditions than would otherwise be obtained.12 One paper noted that the rotating tray agitation technique in headspace sampling resulted in splashing of the fibre, which decreased the fibre’s lifetime.12 Since 2002, a commercial magnetic stirrer module is available as an option for the CombiPAL autosampler,12 and this option also became available in combination with the MPS 2 and CONCEPT autosamplers. Magnetic stirrers that are not controlled by the software can also be used in conjunction with this unit.13 However, the use of a magnetic stirrer bar may introduce contaminants to the sample.

5.1.1.3 New Superelastic Metal Fibre Assemblies One way to improve the robustness and sample throughput of an automated SPME method is to increase the number of analyses that can be performed with a single SPME assembly. For the fibre SPME approach, Supelco recently developed a new generation of superelastic metal fibre assemblies, as shown in Figure 5.3 (see also Section 4.3.1 in Chapter 4). Figure 5.3 New superelastic metal fibre assemblies.

Automated SPME Systems

139

The fibre needle, plunger and fibre core are made of a special type of inert, flexible alloy that makes the assemblies more robust and able to perform several hundred analyses in succession. The fibre tip is of a bevelled design so it can pierce the sample vial septum more effectively. The alloy used in manufacturing these fibre assemblies has excellent shape memory properties and tensile strength; therefore, the fibre remains straight even after several hundred injections. Most of the commercial fibre coatings presently sold attached to the fused silica cores are available also in the superelastic arrangement. Since the fibre’s thin-wall needle size is 23-gauge, septum coring may occur during the injection procedure. Therefore, the use of septum-free injectors, such as the Merlin Microseal, SLH or similar septumless sealing systems, in combination with these new assemblies is strongly recommended. Another potential problem when metal fibres are used in conjunction with automation may arise from the shape of the needles. If there is a slight misalignment between the needle position and the hole in the injection nut, the needle will catch and bend. This can be addressed by doing regular maintenance; that is, changing the driving belt in the autosampler at least once a year to ensure accurate positioning of the needle, in addition to doing frequent alignment checks. An evaluation of superelastic fibre assemblies in SPME GC TOF MS was recently performed at the Department of Chemistry at the University of Waterloo as part of a collaborative project with Supelco.14 The superelastic PDMS fibre with 100 µm coating thickness was shown to have highly reproducible extraction properties for all the volatile compounds studied; that is, benzene, toluene, pyridine, 1-pentanone and 1-nitropropane. The inter-fibre comparison for the 100-µm superelastic PDMS was also very promising. As proof of durability, constant abundances of the target analytes were obtained even after more than 600 injections using the 100-µm superelastic PDMS fibre assembly (Figure 5.4). In addition to PDMS, superelastic fibre assemblies coated with other coating types were tested as well, and the results of these studies demonstrate the following: (i) reproducibilities expressed as standard deviations (n 5 7) were less than 5% in most cases in single-fibre reproducibility studies and (ii) the differences in performance were statistically insignificant in most cases and statistically significant for some types of fibre coatings between fibre reproducibility studies. However, as far as the latter point is concerned, it is worth noting that the performance similarity of the metal fibre assemblies is not as critical as in the case of silica-based assemblies due to the high durability characteristics of metal fibres. A very slight decreasing trend in peak areas of the volatile target analytes, caused by leaking of the vial caps (probably through the septum), was observed (Figure 5.5). An aluminium barrier disk placed on top of the septa can eliminate this effect.

5.1.1.4 Injector Liners Splitless direct injection (DI) glass liners (also called SPI liners or Unilinerss) significantly improve the transfer of the sample onto the chromatographic column.15 DI liners are, to a certain extent, able to simulate on-column injection. Several types of DI liners, with various internal diameters, are available. The

140

Handbook of Solid Phase Microextraction

40,000,000 35,000,000 30,000,000 2-Pentanone 1-Nitropropane Pyridine Benzene/3

Peak areas

25,000,000 20,000,000 15,000,000 10,000,000 5,000,000 0 0

20

40

60

80 100 120 Number of injections

140

160

180

200

Figure 5.4 Results of the 100-µm superelastic PDMS fibre durability study, 200 injections performed (peak areas of benzene divided by 3). (Source: Reprinted with permission from Ref. 5. Copyright Wiley-VCH Verlag GmBH & Co. KGaA.)

6

Standard deviations, various vial freshness, crimp vials, fibre F9 Spiked-closed tray after 1 day Closed-spiked tray after 1 day Spiked-closed fridge after 1 day Spiked-closed fresh

Crimp vial

6

Peak areas

4

(Standard deviations, n=5)

2

0 Standard deviations, various vial freshness, screw vials, fibre F9 benzene

Spiked-closed tray after 1 day Closed-spiked tray after 1 day Spiked-closed fridge after 1 day Spiked-closed fresh

2-pentanone

1-Nitropropane

Pyridine

Toluene

Screw vial

Peak areas

4

2

0 Benzene

2-Pentanone 1-Nitropropane

Pyridine

Toluene

Figure 5.5 A decreasing trend in peak areas of the volatile target analytes, caused by leaking of the vial caps (probably through septum).

Automated SPME Systems

141

internal diameter of the DI liners used for SPME applications should be similar to that of the SPME injection sleeve (typically 0.75 1 mm ID). The new superelastic fibre assemblies, used in conjunction with a septumless injection port and low-volume DI liners, seem to be a very powerful combination, capable of significantly increasing the number of SPME applications in various scientific fields of interest. The long-lasting fibre assembly is of great importance (e.g. in classification studies, requiring analysis of hundreds of individual samples for statistical evaluation of the data or in ‘fingerprint’ chromatographic comparisons). A further improvement to the system by the desorption of analytes in a more focused band could be obtained by the development of an automated micro-PTV injector configuration, as described by Gorecki and Pawliszyn.16

5.1.2

Commercial Autosamplers for GC

The major challenge in designing an SPME autosampler is to incorporate agitation and temperature control, as well as other enhancements, such as fibre internal cooling or dedicated injectors. One improvement is an SPME system that incorporates an agitation mechanism consisting of a small motor and a cam to vibrate the needle. The fibre in this design works as a stirrer. The vibration causes the vial to shake and the fibre to move with respect to the solution; the result is a substantial decrease of equilibration times compared to a static system. This mode of agitation simplifies fibre handling because it does not require the introduction of foreign objects into the sample prior to extraction. Varian has developed an SPME autosampler based on their 8000 GC autosampler system, taking advantage of the fact that the SPME device is analogous to a syringe in its operation and that, after desorption, the coating is cleaned and ready for reuse.1 More recently, Varian adapted their CP-8400 autosampler for SPME, but this system does not have controlled heating/cooling or agitation of the sample. It extracts samples directly in the sample tray, with needle vibration for agitation during extraction. As already mentioned, CTC Analytics incorporated SPME sampling on its CombiPAL autosampler (see Figure 5.1). CombiPAL is a robotic system with a great deal of flexibility for programming SPME analyses. Samples are loaded onto trays accommodating five vial sizes, and samples are heated and agitated during extraction using a separate sample preparation chamber (agitator tray/sample preparation station). To facilitate agitation, the extraction chamber is rotated at a programmable rotation speed during extraction. Additional vials/stations are present to accommodate wash solutions, derivatising agents, temperature control, derivatisation and fibre conditioning in order to facilitate operation of SPME under optimal conditions. While the built-in software can be used to perform basic SPME analyses, extra programming flexibility is provided by the Cycle Composert software, available as an accessory. A number of companies sell autosamplers with the same CTC ‘backbone’. One such company is Gerstel, Inc. (see Figure 5.6). The Gerstel autosamplers can do standard liquid injections, static headspace injections and SPME, but they have the

142

Handbook of Solid Phase Microextraction

Figure 5.6 Gerstel, Inc., MPS 2 autosampler.

added advantage of other extraction techniques’ automation. Some additional options available from Gerstel include the automated tube exchange system (ATEX), thermal desorption system (TDS), automated solid-phase extraction (SPE), automated dynamic headspace (DHS) and the twister thermal desorption unit (TDU). The TDU in particular was designed for exchange of liners containing Twister stir bars between a sampling tray and the thermal desorption unit. This setup has been adapted for use with thin-film microextraction (TFME; see Section 3.2.4 in Chapter 3) as well. PTV-type universal inlet from Gerstel with an SLH can be used with most GC instruments. It can be used for split, splitless, on-column and even large-volume injections (up to 1 mL). Either liquid nitrogen or Peltier cooling system allows for fast cooling of the injector. The syringe holder can be adjusted to hold 1.2 1,000 µL syringes, which eliminates the need to buy different syringe holders for each size of syringe. The heated syringe holder can accommodate both 1- and 2.5-mL syringes. A definite advantage of the Gerstel autosampler is the simple ‘fill-in-the-blank’ SPME-dedicated software. The earlier version of the Master software is very straightforward and easy to follow, and the new Maestro software gives more room for adjusting parameters while retaining the ease of use. In addition, the Gerstel software control provides additional functionality for all the standard injection modes: liquid injection, static headspace as well as SPME. All modes of sample introduction can use the PrepAhead feature, letting the software overlap sample preparation wherever possible, enhancing sample throughput. The software comes with a set of Prep functions that can be used to automate more complex sample preparation steps. New generation TriPlus Autosampler (Figure 5.7) from Thermo Fisher Scientific can also perform liquid injections, headspace analysis and SPME. All

Automated SPME Systems

143

Figure 5.7 TriPlus Autosampler, Thermo Fisher Scientific.

TriPlus movements are step controlled, thus providing reliable operation for thousands of samples. In addition, step-controlled movements allow for self-recognition of all components and automatic alignment procedures. The TriPlus autosampler also provides the capability to inject on two separate GC systems, enabling excellent productivity and high-throughput sample analysis. TriPlus autosampler is compatible with automated sample preparation and sample introduction procedures. All the critical steps involved in the SPME procedure can be successfully accomplished, including sample incubation and extraction, fibre cleaning procedures in a fibre conditioning station, good temperature control in the agitator tray during the agitation processes and successful fibre introduction into the GC inlet. The CONCEPT autosampler from PAS Technology, shown in Figure 5.12 for automated high-throughput solid-phase microextraction liquid chromatography (SPME LC) analysis, can also be used to perform SPME GC analysis. Recently, Cromline Srl. started marketing an SPME multi-fibre system, a kit that can be used in conjunction with the CTC CombiPAL and allows the exchange of fibres during a sequence. The system is suitable for series of samples requiring different coatings for optimal results. It also makes it extremely convenient to choose the best suited coating for a particular application, as method optimisation experiments can be performed in an automated fashion.

5.1.3

Fibre Conditioners

New SPME fibres require initial conditioning, at manufacturer-recommended temperatures ranging from 210 C to 320 C, for time periods ranging from 0.5 to 4 h. Conditioning is also recommended at the beginning of the workday and between runs in cases where analyte carryover is a possibility. Many SPME FastGC applications, such as field sampling, require additional fibre cleaning in order to

144

Handbook of Solid Phase Microextraction

reduce desorption time in the GC injector and eliminate carryover of non-target analytes.17 Typically, fibre conditioning is performed by desorption in a temperaturecontrolled GC injector for the recommended conditioning time. However, this procedure reduces available GC time by occupying an injector, and in many cases, it prevents analytical work from being performed on the instrument. It may also load the GC column with unwanted products of desorption, which in turn may require additional column conditioning and column ‘blank’ determinations. To address these concerns, a stand-alone SPME fibre conditioning station was designed, built and tested.18 The fibre conditioning station allows fast conditioning of SPME fibres, using high temperatures and gas flow for the desorption and subsequent purging of fibre contaminants. The device, intended for laboratory and field sampling applications, was based on a modified commercial syringe cleaner. The performance of the new fibre conditioning station was tested for several types of commercially available SPME fibres and compared with the traditional GC-injector fibre conditioning method. The new device performed as well as or better than GC injectors, for both new fibre conditioning and the desorption of n-alkanes representing a wide range of boiling points. In addition, the use of the fibre conditioning station is more cost-effecctive.

5.1.4

In-Tube and Other SPME Configurations Coupled with GC

The fibre configuration is not the only form of SPME that has been successfully automated for use with GC. An in-tube device in which the extraction phase is coated onto the inside of a needle has also been automated using the CTC autosampler.19 23 This device has been commercially marketed since 2000 under the name ‘solid-phase dynamic extraction (SPDE)’.21 The device is assembled onto a gastight syringe and, by repeated pulling and depression of the plunger, a sample is drawn past the extraction phase. For an effective desorption of analytes into the GC, the gas phase must flow past the extraction phase and into the injection port. This is achieved by using nitrogen, either drawn into the syringe immediately prior to the injection or added through an inlet in the side of the syringe. Greater robustness of the device, a larger volume of extraction phase and a greater surface-to-volume ratio compared to standard fibre configuration of SPME have been quoted as advantages of the technique. In terms of robustness, the use of one needle for 200 or more analyses was reported.19,21,23 For comparison, the traditional (fused silica) SPME fibres are typically anticipated to last for about 100 analyses,19,21,24,25 regardless of whether manual or automated techniques are used. This has relevance for automation because a longer lifetime will allow the instrument to run unattended for longer periods of time. One limitation of the technique is that it requires magnetic stirring for agitation because the needles cannot be bent, as sometimes is encountered when using the rotating tray device of the CombiPAL autosampler19 or the fibre vibration technique. Method development is similar to

Automated SPME Systems

145

that for fibre SPME techniques (Chapter 7). However, the in-tube configuration requires additional optimisation of the number of plunger strokes, plunger speed aspiration volume and desorption flow rate conditions. Due to the complexity of the technique and the large number of precise plunger strokes required, this method is better suited to automated methodology and would be difficult to perform in manual mode. An alternative automated in-tube configuration coupled with GC was reported for the direct analysis of water samples.26 This technique involved the use of a switching valve, a suction pump and an auxiliary gas source to draw the liquid sample past the extraction phase and, respectively, to desorb the extracted analytes onto the head of the GC column. The technique was applied to the analysis of polycyclic aromatic hydrocarbons (PAHs) and pesticides in water samples. Another approach uses a plug of stationary phase inserted into the main body of the syringe instead of the needle.27 The automation of this procedure was achieved, again using the PAL system. The sequence of the automated method is similar to SPDE except that, in the reported application, after the loading cycles, the extraction phase was flushed with water and the analytes desorbed with methanol into the GC injector. This technique was applied to the analysis of local anaesthetics in plasma samples. Advantages over fibre SPME were identified as greater robustness, reduced matrix effects and higher extraction recovery.

5.1.5

Automation of Procedures Involving In-Fibre Derivatisation

The flexibility of currently available technology allows automation of more complicated SPME procedures, such as those involving derivatisation. Automated in-fibre (or, more accurately, in-coating) derivatisation methods have been reported, for example, in the analysis of aldehydes with pentafluorobenzyl hydroxylamine (PFBHA) derivatisation,28,29 cannabinoids with N-methyl-N-trimethylsilyltrifluoroacetamide (MSTFA) derivatisation30,31 and carboxylic acids with pyrenyldiazomethane (PDAM)32,33 derivatisation. The in-tube SPDE approach also has been used for in-coating derivatisation applications.19,21,23 Because of the additional complexity and time costs typically involved in derivatisation methods, the throughput and reproducibility advantages of automation are even more significant than for simpler methodology. Automation allows the use of this approach on a routine basis when it would generally not be feasible by manual means. The automation of previously reported manual in-fibre derivatisation techniques can also highlight difficulties previously not observed or not fully investigated because of the lower throughput possible using manual techniques. For example, difficulties were observed during the automation of an existing procedure for PDAM-based, in-fibre derivatisation of carboxylic acids. As the number of injections increased, the fibre coating was affected and the capacity of the system to derivatise the acids declined, sometimes shortening the lifetime of the fibre to as few as 20 cycles.33 This prevented the system from working without user intervention for extended periods of time.

146

Handbook of Solid Phase Microextraction

A practical aspect to be considered when performing SPME derivatisation is depletion of the derivatisation reagent from the derivatisation reagent vial. This can be a significant issue, particularly with a high-throughput automated system. It was recently shown that, with PFBHA derivatisation, the loading on the fibre after 50 injection cycles had decreased by 30% due to depletion from the single vial used.28 The use of a new derivatisation reagent vial for each sample, as done in the MSTFA procedure, successfully solved this problem, but this approach may not be practical for all methods because it increases the number of vials involved.30 The reproducible loading of the derivatisation reagent onto the fibre may also be important in terms of the derivatisation reaction. The use of an automated technique is ideally suited to minimise the uncertainty in the loading and subsequent steps, as long as reagent depletion issues are removed for successive runs.

5.1.6

Other Automated SPME GC Procedures

5.1.6.1 Multiple SPME Extractions An automated approach to improve detection limits without derivatisation was developed in which multiple SPME extractions from a sample vial are concentrated in the GC injector prior to analysis.13 In order to facilitate this procedure and prevent peak splitting, the flow rate of the GC was programmed to 0.1 mL/min during each extraction and 0.3 mL/min during the desorption steps, with normal flows used only during the analysis step. The technique was applied to the analysis of pesticides in water and improved detection limits by 3 10 times compared to a single extraction. Sakamoto and Tsutsumi classified a series of pesticides into groups based on their optimum extraction temperatures by Headspace Solid Phase Microextraction (HS-SPME) GC.34 To improve detection parameters, they prepared multiple vials of each sample and performed automated analysis of each vial at a different temperature to analyse each group. This would be significantly more time consuming if performed manually.

5.1.6.2 Automated Internally Cooled Fibre SPME An automated internally cooled fibre SPME was developed by Chen and Pawliszyn.35 Internally cooled fibre SPME is a very powerful approach to SPME analysis, first demonstrated by Zhang and Pawliszyn in 1995.36 In this technique, a cooled gas or liquid passes the inside of the fibre during sampling. This cools the extraction phase and thus allows an increase in the percentage of analytes that can be extracted from a sample, in some cases enabling quantitative extraction. It can also allow the analyst to heat a sample to much higher temperatures than normally achievable with SPME, without reducing the distribution constant and equilibrium fibre loading. The new automated device involved the miniaturisation of the technology initially used to demonstrate the principle of the method. The fibre was contained in an 18-gauge needle, to accommodate the additional parts of the device. However, the internally cooled fibre could perform only a limited number of

Automated SPME Systems

147

injections before the GC septum needed to be replaced. Fibre failure was not found to be an issue in test experiments. The device could be mounted and used on the CombiPAL autosampler with only small changes to the GC system required, such as enlarging the injector nut hole and support. The temperature of the fibre coating was maintained using a temperature controller, a solenoid valve and tubing of different inner diameters. With this system, the temperature of the fibre coating could be controlled to within 5 C of the preset value. Full automation of the process was achieved by coupling the external temperature control system with the autosampler through a logic circuit built into the temperature controller. This allowed the controller to be turned on or off as required.

5.1.6.3 Automated SPME Flow-Through Samplers The first automated SPME flow-through sampler was reported by Eisert and Levsen.37 The addition of input and output connections to the GC autosampler vial allows the system to be used to monitor flowing streams continuously. A flowthrough cell for the analysis of liquid samples was mounted into a slot on the sample carousel of a Varian autosampler. This flow-through design facilitates agitation of the sample. The device was later modified to allow the sampling of gases.38 Grote et al.39 developed a stopped flow device for automated analysis of flowing streams that could treat water samples (for pH, internal standard and salt) and perform the SPME extraction and GC analysis. This device was used for the analysis of organic compounds in an industrial wastewater stream. The system could function for a week without any user intervention. The limiting factor was the reduced lifetime of the fibre, which was caused by soiling due to the wastewater matrix. Memory effects were also an issue, and a headspace version of the device was reported.40 This device could be used for significantly longer periods of time without user intervention and displayed less memory effects due to the replacement of plastic tubing with glass.

5.1.6.4 Custom Modifications of Commercial Devices Alternatively, when a connection is added directly to the needle of the autosampler syringe, the system can analyse samples present in the vial without the need to expose the fibre. This modified approach, which would facilitate simplified automated sampling, can be designed to rely on air pressure to push the sample through the needle. Then the fibre containing the extracted analytes in its coating is introduced to the instrument for desorption.

5.1.6.5 Internal Standard Loading A new approach, recently introduced in SPME, involves loading an internal standard onto the fibre coating prior to extraction instead of spiking directly into the sample.41 The standard, preloaded onto the extraction phase prior to the extraction step, partially desorbs to the sample matrix during sampling, and the amount lost can be used as a means of calibration. This method is based on the fact that, in

148

Handbook of Solid Phase Microextraction

SPME, the components of the sample are not exhaustively removed from the sample matrix, but rather, equilibrium is established between the extraction phase and the sample matrix. The main advantage of this approach is that the standard spiking step is eliminated, saving time and reducing the cost of analysis. The internal standard approach is particularly advantageous in cases such as field or in vivo applications, where an internal standard cannot be spiked directly into the sample. It also allows calibration under non-uniform agitation conditions. Theoretical aspects and example applications of this calibration method are described in Chapter 6. In order to maximise reproducibility, the internal standard loading step and extraction procedures should be fully automated.42,43 To obtain a reproducible, workable loading of the internal standard on the fibre, a loading solution of the internal standard in vacuum pump oil is usually used.

5.1.6.6 Dual-Arm Sample Preparation A further dimension to automated SPME analysis has been facilitated by the development of a dual-arm PAL autosampler that can perform sample preparation steps prior to an SPME extraction.44 46 The device consists of two CombiPAL autosamplers stacked one on top of the other (Figure 5.8). One robotic arm is fitted with a syringe, while the other has an SPME fibre installed. This approach has been

Figure 5.8 Diagram of the dual-arm PAL.44 1 sample preparation station/agitator tray 1; 2 needle heater/fibre conditioning station; 3 cooled sample tray; 4 sample tray; 5 fast wash station; 6 GC injector port; 7 sample preparation station/agitator tray 2. The Prep PAL contains a syringe, and the Inject PAL contains an SPME fibre. (Source: Reprinted with permission from Elsevier.)

Automated SPME Systems

149

used for in-vial derivatisation of organometallic compounds using sodium tetraethylborate in aqueous samples, followed by SPME extraction of the derivatives.44 The system was able to fully automate the entire procedure, which involved the addition/mixing of buffer and derivatisation reagent, derivatisation, and finally, SPME extraction. The availability of a cooled tray with the autosampler was beneficial because it increased the lifespan of the derivatisation reagent solution. The system was also able to prepare diluted standards from a stock solution. With ‘Cruise Control’ software that was specifically designed for use with this dual-arm system, it was possible to use the arms simultaneously without their ever colliding. This enabled the maximum sample throughput because there was no time wasted waiting for one arm to finish its tasks prior to the second starting, as highlighted in Figure 5.9. For this application, the reduction in analysis time was 24% compared to the situation when the arms could only be moved one at a time, as was the case with the original CombiPAL software. A precision comparison of the automated and manual methods showed a clear improvement with the automated technique. The dual-arm system was also used to automate the derivatisation of 2-chlorovinylarsonous acid in urine prior to SPME analysis, but in this case, the buffer and internal standard were added manually.45 A method for the determination of ethyl carbamate in alcoholic beverages used one arm of the dual system to deliver a salt solution to the samples prior to SPME extraction, which was then performed by the second.46

Figure 5.9 Time chart of the automated system in Figure 5.8, programmed using Cruise Control software that maximises sample throughput by allowing simultaneous use of both arms.44 (Source: Reprinted with permission from Elsevier.)

150

5.2

Handbook of Solid Phase Microextraction

Automated SPME LC

The efforts to automate SPME have focused primarily on GC instrumentation, due to the excellent efficiency that can be obtained using thermal desorption of the SPME fibre in the GC injection port.5 By contrast, there are relatively few automated SPME LC methods, due to the difficulty of interfacing the commercial fibre format with the LC instrument. Although SPME LC methods are typically very sensitive, accurate and precise (see Section 7.5 in Chapter 7), their widespread use in laboratories that routinely need to analyse large numbers of similar samples is currently limited. The need to automate SPME LC analyses was initially addressed through the development of in-tube SPME.47 This approach can provide full automation of SPME LC using existing commercial HPLC autosamplers. This makes in-tube SPME easy to implement, but the overall sample throughput is limited by the serial processing of the samples and the fact that mass-transfer kinetics are slow in the liquid phase. Within the last year, a novel SPME LC robotic platform based on the parallel extraction and desorption of multiple samples on a multi-well plate format has emerged.48 This new platform addresses the need for automation and the need for high throughput simultaneously. In fact, it can provide the highest SPME throughput to date, as more than 1,000 samples can be processed per day if the unit is operated continuously. It is easily interfaced with common LC instrumentation, and its compatibility with LC MS/MS has also been demonstrated. This advance is expected to have a major impact on SPME LC applications in the upcoming years.

5.2.1

Automated In-Tube SPME LC

The development of an automated in-tube SPME approach by Eisert and Pawliszyn in 1997 has offered a promising alternative for automated LC (Figure 5.10).47 In this method, a commercial HPLC autosampler is modified by replacing a section of tubing or the injection loop with a capillary containing the extractive material. The sample is drawn into and out of the extraction capillary a number of times until equilibrium is reached or the desired level of extraction is obtained. The system is then switched to desorption conditions, the analytes are swept onto the analytical column using a small volume of solvent or mobile phase, and the chromatographic analysis follows. It is important to emphasise that this approach, suitable for the analysis of very small samples, also offers convenient interfacing to micro-HPLC instrumentation. Figure 5.10 illustrates the incorporation of in-tube SPME on a modified Spark Holland micro-LC autosampler.47 Subsequently, in-tube SPME was successfully implemented on a variety of commercial HPLC autosamplers. These different practical configurations are illustrated in Figure 5.11.49 In in-tube SPME, the extraction capillary can simply be a section of commercially available coated fused silica GC columns. The most commonly used capillaries have coatings similar to common commercially available SPME fibres, such as polyethylene glycol (PEG) or porous divinylbenzene (PS DVB).

Automated SPME Systems

151

1

2

Dispenser

HPLC pump

Loop Buffer tubing

3

6

Syringe

Analytical column

4

5

Six-port valve

Waste Wash solvent

SPME capillary

Inject position Load position

Adjustable capillary guide/depth gauge

Adjustable needle guide/depth gauge Vial retainer arm Solvent vial

Septum piercing needle Sample vial

Figure 5.10 SPME with a modified needle to allow in-needle extraction and automated flow-through analysis of small samples and automation of a coated tubing solid-phase microextractor using the Spark Holland Autosampler.

One significant difference between in-tube SPME and manual fibre-based SPME HPLC is the decoupling of desorption and injection that is possible with the in-tube method. In the fibre-based method, analytes are desorbed during injection, as the mobile phase passes over the fibre. With in-tube SPME, the analytes are desorbed either by mobile phase flow or by aspirating a desorption solvent of choice from a second vial, and then, later, by transferring the solvent with desorbed analytes to the injection loop for injection onto the column. While fibre-based HPLC has the advantage of eliminating the need to filter cloudy samples, the method does suffer peak broadening in applications where analytes are slow to desorb from the fibre to the mobile phase due to coating thickness. With in-tube SPME, analytes are completely desorbed prior to injection, so peak broadening is less of a factor. The coating thickness is only a fraction of a micron, so desorption is fast. If analytes are sufficiently solvated by the mobile phase, there is no need to use additional solvent for desorption. The theory of extraction and desorption by this method was investigated and validated with the automated analysis of six phenylurea pesticides in water.47 For the analysis of phenylurea pesticides, a 60-cm section of Carbowax GC capillary (0.25 mm ID, 0.25 µm film thickness) was used for extraction, with a total of 10 sample aspirate/dispense steps, with subsequent desorption into 38-µL methanol. For extraction, the sample was aspirated and dispensed from the capillary a number

152

Handbook of Solid Phase Microextraction HPLC pump

HPLC pump 2

2

1

1

Buffer tubing

Buffer tubing 6

3

Column

Column

Loop

Loop 4

6

3

5

4

5

Six-port valve

Six-port valve

(A) Transfer line in-tube SPME Sample vial

Sample vial In-tube SPME capillary

(B) Standard injection loop in-tube SPME

Column connector Injection loop

Autosampler

Metering pump Injection needle Six-port valve From pump To column

(C) Flush loop in-tube SPME

Figure 5.11 Automated in-tube SPME configurations.49 (Source: Reprinted with permission from Elsevier.)

of times, preferably until equilibrium was nearly achieved. It was expected that the number of aspirate/dispense steps required to reach equilibrium would increase with the K-values of the analytes. Experimentally, equilibrium extraction was not obtained for any of the analytes. It was reasoned that during each dispense step, analyte would at least partially desorb into the mobile phase that follows the sample in the capillary, thus complicating the extraction. An aspirate/dispense rate of 50 100 µL/min was found to be optimal for extraction and desorption. Below this level, aspiration and desorption required an inconveniently long time, and above this level, bubbles formed on the inside of the capillary, reducing extraction/desorption efficiency. Method precision varied between 1.6% and 5.6% standard deviation. The method was linear over the range of 10 10,000 µg/L, with the limit of detection below 5 µg/L. In-tube SPME was used to analyse the thermally labile carbamate pesticides in water samples, using both the FAMOS autosampler from LC Packings and the Hewlett-Packard (HP, now Agilent) 1100 autosampler, in both cases without modification of the autosampler itself.50,51 Subsequently, this method also was transferred

Automated SPME Systems

153

to capillary LC using the HP1100 autosampler.52 Due to the relatively large effective injection volume of the in-tube SPME technique (30 45 µL), it was necessary to employ on-column focusing when interfacing to capillary LC in order to achieve good chromatographic efficiency. The use of capillary LC improved detection limits by a factor of 24 81 over conventional LC, indicating good potential of this technique for trace analysis. In-tube SPME can also be applied for the analysis of more complex samples such as urine,53 57 serum,53,54,56,57 saliva58 and homogenised tissues.59 For example, the in-tube SPME HPLC (fluorescence detection) method for the determination of telmisartan in rat tissue had detection limits of 0.24 1.8 ng/g, depending on the tissue examined (heart, liver and kidney). Method precision was better than 9% standard deviation at all concentrations and in all tissues examined, while accuracy was within 65% of nominal value. In-tube SPME was also successfully applied to food analysis (see also Section 9.2.2.7 in Chapter 9).60,61 For example, mutagenic heterocyclic amines were analysed by in-tube SPME.60 The amines extracted by the capillary column were desorbed by aspiration of 30 µL of methanol prior to injection, and carryover of heterocyclic amines was not observed. The detection limits (S/N 5 3) were 0.2 3.1 ng/mL and concentrations of heterocyclic amines in grilled beefsteak were successfully determined with the optimised method. A limitation of the automated in-tube SPME technique is that only particle-free samples can be analysed because it is very easy to block the tube. When using this technique for complex samples, such as biological fluids or food samples, additional sample dilution steps are typically required in order to achieve good method performance. Therefore, this approach does substantially reduce human intervention, but the addition of sample filtration, centrifugation and/or dilution step(s) is required for most applications. In-tube SPME also does not allow for simultaneous preparation of many similar samples because all calibration standards and samples are prepared and analysed serially. Method development strategies for in-tube SPME are discussed in Chapter 7. In addition to references cited above, the technique has been applied to a number of environmental, clinical, forensic and food applications. For additional information on possible applications of in-tube SPME, the reader is directed to recent reviews by Kataoka.62,63

5.2.2

High-Throughput SPME LC Analysis Using the Concept 96 Autosampler

The extraction and desorption of analytes during SPME LC are typically the most time-consuming events of the analytical process. Many biological and environmental applications generate numerous samples for analysis and the total analysis time may prove to be impractical when these samples are analysed in a linear sequence (as is the case for in-tube SPME). Therefore, it is more efficient to perform parallel extraction and desorption steps of multiple samples on a multi-well plate format. The 96-, 384- and 1,536-well plates can be purchased commercially, and these are

154

Handbook of Solid Phase Microextraction

compatible with the small fibres used in SPME. After the extraction of analytes from the sample onto the fibre, the analytes can be desorbed from the fibre using liquid desorption, followed by injection into the LC instrument. This procedure was recently fully automated in the 96-well plate format by PAS Technology (Magdala, Germany) and is offered commercially as the Concept 96 robotic sample preparation station.48 The design of Concept 96 was based on the results of a preliminary study where several different device and agitation configurations were tested using the extraction of selected PAHs from water samples as a model system.64 A 96-fibre device based on a commercially available pin-tool replicator was found to be the best configuration for automation because of its robust construction and low cost. This initial device consisted of 96 pins that fit directly into the centre of the wells of commercially available multi-well plate. Furthermore, the extraction profiles for fluorene under various agitation conditions (static regime, magnetic stirring, orbital shaking and sonication) were determined. Orbital shaking at 900 rpm, while the fibres are held fixed, was found to provide the most uniform agitation in all wells, good enhancement of extraction rate and minimal evaporative losses.64 The Concept 96 robotic system is shown in Figure 5.12. The main components of the system include three robotic arms, three orbital agitators, one wash station, one 96-fibre (or 96-thin-film) device and one 96-well nitrogen blowdown device. Therefore, the Concept 96 autosampler can perform the following sample preparation steps in a fully automated fashion using Concept software: preconditioning, SPME extraction, fibre rinsing, solvent desorption, agitation, addition of internal standard, solvent evaporation and solvent reconstitution (see also Chapter 13). The system can be further customised based on the user’s exact requirements. Initial Figure 5.12 Concept 96 robotic station from PAS Technology. Main components include (A) a 96-SPME device, (B) a 96-well nitrogen blowdown device, (C) a syringe arm, (D) three orbital agitators and (E) wash station.

Automated SPME Systems

155

prototype designs of Concept 96 also included HPLC injection port, which allowed the direct interfacing of the unit with LC MS/MS instrumentation for online operation.48 However, significantly greater sample throughput is achievable if the unit is operated as a sample preparation station, with a separate conventional HPLC autosampler used to perform LC injection from a multi-well plate. Because the majority of modern HPLC autosamplers are capable of performing sample injections from the multi-well plate format, this feature was omitted from the subsequent design of this unit. The only steps of the analysis that are not automated by Concept 96 are (i) the dispensing of sample, desorption solvent and preconditioning solutions in the multiwell plate and (ii) the transfer of the completed plate to the HPLC autosampler for the analysis. These steps can be performed manually by the user or using appropriate existing commercially available robotic stations and/or plate feeders. Figure 5.13A shows a close-up of the prototype 96-fibre device used during the evaluation experiments of the unit, while Figure 5.13B shows the final commercial device that uses thin-film geometry (TFME) rather than fibre geometry. The main reason for adopting thin-film geometry was to further increase the surface area and volume of the extraction phase.65 This increase of extraction phase volume results in increase in the amount extracted by the device and consequently in improved analytical sensitivity. This is illustrated in Figure 5.14 for the extraction of diazepam. The absolute recovery of diazepam from 1 mL standard solution increased from 30% for fibre geometry to 50% for thin-film geometry. With the use of other extraction phases (e.g. 3M Empore membrane disks), exhaustive extraction was achievable.66 Another advantage of thin-film configuration is that the increased surface area results in increased rates of extraction. Approximately twofold improvement in extraction rates of four benzodiazepines was observed with thin-film geometry.65 This allows the use of shorter extraction times and additional increases in sample throughput. Two methods of liquid desorption using multi-well plates and Concept 96 can be employed. In one method, direct solvent desorption can be performed in a large

Figure 5.13 (A) Prototype 96-fibre SPME device shown with C18 coating on 0.061-in. stainless steel wire; (B) commercial 96-thin-film device.

Handbook of Solid Phase Microextraction

Amount diazepam Extracted (%)

156

100 80 60 40 20 0 RPA or C18 SPME fibre

C18 thin film

3M Empore membrane disks

Type of extraction phase

Figure 5.14 Comparison of automated SPME, automated TFME and 3M Empore membrane extraction in terms of extraction efficiency for the extraction of diazepam from 1-mL sample volume.

volume of desorption solvent (sufficient desorption solvent volume to immerse completely the entire length of the extraction phase). The composition of solvent is directly compatible with DI into LC MS/MS. In the second approach, direct solvent desorption is performed into a large well volume, followed by solvent evaporation to dryness and reconstitution into the minimum amount of solvent suitable for LC injection. Typically, non-volatile analytes are analysed by LC, hence analyte loss during the solvent evaporation step is not an issue. However, if semi-volatile analytes are targeted, an internal standard can be added to quantitate the loss of analyte due to evaporation. Obviously, the use of first approach provides higher sample throughput but significantly reduces analytical sensitivity due to dilution. However, the sensitivity achievable with this approach was still satisfactory for applications developed to date. For example, limits of quantitation (LOQ) for the analysis of ochratoxin A in urine was 0.7 ng/mL which was sufficient for monitoring of human exposure to this mycotoxin through food intake.67 This LOQ is comparable to SPE and liquid-liquid extraction (LLE) methods coupled to LC with fluorescence detection (FD) reported in literature, but not as sensitive as methods relying on immunoaffinity chromatography LC-FD.68 In addition to improved sensitivity, another advantage of the evaporation/reconstitution approach is the flexibility in the choice of solvent to use for desorption. The performance of the parallel extraction/desorption SPME approach is highly dependent on factors such as (i) uniformity of agitation in all the wells; (ii) fibre robustness, to withstand various agitation conditions and robotic manipulations and (iii) similar extraction capacity for all fibres.48 The results of the 96 parallel extractions of four PAHs are shown in Figure 5.15 to demonstrate the effect of analyte volatility on method performance.64 As expected, excellent results were obtained for less volatile analytes, such as fluoranthene because evaporative losses are minimised. SPME of more volatile analytes, such as naphthalene, exhibited high standard deviations because the plate dry technique, which includes an evaporation step, was used. This preliminary study was carried out using a PDMS-based coating, and GC MS analysis was used for the detection of the analytes. In subsequent studies, hollow-tubing PDMS

Automated SPME Systems

157

Moles extracted and desorbed into acetonitrile (n)

1.20E – 09

Naphthalene Fluorene Anthracene Fluoranthene

1.00E – 09

8.00E – 10

6.00E – 10

4.00E – 10

2.00E – 10

0.00E + 00 0

10

20

30

40

50

60

70

80

90

100

Analysis number

Figure 5.15 Amount extracted from 96 identical aqueous samples containing four PAHs at 200 µg/L, on the 96-well plate format, using the 96-well SPME replicator device.64 (Source: Reprinted with permission from Elsevier.)

coating proposed by Hutchinson et al.64 did not perform well for drug analysis due to long equilibration times (20 h for benzodiazepines) and poor intra-fibre reproducibility for repeated extractions.48 To address this issue, thinner (B5 6 µm thickness) coatings based on coated porous silica particles were proposed.48 These coatings had much shorter extraction times (30 min for benzodiazepines) and excellent intra- and inter-fibre reproducibility. Another type of coating that was found to have satisfactory performance is carbon tape coating.67 The main advantages of this coating are (i) commercial availability, (ii) ease of preparation, (iii) good extraction capacity towards more polar analytes and (iv) no preconditioning being required. The extraction efficiency of this coating was found to decrease for subsequent extractions when exposed to biological fluids such as urine. As such, this coating is recommended as a single-use coating. Although the abovementioned coatings were found to have good inter-fibre reproducibility (typically more than 10% standard deviation), it is still important to use an appropriate calibration method in order to compensate for small variations in the amount of extraction phase immobilised on each individual fibre. Internal standard calibration, fibre-constant calibration and standard-in-fibre (kinetic) calibration were evaluated to determine which methods are the most suitable to use for

158

Handbook of Solid Phase Microextraction

% Standard deviation (n = 12 fibres)

this particular application. The results are presented in Figure 5.16. Using the external calibration method, no corrections for the differences in the amount of extraction phase on each fibre were applied. As shown in Figure 5.16, the use of either internal standard calibration or fibre-constant calibration resulted in significant improvement of method precision, as indicated by approximately twofold improvement in standard deviation. Therefore, the use of either of these calibration methods is recommended during the development of high-throughput methods. Contrary to theoretical expectations, the use of standard-in-fibre approach did not improve precision in this study. One plausible reason for this is the possible use of suboptimal standard loading procedures. The ways to improve this procedure further are currently being investigated in our laboratory at the University of Waterloo. Automated multi-well SPME systems, such as Concept 96, have great potential for providing fast methods for high-throughput applications, such as pharmaceutical analysis and therapeutic drug monitoring. Automated SPME/TFME drastically increases sample throughput (to more than 1,000 samples per day) and is capable of handling very complex matrices, such as whole blood and tissue homogenates, with absolutely no sample pretreatment required. In contrast, traditional sample preparation methods such as SPE typically require removal of proteins, cells and particulates to ensure no disruption of the flow through the closed-bed cartridge. In addition, due to parallel extraction format and accurate and reproducible (softwarecontrolled) timing and fibre positioning within wells, the system can be used for pre-equilibrium SPME with no loss of precision. This is shown in Figure 5.17.48 For all time points except for 15 min, method precision was the same (B10% standard deviation) regardless of whether equilibrium or pre-equilibrium extraction time was employed. High standard deviation observed at the initial point was attributed to the fact that it requires some time to establish uniform mixing within all wells.48,65 Once uniform mixing is established, the shortest extraction time that gives sufficient sensitivity can be employed. Another significant advantage of the use of automated SPME/TFME for bioanalysis is the ability to obtain both free and total concentration of the analyte from 16 14 12 10 8 6 4 2 0 External calibration

Lorazepam as internal standard

Diazepam d-5 as internal standard

Standard on the fibre calibration

Fibre constant calibration

Type of calibration

Figure 5.16 Comparison of the ability of various calibration methods to compensate for inter-fibre variability. Results are shown for the extraction of diazepam using n 5 12 fibres.

% Standard deviation of amount extracted (n=96 fibres)

Automated SPME Systems

159

40 Equilibrium

30 20 10 0 15

60

120

330 840 960 1,200 1,440 1,800 Extraction time (min)

Figure 5.17 Effect of pre-equilibrium and equilibrium extraction times on the precision of the automated multi-fibre SPME method.

the same sample, depending on the calibration strategy employed (see Chapters 6 and 11). Traditional methods generally rely on the disruption of the binding between analyte and proteins and thus can provide only total concentration. If free concentration information is required, a different analytical methodology on a new aliquot of sample must be performed (e.g. ultrafiltration). Therefore, automated SPME/TFME can provide not only very high sample throughput (comparable to multi-well SPE and LLE and greater than online SPE) but also increase the amount of useful information acquired. To date, fully validated applications of this system according to the guidelines of the Food and Drug Administration (FDA) include the equilibrium analysis of benzodiazepines in whole blood48 and pre-equilibrium SPME determination of ochratoxin A in urine67 (see Table 7.6 for a summary of the results). Furthermore, the system can be used for automated ligand receptor binding studies, which are very important in drug discovery.69 A parallel desorption device to be used for desorption of in vivo SPME assemblies has also been proposed recently.70 This device allows for compact transportation of up to 96 in vivo probes, followed by easy desorption, simply by switching the top cover holding the SPME probes to a new plate containing the desorption solvent. Further automation of this procedure is currently underway in collaboration with PAS Technology. Range of coating chemistries were studied and evaluated with the aim of developing biocompatible, robust and reusable coatings for the 96-TFME system. When dealing with complex sample matrices, the adhesion of macro-molecules, such as particulate and proteins, to the coating surface can significantly influence the kinetics of the extraction and the amount of analyte extracted by the coating. Therefore, the biocompatible coating materials are preferred as the SPME coating. Polyacrylonitrile (PAN) is one of the important polymers that is widely used in biomedical areas, such as dialysis and ultrafiltration, because of its biocompatible characteristic. In addition to its biocompatibility, PAN provides high chemical and mechanical stability, which makes it suitable to be used as a binder for stationary phase immobilisation. Recently, PAN was used for the immobilisation of

160

Handbook of Solid Phase Microextraction

different extraction phases such as C18-silica-based, polar-modified polystyrene divinylbenzene (PAN PS DVB) and phenylboronic acid (PBA) particles on the surface of 96-TFME devices. Different methods of coating preparation, such as dipping, brush painting and spraying, were tested and evaluated, and the spraying method was found to be the most reliable method of coating preparation because of the physical stability and reusability of the sprayed coating. The C18-PAN 96-TFME system was used for the extraction of benzodiazepines from PBS, and it provided absolute recovery of 97% (6 5) for diazepam, 82% (64) for nordiazepam, 74% (6 6) for oxazepam and 75% (64) for lorazepam.71 In addition, the C18-PAN 96-TFME coating was tested in terms of stability and reusability in PBS and human plasma. The coating demonstrated good physical stability and reproducible recovery for at least 140 usages (standard deviation 5 4% for n 5 12 coatings for n 5 17 experiments) for the extraction of diazepam from PBS. Also, the evaluation of C18-PAN thin films for extraction from human plasma showed that extraction recovery of the coating was reproducible for at least 70 usages (9% standard deviation for n 5 12 coatings and n 5 14 experiments). Because of the complex biological matrix in plasma, there was a drop in the extraction recovery after 70th extraction, but the amount of extraction recovery was still reproducible after the 70th over 140th extractions (standard deviation 5 12% for n 5 8 experiments and n 5 12 coatings). As a result, by using the proper calibration technique the coating could also be reusable 140 times in plasma.71 Application of SPME in the different areas of clinical, biological, environmental and food studies raise the need for the development of different chemistries of SPME coatings, enabling simultaneous extraction of a variety of analytes with diverse polarities from a single sample. The studies showed that modified PAN PS DVB copolymer with a weak anion exchanger is a good SPME stationary phase, capable of extracting compounds from different chemical classes and of widely varying polarities (data not published yet). The PAN PS DVB 96-TFME coating was tested for the extraction of diazepam (log P 5 2.82), oxazepam (log P 5 2.24), caffeine (log P 5 20.07) and riboflavin (log P 5 21.46) from PBS, and the coating resulted in the exhaustive recovery for extraction of diazepam, oxazepam and caffeine (standard deviation , 5%), and a 71% recovery was achieved for the extraction of riboflavin (standard deviation 5 4%). The PAN PS DVB 96-TFME system was also tested in terms of stability and reusability, and it demonstrated reproducible extractions for over 150 usages from PBS (standard deviation 5 5% for n 5 12 coatings and n 5 7 experiments) and more than 100 usages from human plasma (standard deviation 5 14% for n 5 12 coatings and n 5 10 experiments). Another biocompatible coating for the 96-thin-film miroextraction system is PAN PBA, which has also showed good extraction recovery for a wide range of analytes. Specifically, it has been able to extract carbohydrates such as sucrose (4 6 0.4 absolute recovery), which are impossible to extract well with the other coatings because of high polarity (log P 5 23.7). The PBA coating also has the potential to be used for the extraction of nucleic acids and proteins by using the appropriate binding matrix (data not published yet).

Automated SPME Systems

5.3 5.3.1

161

Other Automated Configurations Involving SPME Automated SPME Coupling with Other Analytical Instruments

There are few reports of automated SPME coupling with other analytical instruments. Rodacy et al.72 proposed a prototype device for automated SPME sampling of a water flow stream, followed by desorption and analysis by ion-mobility spectrometry (IMS). Fibre SPME was used, with the flow path bound on either side by a PDMS septum at the sampling point. The SPME needle pierced the first septum, and then the fibre was exposed to allow the exposure of the extraction phase to the flow stream. For the desorption step, the extraction phase was retracted, and the needle pushed through the second septum, which placed the fibre in the injection port of the IMS. This work was part of a long-term project to develop a portable, unmanned, underwater sensor for detecting unexploded ordinance. Kobayashi et al.73 developed an in-tube SPME approach for the complexation and analysis of copper (II), using electrothermal atomic absorption spectrometry. Automated coupling of SPME with other analytical instrumentation, such as capillary electrophoresis and spectroscopic techniques, is anticipated to receive attention in the future.

5.3.2

Automation of SPME Formats Other than Fibre or In-Tube

Complete automation of SPME formats other than fibre or in-tube type configurations has not yet been achieved. Semi-automated methods have been reported for the configuration known as stir-bar sorptive extraction (SBSE),74,75 in which the extractive phase is coated onto a stirrer bar. To maximise sample throughput and minimise the analyst time required using this technique, a number of extractions can be performed concurrently using a multi-position magnetic stirrer. For GC analysis, the loaded stirrer bars are inserted into thermal desorption tubes and liners, which can be loaded and removed automatically from a GC injection port fitted with a thermal desorption unit and a temperature-programmable injector.75,76 However, the user needs to transfer the stirrer bar manually into the sample vial and from the sample vial to the thermal desorption tube tray.75 The latter step creates the possibility of losses or contamination during transfer, and also makes the technique, in its current form, unsuitable for the analysis of volatiles.75 The SBSE technique has also been semi-automated for LC procedures.77,78 In this embodiment of the technique, the loaded stir bar is placed manually into a 2-mL autosampler vial containing a 250-µL insert and then the vial is capped. An autosampler fitted with a syringe is then used to add desorption solvent, heat and agitate the vial and finally remove the solution and inject it into the chromatograph. The full automation of alternative SPME configurations, such as stir-bar-based extractions, would be beneficial because of their greater sensitivity.

162

Handbook of Solid Phase Microextraction

Overall, the percentage of SPME analyses conducted in an automated fashion is certain to increase in the coming years and, with this trend, the use of this sample preparation method will continue to grow.

5.3.3

NT Devices

The in-needle concept can be expanded to the trap in-needle system (see Section 3.3.2 in Chapter 3).79,80 This device is able to perform sampling, sample preparation and sample introduction conveniently in the analytical instrument. The needle-trap (NT) device can be used as an active sampler by drawing the gas or liquid sample through the needle or using a gastight syringe attached to the needle. This can be performed in exhaustive mode by carefully packing the trap and drawing a limited amount of the sample to prevent breakthrough, or it can be used more conveniently in equilibrium- or diffusion-based mode. This device can also perform as a passive sampler when the needle is not exposed to the investigated system directly, allowing components of the sample to diffuse into the needle. The components of the sample that are captured can be chemical compounds as well as matter Figure 5.18 (A) Schematic of a GC autosampler in the desorption mode for NT devices; (B) a more detailed schematic view of the junction of adapter and luer-lock head of the NT device. PTFE disk (B100 µm thickness) is placed in the adapter for sealing purposes.80 (Source: Reprinted with permission from American Chemical Society, r 2008.)

Automated SPME Systems

163

present in the sample. This is because the trap can be constructed either from quartz wool, to collect particulate matter and aerosols, or from a sorbent material, to extract chemicals from the sample matrix. Currently, the concept GC autosampler (PAS Technology) is being tested at the University of Waterloo laboratory for its capability for automating NT procedures. The design of the autosampler is shown in Figure 5.18. Luer-lock adapters are placed manually on Luer-lock NTs, and the assembled devices are then placed manually on the autosampler tray. During analysis, the magnetic arm of the autosampler transferred NTs one by one to the GC injector, while the closure (pneumatic) arm provided carrier gas flow through the luer-lock needle head, as shown in Figure 5.18B. Programmable Concept software started the GC injector and acquisition when appropriate. The use of this autosampler eliminated the injection errors associated with manual injections and also significantly reduced the time and labour of the analyst. The availability of such automated NT autosamplers for GC can facilitate the use of this technology in industrial hygiene applications.

Acknowledgements The authors thank John R. Stuff from Gerstel, Inc., for his input and contribution of Figure 5.2.

References 1. CL Arthur, LM Killam, KD Buchholz, J Pawliszyn & JR Berg, Anal Chem 64 (1992) 1960 2. JR Berg, Am Lab 25 (1993) 18 3. J Pawliszyn, Solid-Phase Microextraction: Theory and Practice (1997) Wiley: New York, NY 4. A Namera, M Yashiki, T Kojima & M Ueki, J Chromatogr Sci 40 (2002) 19 5. J O’Reilly, Q Wang, L Setkova, JP Hutchinson, Y Chen, HL Lord, CM Linton & J Pawliszyn, J Sep Sci 28 (2005) 2010 6. H Pham-Tuan, J Vercammen & P Sandra, LC GC Europe 14 (2001) 215 7. K Maes & PC Debergh, Plant Cell Tissue Org 75 (2003) 73 8. Supelco, Solid-Phase Microextraction Troubleshooting Guide, Bellefonte, PA, Bulletin 928 (2001) 9. Merlin Microseal High Pressure Septum Kit for HP Split/Splitless Capillary and Purged Packed Inlet Systems, Product Bulletin-Merlin Instrument, Half Moon Bay, CA, 1995 10. R Eisert & J Pawliszyn, J Chromatogr A 776 (1997) 293 11. J Pawliszyn, in J Pawliszyn, Ed, Sampling and Sample Preparation for Field and Laboratory (2002) Elsevier Science B.V.: Amsterdam: 389 12. C Kohlert, G Abel, E Schmid & M Veit, J Chromatogr B 767 (2002) 11 13. J Lipinski, Fresenius’ J Anal Chem 367 (2000) 445

164

Handbook of Solid Phase Microextraction

14. L Setkova, S Risticevic, CM Linton, G Ouyang, LM Bragg & J Pawliszyn, Anal Chim Acta 581 (2007) 221 15. G Ouyang, Y Chen, L Setkova & J Pawliszyn, J Chromatogr A 1097 (2005) 9 16. T Gorecki & J Pawliszyn, Anal Chem 67 (1995) 3265 17. JA Koziel, M Jia, A Khaled, J Noah & J Pawliszyn, Anal Chim Acta 400 (1999) 153 18. J Koziel, B Shurmer & J Pawliszyn, J High Resol Chromatogr 23 (2000) 343 19. F Musshoff, DW Lachenmeier, L Kroener & B Madea, Forensic Sci Int 133 (2003) 32 20. J Lipinski, Fresenius’ J Anal Chem 369 (2001) 57 21. DW Lachenmeier, L Kroener, F Musshoff & B Madea, Rapid Commun Mass Spectrom 17 (2003) 472 22. C Bicchi, C Cordero, E Liberto, P Rubiolo & B Sgorbini, J Chromatogr A 1024 (2004) 217 23. F Musshoff, DW Lachenmeier, L Kroener & B Madea, J Chromatogr A 958 (2002) 231 24. M Abdel-Rehim, Z Hassan, L Blomberg & M Hassan, Ther Drug Monit 25 (2003) 400 25. Supelco, Solid Phase Microextraction: Theory and Optimization of Conditions, Bellefonte, PA, Bulletin 923A (1999) 26. H Wang, W Liu & Y Guan, LC GC Europe 17 (2004) 144 27. M Abdel-Rehim, J Chromatogr B 801 (2004) 317 28. Q Wang, J O’Reilly & J Pawliszyn, J Chromatogr A 1071 (2005) 147 29. P Vesely, L Lusk, G Basarova, J Seabrooks & D Ryder, J Agric Food Chem 51 (2003) 6941 30. DW Lachenmeier, L Kroener, F Musshoff & B Madea, Anal Bioanal Chem 378 (2004) 183 31. F Musshoff, P Junker Heike, DW Lachenmeier, L Kroener & B Madea, J Anal Toxicol 26 (2002) 554 32. F Sporkert, F Pragst, S Hubner & G Mills, J Chromatogr B 772 (2002) 45 33. J O’Reilly, I Bruheim, C Goodridge, DR Parkinson, J Pawliszyn & S Penalver, Automation of SPME with In-Fibre Derivatisation for the Analysis of Volatile Compounds in Food, Pittcon 2004, Chicago, IL, March 7 12, 2004 34. M Sakamoto & T Tsutsumi, J Chromatogr A 1028 (2004) 63 35. Y Chen & J Pawliszyn, Anal Chem 78 (2006) 5222 36. Z Zhang & J Pawliszyn, Anal Chem 67 (1995) 34 37. R Eisert & K Levsen, J Chromatogr A 737 (1996) 59 38. R Eisert, J Pawliszyn, G Barinshteyn & D Chambers, Anal Commun 35 (1998) 187 39. C Grote, K Levsen & G Wuensch, Anal Chem 71 (1999) 4513 40. E Belau, C Grote, M Spiekermann & K Levsen, Field Anal Chem Technol 5 (2001) 37 41. Y Chen & J Pawliszyn, Anal Chem 76 (2004) 5807 42. Y Chen, J O’Reilly, Y Wang & J Pawliszyn, Analyst 129 (2004) 702 43. Y Wang, J O’Reilly, Y Chen & J Pawliszyn, J Chromatogr A 1072 (2005) 13 44. DR Parkinson, I Bruheim, I Christ & J Pawliszyn, J Chromatogr A 1025 (2004) 77 45. JV Wooten, DL Ashley & AM Calafat, J Chromatogr B 772 (2002) 147 46. E Jagerdeo & I Christ, Analysis of Ethyl Carbamate in Alcohol Beverages by an Automated Solid-Phase Mictroextraction Using a Dual-Arm Autosampler and Gas Chromatography/Mass Spectrometry, U.S. Department of Treasury Bureau of Alcohol, Tobacco and Firearms and LEAP Technologies, http://www.leaptec.com/assets/12/353/ automated_spme_2002.pdf, accessed September 2004 47. R Eisert & J Pawliszyn, Anal Chem 69 (1997) 3140 48. D Vuckovic, E Cudjoe, D Hein & J Pawliszyn, Anal Chem 80 (2008) 6870 49. H Lord & J Pawliszyn, J Chromatogr A 885 (2000) 153

Automated SPME Systems

50. 51. 52. 53. 54. 55. 56. 57. 58. 59. 60. 61. 62. 63. 64. 65. 66. 67. 68. 69. 70. 71. 72.

73. 74. 75. 76.

77.

78.

79. 80.

165

Y Gou, C Tragas, H Lord & J Pawliszyn, J Microcolumn Sep 12 (2000) 125 Y Gou, R Eisert & J Pawliszyn, J Chromatogr A 873 (2000) 137 Y Gou & J Pawliszyn, Anal Chem 72 (2000) 2774 J Wu, H Lord, J Pawliszyn & H Kataoka, J Microcolumn Sep 12 (2000) 255 H Kataoka, S Narimatsu, HL Lord & J Pawliszyn, Anal Chem 71 (1999) 4237 H Kataoka, H Lord & J Pawliszyn, J Chromatogr B 731 (1999) 353 H Kataoka, H Lord & J Pawliszyn, J Anal Toxicol 24 (2000) 257 H Kataoka, HL Lord, S Yamamoto, S Narimatsu & J Pawliszyn, J Microcolumn Sep 12 (2000) 493 H Kataoka, E Matsuura & K Mitani, J Pharm Biomed Anal 44 (2007) 160 J Nie, Q Zhao, J Huang, B Xiang & Y-Q Feng, J Sep Sci 29 (2006) 650 H Kataoka & J Pawliszyn, Chromatographia 50 (1999) 532 K Mitani, S Narimatsu & H Kataoka, J Chromatogr A 986 (2003) 169 H Kataoka, Anal Bioanal Chem 373 (2002) 31 H Kataoka, Trends Anal Chem 22 (2003) 232 JP Hutchinson, L Setkova & J Pawliszyn, J Chromatogr A 1149 (2007) 127 E Cudjoe, D Vuckovic, D Hein & J Pawliszyn, Anal Chem 81 (2009) 4226 E Cudjoe & J Pawliszyn, J Pharm Biomed Anal (2008) 10.1016/j.jpba.2008.07.014 R Vatinno, D Vuckovic, CG Zambonin & J Pawliszyn, J Chromatogr A 1201 (2008) 215 R Vatinno, A Aresta, CG Zambonin & F Palmisano, J Pharm Biomed Anal 44 (2007) 1014 D Vuckovic & J Pawliszyn, J Pharm Biomed Anal (2008) 10.1016/j.jpba.2008.08.023 X Zhang, A Es-haghi, FM Musteata, G Ouyang & J Pawliszyn, Anal Chem 79 (2007) 4507 FS Mirnaghi, Y Chen, LM Sidisky & J Pawliszyn, Anal Chem 83 (2011) 6018 PJ Rodacy, SD Reber, RJ Simonson & BG Hance, Unexploded Ordinance Classification Sensor for Underwater Applications, Report No. SAND2000-0920, Sandia National Laboratories, Albuquerque, NM, 2000 K Kobayashi, I Nukatsuka, F Miyashita & K Ohzeki, Bunseki Kagaku 52 (2003) 917 E Baltussen, P Sandra, F David & C Cramers, J Microcolumn Sep 11 (1999) 737 P Popp, C Bauer, B Hauser, P Keil & L Wennrich, J Sep Sci 26 (2003) 961 N Ochiai, K Sasamoto, H Kanda, T Yamagami, F David & P Sandra, Multi-Residue Method for Determination of 85 Pesticides in Vegetables, Fruits and Green Tea by Stir Bar Sorptive Extraction and Thermal Desorption GC MS, Application Note 4/2004, Gerstel GmbH, Mulheim an der Ruhr, 2004 B Hauser, PPC Bauer & E Kleine-Benne, Semi-Automated Stir Bar Sorptive Extraction in Combination with HPLC-Fluorescence Detection for the Determination of Polycyclic Aromatic Hydrocarbons in Water, Application Note 1/2002, Gerstel GmbH, Mulheim an der Ruhr, 2002 E Kleine-Benne, A Buhr, B Hauser & P Popp, Possibilities for Automation of Sampler Preparation Steps Prior to HPLC or GC Analysis Using a Common Autosampler, Technical Note S-66, Gerstel GmbH, Mulheim an der Ruhr, 2002 J Koziel, M Odziemkowski & J Pawliszyn, Anal Chem 73 (2001) 47 Y Gong, I-Y Eom, D-W Lou, D Hein & J Pawliszyn, Anal Chem 80 (2008) 7275

6 Calibration Gangfeng Ouyang School of Chemistry and Chemical Engineering, Sun Yat-sen University, Guangzhou, P. R. China

6.1

Introduction

Solid-phase microextraction (SPME) was first introduced in 19901 and was subsequently optimised and automated.2 It is a solvent-free sample preparation technique that integrates sampling, isolation and concentration. Also, its simplicity of use, relatively short sample processing time and fibre reusability have made SPME an attractive choice for many analytical applications.3
5 Since that time, SPME has been widely applied to the sampling and analysis of environmental,6
10 food,11
13 aromatic,14,15 metallic,16
18 forensic19
24 and pharmaceutical samples.25
30 In SPME, micro-quantities of solid sorbent or liquid polymer in an appropriate format are exposed to the sample for a well-defined period of time. Quantification is based on the amount of analyte extracted under the appropriate conditions. Unlike traditional sample preparation methods, such as liquid
liquid extraction (LLE), solid-phase extraction (SPE) and Soxhlet, SPME is a non-exhaustive extraction technique in which only a small portion of the target analyte is removed from the sample matrix. This feature allows the monitoring of chemical changes, partitioning equilibria and speciation in the investigated system because sampling causes minimal perturbation to the system.31,32 Therefore, the use of SPME results in better characterisation and more accurate information about the investigated system or process compared to exhaustive techniques. SPME also provides signal magnitudes that are proportional to the free concentration of target analyte, defining the fraction of the analyte that is bioavailable. This unique feature of SPME allows the measurement of binding constants in complex matrices, providing additional information about the investigated system (see Chapter 11).26,32,33 Non-exhaustive extraction also indicates the need for careful calibration of SPME for quantitative analysis. An appropriate calibration method for SPME quantification is required. Calibration is a process relating the measured analytical signal to the concentration of analyte in the sample matrix. SPME calibration technique suitability depends on the application, the number of samples to be analysed and the availability of a mass spectrometry instrument in the laboratory. The development of SPME calibration methods is based on an understanding of fundamental principles governing the mass transfer of analytes in multiphase systems. Theories have been developed to understand the principal processes involved in SPME by applying the basic fundamentals of thermodynamics and mass transfer kinetics.3,34,35 Several methods have been used for the calibration of SPME, as shown in Figure 6.1. Handbook of Solid Phase Microextraction. DOI: 10.1016/B978-0-12-416017-0.00006-1 © 2012 Elsevier Inc. All rights reserved.

168

Handbook of Solid Phase Microextraction

Figure 6.1 Various calibration methods in SPME

In addition to traditional calibration methods (i.e. external standard, internal standard and standard addition), the existing SPME calibration methods include equilibrium extraction, exhaustive extraction and diffusion-based calibration. In this chapter, details and characteristics of these calibration methods will be discussed.

6.2

Traditional Calibration Methods for the Quantification of SPME

Traditional calibration methods, such as external standard, internal standard or standard addition calibration methods, can be used for the quantification of SPME, but each calibration method presents different advantages and disadvantages. These methods are more suitable for laboratory analysis, although some on-site applications are reported. Using these methods for the quantification of SPME, two approaches, equilibrium and pre-equilibrium, can be employed. In one approach, a partitioning equilibrium between the sample matrix and extraction phase is reached. In this case, convection conditions do not affect the amount of analyte extracted. In a second approach that uses short pre-equilibrium extraction, if the convection/agitation is constant, the amount of analyte extracted is related to time. Quantification can then be performed based on the timed accumulation of analytes in the coating. Equilibrium extraction does not require the convection/agitation to remain constant, and the analytical sensitivity is higher than pre-equilibrium approach. However, if the time for equilibrium extraction is too long, then the pre-equilibrium extraction is better.

6.2.1

External Standard Calibration

For SPME, a calibration curve involves the preparation of several standard solutions in the sample matrix to obtain the relationship between the peak responses and the target standard concentrations. The samples are subsequently analysed with the same extraction conditions. Then, the concentration of the target analyte can be calculated from the equation of the calibration curve, as shown in Figure 6.2.

Calibration

169

Figure 6.2 SPME external calibration method for toluene aqueous solutions.

This calibration method does not require extensive sample preparation, but the sampling procedure and chromatographic conditions must remain constant for both the sample and the standard solutions. If there are matrix effects, a blank sample matrix is necessary. For gaseous samples, a standard gaseous mixture or a standard gas generating system is required. External calibration is useful when there is little variability in test sample composition or where test samples are all normalised prior to extraction. For instance, in urinalysis of drugs, particularly for overdose samples where analyte concentration is very high, a small amount of sample (ca. 50 µL) of urine may be diluted in several millilitres of buffer. In this case, overall sample composition is very constant, and external calibration will give satisfactory results. On the other hand, for analysis of trace amounts of drugs in urine, it is more likely that such dilution is not feasible. Because of the large variability in ionic strength in urine, the impact of this factor on the extraction must be determined. If the ionic strength effect is significant, ionic strength will either have to be normalised by performing all analyses at high or saturated salt concentration or compensated for if the dependence is well established. Some situations may require an alternate means of quantification. For solid sorbents, equilibrium extraction with external calibration is feasible where the concentration of analyte in the sample does not cause a saturation of the active sites in the sorbent. In order to achieve linear calibration, extractions should be designed so that only a small proportion of the active sites on the sorbent are occupied at equilibrium. The external calibration is a widely used calibration method of SPME. It is easy to find the applications of this method for the quantitative analysis of environmental,36
40 food41
45 and biological samples33,46
48 with SPME. The method is also used for on-site sampling,49
54 and the calibration is normally performed in the laboratory with standard gas mixture. Because the convection conditions are difficult to keep the same both on-site and in the laboratory, equilibrium extraction is preferable. Therefore, the method is more suitable for on-site sampling of gaseous samples compared with on-site water sampling because the equilibrium time is short for volatile compounds. Nevertheless, the loss of the extracted analytes during the transportation of the sampler should be avoided.

170

Handbook of Solid Phase Microextraction

6.2.2

Standard Addition

It is sometimes difficult or impossible to acquire a matrix-matched blank sample for calibration. The sources of the samples may be sufficiently different that one blank matrix cannot be identified that is suitable for all samples. Alternatively, the sample may be available in such limited quantities that it is not possible to acquire additional amounts for calibration. In these cases, standard addition calibration may be a reasonable alternative. The standard addition approach involves adding known quantities of the target analyte to the sample matrix, which initially contains an unknown concentration of the analyte, and this mixture is then analysed. A plot of the responses for the range of target analyte concentrations is then developed, and the extrapolation of the response to zero defines the original concentration in the unspiked sample. Figure 6.3 illustrates the standard addition approach. The concentration of the analyte in the sample, Cs, can be calculated as: Cs 5

As a

ð6:1Þ

where As is the peak area of the analytes in the sample and a is the slope of the line plotted by the peak areas and the concentrations of standard added. In Figure 6.3, the value of a is 70.026. It is recommended for best precision (2
5%) to analyse at least three standard addition samples at different concentrations in triplicate. Obviously, the major drawback to this method is the large number of additional samples that must be analysed for each unknown. It can be extremely tedious and time-consuming for a large number of samples. The advantage of this method is that the sample matrix effects can be compensated. It is appropriate when the sample number is small and sample composition is unknown and complex. For most heterogeneous samples, the standard addition calibration method should be a first consideration.55
60 When using the standard addition method for heterogeneous or solid samples, the mass transfer mechanism can be different for the standards added and the native analytes, and thus the pre-equilibrium approach is not suitable. Figure 6.3 SPME standard addition approach for toluene aqueous solutions.

Calibration

6.2.3

171

Internal Standard

The use of internal standards for quantification is routine for many methods and can give satisfactory results for microextraction as well. An internal standard can be used to compensate for the matrix effect, losses of analytes during sample preparation and irreproducibility in parameters, such as injection volume in gas chromatography (GC) or liquid chromatography (LC).61
66 If an appropriate internal standard can be identified, this technique is highly valuable, both in terms of controlling for matrix variation and in limiting the number of analyses required. The internal standard calibration approach involves adding a compound with very similar properties to the analyte to the calibration solutions and samples. The compound should be sufficiently different from the analytes so that it is well resolved during the chromatographic separation. It should mimic the behaviour of the analyte during the extraction. With this approach, a calibration plot is developed by determining the ratio of the peak area of the analyte to the internal standard for calibration solutions that contain different concentrations of the analyte with a fixed concentration of the internal standard. This ratio is subsequently used to calibrate the sample, as shown in Figure 6.4. It is preferable that the internal standard selected relates closely to the analyte of interest, particularly in terms of the partition coefficient for the extraction phase and any competing phases. If the internal standard is extracted to a significantly different extent than the analyte, error in the analysis will be either understated or overstated. Also, extraction time profiles must be determined for both compounds. If the analyte of interest is extracted at equilibrium (and therefore extraction time is not closely controlled), large errors will result if the internal standard is not also extracted at equilibrium. This can occur if the internal standard has a higher partition coefficient than the analyte of interest. Finally, if the sample composition is optimised for the analyte of interest, extraction conditions for a poorly matched internal standard can be suboptimal. For instance, if the pK of the analyte of interest is very different from that of the internal standard, and the sample pH is adjusted to favour the analyte of interest but not the internal standard, results will be poor. Figure 6.4 SPME internal standard calibration method.

172

Handbook of Solid Phase Microextraction

A limitation to the use of an internal standard exists where there is a significant and variable competing phase present in the sample. The internal standard may have a very different affinity for the competing phase than the analyte of interest does. Where the composition of the competing phase is consistent, quantification by the internal standard may still be possible because the absolute amounts of binding of both compounds to the competing phase would be linearly correlated with the amount of the phase present. Where the composition of the competing phase is variable, however, the absolute amounts of internal standard and analyte of interest bound may not correlate linearly, and peak ratio analysis would not give an adequate quantification. Where the amount of the competing phase is small, the absolute amount bound to the phase would be insignificant and could be ignored.

6.2.3.1 Isotopically Labelled Standards By far, the most accurate and simplest method of quantification is the use of an isotopically labelled standard. Unfortunately, these are commercially available for only a small number of compounds, are expensive and can be unstable. In some cases, it may be possible to prepare deuterated standards in-house. The use of isotopically labelled standards of course requires mass spectral detection. The only caution for using isotopically labelled standards is that the standard should be added at roughly the same concentration (within 1
2 orders of magnitude) as the unknowns. The tendency is to add very little isotopically labelled standard, due to cost considerations. However, in this way, the error in analysis of the standard may be much larger than for the unknown, and this error will be reflected in a larger than normal error for unknown quantification.

6.3

Equilibrium Extraction

The equilibrium extraction method is a widely used quantification method for SPME, especially for on-site sampling. In this method, a small amount of extraction phase (SPME fibre coating or other sorbent or polymer in appropriate format) is exposed to a sample matrix until an equilibrium is reached. The equilibrium conditions can be described by Eq. (6.2), according to the law of mass conservation, if only two phases are considered (e.g. the sample matrix and the fibre coating)3: n5

Kfs Vf Vs C0 Kfs Vf 1Vs

ð6:2Þ

where C0 is the initial concentration of the target analyte in the sample, Vs is the sample volume, Vf is the fibre coating volume and Kfs is the distribution coefficient of the analytes between the fibre coating and the sample. Equation (6.2) indicates that the amount of analyte extracted into the coating (n) is linearly proportional to the analyte concentration in the sample (C0). Equation (6.2) is the analytical basis for quantification using SPME.

Calibration

173

When the sample volume is very large, as in field sampling, then Vs cKfs Vf and Eq. (6.2) can be simplified to Eq. (6.3): n 5 Kfs Vf C0

ð6:3Þ

Equation (6.3) illustrates the advantage of the equilibrium extraction method for field applications. In this equation, the amount of extracted analyte is independent of the sample volume. In practice, there is no need to collect a defined sample prior to analysis because the fibre can be exposed directly to the ambient air, water, and production stream, or to study biological samples, as described in Chapter 12. The amount of extracted analyte will correspond directly to its concentration in the matrix without depending on the sample volume. When the sampling step is eliminated, the whole analytical process can be accelerated, and errors associated with analyte losses through the decomposition or adsorption on the sample container walls will be prevented. Equation (6.3) also illustrates another characteristic of field sampling with SPME: the concentration of target analytes can be determined from the amount of analytes on the fibre under extraction equilibrium, by knowing the distribution coefficients of the analytes between the fibre coating and the sample matrix. The distribution coefficients of the target analytes between the coating material and the sample matrix can be directly determined by experimentation. Several methods have been developed to measure the fibre coating distribution coefficients. The selected analytes include polycyclic aromatic hydrocarbons (PAHs),67
71 polychlorinated biphenyls (PCBs),72
77 polybrominated diphenyl ethers (PBDEs),78 polybrominated biphenyls (PBBs),79 pesticides,80 phthalates81 and so on.82 The reported data is very useful for the application of the SPME equilibrium extraction approach. Determining fibre coating distribution coefficients using dynamic systems is more accurate than with static systems because analyte losses in the system (due to the fibre uptake, sorption on the walls and so on) can be compensated.72 A standard gas/aqueous solution generating system is very useful for determining the distribution coefficients and the calibrations of SPME and other sampling devices because constant concentrations of analytes can be generated.67,83,84 A dynamic system was also proposed for determining the apparent distribution constants between an SPME fibre and a blood matrix/buffer sample matrix.33 In addition to direct partitioning measurements, distribution constants can also be estimated from physicochemical data and chromatographic parameters. For example, distribution constants between a fibre coating and a gaseous matrix (e.g. air) can be estimated with retention indexes from a linear temperature-programmed capillary GC on a column with stationary phase identical to the fibre coating material.85,86 SPME equilibrium extraction method is commonly used for on-site air87
89 or water sampling.90
94 For air sampling, the extraction can be performed in the static and dynamic model. In the dynamic model, the extraction rate can be increased by using an air pump.87
89 The equilibrium extraction time for sampling in water is much longer than in air, sometimes requiring several weeks.90
92 Using a portable drill to rotate the sampler can obviously shorten the equilibrium time.94

174

6.4

Handbook of Solid Phase Microextraction

Exhaustive Extraction

SPME is mainly an equilibrium extraction method instead of an exhaustive extraction technique. In most cases, analytes are not completely extracted. The concentration of an analyte is determined by its linear relationship with the amount of the analyte extracted by the fibre coating. When the sample volume is very small, and the distribution coefficient of the analyte between the fibre coating and the sample matrix is very large (Vs {Kfs Vf ), such as during the sampling of semi-volatile organic compounds (semi-VOCs) in small volumes or the sampling of VOCs in small volumes using a cold fibre, then Eq. (6.2) can be simplified to: n  Vs C0

ð6:4Þ

Equation (6.4) implies that all the analytes in the sample matrix are extracted into the fibre coating and the concentration of the target analyte can be easily calculated from the amount of analyte extracted by the fibre coating and the volume of the sample. The calibration for exhaustive extraction is not often used in SPME because it is typically only suitable for small sample volumes and very large distribution coefficients. Utilising special devices or methods, SPME exhaustive extraction is possible to achieve. An internally cooling fibre device, in which the distribution coefficient is significantly increased by simultaneously heating the sample matrix and cooling the fibre coating with CO2, provides an opportunity to obtain the total amount of analyte in the sample.95 The device has been miniaturised and automated96 and used for the analysis of environmental97, food98,99 and fragrance samples.100
102 Another application of exhaustive extraction is multiple SPME, in which the sample is repeatedly extracted with the fibre and the total amount of the analyte can be extrapolated from only a few extractions, even if the analyte in the sample matrix is not extracted exhaustively.103 Equation (6.5) described the relationship between AT, the total peak area, and A1, the first peak area104: AT 5

N X i51

Ai 5

A1 12β

ð6:5Þ

where β is a constant. According to Eq. (6.5), the total peak area can be calculated from only two values: the first peak area, A1, and the constant β. The former is a measured value, whereas the latter can be obtained from the following equation104: ln Ai 5 ði 21Þln β 1ln A1

ð6:6Þ

The β value can be calculated from the slope of the linear plot ln Ai versus (i
1) obtained from a few (three or four) determinations. Figure 6.5 shows the chromatographic peaks obtained from four HS
SPME
GC determinations of nonanal in hexadecane and the linear plot ln Ai versus (i
1).

Calibration

175

Figure 6.5 Chromatographic peaks of four successive HS
SPME determinations of nonanal in a single hexadecane solution and linear plot ln Ai versus (i 21).104

Multiple SPME has been used for the analysis of solid105
109 and liquid samples.103,110,111 It has also been developed as an alternative method for the determination of distribution coefficient (Kfs) of the analytes between the fibre coating and the sample matrix.78,80,112 The advantage of this method is that the matrix effects can be avoided by determining the total amount of the analyte in the sample. However, the usefulness of this method is limited. The β value should be less than 0.95.113 If the β value is higher than 0.95, it will result in negligible depletion of SPME. The amount of analyte in the vial almost remains constant, and the area for successive extractions is the same. In addition, the adsorption phenomena114 and fibre coating saturated by matrix components can invalidate multiple SPME for quantification.115

6.5

Diffusion-Based Calibration

The extraction process of SPME generally follows the profile shown in Figure 6.6.116 When the sampling time is longer than t95, the extraction nearly reaches equilibrium. If the sampling time is less than t95, the extraction is a kinetic process, and there is almost a linear mass uptake when the sampling time is less than t50. To describe the kinetic process of SPME, the diffusion coefficient is essential. Several diffusion-based calibration methods were developed for the quantification of SPME in recent years. These calibration methods are developed from Fick’s first law of diffusion, the interface model, the cross-flow model and the kinetic processes of absorption/adsorption and desorption. They are mainly used for on-site and in vivo sampling, including grab sampling and long-term monitoring.

176

Handbook of Solid Phase Microextraction

Figure 6.6 Typical extraction profile of SPME.116

Figure 6.7 Schematic diagram of fibre-retracted SPME device and concentration gradient.

6.5.1

Fick’s First Law of Diffusion

These types of SPME on-site sampling devices are unlike conventional sampling with SPME in which the fibre is retracted a known distance into its needle housing during the sampling period. Later, these types of samplers were termed as fibre-retracted SPME devices.117 When sampling with fibre-retracted SPME devices, the analyte molecules access the fibre coating only by means of diffusion through the static air/water gap between the opening and fibre coating. Therefore, Fick’s first law of diffusion can be used for the calibration (Figure 6.7). For these types of samplers, the diffusion paths are well defined. The sampling rate, Rs, is proportional to the molecular diffusion coefficient and the ratio of the opening area (A) to the diffusion path length (Z). If the sorbent is ‘zero sink’ for the target analyte and the sampling is limited in the linear regime (Figure 6.6), the concentrations of analyte in the sample can be calculated with Eq. (6.7)118:

Calibration

177

ð ð A CðtÞdt n 5 Rs CðtÞdt 5 D Z

ð6:7Þ

where C is the concentration of the target analyte in air or water, A is the crosssectional area of the needle, D is the diffusion coefficient of the target analyte in air or water and n is the amount of the analyte extracted by the fibre during time t. Equation (6.7) can be simplified to: C5

nZ ADt

ð6:8Þ

where C is the time-weighted-average (TWA) concentration of the target analyte in air or water. With the well-defined diffusion path, Fick’s first law of diffusion can be directly used for calibration, and the calculation of C is very simple. The samplers based on this calibration method are mainly used for air or water sampling because the diffusion coefficients of the analytes should be known, and the parameters in air or water are easy to find in the literature or calculate with empirical equations.112,119 The desirable feature of the samplers is that the performance of the device is independent of the face velocity, especially for field sampling, where the convection conditions of water are very difficult to measure and calibrate due to the extremely small inner diameter (i.d.) of the diffusion path.120,121 The diffusion of molecules in air happens very fast. The sampling times can be selected, from less than 1 min to days, by adjusting the length of the diffusion path.118,122 The outside needle of the SPME fibre assembly can be used directly as the diffusion path. These types of samplers have been used for TWA sampling of formaldehyde,50,118 alkanes,122 dodecane49 and other analytes in air120,123
125 (Figure 6.8). A more convenient fibre-retracted SPME device was designed by Chen and Pawliszyn126 for field TWA air sampling (Figure 6.9). There are two problems that should be solved when these types of devices are used for TWA water sampling. First, the air in the diffusion path should be

Figure 6.8 SPME TWA passive air sampling device.

178

Handbook of Solid Phase Microextraction

Figure 6.9 SPME field air sampler. (A and B) Fibre holders. (C) Commercially available. (D-1) Cross view of the adjustable cylinder. (D-2) Side view of the adjustable cylinder. (E) Protecting shield. (F) Replaceable Teflon cap. (G) Cross-sectional diagram of the sampler. SPME fibre inside a removable deactivated needle

Additional holes

Z

Hamilton 500 μL gas-tight syringe

PDMS ferrule

Teflon ferrule Glass barrel Stainless steel nut

Adjustable silcosteel-treated needle

Figure 6.10 Schematic diagram of the fibre-in-needle SPME device for TWA water sampling and the adjustable/removable needle.

completely replaced with water; otherwise, calibration with Eq. (6.8) will be improper. In addition, the outside needle of SPME assembly cannot be used as the diffusion path directly because analytes will be adsorbed by the outside wall of the needle. The diffusion of the molecule in water is much slower than in air; therefore, the adsorption of the analytes by the outside wall of the needle should not be ignored. Ouyang et al.121 modified this type of SPME device for TWA water sampling using a 500 µL gas-tight syringe and a removable needle (Figure 6.10). A field TWA water sampling device was designed with a commercial SPME fibre assembly and has been used to monitor PAHs in Hamilton Harbour and Laurel Creek, Ontario, Canada.117,127

Calibration

179

The disadvantage of this device for water sampling is that the amount of analyte acquired is not as high as typical passive samplers because of the low sampling rate.128 However, this disadvantage is not as important for TWA sampling because it can be enhanced by increasing the sampling time, which is the ultimate purpose of TWA sampling.

6.5.2

Interface Model and Cross-Flow Model

When the diffusion path is not well defined, for example, the SPME fibre is directly exposed in the matrix for sampling, the calibration can be performed with the interface model and cross-flow model. Both the interface and cross-flow models are limited to the linear sampling regime, and the convection of air and water should be kept constant. The flow velocity of air and water should be controlled or determined when using these models for calibration.

6.5.2.1 Interface Model One way to overcome the fundamental limitation of porous coatings in a microextraction application is to use an extraction time that is much less than the equilibration time, so that the total amount of analytes accumulated by the porous coating is substantially below the saturation value. At saturation, all surface sites available for adsorption are occupied. When performing such experiments, not only is it critical to control extraction times precisely but also convection conditions must be controlled because they determine the thickness of the diffusion layer. One way of eliminating the need to compensate for differences in convection is to normalise (i. e. use consistent) agitation conditions. For example, by the use of stirring (i.e. a well-defined rate of rotation in the laboratory) or the use of fans for field air monitoring, consistent convection will be ensured.128,129 The short-term exposure measurement described above has an advantage in that the rate of extraction is defined by the diffusivity of analytes through the boundary layer of the sample matrix, and thus the corresponding diffusion coefficients, rather than by distribution constants. This situation is illustrated in Figure 6.11 for a cylindrical geometry of the extraction phase dispersed on the supporting rod. Analyte concentration in the bulk of the matrix can be regarded as constant when a short sampling time is used and there is a constant supply of analyte as a result of convection. These assumptions are true for most types of sampling in which the volume of the sample is much greater than the volume of the interface, and the extraction process does not affect the bulk sample concentration. In addition, the solid coating can be treated as a ‘perfect sink’ for analytes. Adsorption binding is frequently instantaneous and essentially irreversible. The analyte concentration on the coating surface is far from saturation and can be assumed to be negligible for short sampling times and the relatively low analyte concentrations in a typical sample. The analyte concentration profile can be assumed to be linear from Cs to C0. In addition, the concentration of analyte on the coating surface (C0) can be assumed to be zero when extraction begins. Diffusion of analytes inside the

180

Handbook of Solid Phase Microextraction

Silica rod Pores Bulk air movement Ds

Figure 6.11 Diffusion model for a cylindrical geometry of the extraction phase dispersed on the supporting rod, for short extraction time, under normalised stirring conditions.

us Solid coating surface (A) Boundary layer

δ d+b d

Cs Concentration profile

C0

pores of a solid coating controls mass transfer from the outer to inner surfaces of the coating. The function describing the mass of extracted analyte with sampling time can be derived by use of Eq. (6.9)130: nðtÞ 5

B3 ADs δ

ðt

Cs ðtÞdt

ð6:9Þ

0

where n is the mass of analyte extracted (ng) in a sampling time (t), Ds is the gas-phase molecular diffusion coefficient, A is the outer surface area of the sorbent (e.g. outer surface area of the coated rod in Figure 6.11 defined as 6.28(d 1 b)L, where L is the length of the coated portion of the rod), δ is the thickness of the boundary layer surrounding the extraction phase, B3 is a geometric factor and Cs is the analyte concentration in the bulk of the sample. It can be assumed that the analyte concentration is constant for very short sampling times and, therefore, Eq. (6.9) can be further reduced to: nðtÞ 5

B3 Ds A Cs t δ

ð6:10Þ

where t is the sampling time.131 It can be seen from Eq. (6.10) that the mass extracted is proportional to the sampling time, Ds for each analyte and the bulk sample concentration, and inversely proportional to δ. This is consistent with the fact that an analyte with a greater Ds will cross the interface and reach the surface of the coating more quickly. Values of Ds for each analyte can be found in the literature or estimated from

Calibration

181

physicochemical properties.132 This relationship enables quantitative analysis. As mentioned above, the discussion assumes non-reversible adsorption. Equation (6.10) can be modified to enable the estimation of the concentration of analyte in the sample for rapid sampling with solid sorbents: Cs 5

nδ B1 Ds At

ð6:11Þ

The amount of extracted analyte (n) can be estimated from the detector response. The thickness of the boundary layer (δ) is a function of sampling conditions. The most important factors affecting δ are the geometric configuration of the extraction phase, sample velocity, temperature, and Ds for each analyte. The effective thickness of the boundary layer can be estimated for the coated fibre geometry (Figure 6.11) by use of Eq. (6.12), an empirical equation adapted from heat transfer theory: δ 5 9:52

d Re0:62 Sc0:38

ð6:12Þ

where Re is the Reynolds number 5 2usd/ν; us is the linear sample velocity; ν is the kinematic viscosity of the matrix; Sc is the Schmidt number 5 ν/Ds and d is the fibre diameter. The effective thickness of the boundary layer in Eq. (6.12) is a surrogate (or average) estimate and does not take into account changes of the thickness that can occur when the flow separates and/or a wave is formed. Equation (6.12) indicates that the thickness of the boundary layer will decrease with increasing linear sample velocity. Similarly, when sample temperature (Ts) increases, the kinematic viscosity decreases. Because the kinematic viscosity term is present in the numerator of Re and in the denominator of Sc, the overall effect on δ is small. Reduction of the boundary layer and an increased rate of mass transfer for an analyte can be achieved in two ways
by increasing the sample velocity and by increasing the sample temperature. Increasing the temperature, however, will reduce the efficiency of the solid sorbent (reduced Kes). As a result, the sorbent coating might not be able to adsorb all molecules reaching its surface and, therefore, it might stop behaving as a ‘perfect sink’ for all the analytes. Equation (6.10) indicates that the initial extraction rate is proportional to the planar surface area of the extraction phase. The equilibration time, therefore, can be reduced by increasing the interfacial contact between the phases by designing the extraction phases with appropriate configurations
thin, flat films with high surface area, as discussed in Section 3.2.4 in Chapter 3.133 This model enables one to calibrate the extracted analyte mass as a function of the molecular diffusion coefficient, the analyte concentration, the sampling time, the air velocity, the air temperature and the fibre geometry. The use of short sampling times also minimises the effects of competitive adsorption for solid coating. The use of forced air increases the velocity. This calibration method was used for

182

Handbook of Solid Phase Microextraction

rapid sampling of VOCs in air with solid- or liquid-coated fibre.88,134,135 Based on this calibration method, a portable SPME device for rapid field air sampling was developed with controlled air flow by a modified hairdryer.128 The applications of this model for the calibration of aqueous samples were also reported.129,136

6.5.2.2 Cross-Flow Model Chen et al.137 proposed a different diffusion-based calibration method, using a cross-flow model (Figure 6.12). In this model, the target analyte concentration can be calculated from Eq. (6.13): C0 5

n nd 5 ERem Sc1=3 ADt hAt

ð6:13Þ

where A is the surface area of the fibre, n is the mass uptake onto the fibre during sampling time t, D is the diffusion coefficient of the analyte molecule, h is the average mass transfer coefficient, d is the outer diameter of the fibre, Re is the Reynolds number and Sc is the Schmidt number. Constants E and m depend on the Reynolds number and are available in the literature.137 This diffusion-based calibration method was validated by on-site analysis of benzene, toluene, ethylbenzene and xylene (BTEX) in both aqueous samples and indoor air. The results were confirmed with the NIOSH-1501 method. The main advantage of this method is that the quantification is diffusion-based, as in the interface model, which means no calibration curves or internal standards are needed because the necessary constants (e.g. the diffusion coefficient) can be found in the literature or estimated with empirical equations. This characteristic makes this method especially suitable for on-site analysis, where the construction

Figure 6.12 Schematic diagram of rapid extraction with an SPME fibre in a cross-flow model.

Calibration

183

of calibration curves or the addition of internal standards proves to be very difficult. However, calibration with the interface or cross-flow model requires samples with a flowing medium. The sample velocity must be known or controlled and as such, additional equipment is needed. Using the interface or cross-flow model for calibration, the sampling rate will be much higher than with the fibre-retracted devices because the diffusion boundary layer of this sampling method is much thinner than that of the fibre-retracted devices. However, these models are more suitable for fast sampling and cannot be used for TWA sampling.

6.5.3

Kinetic Calibration

Typically, the calibration methods are limited to either the equilibrium regime or the linear regime (Figure 6.6). In 1997, Ai34,35 proposed a theoretical model based on a diffusion-controlled mass transfer process to describe the entire kinetic process of SPME: n 5 ½12 expð2atÞ

Kfs Vf Vs C0 Kfs Vf 1 Vs

ð6:14Þ

where a is a rate constant that is dependent on the extraction phase, headspace and sample volumes, the mass transfer coefficients, the distribution coefficients and the surface area of the extraction phase. This dynamic model suggests that a linearly proportional relationship exists between the adsorbed analyte and its initial concentration in the sample matrix. When the sampling time is long enough (e.g. the extraction has reached equilibrium), then Eq. (6.14) changes to Eq. (6.2), which proves that Eq. (6.14) can be used in the entire process of SPME, including for the kinetic and equilibrium regime. Based on this theoretical model, the two calibration methods of SPME, the kinetic calibration with standard or in-fibre standardisation technique138,139 and the standard-free kinetic calibration,140 were proposed.

6.5.3.1 Kinetic Calibration with Standard/In-Fibre Standardisation Technique The absorption of an analyte into an SPME liquid coating from the sample matrix can be described by Eq. (6.15), according to the dynamic model of SPME proposed by Ai34,35: n 5 12 expð2atÞ ne

ð6:15Þ

where n is the amount of the extracted analyte at time t, ne is the amount of the extracted analyte at equilibrium and a is a constant that depends on the volumes of

184

Handbook of Solid Phase Microextraction

the extraction phase, headspace and sample, mass transfer coefficients, distribution coefficients and the surface area of the extraction phase. Based on this model, Chen and co-workers138,139 demonstrated the isotropy of absorption and desorption in the SPME liquid coating fibre, and a new calibration method, kinetic calibration, was proposed. This kinetic calibration method, also referred to as the in-fibre standardisation technique, uses the desorption of the standards, which are preloaded in the extraction phase, to calibrate the extraction of the analytes.141 For field sampling, the desorption of standard from an SPME fibre can be described by139: q 5 expð2atÞ q0

ð6:16Þ

where q0 is the amount of preadded standard in the extraction phase and q is the amount of the standard remaining in the extraction phase after exposure of the extraction phase to the sample matrix for the sampling time, t. When the constant a has the same value for the absorption of target analytes and the desorption of preloaded standards, the sum of q/q0 and n/ne should be 1 at any desorption/absorption time139: n q 1 51 ne q 0

ð6:17Þ

Then the initial concentrations of target analytes in the sample, C0, can be calculated with Eq. (6.18)142,143: C0 5

q0 n Kes Ve ðq0 2qÞ

ð6:18Þ

where Ve is the volume of the extraction phase, Kes is the distribution coefficient of the analyte between the extraction phase and the sample and q is the amount of the standard remaining in the extraction phase after exposure of the extraction phase to the sample matrix for the sampling time. The applicability of this technique to TWA water sampling was demonstrated by both theoretical derivations and field trials.127,143,144 It was reported that the effect of environmental factors, such as biofouling, temperature or turbulence, can be calibrated using this approach. The in-fibre standardisation is a pre-equilibrium method and can be used for the entire sampling period. The determined concentration before equilibrium is a TWA concentration because the desorption of the preloaded standard calibrated the extraction of the analytes, and the extraction is an integrating process. If the sampling reaches equilibrium, the determined data are the concentrations of the analytes in the sample at the time that the samplers were retrieved. The in-fibre standardisation technique makes it possible to use a simple Polydimethylsiloxane (PDMS) rod or PDMS membrane as a passive sampler to obtain the TWA concentrations of target analytes in a sampling environment (Figure 6.13).127,143,144

Calibration

185

Figure 6.13 SPME PDMS rod; (A) PDMS membrane, (B) inlet liner of GC and (C) passive samplers for automated injection with the ATAS system.

A pure PDMS-rod, which was used as a passive sampler, was 1 cm long with a diameter of 1 mm. The SPME PDMS-membrane passive sampler was constructed using a 127 µm thick PDMS thin film as the extraction phase. The thin-film was cut into a specific house-like shape, and the dimension of the thin-film was 2 3 2 cm with a 1-cm high triangle on the top of the square. The volumes of the SPME PDMS-rod (7.85 µL) and the SPME PDMS-membrane (62.5 µL) samplers are much larger than commercial SPME PDMS fibres (0.61 µL), thus increasing the sensitivity of the passive samplers. In the kinetic calibration method, the samplers require the preloading of a certain amount of standard and are exposed in the sample matrix for a defined period of time. The TWA concentrations of the target analytes in the sample can then be calculated with Eq. (6.18). The deployed samplers should be retrieved before all the preloaded standards are lost. Normally, the sampler should be retrieved in 1 month or less, and this period is determined by the flow velocity of the sample matrix. Figure 6.14 illustrates the symmetry of the absorption of a target analyte and the desorption of a standard in a PDMS rod. The usefulness of these samplers was demonstrated experimentally under laboratory as well as field conditions (in Hamilton Harbour, Hamilton, Ontario, Canada). The effect of environmental factors, such as temperature and turbulence, were successfully calibrated in the field studies and the results are comparable with those achieved by traditional methods.127 The SPME PDMS-rod and PDMS-membrane passive samplers have all the advantages of SPME: they are solvent-free, combine sampling, isolation and enrichment into one step, and can be directly injected into a GC for analysis without further treatment.

186

Handbook of Solid Phase Microextraction

Figure 6.14 Symmetry of absorption and desorption in a PDMS rod. Simultaneous absorption of pyrene (’) and desorption of deuterated pyrene (d10) (K) onto and from the PDMS rod from and into the flow-through.

1.2

n/ne or q/q0

1 0.8 0.6 0.4 0.2 0 0

20

40

60 Time (h)

Sampling rate (mL/d)

100

80

100

84.6

80 60 40

24.6 6.6

20 0.006 0

2.6

3.4 0.00 5

1m

25. 5 0.006

11 m

21 m

Figure 6.15 Sampling rates of three types of SPME-passive samplers at different sampling depths. Black, fibre-retracted device; gridding, PDMS rod; grey, PDMS membrane.

Both samplers are simple and easy to deploy and retrieve. They have large sampling rates, and the sensitivity is much higher than that of the fibre-retracted SPME device because the samplers are in direct contact with the sample matrix. The SPME PDMS membrane has a higher surface-to-volume ratio, compared with the PDMS rod, which results in higher sensitivity and mass uptake, leading to the detection of lower levels of analytes than by many other techniques. Figure 6.15 illustrates the sampling rates for three types of SPME-passive samplers at different sampling depths in Hamilton Harbour, Hamilton, Ontario. The SPME PDMS-membrane sampling rate is about 10 times higher than the one achieved by the PDMS-rod sampler and more than 1,000 times higher than the rate achieved by the fibre-retracted SPME device. An SPME PDMS-membrane passive sampler requires a large volume injector for analysis, but the PDMS-rod sampler can be analysed by a GC
MS with a normal injector. The concept of calibrants in the extraction phase has been extended to determine the concentrations of target analytes directly in the veins of animals, indicating that

Calibration

187

this approach is useful for in vivo studies as well.33,145 The applications of this technique for quantitative analysis of liquid-phase microextraction (LPME) were also reported.142,146 The on-fibre standardisation technique for adsorbent SPME coatings was also validated with the use of a commercially available fibre (Carbowaxs/templated resin) and a home-made fibre (polypyrrole).147 Experiments demonstrated that this calibration corrected for the sample matrix effects and minimised the displacement effects by the use of pre-equilibrium extraction. The technique was successfully applied to the analysis of pesticides in river water and white wine, as well as drug analysis in clinical plasma and whole blood samples.147 The pharmacokinetic profiles of diazepam, nordiazepam and oxazepam obtained by kinetic calibration based on deuterated standards are quite similar to those determined by the calibration curves method (Figure 6.16).33 Deuterated compounds are expensive and sometimes not available. Zhou et al.148 proposed an alternative method that employs the target analytes as the internal standards by dominant desorption. Dominant pre-equilibrium desorption not only

Figure 6.16 Averaged pharmacokinetic profiles of diazepam, nordiazepam and oxazepam, which were monitored by in vivo SPME over 8 h on three dogs.33 (A) Pre-equilibrium extraction and kinetic calibration based on deuterated standards during the experimental course. (B) Equilibrium extraction and external calibration curve method. (Source: http://s100.copyright.com/CustomerAdmin/PLF.jsp?lID=2009030_1235987737101)

188

Handbook of Solid Phase Microextraction

offers a shorter sample preparation time but also provides time constants for the purpose of quantitative analysis. This kinetic calibration method was successfully applied to on-site PAH sampling in a flow-through system and in vivo pesticide sampling in a jade plant.148 Using the kinetic calibration with standard in the extraction phase method, the samplers require the preloading of a certain amount of standard, either deuterated compounds or target analytes. Zhao et al.149 reported several standard loading approaches, which include (i) headspace extraction of the standard dissolved in a solvent or pumping oil, (ii) headspace extraction of pure standard in a vial, (iii) direct extraction in a standard solution and (iv) direct transfer of the standard solution from the syringe to the fibre. The existing SPME kinetic calibration technique, using desorption of preloaded standards to calibrate the extraction of the analytes, requires that the physicochemical properties of the standard be similar to those of the analyte, which limited the application of the technique. Recently, a new method, termed the one-calibrant technique, which can use only one standard to calibrate all extracted analytes, was proposed.150 The theoretical considerations were validated in a flow-through system, using PDMS SPME fibres as passive samplers. The newly proposed one-calibrant technique makes the SPME kinetic calibration method more convenient and more applicable.

6.5.3.2 Standard-Free Kinetic Calibration Kinetic calibration with standard in the extraction phase can be used for both grab sampling and long-term monitoring. For fast on-site or in vivo analysis, preloading standards is inconvenient, and this calibration method may not work in some fast sampling situations because the loss of the standard will be too small to detect. Recently, a standard-free kinetic calibration method was proposed for fast onsite and in vivo analysis.140 With this calibration method, all analytes can be directly calibrated with only two samplings. Equilibrium extraction results in the highest sensitivity in SPME because the amount of analyte extracted onto the fibre coating is maximised when equilibrium is reached. If sensitivity is not a major concern in analysis, reduction of the extraction time is desirable. When the extraction conditions are kept constant (e.g. fast sampling), Eq. (6.19) can be used for the calculation of ne, the amount of analyte extracted at equilibrium140:     t2 n1 n2 5 ln 1 2 ln 1 2 t1 ne ne

ð6:19Þ

where n1 and n2 are the amount of analyte extracted at sampling times t1 and t2, respectively. Then the concentration of the analyte in the sample can be calculated with Eq. (6.2) or (6.3). The feasibility of this calibration method was validated in a standard aqueous solution flow-through system and a standard gas flow-through system. Using this standard-free kinetic calibration method, the sampling time for equilibrium

Calibration

189

extraction can be markedly shortened. In the reported study,140 typical sampling times for the equilibrium extraction of PAHs in a water environment, which typically range from 2 h to 24 h, can be shortened to 2
5 min, and sampling times for BTEX in air can be shortened from 5
10 min to 5
10 s. This calibration method can be used for the entire sampling period without considering whether the system reaches equilibrium.140 This aspect of the technique is desirable for systems when the equilibrium time is not known, and particularly useful for instances when a number of compounds are measured simultaneously. The method is unsuitable for long-term monitoring of pollutants in the environment because the method requires that the sampling rate remains constant and the determined concentration is therefore representative of a spot sampling.

6.6

Calibration of SPME by Liquid Injection

Most quantitative approaches of SPME, especially for on-site and in vivo sampling, require the absolute amounts of the analytes to be known. In this case, the analysis is most often calibrated by the injection of liquid standards. For both liquid injection and SPME fibre injection, usually more than 95% of the sample is supposed to be transferred into the capillary column by splitless GC injection. The applicability of this calibration method relies on the assumption that sample transfer efficiencies are the same for both the liquid injection and the SPME injection. However, the rate of sample transferred into the column for liquid injection is affected by many factors, such as the dimensions of the liner, the presence of wool, the temperature of the injector, and so on. The sample transfer efficiency for the SPME fibre injection is also affected by the carrier gas flow rate, the cross-sectional area of the space between the column and the liner and the length of the column inside the liner. The expansion associated with the solvent vapourisation will cause analyte loss when liquid standards are injected by common injection methods, especially with high-temperature liquid injection. However, SPME is a solvent-free sample preparation and sampling technique, and the sample transfer efficiency for SPME fibre injection is affected by different factors. When liquid injection is used to calibrate SPME fibre injection, the sample transfer efficiency of the liquid injection might be different from that of the SPME fibre injection if the analytes are not completely transferred into the GC column. The best scenario is, of course, that the analytes are completely transferred into the GC column for both the liquid injection and the SPME fibre injection. When this is difficult to achieve, the analytes should be transferred into the GC column at the same rate to ensure correct calibration. An equation has been proposed to estimate the sample loss rate for the SPME injection by Ouyang et al.151 (Figure 6.17): nL DA 5 nT Zv

ð6:20Þ

190

Handbook of Solid Phase Microextraction

SPME Fibre Varian 1079 ID 0.8 mm SPME liner

Varian 1093 ID 0.8 mm SPI liner

Split outlet

1 cm 0.2 cm

ID 0.53 mm OD 0.68 mm Column

ID 0.25 mm OD 0.34 mm Column (A)

(B)

Figure 6.17 Comparison of the SPME liner with the SPI liner for SPME injection. (A) SPME liner; (B) SPI liner.

where nT and nL are the total amount of the analyte in the sample and the amount of any analyte lost in the process, respectively; D is the diffusion coefficient of the analyte; A is the cross-sectional area of the space between the liner and the column; Z is the length of the column inside the liner; and v is the flow rate of the carrier gas. Equation (6.20) was validated by experiments examining change in the values of A, Z and v.151 Equation (6.20) indicates that the sample transfer efficiency of the SPME injection is affected by the cross-sectional area of the space between the column and the liner, the carrier gas flow rate and the length of the column inside the liner. The sample transfer efficiencies of the liquid injection were affected by the dimensions of the liner and the presence of wool. Programmed temperature vaporising (PTV) injection using a small i.d. liner with wool can result in good sample transfer efficiency. It was found that less than 70% of the analyte amount was transferred into the column when a high-temperature liquid injection method was used,151 which suggests that high-temperature liquid injection may not be suitable for SPME calibration. No obvious difference in the response factors was found when methanol, acetone or toluene was used as the solvent. Figure 6.18 describes the sample transfer processes for three liquid injection methods. For both liquid injection and SPME injection, the cross-sectional area between the column inserted into the liner and the inside of the liner is the most important factor in sample transfer. Figure 6.19 is the schematic diagram of the cross-sectional

Calibration

191 On-column syringe, 7.5 cm × 0.21 mm needle

Hamilton 1701 10 µL syringe

Varian 1093 ID 0.8 mm SPI liner

Varian 1093 ID 0.8 mm SPI liner

Varian 1079 ID 0.8 mm SPME liner Split outlet

Wool

ID 0.53 mm OD 0.67 mm Pre-column

(A)

ID 0.53 mm OD 0.68 mm Pre-column

ID 0.25 mm OD 0.34 mm Column

(B)

(C)

Figure 6.18 Comparison of different liquid injection methods: (A) on-column injection, (B) SPME liner injection and (C) SPI liner injection.

Figure 6.19 Cross-sectional area of the space between the column and the liner with different o.d. columns inside an SPME liner.

area of the space between the column and the liner, when different outside-diameter (o.d.) columns are inserted inside an SPME liner. The best way to avoid the deviation caused by calibration with liquid injection in SPME quantitative analysis is the complete transfer of the analytes into the GC column. PTV injection with a direct-injection (DI) liner (SPI, Uniliners, etc.) is considered the best choice to achieve this objective in the current study, as shown in Figure 6.18C. A sample introduced using the DI liner can have high sample

192

Handbook of Solid Phase Microextraction

Table 6.1 Average Sample Transfer Efficiencies for Liquid Injection and SPME Injection with i.d. 0.8-mm SPME Liner151 Injection Method

Average sample transfer efficiency

Liquid Injection

SPME Injection

With Wool, With Wool, PTV, Column PTV, o.d. 0.34 mm Column o.d. 0.67 mm

Z 5 0.8 cm, Column Flow 1 mL/min, Column o.d. 0.67 mm

Z 5 0.8 cm, Column Flow 2 mL/min, Column o.d. 0.34 mm

0.93

0.95

0.95

0.94

transfer efficiency, for both the liquid injection and the SPME fibre injection, and sample introduction can be performed with a common autosampler and a regular syringe for liquid injection. When DI liners are not available for the injectors, Table 6.1 summarises the average sample transfer efficiencies for liquid injection and SPME injection when an SPME liner was used. The average sample transfer efficiencies were close to 95%, for both liquid injection and SPME injection. The close sample transfer efficiencies suggest that it is acceptable to calibrate SPME injection by liquid injection with common SPME liners.

6.7

Summary

Devices based on SPME are able to integrate sampling with sample preparation and sample introduction. They are simple, but they require careful calibration to deliver quantitative data. There are several calibration strategies that can be used to obtain accurate and precise analytical data. Table 6.2 presents SPME calibration methods and their main advantages and disadvantages. Quantification using calibration curves does not require extensive sample preparation, but the sampling procedure and chromatographic conditions must remain constant for both the sample and the standard solutions, and if there are matrix effects, a blank sample matrix is necessary. The standard addition method needs only a small number of samples and is suitable for the analysis of unknown samples with variable compositions, but it requires extensive sample preparation, particularly if the number of target analytes is large. An internal standard can be used to compensate for matrix effects, losses of analytes during sample preparation and irreproducibility in parameters, such as sample injection in GC. However, sometimes a suitable internal standard for complex samples is not easy to find because the compound should be different from the analytes but well resolved during the chromatographic separation. A separate sampling step is typically necessary to place the sample in a container prior to the addition of the calibrant.

Calibration

Table 6.2 The Proposed SPME Calibration Methods and Their Main Advantages and Disadvantages Calibration Method

Advantages

Disadvantages

Traditional

External standard

Do not require extensive sample preparation

Standard addition

Blank sample matrixes for calibration should be available Sampling procedure, chromatographic conditions must remain constant Extensive sample preparation and analysis for large number of samples

Appropriate for the sample compositions, unknown and complex (correct sample matrix effects) Matrix effects, losses of analytes during sample Suitable internal standards for complex samples preparation and irreproducibility in parameters are not easy to find. Isotope-labelled standards such as sample injection in GC/LC can be are expensive and not available for all compensated analytes of interest The concentration of the analytes can be The distribution coefficients of the analytes calculated by the amount of the analytes between the fibre coating and the sample extracted by SPME matrix (K) should be known or determined When sample volume is very large, e.g. field sampling, the amount of extracted analytes is independent on the sample volume The concentration of the target analyte can be Only suitable for small sample volumes and very large distribution coefficients, or requires calculated with the amount of analyte special devices or methods to achieve extracted by the fibre coating and the volume of the sample

Internal standard

Equilibrium extraction

Exhaustive extraction

(Continued) 193

194

Table 6.2 (Continued) Calibration Method

Advantages

Disadvantages

Diffusionbased

Suitable for TWA sampling The sampling rate is independent of the face velocity High sampling rate and short sampling time minimised the competitive effect for solid coating Suitable for on-site sampling, where the construction of calibration curve and addition of standard are difficult Suitable for in vivo and TWA sampling, especially where the convection and concentrations of analytes always change Does not need standard loading The concentrations of all extracted analytes in the sample can be calculated

The sorbent should be zero sink for target analytes The sample rate for water sampling is very low

Fick’s first law of diffusion

Interface model and cross-flow model

Kinetic calibration with standard

Need standard loading K value should be known or determined Need sampling two times, and the conditions for sampling should keep constant Unsuitable for long-term monitoring K value should be known or determined

Handbook of Solid Phase Microextraction

Standard-free kinetic calibration

The flow velocity of sampling matrix should be controlled or determined Limited to the linear sampling regime

Calibration

195

Equilibrium extraction of SPME is widely used for on-site sampling. The advantages of this method are lower variability induced by the sampling times and higher sensitivity compared to the non-equilibrium method. However, the disadvantage of this method is the length of time required to reach equilibrium, especially for compounds with lower volatility or for field water sampling. An air pump or fan is normally used for dynamic air sampling, to shorten the sampling time. For field water sampling, because agitation of the sample is inconvenient, rotated extraction phase samplers were developed. The advantage of the SPME diffusion-based calibration methods (including the calibration methods based on Fick’s first law of diffusion, on the interface model and on the cross-flow model) is that the concentration of target analytes in the sample can be directly calculated using the amount of analytes extracted by the extraction phase and no calibration curve or internal standard is required. This characteristic makes diffusion-based calibration methods more suitable for on-site quantitative analysis, compared with traditional calibration methods. The fibre-retracted (fibre-in-needle) SPME devices can be easily calibrated based on Fick’s first law of diffusion. The results obtained with this type of SPME device are independent of the face velocity, affected only slightly by temperature, and the position of the SPME fibre can be easily adjusted to different diffusion path lengths in order to adapt to a wide range of analyte concentrations. This approach is more desirable for long-term field sampling, especially where the convection conditions of air and water are difficult to measure and calibrate. But when the concentration of the target analyte is very low (e.g. in water), the necessary sampling time for this type of SPME device can be very long, in order to obtain an adequate amount of extracted analyte, because the diffusion coefficients of molecules are much lower in water than in air. Diffusion-based calibration through an interface model and a cross-flow model are more suitable for fast sampling. These calibration methods require samples with a flowing medium. The sample velocity must be known and controlled for which requires additional equipment. Another approach is to use the rotated extraction phase devices to fix the convection conditions. Recently proposed calibration methods, including the in-fibre standardisation technique and standard-free kinetic calibration method, can be used for the quantification of passive sampling for the entire regime. Both approaches are preequilibrium techniques, and the analyte concentrations are calculated with an equilibrium equation, which requires that distribution coefficients of analytes between the extraction phase and sample matrix to be known. When the in-fibre standardisation technique is used for quantification, the physicochemical properties of the preloaded standard and the target analyte should be similar. The results obtained before the sampling reaches equilibrium are representative of the TWA concentrations of the target analytes. Using the standard-free kinetic calibration method, the sampling time for equilibrium extraction can be markedly shortened. This calibration method can be used for the entire sampling period without considering whether the system reaches equilibrium. This makes it desirable for situations when the equilibrium time is unknown or when several compounds are measured simultaneously.

196

Handbook of Solid Phase Microextraction

References 1. CL Arthur & J Pawliszyn, Anal Chem 62 (1990) 2145 2. CL Arthur, LM Killam, KD Buchholz, J Pawliszyn & JR Berg, Anal Chem 64 (1992) 1960 3. J Pawliszyn, Solid Phase Microextraction
Theory and Practice (1997) Wiley
VCH: New York, NY 4. J Pawliszyn, Ed, Applications of Solid Phase Microextraction (1999) RSC: Cornwall 5. EY Zeng & JA Noblet, Environ Sci Technol 36 (2002) 3385 6. G Ouyang & J Pawliszyn, Trends Anal Chem 25 (2006) 693 7. G Ouyang & J Pawliszyn, Anal Bioanal Chem 386 (2006) 1059 8. B Tang & U Isacsson, Energy Fuel 22 (2008) 1425 9. JA Koziel & I Novak, Trends Anal Chem 21 (2002) 840 10. LJ Krutz, SA Senseman & AS Sciumbato, J Chromatogr A 999 (2003) 103 11. HH Jelen, J Chromatogr Sci 44 (2006) 399 12. JS Camara, MA Alves & JC Marques, Anal Chim Acta 555 (2006) 191 13. W Wardencki, M Michulec & J Curylo, Int J Food Sci Technol 39 (2004) 703 14. F Belliardo, C Bicchi, C Cordero, E Liberto, P Rubiolo & B Sgorbini, J Chromatogr Sci 44 (2006) 416 15. AD Harmon, Food Sci Technol 115 (2002) 75 16. S Diez & JM Bayona, J Chromatogr Sci 44 (2006) 458 17. V Kaur, AK Malik & N Verma, J Sep Sci 29 (2006) 333 18. AK Malik, K Ashok, V Kaur & N Verma, Talanta 68 (2006) 842 19. F Pragst, Anal Bioanal Chem 388 (2007) 1393 20. MC Duff, SL Crump, RJ Ray, WE Cotham, S LaMont, D Beals, K Mount, RD Koons & J Leggitt, J Radioanal Nucl Chem 275 (2008) 579 21. F Musshoff & B Madea, Forensic Sci Int 165 (2007) 204 22. H Lai, P Guerra, M Joshi & JR Almirall, J Sep Sci 31 (2008) 402 23. N Alizadeh, A Mohammadi & M Tabrizchi, J Chromatogr A 1183 (2008) 21 24. KG Furton, J Wang, Y Hsu, J Walton & JR Almirall, J Chromatogr Sci 38 (2000) 297 25. FM Musteata & J Pawliszyn, J Biochem Biophys Methods 70 (2007) 181 26. FM Musteata & J Pawliszyn, Trends Anal Chem 26 (2007) 36 27. G Theodoridis & GJ de Jong, Adv Chromatogr 43 (2005) 231 28. T Kumazawa, XP Lee, K Sato & O Suzuki, Anal Chim Acta 492 (2003) 49 29. Z Liu & J Pawliszyn, J Chromatogr Sci 44 (2006) 366 30. EE Stashenko & JR Martinez, Trends Anal Chem 23 (2004) 553 31. B Zygmunt, A Zaborowska, J Swiatlowska & J Namiesnik, Curr Org Chem 11 (2007) 241 32. MB Heringa & JL Hermens, Trends Anal Chem 22 (2003) 575 33. X Zhang, A Es-Haghi, FM Musteata, G Ouyang & J Pawliszyn, Anal Chem 79 (2007) 4507 34. J Ai, Anal Chem 69 (1997) 1230 35. J Ai, Anal Chem 69 (1997) 3260 36. H Bagheri, E Babanezhad & F Khalilian, Anal Chim Acta 616 (2008) 49 37. J Zeng, B Yu, W Chen, Z Lin, L Zhang, Z Lin, X Chen & X Wang, J Chromatogr A 1188 (2008) 26 38. K Saito, K Okamura & H Kataoka, J Chromatogr A 1186 (2008) 434 39. MM Zheng, B Lin & YQ Feng, J Chromatogr A 1164 (2007) 48

Calibration

197

40. P Mocho, V Larroque & V Desauziers, Anal Bioanal Chem 388 (2007) 147 41. J Januszkiewicz, H Sabik, S Azarnia & B Lee, J Chromatogr A 1195 (2008) 16 42. LM Ravelo-Perez, J Hernandez-Borges, TM Borges-Miquel & MA Rodriguez-Delgado, Electrophoresis 28 (2007) 4072 43. H Kataoka, Y Terada, R Inoue & K Mitani, J Chromatogr A 1155 (2007) 100 44. S Insa, E Besalu, V Salvado & E Antico, J Sep Sci 30 (2007) 722 45. K Mitani, M Fujioka, A Uchida & H Kataoka, J Chromatogr A 1146 (2007) 61 46. NC Bouvier-Brown, R Holzinger, K Palitzsch & AH Goldstein, J Chromatogr A 1161 (2007) 113 47. B Lin, MM Zheng, SC Ng & YQ Feng, Electrophoresis 28 (2007) 2771 48. CD Rodrigues de Oliveira, M Yonamine & RLM Moreau, J Sep Sci 30 (2007) 128 49. J Koziel, M Jia, A Khaled, J Noah & J Pawliszyn, Anal Chim Acta 400 (1999) 153 50. J Koziel, J Noah & J Pawliszyn, Environ Sci Technol 35 (2001) 1481 51. M Jia, J Koziel & J Pawliszyn, Field Anal Chem Technol 4 (2000) 73 52. A Sanusi, F Ferrari, M Millet & M Montury, J Environ Monit 5 (2003) 574 53. F Ferrari, A Sanusi, M Millet & M Montury, Anal Bioanal Chem 379 (2004) 476 54. E Davoli, ML Gangai, L Morselli & D Tonelli, Chemosphere 51 (2003) 357 55. SN Zhou, G Ouyang & J Pawliszyn, J Chromatogr A 1196
1197 (2008) 46 56. LS De Jager, GA Perfetti & GW Diachenko, J Chromatogr A 1192 (2008) 36 57. D Saison, DP De Schutter, F Delvaux & FR Delvaux, J Chromatogr A 1190 (2008) 342 58. L Carrasco, S Diez & JM Bayona, J Chromatogr A 1174 (2007) 2 59. N Campillo, R Penalver & M Hernandez-Cordoba, J Chromatogr A 1125 (2006) 31 60. X Li, Z Zeng & Y Xu, Anal Bioanal Chem 384 (2006) 1428 61. J Iglesias & I Medina, J Chromatogr A 1192 (2008) 9 62. LK Silva, CR Wilburn, MA Bonin, MM Smith, A Reese, L Ashley & BC Blount, J Anal Toxicol 32 (2008) 273 63. B Plutowska & W Wardencki, Anal Chim Acta 613 (2008) 64 64. A Sanchez, S Millan, MC Sampedro, N Unceta, E Rodriguez, MA Goicolea & RJ Barrio, J Chromatogr A 1177 (2008) 170 65. C Alves, AJ Santos-Neto, C Fernandes, JC Rodrigues & FM Lancas, J Mass Spectrom 42 (2007) 1342 66. J Palau, A Soler, P Teixidor & R Aravena, J Chromatogr A 1163 (2007) 260 67. B Shurmer & J Pawliszyn, Anal Chem 72 (2000) 3660 68. VH Niri & J Pawliszyn, Analyst 132 (2007) 425 69. K Kolar, M Ciganek & J Malecha, J Chromatogr A 1029 (2004) 263 70. RA Doong & SM Chang, Anal Chem 72 (2000) 3647 71. A Paschke & P Popp, J Chromatogr A 999 (2003) 35 72. J Poerschmann, T Gorecki & FD Kopinke, Environ Sci Technol 34 (2000) 3824 73. TL der Laak, FM Busser & JLM Hermens, Anal Chem 80 (2008) 3859 74. ZY Yang, Ze-Yu, EY Zeng, H Xia, JZ Wang, BX Mai & KA Maruya, J Chromatogr A 1116 (2006) 240 75. EY Zeng, D Tsukada, JA Noblet & J Peng, J Chromatogr A 1066 (2005) 165 76. P Mayer, WHJ Vaes & JLM Hermens, Anal Chem 72 (2000) 459 77. ZY Yang, D Greenstein, EY Zeng & KA Maruya, J Chromatogr A 1148 (2007) 23 78. M Polo, V Casas, M Llompart, C Garcia-Jares & R Cela, J Chromatogr A 1124 (2006) 121 79. H Xia, M Xie, Z Yang & Y Zeng, Fenxi Ceshi Xuebao 27 (2008) 148 80. T Zimmermann, WJ Ensinger & TC Schmidt, J Chromatogr A 1102 (2006) 51 81. U Kotowska, K Garbowska & VA Isidorov, Anal Chim Acta 560 (2006) 110

198

82. 83. 84. 85. 86. 87. 88. 89. 90. 91. 92. 93. 94. 95. 96. 97. 98. 99. 100. 101. 102. 103. 104. 105. 106. 107. 108. 109. 110. 111. 112. 113. 114. 115. 116. 117. 118. 119. 120. 121. 122. 123. 124. 125. 126.

Handbook of Solid Phase Microextraction

VA Isidorov & VT Vinogorova, J Chromatogr A 1077 (2005) 195 JA Koziel, PA Martos & J Pawliszyn, J Chromatogr A 1025 (2004) 3 G Ouyang, Y Chen & J Pawliszyn, J Chromatogr A 1105 (2006) 176 P Martos & J Pawliszyn, Anal Chem 69 (1997) 206 P Martos, A Saraullo & J Pawliszyn, Anal Chem 69 (1997) 402 S Isetun, U Nilsson & A Colmsjo, Anal Bioanal Chem 380 (2004) 319 S Isetun & U Nilsson, Analyst 130 (2005) 94 V Larroque, V Desauziers & P Mocho, J Environ Monit 8 (2006) 106 P Mayer, W Vaes, F Wijnker, K Legierse, R Kraaij, J Tolls & J Hermens, Environ Sci Technol 34 (2000) 5177 L van der Wal, T Jager, R Fleuren, A Barendregt, T Sinnige, C Gestel & J Hermens, Environ Sci Technol 38 (2004) 4842 EY Zeng, D Tsukada & DW Diehl, Environ Sci Technol 38 (2004) 5737 J Conder, T La Point, G Lotufo & J Steevens, Environ Sci Technol 37 (2003) 1625 Z Qin, L Bragg, G Ouyang & J Pawliszyn, J Chromatogr A 1196
1197 (2008) 89 Z Zhang & J Pawliszyn, Anal Chem 67 (1995) 34 Y Chen & J Pawliszyn, Anal Chem 78 (2006) 5222 AR Ghiasvand, S Hosseinzadeh & J Pawliszyn, J Chromatogr A 1124 (2006) 35 E Carasek, E Cudjoe & J Pawliszyn, J Chromatogr A 1138 (2007) 10 E Carasek & J Pawliszyn, J Agric Food Chem 54 (2006) 8688 Y Chen, F Begnaud, A Chaintreau & J Pawliszyn, J Sep Sci 30 (2007) 1037 Y Chen, F Begnaud, A Chaintreau & J Pawliszyn, Flavour Frag J 21 (2006) 822 AR Ghiasvand, L Setkova & J Pawliszyn, Flavour Frag J 22 (2007) 377 JD Carrillo & MT Tena, Anal Bioanal Chem 385 (2006) 937 O Ezquerro, B Pons & MT Tena, J Chromatogr A 999 (2003) 155 M Flores & D Hernandez, J Agric Food Chem 55 (2007) 8688 O Ezquerro, G Ortiz, B Pons & MT Tena, J Chromatogr A 1035 (2004) 17 O Ezquerro, B Pons & MT Tena, J Chromatogr A 985 (2003) 247 X Deng, J Peng, B Luo, M Wei, W Hu & J Du, Anal Bioanal Chem 380 (2004) 950 JD Carrillo & MT Tena, Flavour Fragr J 21 (2006) 626 A Martinez-Urunuela, JM Gonzalez-Saiz & C Pizarro, J Chromatogr A 1089 (2005) 31 JL Gomez-Ariza, T Garcia-Barrera, F Lorenzo & R Beltran, J Chromatogr A 1112 (2006) 133 EN Fuller, PD Schettler & JC Giddings, Ind Eng Chem 58 (1966) 19 MT Tena & JD Carrillo, Trends Anal Chem 26 (2007) 206 M Groening & M Hakkarainen, J Chromatogr A 1052 (2004) 61 O Ezquerro & MT Tena, J Chromatogr A 1068 (2005) 201 G Ouyang & J Pawliszyn, J Chromatogr A 1168 (2007) 226 G Ouyang, W Zhao, M Alaee & J Pawliszyn, J Chromatogr A 1138 (2007) 42 PA Martos & J Pawliszyn, Anal Chem 71 (1999) 1513 W Hayduk & H Laudie, AIChE J 20 (1974) 611 Y Chen & J Pawliszyn, Anal Chem 75 (2003) 2004 G Ouyang, Y Chen & J Pawliszyn, Anal Chem 77 (2005) 7319 A Khaled & J Pawliszyn, J Chromatogr A 892 (2000) 455 IS Lee & SW Tsai, Anal Chim Acta 610 (2008) 149 CY Chen, C Hsiech & JM Lin, J Chromatogr A 1137 (2006) 138 A Paschke, B Vrana, P Popp & G Schueuermann, Environ Pollut 144 (2006) 414 Y Chen & J Pawliszyn, Anal Chem 76 (2004) 6823

Calibration

199

127. G Ouyang, W Zhao, L Bragg, Z Qin, M Alaee & J Pawliszyn, Environ Sci Technol 41 (2007) 4026 128. F Augusto, J Koziel & J Pawliszyn, Anal Chem 73 (2001) 481 129. K Sukola, J Koziel, F Augusto & J Pawliszyn, Anal Chem 73 (2001) 13 130. H Carslaw & J Jaeger, Conduction of Heat in Solids (1986) Clarendon Press: Oxford 131. J Koziel, M Jia & J Pawliszyn, Anal Chem 72 (2000) 5178 132. J Pawliszyn, J Chromatogr Sci 31 (1993) 31 133. X Liu, I Bruheim, J Wu & J Pawliszyn, Anal Chem 75 (2003) 1002 134. L Tuduri, V Desauziers & JL Fanlo, Analyst 128 (2003) 1028 135. S Isetun, U Nilsson, A Colmsjo & R Johansson, Anal Bioanal Chem 378 (2004) 1847 136. A Paschke & P Popp, J Chromatogr A 1025 (2004) 11 137. Y Chen, JA Koziel & J Pawliszyn, Anal Chem 75 (2003) 6485 138. Y Chen, J O’Reilly, Y Wang & J Pawliszyn, Analyst 129 (2004) 702 139. Y Chen & J Pawliszyn, Anal Chem 76 (2004) 5807 140. G Ouyang, J Cai, X Zhang, H Li & J Pawliszyn, J Sep Sci 31 (2008) 1167 141. Y Wang, J O’Reilly, Y Chen & J Pawliszyn, J Chromatogr A 1072 (2005) 13 142. G Ouyang, W Zhao & J Pawliszyn, Anal Chem 77 (2005) 8122 143. W Zhao, G Ouyang, M Alaee & J Pawliszyn, J Chromatogr A 1124 (2006) 112 144. L Bragg, Z Qin, M Alaee & J Pawliszyn, J Chromatogr Sci 44 (2006) 317 145. FM Musteata, ML Musteata & J Pawliszyn, Clin Chem 52 (2006) 708 146. G Ouyang & J Pawliszyn, Anal Chem 78 (2006) 5783 147. SZ Zhou, X Zhang, G Ouyang, A Eshaghi & J Pawliszyn, Anal Chem 79 (2007) 1221 148. SN Zhou, W Zhao & J Pawliszyn, Anal Chem 80 (2008) 481 149. W Zhao, G Ouyang & J Pawliszyn, Analyst 132 (2007) 256 150. G Ouyang, S Cui, Z Qin & Pawliszyn, Anal Chem 81 (2009) 5629 151. G Ouyang, Y Chen, L Setkova & J Pawliszyn, J Chromatogr A 1097 (2005) 9

7 Solid-Phase Microextraction Method Development

Lucie Kudlejovaa, Sanja Risticevica and Dajana Vuckovicb a

Department of Chemistry, University of Waterloo, Waterloo, Ontario, Canada Donnelly Centre for Cellular and Biomolecular Research, University of Toronto, Toronto, Ontario, Canada

b

7.1

Introduction

This chapter describes the analytical methods used for the isolation of analytes of interest from various types of matrices, with a special emphasis on solid-phase microextraction (SPME) technique. The optimization and validation steps required for the development of various applications are described and thoroughly discussed. The sampling and sample preparation step is one of the most important and frequently the most time- and labour-consuming steps in the entire analytical method. Obtaining a representative sample is a basic prerequisite for objective analysis. For a good subsequent extraction, solid samples typically need to be properly homogenised, which often requires either drying the sample in an oven or blending it with water, ice or desiccant (e.g. Na2SO4), depending on the type of application. The isolation of the target analytes from samples (sample extraction) is the next step of the analytical method, and careful selection of an appropriate extraction method at the beginning of the analysis is very important for obtaining reliable analytical data in a reasonable time. Considering that fast separation and detection systems are more and more available in analytical laboratories, the isolation of the target analytes often becomes the time-limiting step of the whole procedure. Solvent extraction (liquid
liquid extraction, LLE, or liquid-phase microextraction, LPME),1
5 supercritical fluid extraction (SFE),3,4,6
9 solid-phase extraction (SPE) performed in the column, cartridge, disc or microwell plate format,1
3,6,7,9 stir-bar sorptive extraction (SBSE),3 ultrasonic extraction (USE),3,4,10,11 Soxhlet extraction (SOX),3,4,12 steam/solvent or vacuum distillation,4,13 cryo-extraction,14 pressurised liquid extraction (PLE), such as accelerated solvent extraction (ASE)3,4,9,10,12 and microwave-assisted extraction (MAE)3,4,7,9
11 are the most commonly used and well-established extraction techniques described or reviewed in the respective cited publications. Most of them require the use of large amounts of sample and solvents (often toxic and relatively expensive) and long extraction times to maximize recovery. In many cases, the automation of such procedures is rather problematic and requires special instrumentation. Relevant parameters of these ‘classical’ extraction techniques were reviewed elsewhere.7,12,15 Handbook of Solid Phase Microextraction. DOI: 10.1016/B978-0-12-416017-0.00007-3 © 2012 Elsevier Inc. All rights reserved.

202

Handbook of Solid Phase Microextraction

One or more clean-up and pre-concentration steps must be carried out to fulfill the sensitivity requirements of most of the detection systems. Filtration,16,17 centrifugation,16,17 adsorption fractionation,3,9 SPE,3 size-exclusion techniques, such as gel permeation chromatography (GPC),3,8,9,11,12 or combinations of these techniques are the most commonly used clean-up procedures. Reducing water or solvent volume by freeze-drying (lyophilisation),18
20 reducing the sample volume under nitrogen flow18 or rotary vacuum evaporator11,19 prior to the clean-up step or prior to the actual final analysis are the most typical pre-concentration techniques used. All these approaches are ill-suited for highly volatile or unstable compounds, which are likely to degrade or escape from the system. The recent U.S. Environmental Protection Agency (EPA) regulations concerning solvent toxicity, particularly for halogenated solvents, led to the development of solventless extraction techniques, such as headspace extraction (static, SHS or dynamic, DHS), purge and trap (P&T) and SPME. The main advantage of the SPME procedure with respect to other techniques is the fact that it uses no extraction solvent. The SPME approach is relatively simple (sampling, extraction and sample concentration are integrated in one step), fast and cheap. SPME methods are suitable for liquid, solid and gaseous samples and typically use only small sample volumes. The fibre housed in the commercial holder is portable, and the whole technique is easily automated. Except for special applications, SPME methods use existing injectors on gas chromatography (GC) instruments, and analytes are thermally desorbed into the inlet. For high-performance liquid chromatography (HPLC) or capillary electrophoresis (CE) applications, interfaces need to be adjusted when used in combination with SPME or offline desorption using a small volume of appropriate solvent can be employed. In HPLC analysis, the desorption interface typically consists of a standard six-position valve with a special fibredesorption chamber and desorption is performed into the mobile phase.9,21,22 For increased sample throughput and automation, automated in-tube SPME or offline desorption (manual or automated) using a small volume of appropriate solvent can be used. For CE applications, an on-column interface made of a Teflon block is used, enabling direct insertion of an SPME fibre into the inlet end of a separation capillary.23 A different approach for interfacing the SPME and CE techniques for protein analysis has recently been described and applied successfully.24 Analytes desorbed from an SPME fibre were transferred by electrophoretic migration into a short piece of microdialysis hollow fibre, located at the inlet of a CE system, to trap analytes with molecular weights greater than the molecular weight cut-off of the microdialysis material. An electric field with a reverse electrode polarity was then applied, and the analytes trapped in the microdialysis hollow fibre migrated into the separation capillary. In most applications, adequate sensitivity, accuracy and precision for current legislation requirements are achievable using SPME coupled to typical separation and detection systems. This has led to the widespread use of SPME for various applications including environmental (Chapter 8), food (Chapter 9), pharmaceutical (Chapter 10) and in vivo investigations of biological systems (Chapter 12).

Solid-Phase Microextraction Method Development

7.2

203

SPME Method Development
General

Several steps must be followed when developing an SPME method, including selection of the fibre coating, selection of the sampling mode, optimisation of the pre-incubation, extraction and desorption conditions and selection of the calibration method. The parameters that should potentially be considered during SPME method optimization are summarised in Figure 7.1. The following discussion of method development strategies will be divided into three sections. The current section will provide a discussion of parameters typically optimized for all SPME methods, regardless of the type of separation/detection system subsequently used to detect the analytes. Depending on the analyte characteristics, mainly its volatility, SPME can be coupled to all common separation systems [GC, HPLC, CE or supercritical fluid chromatography (SFC)] combined with conventional detectors. The choice of the separation/detection system will largely depend on the goals of the analysis and the instrumentation available within the laboratory. SPME
GC applications predominate in the literature, so Section 7.3 will discuss method development strategies used with such applications. Within the past decade, the use of SPME coupled to LC has grown considerably, so Section 7.4 will provide a detailed discussion of method optimization strategies for use with SPME
LC applications. Additional information related to the SPME method development can be obtained from Refs. 25
28.

7.2.1

Selection of the Fibre Coating

The selection of a suitable fibre coating is the first step in SPME method development. A number of fibre coating types, varying in polarity, thickness of the stationary Conditioning and/or cleaning procedures

Fibrecoating coating Fibre Extraction Extraction mode mode Separation/ Separation/ detection detection system system

Calibration method

organicmodifier modifier %%organic

Agitationmethod method Agitation

SPME

Temperature Temperature

Derivatisation Derivatization

Desorption Desorption conditions conditions

Ionicstrength strength Ionic

pH

Samplevolume volume Sample Extractiontime time Extraction

Figure 7.1 Typical parameters considered in SPME method development and optimisation.

204

Handbook of Solid Phase Microextraction

phase and coating length, are commercially available, in either manual or autosampler versions. Various fibre coatings use either adsorption (solid coatings) or absorption (liquid coatings) mechanisms to extract analytes from samples. Single-phase polydimethylsiloxane (PDMS) and polyacrylate (PA) fibres belong to the group of the absorption-based coatings. Mixed-phase fibres, such as PDMS/divinylbenzene (DVB), Carbowaxs CW/DVB, carboxen (CAR)/PDMS and ‘sandwich’ fibres, such as DVB/CAR/PDMS, use the adsorption mechanism to isolate the target analytes from the sample. CW/template resin (TPR) is a special CW/DVB fibre for HPLC applications, designed to reduce molecular weight discrimination of analytes. The properties of all SPME fibres currently commercially available are listed in Chapter 4. The classification of commercial coatings in terms of polarity and retention mechanism is included in Table 4.2, as well as a recent review article.22 Section 4.4 provides useful guidelines regarding the selection of the appropriate commercial SPME fibre for a particular application. As in the case of chromatographic column stationary phases, fibre suitability for the specific compounds of interest is determined by coating polarity. In contrast to absorptive coatings, diffusion of the analytes into the adsorptive fibre coating does not occur (see Section 2.7 in Chapter 2). Therefore, typical extraction times are significantly shorter for adsorption-based fibres than for absorption-based fibres. Nevertheless, adsorption fibres have a smaller linear dynamic range (LDR) and displacement or carryover effects are more likely to occur. It is also important to consider that larger distribution constants (K) of the target analyte and thicker coatings (df) result in longer times to achieve equilibrium. The thinnest possible fibre coating (provided satisfactory detection limits are achieved) should therefore always be used, especially if compounds of high K values are analysed. In a typical preliminary method development experiment, various types of fibre coatings are compared in terms of extraction efficiency, while extraction time and desorption conditions are kept constant. For SPME
GC applications, some of the mentioned fibres are also available in the Stableflex version (see Section 4.3.1 in Chapter 4). The supporting rod in the Stableflex fibres is constructed to be more flexible in order to minimise fibre breakage. The fibre coating is the same as in the silica-based fibres, but the coating partially bonds to the flexible core; therefore, the extraction selectivity may be slightly different.29 One of the inconvenient attributes of silica-based fibres is a relatively short life of up to B100 extraction/injection cycles.30
33 Recently, new-generation SPME fibres with the same coating phases, but with a needle, a thicker plunger and a fibre core made of a titanium-based inert super-elastic alloy, have been introduced to the market to overcome these durability issues (see Section 4.3.1). This new fibre is much less likely to get damaged and has excellent shape memory properties. For some coatings, the time necessary for fibre conditioning prior to first use is reduced compared to the silica ones. The super-elastic fibres should never be conditioned for more than two consecutive hours because the alloy could change its structure and lose its tensile strength and elasticity.29 These new fibres were evaluated using samples consisting of pump oil spiked with highly volatile McReynold’s probes,

Solid-Phase Microextraction Method Development

205

and the results of headspace (HS)
SPME followed by GC
time-of-flight (TOF) MS analysis indicate significant durability improvement when using metal assemblies.34 It should be noted that since the new fibres are manufactured only in the 23-gauge needle size, a septumless injector replacement, such as a Merlin Microsealt, should be used in the GC injector to prevent septum coring, and the injection position should be very carefully inspected. Although commercial coatings perform very well for SPME
GC applications, the coatings suitable for use with LC applications require different coating characteristics, which are summarised in Figure 7.2. Among commercially available fibres, PDMS (100, 30 or 7 μm), PDMS/DVB (60 μm), PA (85 μm) and CW/TPR (50 μm) are suitable for use with HPLC applications. Among commercial fibres, the CW/TPR fibre is predominantly used for the analysis of HPLC-amenable compounds such as diphenylether herbicides,35 paclitaxel,36 tricyclic antidepressants and anticonvulsants37 and heterocyclic amines.38 Among the coatings evaluated, PDMS
DVB worked the best for the extraction of beta-carotene from foods,39 phenylurea and propanil herbicides,40 and fluoxetine and norfluoxetine.41 Although the above-mentioned coatings may be able to extract target analytes well, none of these coatings meets all (or even most) of the properties outlined in Figure 7.2. For example, these commercial fibres need to be preconditioned using a mobile phase or desorption solvent for a minimum of 30 min, or until a stable baseline is achieved. However, swelling or degradation of such fibres may occur upon exposure to various mobile phases used in HPLC. For example, Volmer and Hui42 observed stripping of CW/DVB coating in the interface due to swelling upon exposure to mobile phase. In addition, these coatings require typically long extraction times to reach equilibrium due to their thickness. To date, this has seriously limited the utility of SPME for HPLC applications, but this issue is being addressed with the emergence and development of new coatings. To address the above limitations, Supelco recently introduced a new generation of biocompatible HPLC fibres. These coatings have excellent solvent resistance (see Table 4.8 in Chapter 4), improved durability and robustness, and similar or improved extraction capacity towards HPLC-amenable compounds. In addition,

Sterilisability Robustness

Biocompatibility

Desired properties

Short extraction times Extraction capacity

Fibre-to-fibre reproducibility Low-cost

Good solvent resistance

Figure 7.2 Desirable properties of SPME coatings for use in in vivo and SPME
LC applications.

206

Handbook of Solid Phase Microextraction

they are biocompatible and sterilisable, which makes them suitable for in vivo SPME applications (Chapter 12). In general, the selection of the fibre coating should be based on the polarity of the compound. The new generation of Supelco HPLC coatings is available in several chemistries, such as octadecylsilane (C18) and a more polar C-16 amide. These perform very well for the extraction of compounds with intermediate polarity. For the extraction of very polar or ionic analytes, such as drug metabolites, peptides and proteins, the preparation of various experimental coatings is described in literature37,43
45 and Section 3.4 in Chapter 3. If the extraction capacity of fibre SPME is insufficient for a particular application, one way to enhance sensitivity is to use other SPME configurations, such as thin-film, or stir bars that contain larger volumes of the extraction phase. The amount of analyte extracted is proportional to the volume or surface area of the extraction phase. For example, Hu et al.46 show 10- to 30-fold improvement in the extraction efficiency of estrogens when using a PDMS/β-cyclodextrin-coated stir bar rather than a fibre coated with the same extraction phase. Another way to increase analytical sensitivity is to use specific coatings, such as immunoaffinity and molecularly imprinted polymer (MIP) coatings, described in Section 3.4.2 in Chapter 3.

7.2.2

Selection of the Sampling Mode

The SPME method offers the analyst three general modes: (i) headspace (HS) extraction, (ii) direct immersion (DI) and (iii) membrane-protected SPME, as shown in Figure 2.1 in Chapter 2. Both DI and HS sampling modes of SPME are used extensively in SPME
GC applications. While DI
SPME is more suitable for gaseous or simple liquid sample matrices, HS
SPME is preferentially used for extraction from complex liquid and solid samples. For very dirty samples, only HS mode or membrane-protected SPME should be used. HS extraction is more suitable for analytes of rather high volatility and low polarity. The opposite is true of DI extraction. Because the analyte’s affinity for the matrix also plays an important role, HS extraction might become more complicated for some analytes in dirty samples. In such cases, matrix modifications, such as pH adjustment or salt addition, should be performed in order to improve transfer of the target compounds from the matrix to the headspace above the sample. Typically, for SPME
LC applications, the analytes of interest are not sufficiently volatile, so DI is used. If pH adjustment or salt addition is used with DI
SPME, it is important to ensure that the coating is stable under the extraction conditions. Most commercial coatings are stable at pH values of 2
11 (Table 4.4). pH adjustments outside the recommended range should be avoided. Membrane-protected SPME is rarely used for SPME
LC applications. The kinetics of membrane-protected SPME is drastically slower as compared to DI
SPME because the analytes must diffuse through the membrane prior to partitioning or adsorption into the coating. Despite the slow kinetics, membrane-protected SPME

Solid-Phase Microextraction Method Development

207

can be a very useful method for some important applications. For instance, membrane-protected SPME was successfully used to determine the free concentration of active drug (paclitaxel) in liposome formulations.36 In this study, a CW/TPR fibre was used and a 10-nm membrane (with a MW cut-off of 15 kDa) placed around it in order to block liposomes from fouling the coating, while the analytes were capable of diffusing through the membrane to adsorb onto the coating. The advantages of SPME over traditional techniques used for this type of analysis consist in simplicity, shorter analysis times and more accurate results because the equilibrium between free and bound species is undisturbed during the analysis. In this particular instance, a comparative study using equilibrium dialysis could not even be carried out because the stability of the liposome formulation was only about 4 h, and the time required to reach dialysis equilibrium was too long.

7.2.3

Agitation Method Selection

Agitation of the sample assists the mass transport between the sample and the fibre coating. The time required to reach equilibrium can be reduced by using an agitation method. The more effective the stirring, the shorter the extraction times required to achieve equilibrium or satisfactory sensitivity in non-equilibrium extractions. Theoretically, if perfect agitation is achieved, the time required to reach equilibrium would depend only on the fibre geometry and the diffusion coefficients of the analyte in the sample (see also Section 2.4 in Chapter 2). This means that thinner coatings will have significantly shorter equilibration times. For thinner coatings, the impact of the boundary layer on the rate of mass transfer is diminished, especially for analytes with a low affinity towards the extraction phase.47 As a result, extraction time profiles obtained using thin polypyrrole (PPY) coatings were equivalent under both static and maximum magnetic stirring conditions, indicating that the use of agitation may not be an important factor for such thin coatings. Various agitation methods can be used in SPME, depending on the type of application: (i) magnetic stirring, (ii) intrusive stirring, (iii) needle vibration, (iv) moving vial (vortex stirring), (v) flow-through stirring, (vi) sonication and (vii) orbital shaking. Each type of agitation has particular advantages and disadvantages. For example, orbital shaking is preferred for high-throughput SPME applications using multi-well plates and Concept 96 autosampler (see Section 5.2.2 in Chapter 5) because it can provide most uniform agitation in all wells without inadvertent heating of samples. Magnetic stirring is simple, with no sophisticated equipment needed. Nevertheless, a stir bar must be inserted into the sample vial, and this can introduce unwanted interferences into the sample. Moreover, the stirring plate can cause heating of the vial that includes the sample, leading to an decrease of the distribution constant value and, consequently, decrease of the extraction efficiency. For intrusive stirring, the performance characteristics are usually very good; however, it is difficult to seal the sample vessel, and therefore, this technique is not very useful for volatile compounds.

208

Handbook of Solid Phase Microextraction

The needle vibration technique is especially useful for samples containing only trace amounts of the target analytes. Such samples are sensitive to interferences potentially added with the stir bar, and fibre movement is a good alternative. This technique is applicable to small sample volumes and might exert excessive stress on the needle and fibre coating. With the alternative approach
moving the sample vial
good performance can be achieved easily, with a significant advantage over the needle vibration technique: the small sample volume limitation does not apply. Super-elastic fibres may be suitable as robust substitutes for silica assemblies in the above applications. Flow-through systems offer excellent agitation, especially at rapid sample flow rates. This technique requires additional equipment, as well as special care to avoid cross-contamination and to ensure constant sample flow rates. Another powerful agitation technique is sonication. Equilibrium can be typically reached using very short extraction times. This technique’s fundamental disadvantage is that the liquid in the sonication bath warms up, heating the sample and decreasing extraction efficiency. Consequently, the liquid needs to be replaced often, which can lead to unwanted reproducibility issues. For example, although sonication provided faster equilibration times, Musteata et al.48 opted to use orbital shaking because they found that sonication significantly reduced the lifetime of the fibres. Static extraction (no agitation applied) is reserved for in vivo or other applications, where agitation or the control of agitation rate is difficult or impossible. For example, for intravenous in vivo applications, blood flow during extraction is variable, so the use of static conditions during in vitro method development represents the worst-case scenario. If a method is being developed for in vivo applications, it is simplest to determine the equilibration time under static conditions. According to SPME theory (see Section 2.4 in Chapter 2), if any agitation is applied, the resulting equilibrium extraction time must be shorter than the equilibrium extraction time under static conditions. This concept is illustrated in Figure 7.3, which shows extraction time profiles for three drugs under a low flow rate (75 mL/min) and a high flow rate (320 mL/min).45 At the low flow rate, a 10-min extraction time is required to reach equilibrium, whereas at the high flow rate, a 5-min time is sufficient to reach equilibrium. However, the total amount of drug extracted at the equilibrium is the same, regardless of the flow rate. Therefore, reliable quantitative results can be obtained from SPME, even if agitation conditions are not kept constant during the extraction, provided that equilibrium is reached. This allows the use of SPME for on-site analysis and in vivo applications for which it is impossible to control the flow rate during the experiment. If the time required to reach equilibrium under the worst-case scenario (static or low flow rate) conditions is impractically long, pre-equilibrium SPME can be used, as long as the kinetic calibration method is applied (see Chapter 6).

7.2.4

Optimization of Sample Volume

A method’s sensitivity is directly proportional to the number of moles extracted from the analysed sample at equilibrium (n). The amount of analyte extracted increases with the sample size up to a point, after which the sensitivity does not

(A)

Extracted amount (pg)

Solid-Phase Microextraction Method Development

350 300 250 200 150 100 50 0 0

5

10 15 Extraction time (min)

Extracted amount (pg)

Diazepam

(B)

209

Nordiazepam

20

Figure 7.3 Extraction time profiles obtained for diazepam, nordiazepam and oxazepam from a PBS buffer using a flow-through system at different agitation rates: (A) 75 mL/min; (B) 320 mL/min.45 (Source: Reprinted with 25 permission from Wiley.)

Oxazepam

350 300 250 200 150 100 50 0 0

5

10 15 Extraction time (min)

Diazepam

Nordiazepam

20

25

Oxazepam

increase with further increases in sample volume (Vs). Theoretically, optimum Vs can be selected based on the estimated sample/HS/coating distribution constant. In practice, the experimental arrangement is often limited by the size of sample vials (dictated by autosampler characteristics) or by the sample volume available. In the case of laboratory experiments using an autosampler with the typical headspace vial size (1.5
20 mL), the amount of analyte extracted is typically influenced by the volume of sample in the vial. The reverse is true for the on-site applications, where the sample (e.g. water in the river) is usually very large and the amount of analyte extracted is negligible and independent of Vs. The sample volume that would result in 50% depletion of the initial concentration can be calculated with Eq. (7.3), derived from the basic SPME Eq. (7.1) (see also Eq. (2.2) in Chapter 2): n5

KUc0 UVs UVf Vs 1 KUVf

ð7:1Þ

where K is the fibre/sample distribution constant, c0 is the initial concentration in the sample, and Vs and Vf are the sample volume and volume of the fibre coating, respectively. Therefore: 0:5Un 5 0:5Uc0 UVs 5

KUc0 UVs UVf Vs 1 KUVf

ð7:2Þ

210

Handbook of Solid Phase Microextraction

Expressing the sample volume Vs, this equation can be simplified to Vs;50% 5 KUVf

ð7:3Þ

Assuming the PDMS 100-μm fibre coating volume is 0.65 μL, after the substitution for Vf, Eq. (7.3) becomes Vs;50% ½mL 5 0:65U1023 UK

ð7:4Þ

Accordingly, the Vs,50% for an analyte with a distribution constant of 1,000 is 0.65 mL, and Vs will be 650 mL for an analyte with a distribution constant of 1,000,000, respectively. Similarly, assuming that the amount extracted is considered negligible when it is #1% of the amount initially present in the sample, the limiting sample volume is 64.5 mL for K 5 1,000 and 64.5 L for K 5 1,000,000. The higher the distribution constant of the analyte of interest, the more pronounced is the sample volume effect. The Vs influence on extraction rate is dramatic mainly for small volumes of sample. It should also be noted that the variation is less significant for thinner coatings and the amount of analyte extracted becomes sample volume-independent, especially for compounds with relatively small K values.

7.2.5

Determination of Extraction Time Profile and Selection of Extraction Time

SPME, as a measure of free concentration of analytes in samples, is an equilibrium extraction technique (see also Section 2.4 in Chapter 2). Compared to exhaustive approaches, such as LLE, SOX and so on, equilibrium techniques are typically simpler. The principle of SPME allows for analyte pre-concentration and, usually, there is no need for clean-up steps. It should be noted that exhaustive SPME extraction could also be achieved for very small sample volumes or using the recently introduced internally cooled SPME fibre (CF).49 In most studies, sample extraction becomes the time-limiting step of the SPME procedure (usually more time demanding than sample pre-incubation, analyte desorption or fibre bake-out). Therefore, selection of the optimum extraction time is one of the critical steps in SPME method development. The time required to reach equilibrium is independent of sample concentration, so any sample concentration can be used to construct the extraction time profile. Extraction time selection is always a compromise between the length, sensitivity and repeatability of the analytical method. Equilibrium extraction provides the highest sensitivity, but in most SPME
LC applications, pre-equilibrium conditions are used because equilibrium extraction times tend to be much longer than for SPME
GC and are thus impractical. Both equilibrium and pre-equilibrium extractions need precise and perfectly repeatable timing, although for the latter procedures, timing is more critical. Perfect timing, necessary particularly for the pre-equilibrium extraction approach, can be guaranteed by using an autosampler. In the case of pre-equilibrium extraction,

Solid-Phase Microextraction Method Development

Amount extracted (ng)

200

211

Pre-equilibrium extraction Equilibrium extraction

160 Relative error B

120

A >> B 80 Relative error A 40 0 0

5

10

15 20 Extraction time (min)

25

30

53

Figure 7.4 Model extraction profile, showing the relative error when the extraction time is on the steep slope of the curve (relative error A) and near equilibrium (relative error B).

the longer the extraction times and the less steep the extraction profile curve slope, the smaller the relative errors that occur. When the chosen extraction time is in the steep area of the extraction profile curve, a small error in timing may cause much higher relative errors in analyte sorption, as compared to extraction times selected from the latter part or the equilibrium part of the curve. For a better understanding of this effect, Figure 7.4 shows a model extraction profile and the relative error caused by a 1-min difference in timing.

7.2.6

Determination of the Distribution Constant

The fibre coating/sample distribution constant (Kfs) is calculated for better understanding the experiment, for helping with method optimisation, or for calibration purposes. For some calibration approaches, it is not necessary to calculate Kfs (see Chapter 6). The Kfs constant for SPME in the DI mode is calculated from Eq. (7.5): Kfs 5

nUVs Vf Uðc0 UVs 2 nÞ

ð7:5Þ

If the sample volume Vs is large, the equation can be simplified to Kfs 5

n Vf Uc0

ð7:6Þ

The Kfs constant for HS
SPME is more complex and can be calculated from Eq. (7.7): n5

Kfs UVf UVs Uc0 Kfs UVf 1 Khs UVh 1 Vs

ð7:7Þ

212

Handbook of Solid Phase Microextraction

therefore: Kfs 5

nUðKhs UVh 1 Vs Þ Vf Uðc0 UVs 2 nÞ

ð7:8Þ

where Khs is the constant characterising the distribution of the analyte between headspace and the fibre coating. For PDMS fibres, the constant can be calculated from the retention times of analytes separated on the PDMS GC column, following Eq. (7.9): Khs 5

KH RT

ð7:9Þ

where KH values are Henry’s constants, easily found in the literature. If this is not the case, or the KH values are not known for the particular analyte, Eqs (7.7) and (7.10) can be combined thus: Khs 5

Kfs Kfh

ð7:10Þ

to obtain Eq. (7.11) and calculate the distribution constant: Kfs 5

nUKfh UVs Kfh UVf Uc0 UVs 2 nUKfh UVf 2 nUVh

ð7:11Þ

For more details on calculating distribution constants, see the discussion in Chapter 2, Section 2.3.

7.2.7

Optimisation of Matrix Conditions

Optimisation of matrix conditions in order to achieve maximum sensitivity of the whole method is a crucial step in SPME method development and an important way of improving method performance. The main parameters that can be optimised are temperature, ionic strength, pH and addition of an organic modifier.

7.2.7.1 Temperature Increasing the extraction temperature can significantly reduce the equilibration time and make the whole procedure faster (see also Section 2.3.3 in Chapter 2). Two opposite effects are taking place if the extraction temperature is raised: (1) the positive effect is that the headspace capacity and/or analyte diffusion coefficient is increased, thus the extraction rate is enhanced and (2) the opposite effect of temperature on the distribution constant occurs (see Chapter 2). These two antagonistic phenomena must be optimised in order to achieve the highest sensitivity. For example, increasing temperatures resulted in a decrease of extraction efficiency for

Solid-Phase Microextraction Method Development

213

ochratoxin A.50 By contrast, increasing the temperature from 40 C to 60 C was found to increase extraction yields for tricyclic antidepressant analysis.37 The loss in sensitivity can also be reduced by using the internally cooled fibre technique (see Section 2.3.4 in Chapter 2 and Section 5.1.6.2 in Chapter 5) From a kinetic viewpoint, an increase in temperature during the SPME extraction process will increase the diffusion coefficients of analytes, resulting in increased rates of mass transfer into the fibre. According to thermodynamic theory, because the extraction process is exothermic, the amount of analytes extracted will decrease with an increase in temperature. Therefore, an increase in temperature will simultaneously decrease analyte recovery at equilibrium and decrease the time required to reach equilibrium. The choice of whether to use the elevated sampling temperatures will then depend on whether the main goal of the analysis is short analysis time or maximum sensitivity achievable. The amount extracted at a particular time (prior to equilibrium) at the higher temperatures will be higher than the amount extracted at the lower temperatures. However, at equilibrium, the amount extracted at low temperatures will be higher than that at high temperatures. For example, as demonstrated in Figure 2.21 in Chapter 2, at temperatures of 22 C and 40 C, methamphetamine does not reach equilibrium even after 90 min of extraction, but at 60 C and 73 C, equilibrium is reached in less than 15 min. Therefore, selecting a 60 C extraction temperature allowed a much shorter analysis time and allowed the extraction to be performed at equilibrium, which in turn resulted in better method precision, as discussed in Section 7.2.5. In this example, the sensitivity obtained at 60 C was adequate for the method’s requirements. If that were not the case, selecting 22 C and much longer extraction times would be necessary to improve the extraction efficiency. Using elevated sampling temperatures can also improve analyte recovery at short sampling times. For example, as shown in the Figure 2.21 if a rapid sampling of 5 min is desired, elevated extraction temperature is preferable because the amount extracted is much higher. From a practical point of view, after choosing 60 C as the extraction temperature, the temperature gradient between the needle tip and the sample resulted in some condensation of water, observed as a small drop on the tip of exposed fibres.51 This reduced method precision and, occasionally, extinguished the flame of the flame ionisation detector (FID). To eliminate this problem, the fibre was positioned in the headspace in such a way that the needle tip of the assembly was resting on the vial septum, and only the inner stainless steel tubing holding the fibre passed through the septum.

7.2.7.2 pH In SPME, only the undissociated/neutral species of analytes are extracted. Full conversion of the analytes into neutral forms by pH adjustment can significantly improve method sensitivity. Therefore, low pH values will improve the extraction of acidic compounds and high pH will improve extraction efficiency for basic compounds. For example, for naproxen (an acidic compound), enhanced extraction is

214

Handbook of Solid Phase Microextraction

observed at pH 3.52 Similarly, Figure 7.5 shows that the extraction efficiency of ochratoxin A (weak acidic with a pKa of 4.4) is highly pH-dependent in the range of 2.9
4.5.50 The signal is enhanced threefold simply by adjusting sample pH from 4.5 to 3.0. If the sample pH was set to B3.8, reproducible results would not be obtained because small variations in pH would cause significant differences in the amount extracted. For nearly neutral analytes, such as phenylureas and propanil, the extraction efficiency is not affected to a significant degree by variations in pH.40 Figure 7.6 shows the dependence of the amount extracted on sample pH for nine β-blockers, and alkaline pH (8.5) clearly provides improved extraction efficiencies for these analytes.53,54 Figure 7.5 pH dependence of the ochratoxin A (2 ng/mL) amount extracted by a PDMS/ DVB fibre from wine samples.50 (Source: Reprinted with permission from Elsevier.)

Area counts (a.u.)

1.0

0.8

0.6

0.4

0.2 2.8 3.0 3.2 3.4 3.6 3.8 4.0 4.2 4.4 4.6 pH

Nadolol

5.5

Pindolol Acebutolol Timolol

7

Metoprolol

pH

Oxprenolol Labetalol

8.5

Propranolol Alprenolol

10 0.0

2.0

4.0 6.0 Area counts (×106)

8.0

10.0

Figure 7.6 Effect of sample pH on the extraction efficiency of several β-blockers using in-tube SPME with Omegawax capillary.53 (Source: Reprinted with permission from American Chemical Society (r 1999).)

Solid-Phase Microextraction Method Development

215

When adjusting sample pH, HS sampling is the preferred extraction mode because direct contact of the fibre coating with sample at very low or high pH levels can damage the fibre coating. When DI sampling mode is used, extreme pH values should be avoided because they can cause degradation of the coating. For commercially available fibres, the recommended pH range is 2
11, with the exception of PDMS (100 μm), for which pH should be between 2 and 10, and polyethylene glycol (PEG), with a recommended pH range of 2
9. For example, for the tricyclic antidepressants with pKa values of B10, a pH of 12 would theoretically be optimal for extraction.37 However, because commercial fibres are not stable at such high pH values, the analysis was performed at pH 10. If pH adjustment of the samples is desired, it is important to ensure that the buffers added for pH adjustment have adequate buffering capacity. For instance, the addition of 1 mL of 0.5 M phosphate buffer, pH 12, to urine samples was not enough to adjust the urine sample pH to 12, as desired for the extraction of basic analytes, amphetamine and methamphetamine.51 pH adjustment using 1M KOH was necessary in order to achieve the desired pH adjustment. Saturation with salt or pH adjustment/buffering of samples can also help eliminate small sample-to-sample variations in ionic strength and pH, often encountered with biological samples and other complex samples, thus improving method reproducibility.

7.2.7.3 Ionic Strength Salt addition increases the ionic strength of the sample solution; thus, it increases the Kfs constant and improves sensitivity in most applications, except those involving very polar analytes. The salting-out effect causes the analyte molecules to pass more readily from the sample matrix to headspace and is commonly used for HS
SPME (see also Section 2.3.3 in Chapter 2). For many organic compounds, aqueous solubilities decrease in the presence of large amounts of salt. For compounds whose aqueous solubility does not change, the addition of salt may decrease the amount extracted by decreasing the activity coefficients of the analytes, which adversely affects the partition coefficient between the sample and the SPME coating. For example, the addition of 30% NaCl enhanced the extraction of pesticides from fruit juice,55 and the addition of 35% salt improved the extraction efficiency of five benzodiazepines from aqueous and biological samples when a CW/DVB fibre was used.56 The addition of 25% salt strongly enhanced the extraction of 11 corticosteroids from urine.42 In some cases, an increase in ionic strength improves extraction efficiency of the target analyte, and also improves extraction of interfering compounds, which is not desirable, especially if a solid (adsorbent) type of coating is used.57 Adding salt decreased the extraction of amphetamine and methamphetamine when a PDMS/DVB coating was used,58 and the extraction of ibuprofen from urine samples.59 NaCl is most often used to adjust ionic strength, but other salts can also be used. Interestingly, Sanchez-Ortega et al.60 found that addition of sodium chloride did not improve extraction efficiency for fenitrothion and its metabolites, but addition of sodium sulphate resulted in significant increases for two out of three compounds tested.

216

Handbook of Solid Phase Microextraction

For the analysis of biological samples, the addition of saturating amounts of sodium chloride minimises the variability in ionic strength between samples and yields more reliable quantitative results. Lord and Pawliszyn51 added a saturating amount of NaCl for urine analysis because they found the amount extracted to be dependent on the exact salt concentration of a particular sample. When working with whole blood samples, the addition of salt may promote clot formation, which would negatively impact the extraction due to reduced mass transfer. For such samples, a deproteinisation step should precede the addition of salt. Alternatively, the addition of sodium hydroxide to the whole blood sample promotes haemolysis and prevents clot formation.61

7.2.7.4 Presence of Organic Solvent The amount of organic solvent in samples should be kept to a minimum (see also Section 2.3.3 in Chapter 2). For example, Wu et al.62 report a drop of up to 90% in the extraction efficiency of phenylurea pesticides for samples containing 40% ethanol versus samples containing no ethanol. Typically, for optimal extraction efficiencies, organic solvent should not exceed 1
5% of the sample volume. In some cases, a variation in organic solvent content of even below 0.5% of sample volume seriously affected recoveries.51 This particular study highlights the importance of keeping the amount of organic solvent in spiked samples during method validation constant for all samples and all concentration levels tested. Otherwise, significant recovery and precision problems may be observed. For the analysis of drugs from biological fluids, the addition of small amounts of an organic modifier actually improves extraction efficiency because it releases some of the drug that is bound to matrix proteins.63 However, if the aim of the method under development is to determine free drug concentration, or to study ligand
receptor binding (Chapter 11), the addition of an organic modifier must be avoided in order to avoid disrupting the equilibrium between the free analyte fraction and the macromolecule-bound fraction in the sample.

7.2.7.5 Matrix Modifications
Concluding Comments At this stage in method development, it is advisable to readdress the decisions made earlier. Matrix modifications made in order to improve extraction efficiency could cause the new matrix to no longer be compatible with the selected fibre coating. The extraction time profile may now be different from the original one, and the headspace extraction mode might become more attractive under the modified sample conditions.

7.2.8

Analyte Derivatisation

Analyte derivatisation can be used for various reasons, such as (i) to enhance extraction efficiency due to higher partition coefficients of derivatives to the coating, (ii) to enhance detection sensitivity or (iii) to make compounds more amenable

Solid-Phase Microextraction Method Development

217

to a particular mode of analysis. For instance, if the analytes are not volatile enough to be extracted from the headspace, they can be derivatised to improve their volatility. The derivatisation agent can bring an additional source of interferences and errors into the system and the derivatisation process should be carried out only when necessary. Derivatisation is more often used in SPME
GC applications, due to the availability of a wider range of derivatising agents suitable for GC methods and the attractiveness of HS
SPME for the analysis of complex samples. The derivatisation agent can be added to the sample before sample extraction, during sample extraction or after sample extraction. Post-extraction derivatisation can only improve the chromatographic behaviour and detection properties, not the extraction efficiency. In combination with SPME, derivatisation can be performed directly in the sample, followed by SPME extraction of the derivatised species, directly within the coating as on-fibre derivatisation following extraction64 or as simultaneous derivatisation and extraction. These three approaches are summarised in Figure 7.7. Herraez-Hernandez et al.65 compared all three approaches for the derivatisation of aliphatic amines using 9-fluorenylmethyl chloroformate and SPME
LC with fluorescence detection. The best results were obtained by simultaneous extraction/ derivatisation on the fibre. This approach also offered significant time savings because the preloading of the derivatising agent required only 3 min, and the unreacted reagent did not interfere with LC analysis because it was well separated

Derivatisation strategies and SPME

Sample derivatisation followed by SPME

SPME extraction of analytes followed by on-fibre derivatisation

Simultaneous extraction/on-fibre derivatisation

Step 1 Add derivatising agent to sample solution and allow the reaction to proceed

Step 1 Expose fibre to sample solution to extract analyte(s)

Step 1 Expose fibre to derivatising agent solution to preload the reagent

Step 2 Expose fibre to sample solution to extract derivatised analytes

Step 2 Expose fibre to derivatising agent solution

Step 3 Desorb and analyse

Step 2 Expose fibre to sample solution and allow derivatisation reaction to proceed on the fibre

Step 3 Desorb and analyse

Figure 7.7 Summary of derivatisation strategies used in SPME methods.

Step 3 Desorb and analyse

218

Handbook of Solid Phase Microextraction

from the analytes of interest. A similar study of various derivatisation approaches, conducted using o-phthaldialdehyde-N-acetyl-L-cysteine as the derivatising reagent for the enantiomeric determination of amphetamines, showed that the best results were obtained when derivatisation was carried out in solution, followed by extraction of derivatives onto CW/TPR fibre.66 This result can be explained by a low affinity of the derivatising reagent for the coating. To achieve a good derivatisation reaction yield, an excess of reagent is typically desired and the low affinity of the reagent for the coating may prevent sufficient accumulation of the reagent in the coating, thus resulting in lower efficiencies of on-fibre derivatisation techniques.

7.2.9

Selection of the Calibration Method

Calibration is a process relating the measured analytical signal to the concentration of analyte. The suitability of the SPME calibration technique depends on the application, the number and complexity of samples to be analysed, as well as the availability of an MS instrument in the laboratory. Various calibration approaches suitable for use with SPME are discussed in detail in Chapter 6. Comparison of external calibration and standard addition calibration methods was carried out for the analysis of amphetamines.51 External calibration resulted in poor average accuracy, with a relative error of about 6 20%, while the standard addition method produced an average accuracy of 6 6%. Similarly, Zambonin et al.57 obtained excellent quantitative results for drug analysis from serum using the standard addition method, with excellent linearity and intra-day and inter-day precision. The proposed SPME
LC method did not even require a deproteinisation step, and excellent results were obtained despite the complex matrix studied. Kinetic calibration and external calibration methods were compared for an in vivo analysis.44 Both methods provided comparable pharmacokinetic profiles and correlated well with the results obtained for the conventional blood draw method, followed by in vitro chemical assay, as indicated by correlation factors of 0.99.44,67 An equilibrium method is insensitive to variability in blood flow and has good reproducibility and sensitivity, but it is useful only for in vivo determinations when the required equilibration time is very short. Equilibrium methods are also not recommended for use with adsorptive coatings because analyte displacement and competition may occur, causing deviations from linearity. For such coatings, short pre-equilibrium exposure times are preferable. The kinetic calibration method compensates well for variable agitation conditions during sampling and is particularly suitable for coatings requiring long equilibration times, but it suffers from poorer reproducibility. In bioanalytical applications, an important advantage of SPME-based methods over traditional sample preparation methods is the ability to quantitate both free concentration and total concentration of the drug in the biological fluid in the same experiment.68 This is accomplished by constructing two different calibration curves. One curve is constructed using a set of standard solutions dissolved in phosphate-buffered-saline buffer, pH 7.4, and allows the determination of the free concentration of the drug. The second set of drug standards is prepared in whole

Solid-Phase Microextraction Method Development

219

blood and is used to determine total drug concentration. An example of this type of application is shown in Figure 3.18 in Chapter 3.

7.2.10 Multivariate Experimental Design The systematic method development approaches discussed so far are all univariate approaches. One parameter is optimised at a time, while the remaining parameters are kept at a fixed value. Although this type of experimental approach yields satisfactory results, it does not take into account possible interactions between factors. In the presence of such interactions, a true optimal value for the two interacting variables is unlikely to be attained using an univariate optimisation approach. Therefore, to obtain truly optimal method performance while performing a minimum number of optimisation experiments, the use of multivariate experimental approaches has emerged. Multivariate optimisation is typically performed in two stages. In the first stage, factorial design is used in order to determine which factors are significant. In the second stage, the optimum values for the significant factors are determined using approaches such as Doehlert design and response surfaces. For example, LopezMonzon et al.69 used factorial design to examine the effects of extraction time, temperature and salt concentration. Temperature and salt addition were found to have the most significant effect on recovery. The authors then used response surfaces to determine the optimal values for each significant parameter. Similarly, Plackett-Burman design was used to evaluate six factors affecting the efficiency of the simultaneous on-fibre derivatisation/SPME extraction.65 Fernandes et al.41 used fractional factorial design to optimise an SPME
LC
UV method for the determination of fluoxetine and norfluoxetine in plasma. In eight experiments, the authors evaluated the influence of extraction time, temperature, ionic strength and pH. Increase in ionic strength decreased extraction efficiency, so this parameter was removed from the experimental design and a full factorial design was then carried out. Time, temperature and the interaction between time and temperature were significant for this analysis, based on the Pareto diagram obtained. Factorial design and simplex methodology were also used to optimise an SPME
LC method for the analysis of seven anticonvulsants and tricyclic antidepressants in human plasma.37 The effects of extraction time, extraction temperature and ionic strength at two levels were evaluated in a 23 factorial design. Increasing extraction time and extraction temperature had a positive effect on extraction efficiency. Optimum values for these two variables were then found using simplex experimental design. Extraction time, static desorption time and dynamic desorption time were the three variables evaluated and optimised for the analysis of heterocyclic amines by SPME
LC.38 In conclusion, the use of multivariate techniques during the method development stage of SPME methods is highly recommended because it is faster and more effective than traditional method development approaches. It requires a much smaller number of experiments and increases the likelihood of finding true optimum values for all factors tested.

220

Handbook of Solid Phase Microextraction

7.3

SPME Method Development for GC Applications

7.3.1

Optimisation of Desorption Conditions

There are several main factors influencing the time needed to desorb analytes of interest and transfer these onto the separation/detection system. In the case of thermal desorption into the GC inlet, carrier gas flow rate (defined by the geometry of the injector or, more precisely, the glass liner) and injector temperature are the main factors influencing the rate of transfer. In accordance with the following simplified equation,   1 1b log K 5 a T

ð7:12Þ

the gas-coating distribution constant K decreases as the temperature increases. For example, the K value of benzene at 25 C is B300, compared to a value of 0.4 at 250 C. Therefore, analyte affinity for the fibre placed in the injector (kept at relatively high temperatures, typically 220
300 C) is lower, and analyte transfer onto the GC column occurs immediately.

7.3.2

Determination of the LDR

Once all the previously optimised steps in the method development have been reinspected, it is time to determine the LDR. Especially in the case of classification (food) studies where analyte concentrations may significantly vary in long sequences of samples, the LDR determination is one of the important factors that should be included in method development. Due to the limited number of sorption sites on the adsorption-type coatings, the LDR of adsorption fibres is typically much smaller than that of the absorption liquid-phase fibres. However, the capacity of the fibre coating is usually sufficient. It is mostly the detector linear range that limits the LDR, not the fibre range. Comparing the two most commonly used detectors, the linearity range of an FID is typically larger (4
6 orders of magnitude) than that of the MS detector (typically 2
3 orders of magnitude). On the other hand, the MS has excellent sensitivity at lower concentrations (ppt levels), especially when operated in the MS(n) mode. Complex samples of unknown composition are typically examined using a GC(xGC)
MS(n) system.

7.3.3

Improving Method Precision

Several factors that directly influence method quality and precision will be discussed in this section, in order to give the reader several suggestions that might be useful when developing an SPME method. Most comments are related to the more-common GC analysis.

Solid-Phase Microextraction Method Development

221

1. The fibre coating should occasionally be inspected under a microscope to detect possible coating deterioration caused by conditioning of the fibre at high temperatures or presence of undesirable sample matrix components on the coating. 2. A swelling effect, caused by the presence of oily suspensions, dispersed hydrophobic humic material and so on, might occur. Swelling typically occurs once the concentration of interferences in the sample exceeds 1%. It should be noted that solid sorbents tend to be more prone to swelling than liquid coatings. 3. Vial shape, vial volume and, consequently, HS volume are other critical factors influencing method robustness and reproducibility, especially when small sampling vials are used. 4. Another issue related to vial characteristics is the possible adsorption of target analytes to the vessel walls. Silanisation or adding small amounts of methanol can solve this problem; however, these procedures may also cause a change in the distribution constant level and a subsequent sensitivity decrease. Other avenues for analyte losses caused by adsorption, absorption, permeation or desorption from fibre must be eliminated as well. 5. Sample agitation and temperature must be kept as constant as possible to avoid reproducibility issues. For magnetic stirring, heating of the stirring plate may heat up the sample, but this problem can easily be solved by isolating the vial from the plate. 6. Pre-incubation and extraction timing is critical, especially at non-equilibrium conditions. The time between extraction and analysis must also be kept as short as possible. If necessary, the fibre can be cooled using dry ice for transport to the laboratory. Portable devices are also available for in-field applications to keep analyte losses at a minimum. 7. Caution must be taken to protect the fibre needle from moisture and to keep the needle far from the sample surface to avoid splashing, especially when the HS mode is applied. The fibre’s position in the sampling vial is very important. The fibre needle should penetrate the vial septum at the place with the highest linear velocity of the agitated sample; that is, halfway between the vial centre and vial wall (or the end of the stirring bar). 8. Desorption temperature is one of the main factors optimised during method development. It should be noted that the temperature might not be identical in the whole injector chamber. There is typically a temperature gradient, with the highest level occurring in the middle of the inlet. Therefore, fibre exposure during the thermal desorption of analytes into the GC inlet should be calculated carefully so that the final coating position is in the middle of the injector. 9. Because no solvent evaporation takes place in the injector, low-volume liners should be placed in the inlet to achieve high carrier gas flow rates and to accelerate analyte transfer onto the column. Commercial products, such as SPME injection sleeves or direct injection liners (SPI liner or Uniliners), all with an internal diameter (i.d.) of 0.75
1 mm, are suitable for this purpose. The SPI liner or the Uniliner, respectively, with the narrow part in the lower section of the assembly, allows direct connection of the column to the liner, simulating on-column injection for SPME exposure. The issue of comparing liquid injection to SPME exposure using various liners is discussed in detail in Ref. 70. 10. The GC injector should be conditioned prior to fibre desorption. Either pre-punctured septa or septumless injectors should be used, especially with the 23-gauge fibre needle size, to avoid contamination from septum coring. 11. Exposure of the fibre in the GC injector must be carried out immediately after insertion of the fibre into the hot injector, and quickly enough to avoid split peaks and peak broadening, especially for volatile analytes.

222

Handbook of Solid Phase Microextraction

12. The carryover effect should be carefully inspected as a part of method development (fibre is injected, sealed with septum until the system is prepared for the next analysis and injected again). Fibre bake-out should be performed after each sample injection. 13. A thick-film stationary phase column (df of 1 μm) installed, cryo-trap or retention gap in front of the separation phase may enhance analyte focusing on the GC column. 14. The stability of the detector’s response must be established as a final prerequisite for a reproducible and reliable analytical method.

7.3.4

SPME
GC Automation

SPME can be easily automated using commercially available autosampler systems with robotic arms. The autosampler combines sample preparation, pre-incubation and extraction steps. It also includes fully automated desorption of analytes, as well as fibre bake-out in a separate chamber (needle heater) between the runs to avoid the carryover effect. For a detailed review, the reader is referred to an article by O’Reilly et al.71 which discusses developments and future challenges in the automation of SPME, and to Chapter 5 of this book.

7.4

SPME Method Development for HPLC Applications

Although GC applications of SPME currently predominate in the literature, SPME is increasingly being used for analysis of semi-volatile, non-volatile and/or thermally labile compounds, which are more amenable to HPLC analysis with ultraviolet (UV), fluorescence and electrochemical or MS detection. The main obstacle in the use of SPME for such applications has been the lack of commercially available, fully automated systems capable of performing extraction, desorption and sample introduction in a continuous fashion, without requiring operator input. As such systems become available, LC applications of SPME are expected to increase drastically. Currently, SPME
LC methods for various drugs, pesticides, contaminants, peptides, food components and additives exist in the literature and are summarised in several review articles.9,67,68,72,73 In addition, SPME methods have been developed for various in vivo applications, such as pharmacokinetic profiling and metabolomics (Chapter 12). SPME was also found to be an excellent method for drug
protein binding and plasma protein binding studies. Although method development for SPME
LC methods is generally similar to method development for SPME
GC applications, some important differences and considerations exist and will be discussed in next section. In particular, the analyst is faced with a choice of interfacing strategies. Furthermore, the optimisation of desorption conditions is of critical importance for developing a successful SPME
LC method.

Solid-Phase Microextraction Method Development

7.4.1

223

Selection of the Coupling Mode to HPLC Instrumentation

Selection of the coupling mode to HPLC instrumentation is the first step in SPME
LC method development. Currently, there are four main approaches to coupling SPME to LC: (i) a manual SPME interface, available from Supelco, or a similar home-made interface; (ii) automated in-tube SPME; (iii) manual offline desorption followed by liquid injection and (iv) high-throughput automated offline desorption using the Concept 96 autosampler, which is available commercially from PAS Technology. The advantages and disadvantages of the four SPME
LC coupling strategies are summarised in Table 7.1. Chapters 3 and 4 and a recent review article by Lord21 offer more detailed treatment of SPME
LC coupling modes and the experimental setup of these interfaces. Because method development for in-tube SPME is somewhat different than it is for fibre configurations of SPME, it will be discussed separately in Section 7.4.4.

7.4.1.1 Manual SPME Interface The commercially available, manual SPME interface consists of a standard six-port injection valve and allows for static and/or dynamic desorption of the fibre into the mobile phase or another suitable desorption solvent. This is the most suitable approach for methods requiring high sensitivity because all the analytes desorbed from the fibre are directly injected into the LC system. In current SPME
LC literature, there is some ambiguity regarding the desorption volume actually injected in the HPLC system when the commercially available Supelco interface is used. Depending on the publication, some authors quote the internal volume of the commercial manual interface to be 60 μL, while others state 200 μL.21 The actual physical cell volume is, in fact, 75 μL, but the effective desorption volume may be less than this, which is the explanation for the 60 μL value often quoted. 200 μL quotes likely correspond to the internal volume of the interface plus the volume of the tubing connecting the desorption chamber to the valve. Of course, for maximum sensitivity, it is desirable to keep the internal volume as low as possible, and lab-made interfaces with volumes as low as 14 μL have been reported.21

7.4.1.2 Automated In-Tube SPME In-tube SPME provides a fully automated approach for coupling SPME to LC, using a piece of commercially available GC capillary column or other suitable piece of coated tubing as the extraction device (see Section 5.2.1 in Chapter 5). All the steps are fully automated and controlled by the HPLC autosampler and the software. This is the best option when developing methods for analysis of relatively clean, particulate-free aqueous samples and for the analysis of semi-volatile analytes.

7.4.1.3 Offline Desorption Offline desorption is the simplest and most flexible way to couple SPME to LC analysis. This can be a very low-cost, easy-to-implement approach because no

224

Table 7.1 Summary of Advantages and Disadvantages of Various SPME
LC Coupling Strategies Advantages

Disadvantages

Manual direct SPME
LC interface

High sensitivity since all the analyte extracted is introduced into the LC instrument No solvent front peak if dynamic desorption is used Applicable to complex samples or samples containing particulates

Coating may be damaged or stripped in the interface Possible leaks and loss of sample Labour-intensive

In-tube SPME

Automated Reduced total analysis time Increased throughput Better precision and accuracy Decreased handling of biological and other hazardous sample Good sensitivity (same or better than for direct manual interface) Less likelihood of carryover

Low-throughput Not easily automated Peak broadening Possible carryover problems due to slow desorption kinetics Limited to particulate-free samples Sample preparation takes up instrument time Not applicable to in vivo and on-site analysis

Manual offline desorption

Does not require additional instrumentation Wide choice of desorption solvents available Physical damage to fibre is not likely

Offline Lack of automation

Handbook of Solid Phase Microextraction

SPME
LC Coupling Mode

Concept 96 robotic station

Full automation Highest sample throughput Wide choice of desorption solvents available Suitable for use with pre-equilibrium SPME with no loss of precision Applicable to complex samples such as whole blood with no sample pretreatment required

Possible loss of analytes due to adsorption to walls of vial or insert used for desorption Poorer sensitivity since 100% of analyte is not injected Poorer precision if pre-equilibrium SPME is used Poorer sensitivity since 100% of analyte is not injected Applicable to non-volatile analytes only Possible loss of analytes due to adsorption to walls of 96-well plates Cost

Solid-Phase Microextraction Method Development

Ability to carry out desorption of multiple fibres in parallel, reducing total analysis time

225

226

Handbook of Solid Phase Microextraction

additional instrumentation is required if the manual method is used. It also offers maximal flexibility because a wide variety of solvents and/or agitation methods can be used in order to achieve effective analyte desorption from the SPME fibre. Very recently, the full automation of offline desorption was achieved for the first time using the 96-well plate format and a commercial robotic station (Concept 96 from Pas Technology, Germany).74 This device offers significant advantages over manual desorption because it improves method precision and drastically increases sample throughput. In fact, sample throughput of more than 1,000 samples per day is possible, as demonstrated by example applications of the analysis of benzodiazepines in whole blood74 or ochratoxin A in urine.75 The availability of Concept 96 enables the use of SPME
LC in high-throughput applications for the first time, so this type of application can be expected to grow in the future. The main disadvantage of the unit is that it is applicable only to non-volatile samples because the wells are not sealed during the extraction. In comparison to automated in-tube SPME, Concept 96 is able to handle more complex samples (e.g. whole blood) without the need to perform any sample pretreatment steps. Furthermore, the extraction/desorption is performed in parallel, thus increasing sample throughput up to 100-fold.

7.4.2

Optimisation of Desorption Conditions

The optimisation of desorption conditions is a crucial part of SPME
LC method development. It is typically more challenging and time-consuming than the optimisation of the desorption process for SPME
GC applications. The process involves the optimisation of (i) desorption solvent, (ii) solvent volume and (iii) desorption time, as well as (iv) the evaluation of carryover. Complete desorption of the analyte from the fibre is more difficult to achieve in SPME
LC applications than in SPME
GC applications because the kinetics of the desorption process in the liquid phase are significantly slower than in the gas phase. The slow kinetics of desorption result in three main problems: (i) band broadening, (ii) carryover and (iii) the need for longer desorption times. Therefore, it is important to consider these potential issues when designing desorption optimisation experiments.

7.4.2.1 Desorption Using Manual SPME Interface Desorption using the manual SPME interface can be carried out in either static or dynamic mode. In dynamic mode, the flowing mobile phase is continuously passed through the interface in order to desorb the analytes from the coating. The efficiency of dynamic desorption is determined by the eluent composition (usually the mobile phase) and the flow rate. Because the kinetics of the desorption process tend to be slow, dynamic desorption may not completely desorb all the analyte in a reasonable amount of time. Thus, static desorption for a preset amount of time is often used, followed by a brief dynamic desorption step, in order to sweep the desorbed analytes onto the analytical column. In static desorption, the interface is filled with a certain amount of solvent and the fibre is allowed to sit in this solvent

Solid-Phase Microextraction Method Development

227

for a predetermined amount of time. The efficiency of static desorption is determined by desorption time and the composition of the desorption solvent. More efficient desorption will occur if a higher eluent strength is used. However, chromatographic peak shape may be distorted if the desorption solvent has a higher eluent strength than the mobile phase used for chromatographic separation. Table 7.2 shows optimised desorption and extraction conditions for several selected SPME
LC applications chosen from current literature. In most cases, static desorption times of 1
5 min are sufficient. For example, 2 min is required to desorb paclitaxel.36 However, the desorption of carbamate and phenylurea pesticides required much longer static desorption times of 15 min using 70% methanol:30% water mixture.55 Rellan et al.64 used 12-min static desorption using 60% acetonitrile:40% water to desorb anatoxin-a from a PDMS/DVB fibre. Despite extensive optimisation of the desorption process, Chou and Lee58 were not able to achieve carryover values less than 4
5%, even with long static desorption times of 15 min. To eliminate this carryover, they introduced an additional cleaning step, where the fibre was flushed with two portions of 500 μL acetonitrile. Peak broadening and tailing is more likely to occur with dynamic desorption. To minimise this problem, one can reduce the eluent strength of the mobile phase, add a static desorption step or use a trapping column between the SPME interface and the analytical column. Aresta et al.52 initially obtained quantitative recovery when using dynamic desorption for the analysis of naproxen, but the peaks obtained were too broad. The use of 10-min static desorption with the mobile phase resulted in quantitative recovery while eliminating peak broadening. If peak broadening and/or tailing remains an issue despite the optimisation of desorption conditions, a pre-column can be placed between the SPME
LC interface and the analytical column and used to focus the analytes.65,66,76 Another alternative is to heat the SPME interface in order to improve desorption mass transfer and reduce carryover.41 When interfacing manual SPME with highly sensitive detection systems such as LC
MS/MS, it is important to evaluate two types of carryover: (1) the carryover on the fibre and (2) the carryover in the interface itself. If carryover is observed in the interface, it can be reduced or eliminated by introducing a washing step to flush the interface between analyses.47 In contrast, Cardenes et al.38 found carryover of up to 15% even after 15-min static desorption and four flushes of interface and fibre. They found that a minimum of six flushes of both interface and fibre were necessary to reduce carryover for all compounds to insignificant levels.

7.4.2.2 Offline Desorption The amount of solvent chosen for desorption will depend on several factors. If high sensitivity is required, the amount of solvent should be as small as possible but sufficient to immerse the coating completely. To facilitate the use of small volumes, various inserts can be placed within multi-well plates or into HPLC vials. Larger desorption solvent volumes, on the other hand, provide a more efficient desorption process, minimise or eliminate carryover and allow for multiple injections of the

228

Table 7.2 Selected SPME
LC Applications Using Manual Interface Analysis

Type of Fibre Coating

Extraction Time (min)

Matrix Modification

Desorption Conditions

Reference

Anatoxin-a in water and biological samples

PDMS/DVB 60 μm

pH 10, 20% NaCl, derivatisation 45 C

PDMS/DVB 60 μm

Mycophenolic acid in serum Amphetamine and methamphetamine in serum Paclitaxel in liposome formulations Fenitrothion and main metabolites in environmental water samples Phenylurea and propanil herbicides

CW/TPR 50 μm

30

pH 3

PDMS/DVB 60 μm

30 pre-equilibrium

pH 10 dilution 1:2

12 min static desorption using 60% acetonitrile
40% water 60 s static desorption in mobile phase 60 s static desorption in mobile phase 1 min static desorption

64

Ochratoxin A in wine

20 min extraction and 20 min on-fibre derivatisation 60

Carbowax/TPR 50 μm

60 pre-equilibrium

2 min static desorption

36

PDMS/DVB 60 μm

60 pre-equilibrium

Membrane-protected SPME 15% Na2SO4

5 min dynamic desorption

60

PDMS/DVB 60 μm

50 pre-equilibrium

10% w/v Na2SO4, 55 C

40

Carbowax/TPR 50 μm

30 pre-equilibrium

50 C

Static desorption for 4 min using 50% ACN
50% water, then 5 min dynamic desorption in mobile phase Static desorption for 4 min in mobile phase then 3 min dynamic desorption in mobile phase

50 57 58

35

Handbook of Solid Phase Microextraction

Diphenylether herbicides in water samples

pH 3

Solid-Phase Microextraction Method Development

229

same sample during the analysis. The agitation methods discussed in Section 7.2.3 can be applied to increase the rate of the desorption process. Direct desorption: Typically, manual offline desorption methods use the minimum amount of desorption solvent required to cover the entire length of the coating. Cantu et al.37 used 70 μL of mobile phase as desorption solvent and examined desorption times of 5
30 min. It turned out that 20 min was sufficient for the efficient desorption of amitriptyline, imipiramine, nortriptyline and desipramine. Carryover was also tested, and no drug peaks were found during blank assays. For anticonvulsant analysis, 50 μL of desorption solvent and 10 min of desorption time were sufficient for an effective desorption. Lopez-Monzon et al.69 used 50 μL of methanol in a 200-μL glass vial to desorb benzimidazole fungicides efficiently from a carboxen
PDMS 75-μm fibre within 10 min. The use of 200 μL of methanol with 10% acetic acid achieved desorption of clenbuterol from MIP SPME fibres in 5 min.77 Theodoridis et al. used 200 μL of methanol placed in HPLC inserts to desorb various drugs from PA and PDMS fibres. Offline desorption allowed them to perform multiple extractions from the same sample while performing desorption in the same portion of solvent, thus improving the sensitivity of the proposed method.78 Alternatively, several fibres can be exposed to the same sample solution and desorbed in one HPLC insert, resulting in higher extraction yields and enhanced sensitivity. Hu et al.46 used ultrasonic treatment to achieve complete desorption of analytes from a coated stir bar (with film thickness of 150 μm or 60 μm) using 100 μL of methanol as desorption solvent and a 3-min desorption time. Musteata et al.79 were able to use only 20 μL of acetonitrile:water:acetic acid desorption solvent in plastic inserts to desorb benzodiazepines successfully. As shown clearly by these examples, smaller desorption volumes result in maximum sensitivity but require longer desorption times and/or agitation. Using larger desorption volumes sacrifices some of the sensitivity of the method but significantly shortens the time necessary for complete desorption of the analytes. It is important to note that multiwells can be used to perform manual desorption. For maximum sensitivity, a multiwell plate with a very small well volume (20 μL well volume in Teflon multi-well plate) directly compatible with LC injection can be used to avoid dilution effects and to improve the sensitivity of the technique.71 Ideally, in this case, the entire well contents will be injected onto the LC column for separation and detection. Desorption 1 evaporation/reconstitution: During initial method development, it is recommended to test desorption solvents compatible with the mobile phase conditions used for the analysis (the direct desorption method). As shown in the previous examples, the most common desorption solvents are methanol:water and acetonitrile:water mixtures. If these solvents do not have sufficient desorption strength, a strong solvent, such as pure methanol, acetonitrile or dichloromethane, can be used and subsequently evaporated to dryness. In the final step, the sample can then be reconstituted in a solvent compatible for HPLC analysis. For example, successful offline desorption of benzodiazepines is achieved using 500 μL of 75% methanol
25% water contained in a 96-deep-well plate, for 30 min, with shaking. To obtain optimal sensitivity, the desorption solvent is then evaporated to dryness,

230

Handbook of Solid Phase Microextraction

and the wells are reconstituted using 25
75 μL fresh solvent.80 Es-Haghi et al.45 used a similar strategy, consisting of desorption in a large solvent volume, followed by nitrogen evaporation and reconstitution in a small solvent volume, but they used 100% methanol as the initial desorption solvent. The desorption time was reduced to 10 min, and the evaporation of solvent was faster. In another study, the desorption time was further reduced to 2 min by using 115-rpm agitation, and these conditions were adequate to desorb 99% of analytes from the fibres.44 Automated offline desorption: Concept 96 robotic station can perform both direct desorption and evaporation/reconstitution steps in parallel on up to 96 samples. In contrast to the manual desorption technique, a larger desorption solvent volume is needed in order to ensure that the entire coating is fully immersed (800
1,000 μL of solvent). The use of inserts to reduce desorption solvent volume is not feasible due to the dimensions of the extraction phase. However, even with the use of these relatively large desorption solvent volumes, the analytical sensitivity was found to be sufficient for most applications with typical limits of detection in sub ng/mL to low ng/mL range when LC
MS/MS was employed.74,75 The main reason for good sensitivity (without employing additional pre-concentration) is the fact that extraction geometry was modified to maximise extraction efficiency. The typical desorption times required to reach 98
99% desorption range from 15 to 30 min.

7.4.3

Fibre Preconditioning and Rinsing Procedures

7.4.3.1 Fibre Preconditioning The presence of organic solvents from preconditioning or fibre cleaning may affect the extraction efficiency of the fibre. This effect should be evaluated by comparing extraction results from a fibre that was not preconditioned with results for the extraction where the fibre was appropriately preconditioned. For some recently introduced fibres (e.g. Supelco biocompatible HPLC fibres), such as fibres based on C18 particles, a preconditioning step may be necessary to wet the C18 chains and obtain the best extraction efficiency. This is illustrated in Figure 7.8.81

% Extracted

Effect of conditioning procedure on the amount extracted 0.6 0.5 0.4 0.3 0.2 0.1 0.0 Carbamazepine

Diazepam Analyte

No conditioning

Water

Methanol/water (1/1, v/v)

Figure 7.8 Effect of preconditioning procedure on the extraction efficiency of C18 coatings (n 5 6 fibres) for the extraction of carbamazepine and diazepam. The conditioning procedure prior to the extraction was varied as follows: (i) no preconditioning performed, (ii) 30-min conditioning in purified water and (iii) 30-min conditioning in methanol/water (1/1, v/v).81

Solid-Phase Microextraction Method Development

231

The omission of preconditioning procedure significantly decreased the amounts extracted for both analytes tested, but the effect observed was more pronounced for carbamazepine. Similarly, Lord et al.47 found B2.5-fold increase in extraction efficiency of PPY fibres if the probes were preconditioned by rinsing with methanol for 15 s, followed by rinsing with water for 15 s.

7.4.3.2 Fibre Rinsing In some instances, rinsing the fibres prior to desorption improved reproducibility of the analysis by removing any analyte loosely attached to the surface of the fibre by surface tension.45,47 If such a rinsing step is used, it is important to choose a weak solvent and keep rinsing times short to ensure that the analyte is not accidentally desorbed from the fibre. The incorporation of a short rinsing step (10
30 s) is recommended for automated SPME methods for applications involving complex, heterogeneous matrices and for in vivo SPME methods in order to remove any particulate matter or cells from the surface of the fibre prior to the desorption. It is important to perform this step immediately after extraction for best results.

7.4.4

Method Development for In-Tube SPME

Although it was originally developed for the extraction of organic compounds from aqueous matrices, the use of in-tube SPME was successfully extended to more complex matrices, such as wine and biological fluids. Currently, in-tube SPME methods are used for environmental analysis, clinical and forensic analysis and food analysis, as described in several recent reviews.61,82 The main advantages of using in-tube SPME are convenient automation, simplicity, minimised handling of biological samples and better precision than what is achievable with manual techniques. The instrumental setup for in-tube SPME is described in Chapter 5. In this section, method development strategies for developing in-tube SPME methods will be discussed in detail. Figure 7.9 provides an overview of the typical parameters optimised during the development of an in-tube SPME method.

7.4.4.1 Selection of Capillary Coating The selection of the capillary coating to be used in in-tube SPME is an extremely important step in order to achieve the best extraction efficiency and, thus, good method sensitivity. Typically, a piece of a commercially available GC capillary column is used as the extraction capillary for in-tube SPME. One capillary column can be reused for more than 100 analyses, and Kataoka et al.83 report the successful use of the same capillary for more than 500 analyses over a 3-month period. Various commercial capillaries, such as Omegawax 250, Supel-Q PLOT, Supelcowax, SPB-1 and SPB-5, have successfully been used as in-tube extraction phases. For most polar compounds, Supel-Q PLOT capillaries (made of porous DVB polymer) showed the best performance among commercially available

232

Handbook of Solid Phase Microextraction

Draw/eject speed No. of draw/eject cycles

Sample pH

Sample volume used for draw/eject cycles

In-tube SPME method development

Addition of salt

Type of capillary coating

Coating thickness Capillary length

Figure 7.9 In-tube SPME method development scheme. Capillary column Nadolol

No coating (nonpolar fused silica)

Pindolol Acebutolol Timolol

SPB-1

Metoprolol Oxprenolol Labetalol

SPB-5

Propranolol Alprenolol

Omegawax 250 0.0

2.0

4.0

6.0

8.0

10.0

12.0

Area counts (×106)

Figure 7.10 Evaluation of four capillary columns for in-tube SPME/LC/MS analysis of β-blockers.53 (Source: Reprinted with permission from American Chemical Society (r 1999).)

GC columns for compounds such as cortisol.84 Omegawax (PEG liquid polymer) was optimal for carbamate pesticides85 and β-blockers,53 as shown in Figure 7.10. Various additional custom-made coatings have also been proposed to increase sensitivity of specific analytes. For example, PPY capillary offered superior extraction efficiency for some among 12 phenylurea and carbamate pesticides tested.62 Mullett et al.63 reported the successful use of restricted access material (RAM) alkyl-diol-silica for in-tube SPME
LC determination of benzodiazepines in human serum. Using this bifunctional material prevented protein adsorption and resultant

Solid-Phase Microextraction Method Development

233

fouling of the extraction phase. It also allowed analysis of serum samples without a prior deproteinisation step, thus further simplifying sample preparation procedure. A monolithic polymethacrylic acid
ethylene glycol dimethacrylate capillary provided high extraction efficiency towards basic analytes.86 To enhance selectivity further, an MIP was used as the extraction phase for in-tube SPME for the analysis of propranolol in biological fluids. The use of this material offered increased sensitivity, cleaner chromatograms and reusability for more than 500 injections.87 The amount of target analytes extracted by a given capillary is directly proportional to the length of the coating. Experimentally, a capillary length of 60 cm is found to work best because longer capillaries tend to result in peak broadening, while shorter capillaries exhibit reduced efficiency.

7.4.4.2 Number of Draw/Eject Cycles and Capillary Conditioning Procedure The online automation of in-tube SPME is accomplished by writing an appropriate HPLC autosampler programme to control extraction, desorption and injection steps. Typical steps of an in-tube SPME autosampler programme are shown in Table 7.3. Using larger sample volumes is recommended for in-tube SPME in order to ensure that no depletion of analytes occurs during repeated draw/eject cycles. The exact volume of the capillary should be determined and used as the volume to aspirate/ eject in each draw/eject cycle. This volume can be (i) estimated through calculation using the dimensions of the capillary or (ii) experimentally determined by replacing the analytical column with the capillary and measuring the retention volume of an injected acetone sample.63 The optimal preconditioning procedure will depend on the nature of the extraction phase. MIP coatings require preconditioning with a protic solvent such as acetonitrile.87 Optimising the conditioning procedure for a PPY capillary significantly improved extraction capacity by three- to fourfold and improved relative standard deviation (standard deviation).63 For in-tube SPME methods, the equilibrium extraction profile is obtained by varying the number of draw/eject cycles and plotting it against the amount extracted. Example of an extraction time profile for in-tube SPME is shown in Figure 7.11. Table 7.3 Typical Autosampler Programme for In-Tube SPME Autosampler Step Divert mobile phase flow away from the capillary (capillary bypass where capillary is under ambient pressure) Aspirate/dispense cycle(s) of one or more conditioning solvents Aspirate/dispense cycles of sample Wash injection needle tip Allow mobile phase through the capillary to desorb the analytes (main pass where capillary is under high pressure)

234

Handbook of Solid Phase Microextraction 1.20E + 07

AM MA MDMA Pindolol Propranolol Acebutolol

Area counts

1.00E + 07 8.00E + 06 6.00E + 06 4.00E + 06 2.00E + 06 0.00E + 00

0

5 10 15 Number of draw/eject cycle

20

Figure 7.11 Extraction time profile obtained for selected stimulants and β-blockers using in-tube SPME. AM, amphetamine; MA, methamphetamine; MDMA, 3,4-methylenedioxy methamphetamine.83 (Source: Reprinted with permission from RSC.)

A total of 15 draw/eject cycles were sufficient to reach equilibrium for stimulants (AM, MA and MDMA) when using the Omegawax capillary.83 However, for the remaining compounds, equilibrium was not achieved, as indicated by the continually increasing extracted amount as the number of draw/eject cycles increases. It is worth noting that equilibrium is often not reached when in-tube SPME is used83 because increasing the number of draw/eject cycles beyond 20
30 results in significant peak broadening and impractically long total analysis times,82 and the equilibrium for many analytes is not reached in fewer than 20 draw/eject cycles. However, because all the steps of in-tube SPME are performed by an autosampler, which affords precise timing control, good method precision is still achievable, despite the use of pre-equilibrium extraction times.

7.4.4.3 Draw/Eject Rate A draw/eject rate of about 100 μL/min was found to be optimal for in-tube SPME.88 The use of lower rates enhanced sensitivity by improving mass transfer, but it required very long extraction times. The use of higher rates tended to cause the formation of bubbles within the capillary, thus adversely affecting method precision and reducing extraction efficiency.

7.4.4.4 Desorption For most analyses using in-tube SPME, desorption of the analyte from the capillary is achieved by dynamic desorption
directing the flow of mobile phase through the capillary once the extraction process is completed. Carryover is not often observed because the capillary is washed continually with mobile phase during the chromatographic run and the coating is very thin (0.1
0.5 μm), enabling facile desorption in shorter times.83 By comparison, the thickness of the fibre SPME

Solid-Phase Microextraction Method Development

235

Table 7.4 Optimisation of Desorption Conditions of In-Tube SPME for Propranolol from an MIP In-Tube SPME Coating, Using Different Desorption Solvents87 Desorption Solvent

Slope of Regression Line

Correlation Coefficient n 5 6

Average % Standard Deviation

90/10 water/methanol 80/20 water/methanol 80/20/0.2 water/methanol/TFA

9,578 9,771 16,560

0.8345 0.8373 0.9973

18 17 4.7

Source: Reproduced with permission from American Chemical Society (r 2001).

coatings is up to 100 μm, so carryover is much more likely to be observed for fibre SPME
LC than for in-tube SPME. The efficacy of the elution solvent can be evaluated by analysing several standard spiked samples in triplicate and evaluating the standard deviation and linearity obtained. Poor precision and linearity indicate incomplete desorption, requiring further optimisation. Table 7.4 compares the ability of three different desorption solvents to elute propranolol from an MIP in-tube SPME coating.87 Water/methanol/ trifluoroacetic acid (TFA) mixture performed best for this application. Further confirmation that no carryover is present is obtained by performing blank analyses. Mullett et al.63 confirmed that a small amount of mixing of sample with mobile phase (about 2.9%) occurred during in-tube SPME. This caused an increase in background and pressure when analysing serum samples, but this side effect was completely eliminated by addition of a rapid capillary-wash step with a small amount of low-strength wash solvent prior to the elution of analytes. Raghani and Schultz89 found that this mixing caused systematic error by overestimating the amount of analyte extracted, but the error could be significantly minimised by inserting a 100-μL air plug prior to the extraction step. Kataoka et al.84 used 50 μL to prevent such mixing problems.

7.4.4.5 Matrix Modification Dilution is commonly required to use in-tube SPME for complex matrices. Table 7.5 gives an overview of various complex matrices and the conditions employed for their successful analysis using in-tube SPME. Other matrix modification procedures employed with in-tube SPME were described in Section 7.2.7. It is important to note that salt addition is used more rarely with in-tube SPME because high salt content of the sample may cause plugging of the extraction capillary.

7.5

Method Validation

After selection and optimisation of all relevant SPME parameters, the method should be validated in order to demonstrate its suitability for a particular application.

236

Handbook of Solid Phase Microextraction

Table 7.5 Typical Sample Modifications Used to Perform In-Tube SPME for the Analysis of Complex and Heterogeneous Samples Matrix Analyte

Matrix Modification

Reference

Serum

Benzodiazepines

63

Urine

Ketamines

Serum

Propranolols

Saliva

Cortisols

Urine

Stimulants and β-blockers

10-fold dilution with PBS pH 7.4, methanol (95:5 v/v) Centrifugation followed by 1:1 dilution with 10 mmol phosphate buffer pH 6 1:5 dilution using 1% acetic acid followed by ultrafiltration 1:5 or 1:2.5 dilution with acetate buffer pH 4, followed by protein precipitation and centrifugation 1:10 dilution with water followed by filtration and additional 1:5 dilution and pH adjustment to 8.5 Centrifugation

Juice Bisphenol A, and tea alkylphenols and phthalate esters

86 87 84

83

88

Validation experiments should be conducted to evaluate the selectivity and LDR of the method, the limits of detection and quantitation, and method precision and accuracy. Depending on the intended use of the proposed SPME method, the performance of SPME may be validated further by comparison of the results to those obtained from standard methods for that particular application, the analysis of certified reference materials (CRMs) and/or the participation in inter-laboratory studies. With the availability of commercial SPME robotic stations such as Concept 96 and commercial availability of new coatings suitable for bioanalysis, the use of SPME can be expected to expand rapidly in the future. SPME is capable of meeting the most stringent regulatory requirements such as Food and Drug Administration (FDA) guidelines for the validation of bioanalytical methods.90 Examples of validation results for several compounds using automated SPME
LC
MS/MS are shown in Table 7.6.74,75 Although SPME does not provide exhaustive extraction, it still can be employed in regulated pharmaceutical and clinical environments because the regulatory requirements state that absolute recovery must be constant across all concentration levels, but not necessarily 100%. For SPME, absolute recovery is calculated against the calibration curve obtained by direct liquid injection of standards using a given analytical method. Relative recovery, which is used for quantitation, is calculated using one of the calibration methods described in Chapter 6.

7.5.1

Method Sensitivity

There are several different definitions of the limit of detection (LOD) and limit of quantitation (LOQ). LOD can be defined as (i) three times the noise level, (ii) three times the standard deviation (SD) of blank analysis,32 (iii) 2.33 times the SD of the

Solid-Phase Microextraction Method Development

237

Table 7.6 Method Validation Results for Automated SPME
LC
MS/MS Methods Validation Parameter

Diazepam Oxazepam Nordiazepam Lorazepam Ochratoxin A

Matrix

Whole blood 4

Whole blood 4

Whole blood 4

Whole blood 4

0.7

4
1,000 94
103 2
8

4
500 91
98 8
20

4
1,000 98
106 5
6

4
500 97
106 7
11

0.7
50 91
109 2
12

3
6

7
12

2
4

7
12

4
14

Lower limit of quantification (LLOQ) (ng/mL) Linear range (ng/mL) Accuracy (%) Intra-batch precision (%) Inter-batch precision (%)

Urine

sample measurement value at the 95% confidence level (assuming that, close to detection limit, the SD of the blank and the sample measurement are similar) or (iv) three times the SD of the sample measurement value (seven replicate samples processed through the entire analytical method) at a concentration level not higher than 10 times the estimated detection limit.91,92 The last definition represents the EPA procedure for determining the method detection limit and calculating the LOD according to this definition is recommended.25 For bioanalytical method validation, the guidelines set out by the FDA should be followed.90 Excellent limits of detection are usually achievable by SPME methods, comparable to the sensitivity obtained by other sample preparation techniques. Benzodiazepines are widely used as tranquillisers, muscle relaxants and sleep inducers, and are often drugs that are abused. Therefore, the analysis of benzodiazepines from biological fluids is of clinical and forensic importance. Numerous SPME methods were developed for the analysis of this drug class and an overview of these methods for diazepam, a common benzodiazepine, is shown in Table 7.7. Although the various proposed methods differ in their selectivity, sensitivity and total analysis times as shown in the table, they all have excellent precision. The analytical sensitivity is mainly dependent on three factors: (i) analytical instrument employed for detection/separation, (ii) choice of SPME coating and (iii) choice of equilibrium versus pre-equilibrium SPME extraction times. Best sensitivity is achievable by the use of affinity-based coatings such as antibody-based coatings. The optimisation of various factors discussed in Sections 7.2
7.4 helps further improve the performance of the SPME method. In general, and as illustrated in Table 7.7, the best performance in terms of sensitivity is achieved using immunoaffinity coatings and the highly sensitive LC
MS/MS instruments. Under these conditions, the LOD for diazepam is 0.007 ng/mL. The coupling of in-tube SPME to capillary LC improved the limits of detection by factors of 24
81 for the analysis of carbamate pesticides in water samples.85 The sensitivity achievable for carbaryl using the proposed method was 20 parts

238

Table 7.7 Various SPME Methods Developed to Date for the Analysis of Diazepam Method Type

Biological Fluid

SPME Coating Extraction Time

Matrix Modification

LC
MS

Urine

PA fibre 85 μm

LC
MS

Serum and urine GC
MS Plasma

LC
UV

Serum

LC
MS/ MS LC
MS/ MS LC
MS/ MS

Whole blood and in vivo Whole blood and in vivo Spiked PBS standard

Precision Reference

1.0

13.0%

93

10 draw/ eject cycles 30 min

Saturated NaCl, 10 3 dilution, 5
400 ng/mL pH 9.1 pH 8.5 1
500 ng/mL

1.0

NR

94

10% NaCl, pH 6.9, 55 C

1
1,000 ng/mL

0.3

6
15%

95

4 draw/eject cycles 60 min

Centrifugation 1:10 dilution, pH 7.4, 5% MeOH Saturated salt, pH 7, 45 C

50
50,000 ng/mL

24

4.6%

63

CW/DVB fibre 65 μm RAM alkyl diol 30 min silica PPY 30 min C18/PEG

5 min

Immunoaffinity 30 min SPME

6%

56

None

0.2
5 μg per 3 mL 0.02 of sample 25
1,000 ng/mL 20

12%

96

None

1
1,000 ng/mL

7

NR

47

None

4
2,000 ng/mL

1.7

NR

45

None

0.007
0.2 ng/mL

0.007

10.4%

80

Handbook of Solid Phase Microextraction

GC
MS Urine and serum LC
MS Blood

In-tube SPME Supel-Q plot PDMS fibre 100 μm RAM diol silica

LOD (ng/mL)

180 min

Linear Dynamic Range

Solid-Phase Microextraction Method Development

239

per trillion. The main difficulty in coupling in-tube SPME to capillary LC was the limited injection volume (,100 nL) that can be handled by capillary LC systems, whereas in-tube SPME technique requires at least 30
45 μL of desorption solvent. On-column focusing overcame this issue successfully. If complex instrumentation (such as MS/MS detection or capillary LC systems) is not available, sensitivity can be improved by the use of the internally cooled fibre. Multiple extraction SPME can also be used to enhance extraction yields (and thus method sensitivity), although the effect of this approach on precision has not been thoroughly evaluated yet.81 A detailed method validation and comparison of fibre SPME versus in-tube SPME for the screening analysis of triazine herbicides in water samples was published.76 In this chapter, better limits of detection were achieved by in-tube SPME, but unfortunately, different LC systems were used for analysis (conventional versus capillary), so the SPME efficiency cannot be compared directly.

7.5.2

Method Precision

Good method precision can be achieved by SPME methods as long as the factors affecting the SPME extraction process are well controlled. In general, equilibrium SPME methods will have better method reproducibility than pre-equilibrium methods if manual extraction is employed. For automated methods, method precision tends to be similar whether equilibrium or pre-equilibrium SPME is employed because of accurate and reproducible timing and fibre positioning due to computer control. For SPME
GC applications, typical standard deviation values should be below 5% standard deviation. For SPME
LC applications, standard deviations below 10
15% can be expected for properly developed methods. For bioanalytical methods, it is recommended that the precision around the mean value should not exceed 15%,90 so SPME methods are well suited for these applications. LC methods using fluorescence and UV detection have better method precision than the methods employing mass spectrometric detection, due to the high degree of variability inherent in atmospheric ionisation processes. For example, Volmer and Hui report excellent validation results for the analysis of 11 corticosteroids from urine using SPME, with precision values between 5 and 11% standard deviation, which is excellent for LC
MS/MS analysis.42 Intra-day and inter-day reproducibility for in-tube SPME
LC is typically well below 10% standard deviation for most applications reported in the literature. In-tube SPME
LC methods and Concept 96 multi-fibre SPME methods have better method precision than manual SPME
LC methods, due to a higher degree of automation and better reproducibility of timing and/or fibre positioning (see also Table 7.6). For example, intra-day precision of 1.6% standard deviation was obtained for the determination of ketamine,86 of ,4.6% standard deviation for cortisol,84 of ,7% standard deviation for benzodiazepines,63 ,6.3% standard deviation for carbamate and phenylurea pesticides62 and ,9.8% for fenitrothion and its main metabolites.60 To improve precision, factors such as the immersion depth of the fibre, the position of the fibre within the vial or well, agitation speed and temperature should be kept constant. An in-depth

240

Handbook of Solid Phase Microextraction

discussion of factors affecting method precision in SPME
GC applications is provided in Section 7.3.3. SPME methods that include derivatisation tend to have higher standard deviations because the sample preparation procedure contains additional steps and factors such as reaction yields which become important and decrease reproducibility. For example, Herraez-Hernandez et al.65 obtained inter-day precision values of up to 24% standard deviation for their SPME on-fibre derivatisation method coupled to LC
fluorescence detection, which is undesirable for quantitative analysis. SPME
LC
fluorescence method for the enantiomeric analysis of amphetamines in urine, which also included a derivatisation step, showed somewhat improved performance, with intra- and inter-day reproducibilities of less than 20% standard deviation, and acceptable accuracy, with relative errors less than 20%.66

7.5.3

Method Accuracy

The accuracy obtained by SPME methods is also very good, provided that appropriate care is taken to control appropriate experimental parameters and the main principles of the technique are well understood. In the simplest experiment, accuracy can be evaluated by spiking a blank sample matrix with known amounts of analyte and then subjecting these spiked samples to the entire analytical procedure. The selection of an appropriate calibration method, depending on the sample complexity, is crucial for achieving adequate accuracy. In most applications, relative errors of 620% or less for spiked samples are deemed acceptable, and SPME is quite capable of meeting these requirements. For example, recoveries between 86 and 105% are obtained from ground water samples for the analysis of monolinuron, diuron, propanil, linuron and neburon.40 Similarly, relative recoveries for diphenylether herbicides were between 81 and 104%.35 Accuracies with relative errors of less than 66% and precision of less than 12% standard deviation were obtained for the method validation of chiral analysis of ibuprofen in urine.59 Kataoka et al.84 reported excellent intra- and inter-day reproducibility for the analysis of cortisol in saliva at all spiked sample concentrations (1.5
8.9% standard deviation) and accuracies of 95.5
97.2%. The study by Sagratini et al.55 demonstrates a crucial point. The extraction efficiency of the pesticides tested varied greatly depending on the type of juice. For the pesticide diuron, for instance, the amount extracted ranged from 40% from apple juice to 70% from water and strawberry/cherry juices. The authors found that the observed matrix effect could be reduced by a 1:1 dilution of the samples. Typical accuracy and precision values for the proposed method in the orange juice matrix are shown in Table 7.8. Therefore, for accurate quantitative results, it is important that the calibration standards be prepared in the matrix that closely resembles that of the samples. If the proposed method will be used for various sample matrices, then the performance of the method should be evaluated in all these matrices. Matrix suppression is a common problem encountered in LC
MS/MS methods that rely on electrospray ionisation. As such, it is necessary to evaluate matrix

Solid-Phase Microextraction Method Development

241

Table 7.8 Method Validation Data for the Analysis of Carbamate and Phenylurea Pesticide Residues in Orange Juice Spiked at Two Fortification Levels55 Compound

Carbofuran

Amount Intra-daya Added (mg/kg) Amount Accuracy Detected (%) (mg/kg)

0.5 0.2 Monuron 0.5 0.2 Primicarb 0.5 0.2 Monolinuron 0.5 0.2 Diuron 0.5 0.2 Diethofencarb 0.5 0.2 Benfuracarb 0.5 0.2 Carbosulfan 0.5 0.2 a

0.48 0.21 0.49 0.19 0.52 0.2 0.50 0.21 0.51 0.2 0.49 0.19 0.52 0.19 0.51 0.2

96 105 98 95 104 100 100 105 102 100 98 95 104 95 102 100

Inter-daya Standard Deviation (%)

Amount Detected (mg/kg)

Accuracy (%)

Standard Deviation (%)

9 7 4 3 6 10 5 14 12 7 8 2 15 5 4 11

0.5 0.19 0.49 0.21 0.51 0.21 0.48 0.19 0.5 0.21 0.48 0.20 0.53 0.21 0.50 0.19

100 95 98 105 102 105 96 95 100 105 96 100 106 105 102 95

3 10 7 12 2 5 14 9 6 4 8 15 10 11 5 4

The values are the means of three replicate determinations.

Source: Reproduced with permission from Elsevier.

effects during validation of an LC
MS/MS method to ensure that the analytical data will be accurate and reproducible. The degree of matrix effects encountered in a method will depend on various factors, but the choice of sample preparation method will be one of the major contributing factors. Among traditional sample preparation techniques, protein precipitation typically provides the least amount of sample clean-up, which in turn can result in significant matrix effects. SPE and LLE methods provide much better clean-up, which results in less potential for matrix interferences, provided that the analyte is sufficiently resolved from the solvent front. From a theoretical viewpoint, SPME can be expected to provide even cleaner sample extracts because of the limited amount of extraction phase employed. Table 7.9 shows sample results for the evaluation of absolute matrix effects for the determination of benzodiazepines in whole blood.74,97 The absolute matrix effects are evaluated by the comparison of the signal obtained for a blank whole blood sample extract spiked post-extraction with a known analyte concentration versus the signal for the liquid standard prepared directly in the desorption solvent at the same concentration. If no significant matrix effect is observed, the signal should be 85
115%. In other words, the signal for the two samples should be identical within the experimental margin of error. As shown in Table 7.9, if

242

Handbook of Solid Phase Microextraction

Table 7.9 Evaluation of Absolute Matrix Effects for the Determination of Benzodiazepines in Whole Blood by SPME: Comparison of SPME Coupled to Direct Injection into MS Versus SPME
LC
MS/MS Analyte

Direct Infusion-MS/MS of 1 µg/mL Benzodiazepine Standard % Matrix Suppression

LC
MS/MS (20 µL Injection of 20 ng/mL Benzodiazepine Standard) % Matrix Suppression

Lorazepam Oxazepam Nordiazepam Diazepam Diazepam d5 (IS)

97 79 46 49 45

95 101 100 102 102

SPME extract is analysed by direct infusion into MS (in other words, if LC separation is omitted), most of the analytes except lorazepam exhibit significant matrix suppression. However, when LC
MS/MS analysis is employed, no matrix suppression or enhancement is observed. This indicates the importance of incorporation of the LC separation step in the analytical method when using general sorbent SPME coatings, as well as the excellent sample clean-up capability that SPME can provide.

7.5.3.1 Inter-Laboratory Studies and Comparison to Other Preparation Techniques In a study reported by Gorecki et al.,98 several laboratories, some using SPME on a regular basis and some using this technique for the first time, took part in a roundrobin test on the determination of semi-volatile pesticides in water. No significant differences in performance were observed between the two groups, and the interlaboratory repeatability and reproducibility SD values were less than 5.5% for all target analytes. Another SPME
GC
MS method was successfully applied in a food study for the determination of methylmercury in tuna fish samples, and the laboratory participated in the CCQM-P39 inter-laboratory test organised by the Institute for Reference Materials and Measurements (IRMM).99 Extraction techniques for the isolation of methylmercury from the sample matrix were optional. Results were in excellent agreement with the average results of other participants. Another inter-laboratory comparison study reported the results of benzene and toluene biomonitoring in human blood samples.33 The reliable determination of benzene in blood at very low levels was rather difficult; however, this comparison demonstrated that the SPME method performance characteristics and results were in good agreement with existing extraction methods (i.e. DHS and P&T). Subsequently, Blount et al.100 developed an automated SPME
GC
MS for quantitation of 31 volatile organic compounds (VOCs) in whole blood to support large

Solid-Phase Microextraction Method Development

243

Table 7.10 Comparison of Automated SPME
GC
MS Method to the P&T-High Resolution MS Method for the Analysis of VOCs in Human Whole Blood Based on Results Reported in Ref. 100 SPME
GC
MS

P&T-High Resolution MS

4 3 sample throughput Average standard deviation 5 9% $200

1 3 sample throughput Average standard deviation 5 22% $400

epidemiological studies. The comparison of this new method versus the P&T-high resolution MS method initially used for this analysis is shown in Table 7.10. SPME significantly increased throughput, reduced the cost of analysis and improved method precision, thus making it ideally suited for the analysis of large number of samples. The authors also evaluated long-term reproducibility of SPME performance over 3 years, using appropriate low- and high-concentration quality samples. The results of this study are shown in Table 7.11 and illustrate excellent long-term reproducibility of SPME. Similar in-depth quantitative results were reported for the analysis of volatile n-alkanes in blood by SPME
GC
MS.101 However, in this study, long-term reproducibility was slightly worse (standard deviation range from 12% to 47%) because of the analyte losses during handling (despite many precautions to minimise these losses as much as possible) and the ease of sample contamination. Excellent results were also obtained by Guillot et al.102 and Schuhmacher et al.103 The former compared SPME to the official European Union (EU) LLE method for the determination of organic micropollutants migrating from polyethylene to drinking water. In this study, SPME proved to be the faster, more-sensitive method, with a much larger extraction range compared to LLE. If approved, it could be used as an alternative to the existing official method or as an alert system in the routine analysis of materials used to transport drinking water. A total of 28 laboratories participated in the inter-laboratory comparison discussed in the latter study, evaluating the determination of methyl tert-butyl ether (MTBE) in drinking water.103 The optimised SPME method favourably compared with the SHS and P&T extraction techniques or direct aqueous injection (DAI). Cantu et al.37 compared the performance of SPME to the protein precipitation method for the analysis of anticonvulsants in plasma, and obtained regression coefficients of 0.993 or higher when the results of one method were plotted against the other. In the same publication, SPME performance was also compared against LLE for the analysis of four antidepressants in plasma, and correlation coefficients of 0.987 or higher were obtained for all compounds tested. In another study, the performance of multi-well SPME (semi-automated using lab-built components) was compared to multi-well LLE.104 The main findings of the study are summarised in Table 7.12 and show the potential of SPME as a very useful bioanalytical tool. Rellan et al.64 compared the performance of SPME to SPE for the analysis of anatoxin-a in biological samples, using LC with fluorescence detection. The

244

Handbook of Solid Phase Microextraction

Table 7.11 Long-Term Reproducibility of Automated SPME Analysis for the Determination of VOCs in Human Blood Over a Two-Year Period (n . 120)100 Analyte

Low QCa mean 6 SDb (µg/L)

Standard Deviationc (%)

High QC mean 6 SD (µg/L)

Standard Deviationc (%)

1,1,1-Trichloroethane 1,1,2,2Tetrachloroethane 1,1,2-Trichloroethane 1,1-Dichloroethane 1,1-Dichloroethene 1,2-Dichlorobenzene 1,2-Dichloroethane 1,2-Dichloropropane 1,3-Dichlorobenzene 1,4-Dichlorobenzene 2,5-Dimethylfuran Benzene Bromodichloromethane Bromoform Carbon tetrachloride Chlorobenzene Chloroform cis-1,2-Dichloroethene Dibromochloromethane Dibromomethane Ethylbenzene Hexachloroethane m/p-Xylene Methylene chloride o-Xylene Styrene Methyl tert-butyl ether Tetrachloroethene Toluene trans-1,2Dichloroethene Trichloroethene

0.086 6 0.011 0.028 6 0.002

13 6

0.436 6 0.055 0.135 6 0.005

13 4

0.042 6 0.002 0.021 6 0.002 0.026 6 0.003 0.050 6 0.006 0.052 6 0.005 0.037 6 0.003 0.053 6 0.014 0.264 6 0.029 0.064 6 0.006 0.064 6 0.010 0.065 6 0.004 0.121 6 0.005 0.017 6 0.003 0.028 6 0.002 0.058 6 0.013 0.045 6 0.004 0.027 6 0.001 0.102 6 0.008 0.212 6 0.017 0.034 6 0.002 2.24 6 0.147 0.156 6 0.053 0.242 6 0.016 0.200 6 0.023 0.115 6 0.008 0.201 6 0.027 0.329 6 0.041 0.039 6 0.005

5 11 12 13 9 8 26 11 9 16 6 4 19 7 23 9 5 8 8 7 7 34 7 11 7 13 12 12

0.207 6 0.007 0.105 6 0.009 0.138 6 0.025 0.216 6 0.012 0.253 6 0.014 0.181 6 0.011 0.204 6 0.017 1.13 6 0.054 0.318 6 0.027 0.293 6 0.024 0.320 6 0.013 0.577 6 0.016 0.087 6 0.015 0.134 6 0.006 0.230 6 0.016 0.221 6 0.019 0.130 6 0.004 0.497 6 0.025 0.484 6 0.025 0.166 6 0.010 2.78 6 0.153 0.483 6 0.057 0.381 6 0.030 0.779 6 0.035 0.568 6 0.027 0.809 6 0.105 0.591 6 0.048 0.198 6 0.025

3 9 18 6 6 6 8 5 8 8 4 3 17 4 7 9 3 5 5 6 6 12 8 5 5 13 8 13

0.038 6 0.004

11

0.161 6 0.017

11

a

Quality control serum. Standard deviation. Relative standard deviation. Source: Reprinted with permission from Elsevier.

b c

methods exhibited comparable limits of detection, but SPME required smaller sample volumes. More importantly, the extracts obtained by SPME from complex matrices (e.g. fish, carp or mussel tissue) were much cleaner than the extracts obtained by the SPE sample preparation procedure. In terms of reproducibility,

Solid-Phase Microextraction Method Development

245

Table 7.12 Comparison of Method Validation Results for SPME and LLE in a HighThroughput Multi-well Plate Format Based on the Results Reported in Ref. 104 Analytical Parameter

Multi-well SPME

Multi-well LLE

Linear range in plasma (ng/mL) Accuracy (%) Precision (% CV)

1
500 95.2
103.7 0.5
6.9

1
500 98.0
101.8 0.8
3.3

SPE performed better ( , 6% for SPE and , 14% for SPME), but it is important to note that this particular reproducibility experiment was performed on samples of river water. Considering the interferences observed in the SPE extract chromatograms for biological samples, the standard deviation of the SPME technique would probably be equivalent or better than that of SPE. In another study, Lopez-Monzon compared the accuracies obtained by SPME versus SPE for the analysis of fungicides in seawater, and the performance of SPME was superior.69 Harper et al.105 evaluated the performance of SPME
GC for the trace determination of ethylene oxide in sterilised medical devices in a three-laboratory roundrobin study. SPME was found to be a reliable quantitative method and to improve LOQ by 2
3 orders of magnitude over the corresponding direct injection method. Inter-laboratory standard deviation ranged from 21% at the LOQ level to 5%, indicating excellent performance of SPME.

7.5.3.2 Analysis of CRMs The high accuracy achievable by SPME methods has been confirmed by successful analysis of CRMs. For example, the performance of an SPME
GC method for the analysis of lead in blood and urine was confirmed by the analysis of CRMs.106 The values obtained for the headspace SPME method from urine were 19 6 2 μg/dL and from blood 29 6 3 μg/dL, while the certified values for these CRMs (obtained using three different standard techniques) were 19 6 2 μg/dL for urine and 25 6 4 μg/dL for blood.

7.6

Concluding Remarks

This chapter focused on SPME method development. SPME offers important benefits in terms of speed and solventless use, resulting in low-cost analytical methods free of health hazards. An added advantage is the relatively small sample size needed to achieve satisfactory sensitivity, compared to other ‘classical’ extraction techniques. The SPME method has proven to provide enough sensitivity to fulfill regulatory requirements and to give accurate and precise results in inter-laboratory tests. The achieved sensitivity, accuracy and reproducibility are typically either comparable or complementary with other extraction techniques, depending on the specific application.

246

Handbook of Solid Phase Microextraction

In addition, the use of SPME for unique applications, such as in vivo sampling and direct determination of free analyte concentrations, is gaining in popularity and presents many new and exciting research opportunities, not achievable with traditional sample preparation methods such as LLE.

References 1. A Carrasco-Pancorbo, L Cerretani, A Bendini & A Segura-Carretero, J Sep Sci 28 (2005) 837 2. R Rial-Otero, B Cancho-Grande & J Simal Gandara, J Chromatogr A 992 (2003) 121 3. FE Ahmed, Trends Anal Chem 22 (2003) 170 4. FE Ahmed, Trends Anal Chem 20 (2001) 649 5. E Psillakis & N Kalogerakis, Trends Anal Chem 22 (2003) 565 6. A Go´mez-Hens & MP Aguilar-Caballos, Trends Anal Chem 2 (2003) 847 7. PL Buldini, L Ricci & L Sharma, J Chromatogr A 975 (2002) 47 8. J Poustka, K Holadova & J Hajslova, Eur Food Res Technol 216 (2003) 68 9. G Zambonin, Anal Bioanal Chem 375 (2003) 73 10. L Pallaroni & C von Holst, Anal Bioanal Chem 376 (2003) 908 11. M Tomaniova´, J Hajˇslova´, J Pavelka, V Kocourek, K Holadova´ & I Klı´mova´, J Chromatogr A 827 (1998) 21 12. P Suchan, J Pulkrabova, J Hajslova & V Kocourek, Anal Chim Acta 520 (2004) 193 13. JS Bonvehı´ & FV Coll, J Sci Food Agric 83 (2003) 275 14. A Rizzolo, A Brambilla, S Valsecchi & P Eccher-Zerbini, Food Chem 77 (2002) 257 15. L Wang & CL Weller, Trends Food Sci Technol 17 (2006) 300 16. C Krach, G Sontag & S Solar, Food Res Int 32 (1999) 43 17. I Ferreira, E Mendes, P Brito & MA Ferreira, Food Res Int 33 (2000) 113 18. FM Amaral & MSB Caro, Food Chem 93 (2005) 507 19. A Schieber, P Hilt, P Streker, H Endresz, C Rentschler & R Carle, Innovative Food Sci Emerg Technol 4 (2003) 99 20. European Commission document, Integrated Pollution Prevention & Control Reference Document on Best Available Techniques in the Food, Drink & Milk Industries (August 2006) 21. HL Lord, J Chromatogr A 1152 (2007) 2 22. H Kataoka, HL Lord & J Pawliszyn, J Chromatogr A 880 (2000) 35 23. C Whang & J Pawliszyn, Anal Commun 35 (1998) 353 24. Z Liu & J Pawliszyn, Analyst 131 (2006) 522 25. J Pawliszyn, Solid Phase Microextraction, Theory & Practice (1997) Wiley-WCH, Inc.: New York, NY 26. SAS Wercinski, Solid Phase Microextraction, A Practical Guide (1999) Marcel Dekker, Inc.: New York, NY 27. M Alpendurada, J Chromatogr A 889 (2000) 3 28. W Wardencki, M Michulec & J Curyło, Int J Food Sci Technol 39 (2004) 703 29. Supelco, Sigma-Aldrich Co, Solid-Phase Microextraction (SPME) Metal Fibre Assemblies, Bellefonte, PA, Product Data Sheet (2005) 30. ML Perkins, BR D’Arcy, AT Lisle & HC Deeth, J Sci Food Agric 85 (2005) 2421 31. Supelco, Solid Phase Microextraction: Theory & Optimisation of Conditions, Bellefonte, PA, Bulletin 923 (1998)

Solid-Phase Microextraction Method Development

247

32. C Basheer & HK Lee, J Chromatogr A 1047 (2004) 194 33. E Schimming, K Levsen, C Ko¨hme & W Schu¨rmann, J Anal Chem 363 (1999) 88 34. L Setkova, S Risticevic, CM Linton, G Ouyang, LM Bragg & J Pawliszyn, Anal Chim Acta 581 (2007) 221 35. H Sheu, Y Sung, MB Melwanki & S Huang, J Sep Sci 29 (2006) 2647 36. F Musteata & J Pawliszyn, J Pharm Pharm Sci 9 (2006) 231 37. MD Cantu, DR Toso, CA Lacerda, FM Lancas, E Carrilho & M Queiroz, Anal Bioanal Chem 386 (2006) 256 38. L Cardenes, A Martin-Calero, JH Ayala, V Gonzalez & AM Afonso, Anal Lett 39 (2006) 405 39. HY Tsai, CM Chang, JL Shen, LJ Chen, TF Yang & CB Fuh, J Chromatogr Sci 44 (2006) 354 40. AR Mughari, PP Vazquez & MM Galera, Anal Chim Acta 593 (2007) 157 41. C Fernandes, JDS Neto Alvaro, J Rodrigues, C Alves & FM Lancas, J Chromatogr B 847 (2007) 217 42. DA Volmer & JPM Hui, Mass Spectrom 11 (1997) 1926 43. J Wu & J Pawliszyn, Anal Chim Acta 520 (2004) 257 44. X Zhang, A Es-Haghi, FM Musteata, G Ouyang & J Pawliszyn, Anal Chem 79 (2007) 4507 45. A Es-haghi, X Zhang, FM Musteata, H Bagheri & J Pawliszyn, Analyst 132 (2007) 672 46. Y Hu, Y Zheng, F Zhu & G Li, J Chromatogr A 1148 (2007) 16 47. HL Lord, RP Grant, M Walles, B Incledon, B Fahie & JB Pawliszyn, Anal Chem 75 (2003) 5103 48. FM Musteata, M Walles & J Pawliszyn, Anal Chim Acta 537 (2005) 231 49. E Carasek & J Pawliszyn, J Agric Food Chem 54 (2006) 8688 50. A Aresta, R Vatinno, F Palmisano & CG Zambonin, J Chromatogr A 1115 (2006) 196 51. HL Lord & J Pawliszyn, Anal Chem 69 (1997) 3899 52. A Aresta, F Palmisano & CG Zambonin, J Pharm Biomed Anal 39 (2005) 643 53. H Kataoka, S Narimatsu, HL Lord & J Pawliszyn, Anal Chem 71 (1999) 4237 54. J Wu, HL Lord, J Pawliszyn & H Kataoka, J Microcolumn Sep 12 (2000) 255 55. G Sagratini, J Manes, D Giardina, P Damiani & Y Pico, J Chromatogr A 1147 (2007) 135 56. Y Luo, L Pan & J Pawliszyn, J Microcolumn Sep 10 (1998) 193 57. CG Zambonin, A Aresta & F Palmisano, J Chromatogr B 806 (2004) 89 58. C Chou & M Lee, Anal Chim Acta 538 (2005) 49 59. ARM de Oliveira, EJ Cesarino & PS Bonato, J Chromatogr B 818 (2005) 285 60. A Sanchez-Ortega, MC Sampedro, N Unceta, MA Goicolea & RJ Barrio, J Chromatogr A 1094 (2005) 70 61. T Kumazawa, X Lee, K Sato & O Suzuki, Anal Chim Acta 492 (2003) 49 62. J Wu, C Tragas, H Lord & J Pawliszyn, J Chromatogr A 976 (2002) 357 63. WM Mullett, K Levsen, D Lubda & J Pawliszyn, J Chromatogr A 963 (2002) 325 64. S Rella´n, J Osswald, V Vasconcelos & A Gago-Martinez, J Chromatogr A 1156 (2007) 134 65. R Herraez-Hernandez, C Chafer-Pericas, J Verdu-Andres & P Campins-Falco, J Chromatogr A 1104 (2006) 40 66. C Chafer-Pericas, P Campins-Falco & R Herraez-Hernandez, J Pharm Biomed Anal 40 (2006) 1209 67. FM Musteata & J Pawliszyn, J Biochem Biophys Methods 70 (2007) 181 68. FM Musteata & J Pawliszyn, Trends Anal Chem 26 (2007) 36

248

Handbook of Solid Phase Microextraction

69. A Lopez Monzon, D Vega Moreno, ME Torres Padron, Z Sosa Ferrera & JJS Rodriguez, Anal Bioanal Chem 387 (2007) 1957 70. G Ouyang, Y Chen, L Setkova & J Pawliszyn, J Chromatogr A 1097 (2005) 9 71. J O’Reilly, Q Wang, L Setkova, JP Hutchinson, Y Chen, HL Lord, CN Linton & J Pawliszyn, J Sep Sci 28 (2005) 2010 72. G Vas & K Vekey, J Mass Spectrom 39 (2004) 233 73. F Pragst, Anal Bioanal Chem 388 (2007) 1393 74. D Vuckovic, E Cudjoe, D Hein & J Pawliszyn, Anal Chem 80 (2008) 6870 75. R Vatinno, D Vuckovic, CG Zambonin & J Pawliszyn, J Chromatogr A 1201 (2008) 215 76. C Chafer-Pericas, R Herraez-Hernandez & P Campins-Falco, J Chromatogr A 1125 (2006) 159 77. EHM Koster, C Crescenzi, W den Hoedt, K Ensing & GJ de Jong, Anal Chem 73 (2001) 3140 78. G Theodoridis, M Aikaterini Lontou, F Michopoulos, M Sucha & T Gondova, Anal Chim Acta 516 (2004) 197 79. FM Musteata, ML Musteata & J Pawliszyn, Clin Chem 52 (2006) 708 80. HL Lord, M Rajabi, S Safari & J Pawliszyn, J Pharm Biomed Anal 44 (2007) 506 81. D Vuckovic, R Shirey, Y Chen, L Sidisky, C Aurand, K Stenerson & J Pawliszyn, Anal Chim Acta 638 (2009) 175 82. H Kataoka, Anal Bioanal Chem 373 (2002) 31 83. H Kataoka, HL Lord, S Yamamoto, S Narimatsu & J Pawliszyn, J Microcolumn Sep 12 (2000) 493 84. H Kataoka, E Matsuura & K Mitani, J Pharm Biomed Anal 44 (2007) 160 85. Y Gou & J Pawliszyn, Anal Chem 72 (2000) 2774 86. Y Fan, Y Feng, S Da & X Gao, Analyst 129 (2004) 1065 87. WM Mullett, P Martin & J Pawliszyn, Anal Chem 73 (2001) 2383 88. H Kataoka, M Ise & S Narimatsu, J Sep Sci 25 (2002) 77 89. AR Raghani & KN Schultz, J Chromatogr A 995 (2003) 1 90. Food & Drug Administration, Bioanalytical Method Validation, Guidance for Industry (2001) 91. LJJ Catalan, V Liang & CQ Jia, J Chromatogr A 1136 (2006) 89 92. Analytical Methods Committee, Analyst 112 (1987) 199 93. K Jinno, M Taniguchi & M Hayashida, J Pharm Biomed Anal 17 (1998) 1081 94. H Yuan, Z Mester, H Lord & J Pawliszyn, J Anal Toxicol 24 (2000) 718 95. MH de Oliveira, MEC Queiroz, D Carvalho, SM Silva & FM Lancas, Chromatographia 62 (2005) 215 96. M Walles, WM Mullett & J Pawliszyn, J Chromatogr A 1025 (2004) 85 97. D Vuckovic, E Cudjoe, FM Musteata, J Pawliszyn, Nat. Protoc 5 (2010) 140
161 98. T Gorecki, R Mindrup & J Pawliszyn, Analyst 121 (1996) 1381 99. G Centineo, EB Gonzalez, JI Garcia Alonso & A Sanz-Medel, J Mass Spectrom 41 (2006) 77 100. BC Blount, RJ Kobelski, DO McElprang, DLA Ashley, JC Morrow, DM Chambers & FL Cardinali, J Chromatogr B 832 (2006) 292 101. DM Chambers, BC Blount, DO McElprang, MG Waterhouse & JC Morrow, Anal Chem 80 (2008) 4666 102. S Guillot, MT Kelly, H Fenet & M Larroque, J Chromatogr A 1101 (2006) 46 103. R Schuhmacher, M Fuhrer, W Kandler, C Stadlmann & R Krska, Anal Bioanal Chem 377 (2003) 1140

Solid-Phase Microextraction Method Development

249

104. W Xie, J Pawliszyn, WM Mullett & BK Matuszewski, J Pharm Biomed Anal 45 (2007) 599 105. T Harper, L Cushinotto, N Blaszko, J Arinaga, F Davis & C Cummins, Biomed Chromatogr 22 (2008) 136 106. X Yu, H Yuan, T Gorecki & J Pawliszyn, Anal Chem 71 (1999) 2998

8 SPME and Environmental Analysis Gangfeng Ouyang School of Chemistry and Chemical Engineering, Sun Yat-sen University, Guangzhou, P. R. China

8.1

Introduction

The SPME technique has been widely used for the analysis of environmental pollutants in air, water, soil and sediment samples, using on-site or off-site analytical approaches.1,2 Hundreds of papers addressing environmental analysis by solid phase microextraction (SPME) were published in recent years. With the development of SPME theory, several calibration methods have been proposed for the quantification of SPME and several SPME devices have been developed for on-site fast field sampling or the long-term monitoring of environmental pollutants.3,4 A number of fibre coatings have been developed in the laboratory to extend the applications of SPME for environmental and other analyses (see also Section 3.4). For example, polypyrrole (PPY) coating was developed to extract polar analytes5 from samples, and this coating was used for both fibre SPME5 and in-tube SPME.6 Several coatings based on the sol
gel technology have been developed for the determination of environmental pollutants in samples, including polyethylene glycol (PEG),7 Carbowaxs 20M-modified silica,8 polydimethylsiloxane polyvinyl alcohol (PDMS/ phenyltrimethoxysilane (PTMOS) and methyltrimethoxysilane PVA),9,10 (MTMOS),11 Low-temperature glassy carbon (LTGC)12,13, calix[4]arene,14
17 and a variety of crown ethers.18
22 The SPME derivatisation technique has been widely used for the treatment of polar compounds and to enhance the recovery, selectivity and sensitivity of the SPME method. This approach makes it is possible to identify substances with poor chromatographic behaviour, high reactivity or thermal instability (see also Section 7.2.8).23 In-tube SPME is an effective sample preparation technique that is based on a fused-silica capillary column for the extraction device (see Section 5.2.1). Target analytes in aqueous matrices are directly extracted and concentrated by the coating in the capillary column through repeated drawing and ejection of the sample solution, and can be transferred directly to a liquid chromatograph (LC) or gas chromatograph (GC) for analysis. The procedures of in-tube SPME include extraction, concentration, desorption and injection, and can be easily automated using a common autosampler.24,25 This technique has been successfully applied to the determination of various environmental pollutants.24
33 A new SPME approach, thin-film microextraction (TFME) or membrane SPME, was recently developed to achieve higher extraction efficiency and sensitivity (see Handbook of Solid Phase Microextraction. DOI: 10.1016/B978-0-12-416017-0.00008-5 © 2012 Elsevier Inc. All rights reserved.

252

Handbook of Solid Phase Microextraction

Section 3.2.4).34 This approach has been successfully applied to field water sampling.35
37 In the calibration chapter (Chapter 6) and SPME devices chapter (Chapter 3) of this book, the available SPME calibration methods and some devices for environmental analysis have been introduced. In this chapter, environmental analysis with different configurations of SPME will be summarised, and some of the environmental applications of SPME in gaseous, aqueous and solid matrix are presented.

8.2 8.2.1

Fibre SPME Fibre Coatings

Several coatings are commercially available for SPME analyses, including PDMS, polyacrylate (PA), divinylbenzene (DVB), carboxen (CAR) and Carbowax (CW), and fibres are available in different thicknesses for single coatings, mixtures or co-polymers (see Chapter 4 for detailed discussion). These fibres are suitable for applications of SPME for non-polar organic compounds, such as benzene, toluene, ethylbenzene and xylenes (BTEX), polynuclear aromatic hydrocarbons (PAHs), pesticides and so on and polar organic compounds, such as phenols, alcohols, ketones, nitroaromatics and so on. Supelco (Bellefonte, PA) has developed metal fibre assemblies where the fibre core is made of metal alloy to extend the lifetime of the fibre assembly and improve the durability and reproducibility of the extraction. It is important to use the appropriate coating for a given application. The commercially available stationary phases of SPME fibre are limited and restrict the wide application of SPME.38 PPY coating was developed by Wu et al. to extract polar or ionic analytes.5 This coating addressed the needs of the SPME/LC technique because the extraction phase of SPME fibre should not only have a high extraction ability for polar analytes, but it also must be stable in various solutions.39 Sol
gel technology provides efficient incorporation of organic components in inorganic polymeric structures in solution.19,40,41 These fibres are stable in both strong organic solvents and acidic solutions. The hydrolytic stability of these sol
gel SPME fibres towards organic solvents and high and low pH solutions can be attributed to the fact that the coating is chemically bonded to the surface of the fused-silica substrate. Thermal stability shows that sol
gel PDMS fibre could be used up to 320 C, whereas commercial PDMS fibres began to bleed at lower temperatures (200 C). The high degree of porosity of sol
gel fibre resulted in higher sensitivity and faster extraction times relative to commercial fibres.42 Several coatings have been developed by sol
gel technology for the determination of environmental pollutants. Table 8.1 presents some applications of sol
gel-coated fibres for the environmental analysis.

8.2.2

Derivatisation

One of the challenges for SPME is the sampling and analysis of polar organic compounds. They are difficult to extract from environmental samples and difficult to separate on the chromatographic column. Derivatisation approaches can be used to

SPME and Environmental Analysis

253

address this challenge (see also Section 7.2.8). Good chromatographic performance and detection can be facilitated by in-coating derivatisation following extraction. In addition, selective derivatisation to analogues containing high detector response groups will result in an enhancement of the sensitivity and selectivity of detection. Figure 8.1 summarises various derivatisation techniques of SPME.43 For the direct derivatisation technique, the derivatising agent is added to the sample vial and then the derivatives are extracted by the SPME fibre coating and introduced into the analytical instrument. Derivatisation in the GC injector is an analogous approach, but it is performed at high injection port temperatures. The most interesting and potentially useful technique is the simultaneous derivatisation and extraction method, also referred to as on-fibre derivatisation. This approach is highly efficient and can be used in remote field applications. For on-fibre derivatisation, the derivatising reagent is loaded on the fibre and the fibre is subsequently exposed Table 8.1 Applications of Sol
Gel-Coated Fibres for Environmental Analysis Coating

Analytes

Reference

PEG Carbowax PDMS/PVA PTMOS/MTMOS

BTEX, phenols, diesters and pesticides BTEX Pesticides, PCBs Benzene, toluene, ethylbenzene, 2-octanone, benzaldehyde, acetophenone, dimethylphenol and tridecane BTEX, monohalogenated benzene Organochlorine pesticides, chlorophenols, phenolic compounds Phenols, organochlorine pesticides, aromatic amines, BTEX, chlorobenzenes, arylamines

7 8 9,10 11

LTGC Calix[4]arene Crown ether

Derivatisation SPME

Direct derivatisation in sample matrix

Derivatisation in GC injector port Derivatisation in SPME fibre coating

Simultaneous derivatisation and extraction

Figure 8.1 SPME derivatisation approaches.

Derivatisation following extraction

12,13 14
17 18
22

254

Handbook of Solid Phase Microextraction

to the sample. The analytes are then extracted and simultaneously converted to analogues that possess a high affinity for the coating. Derivatisation SPME was first introduced for the analysis of C1
C10 carboxylic acid in water at trace levels.43 Currently, this approach is widely used for the analysis of organic pollutants in the environment. A PA fibre has been used for the determination of acidic herbicides by derivatisation with diazomethane.44 Acidic herbicides also have been identified with the PA fibre and N-methyl-N-tert-butyldimethylsilyltrifluoroacetamine derivatisation.45 o-2,3,4,5,6-Pentafluorobenzylhydroxylamine hydrochloride (PFBHA) was used as a derivatising agent for the determination of formaldehyde in air46 and C1
C10 aldehydes in particleboard.47

8.3

In-Tube SPME

The two fundamental approaches to in-tube SPME are the static and dynamic approaches. Static in-tube SPME normally is achieved with a fibre-retracted SPME device for time-weighted-average (TWA) sampling. It has been discussed in the calibration chapter (Chapter 6) and SPME devices chapter (Chapter 3) of this book. Currently, most of the in-tube SPME techniques refer to dynamic in-tube SPME, which involves an open tubular fused-silica capillary column as the extraction device. Organic compounds are directly extracted and concentrated into the stationary phase of the capillary column by repetitive sample introduction into the extraction capillary with a programmed autosampler until equilibrium or a suitable extraction level has been reached. The analytes are then directly transferred to an LC column. The selectivity and efficiency of the extraction are determined by the type of stationary phase and the internal diameter, the length and the film thickness of the capillary column. Several commercially available capillary columns are generally used for in-tube SPME. A PPY coating has also been used for the in-tube SPME technique (Figure 8.2).6 PPY-coated capillary column was easily prepared by an oxidative polymerisation method. The experimental results indicated that the in-tube SPME technique with a PPY coating has a higher extraction efficiency than commercial capillary coatings for PAHs and polar aromatics.48 This method was successfully used for the analysis of pesticides26, PAHs and aromatic amines30 in water samples. Several types of coatings for the in-tube SPME column were prepared with the sol
gel technique and used for the identification of organic pollutants in the environment. PAHs, ketones and aldehydes were efficiently extracted and preconcentrated from aqueous samples using sol
gel zirconia-polydimethyldiphenylsiloxane (PDMDPS)-coated capillaries and the in-tube SPME technique.49 Amphiphilic and hydrophilic oligomers were synthesised and coated on fused silica capillaries using the sol
gel technique, and these types of capillaries were used for the analysis of organochlorine pesticides, triazine herbicides and phenols.50 The sol
gel titaniaPDMS coated capillaries were used for the online extraction and analysis of PAHs, ketones and alkyl-benzenes.51

SPME and Environmental Analysis

255

Inner surface

Capillary wall

(A)

(B)

(C)

(D)

Figure 8.2 Scanning electron micrographs of the PPY-coated capillary [(A) cross-sectional view, (C) enlarged inner surface view] and the host silica capillary [(B) cross-sectional view, (D) enlarged inner surface view].6 (Source: Figure reprinted with permission.)

8.4

TFME

There are some limitations in terms of efficiency for SPME fibre, due to the small volume of the extraction phase. For this reason, TFME technology was developed based on the SPME technique. Bruheim et al. proposed a membrane SPME approach in which a thin sheet of PDMS membrane was used as an extraction phase and was directly analysed by GC.34 The main advantage of this technique is that high sensitivity can be achieved without sacrificing the overall analysis time, due to the large surface-area-to-extraction-phase volume ratio. This approach was used for the determination of PAHs in spiked lake water samples, and detection limits in the low ppt level were achieved. This new SPME approach is currently being developed for a field water sampling device, for both grab sampling and long-term monitoring.35
37 More details about this technology can be found in Chapters 3 and 6 of this book.

8.5 8.5.1

Applications of SPME in Various Environmental Sample Matrices Air Samples

Currently, a number of commercially available approaches can be used for the determination of trace contaminants in air. These include grab sampling with stainless steel canisters or nylon bags followed by concentration over a sorbent bed, direct concentration over sorbent using portable pumps, and passive diffusion monitors and others. These techniques are expensive and time consuming to operate. They require the use of adsorbents such as charcoal, silica or other polymers,

256

Handbook of Solid Phase Microextraction

whose breakthrough volumes are strongly affected by humidity. The adsorption of contaminants is followed by a desorption step, principally using solvents and the introduction to an analytical instrument for analysis.52 SPME has the potential to dramatically improve cost efficiency of air analysis because it can integrate the first steps of the analytical process: sampling, extraction and concentration, and convenient introduction to the analytical instrument. Model studies of air extraction have been performed by using two methods, the static and the more realistic dynamic approaches. In the static method, the target compound is introduced to the sealed bulb. In practical ambient air measurements, the system is not static, and convection is present. Therefore, it is more appropriate to use dynamic flowing gas chambers for the modeling studies. In a 20-L chamber operated at 200 mL/min, the equilibration times are between 20 and 100 s for compounds ranging from benzene to 1,3,5-trimethyl benzene, respectively. They are close to theoretical values for perfect agitation conditions. Increased airflow will decrease the equilibration times only for less-volatile analytes. Air is a simple matrix for sampling by SPME. A fibre coating can be selected not to concentrate major air components such as nitrogen, oxygen, carbon dioxide and moisture, but to selectively extract dissolved organics. PDMS has properties close to these characteristics. Only at humidities approaching 100% does it absorb sufficient amounts of moisture to cause a change in the coating polarity resulting in a small change in the distribution constant. This effect can be conveniently compensated for, if necessary, by measuring the relative humidity of the sample and adjusting the response appropriately. The experimental parameter primarily affecting response is temperature. However, the effect of the temperature change on the distribution constant can be conveniently predicted because log Kfg is linearly related to 1/T and the heat of vaporisation of the pure solute ΔH v :     ΔH V RT ΔH V ð8:1Þ 1 log10 log10 Kfg 5 2:303RT 2:303RT  γ i p where p is the analyte vapor pressure at a known temperature T for a pure solute, and γ i is the activity coefficient of the solute in the coating. In other words, Kfg can be calculated for given extraction conditions by measuring the temperature of the sample and knowing the heat of vaporisation for the target compound. In addition, the heat of vaporisation and the activity coefficient are related to the retention time of the compound on the GC column using the same coating material.53 Therefore, retention times of the eluting compound should directly provide the appropriate distribution constants. A more universal approach to describe the relationship is to use the Kovats retention indices because they are available in the literature. However, calculation of Kovats indices requires isothermal chromatographic operation. This limits the range of compounds that can be studied in one run. The more appropriate approach is to use the linear temperature programmed retention index (LTPRI) to estimate the retention indices of a wide range of analytes varying in volatility.54 This scale

SPME and Environmental Analysis

257

allows the assignment of Kfg values because a well-defined relationship exists between the index and Kfg. The linear relationship between the PDMS/gas distribution constant and LTPRI obtained on the PDMS column for a range of linear hydrocarbons is given by55 log10 Kfg 5 0:00415 3 LTPRI 2 0:188

ð8:2Þ

This relationship can be applied to calculate the PDMS/gas distribution constant of any analyte so long as the retention index for a given compound on a PDMS column is available from the literature or by experiment. Table 8.2 summarises the distribution constant data obtained using this approach and compares it to the direct SPME experimental values for a range of hydrocarbons. The observed differences are small, indicating that Eq. (8.2) can be used to determine PDMS/gas distribution constants. This approach can be extended to other coatings so long as appropriate columns are available. In addition, the compounds do not need to be identified to have appropriate distribution constants assigned to them. This approach allows quantitation without identification and can be applied, for example, to determining the total petroleum hydrocarbon (TPH) content of air samples. SPME has superior sensitivity for short-term monitoring compared to traditional devices, which are limited by the gas throughput. The sensitivity of the SPME technique can be further improved by using thicker, more selective coatings or by cooling the fibre to increase the PDMS/gas distribution constant. The other approach is to incorporate the derivatisation reagent in the coating to allow ‘trapping’ of the analyte in the coating. All these modifications will result in SPME sensitivity enhancement for air monitoring. The SPME devices for long-term air sampling have been developed. Typically, the fibre reaches equilibrium with the air components in the first few minutes and then does not accumulate any more analytes, independent of the exposure time. However, SPME assembly also can work as an integrated sampling device. The simplest way to accomplish this task is to retract the fibre into the needle. In this position, the fibre is surrounded by the needle, resulting in a very slow extraction because the analytes need to first diffuse through the needle opening before they can reach the fibre. The depth of the fibre retraction into the needle defines the diffusion length and the integrating factor (see Section 3.3.3). Table 8.3 summarises a field experiment designed to monitor styrene in industrial air using both a 5-min grab sample and 30 min of integrated measurement. Two standard techniques are compared to SPME with a PDMS fibre exposed for grab sampling and retracted into the needle for integrated sampling. The agreement between the methods is very good, considering the preliminary nature of the experiments.56 Frequently, the fibres cannot be immediately analysed after the extraction, particularly in field analysis, because a portable instrument might not be available. In such a situation, the appropriate storage method for the fibres needs to be designed. This can be simply accomplished by retracting the fibre into the needle and sealing its

258

Handbook of Solid Phase Microextraction

Table 8.2 Summary of Kfg Data for Hydrocarbons at 25 C Compound

Kfg For LTPRI

3-Methylpentane 2,4-Dimethylpentane 2,2,3-Trimethylbutane 2-Methylhexane 2,3-Dimethylpentane 2,2-Dimethylhexane 2,5-Dimethylhexane 2,2,3-Trimethylpentane 2,3-Dimethylhexane 2-Methylheptane 4-Methylheptane 3-Methylheptane 3-Ethylhexane 2,5-Dimethylheptane 3,5-Dimethylheptane (D) 3,3-Dimethylheptane 3,5-Dimethylheptane (L) 2,3-Dimethylheptane 3,4-Dimethylheptane (D) 3,4-Dimethylheptane (L) 2-Methyloctane 3-Methyloctane 3,3-Diethylpentane 2,2-Dimethyloctane 3,3-Dimethyloctane 2,3-Dimethyloctane 2-Methylnonane 3-Ethyloctane 3-Methylnonane Benzene Toluene Ethylbenzene m-Xylene p-Xylene o-Xylene Isopropylbenzene n-Propylbenzene 1-Methyl-3-ethylbenzene 1-Methyl-4-ethylbenzene 1,3,5-Trimethylbenzene 1-Methyl-2-ethylbenzene Isobutylbenzene

157 246 258 351 358 616 672 672 877 933 947 1,010 1,020 1,880 1,880 1,900 1,900 2,270 2,310 2,330 2,490 2,670 2,690 4,240 4,990 6,020 6,690 6,890 7,150 296 815 2,020 2,190 2,220 2,710 3,780 4,960 5,340 5,440 6,150 6,260 8,450

For SPMEa 159 262 280 387 412 673 587 569 968 993 1,060 1,090 990 1,970 1,960 2,090 2,100 2,390 2,420 2,620 2,600 2,890 2,610 4,320 5,050 6,100 6,690 6,970 7,100 301 818 2,070 2,090 2,500 2,900 3,880 5,040 4,750 6,230 6,480 6,580 8,360 (Continued)

SPME and Environmental Analysis

259

Table 8.2 (Continued) Compound

sec-Butylbenzene 1-Methyl-3-isopropylbenzene 1-Methyl-4-isopropylbenzene 1-Methyl-2-isopropylbenzene 1-Methyl-3-n-propylbenzene 1,3-Dimethyl-5-ethylbenzene 1-Methyl-2-n-propylbenzene 1,4-Dimethyl-2-ethylbenzene 1,2-Dimethyl-4-ethylbenzene 1,3-Dimethyl-2-ethylbenzene 1,2-Dimethyl-3-ethylbenzene 1,2,4,5-Tetramethylbenzene 2-Methylbutylbenzene tert-1-Butyl-2-methylbenzene n-Pentylbenzene t-1-Butyl-3,5-dimethylbenzene t-1-Butyl-4-ethylbenzene 1,3,5-Triethylbenzene 1,2,4-Triethylbenzene n-Hexylbenzene Pentane Hexane Heptane Octane Nonane Decane Undecane Dodecane Tridecane Tetradecane a

Kfg For LTPRI

For SPMEa

8,680 9,660 9,920 11,200 12,800 13,700 14,700 16,200 17,400 18,300 20,900 23,400 24,000 27,100 35,300 43,900 44,900 66,800 77,600 95,700 77 201 521 1,356 3,525 9,166 23,834 61,973 161,139 418,986

8,590 10,100 10,200 12,000 13,200 15,000 14,900 15,900 17,400 18,100 20,000 24,700 24,100 26,200 34,500 45,600 43,700 67,300 75,600 90,100          

Asterisks indicate compounds used as a calibration standard.

open end with a piece of a septum. Cooling the needle provides additional protection against analyte losses. Table 8.4 summarises the initial data about the suitability of this simple approach for several solvents varying in volatility. One-hour storage with a fibre placed in solid carbon dioxide preserves all analytes well. However, the loss of volatiles is observed for longer times. The loss of analyte can be traced to sorption by the septum, which is typically made of PDMS polymer. For better field performance, the SPME syringe needs to be modified to provide better sealing of the fibre for example. Also, the current design will benefit from substantial structural changes to facilitate more convenient handling by field sampling staff.

260

Handbook of Solid Phase Microextraction

Table 8.3 Comparison of SPME with Other Samplers Sample Type

Industrial Concentrations of Styrene (µg/L)

Grab (5 min) Integrated (30 min)

SPME with PDMSa

Charcoal Tubeb

Passive Badgec

130 56

97 54

90 72

a

Observed concentration of styrene with SPME 100-μm PDMS at 296 K and 25% relative humidity for both sampling types. Active sampling. c Integrated value over the sampling time. b

Table 8.4 Effects of Temperature on Sample Storagea Compound

25 C Uncapped

Capped

5 C

270 C

Capped

Capped

2 min 60 min 30 min 60 min 30 min 60 min 60 min 24 h 48 h Chloroform 1,1,1Trichloroethane Carbon tetrachloride Benzene Toluene Tetrachloroethylene 1,1,2,2Tetrachloroethane

97 95

51 49

70 67

62 59

93 93

89 91

95 92

50 76

30 45

97 94 96 96 98

54 69 77 79 70

73 89 86 85 90

70 81 89 88 80

95 94 95 95 94

93 92 95 94 92

95 97 94 99 98

79 83 92 97 95

66 82 85 90 93

a

Values in the table correspond to relative amounts of analyte (in %) remaining on the fibre. Source: Table reprinted with permission.57

Most SPME air sampling is performed on-site or in the laboratory by collecting the air sample in a bag or solid phase extraction (SPE) device; the analytes are extracted by the SPME fibre either by direct exposure or in the headspace (HS). Most applications involve the use of a commercial SPME fibre, but an Al2O3coated fibre was also used for volatile organic contaminant (VOC) sampling. Table 8.5 presents some recent applications of SPME for the environmental analysis of airborne contaminants.

8.5.1.1 Standard Gas-Generating System A standard gas-generating system is very useful for the development and validation of the SPME air sampling method and device. Figures 8.3 and 8.4 are schematic diagrams of the standard gas-generating system and sampling chamber for SPME

SPME and Environmental Analysis

261

Table 8.5 Applications of SPME for Environmental Analysis in Gas Samples Analytes

Extraction Method

Fibre/Capillary

Detection

Reference

Organic pollutants Toluene

Static indoor air sampling Static indoor, outdoor air sampling Static indoor air sampling

PDMS

GC/MS

58

CAR, fibre-retracted device

GC/MS, FID

59

PDMS/DVB

60

61 62

BTEX, hexane

BTEX BTEXs VOCs VOCs VOCs

VOCs VOCs, formaldehyde Formaldehyde Pesticides Odorants

Static air sampling Static outdoor air sampling Static indoor air sampling Static indoor air sampling Static indoor, outdoor air sampling Static indoor air sampling Static indoor air sampling

CAR/PDMS CAR/PDMS

GC/PID, FID, DELCD GC/MS GC/FID

PDMS/CAR

GC/MS

63

PDMS, CAR/PDMS PDMS/DVB, CAR/PDMS, fibreretracted device

GC/FID

64

GC/MS

65

γ-Al2O3-coated fibre

GC/FID

66

PDMS, PDMS/DVB

67

Static indoor air sampling Dynamic indoor air sampling Static air sampling (landfills) Static air sampling

PDMS/DVB, derivatisation PDMS

GC/PID, FID, DELCD GC/FID GC/MS

68,69

GC/MS

70

GC/PFPD

71

GC/NPD

72
74

GC/PFPD

75

GC/PID, FID, DELCD HPLC/MS

67

DVB/CAR/PDMS

CAR/PDMS Malodorous sulphur compounds Organophosphate Dynamic indoor PDMS trimesters air sampling VSCs Static air sampling PDMS/CAR, fibreretracted device Static indoor air PDMS, fibre-retracted Dodecane sampling device Isocyanates

Static air sampling PDMS/DBV, fibreretracted device, derivatisation

46

76

(Continued)

262

Handbook of Solid Phase Microextraction

Table 8.5 (Continued) Analytes n-Valeraldehyde

Extraction Method

Fibre/Capillary

Static air sampling PDMS/DVB, fibreretracted device, derivatisation Chlorobenzenes SPE followed by CW/PDMS, PDMS/DVB HS-SPME PCBs SPE followed by PDMS, PDMS/DVB HS-SPME Sarin Dynamic air PDMS/DVB sampling C5-C15 n-alkanes Static air sampling PDMS/DVB Hydrocarbons, Static air sampling PDMS, PDMS/DVB, fibre-retracted device formaldehyde BTEX Controlled air CAR/PDMS, PDMS/DVB flow rate VOCs Controlled air PDMS/DVB flow rate PDMS Organophosphate Controlled air trimesters flow rate VOCs Controlled air PDMS/DVB flow rate CAR/PDMS, PDMS/DVB BTEX Controlled air flow rate

Detection

Reference

GC/FID

77

GC/ECD, GC/MS GC/ECD, GC/MS GC/MS

78

80

GC/FID GC/FID

65 81

GC/FID

82

GC/MS, FID GC/NPD

83

GC/FID

84

GC/MS, FID

85

79

73,74

extractions introduced by Koziel et al.86 The system consists of three separate and continuous standard gas generators and air sampling chambers, capable of providing wide ranges of concentrations for VOCs and semi-VOCs in the air, at controlled air temperatures and relative humidities. Standard gas generation for BTEX, chlorinated VOCs and formaldehyde was achieved using permeation tubes. For all n-alkanes, a direct injection of the analytes into an air stream was used. Additional components of this system are an ozone generator, a graduated flow-through chamber and a mixing flow-through chamber with heating and cooling. Heating was achieved with a heating tape and temperature control device. The cooling was achieved with 10 mm inside diameter (i.d.) plastic tubing wrapped around the bulb supplied with chilled water. The range of temperatures for standard gases was from 5 C to 40 C. Sampling chambers were constructed and installed downstream from the standard gas generators. A schematic of an individual sampling chamber is provided in Figure 8.4. These sampling chambers facilitate a steady-state mass flow of all VOCs and semi-VOCs, at constant temperatures. Each sampling chamber consists of a custom-made 1.5-L glass bulb with several sampling ports plugged with halfhole-type Thermogreen LB-1 septa. Omega 120 W heating tape was wrapped

SPME and Environmental Analysis

Figure 8.3 Schematic diagram of standard gas-generating system.

Figure 8.4 Schematic diagram of the sampling chamber for SPME extractions.

263

264

Handbook of Solid Phase Microextraction

around the glass bulb to control temperature inside the bulb. Standard gas flow rates ranged from 50 to 3000 mL/min, resulting in mean air velocities inside the glass bulb of less than 10 mm/s. The standard gas system was proved to be useful in generating standard gas concentrations for a wide variety of experiments, involving SPME and other, conventional techniques used to validate SPME methods. The system was flexible, allowing to combine flows from three standard gases and to condition the standard gas. The major advantage of this system was the continuous availability of the standard gas which, besides in experiments, was used to (i) check the performance of the standard gas system, and (ii) daily quality assurance/quality control runs on several gas chromatographs. The standard gas-generating systems presented permit the study of the physical-chemical properties of various SPME systems and the investigation of various concepts for air sampling. The flexibility of the system enabled researchers to study a number of complex variables encountered in air sampling, such as rapid modifications in gas temperature, relative humidity, ozone concentrations, and, most importantly, a wide range of airborne organics over extended concentration ranges.

8.5.1.2 Distribution Coefficients Between Fibre Coating and Air (Kfa) Tables 8.6 and 8.7 present some log Kfa data for different organic compounds. Some Kfa data also can be found in Table 8.2.

8.5.2

Water Samples

Aqueous samples can be analysed in the laboratory with the direct-immersing model, the HS model or the in-tube model. On-site sampling with SPME devices is more convenient because the method eliminates the requirement of sample transportation. Some SPME water-sampling devices and the calibration method have Table 8.6 Distribution Coefficients (Kfa) Between 100-μm PDMS Coating and Air at 298 K Compound

log Kfa87

Compound

log Kfa88

Isoprene α-Pinene Myrcene 3-Carene Limonene γ-Terpinene Terpinolene Acetone 2-Butanone 2-Pentanone 3-Hexanone 2-Hexanone

1.76 3.66 3.93 3.94 4.04 4.14 4.25 2.06 2.4 2.99 3.25 3.27

2-Heptanone 3-Octanone Naphthalene Biphenyl Acenaphtylene Acenaphthene Fluorene Phenanthrene Anthracene Fluoranthene Pyrene

3.60 3.91 5.5 6.4 7.2 6.8 7.0 7.8 8.2 8.4 8.2

SPME and Environmental Analysis

265

Table 8.7 Log Kfa Values Between CAR/PDMS Coating and Air at 298 K89 Compound

log Kfa

Compound

log Kfa

Acetone 2-Butanone 2-Pentanone 2-Heptanone Heptanal Acroleine Methylacroleine 3-Buten-2-one 3-Penten-2one Benzaldehyde

2.54 2.94 3.37 4.33 4.42 2.72 2.91 3.09 4.03 5.09

Ethyl acetate Butyl acetate Methyl butyrate Pentyl butyrate Hexyl butyrate Ethanol 1-Propanol 1-Butanol 3-Methyl-1-butanol

2.90 3.78 3.38 4.37 4.87 2.57 2.82 3.42 3.54

Source: Table reproduced with permission.

been discussed in previous chapters, so the discussion in this section will be limited to topics not already covered. The fibre coating/water distribution constants can be calculated from the following equation: Kfw 5 Kfg Kgw

ð8:3Þ

where Kfg can be calculated from chromatographic data, as discussed above, by using LTPRI. Kgw is the gas/water distribution constant for a given analyte and can be found in the Henry’s constant tables. For example, the equation for a PDMS coating and aqueous matrix can be calculated from the equation below (see Section 2.3.2 and Table 2.1): logKfw 5 0:00415 3 LTPRI 2 0:188 1 log10 Kgw

ð8:4Þ

Therefore, by finding the relationship between Kfg and LTPRI for a given coating, the appropriate distribution constant can be calculated conveniently from chromatographic data and literature values of Henry’s constants. In addition, the Henry’s constants are similar for compounds that are closely related, resulting in a single linear relationship between Kfw and LTPRI that is characteristic for a group of different types of analytes. For example, as expected from their high Henry’s constant values, the Kfw values for paraffins are larger than those of aromatic analytes. Because of this linear relationship, quantitation with minimum identification is possible so long as a detector can assign extracted analytes selectively to appropriate groups of compounds.90 Table 8.8 summarises the distribution constants obtained for a number of aromatic and aliphatic hydrocarbons. Several reports indicated the existence of a linear relationship between the coating/water distribution constant and the octanol
water distribution constant,

266

Handbook of Solid Phase Microextraction

Table 8.8 Summary of Kfw Data for Hydrocarbons Obtained by SPME at 22 C Compound

Kfw

Cyclopentane Methylcyclopentane Cyclohexane cis-1,3-Dimethylcyclopentane trans-1,2-Dimethylcyclopentane Methylcyclohexane cis-trans-cis-1,2,4-Trimethylcyclopentane cis-trans-cis-1,2,3-Trimethylcyclopentane 1-Ethyl-1-methylcyclopentane trans-1,2-Dimethylcyclohexane cis-cis-cis-1,2,3-Trimethylcyclopentane cis-1,2-Dimethylcyclohexane 2,3-Dimethylbutane 2-Methylpentane 3-Methylpentane 2,2-Dimethylpentane 2,4-Dimethylpentane 2,2,3-Trimethylbutane 3,3-Dimethylpentane 2-Methylhexane 2,3-Dimethylpentane 3-Methylhexane 3-Ethylpentane 2,2-Dimethylhexane 2,5-Dimethylhexane 2,2,3-Trimethylpentane 2,4-Dimethylhexane 2,3-Dimethylhexane 2-Methylheptane 4-Methylheptane 3-Methylheptane 3-Ethylhexane 2,5-Dimethylheptane 3,5-Dimethylheptane (D) 3,3-Dimethylheptane 3,5-Dimethylheptane (L) 2,3-Dimethylheptane 3,4-Dimethylheptane (D) 3,4-Dimethylheptane (L) 2-Methyloctane 3-Methyloctane 3,3-Diethylpentane 2,2-Dimethyloctane 3,3-Dimethyloctane

712 1,356 1,592 4,289 3,372 4,657 5,621 6,556 6,831 6,638 7,109 7,826 2,359 3,224 3,270 7,349 8,989 9,802 10,963 10,202 13,074 11,146 10,816 24,504 23,519 21,205 41,133 33,749 25,806 27,274 31,856 28,370 84,142 78,829 76,013 72,856 68,675 93,292 89,182 45,267 66,682 63,718 72,155 82,430 (Continued)

SPME and Environmental Analysis

267

Table 8.8 (Continued) Compound

Kfw

Benzene Toluene Ethylbenzene m-Xylene p-Xylene o-Xylene Isopropylbenzene n-Propylbenzene 1-Methyl-3-ethylbenzene 1-Methyl-4-ethylbenzene 1,3,5-Trimethylbenzene 1-Methyl-2-ethylbenzene 1,2,4-Trimethylbenzene tert-Butylbenzene Isobutylbenzene sec-Butylbenzene 1-Methyl-3-isopropylbenzene 1-Methyl-2-isopropylbenzene 1-Methyl-3-n-propylbenzene 1-Methyl-4-n-propylbenzene n-Butylbenzene 1,4-Dimethyl-2-ethylbenzene 1,2-Dimethyl-4-ethylbenzene 1,3-Dimethyl-2-ethylbenzene 2-Methylbutylbenzene tert-1-Butyl-3,5-methylbenzene n-Pentylbenzene tert-1-Butyl-3,5-dimethylbenzene 1,3,5-Triethylbenzene 1,2,4-Triethylbenzene

58 189 566 533 564 485 1,412 1,664 1,231 1,581 1,451 1,321 2,183 2,185 4,197 4,011 3,284 3,003 3,772 3,870 3,872 3,628 3,984 4,345 9,099 6,059 8,195 18,260 18,517 16,253

Kow.91,92 Considering the discussion above, this relationship is expected to exist only for a group of related compounds, such as isoparaffins, cyclohexanes or substituted benzenes. Because the activity coefficients, and therefore the selectivity corresponding to unrelated groups of analytes in octanol, are expected to be different compared to PDMS or other fibre coating, the relationship between the two extracting phases will vary with the change of the chemical properties of analytes. The above discussion pertains to a pure water sample as a matrix. The presence of other components in water will modify the distribution constants for given analytes. For example, the addition of salt would generally result in an increase of the distribution constant for neutral organics, but the change is expected to be

268

Handbook of Solid Phase Microextraction

noticeable only if the concentration of salt exceeds 1%. Also, the presence of water-miscible polar organic solvents would result in changing properties of the matrix by reducing its polarity. In addition, swelling of the polymer with the solvent might occur for a polar coating, resulting in a change of Kfs. However, the change is not expected to be substantial when the concentration of the solvent is below 1%.93 When samples contain more salt and dissolved organics but are well defined so that a pure matrix can be prepared, external calibration may still be appropriate. Otherwise, a standard addition of isotopically labeled analytes should be used to compensate for variations in matrix composition. A major analytical challenge is always associated with the analysis of samples containing solids, such as sludge. Several approaches can be implemented with SPME. Sometimes modification of the extraction conditions, such as temperature, pH, salt and other additives, facilitates the displacement of analytes into the aqueous phase or HS, resulting in similar distribution constants as those obtained for pure water. Table 8.9 illustrates the quantitation of spiked sludge using salt addition and low pH conditions. Salt, in this case, appears to work well as a displacing reagent for the majority of analytes. In many cases, direct extraction is not possible because of a very dirty matrix or extreme pH conditions, which may damage the fibre. In such situations, the HS mode is suitable for many applications, as described briefly in the balance of this chapter. Even semivolatiles can be analysed Table 8.9 Analyte Recovery from Sewage Sludge94 Compound

Recovery (%)a

Recovery (%) Acid 1 Salt

Phenol 2-Chlorophenol o-Cresol m-Cresol p-Cresol 2,4-Dimethylphenol 2,4-Dichlorophenol 2,6-Dichlorophenol 4-Chloro-3-methylphenol 2,3,5-Trichlorophenol 2,4,6-Trichlorophenol 2,4,5-Trichlorophenol 2,3,4-Trichlorophenol 2,4-Dinitrophenol 4-Nitrophenol Tetrachlorophenol isomers 2-Methyl-4,6-dinitrophenol Pentachlorophenol

74.2 128 118 95.8 95.8 104 105 21.3 91.5 56.1 21.5 64.5 66.8 2.7 38.4 16.7 0 8

92 b 92 b 95.9 97.8 97.8 96.5 78.8 83.4 85 71.3 66.1 66.1 71.9 111 118 61.4 83.1 32.2

a

100% recovery represents the amount recovered from laboratory water sample spiked at the same level as the sewage sample. b Coeluted on GC column. Source: Table reprinted with permission.

SPME and Environmental Analysis

269

by this method so long as the extraction temperature is sufficiently high and good agitation conditions are provided. Application of SPME with membrane protection can assist in the extraction of non-volatile species in the presence of high molecular weight interferences, which are able to passivate the coating, such as proteins or humic material. Table 8.10 presents some recent applications of SPME for the analysis of aqueous environmental samples.

8.5.2.1 Standard Aqueous Solution-Generating System A flow-through standard water system was introduced by Shurmer and Pawliszyn.149 The system was constructed using an LC pump to deliver water to a stirred SPME sampling cell. The SPME sampling cell consists of a modified 40-mL vial with 1/4-in. glass inlet and outlet tubes added (Figure 8.5). A standard mixture of naphthalene, anthracene, fluoranthene, pyrene and benzo [a]pyrene in acetone was injected into the water, upstream of the SPME sampling cell, using a syringe pump and 1-mL Hamilton gastight syringe. To ensure adequate mixing of the water and PAHs, a 15-mL in-line glass impinger filled with 4-mm glass beads was inserted immediately after the PAH injection tee (Figure 8.5). The flow-through standard water generator eliminates errors associated with both losses of analytes to the system surfaces and limited sample volume. The system was successfully used to determine the distribution constants of PAHs between the SPME fibre coating and water. A good agreement between Kfw and Kow was observed. Figure 8.6 depicts another flow-through system for standard aqueous solutions. It is a modified system based on the one described earlier. A high-performance LC (HPLC) pump and a syringe pump were used to deliver water and standards in organic solvent, respectively. The results of the 20-month experiment showed that the change in concentration of naphthalene, acenaphthene and fluorene generated in the flow-through system was about 15%. A new flow-through standard water system was developed recently.150 Figure 8.7 presents the schematic diagram of the new flow-through standard water system based on permeation. It consists of a permeation chamber, a mixing chamber (also used as sampling chamber) and a sampling cylinder and chamber (the sampling cylinder is used for determining the effect of variable linear velocity of water). Water was filled in an approximately 12-L glass reservoir and delivered by a digital pump. The key parts of the generator are the DispoDialyzers in the permeation chamber. Each DispoDialyzer was partially filled with pure standards (naphthalene, acenaphthene, fluorene, anthrancene, fluoranthene and pyrene) and pure water (Figure 8.8). The DispoDialyzers were then tightly sealed and placed into a 400-mL beaker filled with 300 mL of water and sonicated for about 30 min. The prepared DispoDialyzers were deployed in the permeation chamber with an inlet

270

Table 8.10 Applications of SPME for Environmental Analysis in Aqueous Samples Extraction Method

Fibre/Capillary

Detection

Reference

BTEX BTEX BTEX, benzenes BTEX and ethers BTEX, naphthalene, chlorinated hydrocarbons PAHs

HS HS HS HS HS

CAR/PDMS PPY-coated gold wire PDMS/DVB/CAR PDMS/DVB PDMS, PA

GC/FID GC/FID GC/FID GC/MS MCC/UV-IMS

95 96 97 98 99

PDMS

GC/MS

100

PDMS-coated capillary PDMS PDMS-coated capillary PPY-coated capillary PDMS, fibre-retracted device PPY-DS PDMS

HPLC HPLC/FLD GC/MS HPLC/UV GC/MS GC/MS, FID GC/ECD

101 102 25 48 103 104 105

PCBs, pesticides Pesticides Pesticides Pesticides Pesticides Pesticides Pesticides Pesticides

DI, ultrasound treatment In-tube SPME DI In-tube SPME In-tube SPME Static water sampling HS HS, microwave assisted Static water sampling In-tube SPME In-tube SPME In-tube SPME HS DI DI DI

GC/MS HPLC/ESI/MS HPLC/UV LC/UV GC/MS GC/MS/MS GC/MS GC/ECD, GC/MS

106 26 24,27
29 30 107 108 109 110

Pesticides

DI

PDMS PPY-coated capillary Omegawax 250 Super-Q PLOT PA PDMS/DVB PDMS/DVB CW/DVB, CAR/PDMS, DVB/ CAR/PDMS PDMS, PA

GC/MS, GC/ICP-MS

111

PAHs PAHs PAHs PAHs PAHs PAHs PCBs

Handbook of Solid Phase Microextraction

Analytes

Table 8.10 (Continued) Extraction Method

Fibre/Capillary

Detection

Reference

Pesticides Pesticides Pesticides Pesticides

DI HS, microwave DI HS

GC/MS GC/ECD HPLC GC/ECD

112 113 114 115

Herbicides Herbicides Herbicides Herbicides Phenols Phenols Phenols Organotin Organotin Organotin Organometallic compounds

DI DI DI In-tube SPME HS HS HS HS HS HS HS

HPLC GC/MS MEKC LC/MS GC/FID GC/FID GC/FID GC/PFPD GC/FID GC/MS GC/MS

116 117 118 31,32 119 120 121 122 123 124,125 126

Organometallic compounds Methylmercury, mercury(II) Chlorinated hydrocarbons Chlorohydrocarbons Explosives Explosives VOCs, MTBE, etc. MTBE Triazines

HS DI HS HS DI DI HS HS DI

PA DVB/CAR/PDMS PDMS/DVB Polymethylphenylvinylsiloxane, sol
gel PDMS/DVB, CW/TPR PA PDMS/DVB DB-WAX Polyaniline coated fibre C4/OH-TSO coated fibre PDMS/DVB, DVB, derivatisation CAR/PDMS, derivatisation PDMS, derivatisation PDMS, derivatisation PDMS, DVB/CAR/PDMS, derivatisation PDMS, derivatisation PDMS Actived carbon fibre Actived carbon fibre CW/DVB CAR PDMS, CAR/PDMS, DVB/PDMS CAR/PDMS CW/templated resin, PDMS/DVB

GC/MS GC/MS GC/MS GC/MS GC/ECD HPLC/UV GC/MS GC/MS HPLC

127 128 129,130 131 132 133 134 135 136

SPME and Environmental Analysis

Analytes

(Continued)

271

272

Table 8.10 (Continued) Extraction Method

Fibre/Capillary

Detection

Reference

Triazines Methylamine Iodophenols Chromium

DI DI DI DI, HS

PDMS/DVB CW/TPR, derivatisation CAR/PDMS PDMS, derivatisation

137 138 139 140

Alkanethiols, dihydrogen sulphide Aromatic amines Odorous trihalogenated anisoles DCP, MITC Chlorophenols Aldehydes PBDEs, PBBs Organic pollutants BTEX

Direct DI HS-SPME Direct HS HS HS Static water sampling Controlled flow rate

PDMS/DVB, derivatisation PDMS/DVB, derivatisation PDMS PA PA PDMS/DVB, derivatisation PDMS, PA PDMS CAR/PDMS, PDMS/DVB

GC/MS HPLC CE/ICP-MS GC/ECD, MS, GC/ICPMS GC/MS GC/MS GC/MS GC/ECD/NPD GC/FID GC/MS GC/MS/MS GC/MS GC/MS, FID

141 142 33 143 144 145 146 147 85,148

Handbook of Solid Phase Microextraction

Analytes

SPME and Environmental Analysis

273

Figure 8.5 Flow-through system for sampling PAHs using SPME.

Figure 8.6 Schematic diagram of modified flow-through system.

close to its bottom and an outlet near its top. A little air was left inside the DispoDialyzers to suspend them in the top of the chamber. A magnetic stir-bar was placed in the bottom of the permeation chamber, and a digital magnetic stirrer was used for agitation. Analytes were dissolved in the water inside the DispoDialyzer. The dissolved analyte molecules further diffused through the membrane of the DispoDialyzer and were carried out by the water to the sampling cylinder and the sampling chamber. As the solids (PAHs) and liquid (water) coexist inside the DispoDialyzer, the concentrations of the analytes will remain constant (saturated concentration) if the temperature remains constant. If the flow rate of water is kept constant, the diffusion of analyte molecules from inside the

274

Handbook of Solid Phase Microextraction

Figure 8.7 Schematic diagram of the new flow-through system for the generation of standard aqueous solution by permeation.150 (Source: Figure reprinted with permission.)

Figure 8.8 Schematic diagram of the DispoDialyzer.

DispoDialyzer to the water outside will reach steady state, creating constant analyte concentrations. The observed change in the concentration of the six PAH compounds was less than 20% over 3 months. Each DispoDialyzer was filled with 20B50 mg of the respective compound, and the lifetime of the generator is more than 1 year. The main difference between the new permeation system and a typical dilution system is that the syringe pump was replaced with a permeation generator. The permeation-based generator approach offers many advantages, including convenience, inexpensive equipment (the DispoDialyzer is refillable), solvent-free operation, long lifetime and production of high concentrations of the target analytes. A high concentration of analytes in the flow-through system is very important because it will obviously shorten the testing time for passive samplers. It is difficult to achieve a high concentration of analytes when a peristaltic pump or syringe

SPME and Environmental Analysis

275

pump is used for dilution, due to the presence of a high concentration of solvent and the frequent need to refill.

8.5.2.2 Distribution Coefficients Between Fibre Coating and Water (Kfw) Table 8.11 presents some Kfw data for different fibre coating and various organic compounds.

8.5.3

Solid Matrix

Accurate quantitation of target analytes in solids represents a very significant challenge to the analytical community. Although SPME cannot be used directly to extract analytes from solids, several approaches can be taken to facilitate simple sample preparation. For volatiles, the typical approach is to perform HS analysis. To quantitatively release analytes from the matrix, the temperature needs to be increased. This facilitates higher extraction amounts and faster kinetics of the process. Loss of sensitivity associated with a decrease in the distribution constant can be compensated by cooling the fibre. Table 8.12 illustrates extraction data obtained for BTEX spiked into a clay matrix using several extraction conditions.155 When extraction is conducted at room temperature, the recoveries are only 0.01% or less, indicating the small amount of analyte present in the HS. Increasing the temperature to 50 C increases the extracted amount by about an order of magnitude; however, a further increase of temperature results in the reduction of recoveries because Kfs decreases.155 The addition of water helps to displace the analytes from the surface of clay and results in further improvement by a factor of 2. The situation changes dramatically when high extraction temperature is combined with a cold coating (see Figure 3.7). In this system, the analytes are removed to the HS and they are concentrated onto the cold fibre, resulting in high recoveries. Further increase of the recoveries, corresponding to exhaustive values, can be accomplished by adding a displacing reagent, such as water, to facilitate rapid transfer of the analytes to the HS. The above preliminary data illustrate the effect of parameters which can be adjusted to optimise the SPME of volatiles present in solids. Table 8.13 illustrates the comparison between two exhaustive extraction methods, purge-and-trap (P&T) and internally cooled SPME, for the analysis of a real clay sample. The results are similar, considering the preliminary nature of this experiment. The other successful indirect SPME approach to analysis of solids involves the use of water or a polar organic solvent such as methanol. Pure water or solvent is added to remove analytes from the matrix first, followed by SPME. When a polar solvent is used, the extract is spiked into pure water. Low-temperature water extraction followed by SPME is found to be a very useful approach for polar compounds, such as herbicides.156 Application of methanol with water spiking, on the other hand, has been found to be useful for analysis of volatile hydrocarbons. Table 8.14 summarises the quantitation data obtained for the extraction of native BTEX from soils and GC/ion trap MS determination. The SPME results are

276

Handbook of Solid Phase Microextraction

Table 8.11 Distribution Coefficients Between Different SPME Fibre Coating and Water Compound

log Kfw (100 µm PDMS, 25 C)151

log Kfw (85 µm PA, 25 C)151

Naphthalene Acenaphthylene Acenaphthene Fluorene Phenanthrene Anthracene Fluoranthene Pyrene Benz[a]anthracene Chrysene Benzo[b] fluoranthene Benzo[k] fluoranthene Benzo[a]pyrene Dibenzo[a,h] anthracene Benzo[ghi]perylene Indeno[1,2,3-cd] pyrene

3.02 3.40 3.63 3.71 3.96 3.98 4.71 4.86 5.26 5.69 5.17

3.37 4.01 4.09 4.32 4.39 4.66 4.87 4.84 5.34 4.95 4.34

5.33

4.39

5.39 4.86

5.62 4.91

4.28 4.43

4.03 4.43

Compound

log Kfw (7 µm PDMS, 25 C)152

log Kfw (100 µm PDMS, 25 C)152

PCB 1 PCB 15 PCB 28 PCB 47 PCB 101 PCB153 PCB 202 PCB 180 PCB 206 PCB 209

4.44 5.11 5.47 5.86 6.21 6.68 6.77 6.76 7.04 6.84

4.09 4.83 5.18 5.64 6.08 6.45 6.20 6.54 6.16 5.59

Compound

Kfw (30 µm PDMS, 22 C)153

Kfw (85 µm PA, 22 C)153

Dimethylphthalate Diethylphthalate Di-n-propylphthalate Diisobutylphthalate Di-n-butylphthalate Di-2-ethylhexylphthalate

20 59 320 2,090 2,230 128,080

33 218 1,030 3,340 4,945 9,535 (Continued)

SPME and Environmental Analysis

277

Table 8.11 (Continued) Compound

Kfw (65 µm PDMS-DVB, 24 C)154

Kfw (65 µm CW-DVB, 25 C)154

1,2,3,4-Tetrachlorobenzene (TeCB) Pentachlorobenzene (PeCB) Hexachlorobenzene (HCB) α-Hexachlorocyclohexane (α-HCH) β-Hexachlorocyclohexane (β-HCH) γ-Hexachlorocyclohexane (γ-HCH) δ-Hexachlorocyclohexane (δ-HCH) 1,1-Dichloro-2,2-bis(p-chlorophenyl)ethylene (p, p0 -DDE) 1,1-Dichloro-2,2-bis(p-chlorophenyl)ethane (p, p0 -DDD) PCB 28 PCB 52 PCB 101 PCB 138 PCB 153 PCB 180 Phenanthrene Anthracene Fluoranthene Pyrene Benzo[a]pyrene

304,000 468,000 407,000 120,000 74,100 158,000 52,500 275,000

189,000 281,000 281,000 253,100 266,300 24,000 11,000 457,000

256,000

226,000

516,000 477,000 494,000 427,000 504,000 320,000 275,000 227,000 240,000 206,000 20,000

620,000 674,000 860,000 769,000 942,000 367,000 238,000 206,000 240,000 212,000 27,500

compared with the standard P&T approach. In the majority of cases, the agreement between both methods is very good.157 An interesting modification to the above procedure involves volatising the extract followed by fibre extraction from the gaseous phase158 because the quantitation of analytes in the gas phase is easier, as discussed in Section 8.5.1. For less-volatile analytes, the methanol approach described earlier can give good results. In addition, hot water extraction is a very suitable solvent-free alternative. Table 8.15 summarises preliminary data obtained for a dynamic system indicating two important factors. It is important that the fibre be immersed in the aqueous phase when it is cooled down. Extraction from the hot aqueous phase results in loss of more-volatile analytes. On the other hand, when the fibre is not immersed continuously in water during collection or cooldown, loss of poorly soluble analytes occur. Both static and dynamic approaches of hot water extraction have their advantages. The static method is very simple and inexpensive because it does not use high-pressure pumps. The dynamic approach, on the other hand, provides extract that is much cleaner than the original matrix. However, it might be possible to extract many semivolatile target analytes from very dirty matrices with the high-temperature static system when using the HS mode of SPME.

278

Handbook of Solid Phase Microextraction

Table 8.12 Recoveries (%) of BTEX in a Clay Matrix Under Different Extraction Conditions154 Compounds

Benzene Toluene Ethylbenzene m,p-Xylene o-Xylene

Methods Ia

II b

III c

IV d

Ve

0.01 0.01 0.01 0.01 0.01

0.08 0.06 0.06 0.07 0.08

0.11 0.14 0.21 0.22 0.24

27 11 6 11 7

80 86 91 93 98

a

Room temperature HS sampling with a normal SPME device (without cooling). 50 C HS sampling with a normal SPME device (without cooling). Adding 15% water to clay matrix and 50 C HS sampling with a normal SPME device (without cooling). d 250 C HS sampling with an internally cooled SPME device. e Adding 5% water to clay matrix and 170 C HS sampling with an internally cooled SPME device; 2 min extraction in all cases. Source: Table reprinted with permission. b c

Table 8.13 Concentrations (ng/g) of Native BTEX in a Real Clay Sample, Analysed by Both P&T GC/MSD and Internally Cooled SPME GC/Ion Trap MS Compounds

P&T/GC/MSD a

IC-SPME/GC/ITMS b

Benzene Toluene Ethylbenzene m,p-Xylene o-Xylene



7.4 35.5 4.8


4.5 10.6 22.8 7.1

a

Analysis conducted by the Wastewater Technology Center, Burlington, Ontario. Extracted by internally cooled SPME with sampling temperature at 170 C, analysed by a GC ion trap mass spectrometer.

b

Table 8.14 Results of Two Analytical Methods for BTEX Applied to Four Contaminated Soils Soil

S1 S3 S4 S2

Method

P&T/MSD SPME/ITMS P&T/MSD SPME/ITMS P&T/MSD SPME/ITMS P&T/MSD SPME/ITMS

Compound Concentration (µg/g) B

T

E

m/p-Xylene

o-Xylene

2.92 2.63 t 0.32 0.82 0.78 w 0.02

3.07 3.40 0.33 0.47 1.46 1.25 t 0.06

42.5 43.7 24.9 19.8 27.8 20.4 0.19 0.17

169 180 60.5 47.0 104 96.9 0.85 0.78

48.3 43.0 2.23 1.69 2.62 2.28 0.10 0.09

w, not detected; t, detected but below the statistical instrument detection limit.

SPME and Environmental Analysis

279

Internally cooled SPME fibre device has been automated159 and additional applications have been reported.160
165 Some applications of SPME for sampling in soils and sediments were assisted by sonication132,166,167 and microwaves.168
170 A hollow-fibre-membrane-protected SPME technique also was reported for the determination of herbicides in sewage sludge samples.171 Table 8.16 presents some applications of SPME for environmental analysis in soil and sediment samples in recent years.

8.6

Applications of SPME for Various Analytes in Environmental Samples

A lot of research has been done on samples of environmental origin: air, water, sludge and soils. The majority of applications have been developed for aqueous matrices. The results obtained for priority pollutants in water are very encouraging, indicating that the performance of SPME can meet US Environmental Protection Agency (EPA) method requirements. Table 8.17 summarises the limit of detection (LOD) results obtained for SPME of volatiles and semivolatiles, including phenols. Similar data for pesticides is available in Table 8.21. In each case, the SPME method can be optimised to meet regulatory agency requirements. The low detection limits reflect the fact that all extracted analytes are introduced to an analytical instrument for determination. The discussion below Table 8.15 Amount Extracted by SPME from Collected Water Under Different Conditions After Dynamic High Temperature Water Extraction as Percent of Amount Extracted by SPME from Spiked Water (% RSD) a Analytes

Direct SPME from Spiked Water A

No Cooling, SPME After Water Extraction B

Cooling, SPME After Water Extraction C

No Cooling, SPME Simultaneous with Water Extraction D

Cooling, SPME Simultaneous with Water Extraction E

Naphthalene Anthracene Fluoranthene Pyrene Benzo(a) pyrene

100 (4) 100 (5) 100 (6) 100 (9) 100 (4)

19 (18) 70 (11) 101 (6) 102 (6) 158 (29)

110 (8) 99 (10) 107 (6) 103 (6) 50 (8)

19 (18) 76 (8) 102 (6) 102 (10) 162 (18)

96 (7) 97 (5) 111 (8) 111 (8) 102 (7)

a

Sample: spiked sand. All experiments were performed in triplicate. Amount spiked to sand equals amount spiked to water. A: Extraction with SPME fibre was performed on spiked 30 mL of water, without high-temperature water extraction. B: Collection device was without cooling bath and extraction with SPME fibre after the water extraction was complete. C: Collection device was with cooling bath and extraction with SPME fibre after water extraction was complete. D: Collection device was without cooling bath and with simultaneous extraction with SPME fibre. E: Collection device was with cooling bath and simultaneous extraction with SPME fibre. For B, C, D and E, spiking was performed onto sand placed in the extraction cell. Extraction was performed at 300 C, 300 atm, and 1.2 mL/min for 15 min.

280

Table 8.16 Applications of SPME for Environmental Analysis in Soil and Sediment Samples Analytes

Extraction Method

Fibre/Capillary

Detection

Reference

PBPs TNT and its degradation products Organic pollutants BTEX BTEX BTEX Organotin PAHs PAHs DCP, MITC

Static sampling (sediment) Static sampling (sediment)

PDMS PA

GC/ECD HPLC

172 173

Static sampling (soil) HS, cooled fibre HS (sand) Multiple HS (soil) HS (sediment) Direct SPME (sediment) DI, microwave-assisted (sediment) HS (soil)

PDMS PDMS CAR/PDMS CAR/PDMS PDMS, derivatisation PDMS PDMS, PA PA

HS (soil)

Explosives Methylphosphonates Fungicides Fungicides Herbicides

DI (sediment), sonication HS (soil) HS (soil) DI
SPME followed water extraction (soil) HFM (hollow fiber membrane-protected)-SPME (sewage sludge) DI (soil) HS (soil), microwave assisted

Chlorobenzenes Chlorophenols

HS (soil), sample heating, ultrasonic device HS (sediment), sonication

174 155 175 176 124 177 168 147

PDMS PA

181 170

GC/ECD GC/ECD

178 166 167 132 179 180 169 171

Handbook of Solid Phase Microextraction

2-Chloroethyl ethyl sulphide PCDD/F Butyltin

GC/MS GC/MS GC/MS GC/FID GC/MS GC/MS GC/MS GC/ECD/ NPD Acrylate/silicon co-polymer coating, GC sol
gel PDMS, cooling device GC/MS/MS PDMS GC/MIP AED CW/DVB GC/ECD PDMS, PDMS/DVB IMS PA GC/MS PA, ultrasonic GC/MS PDMS/DVB GC/MS

SPME and Environmental Analysis

281

Table 8.17 Summary of Detection Limits for SPME Analysis of Various Classes of Analytes Analyte

SPME LOD (pg/mL)

EPA (pg/mL)

BTEX Polychlorinated HC solvents PAHs PCBs Phenols

1
15 1
100 1.2
20 3 10
800

30
90 10
100 40 60
100 1,500
42,000

summarises the most important accomplishments obtained for several groups of analytes.

8.6.1

BTEX

BTEX has been the most investigated group of organic compounds of environmental interest. In many cases, performance of SPME for various matrices has been tested using these analytes. The sensitivity of the SPME method using PDMS coating for these compounds is very high, allowing mid- to low ppt determinations in water. This facilitated immediate application of the technique to practical environmental samples, which provided a significant momentum to the early investigations. BTEX compounds are volatile, so the HS
SPME mode can be used successfully in many applications. Careful validation research has been conducted to compare performance of this method with more established alternatives.182 The results are summarised in Tables 8.12, 8.13 and 8.18. Considering the complexity of the matrices, a good correlation between the P&T and HS
SPME techniques can be noted. Validation work with other VOCs, such as chlorinated hydrocarbons, produces similar conclusions.183

8.6.2

TPH

The measure of TPHs is frequently used to estimate the contamination level of a sample. For air samples, this value can be calculated directly from chromatographic data without need for identification of all individual components, as emphasised in Section 8.5.1. TPH can then be reported, together with more specific information related to distribution and type of contaminants present in the sample, without performing additional experiments. TPH of water samples is possible if the groups of compounds are separated and identified by a chromatographic technique combined with a quantitation method such as mass spectrometry (see the discussion in Section 8.5.2). For solids, this value can be obtained after the analytes are transferred into the gaseous or aqueous phase during the extraction process, as described in Section 8.5.3.

282

Handbook of Solid Phase Microextraction

Table 8.18 Results of Two Analytical Methods for BTEX Applied to Four Non-Spiked Municipal Sludge Samples Sludge

Method

Guelph

P&T/MSD SPME/ITMS P&T/MSD SPME/ITMS P&T/MSD SPME/ITMS P&T/MSD SPME/ITMS

Calgary Winnipeg Halton

Compound Concentration (µg/L) B

T

E

m/p-Xylene

o-Xylene

w 2.70 2.52 3.18 w 2.04 3.00 2.94

24.6 17.7 50.0 31.8 5.96 4.88 900

6.48 10.9 36.4 35.1 5.22 6.86 20.6 19.8

13.8 10.5 231 167 21.7 27.9 62.4 31.4

4.50 7.89 103 123 10.5 12.1 29.4 31.5



w, not detected;  , not calculated/out of range for isotope dilution.

Table 8.19 Analytical Characteristics for the Analysis of Selected Polar Compounds in Water by SPME Compound

RSD (%) (100 ng/mL)

LOD (ng/mL)

Acetone Tetrahydrofuran Methylethylketone i-Propanol t-Butanol Methylisobutylketone

3.9 0.8 0.8 1.9 3.2 2.9

5 1 1 2 5 0.1

8.6.3

Polar Volatile Solvents

Polar volatile solvents like alcohols, ketones and aldehydes are difficult to quantify at trace levels in aqueous matrices. Extraction with non-polar solvents leads to poor recoveries. More polar solvents cannot be used because they are miscible with water. SPME with a polar polymeric coating has the potential to provide a needed alternative for this application. Table 8.19 illustrates that low ppb detection limits are possible even with flame ionisation detectors (FIDs) when PDMS/DVB coating and 35% salt concentration are used.184

8.6.4

Non-Polar Semivolatile Compounds

Non-polar semivolatiles, such as PAHs and polychlorinated biphenyls (PCBs), are characterised by very large Kfs, which result in very high sensitivities of SPME determination, frequently reaching low ppt levels.185,186 The major advantage of SPME over alternative techniques is its field portability. Frequently, it is very

SPME and Environmental Analysis

283

Table 8.20 The Measured Distribution Constants K Between PA Fibre and Water, the Octanol
Water Partition Coefficients (log Kow), Aqueous Solubility (mg/L) and LODs of Heteroaromatic Compounds by SPME/GC/FID and SPME/GC/ITMS in μg/L and % RSD Compound

K

log Kow

Sw

LOD-FID (µg/L)

LOD-MS % RSD (µg/L) ITMS

Thiophene 1-Methylpyrrole Pyrrole 2-Methylpyridine 2,4-Dimethylpyridine Benzofuran Benzothiophene Quinoline Indole 2-Methylquinoline Dibenzofuran Dibenzothiophene Acridine Carbazole DBT-sulphone

0.34 0.04 0.1 0.02 0.09

1.81 NA 0.75 1.06 NA

3,600 Soluble 16,000 Soluble Soluble

n.d. n.d. n.d. n.d. n.d.

1 2.5 10 10 10

14 13 12 5.9 14

3.1 7.9 0.46 3.2 0.53 22 27 7.7 29 4.9

2.67 3.12 2 2.03 2.23 4.12 5.45 3.71 3.5 NA

224 130 3,000 6,500 NA 10 1 1.2 38 NA

3 2 15 2 10 2 2 0.5 0.5 0.5

0.03 0.02 0.3 0.02 0.2 0.03 0.02 0.02 0.02 0.04

4.2 5.8 10 6.9 3.3 10 11 9.2 10 9.1

NA (not available), n.d. (not detected).

difficult to obtain and preserve a representative sample because non-polar semivolatiles have a tendency to adsorb on solid surfaces associated with the sampling vessel or on particulate matter. By sampling directly in the field, these limitations can be substantially reduced.

8.6.5

Polar Semivolatile Compounds

Polar semivolatiles in natural matrices represent a very significant analytical challenge. Adjustments of pH and salt addition help to reach the required detection limits. For example, LODs obtained for phenols with direct extraction from an acidified solution with a pH stable PA coating are lower than values specified by the EPA (Table 8.20).94 Derivatisation with acetic anhydride allows further improvement in sensitivity and chromatographic performance. The derivatisation approach also enhances the extraction performance of carboxylic acids43 and amines, particularly for lower molecular weight species. Table 8.20 shows the results obtained for very polar heteroaromatic compounds frequently present in creosote contaminated groundwater.187 Extraction with a PA coating, combined with sensitive ion trap mass spectrometry detection, allows low to sub-ppb detection limits for these very water-soluble compounds.

284

Handbook of Solid Phase Microextraction

Table 8.21 Detection Limits for SPME Coupled with Various Detectors Pesticide Class

Target Analyte Subgroups

Nitrogencontaining herbicides SPME with 85-μm PA fibre

Thiocarbamates Triazines Nitroanilines Substituted uracils Substituted amides Acetanilides Diphenyl ethers Triazoles

Organochlorine pesticides SPME with 100-μm PDMS fibre

Benzene hexachlorides 0.9
9 (BHCs) Hexachlorocyclodienes 0.06
1.6 Diphenyl aliphatics 0.05
4.7

Phosphorouscontaining pesticides SPME with 85-μm PA fibre

Phosphate Phosphorothiolate Phosphonothiate Pyrophosphate Phosphorodithioates Phosphorothioates

8.6.6

ECD (ng/L)

NPD (ng/L)

MS (ng/L)

EPA Methods (ng/L)

20
60 40
6,000 10
30 200
400 800 200 300 30

0.05
0.8 0.4
3 0.02
0.4 0.1
1 15 0.01 6 0.01

100
200 100
800 200 2,500
4,500 500 700 N/A N/A

0.01
0.04 10
25

500 130 15 16 9
320 11
280

0.02
0.6 0.06
4.5

2.5
50 10
75

6 2 8 1 0.4
9 0.7
100

100
2,500 N/A 9
200 6 9
1,500 4
2,000

Pesticides

Pesticides are an analytically difficult group of compounds because of their wide range of chemical structures and properties. Some of them are classified as nonpolar (organochlorine pesticides) and others as very polar semivolatiles (herbicides). Despite this diversity, SPME has been very successfully applied for their determinations in aqueous matrices by several research groups.188
192 Table 8.21 summarises the SPME detection limits obtained for several groups of pesticides. They are lower compared to EPA method LODs. It is even possible to develop single extraction conditions to facilitate screening of 60 target pesticides combining three classes: organochlorines, organophosphorous and nitrogen-containing herbicides.190 In addition, the round-robin test discussed in Chapter 7 demonstrates the ruggedness of the SPME/GC pesticide method and its ability to be rapidly implemented in different laboratories. Many pesticides are non-volatile compounds requiring direct extraction, which results in relatively long equilibration times (frequently exceeding 60 min with magnetic stirring). Increasing temperatures will decrease the equilibration times, but cooling the fibre might be necessary to maintain good detection limits. GC separation methods limit the application of the SPME technique to

SPME and Environmental Analysis

285

thermally stable pesticides. Application of HPLC would allow analysis not only of less stable species such as carbamates, but also products of pesticide metabolism and degradation.

8.6.7

Metal Ions and Organometallic Compounds

The application of sodium tetraethylborate as an in situ derivatisation reagent results in quantitation of some metal ions and organometallic species in aqueous samples by SPME. This approach was first used for the determination of methylmercury in water and biota,193 followed by the application of this approach to quantitation of lead ion and organolead compounds.194 In the method, analytes are first derivatised to fully ethylated forms before they are extracted by the PDMS fibre. These species are volatile; therefore, the HS extraction mode of SPME can be applied, facilitating rapid analysis of very complex matrices. The reaction time, which takes about 15 minutes, typically determines the extraction time. Stirring must be very efficient during derivatisation because the reagent decomposes rapidly after contact with the sample. The affinity of neutral forms of organometallic compounds towards a PDMS coating is very high, resulting in very low detection limits approaching sub-ppt levels for determination of tetraethyllead in water by SPME/ GC/ITMS. This method can be applied to partial speciation of various forms of these metals present in a sample, first by extracting native neutral organometallic species followed by determination of the total metal content in the sample. Full speciation can be accomplished by isotopically labeling the derivatisation reagent and then differentiating between various forms based on their mass spectra.195

8.7

Concluding Remarks

A wide range of applications of SPME for environmental analysis have been reported. It is evident that the SPME technique is very useful for extraction of a variety of chemicals from environmental samples, including very complex matrices. Off-site analysis of environmental samples is now performed in a fully automated fashion using autosamplers or robots. The main remaining challenge is development of the certified methods that would facilitate broader application in high-throughput environmental laboratory applications of this green extraction technology. On-site sampling and analysis of environmental samples with SPME is facilitated by development of sampling devices and calibration methods, for both grab sampling and long-term monitoring. The development of new SPME calibration methods has accelerated the applications of SPME for on-site sampling and analysis. On-site analysis requires portable and solvent-free sample preparation approaches, which work well with micro-instruments.196 SPME on-site sampling techniques, especially the techniques for field water and solid sampling, is the main scientific opportunity for SPME in the near future.

286

Handbook of Solid Phase Microextraction

References 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34. 35. 36. 37. 38. 39. 40. 41. 42. 43. 44. 45.

http://www.spme.uwaterloo.ca/SPMEdata/spmedata.html G Ouyang & J Pawliszyn, Anal Bioanal Chem 386 (2006) 1059 G Ouyang & J Pawliszyn, Trends Anal Chem 25 (2006) 692 G Ouyang & J Pawliszyn, Anal Chim Acta 627 (2008) 184 J Wu, X Yu, H Lord & J Pawliszyn, Analyst 125 (2000) 391 J Wu & J Pawliszyn, J Chromatogr A 909 (2001) 37 Z Wang, C Xian, C Wu & H Han, J Chromatogr A 893 (2000) 157 R Gomes da Costa, Silva & F Augusto, J Chromatogr A 1072 (2005) 7 VG Zuin, AL Lopes, JH Yariwake & F Augusto, J Chromatogr A 1056 (2004) 21 AL Lopes & F Augusto, J Chromatogr A 1056 (2004) 13 M Azenha, C Malheiro & AF Silva, J Chromatogr A 1069 (2005) 163 M Giardina & SV Olesik, Anal Chem 73 (2001) 5841 M Giardina, L Ding & SV Olesik, J Chromatogr A 1060 (2004) 215 C Dong, Z Zeng & X Li, Talanta 66 (2005) 721 X Li, Z Zeng, S Gao & H Li, J Chromatogr A 1023 (2004) 15 X Li, Z Zeng & J Zhou, Anal Chim Acta 509 (2004) 27 F Zhou, X Li & Z Zeng, Anal Chim Acta 538 (2005) 63 L Yun, Anal Chim Acta 486 (2003) 63 Z Zeng, W Qiu & Z Huang, Anal Chem 73 (2001) 2429 D Wang, J Peng, J Xing, C Wu & Y Xu, J Chromatogr Sci 42 (2004) 57 Z Zeng, W Qiu, M Yang, X Wei, Z Huang & F Li, J Chromatogr A 934 (2001) 51 D Wang, J Xing, J Peng & C Wu, J Chromatogr A 1005 (2003) 1 EE Stashenko & JR Martinez, Trends Anal Chem 23 (2004) 553 R Eisert & J Pawliszyn, Anal Chem 69 (1997) 3140 D Globig & C Weickhardt, Anal Bioanal Chem 381 (2005) 656 J Wu, C Tragas, H Lord & J Pawliszyn, J Chromatogr A 976 (2002) 357 Y Gou, R Eisert & J Pawliszyn, J Chromatogr A 873 (2000) 137 Y Gou & J Pawliszyn, Anal Chem 72 (2000) 2774 Y Gou, C Tragas, H Lord & J Pawliszyn, J Microcol Sep 12 (2000) 125 Y Hirayama, M Ohmichi & H Tatsumoto, J Health Sci 51 (2005) 526 M Takino, S Daishima & T Nakahara, Analyst 126 (2001) 602 H Kataoka, K Mitani & M Takino, Method Biotechnol 19 (2006) 365 A Diaz, F Ventura & M Teresa Galceran, J Chromatogr A 1064 (2005) 97 T Bruheim, X Liu & J Pawliszyn, Anal Chem 75 (2003) 1002 G Ouyang, W Zhao, L Bragg, Z Qin, M Alaee & J Pawliszyn, Environ Sci Technol 41 (2007) 4026 Z Qin, L Bragg, G Ouyang & J Pawliszyn, J Chromatogr A 1196
1197 (2008) 89 L Bragg, Z Qin, M Alaee & J Pawliszyn, J Chromatogr Sci 44 (2006) 317 C Dietz, J Sanz & C Camara, J Chromatogr A 1103 (2006) 183 H Lord & J Pawliszyn, J Chromatogr A 885 (2000) 153 SL Chong, D Wang, JD Hayes, BW Wilhite & A Malik, Anal Chem 69 (1997) 3889 J Yu, C Wu & J Xing, J Chromatogr A 1036 (2004) 101 AK Malik, V Kaur & N Verma, Talanta 68 (2006) 842 L Pan & J Pawliszyn, Anal Chem 69 (1997) 196 MR Lee, RJ Lee, YW Lin, CM Chen & BH Hwang, Anal Chem 70 (1998) 1963 I Rodriguez, E Rubi, R Gonzalez, JB Quintana & R Cela, Anal Chim Acta 537 (2005) 259

SPME and Environmental Analysis

46. 47. 48. 49. 50. 51. 52. 53. 54. 55. 56. 57. 58. 59. 60. 61. 62. 63. 64. 65. 66. 67. 68. 69. 70. 71. 72. 73. 74. 75. 76. 77. 78. 79. 80. 81. 82. 83. 84. 85. 86. 87. 88.

287

J Koziel, J Noah & J Pawliszyn, Environ Sci Technol 35 (2001) 1481 Q Wang, J O’Reilly & J Pawliszyn, J Chromatogr A 1071 (2005) 147 J Wu & J Pawliszyn, Anal Chem 73 (2001) 55 K Alhooshani, TY Kim, A Kabir & A Malik, J Chromatogr A 1062 (2005) 1 C Basheer, S Jegadesan, S Valiyaveettil & HK Lee, J Chromatogr A 1087 (2005) 252 TY Kim, K Alhooshani, A Kahir, DP Fries & A Malik, J Chromatogr A 1047 (2004) 165 J Adkins & N Henry III, Anal Chem 65 (1993) 133R Z Zhang & J Pawliszyn, J Phys Chem 100 (1996) 17648 The Sadtler Capillary GC Standard Retention Index Library and Data Base, Sadtler Research Laboratories, Philadelphia P Martos, A Saraullo & J Pawliszyn, Anal Chem 69 (1997) 402 P Martos & J Pawliszyn, Anal Chem 69 (1997) 206 M Chai & J Pawliszyn, Environ Sci Technol 29 (1995) 693 D Gorlo, B Zygmunt, M Dudek, A Jaszek, M Pilarczyk & J Namiesnik, Fresenius J Anal Chem 363 (1999) 696 Y Chen & J Pawliszyn, Anal Chem 76 (2004) 6823 M Jia, J Koziel & J Pawliszyn, Field Anal Chem Technol 4 (2000) 73 S Tumbiolo, J Gal, P Maria & O Zerninati, Anal Bioanal Chem 380 (2004) 824 G Xiong, Y Chen & J Pawliszyn, J Chromatogr A 999 (2003) 43 V Larroque, V Desauziers & P Mocho, J Environ Monit 8 (2006) 106 L Muller, T Gorecki & J Pawliszyn, Fresenius J Anal Chem 364 (1999) 610 A Khaled & J Pawliszyn, J Chromatogr A 892 (2000) 455 L Wei, Q Ou, J Li & B Liang, Chromatographia 59 (2004) 601 J Koziel, M Jia, A Khaled, J Noah & J Pawliszyn, Anal Chim Acta 400 (1999) 153 A Sanusi, F Ferrari, M Millet & M Montury, J Environ Monit 5 (2003) 574 F Ferrari, A Sanusi, M Millet & M Montury, Anal Bioanal Chem 379 (2004) 476 E Davoli, ML Gangai, L Morselli & D Tonelli, Chemosphere 51 (2003) 357 F Lestremau, V Desauziers, J Roux & J Fanlo, J Chromatogr A 999 (2003) 71 S Isetun, U Nilsson & A Colmsjo, Anal Bioanal Chem 380 (2004) 319 S Isetun & U Nilsson, Analyst 130 (2005) 94 S Isetun, U Nilsson, A Colmsjo & R Johansson, Anal Bioanal Chem 378 (2004) 1847 F Lestremau, FAT Andersson, V Desauziers & J Fanlo, Anal Chem 75 (2003) 2626 R Batlle, A Colmsjo & U Nilsson, Fresenius J Anal Chem 371 (2001) 514 SW Tsai & TA Chang, J Chromatogr A 954 (2002) 191 R Barro, S Ares, C Garcia-Jares, M Llompart & R Cela, J Chromatogr A 1045 (2004) 189 R Barro, S Ares, C Garcia-Jares, M Llompart & R Cela, J Chromatogr A 1072 (2005) 99 GL Hook, CJ Lepage, SI Miller & PA Smith, J Sep Sci 27 (2004) 1017 PA Martos & J Pawliszyn, Anal Chem 71 (1999) 1513 J Koziel, M Jia & J Pawliszyn, Anal Chem 72 (2000) 5178 F Augusto, J Koziel & J Pawliszyn, Anal Chem 73 (2001) 481 L Tuduri, V Desauziers & JL Fanlo, Analyst 128 (2003) 1028 Y Chen, J A Koziel & J Pawliszyn, Anal Chem 75 (2003) 6485 JA Koziel, PA Martos & J Pawliszyn, J Chromatogr A 1025 (2004) 3 K Kolar, M Ciganek & J Malecha, J Chromatogr A 1029 (2004) 263 VA Isidorov & VT Vinogorova, J Chromatogr A 1077 (2005) 195

288

Handbook of Solid Phase Microextraction

89. V Isidorov, A Purzynska, A Modzelewska & M Serowiecka, Anal Chim Acta 560 (2006) 103 90. A Saraullo, P Martos & J Pawliszyn, Anal Chem 69 (1997) 1992 91. C Arthur, L Killam, S Motlagh, M Lim, D Potter & J Pawliszyn, Environ Sci Technol 26 (1992) 979 92. J Dean, W Tomllnson, V Makovskaya, R Cumming, M Hetheridge & M Comber, Anal Chem 68 (1996) 130 93. C Arthur, L Killam, K Buchholz & J Pawliszyn, Anal Chem 64 (1992) 1960 94. K Buchholz & J Pawliszyn, Anal Chem 66 (1994) 160 95. HJ Cho, K Baek, HH Lee, SH Lee & JW Yang, J Chromatogr A 988 (2003) 177 96. D Djozan, MH Pournaghi-Azar & S Bahar, Chromatographia 59 (2004) 595 97. CMM Almeida & LV Boas, J Environ Monit 6 (2004) 80 98. I Arambarri, M Lasa, R Garcia & E Millan, J Chromatogr A 1033 (2004) 193 99. G Walendzik, JI Baumbach & D Klockow, Anal Bioanal Chem 382 (2005) 1842 100. E Psillakis, A Ntelekos, D Mantzavinos, E Nikolopoulos & N Kalogerakis, J Environ Monit 5 (2003) 135 101. H Bagheri & A Salemi, Chromatographia 59 (2004) 501 102. HW Chen, Anal Sci 20 (2004) 1383 103. G Ouyang, Y Chen & J Pawliszyn, Anal Chem 77 (2005) 7319 104. A Mohammadi, Y Yamini & N Alizadeh, J Chromatogr A 1063 (2005) 1 105. YY Shu, SS Wang, M Tardif & Y Huang, J Chromatogr A 1008 (2003) 1 106. EY Zeng, D Tsukada & DW Diehl, Environ Sci Technol 38 (2004) 5737 107. M Sakamoto & T Tsutsumi, J Chromatogr A 1028 (2004) 63 108. CGoncalves & MF Alpendurada, J Chromatogr A 1026 (2004) 239 109. TA Albanis, DG Hela, DA Lambropoulou & AS Vasilios, Int J Environ Anal Chem 84 (2004) 1079 110. JP Perez-Trujillo, S Frias, JF Conde & MA Rodriguez-Delgado, J Chromatogr A 963 (2002) 95 111. N Fidalgo-Used, M Montes-Bayon, E Blanco-Gonzalez & A Sanz-Medel, J Anal Atom Spectrom 20 (2005) 876 112. SHG Brondi & FM Lancas, J Environ Sci Health B 40 (2005) 513 113. HP Li, GC Li & JF Jen, J Chromatogr A 1012 (2003) 129 114. A Sanchez-Ortega, MC Sampedro, N Unceta, MA Goicolea & RJ Barrio, J Chromatogr A 1094 (2005) 70 115. C Dong, Z Zeng & M Yang, Water Res 39 (2005) 4204 116. HH Lin, YH Sung & SD Huang, J Chromatogr A 1012 (2003) 57 117. R Carabias-Martinez, C Garcia-Hermida, E Rodriguez-Gonzalo, FE Soriano-Bravo & J Hernandez-Mendez, J Chromatogr A 1002 (2003) 1 118. S Frias-Garcia, MJ Sanchez & MA Rodriguez-Delgado, J Sep Sci 27 (2004) 660 119. H Bagheri, A Mir & E Babanezhad, Anal Chim Acta 532 (2005) 89 120. F Zhou, X Li & Z Zeng, Anal Chim Acta 538 (2005) 63 121. A Urbanczyk & J Staniewski, Chem Anal (Warsaw) 50 (2005) 749 122. M Le Gac, G Lespes & M Potin-Gautier, J Chromatogr A 999 (2003) 123 123. I Arambarri, R Garcia & E Millan, Chemosphere 51 (2003) 643 124. C Devos, M Vliegen, B Willaert, F David, L Moens & P Sandra, J Chromatogr A 1079 (2005) 408 125. CC Chou & MR Lee, J Chromatogr A 1064 (2005) 1 126. G Centineo, E Blanco Gonzalez & A Sanz-Medel, J Chromatogr A 1034 (2004) 191 127. D Parkinson, I Bruheim, I Christ & J Pawliszyn, J Chromatogr A 1025 (2004) 77

SPME and Environmental Analysis

289

128. S Mishra, RM Tripathi, S Bhalke, VK Shukla & VD Puranik, Anal Chim Acta 551 (2005) 192 129. TH Sun, LK Cao & JP Jia, Chromatographia 61 (2005) 173 130. TH Sun, JP Jia, NH Fang & L Wang, Anal Chim Acta 530 (2005) 33 131. T Sun, N Fang, Y Wang, J Jia & J Yu, Anal Lett 37 (2004) 1411 132. F Monteil-Rivera, C Beaulieu & J Hawari, J Chromatogr A 1066 (2005) 177 133. F Monteil-Rivera, C Beaulieu, S Deschamps, L Paquet & J Hawari, J Chromatogr A 1048 (2004) 213 134. S Nakamura & S Daishima, Anal Chim Acta 548 (2005) 79 135. F Fang, CW Hong, S Chu, W Kou & A Bucciferro, J Chromatogr A 1021 (2003) 157 136. SD Huang, HI Huang & YH Sung, Talanta 64 (2004) 887 137. S Frias, MA Rodriguez, JE Conde & JP Perez-Trujillo, J Chromatogr A 1007 (2003) 127 138. R Herraez-Hernandez, C Chafer-Pericas & P Campins-Falco, Anal Chim Acta 513 (2004) 425 139. SS Kannamkumarath, RG Wuilloud, S Jayasinghe & JA Caruso, Electrophoresis 25 (2004) 1843 140. L Abranko, L Yang, RE Sturgeon, P Fodor & Z Mester, J Anal Atom Spectrom 19 (2004) 1098 141. C Salgado-Petinal, R Alzaga, C Garcia-Jares, M Llompart & JM Bayona, Anal Chem 77 (2005) 6012 142. T Zimmermann, WJ Ensinger & TC Schmidt, Anal Chem 76 (2004) 1028 143. S Fuster, J Beltran, FJ Lopez & F Hernandez, J Sep Sci 28 (2005) 98 144. M Portillo, N Prohibas, V Salvado & BM Simonet, J Chromatogr A 1103 (2006) 29 145. SW Tsai & CM Chang, J Chromatogr A 1015 (2003) 143 146. M Polo, G Gomez-Noya, JB Quintana, C Garcia-Jares & R Cela, Anal Chem 76 (2004) 1054 147. T Nilsson, L Montanarella, D Baglio, R Tilio, G Bidoglio & S Facchetti, Int J Environ Anal Chem 69 (1998) 217 148. K Sukola, J Koziel & J Pawliszyn, Anal Chem 73 (2001) 13 149. B Shurmer & J Pawliszyn, Anal Chem 72 (2000) 3660 150. G Ouyang, Y Chen & J Pawliszyn, J Chromatogr A 1105 (2006) 176 151. R Doong & S Chang, Anal Chem 72 (2000) 3647 152. ZY Yang, EY Zeng, H Xia, JZ Wang, BX Mai & KA Maruya, J Chromatogr A 1116 (2006) 240 153. U Kotowska, K Garbowska & VA Isidorov, Anal Chim Acta 560 (2006) 110 154. A Paschkea & P Popp, J Chromatogr A 999 (2003) 35 155. Z Zhang & J Pawliszyn, Anal Chem 67 (1995) 34 156. A Boyd-Boland & J Pawliszyn, J Chromatogr 704 (1995) 163 157. B MacGillivray, Analysis of Substituted Benzenes in Environmental Samples by Headspace Solid Phase Microextraction, M.Sc. Thesis, University of Waterloo, Waterloo (1996) 158. A Saraullo, Determination of Petroleum Hydrocarbons in the Environment by Solid Phase Microextraction, M.Sc. Thesis, University of Waterloo (1996) 159. Y Chen & J Pawliszyn, Anal Chem 78 (2006) 5222 160. AR Ghiasvand, S Hosseinzadeh & J Pawliszyn, J Chromatogr A 1124 (2006) 35 161. E Carasek, E Cudjoe & J Pawliszyn, J Chromatogr A 1138 (2007) 10 162. E Carasek & J Pawliszyn, J Agr Food Chem 54 (2006) 8688 163. Y Chen, F Begnaud, A Chaintreau & J Pawliszyn, J Sep Sci 30 (2007) 1037

290

164. 165. 166. 167. 168. 169. 170. 171. 172. 173. 174. 175. 176. 177. 178. 179. 180. 181. 182. 183. 184. 185. 186. 187. 188. 189. 190. 191. 192. 193. 194. 195. 196.

Handbook of Solid Phase Microextraction

Y Chen, F Begnaud, A Chaintreau & J Pawliszyn, Flavour Frag J 21 (2006) 822 AR Ghiasvand, L Setkova & J Pawliszyn, Flavour Frag J 22 (2007) 377 KJ Chia, TY Lee & SD Huang, Anal Chim Acta 527 (2004) 157 J Carpinteiro, I Rodriguez & R Cela, Anal Bioanal Chem 380 (2004) 853 V Pino, JH Ayala, AM Afonso & V Gonzalez, Anal Chim Acta 477 (2003) 81 DA Lambropoulou & TS Albanis, Anal Chim Acta 514 (2004) 125 MC Wei & JF Jen, J Chromatogr A 1012 (2003) 111 C Basheer & HK Lee, J Chromatogr A 1047 (2004) 189 P Mayer, W Vaes, F Wijnker, K Legierse, R Kraaij, J Tolls & J Hermens, Environ Sci Technol 34 (2000) 5177 J Conder, T La Point, G Lotufo & J Steevens, Environ Sci Technol 37 (2003) 1625 L van der Wal, T Jager, R Fleuren, A Barendregt, T Sinnige, C Gestel & J Hermens, Environ Sci Technol 38 (2004) 4842 R Dungan, Anal Lett 38 (2005) 2393 O Ezquerro, G Ortiz, B Pons & MT Tena, J Chromatogr A 1035 (2004) 17 SB Hawthorne, CB Grabanski, DJ Miller & JP Kreitinger, Environ Sci Technol 39 (2005) 2795 M Liu, Z Zeng & H Fang, J Chromatogr A 1076 (2005) 16 P Rearden & PB Harrington, Anal Chim Acta 545 (2005) 13 A Navalon, A Prieto, L Araujo & JL Vilchez, Anal Bioanal Chem 379 (2004) 1100 L van der Wal, CAM van Gestel & JLM Hermens, Chemosphere 54 (2003) 561 B MacGillivray, P Fowlie, C Sagara & J Pawliszyn, J Chromatogr Sci 32 (1994) 317 T Nilsson, F Pelusio, L Montanarella, B Larsen, S Facchetti & J Madsen, HRC 18 (1995) 617 T Gorecki, P Martos & J Pawliszyn, Anal Chem 70 (1998) 19 D Potter & J Pawliszyn, Environ Sci Technol 28 (1994) 298 J Langenfeld, S Hawthorne & D Miller, Anal Chem 68 (1996) 144 S Johansen & J Pawliszyn, J High Res Chromatogr 19 (1996) 627 R Eisert & K Levsen, J Am Soc Mass Spectrom 6 (1995) 1119 I Barnabas, J Dean, I Fowlis & S Owen, J Chromatogr 705 (1995) 305 K Graham, L Sarna, G Webster, J Graynor & H Ng, J Chromatogr 725 (1996) 129 A Boyd-Boland, S Magdic & J Pawliszyn, Analyst 121 (1996) 929 P Popp, K Kalbitz & G Oppermann, J Chromatogr 687 (1994) 133 Y Cai & J Bayona, J Chromatogr 696 (1995) 113 T Gorecki & J Pawliszyn, Anal Chem 68 (1996) 3008 X Yu, H Yuan, T Gorecki & J Pawliszyn, Anal Chem 71 (1999) 2998 J Pawliszyn, Anal Chem 75 (2003) 2543

9 Application of Solid-Phase

Microextraction in Food and Fragrance Analysis Lucie Kudlejova and Sanja Risticevic Department of Chemistry, University of Waterloo, Waterloo, Ontario, Canada

9.1

Introduction and Method Development Considerations

Food analysis is crucial for food quality and safety control. Studies in food analysis are focused mainly on the nutritional value of the final product,1
3 food freshness,4 supplementary materials added5,6 or toxic components spontaneously occurring in the product or during the food processing,7,8 and the effect of processing or storage on food composition, texture and microbiological quality.9
11 Photodegradation,12 thermal degradation13 and/or (auto)oxidation14 of components may occur during food processing or storage. Pesticide contamination of foodstuffs from the primary ingredients15,16 or contamination from the packaging materials17 have also been proved and discussed in earlier studies. The food aroma profile, consisting mostly of volatile and semi-volatile components, is another factor that plays an important role in consumer acceptance. It can often reveal the freshness and processing or storage history of the product and, therefore, is a subject of many chemical or sensory studies.18
21 Recently, consumers are also more and more concerned with the characterisation and differentiation of various crops according to their origin or processing technology. Therefore, food authentication and adulteration control are becoming issues to be considered in terms of consumer health, industrial concerns, religious traditions and food-related habits.22
29 Food is a very complex matrix, and several procedures are typically used in order to prepare a food sample for the final gas chromatography (GC), liquid chromatography (LC) or other analysis. Sampling and sample preparation depends on the type of matrix. In order to prepare a representative sample, solid matrices need to be homogenised and liquid or gaseous samples must be properly stirred prior to the isolation of the target analytes from the examined matrix. Fast isolation of the analytes from food matrices is particularly important in order to minimise or prevent changes in sample associated with enzyme activity, lipid oxidation, microbial growth and physical changes that are likely to occur in the unstable food systems. Recently, solid-phase microextraction (SPME) has become one of the most generally used extraction procedures. When it comes to SPME method development for food/fragrance samples, optimisation of the extraction process must be made considering the Handbook of Solid Phase Microextraction. DOI: 10.1016/B978-0-12-416017-0.00009-7 © 2012 Elsevier Inc. All rights reserved.

292

Handbook of Solid Phase Microextraction

complexity of examined samples, as well as the fact that trace target analytes may be co-extracted with interfering matrix components. The parameters that are typically critically evaluated involve the implementation of selective extraction phases towards a specific group of target analytes that simultaneously demonstrate substantially reduced selectivity and sensitivity for interfering non-targeted matrix components. In those cases involving non-targeted fingerprinting and profiling food applications, the objective of coating selection is the opposite: comprehensive metabolite coverage requiring a non-selective extraction phase. These critical requirements have been recently addressed in order to (i) standardise SPME coating selection for a wide range of analytes frequently encountered in food/fragrance matrices and having differing physico-chemical properties and (ii) evaluate the performance of commercial coatings in non-targeted studies of food/plant systems (metabolomics). The study is being performed by considering both polar and non-polar analytes and their distribution on the two-dimensional chromatographic plane generated by comprehensive two-dimensional gas chromatography/time-of-flight mass spectrometry (GCxGC
TOFMS), therefore taking the advantage of GCxGC structurally ordered chromatograms in order to depict the particular sorbent
analyte interactions. For the first objective mentioned above, the fibres were evaluated both in terms of extraction and desorption efficiency in headspace (HS)
SPME mode using water samples spiked with 52 components belonging to ethyl esters, aldehydes, 2-ketones, alkanes, 1-alcohols, 2-alcohols, terpene hydrocarbons and oxygenated terpenes. The enrichment factors were evaluated by considering absolute recovery data in a system operated either in equilibrium conditions or extraction conditions close to equilibrium for most of the examined compounds. Briefly, the study demonstrated that the choice of best coating is highly dependent on the objective of study and target analytes of interest, and no single fibre coating performed best for all the 52 components in the system. For the second aforementioned objective, HS
SPME analysis of apple samples was performed with all commercially available fibre coatings involving polydimethylsiloxane (PDMS), polydimethylsiloxane/divinylbenzene (PDMS/DVB), carboxen/polydimethylsiloxane (CAR/PDMS), polyacrylate (PA), carbowax (CW), divinylbenzene/ carboxen/polydimethysiloxane (DVB/CAR/PDMS) and carbopack Z
PDMS. After performing HS
SPME extraction with each of the coatings, the coatings containing target analytes were thermally desorbed in injector of GCxGC
TOFMS instrument. Alignment of analytes was performed automatically and included signal to noise (S/N) threshold of 50 for peak finding and minimum forward library match factor of 750, followed by manual filtering of column bleed and fibre contaminant peaks. Figure 9.1 illustrates the peak apex plots obtained using PDMS and PA fibre coatings and shows the distribution of peaks in the two-dimensional plane with the x-axis and y-axis representing retention times in first (non-polar column) and second (polar column) dimensions, respectively. As can be seen, the polar analyte coverage (higher retention times in second dimension) is substantially reduced with the use of non-polar PDMS phase. On the other hand, when using moderately polar PA coating, the polar analyte extraction coverage is significantly improved. Solid sorbents incorporating porous polymer such as DVB and porous carbon extract analytes by physical trapping. As a result, the extraction efficiency depends on the strength of adsorbent to retain the extracted analytes (see also Chapters 2

Application of Solid-Phase Microextraction in Food and Fragrance Analysis

PDMS Figure 9.1 GCxGC
TOFMS peak

5 Second dimension RT (s)

293

apex plots illustrating retention positions of volatile and semivolatile metabolites extracted with PDMS and PA coatings and HS
SPME in apple samples.

4 3 2 1 0

Second dimension RT (s)

0

5

1,000 2,000 3,000 4,000 First dimension RT (s) PA

4 3 2 1 0 0

1,000 2,000 3,000 4,000 First dimension RT (s)

and 4), which is a function of available surface area, the amount of porosity, the size of pores and the size of analytes to be retained. The comparison of two peak apex plots corresponding to CAR/PDMS and PDMS/DVB apple extracts is given in Figure 9.2. Based on the results illustrated, it can be seen that small molecular weight analytes (both polar and non-polar) are as expected more efficiently extracted with both solid sorbents as compared to the relevant extraction coverage that was obtained with the two absorbent phases. However, as the molecular weight of analytes increases, the extraction efficiency of the CAR/PDMS coating decreases. This is mainly due to incomplete desorption for high molecular weight compounds from the thick carboxen phase. The desorption efficiency for thickest commercially available CAR/PDMS coating (Figure 9.3) was evaluated in the aforementioned targeted study as well and percentage carryover was on average 25% for the components belonging to homologous series of ethyl esters (from ethyl heptanoate to ethyl tridecanoate). Consequently, higher molecular weight ethyl esters were not detected due to poor desorption efficiency even though aggressive desorption conditions were employed. As illustrated from these examples, these particular studies were able to confirm already established theoretical principles of SPME in a more sophisticated way, at the same time providing standardisation in the SPME coating selection process and insight into the gaps that were missing so far.

294

Handbook of Solid Phase Microextraction

Figure 9.2 GCxGC
TOFMS peak

5

CAR/PDMS apex plots illustrating retention

Second dimension RT (s)

4.5

positions of volatile and semivolatile metabolites extracted with CAR/PDMS and PDMS/DVB coatings and HS
SPME in apple samples.

4 3.5 3 2.5 2 1.5 1 0.5 0 0

1,000

5

PDMS/DVB

4.5 Second dimension RT (s)

2,000 3,000 4,000 First dimension RT (s)

4 3.5 3 2.5 2 1.5 1 0.5 0 0

1,000

2,000 3,000 4,000 First dimension RT (s)

The rest of this chapter provides a summary of the recently published SPME studies and reviews in the field of food and fragrance analysis, together with several interesting case studies highlighting the advantages and disadvantages of this technique compared to other commonly used isolation techniques. Classification studies, typically performed as multi-component analysis, and the suitability of using the SPME approach in such studies as proof of sample authenticity are also discussed in this review.

9.2

Reviews and Case Studies Involving SPME as an Extraction Procedure

Food studies involving SPME have often been summarised and reviewed. The objective of this chapter was to present the most recent SPME food and fragrance applications. In Table 9.1, a compilation of recently published studies (mostly

Application of Solid-Phase Microextraction in Food and Fragrance Analysis

295

30

% Carryover

25 20 15 10 5 0 Ethyl heptanoate

Ethyl nonanoate

Ethyl undecanoate

Ethyl tridecanoate

Figure 9.3 Percentage carryover detected with CAR/PDMS fibre coating following HS
SPME extraction of spiked water samples.

within the last 5 years) in the field of food and fragrance analysis is presented. Two environmental applications are listed as well. It should be noted that Table 9.1 covers the most frequently examined food and fragrance matrices; however, this list is far from complete.

9.2.1

Summary of Recent Reviews in the Field of Food and Fragrance Chemistry

Probably the most recent general review of SPME in food analysis is the study by Wardencki et al.104 The main parameters affecting the extraction efficiency are discussed and exemplified by chromatograms. Many examples of SPME suitability for different food matrices are listed, and the review is very useful for readers who are new to the field of SPME or its use in food chemistry. Flavor, Fragrance and Odor Analysis, a book published by Marsili105 in 2002, summarises the recent sample preparation techniques for isolating and concentrating flavour and odour chemicals from various types of foods and beverages prior to the GC
MS analysis. This book is very informative and presents many practical suggestions, and special emphasis has been given to SPME technique. Kataoka et al.106 summarised the coupling of SPME to various analytical instruments in their comprehensive review on food analysis, and Buldini et al.107 compared various sample preparation techniques, including sorbent extraction techniques such as SPME, in food analysis. The review by Vas and Vekey108 summarises performance characteristics and types of applications of SPME in combination with MS. Food analysis is highlighted in this review, and several food applications, such as an automated in-tube extraction system, are illustrated and discussed in more detail.

Table 9.1 Experimental Conditions in the Recently Published Food and Fragrance Studies Involving HS
and DI
SPME Analysed Matrix

Cheese

Target Analytes

Aroma profile

Extraction Technique

a. HS
SPME

Extraction Procedure Detailsa

S: 10 g sample, INC: 45 C (1 h), EX: 45 C (1 h), DVB/CAR/ PDMS SF (2 cm), D: 260 C (10 min)

b. P&T

Separation and Detection

Compound

System

Identification

a. GC
qMS/EI-FID (parallel,

Spectral match with

1:1), DB-Wax column

LOD

Year of

Reference

Publication Not specified

2005

30

Not specified

2004

31

Standard solutions

Not specified

2005

32

Standard solutions

50
100 ppb

2002

33

library, RIsb

(60 m 3 0.32 mm i.d. 3 1 μm) b. GC
O, see above

S: 5 g sample, INC: 35 C (5 min), EX: DHS, 35 C (15 min), N2 (40 mL/min), Tenax trap (36 C), D: 230 C (4 min) into cryofocusing unit (2125 C) and 230 C (1.5 min) into the GC system

Cheese

Aroma profile (sulphur

HS
SPME

S: 7 g sample, EX: 22 C (16 h),

a. GC
itMS (EI, CI), BP21

compounds, pyrazines,

CAR/PDMS, D: 250 C (time

column, polyethylene glycol

furanones,

not specified)

(PEG) terephthalic acid treated

Spectral match with library, RIs

(30 m 3 0.32 mm i.d.

sesquiterpenes)

3 0.25 mm) b. GC
O
FID (parallel, 10:1), see above Cheese

Aroma profile

HS
SPME

S: 2 g sample 1 0.2 g vegetable oil, EX: CAR/

GC
PFPD, DB
FFAP column (30 m 3 0.32 mm i.d. 3 1 μm)

PDMS SF, 50 C (30 min), D: 300 C (10 min) Cheese

Mycophenolic acid (mycotoxin)

DI
SPME

S: 0.5 g 1 5 mL potassium

HPLC
UV/DAD, 5-μm

bicarbonate buffer (0.2M, pH

Supelcosil LC-NH2 column

9.7), sonicated for 30 min,

(250 3 2.1 mm i.d.), mobile

filtered, acidified to pH 3 with

phase: acetonitrile/methanol/

6N HCl, EX: CW/TPR-100,

ammonium acetate buffer

room temperature (30 min),

(50 mM, pH 7) mixture

D: soaking in acetonitrile/

(78:2:20, v/v/v), detection

ammonium acetate buffer

wavelength 254 nm

(50 mM, pH 7) mixture (80:20, v/v) for 60 s and exposing to the mobile phase for 20 s Cheese

Aroma profile

a. SPME (mode not specified)

S: 5 g sample, EX: 80 C (30
35 min), DVB/CAR/

GC
qMS/EI, SE-54 column (25 m 3 0.2 mm i.d.)

Spectral match with

Not specified

2003

34

2002

35

library

PDMS SF (2 cm), D: 280 C (time not specified) b. P&T

S:100
150 mg sample, EX: drying tube with anhydrous calcium chloride and tube with Carbotrap 300, 50 C (30 min), He (2 mL/min), D: 280 C (time not specified)

Cheese

Aroma profile of the surface-

a. Vacuum distillation

S: 100 g sample, EX: distillation under vacuum (1022 mbar),

ripened cheese

room temperature (2 h), distillate extracted (3 3 LLE: 10 mL dichloromethane, 20 min), desiccation, filtration

a. GC
FID, Supelcowax 10 column (30 m 3 0.32 mm

Spectral match with library, RIs

i.d. 3 1 μm) b. GC
qMS/EI, FFAP column (30 m 3 0.32 mm i.d. 3 1 μm) c. GC
O
FID (parallel, 1:1),

and concentration by

Supelcowax 10 column

distillation (40 C) to 200 μL,

(30 m 3 0.32 mm i.d. 3 1 μm)

INJ: 1 μL b. HS
SPME

S: 0.4 g of rind or 0.4 g of paste, INC: 25 C (1 h) EX: CAR/ PDMS 25 C (30 min), D: 260 C (5 min)

(Continued)

Table 9.1 (Continued) Analysed Matrix

Goat cheese

Target Analytes

Aroma profile

Extraction Technique

HS
SPME

Extraction Procedure Detailsa

S: 3 g sample, INC: 60 C (10 min), EX: DVB/CAR/ PDMS, 60 C (50 min), D: 250 C (5 min)

Cheese and curdled milk

Volatile free fatty acid profile influenced by

SPME (mode not specified)

coagulant and native

S: 10 g sample, INC: 45 C

Separation and Detection

Compound

System

Identification

LOD

Year of

Reference

a. GC
FID, RTX-1301 column

Standard solutions, RIs

Not specified

2005

36

Spectral match with

Not specified

2006

37

Standard solutions

0.5
1 mg TMA/kg milk

2002

38

Standard solutions

, 2
40 ng/mL

1999

39

Standard solutions

DEHP: 0.31
3.3 ng/g

2005

8

Publication

(30 m 3 0.25 mm i.d. 3 0.25 μm) b. GC
qMS/EI, see above GC
itMS/EI, FFAP CP-Wax 58

(5 min), E: DVB/CAR/PDMS,

column, nitroterephthalic acid

45 C (1 h), D: not specified

modified polyethylene glycol

library

(50 m 3 0.25 mm i.d.

microflora

3 0.39 mm) Milk

Trimethylamine (TMA)

a. P&T

S: 10 g sample, alkalised to pH 9 EX 1 D: DHS, other conditions not specified

b. HS
SPME

S: 10 g sample, alkalised to pH 9, EX 1 D: PDMS/DVB (SF), other conditions not specified

Milk

Tetracycline antibiotics

DI
SPME

S: 3.5 mL sample, saturated with KCl, EX: CW/TPR, CW/

GC
qMS/EI-FID (parallel, 1:1), SPB-1 sulphur (30 m 3 0.32 mm i.d. 3 4 μm) GC
O
NPD, DB-WAXetr column (60 m 3 0.32 mm i.d. 3 1 μm) HPLC
ESI
MS/MS (triple-q), PuroSphere column

DVB, 65 C (15 and 45 min),

(4.0 3 50 mm), stationary

D: soaking in acetonitrile/

phase: 3 μm RP-18e, initial

water (15:85, v/v, 5 min,

mobile phase:

40 C) in desorption chamber,

acetonitrile 1 0.2% formic

exposed to LC mobile phase

acid and water 1 0.2% formic acid, concentration gradient

Milk

Phthalate esters

HS
SPME

S: 5 g sample 1 2.5 g NaCl, INC:

GC
qMS/EI, DB-5 column

90 C (2 min), EX: PDMS

(30 m 3 0.25 mm i.d.

100 μm, 90 C (60 min), D:

3 0.25 μm)

280 C (10 min)

milk (0
10.8% fat)

Milk

Triazine herbicides

DI
SPME (HFM protected)

S: 5 mL sample, NaCl (30%) (w/v), pH 10

GC
qMS/EI, DB-5 column

Standard solutions

0.003
0.013 g/L

2004

40

Standard solutions

0.8
29.0 μg/L

2005

41

Standard solutions,

Not specified

2003

42

3
12 pg (SPME-

2005

43

(30 m 3 0.32 mm i.d. 3 0.25 μm)

EX: PDMS/DVB, 80 C (40 min), D: 280 C (5 min) Milk

Aroma profile (C3
11,C13

HS
SPME

S: 7 mL sample, INC: 40 C

GC
FID, BPX-5 column

methyl ketones, C3
10

(15 min), EX: DVB/CAR/

(50 m 3 0.22 mm i.d.

saturated aldehydes) in

PDMS (2 cm), 40 C (15 min),

3 0.25 μm)

UHT-processed stale

D: 240 C (2 min)

milk Butter

Aroma profile

a. P&T

S: 20 g sample, INC: 45 C

GC
qMS/EI, HP-Innowax

(5 min), EX: DHS, N2

column (60 m 3 0.32 mm

spectral match with

(30 mL/min), Tenax trap

i.d. 3 0.5 μm)

library

(30 min), N2 (5 min) to remove moisture D: 250 C (10 min) into cryofocusing unit (2120 C) and 250 C (1 min) into the GC system b. HS
SPME

S: 10 g sample, INC: 45 C (5 min), EX: DVB/CAR/ PDMS (2 cm), 45 C (30 min), D: 270 C (3 min)

Butter

Aroma profile

a. SPE

S: aqueous solution of butter

GC
qMS/EI, BP-21

Standard solutions,

(volume not specified), EX:

polyethyleneglycol TPA-

spectral match with

GCxGC
FID),

SDB-1, PS-DVB copolymer

treated column

library

25
50

cartridge, washing with 1 mL

(30 m 3 0.25 mm i.d.

water, drying (room

3 0.25 μm),

temperature, 15 min), D: 1 mL

GCxGC
TOFMS, 50 Hz,

methyl acetate, INJ: 1 μL

1-D: BP-21 column

(SPME
GC
MS)

(Continued)

Table 9.1 (Continued) Analysed Matrix

Target Analytes

Extraction Technique

b. HS
SPME

Extraction Procedure Detailsa

S: 8 g sample 1 30% (w/w) NaCl,

Separation and Detection

Compound

System

Identification

LOD

Year of

Reference

Publication

(30 m 3 0.25 mm i.d.

INC: 170 C (5 min), EX:

3 0.25 μm), 2-D: BPX-35

40 C, combination of CAR/

column (1 m 3 0.1 mm

PDMS (20 min) and CW/DVB

i.d. 3 0.1 μm), CO2-cooled

(60 min), D: 250 C (1 min)

dual-jet modulation, 3 s, 100 C below the oven temperature, GCxGC
FID, 200 Hz, columns and modulation: see above

Iodised salt, milk

Iodine

a. HS
SPME

S: 0.5
2 mL aliquot of sample

GC
MS/EI, HP-5 column

powder and

solution 1 200 μL phosphate

(30 m 3 0.25 mm i.d.

vegetables

buffer 1 250 μL N,N-

3 0.25 μm)

Standard solutions

25 ng/L iodine

Standard solutions

10 ng/L iodine

dimethylaniline (derivatisation agent), 400 μL 2-iodosobenzoate 1 2 μL IS 1 4 mL water, INC: 26 C (1 min), EX: PDMS 100 μm, 26 C (15 min), D: 250 C (5 min) (milk powder and vegetables: clean-up prior to SPME: SPE on C18 cartridge, eluted with methanol and water) b. SDME

S: see above, EX: 1 μL drop, 26 C (15 min), INJ: into injector (250 C), (milk powder and vegetables: cleanup prior to SDME: SPE on C18 cartridge, eluted with methanol and water)

See above

2004

44

Sausages

Volatile nitrosamines

HS
SPME

S: 2.5 g sample, INC: 45 C (10 min), EX: PDMS/DVB,

GC-TEA, HP-INNOWAX column

Standard solutions

3 μg/kg

2005

45

Standard solutions,

Not specified

2004

46

Standard solutions

0.011
0.357 μg/L

2006

47

Standard solutions,

Not specified

2006

48

28 ng/g

2006

49

(30 m 3 0.53 mm i.d. 3 1 μm)

45 C (25 min), D: 200 C (8 min) Dry-fermented

Aroma profile

HS
SPME

sausages

S: 3 g sample, INC: 30 C (1 h),

GC
qMS/EI, DB-624 column

EX: DVB/CAR/PDMS, 30 C

(30 m 3 0.25 mm i.d.

spectral match with

(90 min) or CAR/PDMS,

3 1.4 μm)

library, RIs

30 C (3 h), D: 220 C (6 min) Model meat

Volatile nitrosamines

On-site DED
SPME

S: 68% chicken meat, 30% water,

GC
qMS/EI, HP-5 column

(DED immersed in

3% NaCl, homogenisation,

(50 m 3 0.32 mm i.d.

sample, “HS”

gelling (70 C, 15 min), INC:

3 1.05 μm)

extraction)

25 C (15 min), E: SPME fibre introduced into DED, CAR/ PDMS, 25 C (60 min), D: 270 C for the whole GC run

Processed meat

Aroma profile

HS
SPME

S: 5 g sample, INC: room

GC
qMS/EI, Rtx-Wax column

products (sausages,

temperature (1 h), EX: CAR/

(30 m 3 0.25 mm i.d.

spectral match with

mortadella, cooked

PDMS (performed best), room

3 0.25 μm)

library, RIs

ham)

temperature (90 min), D: 220 C (5 min)

Tuna fish

Methylmercury

HS
SPME

S: 0.4 g sample 1 15 mL saturated

GC
qMS/EI, HP-5MS column

NaCl solution 1 100 μL HCl,

(30 m 3 0.25 mm i.d.

3 mL for extraction (pH

3 0.25 μm)

Stable isotope dilution analysis (SIDA)

adjusted to 5.3) 1 1 mL sodium tetraethyl borate, EX: DVB/CAR/PDMS, room temperature (15 min), D: 260 C (1 min)

(Continued)

Table 9.1 (Continued) Analysed Matrix

Fish (sea water and

Target Analytes

Formaldehyde

Extraction Technique

HS
SPME

Extraction Procedure Detailsa

Derivatisation: fibre exposed to

Separation and Detection

Compound

System

Identification

GC
qMS/EI, HP-5MS column

Standard solutions

17 μg/kg

2007

50

Standard solutions

Not specified

2001

51

Standard solutions

Not specified

2004

52

2 mL PFBHA solution

(30 m 3 0.25 mm i.d.

shrimps and

(20 mg/L), 60 C (10 min),

3 0.25 μm)

cuttlefish)

S: 6 g cooked sample 1 6 mL

freshwater,

LOD

Year of

Reference

Publication

water, EX: CAR/PDMS, 80 C (30 min), D: 310 C (3 min) Brown and milled rice

2-Acetyl-1-pyrroline

a. HS
SPME

S: 0.75 g sample 1 100 μL water,

GC
MS, DB-5 column

INC: 80 C (25 min), EX:

(30 m 3 0.25 μm df, i.d.

DVB/CAR/DPMS SF, 80 C

not specified)

(15 min), D: 270 C (5 min) b. Solvent extraction

S: 0.3 g sample placed in a vial with IS in methyl chloride

GC
FID, other conditions not specified

(solvent for extraction), E: 85 C (2.5 h), INJ: 2 μL Jasmin rice

2-Acetyl-1-pyrroline

a. Solvent extraction

S: 5 g sample with 50 mL 2,4,6-

GC
FID, HP-5MS column

trimethylpyridine (IS)

(30 m 3 0.25 mm i.d.

solution, filtration, NaOH

3 0.25 μm)

addition, E: 50 mL dichloromethane (2x), concentration (RVO, 28 C), INJ: μL (250 C) Aroma profile

b. HS
SPME

S: 8 g sample powder, IS addition

GC
qMS/EI, HP-1MS column (30 m 3 0.25 mm i.d. 3 0.25 μm)

INC: room temperature (15 min), then 80 C (30 min), EX: PDMS 100 μm: 80 C (30 min), D: 250 C (1 min)

Not specified

Pathogens in potatoes

Volatile compounds

HS
SPME

a. S: individual tuber in a jar,

a. GC
qMS/EI, Chrompack

identifying pathogens in

INC: room temperature

CPSil-5CB column

potato tubers caused by

(30 min), EX: CAR/PDMS

(50 m 3 0.32 mm i.d.

Ralstonia solanacearum

room temperature (30 min),

(brown rot) and

D: 270 C (1 min)

Clavibacter

Spectral match with

Not specified

2006

53

Not specified

2005

54

Not specified

2007

55

library

3 1.2 μm) b. GC
itMS, CPSil-5CB column (42 m 3 0.32 mm i.d.

michiganensis

3 1.2 μm)

sepadonicus (ring rot)

b. S: Individual tuber in a jar,

e-nose: 8-metal oxide sensor array

INC: room temperature

system (data from 25 to 110 s

(30 min), EX: CAR/PDMS

of fibre exposure averaged)

room temperature (5 min), D: fibre exposure to heated sensor array (2 min) Potato crisps

Aroma profile for rancidity

DHS

tests

S: 60 g sample, INC: 50 or 70 C,

a. 12-metal GS e-nose




b. (GC)
qMS/EI e-nose, 5 m

Spectral match with

respectively, 30 min, EX: DHS (10 min), air flow (150 mL/min) HS
SPME

S: 4 g sample, INC: 50 C, time not specified, EX: CAR/

deactivated fused silica column

PDMS, 50 C (20 min),

(250 C)

D: ad (b) 300 C (3 min), ad (c) 250 C (5 min)

library

c. GC
qMS/EI, Equity-5 poly column (30 m 3 0.25 mm i.d. 3 0.25 μm)

Thawed and cooked French beans

Volatile and semi-volatile compounds

a. HS
SPME

S: 10 g sample, INC: 40 C

GC
FID
qMS/EI, two HP-1

(30 min), EX: DVB/CAR/

columns (50 m 3 0.2 mm

PDMS, 40 C (1 h), D: 250 C

i.d. 3 0.33 μm)

Spectral match with library, RIs

(2 min)

(Continued)

Table 9.1 (Continued) Analysed Matrix

Target Analytes

Extraction Technique

b. SDE

Extraction Procedure Detailsa

Separation and Detection

Compound

System

Identification

a. GC
qMS/EI, SPB-1 column

Spectral and RI match

LOD

Year of

Reference

Publication

S: 300 g sample 1 1L water 1 70 mL dichloromethane in the Likens
Nickerson apparatus, EX: boiled (2 h), extract dried over MgSO4, concentration to 250 μL, injection

Mango

Unsaturated fatty acid esters

a. HS
SPME

S: 1 mango, intact INC: room temperature (30 min),

and other volatile

EX: DVB/CAR/PDMS (2 cm),

compounds

b. Solvent extraction

(30 m 3 0.25 mm i.d. 3 1 μm) b. GC
qMS/EI, Supelcowax

room temperature (60 min),

column (30 m 3 0.25 mm

D: 250 C (0.5 min)

i.d. 3 0.25 μm)

Not specified

2006

56

Not specified

2005

57

with the previously analysed (reference) samples

S: Mango pulp (1.2) 1 700 mL

(Kutscher-Steudel

water, EX: continuous

extraction)

extraction in Kutscher-Steudel extractor with 500 mL diethyl ether (5 h), ether phase dried over MgSO4, concentration to 60 mg of solid extract

Strawberries

Aroma profile

HS
SPME

Standard solutions,

a. Whole fruit: S: 4 pieces, INC:

GCxGC
FID, LMCS modulation

room temperature (2 h), EX:

(mod. period 4 s), 1-D column

chromatogram

PDMS 100 μm, room

(2 chiral directly coupled

“fingerprint”

temperature (45 min),

columns): EtTBS-CD

comparison

D: 250 C (5 min)

(20 m 3 0.25 mm i.d.

b. Puree´d fruit: S: 4 puree´d pieces, 5 mL puree´ for

3 0.25 μm) 1 CycloSil B (26 m 3 0.25 mm

analysis 1 15 μL tridecane/

i.d. 3 0.25 μm) 1 short piece

ethanol mix, EX: PDMS

of BPX-5 (0.14 m 3 0.25 mm

100 μm, room temperature

i.d. 3 0.25 μm), 2-D column:

(15 min), D: see above

BPX-50 (1 m 3 0.1 mm i.d. 3 0.1 μm), FID: 100 Hz

Strawberries

Aroma profile in strawberries

HS
SPME

S: Fruits injured by shaking, EX:

affected by Botrytis

dynamic HS sampling from

cinerea following fruit

air (air flow 60 mL/min),

wounding

PDMS 100 μm, 15 min

GC
FID, DB-5 column (60 m 3 0.32 mm i.d. 3 1 μm)

Standard solutions,

Not specified

2003

58

Not specified

2004

59

Not specified

2006

60

spectral match with library, RIs

sampling cycles (0
195 min), D: 240 C (time not specified) Polymer trap procedure

S: Fruits injured by shaking, EX:

GC
qMS/EI, DB-5 column

headspace sampling (air flow,

(25 m 3 0.25 mm i.d., df not

60 mL/min), 50 mg of Super

specified)

Q polymer in a glass trap (3 h), D: eluted with 300 μL hexane, injected (injector: 250 C) Avocado

Aroma profile

HS
SPME

S: 20 g sample puree´, heated in

GC
qMS/EI, HP-FFAP column

Standard solutions,

microwave oven (2,450 MHz,

(30 m 3 0.25 mm i.d.

spectral match with

633 W, 30 s), 5 g of puree´

3 0.25 μm)

library, RIs

taken for analysis (or 0.2 g of leaves), INC: room temperature (24 h), EX: CAR/ PDMS, (30 min), D: 180 C, time not specified Tropical fruits (yellow passion fruit,

Aroma profile

a. HS
SPME b. CF
HS
SPME

S: 0.55 g sample, INC: 60 C (10 min), EX: 60 C (25 min)

a. GC
FID, CP-Sil 8 column (30 m 3 0.25 mm i.d.

Spectral match with library, RIs

3 0.25 μm)

(Continued)

Table 9.1 (Continued) Analysed Matrix

Target Analytes

Extraction Technique

Extraction Procedure Detailsa

a. DVB/CAR/PDMS fibre,

cashew, tamarind,

25 min

acerola and guava)

b. Internally cooled PDMS fibre,

Separation and Detection

Compound

System

Identification

LOD

Year of

Reference

Publication

b. GC
qMS/EI, VF-5MS column (30 m 3 0.25 mm i.d. 3 0.25 μm), cryo-trap system

thickness of the coating: 340 μm, length: 1 cm, fibre temperature 0 C, D: 250 C (3 min) Unfermented olives

Aroma profile

HS
SPME

S: 5 mL olive brines, EX: PDMS/

GC
qMS/EI, HP-5MS column

Not specified

2004

61

Standard solutions

0.006
0.010 mg/kg

2006

62

Standard solutions,

Not specified

2006

63

25.7 ng/kg (carrots)

2006

64

Standard solutions,

subjected to

DVB, room temperature

(30 m 3 0.25 mm i.d.

spectral match with

pasteurisation and

(60 min), D: 250 C (5 min)

3 0.25 μm)

library, RI

sterilisation Olive oil

Organophosphorus

HS
SPME

insecticides

S: 5 g sample, INC: 75 C

GC
FTD, DB-1 column

(10 min), EX: PDMS 100 μm,

(30 m 3 0.32 mm i.d., df not

75 C (60 min), D: 250 C

specified)

(7 min) Olive oil

Aroma profile in thermally

DHS

oxidised olive oil

S: 1 g oxidised oil sample, INC: placed in stripping gas (He, 200 mL/min) receptacle, 37 C (5 min), EX: Tenax trap, room

a. GC
FID, SPB-5 column (30 m 3 0.32 mm i.d., df not

spectral match with

specified)

library

b. GC
itMS, see above

temperature (15 min), D: heated desorption unit (300 C, 8 min) HS
SPME

S: 3.5 mL oxidised oil, INC: 23 C

GC
TOFMS/EI, cryo-trap, HP-5

(2 h), EX: PDMS/DVB,

column (5 m 3 0.1 mm i.d.

PDMS 100 μm, 23 C

3 0.34 mm), 40 spectra s21

(30 min), D: 250 C (2 min) Baby food (carrots,

Furan

HS
SPME

S: 4 g boiled sample,

GC
qMS/EI, HP-INNOWAX

Standard solutions,

homogenised, EX: CAR/

column (60 m 3 0.25 mm

spectral match with

banana, apple,

PDMS, 30 C (10 min),

i.d. 3 0.5 μm)

library

pear, potato, kiwi,

D: 230 C (3 min)

beans, zucchini,

green peas)

Breast milk

Chlorinated organic

HS
SPME

S: 0.5 mL sample 1 0.5 mL

GC
ECD, SE-54 column

compounds (PCBs, HCH,

perchloric acid (1M) 1 0.15 g

(50 m 3 0.32 mm i.d.

HCB, DDT and

Na2SO4, EX: PA, 100 C

3 0.35 μm)

derivatives,

(40 min), D: 280 C (10 min)

Standard solutions

0.06
3.41 μg/L

2000

65

Spectral match with

Not specified

2004

66

Not specified

2007

67

chlorophenols) Honey

Aroma profile

a. SHS

S: 7 g sample 1 1 mL water 1 1.05 g NaCl, INC:

Smart Nose-qMS/EI, 3 scanning cycles (10
160 amu)

library

100 C (15 min), EX : 2.5 mL of headspace, INJ: injected into SMart Nose injector (120 C) b. HS
SPME

S: 7 g sample 1 1 mL water 1 1.05 g NaCl, INC: 90 C (2 min), EX: DVB/ CAR/PDMS (SF), 90 C (30 min), D: 190 C (time not specified)

c. Inside-needle

INC: 90 C (5 min), EX: DVB/

dynamic extraction

CAR/PDMS polymer bed

(INDEX)

(1:1:1), 2.5 mL of headspace (10x aspiration/ejection cycle), INJ: injected into Smart Nose injector (200 C, 10x aspiration/ejection)

Unifloral honey

Aroma profile

HS
SPME

S: 6 mL honey
water solution

GC
qMS/EI, HP-5MS column

Standard solutions,

(3 g/mL), INC: 60 C

(30 m 3 0.25 mm i.d.

spectral match with

(30 min), EX: DVB/CAR/

3 0.25 μm)

library, RIs

PDMS, 60 C (60 min), D: 220 C, time not specified

(Continued)

Table 9.1 (Continued) Analysed Matrix

Matrix affected by trichotheceneproducing fungi

Target Analytes

Aroma profile markers

Extraction Technique

a. HS
SPME

produced by growth of toxigenic Fusarium spp. (mainly trichodiene and

b. HS
SBSE

Extraction Procedure Detailsa

S: Fungi cultivated in SPME

Separation and Detection

Compound

System

Identification

LOD

Year of

Reference

Publication Not specified

2004

68

Standard solutions

10 ng/L

2006

69

Standard solutions

0.5
10 μg/L

2006

70

Standard solutions

0.6 μg/L

2006

71

GC
qMS/EI, HP-5 column

Spectral match with

vials, EX: PDMS 100 μm,

(30 m 3 0.25 mm i.d.

library, RIs

25 C (30 min), D: 250 C

3 0.25 μm), PTV injector

EX: PDMS stir bar, 24 μL stationary phase, 25 C

deoxynivalenol)

(30 min), injector temperature 250 C, D: stir bar in a glass tube, 20 C to 250 C (60 C/ min, 250 C held for 7 min) Drinking water

Methyl tert-butyl ether

HS
SPME

(MTBE)

S: 4 mL sample 1 NaCl content

GC
MS, DB-624 column

of 10%, EX: CAR/PDMS,

(60 m 3 0.32 mm i.d.

18
19 C (30 min), D: 260 C

3 1.8 μm)

(10 min) Drinking water

Organic micropollutants

a. LLE

S: 80 mL sample, EX: 1 3 4 mL

migrating from

and 2 3 2 mL

polyethylene into water

dichloromethane, concentration to 0.5 mL, INJ: 2 μL injected b. DI
SPME

a. GC
FID (for standards), Optima 17 column (15 m 3 0.53 mm i.d. 3 1 μm) b. GC
qMS (for real samples), see above

S: 10 mL sample, EX: PDMS/ DVB, 60 C (30 min), D: 270 C (3 min)

Drinking water

Diethylhexylphthalate (DEHP)

DI
SPME

S: 500 mL sample, EX: PDMS/ DVB, 2
4 C

HPLC
UV, Luna Phenomenex C18 column (4.6 mm 3 30 mm 3 5 μm), mobile phase: acetonitrile, detection wavelength 224 nm

D: acetonitrile (5 min static desorption), 2 min dynamic elution to the column

Freshwater and drinking water

Saxitoxin (paralytic shellfish

DI
SPME

poisoning toxin)

Fibre incubation: soaking fibre in

HPLC
FLD, Beckman C18

0.1 M NaOH solution, S:

reversed phase column

5 mL sample, EX: CW/TPR,

(4.6 mm 3 150 mm i.d.

room temperature, pH 8.1 (40

3 5 μm), 30 C, mobile phase:

min), D: static ion-pairing

sodium 1-heptanesulfonate in

desorption (mixture of 20 mM

ammonium phosphate

sodium 1-heptanesulfonate in

(pH 7.1) and acetonitrile, post

30% aqueous acetonitrile

column fluorescent

acidified with 50 mM

derivatisation: reaction with

sulphuric acid)

periodic acid in potassium

Standard solutions

0.11 ng/mL

2005

72

Standard solutions

2.5
6 μg/L (water),

2005

73

2004

74

phosphate and with acetic acid at 85 C, FLD: 330 (excit.) and 390 nm (emis.) Water, orange juice,

Pesticides

apple juice

DI
SPME
RP
SWMR

1. SPME: S: 6 mL sample 1 1.8 g

CE-UV, 214 nm, CE capillary:

NaCl, pH adjusted to 6, EX:

50 cm detection length, 57 cm

PDMS/DVB, room

total length, 50 mm i.d.

3.1
47 μg/L (juice)

temperature (150 min), D: 200 μL methanol, stirring for 16 min, addition of 200 μL 0.4M acetic acid, injected into CE 2. Online pre-concentration: reversed polarity-stacking with matrix removal (RP
SWMR), 120 kV, field-enhanced sample injection (FESI) Apple juice

Off-flavour compounds

HS
SPME

S: Sample diluted to 10%, 2.5 g

caused by bacteria

Na2SO4, INC: 60 C (5 min for

(Alicyclobacillus

compounds produced by

acidoterrestris and

Actinomycetes and 10 min for

Actinomycetes)

by the other), EX: DVB/CAR/

GC
qMS/EI, HP-5 column (30 m 3 0.25 mm i.d. 3 1 μm)

Standard solutions,

0.08
7.7 μg/L

spectral match with library, RIs

(Continued)

Table 9.1 (Continued) Analysed Matrix

Target Analytes

Extraction Technique

Extraction Procedure Detailsa

Separation and Detection

Compound

System

Identification

GC
qMS/EI, CPSil-5CB column

Spectral match with

LOD

Year of

Reference

Publication

PDMS, 60 C (10 min and 30 min, respectively), D: 270 C (10 min) Strawberry juice and

Aroma profile

HS
SPME, DI
SPME

wine

S: 50 mL sample saturated with NaCl

(25 m 3 0.25 mm i.d.

Not specified

2006

75

0.2
1.8 μg/L

2004

76

0.015
0.081 ng/L

2003

77

library, RIs

3 0.4 μm) EX: PDMS 100 μm, 30 C (30 min), fibre in HS or immersed, D: 250 C (2 min 1 8 min fibre bake-out) Fruit juices and

Aroma profile

HS
SPME

S: 5 mL sample, EX: PDMS

nectars (pear,

100 μm: 40 C (30 min),

apricot, peach)

D: 250 C (5 min)

a. GC
FID, Supelcowa 3 10

Standard solutions,

column (30 m 3 0.25 mm i.d.

spectral match with

3 0.25 μm)

library, RI

b. GC
qMS/EI, see above Tea (Chinese)

Organochlorine pesticides

a. MAE
HS
SPME

1. MAE: S: 0.5 g sample 1 15 mL water,

GC
μECD, HP-5 column (30 m 3 0.25 mm i.d.

EX: 10 min, 80% power level,

3 0.25 μm), GC
qMS, HP-5

volume adjusted to 15 mL,

column (30 m 3 0.25 mm i.d.

SPME extraction followed

3 0.25 μm)

2. SPME: S: 15 mL sample 1 5 g NaCl, EX: sol
gel polyphenylmethylsiloxane (PPMS) fibre, 90 C (40 min), D: 280 C (4 min)

Standard solutions

b. USE
HS
SPME

Not specified

S: 0.5 g sample 1 15 mL water, EX: 1 h sonication, SPME extraction followed (see above)

Tea (green, black,

Catechins, caffeine

In-tube SPME

S: 1 mL sample, EX: 40 μL of

78

Not specified

2002

79

Not specified

2005

80

i.d., 5 μm particle size),

(catechins)

(60 cm 3 0.25 mm i.d.),

gradient elution 0.3 mL/min

ejected back to the vial

(acetonitrile
acetic acid,

(15 cycles) Solvent extraction

2000

0.01 (caffeine),

to the PPY-coated capillary

juice, wine

a. Free amino acids

Standard solutions

, 0.5 ng/mL

sample transferred from vial

Cocoa

HPLC
ES
MS, Supelcosil LC18 column (15 cm, 4.6 mm

herbal), grape

Roasting: 5 g cocoa mass or 5 g cocoa mass 1 50 mg D-glucose

water
acetic acid) Quantification by reaction with o-phthaldehyde

Reaction with ophthaldehyde

or 20.8 mg glycine or both,

(results as mg

roasted (150 C, 30 min),

glycine/g sample)

S: 2 g cocoa powder, 30 mL sodium citrate (20 g/L, pH 2.2), centrifugation, filtration, EX: 70 mL sodium citrate addition b. Pyrazine

HS
SPME

Roasting: see above, S: 1 g cocoa

GC
FID, PE-WAX column

powder 1 10 mL saturated

(30 m 3 0.53 mm i.d.

NaCl solution, INC: 15 min,

3 0.5 μm)

Standard solutions

CAR 100 μm, EX: 60 C (45 min), D: 240 C (4 min) Coffee

Aroma profile

HS
SPME

S: 2 g of roasted sample, INC: 60 C (10 min), EX: DVB/ CAR/PDMS, 60 C (40 min), D: 260 C (5 min 1 5 min fibre

a. GC
FID, Omegawax 250

Spectral match with

(30 m 3 0.25 mm i.d.

library (dual filter

3 0.25 μm)

MS spectra search),

b. GC
qMS, see above

RIs

bake-out)

(Continued)

Table 9.1 (Continued) Analysed Matrix

Coffee (Arabica and

Target Analytes

Aroma profile

Extraction Technique

HS
SPME

Robusta)

Extraction Procedure Detailsa

Separation and Detection

Compound

System

Identification

S: 1 coffee bean, cut in half, INC:

a. GCxGC
TOFMS/EI, LMCS

Spectral match with

60 C (10 min), EX: DVB/

modulation (modulation

library, comparison

CAR/PDMS, 60 C (40 min),

period 5 s, cryo-trap: 0 C),

with previous

D: 250 C (2 min)

data acquisition: 100 Hz;

GCxGC
FID

system A: 1-D column:

results

LOD

Year of

Reference

Publication Not specified

2004

81

Not specified

2006

82

SolGel-WAX (30 m 3 0.25 mm i.d. 3 0.25 μm), 2-D column: BPX-5 (1 m 3 0.1 mm i.d. 3 0.1 μm), system B: 1-D column: BPX-5 (30 m 3 0.25 mm i.d. 3 0.25 μm), 2-D column: BP20 (0.8 m 3 0.1 mm i.d. 3 0.1 μm) b. GCxGC
qMS/EI, LMCS modulation (modulation

Spectral match with library

period 5 s), data acquisition: 20 Hz; 1-D column: Supelcowax 10 (30 m 3 0.25 mm i.d. 3 0.25 μm), 2-D column: SPB-5 (1 m 3 0.1 mm i.d. 3 0.1 μm) Beer

Aroma profile

HS
SPME

S: 5 g sample 1 2 g NaCl, INC:

GC
qMS/EI, SPB-5 column

20 C (30 min),

(60 m 3 0.32 mm i.d.

ultrasonication, EX: CAR/

3 1.0 μm)

PDMS, 20 C (30 min), D: 290 C, 10 min

Spectral match with library

Wine

Aroma profile

HS
SPME

S: 10 mL diluted sample (1:10 or

GC
qMS/EI, ZB-Wax column

100, respectively) 1 2 g NaCl,

(60 m 3 0.25 mm i.d.

EX: CW/DVB, 35 C

3 0.25 μm)

Standard solutions,

Not specified

2005

83

LOQ: 0.2
0.3 ng/L

2006

84

2005

85

2000

86

SIDA

(10 min), D: 30 s 1 9.5 min fibre bake-out Wine

2,4,6-trichloroanisole, 2,4,6-

HS
SPME

tribromoanisole

S: 3 mL wine saturated with

a. GC
HRMS/NCI (methane),

Standard solutions

NaCl, EX: PDMS 100 μm,

Equity-5 column

(GC
LRMS),

21 C (30 min), D: 250 C

(30 m 3 0.25 mm i.d.

0.03 ng/L

(5 min)

3 0.25 μm)

(GC
HRMS)

b. GC
HRMS/EI, see above Wine

2-methoxy-3-(2-

HS
SPME

S: 10 mL sample (diluted to 12%

a. GCxGC
NPD, LMCS

methylpropyl) pyrazine

ethanol) 1 d3-IBMP (29.7 ng/

modulation (modulation period

(IBMP)

L wine)

4 s, data acquisition: 100 Hz),

Standard solutions, SIDA (MS analysis)

0.5 ng/L (NPD), 1.95 ng/L (MS)

1-D column: BPX-5 (30 m 3 0.25 mm i.d. 3 0.25 μm), 2-D column: BP20 (1 m 3 0.1 mm i.d. 3 0.1 μm) b. GCxGC
TOFMS/EI, see above EX: PDMS/DVB, 33 C (83 min), D: system (a): 50
250 C at 10 K/s (3 min), system (b) 250 C Alcoholic beverages

Volatile carbonyl compounds (C1-C6)

LLE

S: Derivatisation: alcoholic beverage sample dilution to 20% ethanol, 2 mL sample,

GC
ECD, Rtx-5 column (30 m 3 0.32 mm i.d. 3 3 μm)

Standard solutions

0.23
3.3 μg/L (LLE), 0.005
0.33 μg/L (SPME)

derivatisation with PFBHA, reaction and INC: 45 C (2 h), EX: 1 mL n-heptane 1 8 mL 0.05M H2SO4, shaken for 30 s, INJ: 1 μL injected

(Continued)

Table 9.1 (Continued) Analysed Matrix

Target Analytes

Extraction Technique

DI
SPME

Extraction Procedure Detailsa

Separation and Detection

Compound

System

Identification

LOD

Year of

Reference

S: 0.5 mL sample, filtration 1 2 g

GC
MS/MS/EI (triple-q),

Standard solutions

0.03 mg/L

2006

87

NaCl 1 4 mL pH 7 buffer

Stabilwax column

solution, INC: 70 C (10 min),

(60 m 3 0.25 mm i.d.

EX: CW/DVB, 70 C

3 0.25 μm)

Spectral match with

Not specified

2006

88

Not specified

2002

89

Publication

S: Derivatisation: 10 mL, pH adjusted to 2, addition of 1 mL PFBHA, reaction and INC: 45 C (2 h), EX: PDMS 100 μm, room temperature (15 min), D: 250 C (10 min)

Stone-fruit spirits

Ethyl carbamate

HS
SPME

(30 min), D: 250 C (2 min 1 8 min fibre bake-out) Sweetened orange spirit

Aroma profile changes

HS
SPME

S: 5 mL sample, INC: 37 C

following after addition

(10 min), EX: DVB/CAR/

of carbohydrates

PDMS, 37 C (5 min), D: 250 C (2 min)

a. GC
FID, SupelcoWax 10 column (30 m 3 0.32 mm

library

i.d. 3 0.3 μm) b. GC
itMS/EI, EC-Wax column (30 m 3 0.25 mm i.d. 3 0.25 μm)

Pepper

Aroma profile of essential oil

HS
SPME

S: Hydrodistillation of seeds

a. GC
FID, RSL 200 column

Standard solutions,

(30 m 3 0.32 mm i.d.

spectral match with

placed in SPME vial, EX:

3 0.25 μm), HP-5MS column

library, RI

DVB/CAR/DPMS SF (2 cm),

(30 m 3 0.32 mm i.d.

(20 g), essential oil dried,

room temperature (4 h),

3 0.25 μm), Stabilwax column

D: 250 C (5 min)

(30 m 3 0.32 mm i.d. 3 0.50 μm) b. GC
O/FID (parallel, 50:1), RSL 200 column (30 m 3 0.32 mm i.d. 3 0.25 μm)

c. GC
qMS/EI, RSL 200 and Stabilwax columns (for dimensions, see above) Vinegar

2-Furfural, 5-methylfurfural

HS
SPME

S: 8 g sample 1 NaCl to obtain

GC
qMS/EI, DB-WAX column

40% (w/v) solution, IS

(60 m 3 0.25 mm i.d.

addition

3 0.25 μm)

15 μg/L

2003

90

0.03
8.60 μg/L

2006

91

RIs

Not specified

2003

92

Standard solutions

0.02
0.26 μg

2003

93

Standard solutions

0.02
0.2 ppb

2006

94

Standard solutions

Not specified

Standard solutions, SIDA

EX: DVB/CAR/PDMS, 50 C (40 min), D: 280 C (10 min) Vinegar

Aroma profile

a. SBSE

S: 25 mL sample 1 5.85 g

GC
qMS/EI, DB-Wax column

NaCl 1 50 μL 4-methyl-2-

(60 m 3 0.25 mm i.d.

pentanol solution, EX: PDMS

3 0.25 μm)

Spectral match with library, RIs

stir bars, 25 C, 1250 rpm (120 min), D: 330 C (10 min)

Cotton, cigarettes,

Smoke volatile components

b. HS
SPME

Conditions not specified

HS
SPME

S: 9 cm3 of material in the tube,

paper and grass

to distinguish the

ignited, EX: 4, 6, 8 and 10 s

smoke

combustion source

for cotton, cigarettes, paper

GC
PFAIMS and GC
qMS/EI, SP 2300 column (23 m)

and grass, resp., D: 30 s, temperature not specified Indian-made bidi cigarettes

Flavour-related toxic

HS
SPME

compounds

S: 0.1 g sample 1 300 μL 3M KCl solution, INC: room

GC
qMS/EI, DB-5MS column (30 m 3 0.1 mm i.d. 3 0.1 μm)

temperature (1 h), then 95 C (5 min), EX: CW/DVB, 95 C (2 min), D: 230 C (4 min) Shampoo

Perfume compounds

HS
SPME

Equilibrium extraction: S: 50 μL shampoo, diluted to 0.01% in water, 3 g salt addition, INC: 45 min, temperature not specified EX: 45 C (45 min), PA, D: 270 C

a. GC
FID, SPB-1 column (30 m 3 0.25 mm i.d. 3 0.25 μm) b. GC
FID, SPB-1 column (15 m 3 0.25 mm i.d. 3 0.25 μm)

(1
2 min) Exhaustive extraction:

GC
qMS/EI, SPB-1 column (30 m 3 0.32 mm i.d. 3 1 μm)

(Continued)

Table 9.1 (Continued) Analysed Matrix

Target Analytes

Extraction Technique

Extraction Procedure Detailsa

Separation and Detection

Compound

System

Identification

GC
itMS/EI, SPB-5 column

Standard solutions,

LOD

Year of

Reference

Publication

S: 50 μL shampoo, diluted to 1% in water, INC: 60 C (45 min), EX: PA, 45 C (45 min), D: 270 C (1
2 min) Lavandula flower

Aroma profile

a. Solid-phase trapping

S: 20 g sample (dried leaves,

(L. stochas,

solvent extraction

flowers or buds), EX: Pasteur

L. dentate,

(SPTE)

(60 m 3 0.25 mm i.d.

spectral match with library, RIs

pipette packed with 500 mg

3 0.25 μm) and

L. angustifolia and

ethylvinylbenzene


Supelcowax-10 column

L. heterophylla)

divinylbenzene copolymer

(60 m 3 0.25 mm i.d.

adsorbent, sample in glass

3 0.25 μm)

syringe, nitrogen flow through syringe and trap (400 mL/ min), room temperature (3 h), D: elution with 2 mL petroleum ether, INJ: 2 μL b. Reduced pressure

S: 300 g sample (see above), EX:

steam distillation

distillation, 100 C (3 h),

(RPSD)

100 mm Hg pressure, distillate

See above

extracted with 5 mL petroleum ether, extract dried, INJ: 2 μL c. Simultaneous steam

S: 10 g sample (see above) 1

distillation-solvent

100 mL water, EX: 10 mL

extraction (SDE)

petroleum ether in

See above

Likens
Nickerson micro-SDE apparatus, 2 h extraction, extract dried, INJ: 2 μL d. SHS

S: 7 g sample (see above), EX: 40 C (60 min), INJ: aliquot (2 μL) of the vapour in headspace injected to GC

See above

Not specified

2002

95

e. HS
SPME

S: 1 g sample (see above), EX: PDMS 100 μm, 40 C (60 min), D: 250 C (60 s)

a. See above b. GC
FID (for determination of partition coefficients), SPB-5 column (60 m 3 0.25 mm i.d. 3 0.25 μm)

Syringa oblata flower

Aroma profile

HS
SPME

S: Fresh flowering parts (white

GC
itMS/both EI and CI modes

Not specified

2006

96

Not specified

2006

97

Not specified

2006

98

Standard solutions

Not specified

2005

99

Standard solutions,

2 ng/L (MeSA standard

2004

100

Some standard

and purple), amount not

(liquid CH3CN CI reagent),

solutions, spectral

specified, INC: applied, but

CP-Sil8 CB column

match with library,

not specified, EX: PDMS/

(30 m 3 0.25 mm i.d.

RIs MW confirmed

DVB, 25 C (30 min), D:

3 0.25 μm)

by CI

250 C (5 min) Brunfelsia australis

Aroma profile

HS
SPME

S: 5 flowers (both young purple

GC
qMS/EI, HP-5 column

Standard solutions,

and Brunfelsia

and mature white flowers),

(60 m 3 0.25 mm i.d.

spectral match with

pauciflora flowers

EX: PDMS 100 μm, 25 C

3 0.25 μm)

library, RIs

(45 min), D: 250 C (5 min) Robinia pseudoacacia

Aroma profile

HS
SPME

L. flower

S: 3 g fresh flower, EX: PDMS

GC
itMS/EI, VF-5MS column

Some standard solution,

100 μm, room temperature

(30 m 3 0.25 mm i.d.

spectral match with

(45 min or 2 h), D: 250 C

3 0.25 μm)

library, RIs

(4 min) Douglas-fir, rosemary

Biogenic enantiomeric and

Dynamic HS
SPME

S: Plant branch closed in plastic

GC
qMS/EI, β-cyclodextrin

and lavender trees

non-enantiomeric

(portable dynamic

cuvette, EX: portable pump

column CYCLODEX-B

and plants

monoterpenes

air sampler, PDAS)

(air velocity 70 cm/s), PDMS/

(30 m 3 0.256 mm i.d.

DVB, 0
10 C (1 min), D:

3 0.25 μm)

250 C (5 min) Tomato plant

Methyl salicylate (MeSA), salicylic acid (defence

HS
SPME

S: 4 tomato plants in 5 L glass bottle

response to Tobacco

GC
qMS/EI, HP-5MS column (30 m 3 0.25 mm i.d.

spectral match with

3 0.25 μm)

library

in ethanol)

Mosaic Virus, TMV) EX: PDMS 100 μm, 25 C (15 min), D: 270 C (2 min)

(Continued)

Table 9.1 (Continued) Analysed Matrix

Tomato plant

Target Analytes

C6 aldehydes (defence

Extraction Technique

HS
SPME

Extraction Procedure Detailsa

On-fibre derivatisation: PFBHA

Separation and Detection

Compound

System

Identification

GC
qMS/EI, HP-5MS column

Standard solutions (gas)

response to Helicoverpa

(17 mg/mL), headspace

(30 m 3 0.25 mm i.d.

armigera)

adsorption, 25 C (5 min), S: 1

3 0.25 μm)

LOD

Year of

Reference

Publication 0.1
0.5 ng/L (in

2005

101

Not specified

2003

69

1 ng (when using D2

2005

102

standard gas)

tomato plant in 5 L glass bottle, EX: PDMS/DVB, 25 C, 6 min, D: 270 C (2 min) Rosemary, grape wine,

Aroma profile

orchid, lathyrus

HS
SPME (compared

S: Single flower head or 0.4 g

GC
qMS/EI, HP-5MS column

Some standard

to the charcoal

dried rosemary leaves or

(30 m 3 0.25 mm i.d.

solutions, spectral

extraction)

30 mL grape wine, EX: CAR/

3 0.25 μm)

match with library

PDMS SF, room temperature (1 h), D: 250 C (5 min) Wasp Trichogramma

Putative sex pheromones

HS
SPME, on-fibre

S: 50
110 wasps in a vial

a. Preliminary studies: GC
qMS/

turkestanica

produced by virgin

derivatisation

(including drop of honey),

EI, DB-5 column

(environmental

females of the minute

(silylation)

EX: PDMS 100 μm, 25 C

(30 m 3 0.25 mm i.d.

application)

parasitoid wasp

(20
50 h), post-extraction

Trichogramma

derivatisation: headspace, vial

3 0.25 μm) b. Model experiments with ginger

turkestanica (compounds

with 5 μL BSTFA with 1%

oil: GC
qMS/EI, DB-1

structures: C17H32 and

TMCS, D1: 250 C

column (20 m 3 0.18 mm i.d.

C17H32O)

(5
15 min), D2: solvent desorption of extracted

3 0.4 μm) c. Accurate mass measurements:

compounds into acetone

GC
msHRMS/EI, PB-5

(15 min), addition of MTAD

column (30 m 3 0.25 mm i.d.

in CH2Cl2, left to evaporate to a volume of 4 μL, injected

3 0.25 μm) d. Retention indices determination: 2 3 GC
FID,

RIs

desorption)

DB-1 column (60 m 3 0.25 mm i. d. 3 0.25 μm), Stabilwax column (60 m 3 0.25 mm i.d. 3 0.25 μm) Freshwater (environmental application)

Anatoxin-a

USE
DI
SPME

In situ derivatisation:

GC
qMS/EI, HP-5MS column

hexylchloroformate to the

(30 m 3 0.25 mm

alkalinised sample (pH 9)

i.d. 3 0.10 μm)

Calibration curve

2 ng/mL

2006

103

S: 2 mL water, 200 μg sodium bicarbonate, EX: (1) USE (10 min), (2) vial incubated at room temperature overnight, (3) SPME: PDMS 100 μm (20 min), D: 250 C (20 min)

List of abbreviations: 1D, 2D…one-dimensional, two-dimensional, ANN. . .artificial neural networks, ANOVA. . .analysis of variance, BSTFA…N,O-bis(trimethylsilyl)trifluoroacetamide, BTEX…a mixture of benzene, toluene, ethylbenzene and xylenes, c0. . .initial concentration of analyte in the sample, CAR…carboxen coating, CCD. . .central composite design, CF. . .cold fiber (internally cooled fiber), CRM…certified reference material, CW…carbowax coating, D. . .desorption of the target analytes into the system, DAD…photodiode array detector, DAI…direct aqueous injection, DDT…dichlorodiphenyltrichloroethane, DED. . .direct extraction device, DEHP. . .diethylhexyl phthalate, DFA. . .discriminant factor analysis, DHS…dynamic headspace extraction, DI. . .direct immersion, dtto…the same as above, DVB…divinylbenzene, ECD. . .electron capture detector, EDTA…ethylenediaminetetraacetic acid, EI…electron impact ionization, EPA… environmental protection agency, EX. . . extraction of the target analytes, FDA…Food and Drug Administration, FESI. . .field-enhanced sample injection, FID…flame ionization detector, FFD. . .full factorial design, FTD. . .flame thermionic detector, GC…gas chromatography, GS. . .gas sensor, HCB. . .hexachlorobenzene, HCH. . .hexachlorocyclohexane, He. . .helium, HFM. . .hollow fiber membrane-protected, HPLC…high-performance liquid chromatography, HS. . .headspace, IBMP. . .2-methoxy-3-(2-methylpropyl) pyrazine, INC…sample pre-incubation, INJ. . .sample injection into the system, it. . .ion trap MS analyzer, IRMM…Institute for Reference Materials and Measurements, IS. . .internal standard, ISI. . .ionspray ionization, K…distribution constant, KH…Henry‘s constant, KSOM…Kohonen‘s self-organizing map, LDA. . .linear discriminant analysis, LDR…linear dynamic range, LLE. . .liquid-liquid extraction, LMCS. . .longitudinally modulated cryogenic system, LOD. . .limit of detection, LOQ. . .limit of quantification, LTPRI. . .linear temperature-programmed retention index, MAE. . .microwave-assisted extraction, MRL…maximum residue level, MS…mass spectrometry, ms…magnetic sector MS analyzer, MTAD. . .4-methyl-1,2,4-triazoline-3,5-dione, MTBE…methyl tert.-butyl ether, n…number of moles extracted from the sample at equilibrium, NCI. . .negative chemical ionization, NPD. . .nitrogen phosphorus detector, OTA…ochratoxin A, OTC…oxytetracycline, P&T. . .purge & trap extraction, PA…polyacrylate coating, PC…principal component, PCA. . .principal component analysis, PCB. . .polychlorinated biphenyl, PDAS. . .portable dynamic air sampler, PDMS…polydimethylsiloxane coating, PDO…protected designation of origin, PFAIMS…planar field asymmetric-waveform ion mobility spectrometry, PFBHA. . .o-(2,3,4,5,6-pentafluorobenzyl) hydroxylamine, PFPD. . .pulsed flame photometric detector, PPMS. . .polyphenylmethylsiloxane, PPY. . .polypyrrole coating, q. . .quadrupole MS analyzer, ref. …reference, RP-SWMR. . .reversed polarity-stacking with matrix removal, RPSD. . .reduced pressure steam distillation, RSD…relative standard deviation, SBSE. . .stir-bar sorptive extraction, SDE. . .simultaneous steam distillation/solvent extraction, SDME. . .single drop microextraction, SF. . .StableFlex, SFC…supercritical fluid chromatography, SHS…static headspace extraction, SIDA…stable isotope dilution analysis, SOX. . .Soxhlet extraction, SPE. . .solid-phase extraction, SPME. . .solid-phase microextraction, SPTE. . .solid-phase trapping solvent extraction, TEA. . .thermal energy analyzer, TMA. . .trimethylamine, TMCS…trimethylchlorosilane, TMV…Tobacco Mosaic Virus, tof…time-of-flight, TPR. . .templated resin, USE. . .ultrasonic extraction, V, Vs. . .sample volume and composition, Vf… fiber coating volume, VOC…volatile organic compound. a If not specified otherwise, 1 cm SPME fibre was utilised, S: sample, INC: sample pre-incubation, EX: extraction of the target analytes, INJ: sample injection into the system, D: desorption of the target analytes into the system, other abbreviations explained in the list of abbreviations. b If not specified otherwise, linear temperature programmed retention indices (LTPRIs) used for compound identification where RI is indicated.

320

Handbook of Solid Phase Microextraction

Namiesnik et al.109 compiled a thorough review on the practical aspects of SPME analysis of organic vapours (air pollutants) in gaseous matrices and the analysis of the headspace of water, soil and food. Calibration-related challenges and an apparatus for the generation of standard calibration gas are described here. Food aroma analysis is the subject of another review by Plutowska and Wardencki.110 Characteristic aroma profiles (aromagrams, also called ‘fingerprints’), unique for individual food products and beverages, obtained by chromatographic methods or electronic nose techniques are discussed. Pillonel et al.111 reviewed rapid and cheap pre-concentration and enrichment techniques for the analysis of volatile compounds, mainly in dairy products. An older review of Mariaca and Bosset112 on instrumental analysis of volatile organic compounds (VOCs) in milk and dairy products offers an overview of sample preparation, extraction and pre-concentration procedures in this field. Purge and trap (P&T) and on-column injection techniques are discussed in more detail. Ampuero and Bosset113 reviewed the state-of-the-art trends in the development of aroma analysis with electronic noses, with special reference to its application to dairy products. Relatively novel sampling techniques, such as SPME or stir-bar sorptive extraction (SBSE) in relation to dairy product analysis, are discussed in this work. The determination of food contaminant, ochratoxin A (OTA), is critically reviewed in the paper of Monaci and Palmisano.114 Because OTA persists in food chains, exposure to this compound is a potential human health hazard and the evaluation of its occurrence is of great importance. Updated information on the HS
SPME
GC use for the analysis of volatile and semi-volatile compounds in biological fluids and materials, such as urine, blood, breast milk, faeces, hair, breath and saliva, is provided in the study of Mills and Walker.115 The more recent coupling of SPME to high-performance liquid chromatography (HPLC) extended the range of applications to non-volatile or thermally unstable compounds. This is a subject of the study presented by Zambonin,116 which is focused on food, as well as environmental and biological fluid samples. The coupling of SPME to electronanospray is discussed in the study of Walles et al.117 Biocompatible SPME devices, used for the analysis of peptides, were prepared using a coating of restricted access materials immobilised on steel and platinum wires. The review of practical applications of SPME in combination with HPLC for the analysis of toxic metallic species of As, Cr, Pb, Hg and Se in food and environmental matrices is presented in the paper of Malik et al.118 Pesticide residue analysis in water, soil, food samples and biological fluids, using SPME coupled to both GC and HPLC, is discussed, and the complete upto-date review of existing literature is given in the study of Beltran et al.119 In their review, Kataoka120 focused on in-tube SPME applications using an open tubular, fused silica capillary column with sorbent to extract organic compounds from aqueous samples. Using an autosampler, online in-tube SPME performed continuous extraction, concentration, desorption and injection. In published pesticide, drug, environmental pollutant and food contaminant analyses, the in-tube SPME technique has typically been coupled to HPLC separation systems.

Application of Solid-Phase Microextraction in Food and Fragrance Analysis

321

Applications of SPME in the analysis of biological samples have been summarised in papers by Theodoridis et al.121 and Ulrich.122 Important aspects of SPME application for the determination of pharmaceuticals, drugs of abuse and compounds of clinical and toxicological interest related to food chains are discussed here. Advantages, disadvantages and practical approaches to the determination of VOCs in plants are discussed in the study of Tholl et al.123

9.2.2

Case Studies

Rather than reviewing the individual food matrices extracted using a standard SPME procedure, the following part of this chapter focuses on several interesting applications in food classification, atypical food or fragrance matrices, comparison of extraction techniques including SPME, enantiomeric separation of compounds extracted by SPME, innovative calibration approaches, non-standard SPME arrangements or new/unusual extraction phases used in food analysis.

9.2.2.1 Case Study I: SPME
GC
TOFMS for Classification of Ice Wines The first case study of Setkova et al.124 focuses on the classification of ice wines using a rapid HS
SPME
GC
TOFMS/electron impact ionisation (EI) analytical method. Ice wine is a sweet dessert wine with very low yields (B5
10%, as compared to regular table wines, with typical yields of 50
70%), which explains its relatively high price and eventual adulteration attempts. Harvesting and pressing is performed while grapes are still frozen, and water in the berries is crystallised, which means that only a few drops of sweet juice are collected and relatively low yields are produced. Canada is the largest ice wine producer in the world, and several requirements defined by the Vintners Quality Alliance must be met, including the prohibition of artificial freezing, the allowance of certain sugar and acidity levels and so on. A new-generation, super-elastic DVB/CAR/PDMS 50/30-μm fibre assembly was used for the HS extraction of analytes in the long sequence of samples for classification purposes (statistical evaluation of the data was performed) and a GC
MS procedure, using the high-speed time-of-flight mass analyser (TOFMS, Leco Corporation, MI), was carried out to speed up the whole process. The ice wine sample consisted of 3 mL of wine with the addition of 1 g NaCl in a 10-mL headspace vial. The pre-incubation of the sample, for 5 min at 45 C with the agitation speed of 500 rpm, was carried out, followed by the sample extraction for another 5 min at the same temperature and stirring conditions. 3-Octanol was used as an internal standard. It was loaded onto the fibre prior to the sample extraction, by exposing the fibre coating to the headspace above the mixture of 4 g of pump oil and 20 μL of 3-octanol solution (concentration 1.09 mg/mL in methanol), in a 20-mL vial, for 30 s. Analytes were thermally desorbed into the GC injector for 2 min at 260 C, and then, the fibre was placed into the needle heater for a 2-min

322

Handbook of Solid Phase Microextraction

4.00 3.00

PC 1

2.00 1.00 0.00 –1.00 –2.00 –2.00

–1.00

0.00

1.00

2.00

3.00

PC 2

Figure 9.4 Principal component analysis using the HS
SPME
GC
TOFMS ice wine data to differentiate between the various grape varieties in Ontario ice wines. Principal component 1 (PC 1) and 2 (PC 2) are shown here. x, Vidal; e, Riesling; 3 , Gewu¨rtztraminer; ƒ, Cabernet Franc; &, Cabernet Sauvignon.124 Source: Reprinted with permission.

bake-out period (260 C). It should be noted that the whole GC
MS run was completed in less than 5 min. All screening analyses, searching for the suitable markers for classification purposes, were qualitative, except for the determination of concentrations of oak lactones, typically occurring in the samples fermented and/or aged in oak barrels. A standard addition calibration approach was used in oak lactone determination. The peak areas obtained from the examination of more than 130 ice wine samples (originated from Canada and the Czech Republic) were subjected to the principal component analysis (PCA) evaluation124 and to the Kohonen’s self-organizing map (KSOM) data evaluation technique,125 in order to differentiate and classify ice wines according to their geographical origin, grape varieties and vintage years. Clear differentiation of wines was obtained using both approaches, and the determinant aroma components, which were responsible for the observed differentiations, were identified and further discussed. PCA was applied to a large set of Ontario ice wine data to differentiate the various ice wine grape varieties and the results are shown here. PCA was carried out with 98 samples of common grape varieties included in the data set (white wines: Vidal, Riesling and Gewu¨rtztraminer; red wines: Cabernet Franc and Cabernet Sauvignon) and 21 aroma components. Principal component 1 (PC 1, explaining 41.37% of the variance) and PC 2 (explaining 14.07% of the variance), which best described the variance, are plotted in Figure 9.4.

Application of Solid-Phase Microextraction in Food and Fragrance Analysis

16 17

18 23

24

13 97

100

54

91

56

3

128 132

90

55

48 64 65

323

134 8 69

Canada

129

133

123

124 137

Czech Republic

Figure 9.5 A 35-cell, self-organising map computed from the signatures relative to ice wine labelled samples elaborated from a unique grape variety (Riesling) produced in Canada and the Czech Republic (27 samples, 58 chemical compounds).125 Source: Reprinted with permission.

The chemical volatile signatures were then interpreted with a non-supervised algorithm, the KSOM approach. A clear discrimination of ice wine samples, according to their Canadian or Czech origin and according to grape varieties, was achieved from a trained map-in, confirming a good classification efficacy of the method.125 The differentiation according to the ice wine origin using the KSOM approach is illustrated in Figure 9.5. It should be noted that the classification studies typically deal with a large number of samples to enable statistical data evaluation; therefore, the use of automated methods is almost unavoidable.

9.2.2.2 Case Study II: SPME
GCxGC
TOFMS for the Analysis of Honey Aroma The determination of honey origin is typically carried out by a relatively complex analysis of the pollen.126 As an alternative, an automated SPME technique using the DVB/CAR/PDMS 50/30-μm fibre in HS mode, combined with the comprehensive two-dimensional (2-D) GC separation and high-speed TOFMS (Leco ˇ Corporation, MI) analysis, was used in this study by Cajka et al.127 to analyse volatiles directly in honey aroma. The selected fibre provided the best sorption capacity and extracted the broadest range of volatile compounds from headspace above the honey sample. Two grams of honey mixed with distilled water (2 mL), placed in a 10-mL headspace vial, was a sufficient sample amount to achieve the required sensitivity. The optimised extraction and injection conditions were: pre-incubation

324

Handbook of Solid Phase Microextraction

time 5 min, agitation speed 500 rpm, pre-incubation/extraction temperature 40 C, extraction time 20 min, and desorption temperature and time 250 C and 45 s (splitless), respectively. Because the separation and detection are not a subject of this chapter, the GC and TOFMS conditions will not be discussed in detail. The performance of the optimised procedure using GCxGC separation was compared to the one-dimensional (1-D) separation system. Of all tested combinations, the best resolution of sample components was achieved in the GCxGC system when using the combination of DB-5MS (based on 5% phenyl methylpolysiloxane; 30 m 3 0.25 mm i.d. 3 0.25 μm film thickness) and Supelcowax column (based on PEG; 1.5 m 3 0.1 mm i.d. 3 0.1 μm film thickness) in the first and second dimension, respectively. Cryogenic modulator was mounted in the GC oven. The entire GC run was completed in 19 min. As expected, S/N ratios in the case of GCxGC analyses were improved as compared to the 1-D GC run (due to analyte refocusing in the modulator), and GCxGC proved to have higher peak capacities (due to two different separation mechanisms on two columns). Complementary separation mechanisms occurring on both columns led to the formation of structured chromatograms. The optimised analytical method was applied to the determination of honey aroma profile and examination of 164 volatile compounds representing the honey botanical and geographical origin (honey samples from Austria, the Czech Republic, Slovakia, France, Brazil and Italy). It should be noted that the number of detected and identified volatile compounds is rather impressive compared to the similar studies that use other extraction or detection systems. Spectral match with the National Institute of Science and Technology (NIST) library search and LTPRI approaches were used to identify the volatile aroma compounds.

9.2.2.3 Case Study III: Identification of Irradiated Beef Extract Powder Irradiation has been introduced in food preservation mainly as a response to the restrictions related to more-or-less-hazardous chemical preservation. Although irradiation is a clean and safe preservation technique, it has been banned in some countries, and irradiated food must be labelled in others. Irradiation is extremely cost-effective and is commonly used for wheat flower, spices, poultry and so on. Nevertheless, the amount of some essential nutrients, such as vitamin E, can be partially reduced by irradiation, and disposal of the used radioactive material is relatively complicated. Reliable detection of irradiated food is a great concern. It should be noted that the radiolytic marker must not be present in the non-irradiated sample, its intensity should not change with storage time and its occurrence must be dose-dependent. In a study by Kim et al.,128 the beef extract powder was irradiated with a 60Co source (γ-irradiation) at doses of 0, 1, 3, 5 and 10 kGy and stored for 0
30 days. SPME and P&T techniques were tested for their suitability to isolate volatile compounds from the headspace above the irradiated beef extract powder. The radiolytic marker, specific for irradiation-treated food, was found to be 1,3-bis(1,1dimethylethyl) benzene.

Application of Solid-Phase Microextraction in Food and Fragrance Analysis

325

For the SPME experiments, a 5-g sample with the addition of 5 mL distilled water and 0.3 mL of cyclohexanol (an absolute addition of 14.08 μg) as an internal standard (IS) were placed in a 20-mL headspace vial. A PDMS/DVB (65 μm) fibre was used to isolate volatiles from headspace (extraction temperature and time were 40 C and 40 min, respectively). The analytes were desorbed into the GC injector at 220 C (desorption time, 5 min). The same amount of sample, distilled water and IS as for SPME were transferred into a P&T vessel (25 mL) and connected to the concentrator equipped with a Tenax/silica gel/charcoal trap. The sample and line temperature were kept at 40 C and 200 C, respectively. A purge time of 20 min, desorption temperature of 225 C (5 min) and flow rate (He) of 20 psi were applied. A GC
qMS/EI system equipped with HP-INNOWax column (60 m 3 0.25 mm i.d. 3 0.25 μm film thickness) was used for the analyses. The volatile compound identification was based on the retention index method, in conjunction with the spectral match using the Wiley library. Statistical evaluation of the data using the SPSS software was performed. The results of this study showed that the P&T extraction technique was more efficient for the detection of low-molecular-weight and high-volatility compounds, compared to SPME. Nevertheless, SPME performed much better for compounds eluting at the end of the chromatogram (less-volatile components with a higher molecular weight). Therefore, these two methods can provide complementary data. The radiolytic volatile marker compound, 1,3-bis(1,1-dimethylethyl) benzene, was reliably detected using both extraction techniques. However, the SPME technique was considered by the authors to be more suitable due to its speed and cost.

9.2.2.4 Case Study IV: Volatile Aromatic Profile of Three European PDO Cheeses SPME and P&T techniques were compared and tested for their suitability for the extraction of cheese aroma compounds from raw milk cheeses in Mallia et al.30 In order to ensure the preservation of their history, tradition and diversity, and due to the increasing demand for traditional food, 147 European cheeses were given the protected designation of origin (PDO) label by the European Union (EU) authorities. Cheeses participating in this study were among those labelled as PDO. For SPME of the target analytes, 10 g of finely grated cheese was introduced into a 20-mL vial. Pre-incubation of the sampling vial was carried out for 1 h at 45 C. A DVB/CAR/PDMS 50/30-μm fibre (2 cm long) was inserted and exposed to the cheese headspace for 1 h. The volatile compounds were desorbed into the GC injector for 10 min at 260 C. For the dynamic HS extraction using the P&T system, 5 g of grated sample was pre-incubated for 5 min and then placed into the sparger for 15 min (35 C, N2, 40 mL/min). Analytes of interest were concentrated in a Tenax trap kept at 36 C and desorbed for 4 min at 230 C, into the cryofocusing unit at 2125 C and then desorbed into the GC column at 230 C for 1.5 min. A GC
qMS
flame ionisation detector (FID) mounted in parallel, splitting the flow in ratio of 1:1, and GC
MS
O (also parallel) separation/detection systems

326

Handbook of Solid Phase Microextraction

were used. The identification of the target analytes was carried out by spectral match with the Wiley library using the LTPRI method. The results of the comparison study were similar to those obtained in the previously mentioned case study III,128 i.e. SPME was found to be more effective for the extraction of medium- and high-boiling compounds than P&T extraction, which performed better for the isolation of highly volatile components. Similar to the previous case study, these two extraction techniques could be used together to provide complementary results. The results of analytical methods using each extraction technique separately were subjected to the statistical evaluation of the data and discriminant analysis correctly distinguished all participating samples according to their PDO origin. Analyses of Gruye`re cheeses (originating in Switzerland) resulted in high concentrations of alkenes, aldehydes, methyl ketones, butane-2,3-dione, unsaturated alcohols, branched chain acids and 2,6-dimethyl pyrazine. Manchego cheeses (originating in Spain) contained high concentrations of alkanes, alkanols, prop-2-en1-ol, propan-2-one and butan-2-one and their corresponding reduction products propan-2-ol and butan-2-ol, propyl esters and aromatic compounds. Ragusano cheeses (originating in Italy) contained high concentrations of fatty acids and ethyl and butyl esters.

9.2.2.5 Case Study V: Enantiomeric Composition of Fruit Beverages Chiral terpenes and their occurrence in commercial fruit beverages were the target analytes in the study by Riuz del Castillo et al.129 Although the beverage industry uses various fruits for juice production, orange juice remains the most popular. Limonene, linalool and α-pinene are considered the most prevalent components characterising the orange juice aroma. Adulteration of juices by adding aroma to fruit beverages has become a significant problem in the beverage industry and using the enantiomeric purity could serve as a valuable tool to authenticate the products by detecting the aroma addition. An SPME
GC
FID method with the chiral column Chirasil-β-Dex (25 m 3 0.25 mm i.d. 3 0.25 μm film thickness) installed in the GC oven was used in this study. A sample volume of 1 mL was introduced into a 5-mL vial and extracted with a PDMS 100-μm fibre in HS mode for 2 min at 60 C. Using the optimised method, relative standard deviation (RSD) values for standard solutions of terpenes ranging from 2% to 12% were achieved. It should be noted that because the enantiomeric sample composition can easily change, higher temperatures (more than 60 C) should not be applied in order to prevent such alterations. Table 9.2 summarises the levels of the investigated compounds in five commercially available beverages, as well as the enantiomeric purity. The enantiomeric excess, ee, was calculated from peak areas, and the excess of predominant enantiomer was expressed as a percent, as shown in Eq. (9.1): ee 5

A2B 3 100ð%Þ A1B

ð9:1Þ

Table 9.2 Enantiomeric Excess (ee), Predominant Enantiomer and Level of Chiral Terpenes in Commercial Beverages Examined by SPME
GC
FID in Case Study V129 Terpene

Beverage A ee (%)

(1)-α-Pinene (2)-β-Pinene (2)-α-Phellandrene (1)-Limonene (1)-Linalool (1)-Citronellal Terpinen-4-ol (1)-α-Terpineol (2)-β-Citronellol (2)-β-trans-Caryophyllene

nda nd nd 100 52.2 100 16.3(2)b 51.4 nd 100

Beverage B

Level (µg/mL)

ee (%)

Level (µg/mL)

0.22 nd nd 29.0 3.97 0.12 0.71 0.19 nd 0.22

100 100 100 100 32.9 100 2.6(1) 13.3 100 100

0.15 1.82 0.22 37.0 4.26 0.16 2.12 1.78 0.17 0.38

Beverage C ee (%)

100 nd 100 100 74.0 100 45.7(1) 54.6 100 100

Beverage A: dairy product containing orange juice plus natural and nature-identical aromas. Beverage B: dairy product containing orange juice plus unspecified aromas. Beverage C: dairy product containing various fruit juices, including orange juice, plus unspecified aromas. Beverage D: orange soft drink containing natural aromas. Beverage E: dairy product containing peach juice plus unspecified aromas. a Not detected. b Predominant enantiomer. Source: Reprinted with permission.

Level (µg/mL) 0.32 nd 0.15 25.0 3.15 0.21 1.18 1.52 0.12 0.32

Beverage D ee (%)

100 nd nd 100 100 100 37.0(1) 76.6 nd nd

Beverage E

Level (µg/mL)

ee (%)

Level (µg/mL)

0.17 nd nd 29.5 3.72 0.14 1.95 2.11 nd nd

100 20.7 nd 100 7.2 nd nd nd nd 100

2.32 4.81 nd 2.56 4.59 nd nd nd nd 2.81

328

Handbook of Solid Phase Microextraction

where A is the peak area of the predominant enantiomer and B is the peak area of the minor enantiomer. While the enantiomeric distribution of some terpenes did not change in various fruit beverage samples, significant variation was observed for other terpenes. As can be seen from Table 9.2, (1)-α-pinene, (2)-α-phellandrene, (1)-limonene, (1)-citronellal, (2)-β-citronellol and (2)-β-trans-caryophyllene occurred as pure enantiomers in all samples, when present, whereas β-pinene, linalool, terpinen-4-ol and α-terpineol presence varied depending on the examined beverage. Therefore, the SPME
GC determination of the enantiomeric distribution of chiral terpenes, mainly limonene, linalool and α-terpineol, can be used for the detection of aroma addition. Natural enantiomeric distributions of those compounds are 100%, 100% and 80%, respectively, the (1)-enantiomer being the predominant isomer. Any alteration in these values may be attributable to the addition of aromas. The wide range of enantiomeric excesses (7.2
74.0%) found for (1)-linalool in Beverages A, B, C and E (see Table 9.2) revealed either the employment of a technological process involving heat treatment or, more likely, the addition of aromas to improve the sensorial perception of those beverages (these samples were declared as having been submitted to the same processing procedures as sample D where no racemisation was observed). The non-orange aroma addition led to the presence of (2)-linalool, which is naturally absent in orange juices. Because the ratio of (1)- and (2)-enantiomers of terpinen-4-ol is significantly dependent on the fruit cultivar, this compound is useless in the enantiomeric purity experiments. Nevertheless, α-terpineol seems to be a valuable marker in confirmation of natural origin. Its enantiomeric distribution in Beverage D was very close (76.6%) to that obtained previously in natural orange concentrates (80%). Enantiomeric ratios in other studied beverages again supported the addition of aromas to those samples.

9.2.2.6 Case Study VI: Equilibrium In-Fibre Standardisation in SPME Internal standardisation by loading of deuterated standard onto the fibre coating prior to the introduction of the fibre into the sampling vial for the extraction procedure is described in the study of Wang et al.130 Introducing low amounts (milligram levels) of standards such as benzene, toluene, ethylbenzene and xylenes (BTEX) or naphthalene into a few grams of pump oil proved to be an excellent and reproducible standard generator (RSD levels were less than 4% for over 100 standard loadings from a single solution). The quantification is based on two simultaneous processes: (i) extraction of analytes from the sample matrix into the fibre coating and (ii) desorption of standards from the fibre coating into the sample matrix. Isotropy of absorption and desorption is maintained for in-vial analysis (see Chapter 6). The fully automated SPME part of the method consisted in using a PDMS 100-μm fibre for the loading of deuterated ethylbenzene or naphthalene (IS) by exposure to the headspace of a 20-mL vial containing 4 g pump oil spiked to a concentration of 0.47 mg/g (ethylbenzene) or 4 mg standard in 2 g of pump oil (naphthalene) for 1 min at 35 C. Then the fibre was exposed to the headspace of a

Application of Solid-Phase Microextraction in Food and Fragrance Analysis

329

Table 9.3 Calculated Recoveries of BTEX from a Dry White Wine with and without the Use of In-Fibre Internal Standardisation130 Compound

Benzene Toluene Ethylbenzene o-Xylene

Relative Recovery (%) (Standard Deviation,%, n 5 3) Using External Calibration Data

Using Internal Calibration Data

77 (4) 78 (4) 76 (3) 69 (2)

98 (2) 98 (2) 100 (1) 91 (1)

Source: Reprinted with permission.

10-mL vial including 3 mL of sample for 5 min at the same temperature (an agitation speed of 500 rpm was applied for both processes), after a 6-min pre-incubation period at the same experimental conditions. The desorption was carried out for 1 min at 250 C and was followed by a fibre bake-out. It was calculated that, theoretically, the standard-containing vial can be used 115 times (ethylbenzene) or 300 times (naphthalene) before 1% of the vial contents will be removed. GC
MS/EI and GC
FID systems were used for analyte separation and detection. An RTX-5MS column (25 m 3 0.25 mm i.d. 3 0.25 μm film thickness) was installed in the GC oven. Other experimental conditions are described in the paper. The developed and optimised method corrected for matrix effects and was successfully applied for the analysis of real samples of white wine. Deuterated ethylbenzene-d10 and MS detection were used for this purpose. The linearity over the calibration range for both external and internal standardisation approaches was excellent (r2 $ 0.9998). The recoveries from wine spiked with 7.3 μg/L BTEX are summarised in Table 9.3, and they demonstrate that the in-fibre standardisation provides more accurate determination of these compounds compared to the external calibration. It should be noted that the developed automated procedure, including the in-fibre standardisation, pre-incubation, extraction, desorption and fibre bakeout, requires only a single arm autosampler.

9.2.2.7 Case Study VII: Tetracycline Antibiotics in Fish Muscle by In-Tube SPME
HPLC Simultaneous analysis of four tetracycline antibiotics (tetracycline, oxytetracycline, chlortetracycline and doxycycline) in fish muscle tissues, by in-tube SPME with HPLC and photodiode array detector (PDA), was developed by Wen et al.131 The relatively widespread use of tetracycline antibacterial compounds in veterinary medicine and aquaculture can lead to the occurrence of their residues in animaloriginated food. A biocompatible poly(methacrylic acid-ethylene glycol dimethacrylate) monolithic capillary (15 cm 3 0.25 mm i.d.) was used for extraction. Due to the biocompatibility of the extraction phase, no protein precipitation was necessary prior to extraction. The sample (1 g of edible muscle of crucian) was spiked with variable

330

Handbook of Solid Phase Microextraction

amounts of tetracycline antibiotics and treated with 10 mL of 0.01M ethylenediaminetetraacetic acid (EDTA) in MacIlvaine buffer solution (the sample matrix was adjusted to pH 4), homogenised and left in the dark for 30 min at 4 C. Finally, the sample was centrifuged at 12,000 rpm for 5 min and the supernatant was transferred directly into the extraction device. The fish sample extraction time was 16 min. The HPLC analytical column was packed with Kromasil ODS (250 mm 3 4.6 mm i.d. 3 5 μm). The optimised mobile phase for desorption and separation was 20% methanol
20% acetonitrile
60% 0.02 mol/L oxalic acid solution (pH 3), and the flow rate was kept at 0.8 mL/min. The limits of detection for tetracycline, oxytetracycline, chlortetracycline and doxycycline were 22, 16, 30 and 21 ng/g, respectively. The linearity of calibration curves was satisfactory, with the regression coefficient r2 . 0.9980. The calculated standard deviation levels showed very good precision of the optimised method (standard deviation , 4.2% for 1-day reproducibility and , 5.7% for between-day reproducibility, respectively). As can be seen from the detection limits, the developed method also easily fulfilled the maximum residue limits (MRLs) established by the worldwide authorities (EU and China: 100 ng/g, U.S. Food and Drug Administration: 2 μg/g in muscle). The efficacy of the optimised analytical method was inspected by analysing real fish muscle samples both (i) spiked with the four target tetracycline antibiotics and (ii) orally medicated with a daily dose of 100 mg/kg oxytetracycline (OTC) for 4 days. No interfering peaks were obtained in the chromatograms of any blank fish tissues.

9.3

Concluding Remarks

A summary of recently published studies using SPME, with special emphasis on the food and fragrance analysis, was compiled and several papers were discussed in more detail, mainly with respect to the SPME conditions used and food matrices examined. A brief summary of the reviews related to food and fragrance analysis was given, and recent food classification studies employing SPME were discussed. The purpose of the classification studies is either (i) to find the specific markers characterising the sample group or (ii) the in-between sample comparison of chromatograms or mass spectral fingerprints. For this purpose, the new-generation, super-elastic fibres proved to be more robust and reliable in long sequences of samples (in terms of durability) than the silica-based fibres. SPME extraction proved to be sensitive enough to satisfy legislation requirements related to low detection and quantification limits, as well as method accuracy and precision requirements.

References 1. A Rizzolo, A Brambilla, S Valsecchi & P Eccher-Zerbini, Food Chem 77 (2002) 257 2. A Carrasco-Pancorbo, L Cerretani, A Bendini & A Segura-Carretero, J Sep Sci 28 (2005) 837

Application of Solid-Phase Microextraction in Food and Fragrance Analysis

331

3. BD Fredalina, BH Ridzwan, AA Zainal Abidin, MA Kaswandi, H Zaiton, I Zali, P Kittakoop & AM Jais, Gen Pharmacol 33 (1999) 337 4. G Duflos, VM Coin, M Cornu, J-F Antinelli & P Malle, J Sci Food Agric 86 (2006) 600 5. P Simko, J Chromatogr B 770 (2002) 3 6. A Gramza & J Korczak, Trends Food Sci Technol 16 (2005) 351 7. A Go´mez-Hens & MP Aguilar-Caballos, Trends Anal Chem 22 (2003) 847 8. Y-L Feng, J Zhu & R Sensenstein, Anal Chim Acta 538 (2005) 41 9. AE Karatapanis, AV Badeka, KA Riganakos, IN Savvaidis & MG Kontominas, Int Dairy J 16 (2006) 750 10. GG Rye & DG Mercen, Food Res Int 36 (2003) 167 11. V Huis & HJ Jos, Int J Food Microbiol 33 (1996) 1 12. G Mortensen, J Sorensen & H Stapelfeldt, Eur Food Res Technol 216 (2003) 57 13. C Couteau & L Coiffard, Int J Food Sci Technol 38 (2003) 77 14. DS Mottram, Int J Food Sci Technol 33 (1998) 19 15. R Rial Otero, B Cancho Grande & J Simal Ga´ndara, J Chromatogr A 992 (2003) 121 16. FE Ahmed, Trends Anal Chem 20 (2001) 649 17. O-W Lau & S-K Wong, J Chromatogr A 882 (2000) 255 18. MB Petersen & L Poll, Eur Food Res Technol 209 (1999) 251 19. NJN Yau & JJ Huang, Food Qual Preference 7 (1996) 263 20. T Luckow & C Delahunty, Food Res Int 37 (2004) 805 21. K Morita, K Kubota & T Aishima, J Sci Food Agric 83 (2003) 289 22. MJC Pontes, SRB Santos, MCU Arau´jo, LF Almeida, RAC Lima, EN Gaia˜o & UTCP Souto, Food Res Int 39 (2006) 182 23. S Sumar & H Ismail, Nutr Food Sci 95 (1995) 11 24. IP Hurley, H Elyse Ireland, RC Coleman & JHH Williams, Int J Food Sci Technol 39 (2004) 873 25. M Brenes, A Garcı´a, JJ Rios, P Garcı´a & A Garrido, Int J Food Sci Technol 37 (2002) 615 26. C Durante, M Cocchi, M Grandi, A Marchetti & R Bro, Chemom Intell Lab Syst 83 (2006) 54 27. M Cocchi, C Durante, G Foca, D Manzini, A Marchetti & A Ulrici, Chemom Intell Lab Syst 71 (2004) 129 28. C Bauer-Christoph, N Christoph, BO Aguilar-Cisneros, MG Lo´pez & E Richling, Eur Food Res Technol 217 (2003) 438 29. S Kallithraka, IS Arvanitoyannis, P Kefalas, A El-Zajouli, E Soufleros & E Psarra, Food Chem 73 (2001) 501 30. S Mallia, E Fernandez-Garcia & JO Bosset, Int Dairy J 15 (2005) 741 31. DC Frank, CM Owen & J Patterson, Lebensm Wiss Technol 37 (2004) 139 32. HM Burbank & MC Qian, J Chromatogr A 1066 (2005) 149 33. CG Zambonin, L Monaci & A Aresta, Food Chem 78 (2002) 249 34. F Bellesia, A Pinetti, UM Pagnoni, R Rinaldi, C Zucchi, L Caglioti & G Palyi, Food Chem 83 (2003) 55 35. L Lecanu, V Ducruet, C Jouquand, JJ Gratadoux & A Feigenbaum, J Agric Food Chem 50 (2002) 3810 36. L Mondello, R Costa, PQ Tranchida, B Chiofalo, A Zumbo, P Dugo & G Dugo, Flavour Fragr J 20 (2005) 659 37. FK Tavaria, TG Tavares, AC Silva-Ferreira & FX Malcata, Int Dairy J 16 (2006) 886 38. A Lunden, V Gustafsson, M Imhof, R Gauch & J-O Bosset, J Dairy Res 69 (2002) 383

332

Handbook of Solid Phase Microextraction

39. 40. 41. 42. 43.

CM Lock, L Chen & DA Volmer, Rapid Commun Mass Spectrom 13 (1999) 1744 C Basheer & HK Lee, J Chromatogr A 1047 (2004) 189 ML Perkins, BR D’Arcy, AT Lisle & HC Deeth, J Sci Food Agric 85 (2005) 2421 M Povolo & G Contarini, J Chromatogr A 985 (2003) 117 M Adahchour, J Wiewel, R Verdel, RJJ Vreuls & UAT Brinkman, J Chromatogr A 1086 (2005) 99 P Das, M Gupta, A Jain & KK Verma, J Chromatogr A 1023 (2004) 33 R Andrade, FGR Reyes & S Rath, Food Chem 91 (2005) 173 A Marco, JL Navarro & M Flores, Food Chem 84 (2004) 633 S Ventanas & J Ruiz, Talanta 70 (2006) 1017 LM Chiesa, S Soncin, PA Biondi, P Cattaneo & C Cantoni, Vet Res Commun 30 (2006) 349 G Centineo, EB Gonzalez, JI Garcia Alonso & A Sanz-Medel, J Mass Spectrom 41 (2006) 77 F Bianchi, M Careri, M Musci & A Mangia, Food Chem 100 (2007) 1049 CC Grimm, C Bergman, JT Delgado & R Bryant, J Agric Food Chem 49 (2001) 245 S Wongpornchai, K Dumri, S Jongkaewwattana & B Siri, Food Chem 87 (2004) 407 JA Stinson, KC Persaud & G Bryning, Sens Actuators B 116 (2006) 100 M Vinaixa, A Vergara, C Duran, E Llobet, C Badia, J Brezmes, X Vilanova & X Correig, Sens Actuators B 106 (2005) 67 A Barra, N Baldovini, A-M Loiseau, L Albino, C Lesecq & L Lizzani Cuvelier, Food Chem 101 (2007) 1279 R Naef, A Velluz & A Jaquier, Eur Food Res Technol 22 (2006) 554 A Williams, D Ryan, A Olarte Guasca, P Marriott & E Pang, J Chromatogr B 817 (2005) 97 TR Hamilton-Kemp, DD Archbold, RW Collins & K Yu, J Sci Food Agric 83 (2003) 283 MG Lo´pez, GR Guzma´n & AL Dorantes, J Chromatogr A 1036 (2004) 87 E Carasek & J Pawliszyn, J Agric Food Chem 54 (2006) 8688 T Navarro, C de Lorenzo & RA Perez, Anal Bioanal Chem 379 (2004) 812 C Tsoutsi, I Konstantinou, D Hela & T Albanis, Anal Chim Acta 573
574 (2006) 216 A Kanavouras & RJ Hernandez, Int J Food Sci Technol 41 (2006) 743 F Bianchi, M Careri, A Mangia & M Musci, J Chromatogr A 1102 (2006) 268 L Ro¨hrig & H-U Meisch, Fresenius J Anal Chem 366 (2000) 106 S Ampuero, S Bogdanov & J-O Bosset, Eur Food Res Technol 218 (2004) 198 E Alissandrakis, PA Tarantilis, PC Harisanis & M Polissiou, Food Chem 100 (2007) 396 JCR Demyttenaere, RM Morina, N De Kimpea & P Sandra, J Chromatogr A 1027 (2004) 147 P Bartak, P Bednar, L Cap, L Ondrakova & Z Stransky, J Sep Sci 26 (2003) 715 S Guillot, MT Kelly, H Fenet & M Larroque, J Chromatogr A 1101 (2006) 46 N Kayali, FG Tamayo & LM Polo-Diez, Talanta 69 (2006) 1095 IOM Chan, PKS Lam, RHY Cheung, MHW Lam & RSS Wu, Analyst 130 (2005) 1524 J Hernandez-Borges, A Cifuentes, FJ Garcia-Montelongo & MA Rodriguez-Delgado, Electrophoresis 26 (2005) 980 B Zierler, B Siegmund & W Pfannhauser, Anal Chim Acta 520 (2004) 3 E Kafkas, T Cabaroglu, S Selli, A Bozdogan, M Ku¨rkcu¨oglu, S Paydas & KHC Baser, Flavour Fragr J 21 (2006) 68

44. 45. 46. 47. 48. 49. 50. 51. 52. 53. 54. 55. 56. 57. 58. 59. 60. 61. 62. 63. 64. 65. 66. 67. 68. 69. 70. 71. 72. 73. 74. 75.

Application of Solid-Phase Microextraction in Food and Fragrance Analysis

333

76. M Riu-Aumatell, M Castellari, E Lopez-Tamames, S Galassi & S Buxaderas, Food Chem 87 (2004) 627 77. L Cai, J Xing, L Dong & C Wu, J Chromatogr A 1015 (2003) 11 78. J Wu, W Xie & J Pawliszyn, Analyst 125 (2000) 2216 79. ES de Brito, N Narain, NH Pezoa Garcia, ALP Valente & GF Pini, J Sci Food Agric 82 (2002) 534 80. L Mondello, R Costa, PQ Tranchida, P Dugo, ML Presti, S Festa, A Fazio & G Dugo, J Sep Sci 28 (2005) 1101 81. D Ryan, R Shellie, P Tranchida, A Casilli, L Mondello & P Marriott, J Chromatogr A 1054 (2004) 57 82. O Pinho, IMPLVO Ferreira & LHMLM Santos, J Chromatogr A 1121 (2006) 145 83. TE Siebert, HE Smyth, DL Capone, N Neuwo¨hner, KH Pardon, GK Skouroumounis, MJ Herderich, MA Sefton & AP Pollnitz, Anal Bioanal Chem 381 (2005) 937 84. S Jo¨nsson, T Uusitalo, B van Bavel, I-B Gustafsson & G Lindstro¨m, J Chromatogr A 1111 (2006) 71 85. D Ryan, P Watkins, J Smith, M Allen & P Marriott, J Sep Sci 28 (2005) 1075 86. W Wardencki, J Orlita & J Namie´snik, Fresenius J Anal Chem 369 (2001) 661 87. DW Lachenmeier, U Nerlich & T Kuballa, J Chromatogr A 1108 (2006) 116 88. C Da Porto, F Cordaro & N Marcassa, Lebensm Wiss Technol 39 (2006) 159 89. L Jirovetz, G Buchbauer, MB Ngassoum & M Geissler, J Chromatogr A 976 (2002) 265 90. L Giordano, R Calabrese, E Davoli & D Rotilio, J Chromatogr A 1017 (2003) 141 91. E Duran Guerrero, R Natera Marin, R Castro Mejias & C Garcia Barroso, J Chromatogr A 1104 (2006) 47 92. GA Eiceman, A Tarassova, PA Funk, SE Hughs, EG Nazarov & RA Miller, J Sep Sci 26 (2003) 585 93. SB Stanfill, AM Calafat, CR Brown, GM Polzin, JM Chiang, CH Watson & DL Ashley, Food Chem Toxicol 41 (2003) 303 94. Y Chen, F Begnaud, A Chaintreau & J Pawliszyn, Flavour Fragr J 21 (2006) 822 95. N-S Kim & D-S Lee, J Chromatogr A 982 (2002) 31 96. Z-G Li, M-R Lee & D-L Shen, Anal Chim Acta 576 (2006) 43 97. C Bertrand, G Comte & F Piola, Biochem Syst Ecol 34 (2006) 371 98. J Xie, B Sun & M Yu, Flavour Fragr J 21 (2006) 798 99. N Yassaa & J Williams, Atmos Environ 39 (2005) 4875 100. C Deng, X Zhang, W Zhu & J Qian, Chromatographia 59 (2004) 263 101. C Deng, N Li, W Zhu, J Qian, X Yang & X Zhang, J Sep Sci 28 (2005) 172 102. TA van Beek, IMMS Silva, MA Posthumus & R Melo, J Chromatogr A 1067 (2005) 311 103. V Rodriguez, M Yonamine & E Pinto, J Sep Sci 29 (2006) 2085 104. W Wardencki, M Michulec & J Curyło, Int J Food Sci Technol 39 (2004) 703 105. R Marsilli, Ed, Flavor, Fragrance & Odor Analysis (2002) Marcel Dekker, Inc.: New York, NY, ISBN: 0-8247-0627-7 106. H Kataoka, HL Lord & J Pawliszyn, J Chromatogr A 880 (2000) 35 107. PL Buldini, L Ricci & JL Sharma, J Chromatogr A 975 (2002) 47 108. G Vas & K Vekey, J Mass Spectrom 39 (2004) 233 109. J Namiesnik, B Zygmunt & A Jastrzebska, J Chromatogr A 885 (2000) 405 110. B Plutowska & W Wardencki, Food Chem 101 (2007) 845 111. L Pillonel, JO Bosset & R Tabacchi, Lebensm Wiss Technol 35 (2002) 1

334

112. 113. 114. 115. 116. 117. 118. 119. 120. 121. 122. 123. 124. 125. 126. 127. 128. 129. 130. 131.

Handbook of Solid Phase Microextraction

R Mariaca & JO Bosset, Lait 77 (1997) 13 S Ampuero & JO Bosset, Sens Actuators B 94 (2003) 1 L Monaci & F Palmisano, Anal Bioanal Chem 378 (2004) 96 GA Mills & V Walker, J Chromatogr A 902 (2000) 267 CG Zambonin, Anal Bioanal Chem 375 (2003) 73 M Walles, Y Gu, C Dartiguenave, FM Musteata, K Waldron, D Lubda & J Pawliszyn, J Chromatogr A 1067 (2005) 197 AK Malik, V Kaur & N Verma, Talanta 68 (2006) 842 J Beltran, FJ Lopez & F Hernandez, J Chromatogr A 885 (2000) 389 H Kataoka, Anal Bioanal Chem 373 (2002) 31 G Theodoridis, EHM Koster & GJ de Jong, J Chromatogr B 745 (2000) 49 S Ulrich, J Chromatogr A 902 (2000) 167 D Tholl, W Boland, A Hansel, F Loreto, USR Ro¨ose & J-P Schnitzler, Plant J 45 (2006) 540 L Setkova, S Risticevic & J Pawliszyn, J Chromatogr A 1147 (2007) 224 JL Giraudel, L Setkova, J Pawliszyn & M Montury, J Chromatogr A 1147 (2007) 241 E Anklam, Food Chem 63 (1998) 549 T Cajka, J Hajslova, J Cochran, K Holadova & E Klimankova, J Sep Sci 30 (2007) 534 H Kim, W-J Cho, J-S Ahn, D-H Cho & Y-J Cha, Microchem J 80 (2005) 127 ML Ruiz del Castillo, MM Caja & M Herraiz, J Agric Food Chem 51 (2003) 1284 Y Wang, J O’Reilly, Y Chen & J Pawliszyn, J Chromatogr A 1072 (2005) 13 Y Wen, Y Wang & Y-Q Feng, Talanta 70 (2006) 153

10 Drug Analysis by SPME Heather Lord and Barbara Bojko Department of Chemistry, University of Waterloo, Waterloo, Ontario, Canada

10.1

Introduction

The study of solid-phase microextraction (SPME) technology for the analysis of drugs has been pursued for nearly two decades. In this time, significant advances have been made in the theory and application of the technology for both therapeutic pharmaceutical compounds and drugs that are abused. As has been seen for other applications of the technology, the primary attractants are the simplified nature of SPME, resulting in reduced sample handling and solvent use, time and cost. Initial applications in drug analysis were primarily for forensic drugs. Because these compounds are typically less polar and more volatile, they were readily analysable by PDMS-coated fibres interfaced to gas chromatography (GC). Numerous applications appeared on the analysis of compounds such as the amphetamines. With the introduction of the solid sorbent coatings, primarily the divinylbenzene (DVB)-related products, in the mid-1990s, a broader range of drugs were amenable for analysis, opening up the field of application to the semivolatiles, such as cocaine and the benzodiazepines. Application to non-volatile drugs appeared somewhat later, partially due to the limited availability of extraction phases for these compounds
only the Carbowaxs templated resin (CW/ TPR) fibre was widely applicable
and due to the more cumbersome nature of interfacing SPME with condensed phase separations. In recent years, these latter challenges have spurred much research interest in the development of improved phases for non-volatile extraction and strategies for interfacing SPME with liquid chromatography (LC) and capillary electrophoresis (CE) separations. These developments are now also bearing the fruit of opening the technology to the analysis of other compounds of biomedical and biological interest, such as proteins, peptides, endogenous biological compounds and biomarkers of health status. This chapter presents a review of the application of SPME technology to various aspects of drug and biomedical analysis, along with practical descriptions of its implementation and use.

10.2

Fundamentals of Extraction

While the fundamentals of the extraction process do not differ between the application to drugs and compounds of interest in other areas such as environmental, food, Handbook of Solid Phase Microextraction. DOI: 10.1016/B978-0-12-416017-0.00010-3 © 2012 Elsevier Inc. All rights reserved.

336

Handbook of Solid Phase Microextraction

fragrance, several aspects of SPME theory and practice warrant discussion of the particular relevance to drug analysis. Considerations are generally relevant to the lower diffusion rates in the condensed phase and the non-volatile and sometimes thermal instability of the compounds. Only the extraction fundamentals of particular interest for SPME drug analysis are discussed here. The reader is referred to Chapter 2 for a general overview of SPME extraction fundamentals and Chapter 7 for a discussion of general SPME method development.

10.2.1 Equilibrium Extraction As an equilibrium technique, SPME calibration depends on the establishment of an equilibrium between analyte concentrations in the extraction phase and the bulk of the sample. Where the ratio of amount of sample phase to the amount of extracting phase is high, a substantial proportion of the initial analyte remains in the sample phase after extraction. Indeed, for the condition of negligible depletion where sampling is independent of sample volume, the amount of analyte remaining in the sample phase must not deviate from the initial amount present, within experimental error. The establishment of this equilibrium is followed by the equilibrium time profile or a plot of amount extracted versus extraction time. The equilibrium time is determined as the time required for extraction of 95% of the equilibrium amount. The time required for the analyte to reach an equilibrium state within the extraction phase is affected by both the time required for sufficient mass of the analyte to reach the surface of the extraction phase as well as the time required for the analyte to distribute within that phase. Where diffusion rates within the boundary layer surrounding the extraction phase are slow, complete equilibration may take a long time. As this is commonly the case in the direct extraction of analytes from liquid samples, extraction is often terminated prior to complete equilibration. So long as the extraction time is carefully controlled, the strategy will be successful.

10.2.2 Exhaustive Extraction Exhaustive extraction is a situation where an analyte is completely removed from a sample during extraction. The capacity of the extraction phase for the analytes extracted must be significantly larger than the amount of analyte in the sample (see also Section 6.4 in Chapter 6). This is the mode of extraction employed by most sample preparation techniques [e.g. liquid
liquid extraction (LLE), solid-phase extraction (SPE)]. Although SPME probes were not originally intended to be applied for exhaustive extraction, exhaustive extraction is possible to achieve where the phase ratio (sample:sorbent) is quite low and/or the sorbent capacity is sufficiently high. In practice, exhaustive extraction is considered achieved when the probe capacity is sufficiently high, and sample volume/concentration is sufficiently low, to allow $90% of analyte to be extracted by the probe. A determination of whether exhaustive extraction is possible for a given extraction may be made either experimentally or theoretically. The theoretical determination is described in detail elsewhere.1 Experimentally, a determination must be

Drug Analysis by SPME

337

made for various sample volumes regarding whether the amount extracted is .90% of that originally present in the sample. In the analysis of unknowns using an exhaustive extraction, the initial sample concentration may be determined directly from a measure of the amount extracted, given the volume of the sample used for the extraction. No additional calibration is required. In practice, however, a calibration of the amount extracted to the sample concentration is typically used. Because sample volumes in bioanalysis can often be very limited and the capacity of solid sorbent fibres high, researchers in this field should take care to ensure they understand the proportion of analyte removed during an extraction.

10.2.3 Recovery: Comparison of Definitions It is important to understand the degree of recovery of an SPME method for drug analysis prior to implementing it for a new application. Authors reporting the recovery of a method typically report either the proportion of total analyte extracted by a method or the proportion of analyte extracted from the target matrix relative to extraction from a matrix-free sample. For SPME, the former is useful in determining the degree of exhaustive extraction, or whether negligible depletion has been achieved. The latter indicates whether the sample matrix has an effect on the extraction process. For practitioners of SPE sample preparation, the former descriptor is important because the extraction by definition is assumed to be exhaustive. If it is not, there are important additional considerations for calibration. Some early SPME methods in the literature reported the low observed ‘recovery’ as a deficit in this regard. Care must be taken when reading literature methods to determine which descriptor of recovery is being used. The second descriptor of recovery (degree of matrix influence) is important for decisions on the appropriate calibration for SPME and is particularly important in bioanalysis, where matrices are often complicated. Where the matrix has little or no impact on the amount extracted, external calibration is feasible, and it may not be necessary to ensure a close match between the matrix used for the standards and that of the samples. Where the matrix does have a significant impact on the amount extracted, it is important to either take care to ensure that the matrix used for the standards is closely matched to that of the samples or else employ an internal standard (IS) or standard addition calibration.

10.2.4 Equilibration Time Several factors affect the time required for sufficient mass of the analyte to reach the surface of the extraction phase, including time to diffuse within the boundary layer surrounding the extraction phase and time to reach the boundary layer. For analysis by direct immersion, diffusion in the boundary layer can be relatively slow. Because many drugs of analytical interest are non-volatile, this extraction mode must be employed, with the result that equilibrium extraction for these

338

Handbook of Solid Phase Microextraction

compounds takes much longer than equilibrium extraction of volatile compounds from a gaseous sample, where the diffusion coefficients are much higher. In a well-stirred liquid sample with direct immersion extraction, it is assumed that the analyte concentration at the interface between the boundary layer and the bulk of the sample is always equal to the concentration in the bulk of the sample. This is true of extraction from both liquid and gaseous samples. In this case, time required for a sufficient mass of analyte to reach the boundary layer is inconsequential. Where extraction is from the headspace above a liquid or solid sample, however, mass transfer from the liquid or solid to the headspace may be slow. Again, because of their relatively low volatilities and slow diffusion in the condensed phase, many drugs of analytical interest will suffer from this slow mass transfer to the boundary layer and have relatively long equilibrium times. Finally, diffusion or establishment of an equilibrium within the extraction phase can be much slower for these compounds than for compounds more amenable to GC. In a partitioning sorbent, either the diffusion coefficient itself may be very small or the partition coefficient may be low. In the latter case, a large volume of extraction phase may be required to provide the necessary sensitivity. Because diffusion equilibrium time within a partitioning sorbent depends on the square of the thickness of the extraction phase (Eq. (10.1)), even a small increase in the sorbent thickness will cause a significant lengthening of extraction time (see also Section 2.4 in Chapter 2).2 te 5 t95% 5

ðb 2 aÞ2 2Df

ð10:1Þ

In this equation, (b
a) refers to the thickness of the sorbent material, and Df refers to analyte diffusion in the sorbent (see Figure 2.10). Solid sorbents are widely used for SPME drug analysis, primarily because of the larger amounts of drug that can be extracted by these phases relative to the partitioning (liquid) phases. In the case of a non-porous solid sorbent, equilibration within the sorbent does not occur. As soon as an analyte molecule diffuses through the boundary layer and reaches the active site, it binds there with a rate constant that is usually sufficiently fast that this step is not limiting for the overall extraction rate. In most cases, however, such sorbents would not have sufficient capacity to be of much value. Generally solid sorbents used for SPME have a high degree of porosity, in which case analyte molecules must diffuse through the sorbent pores prior to binding with an active site. Thus, the diffusion path length can be quite long and often determines the overall extraction equilibration time. Finally, although in most cases, the kinetics associated with an analyte binding to an active site are negligible compared to the time required for other steps in an extraction, they can be appreciable and limiting where the active site is a biological recognition unit. This has been demonstrated for immunoaffinity SPME sorbents.1,3 A slow rate of derivatisation may also limit the overall extraction time. Several strategies have been used to address the long equilibration times encountered in many SPME drug analyses. First, effective stirring is essential in any direct immersion extraction. Also, strategies to force a sufficient analyte concentration

Drug Analysis by SPME

339

efficiently to the headspace are often employed, including limiting headspace volume, heating of the sample and the addition of salt to the sample. For porous solid sorbents, limiting of the sorbent volume to just that required for the sensitivity needed can make a substantial improvement in equilibration time.

10.2.5 Selection of Sample Volume Often when one needs to analyse biological samples, the volume of sample available for analysis is very limited. Also, as analytical equipment is reduced in scale, and with the developments of lab-on-a-chip devices, additional pressures are seen for the efficient handling of very small sample volumes. Most conventional sample preparation strategies cannot be used with sample volumes on the microlitre scale. The small dimension of SPME devices can be an advantage in this regard.4 So long as a sampling strategy can be devised where the entire surface of the extraction phase is in contact with the sample, reproducible extraction can be achieved. For instance, Li and Weber5 demonstrated the extraction of barbiturates from just 50 μL of sample contained in a 1.5-mm inside diameter (i.d.) Teflon tube. The application of SPME to in vivo analysis has taken advantage of the devices’ abilities to sample from very small spaces (see Chapter 12).6,7 In such cases, sample volume may not be known. While SPME devices are capable of extraction from very small sample volumes, analysts must carefully consider the consequences on calibration. As discussed above (Section 10.2.1), the equilibrium concentration of analyte in the extraction phase is related to the equilibrium or final concentration in the sample phase according to the partition coefficient. Analysts, however, typically calibrate an extraction by plotting the amount extracted versus the initial sample concentration. If a significant proportion of the analyte is extracted, the initial and final sample concentrations can be significantly different. This is particularly true of very small volume samples. In this case, of course, calibration is still feasible, but the sample volume must be tightly controlled. Variation in the sample volume will directly affect analytical precision. Because liquid handling devices intended for microlitre volumes are inherently less precise than those for millilitre volumes, significant effort must be put into volume measurement in these cases. It is often not possible to achieve acceptable precision with manual devices. Automated sample handling is more appropriate in these cases. These considerations point to the importance of an estimation of the conditions of negligible depletion and an understanding of whether sample volume must be controlled for a particular application.

10.2.5.1 Calculation of Negligible Depletion Sample Volume As discussed, partitioning calculations for SPME are based on the final sample concentration rather than the initial sample concentration. However, where the amount extracted relative to the total amount in the sample is negligible (negligible depletion), the initial sample concentration may be substituted. Typically, depletion is considered negligible when sample concentration changes by 10% or less during

340

Handbook of Solid Phase Microextraction

extraction because this is an amount in line with experimental error. Where analyte depletion is not negligible, sample volume must be carefully controlled during extraction, and the analyte concentration in the sample at equilibrium should be calculated and employed in the calibration regression calculation. This will obviously complicate the calibration, so it is beneficial in method development to determine a sample volume that will permit negligible depletion during extraction. The impact of degree of sample depletion on calibration is investigated below.

10.2.6 Effect of Additional Phases in the Sample When additional phases are present in a sample, analyte can also partition into those phases in addition to the sample matrix and the extraction phase. Any discontinuous phase in the sample having a significant affinity for the analyte of interest competes with the extraction phase for the analyte, and therefore it causes a reduction in the amount extracted. Given the complexity in most biological matrices, consideration should be given to the potential for competing phases. Typically, this is evaluated by assessing recovery by the technique of degree of matrix influence, as discussed above (Section 10.2.3). Where matrix interferences are present but predictable and reproducible from one sample to the next, sensitivity may be compromised, but good results can be obtained. However, method precision suffers significantly if the amount of such a competing phase varies from sample to sample. An understanding of the cause of this variability is useful in assessing viable means to reduce it. Where competing phases are present, the mass balance equation can be rewritten to include the impact of each competing phase, as shown below (see also Section 2.3.5 in Chapter 2). The mass balance equation: C0 Vs 5 CeN Ve 1 C1N V1 1 C2N V2 1 C3N V3 1 ? 1 CnN Vn

ð10:2Þ

The amount extracted: nN e 5

Kes Ve C0 Vs Kes Ve 1 K1s V1 1 K2s V2 1 K3s V3 1 ? 1 Kns Vn 1 Vs

ð10:3Þ

An important observation is that at equilibrium, analyte will be distributed to all phases according to their partition (affinity) coefficients and volumes. Obviously, phases with either very low volume or low affinity can be disregarded. SPME will extract only free analyte, so where a competing phase with high volume or affinity is present, less analyte will be extracted. Where this phase is variable from one sample to the next, accurate analysis is compromised. A detailed discussion of these equations has been published.2 In biological drug extractions, significant competing phases could include dissolved protein, tissue fragments, cells or cell fragments or even receptors or antibodies to the analyte of interest. If the sample is a pharmaceutical compound, several competing phases such as excipients may be

Drug Analysis by SPME

341

present in the drug formulation. The analysis of timed-release formulations must be considered carefully.

10.2.7 Temperature A change in extraction temperature has several effects on microextractions. With a temperature increase, diffusion coefficients and Henry’s constants increase, while partition coefficients to the extraction phase decrease. Because Henry’s constants increase, headspace concentrations also increase. Because diffusion coefficients increase, the time required to reach equilibrium decreases. Finally, due to the lowered partition coefficients to the extraction phase, the equilibrium amount extracted decreases. These effects can be observed in Figure 10.1A, which shows the effect of increased extraction temperature on equilibrium profiles. The shortened equilibration times and lowered total amounts extracted with increased extraction temperature are clearly observed. When the data are replotted with extraction temperature on the x-axis (Figure 10.1B), the maximum extraction sensitivity given these competing forces is easily observed for each extraction time. For the same reasons, temperature can also be an important consideration for optimising desorption.9

10.2.8 Ionic Strength The addition of a salt often can improve recovery when conventional extraction methods are used (see also Section 2.3.3.2 in Chapter 2). Sodium chloride (NaCl) is commonly used for this purpose.10
12 Occasionally, one sees an initial increase in the extraction yield with an increase in salt concentration, with a maximum being reached, followed by a decrease in the amount extracted with a further increase in salt concentration. Hall and Brodbelt noted that the effect varied

(B) 500 450 400 350 300 250 200 150 100 50 0

40°C

500

60°C 22°C 73°C

Amount extracted (ng)

Amount extracted (ng)

(A)

B 400 300 200 15 min

100

60 min 0 0

20

40

60

Extraction time (min)

80

100

0

20

40

60

80

Extraction temperature (°C)

Figure 10.1 Effect of temperature on (A) the methamphetamine equilibrium time profile and (B) the amount of methamphetamine extracted. (Source: Adapted from Ref. 8.)

100

342

Handbook of Solid Phase Microextraction

through a series of barbiturate drugs, apparently related to the pKa of the drug.13 Abdel-Rehim et al.14 noted that for ropivicaine and a series of its metabolites, 20% NaCl provided optimal extraction. Higher levels resulted in reduced extraction efficiency. This behaviour can be explained by considering two simultaneously occurring processes. Initially, analyte recovery increases due to a phenomenon called ‘salting out’, whereby water molecules form hydration spheres around the ionic salt molecules. These hydration spheres reduce the concentration of water available to dissolve analyte molecules; thus, it is expected this phenomenon will drive additional analytes into the fibre coating.15 In competition with this process, however, it is the fact that molecules may participate in electrostatic interactions with the salt ions in solution,16 thereby reducing their ability to move into the fibre coating. Initially, it would be the interaction of the salt molecules with water, which is the predominant process. As salt concentration increases further, salt molecule interactions with analyte molecules will become significant. Thus, it is reasonable that analyte extraction initially increases with increasing salt concentration, followed by a decrease as the salt interaction with the analytes in solution predominates.

10.2.9 Extraction pH The pH of the extraction mixture is important for drugs possessing a pH-dependent dissociable group (see also Section 2.3.3.3 in Chapter 2). It is only the undissociated form of the drug that will be extracted by an absorptive fibre coating, such as polydimethylsiloxane (PDMS) or polyacrylate (PA), or extraction processes governed by non-polar associations in solid sorbent phases. This is important for drug extraction because any drug that has partitioned into the fibre coating cannot participate in the Henderson-Hasselbalch equilibrium between the acid and base forms of the drug in the sample. In the case where the extraction mixture pH is controlled with a buffer, as the undissociated form of the drug is extracted by the fibre, more dissociated drug re-associates and is therefore available for extraction. Thus, in a buffered extraction mixture, more drug can be extracted by an absorptive fibre coating than in an extraction mixture where the pH is not buffered. In an unbuffered extraction mixture, the ratio of undissociated to dissociated forms of the drug can vary. Therefore, one does not achieve the continual transfer of drug from the dissociated to the undissociated form and then to the fibre coating. This effect is seen in Figure 10.2, which compares the amount of methamphetamine extracted from a solution where the pH is base adjusted with KOH versus one where the pH is controlled at pH 12 with a phosphate buffer. Note that there is a significant deviation from linearity in the amount extracted as drug concentration increases in the non-buffered system. Queiroz et al.17 provide a good description and discussion of the effect of pH variation on a series of drugs used therapeutically in the control of seizures. Many drug formulations designed to be taken orally utilise the salt of the active ingredient to aid dissolution in the stomach. Control of extraction pH is important in such cases to ensure that the undissociated form of the drug is present for extraction. In addition, the formulation matrix of a pharmaceutical preparation may itself

Drug Analysis by SPME

343

1,400

Mass extracted (ng)

1,200 1,000 800 600 400 200 0 0

1

2

3

4

5

Mass of spike (µg)

Figure 10.2 Effect of buffering on methamphetamine extraction linearity. Key: open squares, pH 12 buffered samples; closed diamonds, sample adjusted to pH 12 with KOH. (Source: Adapted from Ref. 8.)

act to buffer the extraction matrix, and final extraction pH should be checked before extraction.

10.3

Fibre Selection: Adsorption Versus Absorption

In the extraction of analytes between a sample matrix and a liquid sorbent, the coating solvates analyte molecules partitioning into the extraction phase. The diffusion coefficient in the liquid coating typically allows the molecules to penetrate the whole volume of the coating within a reasonable extraction time. In this case, extraction varies only if the coating property is modified by the extracted components, which occurs only when the amount extracted is a substantial portion (a few percent) of the extraction phase. This is very rarely observed because SPME is typically used to determine trace concentrations of analytes. Extraction calibrations are linear over a wide range of analyte concentrations. In the case of solid sorbents (either crystalline or amorphous), the coating has a high density, which substantially reduces the diffusion coefficients within the structure. Therefore, within the experimental time, the extraction occurs only on the surface of the coating at specific active sites, and there is only limited surface area available for adsorption. If this area is substantially occupied, then saturation and displacement effects occur and the equilibrium amount extracted can vary with concentrations of both the target and other analytes or interferences. Extraction calibrations will deviate from linearity when a significant proportion of the active sites are occupied.18,19

344

Handbook of Solid Phase Microextraction

For all their limitations, one might expect that it is impractical to use solid sorbents at all. On the contrary, solid sorbents have a significant advantage over liquid extraction phases, in that the affinity coefficient (K) for solid sorbents is usually much higher than the partition coefficient (Kes) for liquid extraction phases,4 which typically allows for better sensitivities with solid sorbents.

10.3.1 Partitioning: Immobilised Liquid Fibres For absorptive SPME fibres, extraction phases are immobilised liquids, so the process is analogous to LLE. By the law of conservation of mass, we know that the initial amount of analyte present in the sample will be equal to the sum of the individual amounts of analyte present in all discontinuous phases. This is expressed mathematically as n0 5 ne 1 ns 1 n1 1 n2 1 ?

ð10:4Þ

where n0 is the mass of analyte initially present in the sample, ne is the mass of analyte present in the extraction phase, ns is the amount present in the homogeneous liquid phase and n1, n2 . . . are the amounts present in discontinuous phases. It should be noted that a discontinuous phase is any compartment of a sample with a different affinity for the analyte of interest. This may include both distinct phases, such as immiscible liquids, solids or a headspace, and dissolved phases, such as proteins in solution. At equilibrium, or in fact any time before equilibrium, n0 is equal to the sum of the amounts of analyte present in the sample (ns) and the extraction phase (ne) and any other phases present. A mathematical consideration of the partitioning process provides both a demonstration of the linear dependence of amount extracted on initial sample concentration and an understanding of the impact of other sample variables. The amount extracted by the extraction phase may be derived from Eq. (10.4). To simplify the discussion, it is assumed here that the only significant phases present are the sample matrix and the extraction phase: n0 5 n s 1 ne

ð10:5Þ

Because the partitioning process depends on a concentration equilibrium, concentrations in each of the phases must be introduced into Eq. (10.5): C0 Vs 5 CeN Ve 1 CsN Vs

ð10:6Þ

In Eq. (10.6), C0 is the initial concentration in the sample, Vs is the sample volume, CeN is the concentration in the extraction phase at equilibrium, Ve is the volume of the extraction phase and CsN is the sample concentration at equilibrium. The extraction phase/sample distribution constant for the partitioning is defined as shown here: Kes 5

CeN CsN

ð10:7Þ

Drug Analysis by SPME

345

The amount of analyte extracted by the fibre can be calculated by combining Eqs (10.6) and (10.7) and then rearranging to obtain Eq. (10.8). At equilibrium, therefore, the amount extracted is described as20 nN e 5

Kes Ve C0 Vs Kes Ve 1 Vs

ð10:8Þ

In Eq. (10.8), the equilibrium amount extracted does not depend on extraction time. In practice, this means that once equilibrium is reached, the extracted amount is constant within the limits of experimental error and is independent of any further increase in extraction time. In addition, if VscKesVe (capacity of the sample phase is B10 times larger than the capacity of the extraction phase), Vs cancels out, and the amount extracted no longer depends on sample volume. Equation (10.8) then becomes nN e 5 Kes Vf C0

ð10:9Þ

The foregoing discussion highlights two important considerations for drug analysis. First, sample volumes for drug analyses are commonly on the order of 1 mL. Partition coefficients for drugs into PDMS are typically low. Literature values indicate a lidocaine partition coefficient of 3.3 at a pH allowing for significant ionisation21 and amphetamine/methamphetamine partition coefficients of 80 and 350, respectively, at a pH favouring completely non-dissociated species.12 Given a 100-μm PDMS fibre extraction phase volume of 0.6 μL, KesVe values would be 2 μL, 50 μL and 200 μL, respectively, for the partition coefficients described. Thus, sample volumes would have to be on the order of 20 μL, 500 μL and 2 mL, respectively, to permit the use of Eq. (10.9). Clearly, if 1-mL sample volumes are used, researchers should take care to determine whether or not it is valid to use the simplified Eq. (10.9) for calibration. If it is not, sample volume must be controlled carefully to avoid undue error in analyses.

10.3.2 Solid Sorbent Fibres Solid sorbent fibres were first introduced commercially by Supelco in about 1998. Their utility for analysis of more polar compounds from aqueous matrices was quickly exploited.22 Several versions of fibres based on a styrene/DVB polymer were made available. For GC applications, both Carbowax/DVB and PDMS/DVB were produced. In both cases, the significant extraction phase was DVB, with the Carbowax or PDMS acting as a glue to hold the solid sorbent to the surface. The nature of the glue, however, did impart a degree of selectivity to the fibres, with the Carbowax/DVB performing somewhat better for polar substances and the PDMS/DVB for non-polars. For high-performance liquid chromatography (HPLC) applications, Supelco introduced the Carbowax/TPR fibre. In this case, the TPR extraction phase was chemically the styrene/DVB polymer but polymerised as a

346

Handbook of Solid Phase Microextraction

templated resin to have a tighter distribution of pore sizes. Given the relatively low partition coefficients for drugs with the PDMS coating and higher amounts extracted with the solid sorbents, early research on these new fibres demonstrated the benefits for extraction efficiency of the solid sorbent extraction phases. In 1998, Koster et al.21 reported on the use of the 100 μm PDMS fibre for the analysis of lidocaine from urine. The authors observed a GC/FID limit of detection (LOD) of 5 ng/mL. At that time, the solid sorbent fibres were just being introduced, and although the authors pointed out their potential advantage, they were not investigated in that publication. In 2000, the same group reported on the use of the PDMS/DVB fibre for GC analysis of lidocaine23,24 and noted an approximately sixfold improvement in amount extracted over PDMS, although the solid sorbent fibres had a significantly longer equilibration time24,25 and shorter lifetime26,27 than the PDMS fibres. Others have noted that among the solid sorbent fibres, the CW/TPR generally has the best performance and durability,28 although it has also been noted that it is less selective than the PDMS/DVB fibre,29 potentially resulting in more complicated chromatography. Given their widespread use for extraction of therapeutic drugs, discussion of solid sorbent extraction theory, in relation to that for liquid sorbents, is warranted. Commercial extraction phases incorporating DVB are amorphous solids having negligible diffusion of analyte within the polymer matrix. Analytes bind to surface active sites in an exothermic adsorption process with an activation energy associated in the transfer of analyte between the solution and the coating. Sorbent saturation and displacement effects need to be considered. Gorecki et al.19 have provided a helpful description of the binding isotherm for solid sorbents. The following description of the processes draws on that work. Adsorption generally follows a Langmuir isotherm, with the following assumptions: (i) molecules adsorb into an immobile state, (ii) all sorption sites are equal, (iii) each site can hold only one molecule and (iv) there are no interactions between adsorbed molecules on adjacent sites. The equilibrium amount of analyte on an adsorptive fibre is determined as follows: N nN f 5 C f Vf 5

KCs0 Vs Vf ðCf max 2 CfN Þ Vs 1 KVf ðCf max 2 CfN Þ

ð10:10Þ

where Cf max is the maximum concentration of active sites on the coating, CfN is the equilibrium concentration of analyte on the fibre, K is the analyte’s adsorption equilibrium constant (affinity constant) and Cs0 is the initial concentration of analyte in the sample. The form of Eq. (10.10) is quite similar to that for Eq. (10.8), where the coating extracts by absorption rather than adsorption. The main difference is the presence of the fibre concentration term ðCf max 2 CfN Þ in both the numerator and denominator of Eq. (10.10), which depends on the sample concentration of analyte. This results in a non-linear dependence of the amount extracted relative to sample concentration. It should also be noted that in Eq. (10.10), K is different from Kes in Eqs (10.8) and (10.9). Kes is the partition coefficient, whereas K is the adsorption

Drug Analysis by SPME

347

equilibrium constant. From the equation for the reaction of an analyte ‘A’ with an active site ‘S’ (Eq. (10.11)), the affinity coefficient ‘K’ is derived as shown in Eq. (10.12). ½S 1 ½A-½SA

ð10:11Þ

½SA ½S½A

ð10:12Þ

K5

From Eq. (10.12), we see that the concentration of unoccupied sites is significant in the calculation of affinity. For very low analyte concentrations on the fibre, i.e. where CfN is negligible compared to the total number of active sites, Cf max cCfN , Eq. (10.10) reduces to Eq. (10.13) and the dependence on sample concentration will be linear. N nN f 5 C f Vf 5

KCs0 Vs Vf Cf max Vs 1 KVf Cf max

ð10:13Þ

Equation (10.13) is comparable to Eq. (10.8) for absorptive phases except that the phase capacity term is KVfCf max, and as was seen for the absorptive phases, a determination of volume independence may be made where VscKVfCf max. In this case of negligible depletion, Eq. (10.13) reduces to Eq. (10.14): N 0 nN f 5 Cf Vf 5 KCs Vf Cf max

ð10:14Þ

In practice, it is very rare that only one analyte or compound with affinity for the fibre is present in a sample, and because adsorption is a competitive process, the presence of another compound (B) will affect the amount of analyte (A) extracted by the fibre. The equilibrium concentration of analyte A on the fibre in the presence of a competing compound B is given by the following equation: N 5 CfA

N Cf max KA CsA N N 1 1 KA CsA 1 KB CsB

ð10:15Þ

KA and KB are the adsorption equilibrium constants for compounds A and B, N N and CsB are the equilibrium concentrations of A and B in the respectively, and CsA sample. This can be rearranged to give the amount extracted by the extraction phase (fibre) at equilibrium: V nN fA 5 CfA Vf 5

N CA0 Vs Vf KA ðCf max 2 CfA Þ N NÞ ð1 1 KB CfB ÞVs 1 KA Vf ðCf max 2 CfA

ð10:16Þ

Thus, the amount of analyte A extracted at equilibrium in the presence of a competing compound B must be lower than for the situation where no competing

348

Handbook of Solid Phase Microextraction

compound exists, as the additional term in the denominator of Eq. (10.16) relative to Eq. (10.10) can only be greater than 1. If, however, the interfering compound is either present at a very low concentration or has low affinity to the coating, this first term may be insignificant, and there will be little difference in the amount of analyte A extracted between samples with and without the presence of interfering compound B. The practical application of the foregoing for drug analysis is that samples are almost always complex mixtures where multiple compounds could compete for binding sites on a solid sorbent. It is preferable that the analyte of interest has a high affinity for the sorbent. Where the analyte of interest has a low affinity for the sorbent, precision can suffer due to either displacement effects or a non-linearity of response. Care must be taken in selecting a suitable method or IS. With all their limitations, it would seem that solid sorbents would be significantly less desirable to use than immobilised liquid sorbents. In practice, however, one typically finds that solid sorbent extraction efficiencies are significantly higher than those of liquid sorbents.4,23,24 This advantage often makes their use desirable despite their limitations.

10.3.3 Pre-Equilibrium Extraction for Quantification of High Sample Concentrations As described above, when the capacity of a solid sorbent is approached, the calibration of extracted amount related to sample concentration deviates from linearity. Displacement of analytes with lower affinities may also occur.18,30 Thus the amount extracted can vary with concentrations of both the target and other analytes. This situation is, of course, more problematic with higher sample concentrations, significantly complicating calibration by equilibrium extraction in this case. For this reason, solid sorbents should be used only for trace analysis. A solution to this dilemma is to employ pre-equilibrium extraction times. In this case, the extraction process is governed by extraction kinetics rather than equilibrium states.31 This is described in more detail in Chapter 6 of this book.

10.4

Considerations of Drug Properties

10.4.1 Volatile Drugs Drugs that have inherent volatility are typically analysed with the same techniques as fragrance and environmental compounds. As for the terpenes in fragrance analysis or the benzene-like compounds in environmental samples, these analytes are easily extracted by either headspace or direct immersion extraction, using conventional extraction phases, primarily PDMS. Diffusion coefficients are typically high and polarity is low, resulting in high partition coefficients. Extraction rates are similar between direct immersion and headspace, as there is typically not a significant headspace concentration limitation. An example is the analysis of halothane described by Musshoff et al.32 The method employed HS-SPME with GC/MS to

Drug Analysis by SPME

349

determine halothane in the headspace of blood and other forensic samples. Total analysis time per sample, including extraction, separation and detection, was 40 min.

10.4.2 Presence of Ionisable Groups As discussed above (in Section 10.2.9), many drug compounds have pH-dependent ionisable groups. Amine-containing basic drugs, such as the amphetamines, have a neutral charge state only at very high pH, whereas acidic drugs, such as ibuprofen, are neutral at acidic pH. Ionised species have a very high affinity for aqueous medium and so are poorly extracted by most sorbents. pH adjustment to favour the non-ionised form is typically required to achieve sufficient sensitivity. It should be noted that PDMS will dissolve from the fibre if immersed in a very basic solution (Table 4.4). If adjustment to a highly basic condition is required, headspace extraction may be necessary.12 It is also possible to improve the extraction efficiency of ionised analytes by incorporating an ion-pairing reagent in the sample matrix. Ion pairing allows for the extraction of ionised compounds by partitioning, by combining the analyte ions with counter-ions of opposite charge.33 The inclusion of an ion-pairing reagent in the mobile phase is commonly used to improve HPLC separations. An example is the use of this strategy for the LC separation of antiretroviral drugs extracted from biological fluids by SPE.34 Several articles have appeared in recent years on the use of ion pairing to improve extraction of polar drugs by SPE,35,36 LLE,37 liquidphase microextraction (LPME)38 and liquid-phase membrane extraction.39,40 A few articles have appeared describing this technique’s application to SPME. Most of these do not describe the analysis of drugs, but the methods developed could be instructive in the development of drug analyses. Pan and Pawliszyn41 proposed the use of ion-pair SPME to convert long-chain fatty acids into their methyl esters using in-port derivatisation. This same strategy was employed for the analysis of alkylbenzene sulfonates from water samples using the 100-μm PDMS fibre.42 The ion-pairing reagent employed was tetrabutylammonium hydrogen sulphate (TBAHSO4). In this case, the counter-ion association both promoted the extraction of these ionic compounds using the non-polar PDMS sorbent and facilitated the conversion of the analytes to their butyl esters in the hot injector port of the GC. In an example more closely related to drug analysis, the shellfish toxin saxatoxin was analysed in water samples by SPME-HPLC using post-column derivatisation with fluorescence detection. In the method development, desorption of the toxin from the CW/TPR fibre was observed to have poor efficiency. The problem was resolved by employing static ion pair desorption with heptane sulphonic acid.43 Li et al.44 recently employed in situ derivatisation with ion pairing for the analysis of methylmalonic and glutaric acid in urine as a clinical tool for the diagnosis of organic academia. The organic acids were converted to their ethyl esters using diethyl sulphate. The esters were then easily extracted from headspace using PDMS fibres, and GC-MS was employed for separation and detection. In this work, the ion-pairing reagent TBA-HSO4 was included in the derivatisation

350

Handbook of Solid Phase Microextraction

solution to activate the analytes during derivatisation and increase the derivatisation yield.45,46 A similar strategy was employed for the analysis of haloacetic acids in water.47

10.4.3 Non-Volatile or Semi-Volatile Drugs Drugs with low volatility have low headspace concentrations. They may also be polar and have strong interactions with aqueous sample matrices. Because headspace diffusion coefficients are much higher than liquid diffusion coefficients, such drugs may be extracted by a sorbent located in the headspace faster than the headspace concentration is replenished by mass transfer from the sample. In this case, either direct immersion extraction must be employed or the compound’s affinity for the headspace or fibre must be increased. Headspace affinity may be improved by the addition of salt to the sample as was discussed previously (in Section 10.2.8) or by heating the sample to increase the Henry’s constant and move more analyte to the headspace. If these options do not provide sufficient extraction, the analyte may also be derivatised to a form with a higher volatility or lower polarity.

10.5

Calibration

In a microextraction, the extraction phase should always extract the same proportion of analyte from a sample. Where sample volume is not limited, the amount extracted in a microextraction should be linearly correlated with sample concentration, with the exception described above for adsorptive phases used to extract complicated or high-concentration samples. For quantitative work, one must calibrate the amount extracted relative to the sample concentration. This SPME response calibration is distinct from the calibration of detector response that must be completed for any analysis, although in practice they are often conducted simultaneously. There are a number of options available for calibrating a microextraction analysis (see Chapter 6). External calibration is perhaps the simplest and most widely implemented calibration technique. In other techniques for chemical analysis, a series of calibration standards are prepared in a simple matrix such as water or solvent. These are then processed similarly to unknown standards. It is assumed that the sample matrix does not significantly interfere with the extraction process. The degree to which response from calibration standards differs from response from samples is reported as recovery rate for various sample matrix types. Extraction from the simple matrix is taken as 100%. If sample matrix hinders the extraction of analyte, a recovery of less than 100% is reported for that matrix, and conversely if matrix enhances extraction, a recovery of more than 100% is reported. Sample results are adjusted based on the recovery rate in order to improve accuracy in the reported sample concentration. Typically recoveries of 615% are tolerated.

Drug Analysis by SPME

351

This technique is typically not practical for use in SPME, where the matrix can have a very significant impact on the fundamentals of analyte extraction. Thus, in SPME experiments, it is quite important to use matrix-matched standards for calibration. In this case, microextractions are performed under near-equilibrium conditions and calibrations are calculated for each analyte. In practice, linear calibrations are the simplest, but non-linear calibrations are also possible. As discussed, it is important that calibration samples have the same matrix composition as test samples: if the matrix in test samples differs from that of calibration samples, partitioning will not be the same and the calibrations will not be valid. In addition, where the amount extracted depends on sample volume, this volume must also be closely controlled. If the sample headspace exhibits a significant affinity for the analyte, the headspace volume must also be constant, regardless of whether headspace or direct sampling is performed. For complex matrices such as tissue, external calibration is typically impractical. The variability in sample matrix from one sample to the next will cause unacceptable imprecision in analyses. In addition, it may be impossible to simulate or spike calibration standards. For solid sorbents, equilibrium extraction with external calibration is feasible where the concentration of analyte in the sample does not cause a saturation of the active sites in the sorbent. In order to achieve linear calibrations, extractions should be designed so that only a small proportion of the active sites on the sorbent are occupied at equilibrium. In this case, either Eq. (10.13) or (10.14) may be used for calibration, depending on whether negligible depletion of sample is achieved. Where active sites are significantly occupied, Eq. (10.10) will apply and non-linear response will be observed. Displacement effects may also occur in this situation, so it should be avoided.

10.5.1 Calibration Based on Surface Adsorption Reaction Kinetics One special calibration case, which was not discussed in Chapter 6, is calibration based on surface adsorption kinetics. In most cases, the overall rate of extraction for pre-equilibrium extraction is controlled by the diffusion of analyte through the boundary layer, as this is typically the slowest of all the processes involved. In some cases, however, the diffusion rate may be unusually fast or the adsorption reaction unusually slow, and the overall extraction rate is controlled by the reaction rate. This may be observed where extraction involves simultaneous adsorption and derivatisation, or reaction with a surface immobilised recognition agent such as an antibody. In this case, the overall extraction rate is expressed as shown in Eq. (10.17)4: Frxn 5 n=t 5 Vf kad KCs

ð10:17Þ

where Frxn is the rate of mass transfer to the sorbent controlled by the reaction kinetics. Vf is the sorbent phase volume, kad is the adsorption rate constant, K is the adsorption affinity constant and Cs is the analyte concentration in the sample. The

352

Handbook of Solid Phase Microextraction

overall extraction rate will be optimal for calibration when Cs is matched to affinity (KCs $ 1). An example of a pre-equilibrium extraction rate controlled by interaction with an immobilised antibody has been described for benzodiazepine analysis.1 In this case, it was observed that when no stirring was employed during extraction, the extraction rate was controlled by diffusion through the boundary layer. For any amount of agitation in the sample, the extraction rate was controlled by the drugantibody reaction rate.

10.5.2 Endogenous Presence in Matrix In therapeutic and forensic drug monitoring, the establishment of a ‘clean’ blank is an important aspect of any method development. It is expected that a blank analysis is shown where no peak is observed in the detector signal at the analyte retention time, on a y-axis representative of a signal-to-noise ratio of 3:1. If this cannot be shown, calibration is questioned. In some cases, however, it is not possible to obtain a blank matrix for calibration purposes. This is observed, for instance, in the analysis of herbal preparations from which active ingredients are derived or in the analysis of biological matrices for endogenous small molecules such as steroids or small-molecule products of disease or metabolism. Despite this challenge, strategies are available to address calibration in such cases. Where matrix is available with a low, consistent level of analyte, the endogenous presence may be calculated by standard addition. Then the endogenous amount may be either subtracted from unknowns or added to the amounts of calibration standards to produce a reconstructed external calibration.

10.6

Novel SPME Coatings for LC

Much of the effort around development of novel extraction phases for SPME has been spurred by the need to extract polar and semi-polar molecules from aqueous (biological) matrices more efficiently. The early SPME investigations of forensic drugs successfully employed the PDMS-coated fibres with GC separation, due to the relatively low polarity of most of those compounds. Many drugs cannot be analysed by GC, however. When efforts turned to optimising SPME-LC methods, researchers quickly realised that the somewhat more polar DVB-based fibres were far superior to any of the other commercially available fibres. Because only one commercially available phase performed well for these analyses, the need for more appropriate phases for these analyses became evident. Supelco is now starting to introduce a new line of coatings suitable for SPME-LC, as discussed in Section 4.3.5 in Chapter 4. Several experimental solid sorbent fibres have been reported for specific applications of analysis of polar molecules. For example, ion exchange coatings

Drug Analysis by SPME

353

were used to remove metal ions and proteins from aqueous solutions,18,48 liquid crystalline films were used to extract planar molecules, Carbowax for polar analytes,49 metal rods to electorodeposit analytes,50,51 pencil ‘leads’ to extract pesticides52 and Nafion coatings to extract polar compounds from non-polar matrices.53 Polypyrrole (PPY), used initially for biosensors,54 and restricted access materials (RAMs), originally designed as SPE packings,55 have specifically been investigated for the analysis of drugs from biological matrices. In addition, the development of immunoaffinity fibres has been described, which should allow the application of SPME technology for the extraction of large biomolecules from complex matrices.56,57 Some of the findings from these efforts are summarised below (see also Section 3.4.1 in Chapter 3).

10.6.1 Restricted Access Materials RAMs are particulate sorbents produced by Merck particularly for use in SPE cartridges. They feature an outer surface that is effective at repelling polar biomolecules and pores sufficiently small enough to exclude biomolecules from the inside of the particle. The extraction phase, however, is located exclusively inside the pores. In this way, biomolecules are excluded from the extraction surface. The extraction phase, typically either C-18 or an ion exchange phase, is ideally suited for extracting analytes typically present in biological matrices such as urine or blood. The particle size is similar to the DVB phases already used for SPME. Hence, these phases were evaluated for use in SPME analysis of therapeutic drugs from biological matrices.58
60 While the fibres performed well, equilibration times were very long. This was likely due to the time for diffusion into the interior of the particles and within the C-18 liquid extraction phase. Regardless, the sorbent holds promise for high sensitivity analysis of therapeutic drugs where extraction time is not the primary concern.

10.6.2 C-18 Given the good success seen with the C-18 extraction phase of the RAM particles, it was desirable to determine if a C-18 coated fibre would offer promising results. Because C-18 phases extract absorptively, they are of particular interest for the extraction of polar analytes, which has to date been dominated by the solid sorbents. Initially, technical details for the preparation of such a sorbent directly on a fibre support delayed the investigation. Recently, SPME fibres have been prepared by coating silica particles themselves coated with C-18 onto a stainless steel wire by use of polyethylene glycol (PEG) glue.61 The properties of these fibres were investigated, with the observation that they have good stability, reproducibility and linear range, and that they may be useful for in vivo SPME applications due to biocompatibility of both materials.

354

Handbook of Solid Phase Microextraction

10.6.3 Sol
Gels Sol
gels have been of interest in chemical separations for at least the past decade due to their relative ease of preparation, high stability and reproducibility of performance. Early reports described their use in GC capillaries with PDMS sol
gel,62 CE capillaries with polyalkylene glycol sol
gel63 and SPME fibres with PDMS sol
gel.64 The sol
gel SPME fibres were used for the extraction of polynuclear aromatic hydrocarbons (PAHs), alkanes, anilines and alcohols from water. In 2000, Wang et al.65 demonstrated the use of SPME fibres coated with sol
gel incorporating PEG for the extraction of a range of compounds, including benzene, toluene, ethylbenzene and xylenes (BTEX), phenols, phthalate esters and aromatic amines. The porosity of the resulting phase was studied, and the fibres were demonstrated as useful for extraction of compounds with a range of polarities. In 2005, Bagheri et al.66 described the preparation of SPME fibres incorporating all three strategies (PDMS, polyalkylene glycol and PEG) and their use for the extraction of dextromethorphan and dextrorphan from plasma. In this work, headspace extraction was employed and the polar dextrorphan was derivatised prior to extraction to improve its volatility. Interestingly, the PDMS phase was preferred for these relatively polar compounds. The PEG coating demonstrated a low level of extraction, likely due to its low porosity. The polyalkylene glycol coating extracted more analyte than the PDMS coating but at a longer (.30 min) extraction time than for PDMS (30 min at 60 C). The good performance of PDMS was explained in that although conventional PDMS is quite non-polar, sol
gel PDMS contains a significant number of residual hydroxyl groups after polymerisation, imparting a more polar nature to the sorbent. Recently two articles have appeared describing the use of a calixarene sol
gel SPME fibre for the extraction of propranolol from urine with separation by GC67 and CE.68 The phase was shown to be highly base- and temperature-resistant and was used successfully for extraction by both direct and headspace SPME from highly basic matrix, with heating up to 100 C during extraction and 280 C during desorption in the GC injector port. For coupling to CE, back extraction of the absorbed analytes into 20% acetonitrile, 80% water at slightly acidic pH (6.2) was used to promote protonation of the analyte. Both articles describe excellent reproducibility and durability of the fibres. The same group has also investigated the use of a butyl methacrylate sol
gel extraction phase for the analysis of ephedrine derivatives from water and urine with CE separation.69 For extraction, 5 mL of sample was placed into a 10-mL vial. Optimal headspace extraction conditions included 2.0 g sodium hydroxide and 0.5 g sodium chloride added to the vial, with stirring and heating at 90 C during the 30-min extraction. Back extraction was optimised at 20 C, 20% acetonitrile, for 20 min. The methacrylate fibre performed the best among commercial PDMS and PA fibres and a sol
gel fibre prepared without methacrylate. The methacrylate fibres extracted about twice as much as the two commercial fibres. The sol
gels had good solvent stability, extraction efficiency and durability. Limits of detection were in the low ng/mL range with linearity between 20 and 5,000 ng/mL.

Drug Analysis by SPME

355

The in-tube SPME or capillary microextraction format has also been investigated with sol
gel phases. In 2002, Bigham et al.70 compared sol
gels prepared with PDMS and PEG for extraction of a range of analyte classes, with thermal desorption and separation by GC. PAH, aldehydes and ketones were analysed using the PDMS phase, while alcohols, phenols and amines were analysed with the PEG phase. They observed parts per trillion to parts per quadrillion detection sensitivities. More recently, Fan et al.71 have described the preparation and use of a sol
gel phase incorporating β-cyclodextrin coupled to HPLC for the analysis of the nonsteroidal anti-inflammatory drugs ketoprophen, fenbufen and ibuprofen from urine. A similar phase was described for the fibre SPME extraction of ephedrine and methamphetamine from urine72 by the same group that described the above analyses of propranolol and ephedrine. While the experience with sol
gel fibres for drug analysis is somewhat preliminary at the current time, the ruggedness, reproducibility and customised properties appear to make it a most interesting phase for SPME drug analyses of the future.

10.6.4 Monolithic Extraction Phases Several reports have appeared recently describing the use of polymer monoliths prepared in fused silica capillaries for sample preparation. Although the technique of in-tube SPME is quoted, the authors employ high sorbent capacity and high extraction efficiency. In some cases, near-exhaustive extraction is accomplished and the amount extracted is proportional to the total analyte mass in the sample volume (as per SPE), not initial sample concentration (as per SPME). In some cases, it is not possible to determine with certainty if exhaustive or equilibrium extraction is being used because the authors do not always report extraction efficiency or whether analyte breakthrough is occurring with the sorbent. However, given the general strategy presented of maximising extraction efficiency, it is likely valid to consider these techniques as miniaturised SPE rather than SPME. Regardless of whether a particular extraction is considered as exhaustive or nonexhaustive, the sorbents employed could potentially be employed in either strategy. The sorbents have proved beneficial for variety of bioanalyses, including those of amphetamines in urine,73 ketamine in urine,74 camptothecin in plasma,75 telmisartin in tissues,76 antiotensin II receptor agonists in plasma and urine77 and theobromine, theophylline and caffeine in serum.78 Like the sol
gel phases, the high ruggedness and degree of tunable selectivity make these phases of significant interest in new device development.

10.6.5 Polypyrrole The PPY-based conducting polymers have been used extensively in the biosensor industry, so good data are available on both their biocompatibility and on the abilities of these polymers to extract analytes with polarities similar to those of therapeutic drugs. In addition to its use in sensors for neuroscience applications,79
82 PPY has been recently investigated for its utility as an extraction phase for SPME

356

Handbook of Solid Phase Microextraction

applications.83,84 It was shown to extract as an adsorptive phase.85 In addition to its low toxicity,86 it is insoluble in all solvents, tolerates elevated temperatures and can be used in many chemical environments.87 The known biocompatibility of PPY made it an interesting candidate for initial efforts at in vivo SPME.7 The commercial SPME fibres were unsuitable for this application, due to several factors including their large size, incompatibility with sterilisation processes and large volume of extraction phase, which caused long extraction and desorption times. When PPY was electrochemically polymerised onto the surface of a fine (100-μm) stainless steel wire, the coating was observed to be 10 μm thick, making for a probe with overall diameter of less than 130 μm. The reduced capacity allowed for fast extraction and desorption times (,5 min and ,20 s, respectively, under optimised conditions) but still allowed for quantitative analysis of many drugs in their therapeutic ranges in circulating intravenous blood. Several challenges also attended these efforts. The first was the relatively fragile nature of the polymer and difficulty producing a suitably rugged probe. This appeared to be related to the poor quality of commercially available pyrrole monomer and the difficulty in distilling this to high purity in quantity. Other challenges included the difficulty in obtaining highly reproducible performance between probes and somewhat cumbersome nature of their preparation. Further development of these probes will likely be hampered until the challenges of coating preparation and quality can be addressed. Even so, a significant body of work has been amassed on the subject.54,88
92 The use of conductive polymers, such as PPY, can introduce additional selectivity during extraction by exploiting the electrochemical processes associated with coating properties. Electrons can be supplied to produce redox processes in the coating and convert analytes to more favourably determined species. This has similarity to the derivatisation processes discussed below. In this application, the rod and the polymeric film must have good electrical conductivity. This strategy has been explored for the analysis of some redox active species, including glutamate and dopamine, which are of interest in neuroscience research.93 An important advantage of the technique is the additional extraction selectivity possible by careful control of electrochemical potentials during extraction and desorption. Finally, PPY may also extract as an anion exchange sorbent because of the anionic counterions present in the polymer structure to balance its net positive charge.84 In this work, PPY incorporating perchlorate as a counter-ion was chemically polymerised on the inner wall of an in-tube SPME capillary, and a variety of small anions were determined from aqueous solution, including selinite, selinate and arsenate.

10.6.6 Immunoaffinity The development of selective and sensitive coatings, based on antibodies or antibody binding mimics, is seen as an important goal. SPME devices with immobilised antibodies have been investigated for the highly specific analysis of smallmolecule drugs from complex biological matrices. The fibres were found to be

Drug Analysis by SPME

357

useful for the very high-sensitivity drug analysis, having some of the lowest limits of detection of any SPME device but were limited in linear range and the effort required to prepare them. Immunoaffinity sorbents behave as solid sorbents with a finite number of ‘active sites’ and so will follow essentially the same calibration strategies, with the exception that the active sites may exhibit some heterogeneity. The interested reader is referred to a recent publication for a more detailed description of calibration with immunoaffinity SPME sorbents.3 An immunoaffinity SPME coating was initially prepared via the covalent immobilisation of a theophylline antibody onto a fused silica fibre surface. This surface was modified with 3-aminopropyl triethoxysilane (APTES) and glutaraldehyde for the selective extraction of theophylline from serum samples.56 The specificity of the immunoaffinity SPME fibre was investigated using a fixed concentration of 3 H-theophylline together with various amounts of interference, possessing no crossreactivity with the theophylline antibody. No significant non-specific binding (NSB) was observed. The reproducibility of the fibre preparation and the immunoaffinity SPME analysis were also investigated, resulting in a percent relative standard deviation (% RSD) of 6.1% for five analysis of the same fibre. The antigen
antibody binding isotherm was obtained by analysing various concentrations of theophylline standards (0.005
5 ng/mL) until the saturation values were reached. Initial binding of theophylline was linear with a regression coefficient (r2) of 0.968. The cross-reactivity of the theophylline immunoaffinity SPME fibre for the structural analogue caffeine was investigated by adding various amounts of caffeine in the presence of theophylline at a saturation concentration, and it produced a low cross-reactivity value of 0.1%. Validation of the coating was completed with the successful analysis of theophylline-spiked serum samples (10 and 50 ng/mL). A limitation of this work was that all analyses had to be done with radioactive analytes because the probe did not have sufficient sensitivity for analysis by LC/MS. Subsequently, a technique for purifying the active fraction of a polyclonal antibody was developed. When probes with these antibodies were prepared using an optimised immobilisation strategy, capacity and sensitivity were sufficient for analysis of diazepam and related benzodiazepines from biological fluids by LC/MS/MS.3,57 Monoclonal antibodies performed similarly to purified polyclonals for extraction from aqueous samples. Another important advance described in this work was the development of a routine for drug desorption that maintained antibody activity. The probes could be reused numerous times (.25) and were sufficiently stable that performance was acceptable for at least 6 months with storage under refrigeration. The influence of NSB was evaluated for these probes, with the result that NSB was found to be insignificant for the probes when used in their dynamic range. Above their dynamic range, a high proportion of drug binding was due to NSB, so selectivity was lost. Antibody affinity was evaluated both before and after immobilisation. Immobilised antibodies had specific affinities in the range of 109
1010 M21 with no reduction in affinity observed on immobilisation. Crossreactivity was evaluated both for a range of benzodiazepines and for an unrelated molecule (erythromycin). Calibration was successful for single analytes at sorbent

358

Handbook of Solid Phase Microextraction

saturations of up to 50%. Because of the non-linear nature of the adsorption isotherm, non-linear regression analysis was required for calibration. Accurate calibration of one benzodiazepine in the presence of another, particularly for evaluation of a lower affinity drug in the presence of a higher affinity drug, was problematic. This issue was partially resolved by the use of an antibody with higher specificity for one member of the benzodiazepines class over the others. For analysis of benzodiazepines individually or in the presence of erythromycin, limits of detection were ca. 0.001 ng/mL and the dynamic range (based on 80% antigenic site occupancy) extended to 0.5
1 ng/mL, depending on the antibody. The anti-benzodiazepine probes were successfully employed for the analysis of 7-aminoflunitrazepam (7-AF) from urine.57 The probes were evaluated for utility in analysing for the presence of 7-AF at sub ng/mL concentrations in urine, which would be expected to be found several days after a single oral dose of 2 mg of flunitrazepam. Such analyses are required in monitoring for abuse of this drug, both in terms of ‘club drug’ use and in cases of drug-facilitated sexual assault. In these cases, drug concentrations in blood and urine are much lower than in chronic abuse cases and are difficult to analyse by conventional methods. In the technique, immobilised antibody probes were exposed to a sample containing the drug for 30 min. Extracted drugs were subsequently desorbed from the probes in 500 μL of methanolic desorption solution, which was dried, reconstituted in a small volume of injection solution and analysed by LC
MS/MS. The performance of both polyclonal and monoclonal antibodies was compared, as was the impact of affinity purification of the polyclonal antibody to isolate the drug-specific fraction. Interestingly, monoclonal antibody probes were not useful for analysis of this drug from urine. It appeared that a component of the urine matrix entirely eliminated all affinity of the antibody for the analyte. Presumably the redundancy of affinity inherent in polyclonal antibodies permitted them to retain their performance in the biological matrix. The method developed had an LOD of 0.02 ng/mL, with accuracy ranging from 6% to 25% and precision (% RSD) ranging from 2% to 10% between the lower limit of quantification (LLOQ) and the upper limit of quantification (ULOQ) for the analysis of 7-AF in urine. The dynamic range of the method was from 0.02 ng/mL, which was limited by the instrument sensitivity to 0.5 ng/mL, which was approaching the capacity of the probes. This would allow for quantitative analysis of samples at concentrations below that measurable by many other methods for general benzodiazepines analysis from urine and a highly selective screen for samples at higher concentrations. The method has similar limits of detection to the most sensitive literature methods specifically designed for 7-AF analysis,94
96 but with the advantage of significantly simplified sample preparation, making the technique amenable for use by professionals and non-professionals alike.

10.6.7 Molecularly Imprinted Polymers Molecularly imprinted polymers (MIPs) are extensively cross-linked polymer materials containing synthetic cavities or recognition sites with a predetermined

Drug Analysis by SPME

359

selectivity.97 The method of polymer imprinting involves including a template molecule with sufficient structural similarity to a desired analyte in the polymerisation matrix.98 The resulting so-called antibody mimics can be generated with specificities to virtually any analyte of choice, and a very pronounced difference in selectivity towards the target analyte can be achieved with an MIP material. This MIP material is resistant to mechanical stress, heat, acids, bases and organic solvents, while maintaining a storage life of years with no apparent reduction in its molecular recognition performance. Briefly, the desired affinity can be introduced by adding an amount of the compound of interest or a structural analogue to the polymerisation reaction. This ‘pattern’ chemical may be removed after polymerisation, leaving vacant sites of a specific size and shape, suitable for binding the same chemical again from an unknown sample. Because the cavities are based on the size, shape and chemical complementary of the imprint molecule, chiral selectivity is also possible. These sorbents also first saw application as an SPE packing material prior to application as an SPME sorbent. Several articles and reviews are available discussing the application of MIP for sample preparation for drug analysis.99
102 For SPME application, the polymer is prepared as normal, with a template related to the analyte of interest, and either coated on the inside surface of an in-tube SPME capillary or the outside of a fused silica fibre, or is crushed and sieved to obtain appropriately sized particulate material for gluing to an SPME fibre. For fibre SPME, applications have appeared for the analysis of triazines,103,104 clenbuterol105 and diacetyl morphine.106 In the triazine application, the MIP was formed in a 20- to 53-μm-thick coating on the outer surface of a fused silica fibre. A similar strategy was used in the preparation of the earlier clenbuterol MIP fibres, with a resulting 75-μm sorbent thickness. Finally, the diacetyl morphine fibres were prepared using methacrylic acid/dimethacrylate monolith imprinted with diacetyl morphine. In this work, the authors also performed extensive mathematical modelling of the adsorption isotherms for the final polymers. The in-tube SPME option with MIP coatings has been investigated for analysis of propranolol55 and verapamil107 and related metabolites. As for SPE packings, the main interest in these phases was for a highly selective extraction of the desired analyte from a complex biological matrix. The main limitations have been the amount of work required to prepare the phase and the difficulty in obtaining efficient desorption. If the target analyte is used as the template, it is sometimes difficult to eliminate the target analyte from blank analyses completely due to bleed from the polymer. MIP sorbents are comparable to biological recognition sorbents, such as immobilised antibody sorbents. The significant advantage of MIP is in their physical and chemical stability. In the case of the automated online MIP in-tube SPME extraction method for propranolol determination in biological fluids,55 this process simplified the sample preparation and chromatographic separation of several of the β-blocker compounds. The method developed for propranolol showed improved selectivity in comparison to alternative in-tube SPME coatings. Preconcentration of the sample by the MIP adsorbent increased the sensitivity, yielding an LOD

360

Handbook of Solid Phase Microextraction

suitable for clinical analysis. Excellent method reproducibility (% RSD ,5.0%) and column reusability (.500 injections) were observed over a fairly wide linear dynamic range (0.5
100 μg/mL) in serum samples. The method was inexpensive and simple to set up, and it simplified the choice of SPME adsorbent for in-tube extraction. More important, the approach can potentially be extended to other MIPs for the determination of a wide range of chemically significant analytes.

10.7

Derivatisation

Analyte derivatisation is used in general chemical analysis for several purposes. Derivatisation changes the physicochemical properties of an analyte to a form that can be analysed more easily. This may involve improvements in sample preparation efficiency, chromatographic properties or enhancement of detection sensitivity. In particular, derivatisation can produce a less-polar product from a polar analyte, significantly improving extraction efficiency. Table 10.1 summarises various derivatisation techniques that can be implemented in combination with SPME.41 Some of the techniques, such as direct derivatisation in the sample matrix, are analogous Table 10.1 Summary of SPME Derivatisation Schemes Derivatisation Strategy Derivatisation before extraction

Comments

Reagent added to sample matrix prior to extraction; allow sufficient time for reaction to complete before introducing SPME fibre; extraction proceeds according to the polarity/affinity of the product. Beneficial where analyte has insufficient affinity for the sorbent Derivatisation during Reagent preloaded onto SPME sorbent, simultaneous extraction and derivatisation. Extraction proceeds extraction according to polarity/affinity of the analyte, but sorbent acts as a zero sink until reagent is consumed. Beneficial where analyte has some affinity for the coating but that the product has more, and for maximising the amount of analyte extracted Derivatisation on fibre Fibre with extracted analyte exposed to reagent after being removed from extraction vial; reaction after extraction normally proceeds on the fibre prior to desorption. Beneficial to improve desorption, separation and/or detection Derivatisation after Analytes desorbed from fibre as normal; derivatisation occurs after desorption either in the injection port or desorption on column. Beneficial where separation/detection characteristics of the analyte are suboptimal

Reference

108

109

110

111

Drug Analysis by SPME

361

to well-established approaches used in solvent extraction. In the direct technique, the derivatising agent is added to the vial containing the sample, and the derivatives are then extracted by SPME and introduced into the analytical instrument. For example, this approach has been applied to extract and separate phenols from aqueous samples by first converting the target analytes to their acetate derivatives.112 Where extraction efficiency for underivatised compounds is sufficient to reach the sensitivity required, there may still be problems associated with the separation or detection of these analytes. Good chromatographic performance and detection can be facilitated by in-coating derivatisation following extraction. This has been shown with high-molecular-mass carboxylic acids. After exposing an SPME coating containing extracted analytes to diazomethane, the resulting ester derivatives can be separated as narrow bands on a GC column. For improved detection, selective derivatisation to analogues containing high-detector-response groups will result in enhancement in sensitivity and selectivity of detection. Derivatisation in the GC injector is an analogous approach, but it is performed at high injection port temperatures. For example, long-chain carboxylic acids can be extracted onto the coating as ion pairs when tetramethylammonium hydrogen sulphite is added to the sample. During volatilisation, analytes are converted to methyl esters.41 Another interesting and potentially very useful technique is simultaneous derivatisation and extraction, performed directly in the coating. This approach allows high efficiencies and can be used in remote field applications. The simplest way to execute the process is to dope the fibre with a derivatisation reagent and subsequently expose it to the sample. Then the analytes are extracted and simultaneously converted to analogues having high affinity for the coating. This is no longer an equilibrium process because derivatised analytes are collected in the coating as long as extraction continues. So long as the derivatisation reaction kinetics are faster than the extraction process, the sorbent acts as a zero sink as the concentration of the underivatised analyte in the sorbent is essentially zero at all times. This approach, which is used for low-molecular-mass carboxylic acids, results in the exhaustive extraction of gaseous samples.113 When 1-pyrenyldiazomethane is used as the derivatisation reagent, it is introduced into the coating by first dissolving the reagent in a volatile solvent. The fibre is then immersed in the solution. The fibre coating swells and is doped with the reagent. After evaporation of the solvent, the fibre is ready to perform extraction. The reagent, having low vapour pressure and high affinity towards the coating, remains on the fibre during its exposure to the sample. Volatilities of the pyrenylmethyl esters formed during the reaction are also low, resulting in the accumulation of the product onto the fibre until the analyte or reagent is exhausted or decomposed. At a high injector temperature the derivatised analytes are removed from the coating and the fibre can be reused. A similar approach is used for the analysis of formaldehyde from gaseous samples.114 This simple but powerful procedure, as described above for this analysis, is limited to low-volatility reagents. The approach can be made more general by chemically attaching the reagent directly to the coating. The chemically bound product can then be released from the coating either by high temperature in the injector, light illumination or change of an applied potential and so on. The feasibility of

362

Handbook of Solid Phase Microextraction

this approach was recently demonstrated by synthesising standards bonded to silica gel, which were released during heating.115 This approach allowed solvent-free calibration of the instrument. Numerous applications involving derivatisation SPME, including all the typical strategies, have been reported in the literature. These include derivatisation in the sample prior to extraction,108 simultaneous derivatisation, extraction in sample headspace,109,116,117 derivatisation on the fibre between extraction and desorption110 as well as injector port derivatisation subsequent to desorption.111 Examples of drugs and biologically relevant compounds for which derivatisation has proven useful include amphetamines,108,116,118
120 cocaine,121 methadone and cannabinoids,122 barbiturates,123 clenbuterol,124 ephedrine,69 steroids,110 busulfan125 and methylmalonic acid.44

10.8

Instrumental Configurations

SPME has been successfully coupled to a range of analytical instruments, as discussed in detail in Chapters 3 and 5. Strategies for coupling to the more widely used instrument types, along with important information on practical use for drug analysis, are described below.

10.8.1 Gas Chromatography Most of the examples of the use of microextraction for drug analysis that have been reported to date employ fibre SPME with GC analysis. Virtually any manufacturer’s instrument can be successfully employed for SPME analysis, and publications now exist covering most of the major drug classes. Standard GC injectors, such as split/splitless, can be applied to SPME so long as a narrow insert with an i.d. close to the outside diameter (o.d.) of the needle is used. A 0.7-mm i.d. liner is recommended. The narrow inserts are required to increase the linear flow around the fibre, resulting in efficient removal of desorbed analytes. The split should be turned off during SPME injection or the injector used in the splitless mode. Under these conditions, the desorption of analytes from the fibre is very rapid, not only because the coatings are thin but also because the high injector temperatures produce a dramatic decrease in the coating/gas distribution constant and an increase in the diffusion coefficients. The speed of desorption in many cases is limited by the time required to introduce the fibre into the heated zone. Some authors recommend opening the split valve after ca. 95% desorption has been achieved to permit the removal of the last traces of analyte from the sorbent and eliminate the potential for carry-over. This can be important for the lower volatility drugs that are sometimes slower to desorb from solid sorbent fibres. The carboxen (CAR)/PDMS fibre is particularly prone to carry-over in this regard. In cases where desorption is slower than desired and analyte peaks are unacceptably broad, column cold trapping or the use of a 2- to 5-m uncoated guard column may be sufficient to sharpen peaks.

Drug Analysis by SPME

363

10.8.2 Liquid Chromatography While SPME has seen significant application to the analysis of volatile drugs by GC, the application of SPME technology to LC techniques has lagged behind.126 This has been due to both the above-mentioned challenges of SPME
LC in general as well as the fact that the design of a conventional LC injection port does not lend itself to accepting an SPME fibre. First, conventional LC injection ports do not seal an injection needle in the port at high pressure. Second, while thermal desorption of volatile compounds into a moving gas stream is fast and efficient, the slower diffusion coefficients of compounds in condensed phases makes it difficult to desorb analytes from a fibre into a moving liquid stream efficiently in a compact plug. Research effort has been focused on designing SPME interfaces for liquid phase separation techniques to address these challenges. In the earliest strategy, this was an analogue of the traditional loop injection system consisting of a desorption chamber and a six-port injection valve.127 An updated description of this interface has been published since then.7 The desorption chamber is placed in the position at which the injection loop normally resides on the injection valve. When the injection valve is in the ‘load’ position, the fibre may be introduced into the desorption chamber under ambient pressure. This also allows for the introduction of a desorption solvent if it is different from the mobile phase. The valve is then switched to ‘inject’ to transfer the desorbed analytes to the column. Static or dynamic desorption is possible depending on the time between introducing the fibre to the interface and moving the valve to the inject position. An advantage of this design is that the size of the desorption chamber is easily adjusted. A heater may also be installed in the device to facilitate the desorption process. This interface has performed reasonably well, and its desorption volume is similar to the volume of the typical injection loop. Soon after the initial descriptions of manual interfaces for fibre SPME
HPLC, Supelco introduced a commercial version. It consists of a Valco or Rheodyne 6port injection valve with a desorption chamber (cell volume 75 μL) also in the form of a three-way tee located in the position of the injection loop. Unfortunately, researchers generally have found this interface to be cumbersome to use for more than a handful of analyses; it is also prone to damaging the SPME fibres. An early publication describing the use of the commercial interface provides a comparison of SPME
GC and SPME
LC for the analysis of lidocaine from urine using the 100-μm PDMS fibre.21 In some cases, desorption efficiency may be suboptimal with manual interfaces. Highly efficient desorption is easily obtained in an injection tee having a large internal volume, but this tends to result in overly broad peaks unless a high flow rate is used. When interface volume is reduced peak shape is improved, but desorption efficiency can be compromised. Some authors have addressed this by desorbing in a large internal volume interface but introducing only a portion of this to the analytical column. This of course compromises much of the preconcentration benefit of the SPME device. Finally, in many cases, the analyst needs to analyse dozens

364

Handbook of Solid Phase Microextraction

or hundreds of samples, making automation a requirement. Unfortunately, automation of the fibre introduction to a manual injection tee, desorption and injection is not straightforward. For automated analysis, automated in-tube SPME or automated offline desorption using the Concept 96 (PAS Technology) sample preparation station can be used as discussed in Chapter 5.

10.8.3 Capillary Electrophoresis Several authors have investigated the possibility of coupling SPME to CE separation. For such applications, desorption is typically conducted offline, and in many cases, novel extraction phases are employed. In 1997, Li and Weber5 described the use of SPME as a sample preparation device for CE analysis of barbiturates. In this case, a PVC-coated fibre was used for the extraction, and offline desorption was employed with the desorption solution (basic phosphate buffer) transferred to a sample vial for conventional introduction to CE. A similar technique was applied in 2006 for the analysis of ephedrine in water and urine69 and propranolol enantiomers in urine68 using a sol
gel fibre. In 2001, Jinno et al.128 demonstrated a modified in-tube SPME technique for interfacing with CE for the analysis of tricyclic antidepressants. This involved filling the lumen of the in-tube SPME capillary with sorbent fibre material. Up to 246 filaments were packed into a DB-1 capillary, with the packing density optimised at 52% of the lumen volume. Analytes were optimally desorbed in just 2 μL of acetonitrile in this system, with about 4 nL being injected for CE. Compared to the previous wire-in-tube studies,129 the authors observed between a 3-fold and 76-fold increase in preconcentration depending on whether analytes interacted significantly with the sorbent fibre in the lumen of the capillary. In fact, the preconcentration factors quoted by the authors indicate that they had likely achieved exhaustive extraction using the new technique. Other authors have investigated different options for coupling an SPME fibre to CE separation. In 2003, Stoyanov et al.130 described an interface for the direct coupling of a miniaturised SPME fibre and CE. Injection was accomplished by placing the fibre into the upstream end of the separation capillary and applying the separation voltage. Zone compression was achieved by exploiting the different cross-sectional areas in the two sections of capillary. Analytes moved more quickly in the region containing the fibre, due to the small cross-sectional area, and compressed when they reached the end of the fibre where the cross-sectional area suddenly increased and the electric field strength dropped correspondingly. Liu and Pawliszyn131 employed a novel strategy to address the difficulties of band broadening and analyte carry-over with SPME
CE, caused by slow analyte desorption in liquid phase. The system consisted of a separation channel with an adapter to accommodate an SPME fibre at the inlet. Three electrode locations were used: at the adapter, the start and the end of the separation capillary. The SPME fibre was placed in the adapter in desorption solution and analytes desorbed from the fibre were first transferred by elecrophoretic migration between the adapter and the inlet reservoirs into a short piece of microdialysis hollow fibre, located at the

Drug Analysis by SPME

365

inlet. Analytes with a molecular weight higher than the cut-off of the fibre were trapped at the inlet. When the separation voltage was later switched to the application between the inlet and outlet reservoirs of the separation capillary, trapped molecules were separated. The strategy effected preconcentration and eliminated band broadening and carry-over. The use of the technique was demonstrated by extraction, desorption and separation of β-lactoglobulin A and ovalbumin. A review of current progress in online SPME
CE coupling is also available.132

10.8.4 Direct Introduction to MS Initial reports have appeared for the direct coupling of SPME to mass spectrometry (MS) by electrospray ionisation (ESI) or matrix-assisted laser desorption/ionisation (MALDI). SPME may have an advantage over exhaustive technologies in coupling directly to MS because of the high phase ratio. As discussed above, the small amount of sorbent employed may reduce the amount of unwanted ‘contaminants’ co-extracted along with the analyte. A more highly purified analyte is needed if chromatography is to be circumvented. In an early study, Mo¨der et al.133 reported that they could measure acylcarnitines in water, urine and blood plasma by directly coupling the manual interface described by Chen and Pawliszyn 127 to the eluent flow entering an ESI
MS interface. The desorption chamber (70 μL) was first filled with a desorption solution (methanol:ethanol 80:20). The fibre with desorbed analytes was sealed into the interface for a 2-min static desorption. Upon switching the valve to the inject position, the desorption solution was pumped to the ESI nebuliser at 0.1 mL/min using an eluent containing methanol:water (1:1). The authors reported that the method was simpler than the conventional derivatisation/ chromatographic separation method and gave cleaner results and much better sensitivity than direct injection of the liquid sample to ESI. A similar setup was also employed by McCooeye et al.134 for the analysis of amphetamines in urine for the purposes of a fast screening method for forensic samples. Desorption solution was pumped to an ESI nebuliser with detection by ion-mobility spectrometry (IMS). In-tube SPME has also been directly coupled to MS, for the same reasons given by Mo¨der et al., in terms of convenience, good selectivity and reduction of noise.135 The authors analysed organolead in water samples using a PLOT column for extraction. After extraction, analytes were desorbed dynamically in mobile phase (water:methanol, 88:12) and sent to the ESI nebuliser (0.45 mL/min). By varying the fragmenter voltage, the authors were able to distinguish between elemental and molecular lead to enable speciation. Tong et al.136 have demonstrated the use of MALDI to desorb and ionise analytes from the surface of an SPME fibre with direct introduction to MS. With the use of a conducting polymer as the SPME sorbent, the polymer itself served as the matrix for assisting ionisation. In another strategy, Lokhnauth and Snow137 have reported desorbing analytes from an SPME fibre by placing it directly in the desorber of a conventional IMS with a heated 63Ni ionisation source. The authors applied the technique to the analysis of ephedrine in urine as a fast antidoping

366

Handbook of Solid Phase Microextraction

screening technique. The desorption efficiency was observed as 97%, with an LOD in urine of 0.05 μg/mL. This compares favourably with the World Anti-Doping Agency cut-off of 10 μg/mL. D’Agostino et al.138 reported on the use of desorption electrospray ionisation139 to directly interface an SPME fibre with extracted chemical warfare agents with ESI
MS. The goal of the project was to develop methods to survey typical office environment media (e.g. drywall, fabrics and paper) easily for the presence of chemical warfare agents. In the technique, the fibre was positioned in the region between the electrospray needle and the sampling cone. The spray was produced from an acetonitrile/water mobile phase (10 μL/min). A dedicated mounting device was developed to both position the fibre in the correct location and allow for adequate venting of the electrospray chamber to avoid contamination of the laboratory air with traces of chemical warfare agents. The technique was compared to analyses by more conventional LC
MS. Triethyl phosphate, sarin and soman were successfully analysed by the technique. Although the efficiencies of the recoveries from sample varied widely, sufficient material was recovered to allow for positive identification of the agents when the media were spiked at representative levels. The authors noted that because there was no chromatographic step, chemical interferences and ion suppression may have to be considered for more complex samples. Recently, Walles et al.140 have reviewed several options for direct introduction of SPME fibres via electro-nanospray to MS. The use of an SPME metal wire coated with RAM enabled the extraction of peptides resulting from tryptic digests of protein samples. The RAM coatings provided additional selectivity by eliminating most matrix interferences. The techniques discussed were (i) placing the fibre inside a modified electrospray nebuliser or nanospray tip or (ii) the isolation of the SPME device in a nanospray tip, which was in turn located inside a plastic pipet tip with 2 kV applied to the wire for ionisation. The results were compared to the performance obtained from a commercial nanospray source. The option of placing the fibre inside a modified electrospray nebuliser did not perform satisfactorily, but good data were obtained from the use of the nanospray tip when desorption solvent was added to the tip. The performance of the finalised method compared favourably to zip-tip purification.

10.8.5 Automation The high similarity of the manual SPME (mSPME) device to GC injection syringes permitted the early automation of SPME
GC methods (see also Chapter 5). Probably the first reference to automated SPME
GC was in 1992, when Arthur et al.141 described the optimisation of the SPME analysis of benzene and related compounds from aqueous samples. This involved adapting a Varian model 8100 syringe GC autosampler to accept the SPME device. A year later, Varian released their model 8200 autosampler as the first commercially available SPME
GC autosampler. In 1999, Ugland et al. described the use of this autosampler for the analysis of propylchloroformate derivatised amphetamine and related compounds from

Drug Analysis by SPME

367

urine.108 Unfortunately, this autosampler had some important limitations and researchers moved to the use of more multifunctional instruments.142 In 2000, Sporkert et al.143 reported on the use of a Gerstel MPS 2 autosampler for analysis of methadone and metabolites from hair; and in 2002, Namera et al.144 reported on the use of a CombiPAL autosampler by CTC Analytics (Zwigen, Switzerland) for the analysis of amphetamines from urine. Since that time, GC autosamplers based on the CTC Analytics design and distributed by several companies have been widely used for all automated SPME
GC applications.142 The flexibility of this platform to accommodate virtually any manufacturer’s GC and multiple steps for each injection sequence are important advantages that often offset the relatively high cost. The programming flexibility has been exploited to automate more complex sample preparation strategies, including derivatisation by either single-arm115 or dual-arm145 configurations. For the reasons discussed above, the development of automation strategies for interfacing fibre SPME and HPLC has lagged behind those for SPME
GC (see Section 10.8.2). No strategies have been proposed to automate the process used in manual introduction of a fibre to LC. The most promising strategies to date involve automated offline desorption using a 96-well plate format followed by conventional automated injection of the desorption solution (see Section 5.2.2 in Chapter 5). The obvious disadvantage to this strategy is the loss of some of the preconcentration achieved in the SPME process. Whereas it is possible to desorb a conventional SPME fibre into as little as 50 μL of solvent in a manual interface, often 200
300 μL is the minimum desorption volume required in offline desorption. Several strategies have been investigated for offline desorption of SPME fibres into vials or wells of 96-well plates followed either by direct automated injection of the desorption solution (the direct desorb/inject technique) or a drying of the solvent from the wells, followed by dispensing of a separate injection solution, dissolution of the dried analyte and automated injection (the sample dry technique). While 96-well plates have been used to facilitate the analysis of large numbers of samples, the techniques are also applicable to sample vials with limited volume inserts or conical plastic disposable 1.5-mL centrifuge tubes. Typically a 10- to 20μL injection volume produces acceptable peak shapes, so the volume of reconstitution solution should be targeted to this range.

10.8.5.1 Direct Desorb/Inject Technique To address the issue of loss of preconcentration advantage when fibres are desorbed in a large volume, a new 96-well plate was designed and machined into a block of Teflon.142 Teflon was selected for ease of machining and minimal analyte adsorption. Wells were prepared at standard well plate spacings, with profiles as shown in Figure 10.3. This design allowed for the use of a standard cap mat for sealing the wells and a standard injection syringe for withdrawal of sample after desorption as well as a rise in the level of the desorption solution as the needle moves into the well. The 2-cm well depth at 1 mm diameter allowed the desorption solution to cover the entire extraction phase at a 20-μL minimum well volume. The plate was

368

Handbook of Solid Phase Microextraction

Figure 10.3 Schematic of the well shape used in the custom direct desorption 96-well plate.

∅ 6 mm

20 μL

2 cm

∅ 1 mm diameter

prepared with overall dimensions consistent with a standard deep-well plate. For use, a desorption solution should be selected to allow for efficient desorption in the small volume. For many analytes, 75% methanol was found to be suitable. At higher methanol concentrations, peak broadening became unacceptable. Desorption solution was dispensed to the wells by means of a blunt-tipped Hamilton syringe. An IS was also included in the desorption solution. After desorption, 10
15 μL was injected. The strategy, while more cumbersome than the sample dry technique, is amenable to more volatile or labile analytes.

10.8.5.2 Sample Dry Technique This technique was designed to be used with commercially available supplies and equipment. Polyethylene 96-well plates with 0.5- and 1.0-mL well volumes were tested. In each case, sufficient desorption solution was dispensed to the wells to allow for the entire extraction phase to be immersed (400 or 550 μL, respectively). It was determined that the 0.5-mL well volume plates were not appropriate because it was too easy to lose analyte from the well during the sample dry-down step. Fibre desorption was accomplished either static or with agitation by placing the entire plate with multiple fibres on a rotary shaker at relatively low speed. While desorption was complete within 1 min, longer desorption times were typically employed to allow sufficient time to get all the fibres into the wells before the first had to be removed. Immediately after desorption, the solution was dried using a commercial plate dryer. This was typically achieved in 30 min when neat methanol was used for desorption. Significantly longer drying times were required if an aqueous component was added to the desorption solution. After drying, an injection solution was dispensed to each well (50
75% methanol), plates were vortexed at 500 rpm for 30 min to ensure complete dissolution of the analytes, and samples were analysed. Typically, analytes were redissolved in 40 μL of injection solution,

Drug Analysis by SPME

369

and 15 μL was injected. Desorption and injection solutions were dispensed to the wells by means of a positive displacement-repeating pipettor for optimal reproducibility. Internal standard (IS, lorazepam) was included in both the methanol desorption solution (0.2 ng/mL) and the injection solution (2 ng/mL) to control for pipetting/injection volume variance and evaporation. IS in the desorption solution was at a level 10 times higher than in the injection solution to account for ca. 10 3 concentration factor during drying. In each of these cases, it was not practical to inject more than about 50% of the solution to the LC. While this resulted in a loss of sensitivity, it also had the advantage of preserving the sample in case a re-injection was required. The most significant advantages of the direct desorb/inject technique were the speed of the overall analysis and the ability to analyse volatile or thermally labile compounds. The drawbacks were the tedious nature of dispensing the solvent to the wells and a somewhat higher inherent error, likely due to the combined effects of variability in the volume dispensed, and evaporation of solvent. In practice, these could be addressed by using an automated dispenser, effective sealing of the plate wells and by using an IS included in the desorption solution. This technique is more expensive to implement relative to manual injection because of the need to purchase 96-well plates and an autosampler. It is, however, less expensive than the plate dry technique because no plate dryer is required. The plate dry technique has the lowest inherent error. Another significant advantage of this technique is the flexibility in selecting separate solvents for desorption and injection to allow optimisation of each step independently.

10.9

Applications

10.9.1 Forensics/Toxicology The first applications of SPME technology to drug analysis were directed at monitoring drugs that are abused or toxic substances in clinical and forensic samples. The primary advantage noted by many authors was that new methods could be developed very quickly. This permitted fast turnaround of samples when needed for urgent patient care. A significant volume of literature is available on this subject.8,146
151 SPME has also seen a significant amount of use in profiling illicit drug preparations to determine the synthetic route used or distribution pathways for a particular preparation. Illicit preparations of drugs that are abused may be prepared by different synthetic routes, and for each of these, incomplete and side reactions may also occur. Extensive purification of the active ingredients is not normally performed following synthesis. As a result, several impurities are typically present in these preparations, and impurity profiling is convenient to ‘fingerprint’ these products when they are confiscated. LLE, or occasionally SPE, has been used to prepare samples of confiscated materials for profiling by either GC or HPLC. SPME is often more convenient because of the faster turnaround and typically more detailed fingerprint obtained.

370

Handbook of Solid Phase Microextraction

The use of SPME has been evaluated for impurity profiling in confiscated amphetamine and ecstasy tablets.152 For ecstasy tablets, 10 mg of granulated tablet was mixed with 5 mL of 0.1 M acetate buffer (pH 5.0), sonicated for 10 min and heated to 90 C. SPME headspace extraction was performed for 30 min. For amphetamine powders, 10 mg was mixed with 5 mL of 0.1 M acetate buffer (pH 5.0) and the SPME fibre was immersed in the sample for 30 min at 20 C. Acidic extraction pH was employed to reduce the amount of the basic active ingredient extracted so that impurities would be observed more easily. For ecstasy tablets, it was observed that the PDMS/DVB fibre extracted more impurities than the 100-μm PDMS fibre, substantially increasing the information content of the profile. Headspace extraction was preferred to avoid contamination of the fibre by tablet components. For amphetamine powders, direct immersion extraction was feasible, and the two fibres provided equivalent profiles. The repeatability of profiles was consistent to that attained from LLE, indicating similar abilities in differentiating between closely related but different drug seizures. Interestingly, for ecstasy, isosafrole was identified as a precursor for synthesis. With LLE, this compound was difficult to analyse due to its high volatility. With SPME extraction, however, it could be quantified to 2 ng/g tablet. More recently, Koester et al.153 have demonstrated the use of SPME for impurities profiling for illicit methamphetamine. In an interesting additional application, SPME has been used to determine the odour signature of street cocaine samples to guide the rational training of certified drug dogs.154 To date, SPME has not seen significant application to screening of antidoping samples, likely due to the lack of automation. A first publication on this potential field has appeared recently for the screening of stimulants, narcotics and local anaesthetics.155 The authors identified several advantages of SPME screening over conventional practices, including improved sensitivity, better use of human resources, reduced solvent/consumables and sample volume requirements. Drawbacks identified were the high cost of the fibres and the requirement for a dedicated autosampler, although these were seen as acceptable given the identified benefits.

10.9.2 Therapeutic Drug Monitoring Therapeutic drug monitoring, or the monitoring of circulating drug levels in a patient, is important in some clinical situations. Although it can be prohibitive due to cost and inconvenience, it is advantageous in several situations, such as monitoring for patient compliance, ascertaining appropriate drug levels where the therapeutic window is very small or monitoring in cases when patient physiology is variable or outside of normal ranges, such as may occur in critically ill, very young or elderly patients. In these cases, it allows for optimised pharmacotherapy or monitoring for genetic variability, organ function or the influence of co-medication. Reviews of the clinical and biomedical applications of SPME are available.174,175 Some publications are now appearing where SPME is used in therapeutic drug monitoring. Abdel-Rehim et al.125 demonstrated monitoring of busulphan from plasma using the CW/DVB fibre and GC/MS analysis. This alkylating agent is

Drug Analysis by SPME

371

administered prior to stem cell transplantation for treatment of leukaemia. The drug is characterised by a narrow therapeutic window, where either underdosing or overdosing may result in fatality. The benefits of the SPME method were reduced analysis time and automation of the analysis, with accuracy and precision comparable to conventional analysis. The fast analysis is particularly useful in this case where frequent high dosing is typically required. In another example of therapeutic drug monitoring, Takamoto et al.156 demonstrated monitoring of urinary acrolein during the administration of cyclophosphamide and ifosphamide. Acrolein is a metabolite of these two chemotherapeutic agents and is responsible for mucosal irritation in the urinary tract when this compound is excreted. Left unchecked, severe haemorrhagic cystitis can result. The authors demonstrated that the SPME
GC/MS method developed allows simple frequent monitoring for acrolein, which could improve patient care by quickly signalling when to take heightened preventative measures against haemorrhagic cystitis. Recently clinical studies using SPME coupled with LC/MS were performed to determine the concentration of tranexamic acid in patients undergoing open heart surgery with the use of cardiopulmonary bypass (CPB).176,177 Tranexamic acid is a forefront antifibrinolytic agent used in cardiac surgery to minimise blood loss and also its use has been expanded recently to other subspecialties of medicine. However, the optimal dosing regime to achieve and sustain a therapeutic blood concentration is still unresolved, thus monitoring of tranexamic acid concentration during the surgical procedure is highly recommended. The SPME method developed employed two formats of the coating: commercially available C-18 fibres and PAN
C-18 thin films. In addition, in the latter case an automated SPME CONCEPT 96t system (PAS Technologies, Magdala, Germany) was used to increase throughput of the analysis. The obtained results were validated against two standard methods: protein precipitation (PPT) and ultrafiltration (UF). Statistical analysis showed good agreement between all methods, which indicates that SPME can be used alternatively to conventional techniques in daily clinical practice and automation can reduce time of analysis to less than 3 min per sample. Figure 10.4 presents an example of a pharmacokinetic profile of tranexamic acid obtained by the four aforementioned methods.

10.9.3 Monitoring of Individual Response to Drug
Personalised Medicine Tailored patient care is not established as a standard approach; however, access to methods that simplify analytical procedures enhances the chance of introducing personalised therapy to daily clinical practice. Among other ‘-omics’ studies, metabolomics has been gaining in importance during the past few years in the biomedical field. Although metabolomics analysis usually leads to biomarker discovery, it can also bring valuable information about individual responses on treatment employed or on the medical condition of a patient. Ex vivo studies on metabolic profiling of patients undergoing cardiac surgery and administered with tranexamic acid were performed with the use of prototype mix-mode (C-18 1 benzensulphonic

Handbook of Solid Phase Microextraction

TXA concentration in plasma [μg/ml]

372

350 300 250 200 150 100 50 0 0

50

100

150

200

250

300

350

Time [min] PPT

aSPME

mSPME

UF

Figure 10.4 Pharmacokinetic profiles of tranexamic acid in a plasma patient undergoing open heart surgery with the use of cardiopulmonary bypass surgery. Comparative studies with the use of mSPME, automated solid-phase microextraction (aSPME) in addition to protein precipitation (PPT) and ultrafiltration (UF) for sample preparation.

acid) SPME fibres suitable for extraction of a wide range of compounds at different physical and chemical properties.178 Tentative identification of the analytes that were different between the two profiles enabled researchers to distinguish changes in biochemical pathways induced by the surgical procedure and treatment applied. Moreover, potential factor of neutrophil activation due to blood contact with artificial circulatory system was found. Multivariate principal component analysis not only showed a good clustering of samples obtained from the patients before and during surgery and drug infusion but also the presence of two outliers (Figure 10.5). Further analysis of these samples showed an increased level of the compounds related to tranexamic acid medication in one case (patient 9, Figure 10.5A), suggesting higher sensitivity of the patient to the drug compared to the rest of the group. In the second case (patient 7, Figure 10.5B), identification of the compounds contributing in differentiating between this and the rest of the patients, indicated significant change in bile acids level, what may imply liver disorder occurring during surgery.

10.9.4 In Vivo monitoring In vivo SPME has the potential to complement the range of technologies currently being employed for in vivo analysis of living systems. SPME may provide the selectivity and sensitivity of microdialysis analysis with improved time resolution of sampling. In the case of intravenous monitoring of drug concentrations, minimal disruption to the chemical balance in blood occurs because no blood and only very small amounts of drugs are removed. Ultimately, this should result in lower stress

Drug Analysis by SPME

373

Figure 10.5 Scores plots for PC1 and PC2 (A) and PC1 and PC3 (B) showing clustering of blank and QC samples and the existence of four outliers (duplicates of patient 9 and patient 7 from dosed group samples). Black triangles represent patients before surgery and drugs application, green dots - patients sampled during surgery and drug administration, blue crosses
quality control samples and red triangles
blanks.

levels on animals for pharmacokinetic studies, reduced exposure to blood for analytical personnel and simplified, less disruptive sampling. To date, commercial SPME devices have been used in some applications of direct in vivo analysis of living systems. For example, they have been applied for the in vivo analysis of pheromones and semiochemicals used in chemical communications by insects157,158 and frogs,159 respectively. In these cases, the living animals were non-invasively monitored over time, providing a convenient means to study complicated dynamic processes without interference. The commercially available CW/TPR phase has been used for monitoring pesticides levels in living plant tissues over time.6 Membrane extraction with sorbent interface (MESI) is a technique related to SPME that has been used for monitoring breath volatiles.160 Parallel monitoring of carbon dioxide can be used with MESI to standardise data for degree of mixing with atmospheric air and provide more accurate and precise information.161,162 Additional applications of in vivo SPME are discussed in Chapter 12. In some cases direct exposure of an SPME device to a complicated tissue matrix can result in fouling of the sorbent surface. A membrane made from appropriate material and coated or placed around a sorbent can add a certain degree of selectivity to the extraction process, resulting in membrane-protected SPME.60 Large, unwanted biomolecules may be excluded from a sorbent that would extract them otherwise, or a non-biocompatible sorbent may be rendered biocompatible. However, it is a limitation that the kinetics of membrane extraction are substantially slower than for direct extraction because the analytes must diffuse through the membrane before they can reach the coating. This may be overcome by the use of thin membranes and increased extraction temperatures, which would result in faster extraction times.163,164 Such experimental parameters as ruggedness, binding capacity and biocompatibility of the resultant extraction material need to be evaluated and optimised for each case. Several advantages for such approaches are envisioned for in vivo SPME. The technique would ensure that the drug and metabolite concentrations determined are

374

Handbook of Solid Phase Microextraction

representative of those to which body tissues are exposed. Because sampling, sample preparation and preconcentration occur in vivo, there is no concern that the sampling step would itself promote a change in the drug concentration or composition due to biochemical reactions occurring during sampling, such as the activation of clotting factors. In the case of monitoring of drugs in experimental animals, one would expected that such a technique would shorten and simplify drug dosing studies and significantly reduce the number of animals required per study. It would also reduce the impact of inconsistencies often observed among individual animals. The field of in vivo SPME is still new and doubtless much still needs to be done to evaluate all the potentials it holds. For more information, the interested reader is referred to two recent reviews of the status of in vivo SPME,90,165,179,180 as well as a dedicated chapter in this book (Chapter 12).

10.9.5 Binding Affinity As discussed previously, the presence of competing phases in a sample matrix can compromise analysis of analyte concentration (see Section 10.2.6). Because of the nature of equilibrium extraction, it also provides an important advantage in assessing binding affinities in biological systems. So long as depletion is negligible, the extraction will not disturb the equilibrium in a system. The method, therefore, can be used relatively easily to assess the affinities of various components of a complex system. To summarise, a microextraction of a multi-component system will provide information on the concentration of free analyte present in the system, provided that the amount extracted is insignificant relative to the total free concentration. By adding a standard addition spike and allowing the system to re-equilibrate, one can then assess the degree of binding to system components. By performing this analysis in systems containing the sample phase and one discontinuous phase at a time, one can assess the binding affinities of all system components individually. This technique has been described for analysing protein and membrane binding and determination of the free amount of a drug or other chemical in a system. It is the free concentration that is most toxicologically and pharmacologically relevant, as opposed to the nominal or total concentration most commonly reported. The method is faster and more convenient than alternatives, which include equilibrium dialysis or headspace equilibration analysis. The method is particularly well suited to situations where the discontinuous phase of interest is dissolved (such as proteins and antibodies) or otherwise difficult to separate from the liquid phase. The theory and application of the method for fibre SPME analysis of several polar compounds in biological matrices have been reported.166,167 The authors report the following criteria that must be met to implement the method successfully: G

G

G

The extraction itself must not influence the equilibrium between the aqueous phase of the sample and the discontinuous binding phase. The binding phase must not interfere with the extraction by binding to the extraction phase. The absorption profile of the compound should be the same in a calibration sample as in the sample containing the discontinuous phase.

Drug Analysis by SPME

375

The method was validated by comparing the amount of one of the compounds extracted by microextraction, both inside and outside the membrane in an equilibrium dialysis experiment. There was no difference in the free concentrations determined in the samples from inside and outside the dialysis tube. Bovine serum albumin (BSA) concentrations varied from 2.0 to 200 μM. They therefore concluded that the protein itself is not adsorbed to the fibre and that the protein does not influence the amount of chemical absorbed by the fibre in any other way. In practice, the free fraction (fX) of analyte X in solution is calculated as follows: fX 5

½Xa 1 5 ½Xt 1 1 KM ½M

ð10:18Þ

where [X]a is the concentration of the freely dissolved chemical, [X]t is the total concentration (free and bound to M), KM is the protein
water partition coefficient of X and [M] is the concentration of protein M in the aqueous phase. Experimentally, fX is determined by calculating the slope of the linear plot of [X]a/[X]t. Compound X is first added to an aqueous solution without protein, and the total concentration is determined. Protein is added to the system, and the system is allowed to re-equilibrate. Once equilibrated, the freely dissolved concentration is determined by microextraction. By varying [X]t, the required linear correlation can be determined. The analyte/protein partition coefficient (KM) can then be calculated. Equation (10.18) can be rearranged to give KM 5

ð½X0 2 ½Xa Þ 3 ðVa =VM Þ ½Xa

ð10:19Þ

where [X]0 is the total (nominal) concentration, [X]a is the free concentration in the solution, Va and VM are the volumes of the aqueous and protein phases, respectively. It is well established that a compound’s octanol/water partition coefficient correlates well with the therapeutic potency of many drug compounds and is used as a general parameter to describe effective concentrations and kinetic behaviour in biological systems. However, the molecular structures and overall natures of octanol versus membrane lipids are very different. Perhaps most importantly, octanol is a bulk solvent, whereas membrane lipids exhibit a bilayer structure. A simple means to correlate with the membrane partition coefficient would be much more valuable. Currently, most biological, toxicological and pharmacological effects are expressed based on nominal concentrations, even though it is only freely available concentrations that are responsible for biological effect. Efforts have been made to correlate Kow to actual free concentration in cell cultures. Vaes et al.168 demonstrated that Kow is well correlated with membrane partition coefficients, and not surprisingly, that as partition coefficient increases, the free fraction is reduced. Partitioning was studied in hepatocyte primary culture, S9 post-mitochondrial supernatant and microsomal cell fractions.

376

Handbook of Solid Phase Microextraction

A similar strategy has been employed for the determination of the equilibrium constant of protein binding between diazepam and human serum albumin (HSA).168 In this work, a Scatchard plot was used to evaluate the binding data. An equilibrium constant and an indication of one independent binding site were determined, which was consistent with literature results obtained by gel filtration analysis. For diazepam analysis, direct extraction with a 100-μm PDMS fibre was performed from a 0.5-μg/mL solution of diazepam in 1 mg/mL HSA dissolved in 0.066 M phosphate buffer pH 7.4. Trace amounts of methanol were observed to affect the precision of analysis. Because methanol could not be eliminated entirely due to its use as a solvent for diazepam standards, methanol content was kept consistent for all analyses. Chapter 11 of this book discusses this topic in more detail and provides additional examples of this type of application.

10.9.6 Pharmaceutical Product Analysis A key aspect in the regulatory control of pharmaceutical products is monitoring for impurities in these products. These may arise either from the synthetic route employed or from degradation on storage. Because of the potential for adverse toxicity arising from the presence of such impurities, careful monitoring is essential. SPME has proven ideal for such monitoring because the impurities are often volatile small molecules such as solvents, which are easily extracted by SPME.169 Colon and Richoll170 have recently demonstrated the use of a similar strategy for the monitoring of sulphonic acid ester impurities. In the design of pharmaceutical products, the nature of the active ingredient is only one of several important considerations. It is well known that pharmaceutical products from different manufacturers, containing the same active ingredient and dosage, can produce markedly different circulating concentrations of the active ingredient. Optimal drug formulation is a crucial step in the design of products that produce the optimal therapeutic effect. In addition to meeting regulatory standards for active ingredient identity and purity, the product must have optimised utility, bioavailability and stability. The nature of the formulation procedures and/or auxiliary (non-medical) components of a product can, for example, make a product easier to handle, limit degradation, enhance bioavailability and targeting of the active ingredient to receptor sites, provide for additional routes of administration and target the product for additional therapeutic indications for use.171 These auxiliary inactive ingredients, known collectively as excipients, can also solubilise the active ingredient, suppress the growth of microorganisms, provide bulk or a coating for a tablet or capsule, provide colour or mask an unpleasant taste or odour. A good working definition of an excipient is ‘agents in a medicinal preparation regardless of the nature, purpose or quantities employed, other than the components intended as the active ingredients’.172 With so many tasks to perform, it is obvious that the nature and numbers of excipients in various pharmaceutical products can vary widely. It is essential, therefore, that for the analysis of pharmaceutical products, the nature of these excipients

Drug Analysis by SPME

377

is taken into account, for their potential impact on the process of analysis of the active ingredient. In microextraction, the effects on the partitioning to an extraction phase are of utmost concern. The three most significant effects excipients can have on an extraction are as follows: G

G

G

Providing a phase with competing affinity for the analyte of interest (e.g. polymers, proteins, immiscible liquids and amorphous or crystalline solids). Changing the pH of the sample (e.g. acids, bases and buffers). Changing the polarity of the sample (e.g. miscible liquids).

Because excipients can be present in concentrations far in excess of those of the active ingredient, use of solid sorbents for microextraction is often precluded for direct extractions. For other extraction phases, care should be taken to ensure that the interface between the sample and the extraction phase does not become fouled by an excipient. It is recommended that headspace extraction be used where possible.

10.9.7 Enantiomeric Differentiation Selectivity in analysis based on differentiation of drug enantiomer is important for several reasons. First, different enantiomers may have significantly different therapeutic potency. Quantification on undifferentiated enantiomeric mixtures may give erroneous indications of therapeutic efficacy. Second, clandestine preparations of drugs can have significantly different enantiomeric ratios than legitimate drug preparations. Enantiomeric differentiation can be used both to determine if drug intake arose from legitimate or illicit sources and to determine the most likely synthetic pathways used to prepare illicit drug preparations. This can help investigators determine sources and distribution routes for these products. Enantiomeric selectivity may be introduced into an SPME analysis by employing derivatisation with a chiral reagent, followed by separation of the diasteriomers. A demonstration of this strategy was presented for the analysis of amphetamines.173

10.10

Conclusions

There have been significant advances over the past dozen years in the application of SPME technology to drug analysis and bioanalysis. Much of this has come from innovation in sorbent and instrument design and configuration. At the time of writing, it appears that applications exploiting these innovations are still preliminary or at the proof-of-concept phase. None has achieved wide acceptance or broad application. However, many of the innovations seen are highlighting new possibilities for chemical analysis, which had not been envisioned even 5 years ago. As an enabling tool allowing analysts to ask questions that previously had not been possible, it appears that the field will enjoy continued attention for new discoveries in chemical analysis for both novel tools of investigation and new fields of application.

378

Handbook of Solid Phase Microextraction

References 1. HL Lord, Development of Solid Phase Microextraction Devices for In Vivo Analysis, Ph.D. Thesis, University of Waterloo, Waterloo, ON, Canada (2005) Chapter 9 2. J Pawliszyn, Solid Phase Microextraction: Theory and Practice (1997) Wiley: New York, NY 3. HL Lord, M Rajabi, S Safari & J Pawliszyn, J Pharm Biomed Anal 44 (2007) 506 4. J Pawliszyn, Anal Chem 75 (2003) 2543 5. S Li & S Weber, Anal Chem 69 (1997) 1217 6. HL Lord, M Moder, P Popp & JB Pawliszyn, Analyst 129 (2004) 107 7. HL Lord, R Grant, M Walles, B Incledon, B Fahie & J Pawliszyn, Anal Chem 75 (2003) 5103 8. HL Lord & J Pawliszyn, J Chromatogr A 902 (2000) 17 9. H Daimon & J Pawliszyn, Anal Commun 34 (1997) 365 10. K Yang, R Eisert, H Lord & J Pawliszyn, Determination of Pesticides in Foods by Automated SPME
GC
MS, in J Pawliszyn, Ed, Applications of Solid-Phase Microextraction (1999) Royal Society of Chemistry: Letchworth, Hertfordshire 11. A Boyd-Boland, T Gorecki, M Chai, Y Luo, Z Zhang, M Yang & J Pawliszyn, Environ Sci Technol 28 (1994) 569A 12. H Lord & J Pawliszyn, Anal Chem 69 (1997) 3899 13. BJ Hall & JS Brodbelt, J Chromatogr A 777 (1997) 275 14. M Abdel-Rehim, M Andersson, E Portelius, C Norsten-Ho¨o¨g & LG Blomberg, J Microcol Sep 13 (2001) 313 15. AA Boyd-Boland & J Pawliszyn, J Chromatogr A 704 (1995) 163 16. DD Perrin & B Dempsey, Buffers for pH and Metal Ion Control (1994) Chapman and Hall: London 17. MEC Queiroz, SM Silva, D Carvalho & FM Lanc¸as, J Chromatogr Sci 40 (2002) 219 18. J-L Liao, C-M Zeng, S Hjerten & J Pawliszyn, J Microcol Sep 8 (1996) 1 19. T Gorecki, X Yu & J Pawliszyn, Analyst 124 (1999) 643 20. D Louch, S Motlagh & J Pawliszyn, Anal Chem 64 (1992) 1187 21. EHM Koster, NSK Hofman & GJ de Jong, Chromatographia 47 (1998) 678 22. Y Luo, L Pan & J Pawliszyn, J Microcolumn Sep 10 (1998) 193 23. GJ de Jong & EHM Koster, Chromatographia 52 (Suppl) (2000) S12 24. EHM Koster, C Wemes, JB Morsink & GJ de Jong, J Chromatogr B 739 (2000) 175 25. C-C Chou & MR Lee, Anal Chim Acta 538 (2005) 49 26. CM Lock, L Chen & DA Volmer, Rapid Commun Mass Spectrom 13 (1999) 1744 27. T Kumazawa, H Seno, K Watanabe-Suzuki, H Hattori, A Ishii, K Sato & O Suzuki, J Mass Spectrom 35 (2000) 1091 28. A Aresta, F Palmisano & CG Zambonin, J Pharm Biomed Anal 39 (2005) 643 29. A Aresta, F Palmisano, R Vatinno & CG Zambonin, J Agric Food Chem 54 (2006) 1594 30. T. Gorecki, Solid versus Liquid Coatings, in J Pawliszyn, Ed, Applications of Solid-Phase Microextraction (1999) Royal Society of Chemistry: Letchworth, Hertfordshire, pp. 92
108 31. J Ai, Quantitation by SPME before Reaching a Partition Equilibrium, in J Pawliszyn, Ed, Applications of Solid-Phase Microextraction (1999) Royal Society of Chemistry: Letchworth, Hertfordshire, pp. 22
37 32. F Musshoff, H Junker & B Madea, J Anal Toxicol 24 (2000) 372

Drug Analysis by SPME

379

33. K Blau & JM Halket, Handbook of Derivatives for Chromatography, 2nd ed (1993) Wiley: New York, NY 34. Y Alnouti, CA White & MG Bartlett, Biomed Chromatogr 18 (2004) 641 35. S AbuRuz, J Millership & J McElnay, J Chromatogr B 798 (2003) 203 36. AT Cawley, R Kazlauskas, GJ Trout & AV George, J Chromatogr B 825 (2005) 1 37. K Lan, X Jiang, L Wang, J Zhou, Q Jiang & L Ye, J. Pharm Biomed Anal 44 (2007) 540 38. YC Fiamegos & CD Stalikas, Anal Chim Acta 597 (2007) 32 39. TS Ho, TG Halvorsen, S Pedersen-Bjergaard & KE Rasmussen, J Chromatogr A 998 (2003) 61 40. TS Ho, JLE Ruebsaet, HS Anthonsen, S Pedersen-Bjergaard & KE Rasmussen, J Chromatogr A 1072 (2005) 29 41. L Pan & J Pawliszyn, Anal Chem 69 (1997) 196 42. R Alzaga, A Pen˜a, L Ortiz & JM Bayona, J Chromatogr A 999 (2003) 51 43. IOM Chan, PKS Lam, RHY Cheung, MHW Lam & RSS Wu, Analyst 130 (2005) 1524 44. N Li, C Deng & X Zhang, J Sep Sci 30 (2007) 266 45. PL Nietzel, W Walther & W Nestler, Fresen J Anal Chem 361 (1998) 318 46. X Yu & J Pawliszyn, Anal Chem 72 (2000) 1788 47. MN Sarrio´n, FJ Santos & MT Galceran, Anal Chem 72 (2000) 4865 48. E Out & J Pawliszyn, J Mikrochim Acta 112 (1993) 41 49. RG Belardi & J Pawliszyn, Water Qual Res J Can 24 (1989) 179 50. Z Zhang & J Pawliszyn, Anal Chem 65 (1993) 1843 51. F Guo, T Gorecki, D Irish & J Pawliszyn, Anal Commun 33 (1996) 361 52. HB Wan, H Chi, MK Wong & CY Mok, Anal Chim Acta 298 (1994) 219 53. T Gorecki, P Martos & J Pawliszyn, Anal Chem 70 (1998) 19 54. J Wu & J Pawliszyn, J Chromatogr A 909 (2001) 37 55. W Mullett, P Martin & J Pawliszyn, Anal Chem 73 (2001) 2383 56. H Yuan, WM Mullett & J Pawliszyn, Analyst 126 (2001) 1456 57. HL Lord, M Rajabi, S Safari & J Pawliszyn, Pharm Biomed Anal 40 (2006) 769 58. W Mullett & J Pawliszyn, Anal Chem 74 (2002) 1081 59. M Walles, W Mullett & J Pawliszyn, J Chromatogr A 1025 (2004) 85 60. FM Musteata, M Walles & J Pawliszyn, Anal Chim Acta 537 (2005) 231 61. A Es-haghi, X Zhang, FM Musteata, H Bagheri & J Pawliszyn, Analyst 132 (2007) 672 62. D Wang, SL Chong & A Malik, Anal Chem 69 (1997) 4566 63. JD Hayes & A Malik, J Chromatogr B 695 (1997) 3 64. SL Chong, D Wang, JD Hayes, BW Wilhite & A Malik, Anal Chem 69 (1997) 3889 65. Z Wang, C Xiao, C Wu & H Han, J Chromatogr A 893 (2000) 157 66. H Bagheri, A Es-haghi & MR Rouini, J Chromatogr B 818 (2005) 147 67. X Li, Z Zheng, M Hu & M Mao, J Sep Sci 28 (2005) 2489 68. X Zhou, X Li & Z Zeng, J Chromatogr A 1104 (2006) 359 69. H Fang, M Liu & Z Zeng, Talanta 68 (2006) 979 70. S Bigham, J Medlar, A Kabir, C Shende, A Alli & A Malik, Anal Chem 74 (2002) 752 71. Y Fan, Y-Q Feng, S-L Da & Z-H Wang, Talanta 65 (2005) 111 72. J Zhou & Z Zeng, Anal Chim Acta 556 (2006) 400 73. Y Fan, Y-Q Feng, J-T Zhang, S-L Da & M Zhang, J Chromatogr A 1074 (2005) 9 74. Y Fan, Y-Q Feng, S-L Da & X-P Gao, Analyst 129 (2004) 1065 75. Y Wen, Y Fan, M Zhang & Y-Q Feng, Anal Bioanal Chem 382 (2005) 204

380

Handbook of Solid Phase Microextraction

76. J Nie, Q Zhao, J Huang, B Xiang & Y-Q Feng, J Sep Sci 29 (2006) 650 77. J Nie, M Zhang, Y Fan, Y Wen, B Xiang & Y-Q Feng, J Chromatogr B 828 (2005) 62 78. Y Fan, Y-Q Feng, S-L Da & Z-G Shi, Anal Chim Acta 523 (2004) 251 79. A Witkowski & A Brajter-Toth, Anal Chem 64 (1992) 635 80. K Pihel, QD Walker & RM Wightman, Anal Chem 68 (1996) 2084 81. C Hsueh & A Brajter-Toth, Anal Chem 66 (1994) 2458 82. Z Gao, B Chen & M Zi, Analyst 119 (1994) 459 83. J Wu, H Lord, H Kataoka & J Pawliszyn, J Microcolumn Sep 12 (2000) 255 84. J Wu, X Yu, H Lord & J Pawliszyn, Analyst 125 (2000) 391 85. J. Wu, Ph.D. Thesis, University of Waterloo, Waterloo, ON, Canada (2001) Section 4.3.3 86. V Haase & F Beck, Electrochim Acta 39 (1994) 1195 87. W Su & JO Iroh, Electrochim Acta 46 (2000) 1 88. Y Wang, M Walles, B Thomson, S Nacson & J Pawliszyn, Rapid Commun Mass Spectrom 18 (2004) 157 89. Y Wang, BB Schneider, TR Covey & J Pawliszyn, Anal Chem 77 (2005) 8095 90. FM Musteata, ML Musteata & J Pawliszyn, Clin Chem 52 (2006) 708 91. Y Wang, S Nacson & J Pawliszyn, Anal Chim Acta 582 (2007) 50 92. JK Schubert, W Miekisch, P Fuchs, N Scherzer, H Lord, J Pawliszyn & RG Mundkowski, Clin Chim Acta 386 (2007) 57 93. J Wu, WM Mullett & J Pawliszyn, Anal Chem 74 (2002) 1855 94. MA LeBeau, MA Montghomery, JR Wagner & ML Miller, J Forensic Sci 45 (2000) 1133 95. N Jourdil, J Bessard, F Vincent, H Eysseric & G Bessard, J Chromatogr B 788 (2003) 207 96. A Negrusz & RE Gaensslen, Anal Bioanal Chem 376 (2003) 1192 97. K Ensing & T de Boer, Trends Anal Chem 18 (1999) 138 98. G Vlatakis, LI Andersson, R Mu¨ller & K Mosbach, Nature 361 (1993) 645 99. C Chassaing, J Stokes, RF Venn, F Lanza, B Sellergren, A Holmberg & C Berggren, J Chromatogr B 804 (2004) 71 100. P Dzygiel, E O’Donnell, D Fraier, C Chassaing & PAG Cormack, J Chromatogr B 853 (2007) 346 101. J Courtois, G Fischer, B Sellergren & K Irgum, J Chromatogr A 1109 (2006) 92 102. B Dirion, Z Cobb, E Schillinger, LI Andersson & B Sellergren, J Am Chem Soc 125 (2003) 15101 103. X Hu, Y Hu & G Li, J Chromatogr A 1147 (2007) 1 104. E Turiel, JL Tadeo & A Martin-Esteban, Anal Chem 79 (2007) 3099 105. MEH Koster, C Crescenzi, WD Hoedt, K Ensing & GJ de Jong, Anal Chem 73 (2001) 3140 106. D Djozan & T Baheri, J Chromatogr A 1166 (2007) 16 107. W Mullett, M Walles, K Levsen & J Pawliszyn, J Chromatogr B 801 (2004) 297 108. HG Ugland, M Krogh & KE Rasmussen, J Pharm Biomed Anal 19 (1999) 463 109. Q Wang, J O’Reilly & J Pawliszyn, J Chromatogr A 1071 (2005) 147 110. P Okeyo, SM Rentz & NH Snow, J High Resol Chromatogr 20 (1997) 171 111. M del Olmo, AZB Sua´rez, A Gonzalez-Casado, J Taoufiki & JL Vilchez, J Chromatogr B 817 (2005) 167 112. K Buchholz & J Pawliszyn, Anal Chem 66 (1994) 160 113. L Pan & J Pawliszyn, Anal Chem 67 (1995) 4396 114. P Martos & J Pawliszyn, Anal Chem 70 (1998) 2311

Drug Analysis by SPME

381

115. P Konieczka, L Wolska, E Luboch, J Namiesnik, A Przyjazny & J Biernat, J Chromatogr A 742 (1996) 175 116. K-J Chia & S-D Huang, Anal Chim Acta 539 (2005) 49 117. Q Wang, JM Chong & J Pawliszyn, Flavour Fragr J 21 (2006) 385 118. A Namera, M Yashiki, T Kojima & M Ueki, J Chromatogr Sci 40 (2002) 19 119. K Watanabe, N Okamoto, I Yamagishi, H Nozawa, A Ishii & O Suzuki, Chromatographia 58 (2003) 455 120. C Cha´fer-Perica´s, P Campı´ns-Falco´ & R Herra´ez-Herna´ndez, J Pharm Biomed Anal 40 (2006) 1209 121. FCP de Toledo, M Yonamine, RL de Moraes Moreau & OA Silva, J Chromatogr B 798 (2003) 361 122. DW Lachenemeier, L Kroener, F Musshoff & B Madea, Rapid Commun Mass Spectrom 17 (2003) 472 123. U Staerk & WR Ku¨lpmann, J Chromatogr B 745 (2000) 399 124. MD Engelmann, D Hinz & BW Wenclawlak, Anal Bioanal Chem 375 (2003) 460 125. M Abdel-Rehim, Z Hassan, L Blomberg & M Hassan, Ther Drug Monit 25 (2003) 400 126. HL Lord, J Chromatogr A 1152 (2007) 2 127. J Chen & J Pawliszyn, Anal Chem 67 (1995) 2530 128. K Jinno, M Kawazoe, Y Saito, T Takeichi & M Hayashida, Electrophoresis 22 (2001) 3785 129. Y Saito, M Kawazoe, M Hayashida & K Jinno, Analyst 125 (2000) 807 130. A Stoyanov, Z Liu & J Pawliszyn, Anal Chem 75 (2003) 3324 131. Z Liu & J Pawliszyn, Analyst 131 (2006) 522 132. Z Liu & J Pawliszyn, J Chromatogr Sci 44 (2006) 366 133. M Mo¨der, H Lo¨ster, R Herzschuh & P Popp, J Mass Spectrom 32 (1997) 1195 134. MA McCooeye, Z Mester, B Ells, DA Barnett, RW Purves & R Guevremont, Anal Chem 74 (2002) 3071 135. Z Mester & J Pawliszyn, Rapid Commun Mass Spectrom 13 (1999) 1999 136. H Tong, N Sze, B Thomson, S Nacson & J Pawliszyn, Analyst 127 (2002) 1207 137. JK Lokhnauth & NH Snow, J Sep Sci 28 (2005) 612 138. PA D’Agostino, JR Hancock, CL Chenier & CR Jackson Lepage, J Chromatogr A 1110 (2006) 86 139. Z Takats, JM Wiseman, B Gologan & RG Cooks, Science 306 (2004) 471 140. M Walles, Y Gu, C Dartiguenave, FM Musteata, K Waldron, D Lubda & J Pawliszyn, J Chromatogr A 1067 (2005) 197 141. C Arthur, L Killam, K Buchholz & J Pawliszyn, Anal Chem 64 (1992) 1960 142. J O’Reilly, Q Wang, L Setkova, JP Hutchinson, Y Chen, HL Lord, CM Linton & J Pawliszyn, J Sep Sci 28 (2005) 2010 143. F Sporkert, F Pragst, S Hu¨bner & G Mills, J Chromatogr B 772 (2002) 45 144. A Namera, M Yashiki, T Kojima & M Ueki, J Chromatogr Sci 40 (2002) 19 145. D Parkinson, I Bruheim, I Christ & J Pawliszyn, J Chromatogr A 1025 (2004) 77 146. G Theodoridis, EHM Koster & GJ de Jong, J Chromatogr B 745 (2000) 49 147. F Degel, Clin Biochem 29 (1996) 529 148. H Kataoka, Trends Anal Chem 22 (2003) 232 149. MEC Queiroz & FM Lanc¸as, LC
GC Europe 18 (2005) 145 150. S Ulrich, J Chromatogr A 902 (2000) 167 151. NH Snow, J Chromatogr A 85 (2000) 445 152. KE Kongshaug, S Pedersen-Bjergaard, KE Rasmussen & M Krogh, Chromatographia 50 (1999) 247

382

Handbook of Solid Phase Microextraction

153. CJ Koester, BD Andresen & PM Grant, J Forensic Sci 47 (2002) 1002 154. KG Furton, Y-C Hong, YL Hsu, T Luo, S Rose & J Walton, J Chromatogr Sci 40 (2002) 147 155. S Strano-Rossi, F Molaioni & F Botre`, J Anal Toxicol 29 (2005) 217 156. S Takamoto, N Sakura, A Namera & M Yashiki, J Chromatogr B 806 (2004) 59 157. G Moneti, FR Dani & GTS Pieraccini, Rapid Commun Mass Spectrom 11 (1997) 857 158. B Frerot, C Malosse & AH Cain, J High Resolut Chromatogr 20 (1997) 340 159. BP Smith, CA Zini, J Pawliszyn, MJ Tyler, Y Hayasaka, B Williams & EB Caramao, Chem Ecol 17 (2000) 215 160. HL Lord, Y Yu, A Segal & J Pawliszyn, Anal Chem 74 (2002) 5650 161. Y Yu & J Pawliszyn, J Chromatogr A 1056 (2004) 35 162. X WMa, Liu & J Pawliszyn, Anal Bioanal Chem 385 (2006) 1398 163. H Lord & J Pawliszyn, J Chromatogr A 885 (2000) 153 164. Z Zhang, J Poerschmann & J Pawliszyn, Anal Commun 33 (1996) 129 165. FM Musteata & J Pawliszyn, J Biochem Biophys Meth 70 (2006) 181 166. WHJ Vaes, EU Ramos, HJM Verhaar, W Seinen & JLM Hermens, Anal Chem 68 (1996) 4463 167. WHJ Vaes, EU Ramos, C Hamwijk, I van Holsteijn, BJ Blaauboer, W Seinen, HJM Verhaar & JLM Hermens, Chem Res Toxicol 10 (1997) 1067 168. H Yuan, R Ranatunga, P Carr & J Pawliszyn, Analyst 124 (1999) 1443 169. AR Raghani, J Pharm Biomed Anal 29 (2002) 507 170. I Colo´n & SM Richoll, J Pharm Biomed Anal 39 (2005) 477 171. I Racz, Drug Formulation (1989) Wiley: Budapest 172. M Weiner & IL Bernstein, Adverse Reactions to Drug Formulation Agents: A Handbook of Excipients (1989) Marcel Dekker: New York, NY 173. C Cha´fer-Perica´s, P Campins-Falco & R Herraez-Hernandez, J Pharm Biomed Anal 40 (2006) 1209 174. H Kataoka & K Saito, J Pharm Biomed Anal 54 (2011) 926 175. B Bojko, E Cudjoe, M Wasowicz & J Pawliszyn, Trends Anal Chem 30 (2011) 1505 176. B Bojko, D Vuckovic, F Mirnaghi, E Cudjoe, M Wasowicz, A Jerath, J Pawliszyn, Ther Drug Monit submitted 177. B Bojko, D Vuckovic, E Cudjoe, Md E. Hoque, F Mirnaghi, M Wasowicz, A Jerath, J Pawliszyn, J Chromatogr B, in press, doi:10.1016/j.jchromb.2011.08.003 178. B Bojko, M Wasowicz, J Pawliszyn, manuscript in preparation 179. D Vuckovic, R Shirey, Y Chen, L Sidisky, C Aurand, K Stenerson & J Pawliszyn, Anal Chim Acta 638 (2009) 175 180. JCY Yeung, D Vuckovic & J Pawliszyn, Anal Chim Acta 665 (2010) 160

11 Ligand
Receptor Binding and Determination of Free Concentrations

Florin Marcel Musteata Department of Pharmaceutical Sciences, Albany College of Pharmacy and Health Sciences, Albany, NY, USA

11.1

Introduction

Bioanalytical chemistry is playing an increasingly central role in the fields of academic and industrial science. It overlaps with a diverse range of disciplines, including biotechnology, biopharmaceuticals and diagnostics.1 Bioanalytical chemistry can be defined as the development and application of chemical measurements and instrumentation to problems in biology, biochemistry and medical science. For the pharmaceutical industry, bioanalytical chemistry is often synonymous with measurements in biological samples, typically in support of investigations of drug metabolism and pharmacokinetics.2 Biological materials and pharmaceutical products are very complex mixtures. They often contain proteins, salts, acids, bases and numerous organic compounds that may be similar to the analyte of interest. Furthermore, the analytes often exist at low concentration in these samples. Despite significant advances in the development of highly efficient analytical instruments for the end-point determination of analytes in biological samples and pharmaceutical products, a pre-treatment step is usually necessary to extract and isolate the analytes of interest from complex matrices. The goal of sample preparation is to eliminate interfering compounds from the matrix using a minimum number of steps, resulting in a reproducible methodology. Many of the current sample preparation challenges are addressed by solid-phase microextraction (SPME), which was specifically developed to provide rapid sample preparation both in the laboratory and on-site (where the investigated system is located). This chapter presents recent advances and future trends in SPME method development for the analysis of endogenous and exogenous compounds in biological samples, with a focus on the determination of free concentrations and binding constants.

11.2

Analysis of Biological Samples

Analysis of drugs in biological samples and pharmaceutical products is growing in importance because of the need to understand therapeutic and toxic effects of drugs Handbook of Solid Phase Microextraction. DOI: 10.1016/B978-0-12-416017-0.00011-5 © 2012 Elsevier Inc. All rights reserved.

384

Handbook of Solid Phase Microextraction

and the continuing development of more selective and effective drugs.3 Interest in the field of drug analysis is focusing on improving methodologies, with regard to how quickly, accurately and sensitively the chemicals can be detected. Knowledge of drug levels in body fluids such as serum, saliva and urine allows for the optimisation of pharmacotherapy and provides the basis for studies on patient compliance, bioavailability, pharmacokinetics and genetics, organ function and the influences of co-medication.4 The quantitative and qualitative analysis of drugs and metabolites is extensively applied in pharmacokinetic studies and therapeutic drug monitoring. Drugs of abuse, illicit drugs and intoxications by drugs and poisons are often analysed in clinical and forensic toxicology. Some of the most recent applications of SPME in the analysis of biological samples are included in Table 11.1. Additional applications are presented in Chapter 10. In vitro applications of SPME developed to date include the analysis of drugs from serum, plasma, whole blood, milk, urine, saliva and hair, by headspace, direct immersion SPME and in-tube SPME. Generally, the analysis of biological fluids is encumbered by the presence of dissolved biopolymers. For example, 7
8% of human plasma is composed of proteins. The composition of plasma can be subject to considerable differences due to pathological and non-pathological influences. For example, plasma albumin can be decreased to about 50% of the normal level in hepatic diseases, and the concentration of lipoprotein-bound triglycerides depends on dietary status. The main issues to be considered for qualitative and quantitative analysis in plasma and other biological samples include (i) a change in selectivity, because of interferences from endogenous substances; (ii) analyte binding to biopolymers; and (iii) high viscosity of the sample.15 Analyte binding to biopolymers results in a decrease in the sensitivity for methods based on SPME. However, this also presents a unique advantage of SPME over other sample preparation methods because a direct assay of the free concentration can be performed without separation of the phases. The viscosity of plasma and blood in vitro is about three times higher than the viscosity of water. Because the diffusion coefficients are inversely related to viscosity, the diffusion of analytes in the plasma is approximately three times slower than in a predominantly aqueous phase. When the extraction speed is controlled by diffusion in the sample, an increase in the equilibration time is expected. By increasing the extraction temperature, the distribution constant of the drug between the fibre coating and the sample decreases. At the same time, the analyte diffusion rate increases because of lower viscosity and higher diffusion coefficient. Consequently, SPME methods can be optimised by selecting extraction temperatures that result in satisfactory sensitivity in an acceptable period.16 Applications of SPME in bioanalysis can be divided into eight main groups, according to the type of analyte: toxicological and forensic analysis, drugs that can be abused, clinical chemistry, analysis of pharmaceuticals in biological samples, biochemical analysis, semiochemical analysis and analysis of natural products.17 Commercially available Carbowaxs
TPR fibres have also been successfully used for the direct extraction of chlorhexidine from saliva during a pharmacokinetic

Table 11.1 Selected Recent Applications of SPME in Bioanalysis Analytes

Biological Sample

Type of Investigation

Extraction Time (min)

Extraction Phase(s)

Reference

Ibuprofen, warfarin, caffeine, verapamil and propranolol

Human plasma

Determination of plasma protein binding (PPB)

10

Polypyrrole (PPY) Polydimethylsiloxane (PDMS)

5

Diazepam and metabolites

Whole blood (beagle vein, in vivo)

Quantitative analysis of total and free concentration

0.5 & 2

Hydrophilic PPY Polyethylene glycol (PEG)

6

7-Aminoflunitrazepam Angiotensin I and II

Human urine Whole human blood

Quantitative analysis Quantitative analysis

30 45
60

7 8

Diazepam and isosorbide dinitrate Chlorhexidine and its degradation products

Human serum albumin

8 15

Ibuprofen, naproxen, angiotensin II and neurotensin Diazepam and metabolites

Human urine

Determination of binding parameters Quantitative analysis of total and free concentration Method development

Specific antibodies Exchange diol silica (restricted access) Alkyl diol silica (restricted access) Carbowax
TPR

5

Antibodies

11

Quantitative analysis

30

PPY

12

Quantitative analysis

5
60

13

Determination of free concentrations

5

Alkyl diol silica (restricted access) PDMS

Human saliva

Benzodiazepines

Whole blood (beagle vein, in vivo) Human urine

Valproic acid

Human plasma

9 10

14

386

7e6

Handbook of Solid Phase Microextraction Mass counts

6e6 5e6 4e6 3e6 2e6 (d) (c)

1e6 (a)

0 2

4

6

8

10

(b) min

Figure 11.1 Chromatograms of saliva samples at different time points after administration of chlorhexidine: (a) 0.25 h; (b) 1 h; (c) 4 h; and (d) 8 h. (Source: Reproduced from Ref. 8 with permission of Elsevier. r 2005.)

study (Figure 11.1).10 Such methods of analysis based on SPME are economical and much faster when compared to classical approaches. A general problem with the applications of SPME in bioanalysis continues to be the unavailability of commercial SPME coatings for polar and ionic analytes such as endogenous peptides, pharmaceutical drugs and their metabolites. As a solution, many researchers have decided to develop their own extraction phases. An overview of new extraction phases developed to address these difficulties is provided in Section 10.6.

11.3

Determination of Free Concentrations and Binding Constants

The first step in essentially all biological activities is an interaction between separate molecular constituents, the ligand and the receptor (usually a protein), to form a molecular complex. Such interactions play a vital role in all basic life sciences, including biochemistry, biophysics, pharmacology, physiology, immunology, endocrinology, neurobiology, molecular biology and cell biology.18 The freely dissolved concentration of the ligand is an important parameter in environmental chemistry, pharmacology and toxicology. In the environment, for example, the free concentration is the driving force for the transport, distribution and bioaccumulation of a chemical. In pharmacology and toxicology, it is generally accepted that only freely dissolved molecules can pass through the cell membranes and thus be effective in organisms.19 Drug binding to specific plasma transport proteins (albumin, α1-acid glycoprotein and lipoproteins) is an integral step of many other types of intermolecular interactions in a cell or an organism. If one aspires to carry out a thorough

Ligand
Receptor Binding and Determination of Free Concentrations

387

study of biological responses to molecular stimuli, the strength of the binding of a ligand to its receptor must be investigated. All these binding constants and interactions can be investigated by measuring either the concentrations of the bound or the free form.

11.3.1 Free Concentrations Several methods have been developed to measure the free concentration of compounds in a sample, most of which involve the physical separation of the free and bound fractions followed by a conventional analysis step. Examples of such techniques include equilibrium dialysis, ultrafiltration and gel filtration. These techniques are usually time-consuming, can suffer loss of analyte to membranes or can create a shift in the binding equilibrium during the separation.19,20 Recently developed chromatographic methods21 allow only for the assay of the fraction of the drug that is bound to proteins (a value that is known to fluctuate with drug and protein concentration), and the mobile phase that is used (50 mM ammonium acetate pH 7 with 4% or 20% isopropanol) is very different from physiological conditions. The same is true for methods based on electrospray ionisation mass spectrometry (MS),22 which are effective only when certain buffer solutions are used and when the combination ratio between the receptor and the ligand is 1:1. When this combination ratio is different, the developed equations cannot be applied, and the interpretation of the resulting mass spectra becomes very difficult. While chromatographic methods used for the assay of free concentrations and binding constants assume a very fast equilibrium (less than a few seconds) between the ligand and the receptor, ultrafiltration and ultracentrifugation techniques assume a slow equilibrium (more than 30 min), so the free fraction can pass through a membrane without shifting the equilibrium in the other compartment. The elegant solution offered by the flow-dialysis techniques23 suffers from the necessity of using radiolabelled ligands and is applicable only when the combination ratio is 1:1. Electrophoretic methods have been applied for both 1:124 and 1:n25 combination ratios between the receptor and the ligand, but they are restricted to certain buffer solutions and do not allow for precise control of the temperature. While electrospray ionization mass spectrometry (ESI
MS), capillary electrophoresis (CE) and chromatographic methods are effective only when the samples are dissolved in certain buffer solutions, the SPME method can be used for extraction from any media, and at any concentration range, through the selection of a suitable extraction phase. SPME is proposed as a new technique for the extraction and concentration of the target compounds from a complex matrix in order to determine their free concentration.5,14,19,20,26
29 Because of the small size of the extractive phase and partial extraction approach, this technique allows the simultaneous analysis of the analyte and the matrix. Compared to other methods, SPME offers several advantages: small sample size, short analysis time, possibility to automate and ability to study complex samples (e.g. whole blood) directly. The study of binding equilibria is no longer restricted to certain buffer solutions, but can be performed in any ‘natural’ environment by optimising the experimental conditions. The SPME

388

Handbook of Solid Phase Microextraction

SPME fibre

SPME fibre

Ligand bound to receptor – [R(L)b] Free receptor Free ligand – Cf – [R] Ki

Kfs

+

Vortex

Ligand absorbed onto ADS–SPME fibre – m

Figure 11.2 Schematic representation of experimental setup for the determination of free concentrations and binding constants. (Source: Reprinted from Ref. 13 with permission of American Chemical Society. r 2005 American Chemical Society.)

technique can be applied by exposing the fibre in the headspace above the sample or directly in the sample (direct immersion). The advantage of headspace SPME is that the matrix in the sample cannot interfere with the fibre, but it is applicable only for volatile compounds in a non-volatile matrix. With the introduction of new extracting phases (restricted access materials (RAMs),8,13 molecularly imprinted polymers (MIPs)30,31 and fibres with immobilised antibodies),7,11,32 SPME offers improved accuracy and the selective separation of small ligand molecules from larger receptor molecules (matrix, proteins) in any sample. While most applications of SPME are developed to achieve the highest possible extraction efficiency, Kopinke et al.33 and Vaes et al.27 have independently introduced a specific application of SPME to measure free concentrations based on a negligible extraction of the free concentration. Because this application causes negligible depletion of the free concentration, it has been named negligible depletion SPME (nd-SPME). Nevertheless, measurement of free concentrations with SPME can be performed both in negligible or significant depletion conditions. For example, the free concentration of lidocaine in plasma was measured successfully without applying nd-SPME.26 Many theoretical approaches have been designed for the determination of free concentrations by SPME. One of the simplest methods is based on calculating the free concentration as the ratio between the amount of analyte extracted and the fibre constant.9 Briefly, in the presence of an SPME fibre, an amount m (in moles) of a drug is extracted from the solution, and this amount that is extracted by the fibre will be in equilibrium with the free concentration (Figure 11.2). The free concentration of drug remaining in the solution is then given by Cfree 5

m fc

ð11:1Þ

Ligand
Receptor Binding and Determination of Free Concentrations

389

Concentration (ng/mL)

Free concentration (ng/mL)

30.00 25.00 20.00 15.00

10.00 8.00 6.00 4.00

Diazepam Nordiazepam Oxazepam

2.00 0.00 1.5

2.5

3.5

4.5

5.5

6.5

7.5

8.5

Time (h)

10.00 5.00 0.00 0

1

2

3

4

5

6

7

8

9

Time (h)

Figure 11.3 Comparative pharmacokinetic profile for the free concentration of diazepam, nordiazepam and oxazepam in beagles (obtained by in vivo microextraction). (Source: Reproduced from Ref. 12 with permission of AACC. r 2006.)

where fc is the fibre constant and represents the product of the partition coefficient of the drug (between fibre and solution without binding matrix) and the volume of the fibre (for liquid coatings) or the active surface of the fibre (for solid coatings). By using special materials for the extracting phase, the large receptor molecules are prevented from being co-extracted. Such an approach is equally applicable for negligible and non-negligible depletions, applies for any combination ratio between the ligand and the receptor, and is independent of the analysis method. The quantification of the amount of ligand extracted may be performed by any method that can be coupled to SPME, including liquid chromatography (LC), gas chromatography (GC), MS, electrophoresis or radiometry, to name just a few. The method was applied successfully for the in vivo and in vitro determination of free concentrations during pharmacokinetic studies.6,10 Whole blood concentration and free concentrations, easily obtained with SPME, are of utmost importance in therapeutics because they correlate with the pharmacological effect and are more significant than plasma concentrations. Figure 11.3 presents the free concentration profile for diazepam and two of its metabolites in beagle whole blood and reveals a fast in vivo conversion of diazepam to oxazepam, exhibiting similar free concentration values. The free concentrations were determined by calibration with standard solutions of benzodiazepines in phosphate-buffered saline pH 5 7.4, while the total concentrations were determined by calibration with standard solutions in whole blood. Such results for total concentration are meaningful when the blood composition does not change significantly during a pharmacokinetic study. Conversely, reliable measurements of the free concentration can be obtained even when the concentration of plasmatic proteins changes because the amount of analyte extracted by SPME is inherently related to the free concentration. This is an important advantage because the free concentration is a valuable parameter in pharmacology applications.

390

Handbook of Solid Phase Microextraction

11.3.2 Study of Ligand
Receptor Binding The investigation of binding parameters has received significant attention since its importance was recognised at the beginning of the twentieth century, and dozens of research papers are published yearly on this topic. Different aspects of ligand
receptor interactions have been reviewed, including their molecular nature, biological functions, pharmacological significance, as well as methodological approaches applied and their potential shortcomings.18,34 When several ligand molecules can be bound by a receptor molecule, multiple equilibria are established. These can be described by different types of equilibrium constants, which reflect different perspectives in visualising the equilibria. The multiple equilibria for the binding of a ligand L by a receptor R with binding sites b may be formulated in terms of a stoichiometric analysis or on the basis of a site-oriented analysis.18,35 In the case of stepwise stoichiometric equilibria, the number of ligand molecules bound per receptor molecule (B) can be expressed as a function of the free ligand concentration (Cf), the total ligand concentration (Ct), the stoichiometric binding constants (ki) and the receptor concentration (Cm): B5

Ct 2 Cf k1 Cf 1 2k1 k2 Cf2 1 ? 1 bðk1 k2 ; . . . ; kb ÞCfb 5 Cm 1 1 k1 Cf 1 k1 k2 Cf2 1 ? 1ðk1 k2 ; . . . ; kb ÞCfb

ð11:2Þ

For the site-oriented approach, each site is considered to have a fixed, invariant affinity. Accordingly, stoichiometric binding constants (ki) are replaced with site binding constants (Ki), and the number of moles of ligand bound per moles of receptor is B5

b b X X Ct 2 Cf Ki Cf 5 Bi 5 Cm 1 1 Ki Cf i51 i51

ð11:3Þ

A special case that is often encountered involves a system with two classes of binding sites, each with identical invariant affinities that differ from the identical invariant affinities of the other class.18,25,36 Under these circumstances, Eq. (11.3) can be reduced to B5

Ct 2 Cf b1 K1 Cf b2 K2 Cf 5 1 Cm 1 1 K1 Cf 1 1 K2 Cf

ð11:4Þ

where b1 and b2 are the number of sites in the respective classes (b1 1 b2 5 b, the total number of binding sites), and K1 and K2 are the respective site binding constants. Equations (11.3) and (11.4), although frequently used because of their convenience, often lead to non-real (complex) solutions, especially when the binding constants increase with increased occupancy of the receptor, as is the case with many enzymes or carrier proteins (like haemoglobin). Regardless of the chosen

Ligand
Receptor Binding and Determination of Free Concentrations

391

mathematical model, the binding parameters are determined from pairs of Cf and Ct for a certain Cm. While Cm and Ct are usually known when a standard solution is prepared, Ct is usually determined with Eq. (11.1). The aforementioned method was successfully applied to obtain the binding curves of diazepam (a drug with a high binding constant
more than 105 L/mol) and isosorbide dinitrate (a drug with a low binding constant
less than 105 L/mol) to human serum albumin. The binding constants were subsequently used to calculate receptor and free and total ligand concentrations in synthetic samples.9 A similar approach proved successful for the determination of the binding constant of chlorhexidine to salivary proteins.10 When the amount of sample (receptor) is limited, the method of multiple extractions or multiple additions permits the generation of selected regions of the binding curve with a single small volume of receptor solution. Unlike other methods that assume a short or long equilibration time for the binding of ligands to receptors, SPME methods allow one to observe whether equilibrium has been reached during a certain period for each ligand and receptor pair. The test is performed by determining the period of time after which the amount of ligand extracted from a solution with a binding matrix reaches a constant value. This period represents either the extraction equilibration time or the binding equilibration time, whichever is longer, and can be subsequently used for all experiments. For efficient separation of small ligand molecules from large receptor molecules, the following types of extraction phases may be used: RAMs, MIPs, antibodies and membrane-protected coatings. When a liquid non-polar extraction phase is used, only non-polar molecules of ligand will be extracted. This is of no concern when analysing non-polar ligands but may lead to poor reproducibility and accuracy when analysing polar or ionisable compounds; in this case, care must be taken to ensure that all standard and sample solutions have the same ionic strength and pH. The mathematical model based on Eqs (11.1)
(11.4) works equally well with negligible and non-negligible extractions. In the case of non-negligible extractions, reliable studies may be performed even for drugs with a high binding constant because some of the ligand bound to the receptor will move into the extraction phase, increasing the sensitivity. In this case, a fibre with a high fc for the target ligand should be used. This is an important advantage over dialysis, where the affinity of the acceptor buffer solution cannot be easily changed. Recently, ligand
receptor binding studies have been automated using Concept 96 robotic sample preparation station (PAS Technology) described in detail in Chapter 5.37 The use of this system allows the preparation of all samples simultaneously in order to construct a binding curve, thus increasing sample throughput. Considering the fact that up to 96 samples can be prepared at one time, multiple ligands or receptors can be studied in parallel. To date, the performance of automated SPME for binding studies was shown using diazepam and human serum albumin as a model, but additional studies are currently ongoing. A detailed procedure regarding how to perform automated binding studies properly is included in Chapter 13. The automation of ligand
receptor binding studies using SPME

392

Handbook of Solid Phase Microextraction

is an especially important advance because the full automation of traditional methods such as equilibrium dialysis has not been achieved in 96-well plate format. For example, semi-automated 96-well equilibrium dialysis apparatus proposed by Banker et al.38 required 8 h to reach equilibrium for 10 studied drugs, including diazepam. In contrast, automated SPME required only 60 min of total sample preparation time.

11.3.3 Determination of Drug Plasma Protein Binding Determining the amount of drug binding to plasma proteins is an essential step in both drug discovery and in clinical phases of drug development.38
41 Plasma protein binding (PPB) affects the amount of drug available for diffusion into target tissues, such as the brain,42
44 the calculation of in vivo hepatic clearance45 and the interpretation of the drug’s bioavailability.39 Due to the important clinical implications of PPB data and its role in characterising a drug’s behaviour and proper dosing, there is an increased need to make this measurement as early as possible in the discovery process in order to understand drug disposition and to optimise individual drug therapy. Although the main drugbinding proteins are albumin and α1-acid glycoprotein, plasma contains many other proteins; consequently, there is a high probability that many small molecules will exhibit some levels of binding. To determine the extent of PPB, the molecule should be tested directly in a protein-binding assay using plasma or serum. This is a critical step in characterising the distribution of a small molecule with respect to the plasma compartment.21,42,46 The determination of PPB by SPME is based on determining the free concentration of drug in the presence of plasma proteins.5 When extraction from a volume V of plasma containing binding proteins is performed, an amount mplasma of drug will be extracted by the fibre coating. Considering that the initial concentration of the drug is C0 plasma, the total final concentration of the drug in plasma is then given by Ctotal 5 C0 plasma 2

mplasma V

ð11:5Þ

The free concentration of the drug in plasma is readily calculated by applying Eq. (11.1) to plasma samples: Cfree plasma 5

mplasma fc

ð11:6Þ

Finally, the percentage of drug binding to plasma proteins (PPB) is calculated from the total and free concentrations of the drug: PPB% 5

Ctotal 2 Cfree plasma 3 100 Ctotal

ð11:7Þ

Ligand
Receptor Binding and Determination of Free Concentrations

393

Table 11.2 Changes in Plasma pH with Experimental Conditions Experimental Conditions

pH Value

Raw plasma samples (up to 2 weeks at 4 C) Plasma diluted 1:1 with isotonic PBS buffer Plasma diluted 1:10 with isotonic PBS buffer Plasma incubated 1 h with 10% CO2 Physiological pH of plasma

8.00
8.40 7.60
7.70 7.40
7.50 7.50
7.60 7.35
7.42

Source: Reproduced from Ref. 11 with permission of Wiley (r 2006).

This method of calculation can be easily extended to determine the PPB in diluted plasma samples.5 Highly bound drugs produce low free concentrations when undiluted plasma is used; accordingly, analytical methods for assaying these drugs may have poor reproducibility. When plasma samples are diluted, the free concentration of drug increases, leading to more sensitive assays and better reproducibility. It has long been recognised that plasma pH has a significant influence on the extent of binding to plasmatic proteins. Plasma pH can change considerably during storage and even during short-term incubation. Conversely, most binding studies do not employ any means for controlling the pH of plasma samples. Usually, the binding of acid drugs is minimally influenced by experimental conditions and pH, whereas the binding of basic drugs is more susceptible to changes in the experimental conditions. This large variation of binding values, in the case of basic drugs, may be explained if the influence of pH on the degree of ionisation is considered. At pH 5 7.4, acidic drugs are ionised almost completely, and a small change in pH will not affect the degree of ionisation. In the case of basic drugs, a small change in pH causes a significant change in the ionised fraction of the drug. For analytical methods based on SPME, changes in plasma pH can also induce changes in the amount of analyte extracted. Accordingly, control of the plasma pH is vital for accurate determinations of the target analyte. Table 11.2 presents a summary of the different methods that can be used to stabilise the pH of plasma, with the corresponding pH values. Obviously, pH values closest to physiological conditions are obtained in the case of incubation with 10% CO2 or 1:10 dilution with isotonic phosphate buffered saline (PBS). Incubation in 10% CO2 atmosphere offers an environment that closely mimics physiological conditions and should be used whenever possible. Otherwise, the 1:10 dilution method with isotonic PBS should be employed, in which case the concentration of the drug must be monitored to ensure it is at least 10 times lower than that of the protein concentrations in the diluted plasma. Representation of drug binding as a percentage of the bound fraction leads to substantial compression of results for strongly bound drugs and to broadening of results for weakly bound drugs. The results obtained with different techniques could be compared more easily if PPB values obtained for a specific total

394

Handbook of Solid Phase Microextraction

Experimental pD

Warfarin

9.5

Ibuprofen

8.5

Verapamil Propranolol

7.5

Caffeine

6.5 5.5 5.5

6.5

7.5

8.5

Literature min Standard approach Dilution 1:1 Dilution 1:10 CO2 headpressure Literature max

9.5

Average pD from literature

Figure 11.4 Correlation of experimentally obtained pD values with average values from literature, at a total drug concentration of 1 μM (pD 5 2 log Cfree). (Source: Reproduced from Ref. 11 with permission of Wiley. r 2006.)

concentration are transformed to free concentrations. Figure 11.4 presents the results of a recent PPB study based on SPME, for a total drug concentration of 1 μM. All experimental values are close to the diagonal of the graph, indicating a good correlation with average literature values.5

11.4

Calibration of SPME for Bioanalytical Applications

To date, several calibration approaches have been developed for SPME (see Chapter 6). Equilibrium extraction is the most frequently used method, which involves using a known distribution constant or an external calibration curve to correlate the amount of analyte extracted by the SPME fibre to its concentration in the sample: n 5 Kfs UVf UC0

ð11:8Þ

where n is the number of moles of analyte extracted, Kfs is the distribution coefficient of the analyte between the fibre coating and the sample matrix, C0 is the concentration of a given analyte in the sample and Vf is the volume of the fibre. The product KfsVf, the fibre constant, should be determined in solutions without binding matrix. Subsequently, fc can be used to determine free concentrations, binding constants and PPB when extractions are performed from solutions containing binding matrix. To shorten equilibrium extraction times, address the displacement effects that occur when porous coatings are used, or both, extraction can be interrupted before equilibrium. Even though extraction equilibrium is not reached, there is still a linear relationship between the amount of analyte extracted onto the fibre and the analyte concentration in the sample matrix, provided that the agitation, the extraction time and the extraction temperature remain constant28: n 5 Kfs UVf UC0 ð1 2 e 2 aUt Þ

ð11:9Þ

Ligand
Receptor Binding and Determination of Free Concentrations

395

where t is the extraction time, and a is a time constant, representing the speed at which an equilibrium can be reached. This method eliminates the use of conventional calibration curves. Fast on-site analysis and long-term monitoring are thus possible.3 While performing derivatisation at the same time with extraction, if the reaction is the rate-limiting step, the first-order reaction rate constant (Kr1) can be used for calibration: n 5 Kr1 UC0

ð11:10Þ

In addition to these ‘classic’ calibration methods, the newly developed method of ‘kinetic calibration’ appears to be particularly useful for in vivo determinations. When an SPME coating that is preloaded with a standard compound is exposed to an agitated sample matrix, desorption of the compound from the fibre occurs. The desorbed compound diffuses through the boundary layer into the bulk of the sample matrix. The amount Q of standard remaining on the coating after time t can be described as Q 5 q0 Ue 2 aUt

ð11:11Þ

where q0 is the initial amount of standard present on the fibre. The constant a in Eq. (11.9) for the absorption has the same definition as constant a in Eq. (11.11) for the desorption, and should have the same value for both the absorption and the desorption of an analyte, under the same experimental conditions (sample bulk velocity and temperature). The isotropy of absorption and desorption in SPME allows for the calibration of absorption (n) using desorption (Q): n Q 1 51 n0 q0

ð11:12Þ

where n0 is the amount of analyte extracted at equilibrium.6 This is especially important for the calibration of on-site and in vivo analyses because the control of the agitation conditions of the matrix is sometimes difficult and direct spiking of standards into the matrix is typically not possible. In addition to convenient applications for the determination of total concentrations of drugs and biomolecules, SPME was also introduced as a new technique for the determination of free concentrations (as shown previously in Figure 11.2). In this case, calibration is usually based on the fibre constant, which represents the product of the partition coefficient of the analyte (between fibre and sample) and the volume of the fibre (for liquid coatings) or the active surface of the fibre (for solid coatings). The fibre constant may be easily determined by extracting the analyte from standard solutions in water or buffer when the total drug concentration is considered to be equal to the free concentration.

11.5

Conclusions

SPME is a simple, solvent-free and reliable microextraction technique that has continued to revolutionise sampling and sample preparation since its discovery a

396

Handbook of Solid Phase Microextraction

decade ago. The small dimensions of SPME devices and their solvent-free feature enable convenient sampling for bioanalytical applications, such as the analysis of biological samples, the measurement of free concentrations and the determination of binding constants. The assays described in this chapter can be applied easily to determine the blood to plasma distribution ratio of various drugs. Furthermore, analysis of extracted compounds can be performed with highly specific instruments, such as GC
Ms or LC
Ms/MS. To date, SPME has been successfully applied around the world to a wide range of bioanalytical investigations, clearly demonstrating that the technique provides an excellent alternative to current sample preparation methods. The development of biocompatible extraction phases for SPME has led to significant advances in bioanalysis: all sample preparation steps can be combined into a single one, even for complex biological samples such as whole blood or plasma. Furthermore, biocompatible devices permit the direct extraction of target analytes from the flowing blood of living organisms. Future research in this area should focus on applications for soft tissues, automation of sampling and analysis, and utilisation of highly specific extraction phases. Direct extraction of target analytes from complex biological samples often results in co-extraction of interference compounds. In the case of solid (porous) coatings, co-extraction is an important issue because the interferant may displace the analyte from the extraction phase. In addition, analysis of a complex extract requires a suitable separation method, such as GC or LC, which increases the total processing time. An important aspect of the future application and growth of SPME is the development of new extraction coatings. An obvious choice is the application of extraction phases that are specific for the target compound. SPME devices based on MIPs or antibodies would possess unsurpassed specificity and would be especially useful at very low concentrations of target analyte. Such devices could be interfaced directly with mass spectrometers, resulting in very fast analytical methods.

References 1. SR Mikkelsen & E Corton, Eds, Bioanalytical Chemistry (2004) John Wiley and Sons: Hoboken, NJ, 361 2. CK Larive, Anal Bioanal Chem 382 (2005) 855 3. J Pawliszyn, Ed, Sampling and Sample Preparation for Field and Laboratory: Fundamentals and New Directions in Sample Preparation Vol. 37 (2002) Elsevier: Amsterdam 1131 4. H Kataoka, Trends Anal Chem 22 (2003) 232 5. FM Musteata, J Pawliszyn, MG Qian, JT Wu & GT Miwa, J Pharm Sci 95 (2006) 1712 6. FM Musteata, ML Musteata & J Pawliszyn, Clin Chem 52 (2006) 708 7. HL Lord, M Rajabi, S Safari & J Pawliszyn, J Pharm Biomed Anal 40 (2006) 769 8. FM Musteata, M Walles & J Pawliszyn, Anal Chim Acta 537 (2005) 231 9. FM Musteata & J Pawliszyn, J Proteome Res 4 (2005) 789

Ligand
Receptor Binding and Determination of Free Concentrations

397

10. FM Musteata & J Pawliszyn, J Pharm Biomed Anal 37 (2005) 1015 11. NA Guzman, Electrophoresis 24 (2003) 3718 12. H Lord, R Grant, M Walles, B Incledon, B Fahie & J Pawliszyn, Anal Chem 75 (2003) 5103 13. WM Mullett, K Levsen, D Lubda & J Pawliszyn, J Chromatogr A 963 (2002) 325 14. YL Hsu, S Rose & KG Furton, 223rd ACS National Meeting 2002 143 15. S Ulrich, J Chromatogr A 902 (2000) 167 16. MEC Queiroz & FM Lancas, LC-GC Europe 18 (2005) 145 17. G Theodoridis, EHM Koster & GJ de Jong, J Chromatogr B 745 (2000) 49 18. IM Klotz, Ligand
Receptor Energetics: A Guide for the Perplexed (1997) Wiley: New York, NY, 170 pp 19. MB Heringa & JLM Hermens, Trends Anal Chem 22 (2003) 575 20. MB Heringa, D Pastor, J Algra, WHJ Vaes & JLM Hermens, Anal Chem 74 (2002) 5993 21. Y Cheng, E Ho, B Subramanyam & J-L Tseng, J Chromatogr B 809 (2004) 67 22. A Tjernberg, S Carnoe, F Oliv, K Benkestock, PO Edlund & WJ Griffiths, et al. Anal Chem 76 (2004) 4325 23. G Veldhuis, EPP Vos, J Broos, B Poolman & RM Scheek, Biophys J 86 (2004) 1959 24. M Nilsson, V Harang, M Bergstrom, S Ohlson, R Isaksson & G Johansson, Electrophoresis 25 (2004) 1829 25. J Ostergaard, C Schou, C Larsen & NHH Heegaard, Electrophoresis 23 (2002) 2842 26. EHM Koster, C Wemes, JB Morsink & GJ de Jong, J Chromatogr B 739 (2000) 175 27. WHJ Vaes, EU Ramos, HJM Verhaar, W Seinen & JLM Hermens, Anal Chem 68 (1996) 4463 28. J Pawliszyn, Ed, Applications of Solid Phase Microextraction (1999) Royal Society of Chemistry: Cambridge, 655 pp 29. M Krogh, K Johansen, F Tonnesen & KE Rasmussen, J Chromatogr B 673 (1995) 299 30. WM Mullett & J Pawliszyn, J Sep Sci 26 (2003) 251 31. EHM Koster, C Crescenzi, W den Hoedt, K Ensing & GJ de Jong, Anal Chem 73 (2001) 3140 32. H Yuan, WM Mullett & J Pawliszyn, Analyst 126 (2001) 1456 33. FD Kopinke, J Porschmann & M Remmler, Naturwissenschaften 82 (1995) 28 34. J Oravcova, B Bohs & W Lindner, J Chromatogr B 677 (1996) 1 35. IM Klotz & DL Hunston, J Biol Chem 250 (1975) 3001 36. T Kosa, T Maruyama & M Otagiri, Pharm Res 14 (1997) 1607 37. D Vuckovic & J Pawliszyn, Automated study of ligand–receptor binding using solidphase microextraction J Pharm Biomed Anal 50 (2009) 550
555 38. MJ Banker, TH Clark & JA Williams, J Pharm Sci 92 (2003) 967 39. I Kariv, H Cao & KR Oldenburg, J Pharm Sci 90 (2001) 580 40. RE Olson & DD Christ, Annu Rep Med Chem 31 (1996) 327 41. S Sarre, K Van Belle, SI Smolder, G Krieken & Y Michotte, J Pharm Biomed Anal 10 (1992) 735 42. J Blodgett & J Lynch, Millipore Application Note, AN1732EN00, 2003 43. QR Smith, C Fisher & DD Allen, The role of plasma protein binding in drug delivery to brain, 44th OHOLO Conference (2001) 311 44. H Tanaka & K Mizojiri, J Pharmacol Exp Ther 288 (1999) 912 45. EN Fung, Y-H Chen & YY Lau, J Chromatogr B 795 (2003) 187 46. J Schuhmacher, K Buhner & A Witt-Laido, J Pharm Sci 89 (2000) 1008

12 In Vivo Sampling with

Solid-Phase Microextraction

Florin Marcel Musteataa and Dajana Vuckovicb a

Department of Pharmaceutical Sciences, Albany College of Pharmacy and Health Sciences, Albany, NY, USA b Donnelly Centre for Cellular and Biomolecular Research, University of Toronto, Toronto, Ontario, Canada

12.1

Introduction

Generally, in vivo research is more suited to observe an overall effect than in vitro research, which is better suited to deduce mechanisms of action. In vivo research has the advantage that the experimental system is a more complex biological system and gives a better indication of what will happen in the real world. Usually, in vitro samples do not correspond to the combination of compounds produced by an undisturbed live organism. For example, it has been shown that the composition of the volatile extracts collected from detached or damaged plants can differ significantly from the mixture emitted by the live, undamaged specimen.1 Reliable and accurate analytical methods are indispensable for in vivo research. On the other hand, the development of techniques appropriate for in vivo analysis poses significant difficulties due to the low and unceasingly changing concentrations of target analytes in complex biological samples. An ideal in vivo sampling technique should be portable, solvent-free and offer integration of the sampling, sample preparation and analysis. Simplified procedures and a reduction in sampling errors are important advantages for performing as many chemical analysis steps as possible at the site where a sample or subject is located. Current techniques that are applicable for in vivo analysis include microdialysis, ultrafiltration, solid-phase microextraction (SPME), sensors and arrays of sensors, microfluidics and nanotechnology.2 Microdialysis is a powerful sampling technique based on semipermeable membranes implanted into the tissue of interest and perfused at a low flow rate. Although it has found numerous and various in vivo applications,3 6 microdialysis has many significant problems, of which the most important are the loss of perfusion fluid, complicated calibration and the difficulty to obtain reliable quantitative data.7,8 Ultrafiltration is a filtrate selection method that avoids complicated and time-consuming recovery calculations necessary when in vivo microdialysis is used,9 but the membranes are prone to clogging and the method is not suitable for compounds that are highly bound to plasmatic proteins. Both ultrafiltration and microdialysis require the presence of tubing and pumps. Sensor arrays are unequalled for parallel measurements of many analytes; most of Handbook of Solid Phase Microextraction. DOI: 10.1016/B978-0-12-416017-0.00012-7 © 2012 Elsevier Inc. All rights reserved.

400

Handbook of Solid Phase Microextraction

them contain hundreds of micrometre-sized features. Despite their advantages, such arrays are very difficult to produce, may not be suitable for complex biological samples and are mostly found in research laboratories. Microfluidic systems use small sample and reagent volumes coupled with integrated detection methods. Such chips have been crafted for DNA analysis, immunoassays, cell analysis and enzyme-activity measurements. For now, most of these systems remain more complicated and bulkier than a simple integrated miniature device because in many cases, external optics, pumps and detectors are required to control and read out signals from the chips. In addition, liquid reagent reservoirs must be incorporated into the systems so that they can be used for multiple samples; consequently, reagent stability and applicability for in vivo analysis remain problematic.2 Use of nanomaterials in biosensors allows the use of many new signal transduction technologies in their manufacture. Because of their submicron size, nanosensors, nanoprobes and other nanosystems are revolutionising the fields of chemical analysis.10 However, despite their numerous advantages, nanomaterials have significant cytotoxicity and their applicability to in vivo measurements is insufficiently developed.11,12 One of the most promising techniques for rapid sample preparation and subsequent analysis is SPME. This sampling procedure causes minimal disturbances to the investigated system, as no liquid and only small fractions of analytes are removed. An in vivo sampling approach can eliminate errors and reduce the time associated with sample transport and storage, and therefore it can result in collecting more accurate and precise analytical data.13 In vivo analysis is a special application area where SPME is gaining ground because of its unique characteristics: on-site sampling, easy extraction and analysis of the whole extracted amount. Furthermore, in vivo SPME permits repeated sampling of the same individual over time, which can be invaluable for some applications for example, in vivo SPME was used to monitor pheromone changes in worker bees to become dominant.14 Early in vivo investigations focused on fragrances and odours emitted by insects, fungi and bacteria. These investigations were extended to biogenic volatile organic compounds (VOCs) emitted by animals and plants. Furthermore, SPME can also be a useful tool to study various biotransformation processes as well as for metabolomics-type applications. The applicability of SPME is no longer limited to sampling only volatile compounds, and new applications for sampling polar and non-volatile compounds are now emerging. For example, SPME can be used to study the fate of pesticides in plants. More recently, the SPME technology was used for in vivo analysis of intravenous drug concentrations in living animals15 18 and pharmaceuticals in fish to monitor bioaccumulation and toxicity.18 For the majority of in vivo applications of SPME, the sensitivity and precision provided by SPME were comparable or better than those of the techniques traditionally employed for the same samples. Moreover, some applications would not be feasible using other sample preparation methods because they would cause severe damage to the living organisms or would demand their sacrifice.1 This chapter aims to present recent advances and future trends in SPME method development for analysis of endogenous and exogenous compounds in live organisms, including animals, plants and microorganisms. The main objective of the chapter is to demonstrate that in vivo SPME is an extremely useful tool for life

In Vivo Sampling with Solid-Phase Microextraction

401

sciences research because it can provide more accurate and complete picture of processes occurring in a biological system as well as to give the reader an overview of the variety of types of applications possible.

12.2

In Vivo Method Development

SPME eliminates or minimises the use of organic solvents, substantially shortens the total time of analysis and allows convenient automation of the sample preparation step. It can integrate sampling with sample preparation, which makes it suitable for on-site analysis and process monitoring. The nature of target analytes and complexity of sample matrix determine the level of difficulties in performing a successful extraction. In the case of in vivo experiments, special consideration should be given to the selection of extraction mode, extraction phases and the most appropriate calibration procedure. Table 12.1 presents an overview of the most recent applications of in vivo SPME for the sampling of animal, plant and microbial volatile and semi-volatile emissions along with the main experimental details. Later in this chapter, Table 12.4 shows similar applications of in vivo SPME for sampling human emissions, while Table 12.5 shows applications of in vivo SPME for studying circulating and tissue concentrations of various compounds in plants and animals.

12.2.1 In Vivo Sampling The extraction mode of SPME is selected on the basis of the sample matrix composition, analyte volatility and its affinity to the matrix. For very complex samples, the headspace, fibre protection and biocompatible or restricted-access coatings should be selected. For clean matrices, both direct and headspace sampling can be used. The main factor to be considered when selecting an extraction mode is the volatility of the target analytes. Highly or moderately volatile compounds are usually determined in the headspace, and this has been the preferred method for most applications to date (Tables 12.1 and 12.4). For headspace sampling, the investigated individuals, plant parts or insects, are enclosed in capped vials (Figure 12.1) or special chambers or bags that are usually designed to minimise interferences (Figures 12.2 and 12.3). For example, for sampling of plant volatile emissions leaves of plant under study were enclosed in a customised Teflon bagt that had two ports (one for the insertion of SPME device and one for the insertion of tools to injure the leaves under study in order to study ‘wound volatiles’).34 The bag was held in place by twist ties. This type of enclosure device also allowed the injection of internal standard so that each in vivo SPME sampling could be normalised to the amount of internal standard extracted. Alternatively, an appropriate plant part (e.g. a leaf or a flower) were enclosed using a glass funnel, as shown in Figure 12.2. Headspace sampling of human skin can be performed either using the specially designed device shown in Figure 12.3,37 using an alternative device proposed by

Analytes

Live Sample

Extraction Mode and Time

Extraction Phase

Calibration and/or Identification

CHs

Insects (Periplaneta americana) Insects ants (Leptothorax acervorum, L. gredleri, Dinoponera quadriceps) Insects (Apis mellifera)

Headspace, 12 h Direct contact with abdomen, 0.25 h

PDMS 100 μm PDMS/DVB 65 μm PDMS 30 μm PDMS 7 μm

19 External calibration curve based on standard compounds Qualitative data only, based on 20 23 MS libraries and standard compounds

Headspace, 0.2 h

PDMS/DVB 65 μm

Headspace, 0.25 h

PDMS 100 μm Pencil lead

24 Semi-quantitative based on relative peak area Qualitative data based on MS libraries and standard compounds Qualitative data only, based on 25 MS libraries.

CHs and fatty acids

Volatile compounds

Defensive volatile Insects (Graphosoma chemicals lineatum) Insects (Diprion pini)

Headspace, 0.2 5 h

PDMS/DVB 65 μm

Sex pheromones

Insects (Cossus insularis)

Headspace, 16 h

PDMS 100 μm

Sex pheromones

Insects praying mantid Headspace, 10 12 h (Spodromantis lineola) Defensive volatile Insects viceroy butterfly Headspace chemicals (phenolic) Pheromones Honeybee (Apis mellifera) Direct contact with mandibles, 3 min

PDMS 100 μm Stableflex PDMS/DVB 65 μm Silicone tubing

External calibration curve. Qualitative data based on MS libraries Qualitative data only, based on MS libraries and standard compounds Identification based on MS libraries Qualitative data only, based on MS libraries retention index and standard compounds External calibration

Reference

26,27

28

29 30

14

Handbook of Solid Phase Microextraction

Pheromones

402

Table 12.1 Recent In vivo Non-invasive Sampling Methods for Plant, Microbial and Animal Emissions Based on SPME (Selection)

Volatile organic compounds odourants

Insects (Harmonia axyridis)

Headspace, 24 h 124 h pre-incubation

Pheromones and defensive volatile chemicals

Insects (Paratrechina longicornis)

Targeted metabolomics Volatile organic compounds (VOCs)

Plants (Petunia hybrida)

Direct contact, PDMS/DVB 65 μm 20 60 min (ants freely walk over the fibre) 5 min (fibre attacked by ants) Headspace, 30 min PDMS 100 μm fibre

Metabolomics

Plants (Centaurea solstitialis)

Fungi (Fusarium sambucinum, Fusarium sporotrichiodes and Fusarium graminareum) Biotransformation Fungi (Asperigillus niger, Asperigillus tubingensis, of citronellol to rose oxide Penicillium digitatum and Penicillium roqueforti)

Headspace, 60 min

Headspace, 30 min

Headspace, 30 min

31

Qualitative data only, based on 32 MS libraries and standard compounds

Qualitative analysis using MS libraries and standards PDMS 100 μm fibre Internal standard normalisation and qualitative analysis using MS libraries retention index and standard compounds PDMS 100 μm fibre Qualitative analysis, PDMS stir bar (24 μL) identification based on MS libraries, retention index and standard compounds DVB/Carboxen/PDMS Identification based on MS 50/30 μm libraries and retention index

33

In Vivo Sampling with Solid-Phase Microextraction

In situ derivatisation with bis(trimethylsilyl) trifluoroacetamide DVB/carboxen/PDMS External calibration, 50/30 μm qualitative analysis using MS libraries and standard compounds Olfactory analysis

34

35

36

403

404

Handbook of Solid Phase Microextraction

Figure 12.1 Sampling of surface cultures of Fusarium graminareum using (1) headspace SPME (HS-SPME) and (2) headspace stir-bar sorptive extraction. The stir bar was held in place using a paper clip (magnetic force). (Source: Reprinted with permission from Elsevier. r 2004.35)

Figure 12.2 Experimental design for HS-SPME of petunias in vivo.34 A glass funnel was placed around a developing flower, thereby minimally interfering with its development. Aluminium foil was placed over both sides of the funnel to limit air movement. The SPME device was inserted through the aluminium foil, exposing the PDMS fibre to the volatile compounds present in the floral headspace. (Source: Reprinted with permission from Elsevier. r 2003.34)

Bicchi et al.38 or by using an even simpler setup that places a glass funnel on top of the skin, as proposed by Gallagher et al.39 For most breath analysis applications, the specimen is typically collected using Tedlar bags followed by the exposure of SPME fibre to its contents.40 42

In Vivo Sampling with Solid-Phase Microextraction

405

Figure 12.3 Sampling chamber for volatile compounds. (Source: Reproduced from Ref. 37 with permission of Elsevier. r 2005.)

In the case of polar and non-volatile analytes, only direct extraction is feasible. To sample volatile and non-volatile emissions, the fibre coating can be placed directly in contact with the body part under study and carefully rubbed against the sample 14,19,20,22,23 or placed in direct contact with human skin.38,43 45 Witte et al.32 placed the polydimethylsiloxane/divinylbenzene (PDMS/DVB) fibre directly in the path of an ant colony, so direct contact was achieved by ants walking over the surface of the fibre. This in vivo SPME approach allowed the identification of two main chemicals, which also were identified by traditional sample preparation method (extraction of gland tissue using dichloromethane) as long as in vivo SPME extraction time was sufficiently long. Clearly, in vivo SPME is a very useful alternative sample preparation method for biologists because it does not require subject sacrifice. Furthermore, depending on the size of the living system under study, SPME can permit the study of (i) temporal dynamics and/or (ii) spatial distribution on the same individuals, thus providing researchers with greater insights into the biology of the investigated system. Another useful configuration for direct contact types of application is the use of thin films or membranes, as shown, for example, in Figure 12.4 for sampling of human skin and carcinoma lesions.46 The increased surface area and extraction phase volume of these devices improves sensitivity and extraction kinetics, thus permitting shorter sampling times. For study of non-volatile and circulating concentrations of analytes, SPME coating can be inserted directly into a plant (stem, leaf and so on) or an animal (blood vessel or tissue) with a special in vivo device based on a hypodermic needle, which is shown schematically in Figure 3.16 in Chapter 3.15 This device is now commercially available from Supelco, and its photo is shown in Figure 3.17. The dimensions of this type of device are suitable for direct insertion into blood vessel (using a catheter) when sampling larger animals such as dogs. However, for sampling small animals such as rodents, the size of the blood vessels typically precludes a

406

Handbook of Solid Phase Microextraction

Figure 12.4 Sampling of human skin using PDMS membrane.46 After application with tweezers, a skin patch is covered with a cotton wool pad and secured with microporous tape to the skin surface. (Source: Reprinted with permission of Royal Society of Chemistry from Ref. 46.)

SPME holder Septum

Blood out SPME probe

Or Blood in

Biocompatible tubing

Figure 12.5 Interface for in vivo SPME sampling of rodents. The interface is surgically connected to the carotid artery and permits easy insertion of SPME probe through the septum.17 (Source: Reprinted with permission from Elsevier. r 2008.)

direct insertion of SPME probe. To address this issue, a polyurethane sampling interface for use in rodents has been designed, as shown in Figure 12.5. Y-shaped interfaces, which were designed to allow the recirculation of blood to the animal and were prone to clotting problems, and it was not possible to maintain adequate blood flow through the interface for prolonged periods of time.17 The alternative approach, where only one tube was connected to a catheter, was found to work much better. Blood flow through the interface in this design was provided by manual push/pull with a syringe.17 This latter type of interface has been used successfully in rats17 and in mice after further miniaturisation.47,48 Recently, a new type of spatially resolved SPME device was designed to facilitate sampling of heterogeneous semisolid matrices such as animal tissue and plant compartments.49,50 This device is depicted schematically in Figure 12.6. The length

In Vivo Sampling with Solid-Phase Microextraction

1 mm

407

9 mm

5 mm 3 mm 1 mm (A)

(B)

(C)

Figure 12.6 The segmented, spatially resolved SPME fibre and the two-step desorption process in a 96-well plate for the extracted analyte. (A) The segmented fibre; (B) the fibre placed into 50 μL of methanol (100%) in a well to desorb the analyte from the first (lower) coating segment; (C) the fibre placed into another well containing 200 μL of methanol (100%) to desorb the analyte from the second (upper) coating segment.50 (Source: Reprinted with permission from American Chemical Society. r 2009.)

of the coating in this design is reduced from typical 10 15 mm down to 1 2 mm and the fibre has two or more discontinuous portions of coating in order to facilitate simultaneous sampling of spatially resolved areas within a living system. For example, this design was successfully applied to simultaneous sampling of fish muscle and adipose tissue, thus offering a novel and environmentally friendly way to determine the bioaccumulation of pharmaceuticals and organic pollutants directly in vivo without the need to sacrifice animals.50 Figure 12.7 summarises different extraction modes of SPME and the types of configurations most applicable to a given extraction mode.51 In addition, it provides the main advantages and disadvantages of the various SPME devices in order to aid the reader in the selection of the most appropriate device for a given application. Additional in-depth descriptions of all in vivo SPME configurations and applications to date can be found in recent review articles.49,51

12.2.2 Extraction Phases When headspace sampling is used, any extraction phase is suitable as long as it is robust and innocuous. Most applications to date use commercially available PDMS or PDMS/DVB. Nevertheless, custom-designed extractive phases such as pencillead fibres can prove to be very useful for certain applications.25 For applications relying on the direct contact between fibre and insect, Stableflex fibres are preferred because of the improved robustness and durability, as fused silica fibres can be accidentally broken by the insect during exposure. For applications that require sampling of a large number of individuals, Crewe et al.14 proposed the use of inexpensive silicone tubes. Although the extraction capacity of flexible silicone tubes was better, these tubes could not be used for sampling queen bee signal because

Direct contact (rubbing) Insects Skin

Commercial device (ease of use and availability)

408

Fibre

Increases sampling area Thin film Possible loss of volatile species

Volatile or semivolatile

Fibre Headspace Plants Microorganisms Insects Animals Breath

Internally cooled fibre

Commercial device (ease of use and availability)

Improves sensitivity

Increases sampling area

Type of analyte Thin film Possible loss of volatile species

Non-volatile, polar or thermally labile

Direct extraction Plants Animals

Minimises adverse reaction in vivo Improves spatial resolution for tissue sampling

High spatial resolution fibre Decrease of sensitivity

Figure 12.7 Summary of in vivo SPME sampling modes and device selection.51 (Source: Reprinted with permission from American Chemical Society. r 2011.)

Handbook of Solid Phase Microextraction

Biocompatible fibre

In Vivo Sampling with Solid-Phase Microextraction

409

they would destroy it. The authors successfully addressed this problem by inserting a piece of fused silica capillary to make the silicone tubing more rigid. Another disadvantage of proposed silicone tubing was the need to desorb the devices in solvent such as dichloromethane because they were not compatible with thermal desorption in gas chromatography (GC) due to their geometry and dimensions. In addition to the use of PDMS membranes or thin films46 described in the previous section, Soini et al. proposed the use of commercial PDMS stir bars (24 μL volume) to sample human skin emissions in metabolomics studies.43 45 The rapid sampling was performed by rolling the device over defined surface area of skin followed by automated analysis using thermal desorption GC MS. In the case of direct extraction of target analytes from biological samples or complex matrices, the process is usually hindered by various matrix effects such as fouling and disturbance of uptake kinetics. Interfering compounds or suspended particles can be adsorbed by the fibre coating during direct SPME. Consequently, they cause calibration problems and preclude fibre reusability. Two approaches are currently applied for effective direct extraction of low-molecular-weight compounds from complex liquid matrices. The first approach consists of using restricted-access materials or biocompatible extraction phases that selectively reject proteins.52,53 For the second approach, a hollow membrane is used to form a concentric sheath around a coated SPME fibre. The membrane blocks the access of large particles or proteins to the coating surface, while target analytes with low molecular weight diffuse through the membrane and reach the extraction phase.53 Direct in vivo sampling of flowing blood or animal tissue is much more demanding than conventional sampling: all the materials must be biocompatible and sterilisable, preferably by autoclaving (widely available and accepted). Biocompatibility refers to non-rejection of biological products or artificial devices that are in contact with a living tissue. Incompatibility can lead to toxic reactions or immunological rejection. A material can be considered biocompatible if the sum of adverse humoral and cellular reactions occurring during exposure is lower than for a reference material.54 The creation of non-fouling surfaces is one of the major prerequisites for microdevices for biomedical and analytical applications. Historically, the commercial fibres were not suitable for direct in vivo extraction, so customdesigned coatings based on restricted-access materials, polypyrrole (PPY), polyethylene glycol (PEG) and polyacrylonitrile (PAN; deposited as a thin layer on medical grade stainless steel wires), were applied for their well-known biocompatibility.15 17,53,55 These coatings, their preparation methods and performance characteristics are summarised in Table 12.2. More recently, Supelco has developed a new line of biocompatible SPME coatings suitable for in vivo use, as described in Section 4.3.5 in Chapter 4. Currently, only C18 coating is commercially available in biocompatible format, but additional coating chemistries can be expected to become available in future to help address the range of potential in vivo SPME applications.48,64,65,66 In addition to biocompatibility, one important parameter to consider for in vivo SPME applications is that typically multiple fibres are needed for a single in vivo experiment (e.g. to sample multiple time points or multiple animals or

410

Table 12.2 Overview of Biocompatible SPME Coatings, Their Preparation Methods and Properties.49 Preparation Method

Coating Thickness

Advantages

Disadvantages

Reference

PDMS

Commercial hollow tubing inserted onto stainless steel wire

165 μm

Ease of preparation

Poor extraction efficiency for polar compounds Coating not bonded to support Long equilibration times

18

PAN 1 C18 (or other suitable solid-phase extraction (SPE) sorbent)

Dip method on stainless steel wire

60 μm

Extraction efficiency decreased by approximately 15% after sterilisation by autoclaving

53

PEG 1 C18 (or other suitable SPE sorbent)

Dip method on stainless steel wire

60 μm

Not stable for sterilisation by autoclaving Inter-fibre reproducibility of 15 25%

15,55,56

PPY

Chemical oxidation of pyrrole monomer with ammonium

,5 μm

Limited evaluation performed

57

Excellent inter-fibre reproducibility 6% standard deviation Flexibility in selection of the sorbent Good mechanical and chemical robustness Suitable for sterilisation by autoclaving Excellent extraction efficiency Good inter-fibre reproducibility ,10% standard deviation Flexibility in selection of the sorbent Carryover ,1% Short equilibration times ( ,5 min) Better sensitivity than PPY Short equilibration times (2 min)

Handbook of Solid Phase Microextraction

Type of Coating

Suitable for direct coupling to ion-mobility spectrometry (IMS)

PPY

Electrodeposition on stainless steel or platinum

, 10 μm

Short equilibration times Suitable for autoclaving

PAN as a thin membrane

Covering commercial Carbowaxs templated resin (CW/TPR) or other types with a thin layer of biocompatible PAN

50 μm

Improved mechanical stability Extraction capacity similar to that of CW/TPR without PAN layer Good inter-fibre reproducibility (,10%)

Restricted-access material (RAM) ion-exchange diol silica

Immobilisation of alkyl diol silica with epoxy adhesive on stainless steel fibres

Not reported

Excellent inter-fibre reproducibility 6% standard deviation (n 5 5) Good reusability (.150)

Long equilibration times

RAM silica

alkyl diol

40 μm

alkyl diol

Excellent inter-fibre reproducibility 5% standard deviation (n 5 5) Good reusability (.50) Good reusability (.50)

Long equilibration times

RAM silica

Immobilisation of alkyl diol silica with Locktite 349 adhesive on silica fibres Immobilisation of alkyl diol silica with biocompatible epoxy adhesive in stir-bar format

Not reported

No detectable carryover

Poor inter-fibre reproducibility Adsorptive coating (limited linear dynamic range and possible displacement effects) Extraction efficiency decreased 30% after sterilisation by autoclaving Long equilibration times

15 17,58,59

53

Not tested in vivo

In Vivo Sampling with Solid-Phase Microextraction

persulphate on silica optical fibre

52

Not tested in vivo

Not tested in vivo Evaporation/reconstitution required for optimum sensitivity due to use of large desorption solvent volume Long equilibration times

60 62

63

411

(Continued)

412

Table 12.2 (Continued) Type of Coating

Preparation Method

Coating Thickness

Advantages

Disadvantages

Reference

Supelco biocompatible fibres (proprietary binder 1 C18 silica particles or other suitable sorbent)

Batch coating on flexible metal alloy

45 μm

Good chemical and mechanical stability Good inter-fibre reproducibility (,10%)

Long equilibration times

47,48,64 66

Excellent extraction efficiency Handbook of Solid Phase Microextraction

Source: Reproduced with permission from Elsevier. r 2010.

In Vivo Sampling with Solid-Phase Microextraction

413

plants), so inter-fibre reproducibility becomes a very important parameter in order to ensure accurate quantitative analysis. We evaluated the inter-fibre reproducibility of Supelco prototype biocompatible C18 fibres, and inter-fibre relative standard deviation was consistently 2 7% standard deviation for all analytes tested for the fibres of the same batch (Figure 12.8).64 However, if the results for the extractions performed simultaneously for five different lots (Lots 75, 90, 93, 94 and 96) are evaluated using single-factor ANOVA, the lot of coating is found to be significant factor at 95% confidence level for all analytes except nordiazepam.64 Based on these results, it is recommended to use fibres from a single batch of coating in a given experiment (for both samples and calibration standards) rather than mixing fibres from different lots. Also, considering that the variability of instrumental response for bioanalytical LC MS/MS methods is typically around 5 10%, further correction of inter-fibre reproducibility of proposed coatings for in vivo sampling applications is not necessary and permits single-use fibres for the first time.65 In contrast, PPY coatings exhibited B47 52% standard deviation for linezolid with single-use fibres and required use of three fibres per sampling in order to obtain adequate analytical method performance.58 Therefore, if using other custommade or previously not evaluated types of coatings for in vivo applications, it is extremely important to evaluate inter-fibre reproducibility during method development experiments in order to ensure successful implementation of in vivo SPME. One of the general trends that can be observed from Table 12.2 is that none of the preparation procedures reported to date is able to produce very thin coatings with good inter-fibre reproducibility when using very thin fibre supports such as

Figure 12.8 Comparison of the extraction efficiency obtained for eight independent Supelco C18 lots of coating (n 5 10 fibres per lot) prepared on different days in April May 2009.64 The results for four model analytes: diazepam, lorazepam, nordiazepam and oxazepam are shown. Extraction conditions: equilibrium, extraction from 100 ng/mL drug standard in PBS buffer pH 7.4. (Source: Reprinted with permission from Ref. 64. r 2010.)

414

Handbook of Solid Phase Microextraction

needed for in vivo applications. This is currently an active research area, but until this difficulty is resolved, the analyst is faced with two main options when designing in vivo experiments requiring multiple fibres: (i) the use of thicker fibres with good standard deviation but long equilibration times which necessitates the use of kinetic calibration; or (ii) the use of thin fibres with short equilibration times but poor standard deviation, which permits the use of equilibrium SPME calibration methods but require the incorporation of appropriate experimental design strategies in order to correct for poor inter-fibre reproducibility due to variable extraction phase volume. The type/scope of study plays an important role in the selection of which extraction phase is the most suitable for a given application. For targeted applications, which focus on the extraction of one of or a few a priori selected analytes, the extraction phase with the highest Kfs for the analytes of interest is usually selected in order to ensure maximum achievable sensitivity of the analytical method. However, for untargeted applications such as metabolomics studies, simultaneous extraction of both polar and non-polar compounds is needed. To address this need for general-purpose coatings when using direct extraction mode, we recently evaluated 42 different SPME coatings for the extraction of 36 metabolites from different chemical classes covering wide polarity range (log P range of 7.9 to 7.4).66 Three types of SPME coatings: (i) mixed-mode coatings with C8/C18 1 benzenesulfonic acid; (ii) polar-enhanced polystyrene DVB and (iii) phenylboronic acid (PBA) performed the best for simultaneous extraction of both hydrophilic and

Figure 12.9 Comparison of the extraction efficiency of the three best sorbents for the extraction of common metabolites.66 All extractions were performed in triplicate from metabolite standards prepared in phosphate-buffered saline, pH 7.4, 1.5 mL of sample volume, 16 h of extraction, no agitation. Analytes are arranged in the order of decreasing polarity (increasing log P from left to right according to the PhysProp database) Abbreviations: ATP, adenosine triphosphate; ADP, adenosine diphosphate; AMP, adenosine monophosphate; HBA, hydroxybutyric acid; PBA, phenylboronic acid. (Source: Reprinted with permission from American Chemical Society, r 2011.)

In Vivo Sampling with Solid-Phase Microextraction

415

hydrophobic metabolites at physiological conditions, thus making them suitable for untargeted metabolomic profiling applications. Example results are shown in Figure 12.9. As a general trend, extraction efficiency increases with decreasing compound polarity, but appreciable quantities (in the range of 0.5 5% absolute recovery) could be extracted of even very polar compounds such as sucrose and glutamic acid with log P values of 3.7 and 3.69, respectively. Absolute recoveries of 0.5 5%, although low in comparison to traditional exhaustive techniques, are still sufficient for reliable metabolite detection and quantitation with current state-of-the-art LC MS instrumentation. More polar compounds (log P # 24) than this were poorly extracted (e.g. choline with extraction efficiency of 0.1%) or not extracted at all by any of the coatings within the limit of detection of the instrument (for example, uridine diphosphate glucose), indicating that these compounds may not be observed in biological fluids using the proposed methodology unless they are present at sufficiently high concentrations in the biological sample of interest. The main exception to this general trend were glutathione species, which were efficiently extracted despite predicted log P values of 5.4 and 7.9, respectively.

12.2.3 Calibration of In Vivo Sampling It is important to emphasise that SPME can be successfully used for quantitative analysis, contrary to the perception of some researchers. However, for quantitative analysis, proper selection and setup of calibration procedure are absolutely necessary to ensure correct results. To date, several calibration approaches have been developed for SPME, as discussed in detail in Chapter 6. Equilibrium extraction is the most frequently used method, when a known distribution constant or an external calibration curve are used to correlate the amount of analyte extracted by the SPME fibre to its concentration in the sample. To shorten long equilibrium extraction times and/or address the displacement effects that occur when adsorptive coatings are used, extraction can be interrupted before equilibrium. Even though extraction equilibrium is not reached, there is still a linear relationship between the amount of analyte extracted onto the fibre and the analyte concentration in the sample matrix, provided the agitation, the extraction time and the extraction temperature remain constant.67 Figure 12.10 (top panel) summarises the advantages and disadvantages of equilibrium versus pre-equilibrium SPME calibration methods.51 While performing derivatisation at the same time with extraction, if the reaction is the rate-limiting step, the first-order reaction rate constant can also be used for calibration. The diffusion-based calibration method is very important for field sampling because it eliminates the use of conventional calibration curves. Fast on-site analysis and long-term monitoring are thus possible.68 Besides these ‘classic’ calibration methods, the newly developed method of ‘kinetic calibration’ seems to be particularly useful for in vivo determinations.15,56 When an SPME coating that is preloaded with a standard compound is exposed to an agitated sample matrix, desorption of the compound from the fibre occurs. The desorbed compound diffuses through the boundary layer into the bulk of sample matrix. The symmetry of absorption and desorption in SPME allows for the

416

Handbook of Solid Phase Microextraction

(A)

Use when equilibrium time is short enough for experiment goals Better sensitivity over kinetic methods Good method precision Equilibrium

Simple procedures Exact timing is not crucial May not be suitable for adsorptive coatings due to possible displacement effects

Selection of calibration method

Improved temporal resolution Useful when equilibrium time is too long for timescale of the experiment Kinetic

Useful for unknown matrix composition and agitation conditions Accurate timing is needed Decreased analytical sensitivity May exhibit poorer precision depending on the exact method chosen More complicated procedures and optimization

(B) Isotopically labelled standard available

On-fibre standardisation

Analyte as standard Selection of kinetic calibration method

Dominant desorption calibration

Standard desorbed into sample Not suitable for heterogeneous systems (need two sites) Analyte desorbed intosample

Isotopically labelled standard not available

Diffusion-based calibration

Need to control agitation

Double extraction calibration

Suitable for rapid sampling only

Standard-free methods

On fibre standardization using an analogue with similar time constant a Other calibrant (non-isotopically labelled) available One-calibrant kinetic calibration

Need to confirm similarity of time constant a Standard desorbed into sample Need diffusion constants Standard desorbed into sample

Figure 12.10. Flowchart for the selection of the calibration method for in vivo SPME: (A) equilibrium versus kinetic calibration (B) selection of the most appropriate kinetic calibration method.51 (Source: Reprinted with permission from American Chemical Society, r 2011.)

In Vivo Sampling with Solid-Phase Microextraction

417

Table 12.3 Relative Recovery and Precision Comparison in Whole Blood Between Three Kinetic Calibration Methods, Equilibrium Extraction and Conventional Protein Precipitation Method.69 (n 5 3 at each concentration level.) Fenoterol Concentration (ng/mL)

10 5,000 20,000

Relative Recovery in Whole Blood (%) (standard deviation, %; n 5 3) On-fibre Dominant Standardisation Pre-equilibrium Desorption

DiffusionBased Interface Model

Equilibrium Extraction

Conventional Method: Protein Precipitation

94 (17) 103 (15) 100 (16)

93 (14) 102 (9) 100 (9)

105 (9) 98 (8) 101 (9)

108 (8) 103 (7) 104 (7)

110 (20) 106 (21) 103 (21)

Source: Reprinted with permission from Elsevier. r 2010.

calibration of extraction process of the analyte using desorption of the calibrant from the coating into the sample matrix. This is especially important for the calibration of on-site, in situ or in vivo analysis because the control of the agitation condition of the matrix is sometimes difficult, and direct spiking of standards into the matrix is typically not possible in these cases.68 Obviously, when using this method, a small amount of calibrant is introduced into the living system, but these quantities are usually selected to be negligible (typically #1 ng desorbed per sampling) and are not expected to cause adverse reactions in the living system. Figure 12.10 shows a flowchart to aid in the selection of the most appropriate calibration method for a given in vivo SPME application.51 A recent study also compared the analytical performance of different SPME kinetic calibration methods versus equilibrium SPME using a flow device simulating an animal circulatory system.69 As shown in Table 12.3, all three kinetic methods had acceptable accuracy (93 119%) and yielded correct results when compared to equilibrium SPME and the conventional method based on solvent precipitation. Among kinetic calibration methods, diffusion-based calibration performed the best in terms of method precision and the ease of use, and it does not require isotopically labelled standards for its implementation. However, this method requires known or controlled agitation conditions, a requirement which may not be met by all in vivo experimental designs. In such cases, on-fibre standardisation yielded better method precision than dominant desorption calibration, making it a more suitable choice, provided an appropriate calibrant is available.

12.2.4 Automation Automation of an analytical method provides many advantages, including reduced analyst time both for routine analysis and method development, faster sample throughput and greater reproducibility. Without automation, acceptance and the range of potential applications for a new technique can be reduced significantly.

418

Handbook of Solid Phase Microextraction

All the steps outlined for the development of manual SPME methodology are virtually identical to those required for automated SPME. However, the greater throughput possible when using an automated method allows a considerably faster development process. It can allow greater flexibility in the SPME methodology for a particular application. For example, more accurate extraction timing is achievable when using an automated method, which enables the use of faster, non-equilibrium SPME to increase sample throughput without deterioration of method precision. In contrast, it can be difficult for the analyst to achieve satisfactory reproducibility using a non-automated method, especially when short pre-equilibrium extraction times are employed. The ability of an autosampler to run 24 h a day may make a method that is considered too slow with manual SPME practical.70 A commercially available GC autosampler can be easily programmed for analysis of volatile compounds from live samples that fit in vials, such as insects, plants, tissues or cell cultures. High throughput can be achieved by use of fast GC methods and/or short sampling times. For applications that require LC analysis, Dr. Pawliszyn’s laboratory at the University of Waterloo is currently modifying PAS Concept 96 autosampler described in Section 5.2.2 in order to accommodate an array of 96 SPME commercial devices for in vivo sampling. The main reason why multi-well plate format is preferred for the automation of LC-based methods is the fact that the kinetics of extraction and desorption of analytes are much slower in the liquid phase than in the gas phase. This results in low sample throughput if samples are processed serially. Many biological applications generate numerous samples for analysis, and the total analysis time may prove to be impractical when these samples are analysed manually. It is more efficient to perform parallel extraction and desorption steps of multiple samples on a multi-well plate format. When the dimensions of the in vivo samples are not compatible with the multi-well plate format and the extraction step is performed manually, desorption and analysis can still be carried out in multiwell configurations. Figure 12.11 depicts such a system, where multiple fibres can be stored, sterilised and transported by introducing multiple fibres into the assembly formed by plates 1 and 2. Plate 1 works as a support for the multiple fibres, while plate 2 is used for sealing the tip of the samplers before and after extraction to protect them from contamination. When all the samples are ready for analysis in the laboratory, plate 1 can be superimposed over plate 3, for parallel desorption. Finally, plate 3 can be analysed with any analytical system that handles multi-well plates. Such automated multi-well systems, currently under development, have great potential for providing fast methods for high-throughput applications.

12.3

In Vivo Applications

Although SPME was initially applied only for the analysis of organic compounds from rather clean samples (air and water), it is now increasingly used in bioanalysis (in vitro and in vivo) for the determination of volatile emissions, proteins, polar

In Vivo Sampling with Solid-Phase Microextraction

419

Upper part – support

1

2

Lower part – fit around the tip

….. 12

….. 12

Top

Top 82 mm

82 mm

…..

…..

8

8

126 mm Lateral

126 mm

126 mm

Lateral 3 mm

2 mm D=0.25 in

126 mm

2 mm 3 mm

D=0.027 in (slide fit)

6 mm D=0.25 in

D=0.025 in (tight fit)

The plate – base

3

….. 12 TOP 82 mm ….. 8

126 mm Lateral 19 mm

20 mm

126 mm D=0.25 in

Figure 12.11 Multi-well system for storage, transportation, desorption and analysis of multiple SPME fibres. Plate 1 is used to support the fibres, plate 2 is used to seal the tip of the sampler and plate 3 is for desorption and injection into an LC.

alkaloids, pharmaceuticals and surfactants because of its successful coupling with liquid chromatography and capillary electrophoresis.71 73 SPME is an excellent alternative to classical methods for separating analytes of interest from biological samples. It is simpler, faster and provides markedly cleaner extracts than methods based on liquid liquid extraction (LLE) or solid-phase extraction (SPE). Figure 12.12 summarises typical steps of in vivo SPME workflows in combination with LC MS and GC MS and shows how rapid the entire protocol is.74 In the last 10 years, the total number of reports on application of SPME in bioanalysis increased from 10 to more than 300. Many of these methods can be translated, in the future, in techniques suitable for in vivo extraction. Most of the literature on SPME for in vivo sampling is still related to microbiological applications, monitoring of BVOCs emission from plants and isolation of insect

420

Handbook of Solid Phase Microextraction

Conditioning

In vivo extraction

1–20 min

0.1–0.5 min

0.1–0.5 min

Wash

5–60 min

0.5–10 min

lve

nt

Desorption

So

LC–MS

1–60 min

Th

er

m

al

GC–MS

Figure 12.12 Overall workflow of in vivo SPME in combination with LC MS and GC MS.74 (Source: Reprinted from Wiley-VCH with permission. r 2011.)

semiochemicals where SPME has become a well-established technique. However, since the last reviews,1,13,73 significant improvements have been made in the new fields of direct extraction from flowing blood and tissue, both for targeted analysis and metabolomics applications.49,51,74 There is also increased interest in using SPME to sample human volatile emissions (breath analysis and skin emission analysis) for metabolomics applications in order to identify biomarker profiles indicative of various diseases. The use of breath analysis (by SPME and other methods) for clinical monitoring has also garnered a lot of interest as a non-invasive method of monitoring drug concentrations.

12.3.1 In Vivo Sampling of Plants Using SPME To date, in vivo SPME has been used extensively to sample volatile emissions emitted by plants. The main objectives behind such studies range from (i) better understanding of plant physiology and chemotaxonomic studies, (ii) interaction of plants with the environment, including studies of mechanisms of defence and pollination and (iii) search for new scents in the fragrance industry. A recent review article by Stashenko and Martinez focusing on the methods for sampling flower scents75 states that SPME is currently the predominant sample preparation method in the field of flower scent because it provides a more representative and complete volatile profile than traditional techniques such as solvent extraction and distillation. There are two main reasons behind this: (i) extractive methods, such as solvent extraction, are applied to excised flowers and typically result in extracts containing large amounts of lipophilic substances that require removal prior to chromatographic analysis, which may result in loss of flower

In Vivo Sampling with Solid-Phase Microextraction

421

Figure 12.13 Rhythmic emission of four major volatile benzenoids.33 After flower opening, flower volatiles were analysed four times a day over a 48-h period, using in vivo HS-SPME. The top white bars indicate light periods (L); the top black bars indicate dark periods (D). Each component is plotted as a percentage of its maximum value. (Source: Reprinted with permission from Elsevier. r 2003.)

volatile components; (ii) sampling is performed in vivo, so the same flower can be sampled on multiple occasions in order to study scent development over time. For example, in vivo SPME was used to study scent production in petunias.33 Figure 12.13 shows the circadian rhythm of scent production over 48 h for four of the major benzenoid compounds identified in this study. Comparison of in vivo and in vitro data indicated that volatiles are produced de novo, even during low emission periods rather than stored. Targeted metabolomics data obtained in this study were also correlated to DNA microarray results, which indicated up-regulation of genes preceding the increase of volatile scent emissions. In addition to flower scent analysis applications summarised by Stashenko and Martinez,75 sampling of plant volatile emissions by in vivo SPME can be used for other applications. For example, Beck et al.34 recently used in vivo SPME to sample VOCs emitted by yellow starthistle (a noxious weed that is toxic to horses) in order to look for ways to biologically control this plant by introduction of hostspecific herbivores that would attack only this weed. Such biological control methods are desirable both from the viewpoint of their effectiveness and their minimal environmental impact. Bicchi et al. proposed the use of PDMS membrane for direct sampling of plant leaves, which could be especially useful to study plant stress response by comparing metabolite composition at different distances from the site of stress.38 However, in their preliminary study, they demonstrated only the

422

Handbook of Solid Phase Microextraction

potential of the technique for freshly cut spearmint and rosemary leaves without subsequently proceeding to apply it in vivo. With the development of new coatings suitable for extraction of polar and nonvolatile analytes, the use of in vivo SPME can be extended for sampling of various compounds such as pesticides, endogenous secondary metabolites and allelochemcial directly in plants.76,77 For example, in vivo SPME was recently used to study allelochemical uptake of 1,8-cineole by tomato plants.77 Prior to the availability of in vivo SPME, there was no other suitable method to measure such uptake in vivo, and allelopathic effects had to be inferred from toxicity studies. It was found that the uptake of 1,8-cineole was very rapid and that significant concentrations of this compound persisted in the plant even 72 h after a single cineole application to soil. The study also opens the possibility that brief but repeated exposures to allelochemicals may have an allelopathic effect on plants. Currently, the two main limitations of SPME for this application are (i) the size of commercially available coatings does not permit sampling of most plant leaves due to their current dimensions and (ii) the technique is applicable only to allelochemicals that are translocated in target plants. For this preliminary study, authors used external calibration approach using the phosphate buffered saline (PBS) buffer as a substitute for stem fluid in order to obtain semi-quantitative results. However, further improvement of accuracy can be expected if a more suitable in vivo calibration technique was applied (kinetic calibration discussed in Chapter 6). In vivo SPME was used to sample pesticides from the leaves of the jade plant (Crassula ovata) using the direct insertion method.78 Accurate quantitative results were obtained by using a newly developed kinetic calibration method. A single, one-time pesticide application to soil resulted in detectable concentrations of

Concentration in a leaf (mμ/mL)

0.35 0.30 0.25 0.20 0.15 0.10 0.05 0.00 0

5

10

15 20 25 30 35 40 Days after pesticides were applied to soil

45

50

55

Figure 12.14 Concentration profiles of carbofuran (x), propoxur (e), and carbaryl (¢) in a jade plant leaf. In vivo SPME was performed at 25 C for 20 min using 165-μm PDMS fibres by direct insertion of the fibre into the leaf.78 (Source: Reproduced with permission of American Chemical Society. r 2008.)

In Vivo Sampling with Solid-Phase Microextraction

423

80 Microdialysis

70 Concentration (ng/ml)

Solid-phase microextraction

60 50 40 30 20 10 0

1

3 2 Sample position

4

Figure 12.15 Concentrations of carbofuran in the leaves of a jade plant after 20 days of the pesticide application to soil.79 The sampling time was 20 min. Microdialysis and SPME were used to sample the left and right leaves, respectively. PBS (pH 7.4) and 165 μm PDMS were the perfusate for microdialysis and the SPME coating, respectively. (Source: Reproduced with permission from Elsevier. r 2008.)

carbofuran, propuxur and carbaryl even after 50 days, which confirms literature reports of high pesticide persistence even up to 12 weeks (Figure 12.14). It was also found the concentration of pesticides in leaves depended on the level of the leaf (new or old growth), with the highest concentrations observed in lower level leaves that grew first. A subsequent study by the same authors compared the performance of in vivo SPME to the established in vivo microdialysis technique.79 Both methods were found to have similar limits of detection (LODs) and limits of quantitation (LOQs). For example, the LOD and LOQ were 3.2 and 11 μg/L for carbofuran using in vivo SPME, and 4.5 and 15 μg/L for in vivo microdialysis. The concentrations detected by SPME were similar to the concentrations detected by microdialysis (Figure 12.15). However, SPME results were consistently observed to be slightly higher than microdialysis results. This was due to a small amount of pesticides adsorbing to the microdialysis membrane as well as the small amount of dialysate staying in the outlet tube after microdialysis sampling. This study demonstrates clearly that SPME can also perform in vivo microdialysis. However, SPME offers additional advantages: its implementation does not require a pump (and power supply) that make it more amenable for field sampling, and it offers more accurate results due to the fact that adsorptive losses to the membrane are eliminated.

12.3.2 Sampling of Microorganisms and Biotransformation Studies In vivo SPME can be a useful tool for monitoring the volatile emissions of various microorganisms such as fungi, yeast and bacteria. For example, SPME can be used to monitor production of volatiles during fermentation processes.80 In this study, the production of volatile alcohols, acetate esters and ethyl esters of fatty acids by free and immobilised biocatalyst (Saccharomyces cerevisiae) was monitored. Factors affecting the desired aroma production, such as fermentation time and temperature, were optimised.

424

Handbook of Solid Phase Microextraction

SPME is also applicable for the study of biodegradation or biotransformation pathways of various chemicals.81 Example applications of this type include the elucidation of biodesulfurisation pathway of dibenzothiophene82 and the investigation of biotransformation pathways of hydrocarbons83 and explosives.84 Although these studies were performed using in vitro SPME (bacteria were removed using centrifugation or filtering prior to SPME), the use of in vivo SPME should be equally applicable, especially if the analytes of interest are sufficiently volatile for headspace extraction. In fact, more recently, Demyttenaere et al.36 used in vivo SPME to screen 60 fungal strains for their ability to biotransform citronellol into rose oxide, which is one of the most important materials in the fragrance industry. Stereospecificity of the conversion was also examined. In a similar study, in vivo SPME was used to evaluate the potential of 19 fungi to transform (6)-linalool into lilac alcohols and ketones, which are in high demand by the perfume industry.85 Three strains were found to produce the species of interest and the data obtained in the study was used to help identify new highly stereoselective biosynthetic pathway. Overall, the authors found that SPME was very efficient and convenient method for this type of application because of its speed, simplicity and nondestructive nature. In conclusion, in vivo SPME can play an important role in biotransformation studies in order to help isolate various volatiles and/or short-lived intermediates and clarify mechanistic pathways of production of various compounds of interest.

12.3.3 Sampling of Insect Volatile Emissions The cuticular surface of insects presents a rich reservoir of chemicals, some of which have important informational value. They act as intra- and inter-specific signals for insects. They are involved in nestmate recognition by termites and ants. Cuticular hydrocarbons (CHs) serve as species and caste recognition cues in termites. Despite the importance of CHs in insect chemical communication, direct proof that they are detected and recognised by insects by contact or by olfactory receptors is rare. In Periplaneta americana, CHs induce aggregation. Using SPME and GC, Said et al.19 investigated how CHs are detected by P. americana antennae. They identified the three main CHs of the species profile in the volatiles emitted by these insects. GC coupled with electroantennography recordings demonstrated for the first time that the antennae responded to these three CHs (n-pentacosane, 3-methylpentacosane and 6,9-heptacosadiene). Furthermore, the researchers showed that CHs had an attraction effect in Y-olfactometer bioassays when present at high concentrations. As CHs can be perceived by P. americana, at least from a short distance, they could play a role in attracting members of the same species during aggregation processes, in addition to inducing aggregation when direct contact is possible. The ability of recognising nestmates and distinguishing them from individuals of other colonies is a key property of social insects. After short antennation, social wasps, bees and ants typically show an immediate discriminative response towards nestmates or ants that are not members of the colony, suggesting that non-volatile

In Vivo Sampling with Solid-Phase Microextraction

425

surface chemicals on the cuticle play an important role in nestmate recognition. The profiles of cuticular substances (in particular, hydrocarbons) are often speciesand colony-specific. The quality of interactions among individuals has been found to be correlated with similarities in their hydrocarbon profiles. Tentschert et al. investigated CHs and fatty acids of workers and queens from two ant species, Leptothorax acervorum and Leptothorax gredleri.20 The authors aimed to quantify the variation of CHs between and within colonies and to investigate whether cuticular cues are different enough to allow the separation of individuals from different colonies by multivariate statistics. Cuticular compounds were extracted from single ants by solvent extraction, solid sampling and SPME with two different PDMS fibres and analysed using GC and GC MS. All methods gave similar results, proving that SPME can be applied to very small live ants (body size approximately 3 mm). In this example, direct extraction was achieved by rubbing the abdomen of the ant with a fibre coated with a 30-μm PDMS layer. SPME obviously had an important advantage over conventional extraction: ants did not have to be killed for the extraction, and they could be investigated repeatedly. The hydrocarbon mixtures consisted mostly of branched and unbranched alkanes and alkenes within the range of C25 C33. Dufour glands of both species contained a blend of hydrocarbons different from those found on the cuticle. In addition, terpenoids, especially tetramorenes, were present in the Dufour gland contents. In a principal component analysis (PCA) based on the relative proportions of cuticular compounds, most nestmate workers clustered in four groups corresponding to the original four investigated colonies. Queens and workers were found to differ significantly in their chemical profiles, suggesting that the two castes bear specific labels. Using a similar technique, the same authors showed that queens of the ant Pachycondyla inversa may cooperate during colony founding.23 One of several cofounding queens specialises in foraging, whereas the others remain in the nest and guard the offspring. Division of labour is achieved by aggressive interactions, which result in the formation of dominance hierarchies. GC and mass spectrometry (MS) of CHs obtained from live queens by SPME revealed consistent differences between the patterns of CHs of queens with high versus low rank: only high-ranking queens showed considerable amounts of cuticular pentadecane (n-C15) and heptadecene (n-C17:1). The relatively high volatility of these hydrocarbons might explain why subordinates react to approaching dominants before having physical contact. A similar study was completed by Gilley et al.24 on live honeybees. They used SPME (65 μm PDMS DVB fibres) to sample the volatile compounds emitted by queens in several reproductive states (unmated queens, recently mated queens and established mated queens) and compared them to the volatiles emitted by workers. The authors detected nine compounds that were present in at least 75% of the individuals in at least one type of bee, and which were not present in the sampling environment alone (Figure 12.16). Four of these compounds were present in queens but not in workers. One of these four compounds, identified as E-β-ocimene, was expressed fully only in established mated queens and may be a signal of diploid egg-laying activity. The three remaining queen-specific compounds (including one

426

Handbook of Solid Phase Microextraction

Figure 12.16 Chromatogram of a mated, established queen showing the nine peaks commonly present in at least one type of queen (arrows). (Source: Reproduced from Ref. 24 with permission of Elsevier. r 2006.)

identified as 2-phenylethanol) were associated with unmated queens and may mediate interactions between unmated queens and workers during queen elimination. The five common compounds that were detected in both queens and workers were hydrocarbons and may function as nestmate recognition cues. These discoveries were considered as a first step in determining the potentially important functions of volatile signals and cues within honeybee nests. SPME proved to be a very useful tool for investigating honeybee queen volatiles. A new in vivo SPME method using pencil-lead fibre coupled with GC was developed by Djozan et al. for a fast, easy and reliable monitoring and recognition of a volatile defensive chemical from Graphosoma lineatum.25 Three methods were investigated: (i) in vitro surgical sampling and direct injection in GC, known as a classical method; (ii) in vitro surgical sampling followed by SPME GC analysis and (iii) in vivo SPME followed by GC analysis. Some experimental parameters, such as extraction time and chromatographic conditions, were examined and optimised. The obtained results revealed that the in vivo SPME GC is much better than in vitro pentane extraction of volatile compounds followed by either headspace SPME (HS-SPME) or direct injection and GC analysis (Figure 12.17). From their results, the authors concluded that the amount of chemicals extracted with a pencil-lead fibre is about 1,000 times higher than that obtained with commercially available PDMS fibres. The first identification of a sex pheromone of a pine sawfly (Hymenoptera, Diprionidae) dates back almost 30 years. Since then, female-produced pheromones of over 20 diprionid species have been investigated by solvent extraction followed by separation and identification. However, no study has shown what the females actually release. Collection of airborne compounds using absorption on a charcoal

In Vivo Sampling with Solid-Phase Microextraction

427

Figure 12.17 Representative chromatograms obtained from defensive volatiles compounds of the scent gland of G. lineatum insect, using (a) in vitro classical sampling and (b) homemade pencil-lead fibre. (Source: Reproduced from Ref. 25 with permission of Elsevier. r 2005.)

filter and SPME followed by GC MS analysis revealed an unusual system in Diprion pini, in which the pheromone precursor alcohol, 3,7-dimethyl-2-tridecanol, is released together with acetic, propionic, butyric and isobutyric acids.26 The corresponding acetate, propionate and butyrate esters of 3,7-dimethyl-2-tridecanol were also found in the samples. All esters were electrophysiologically active, and the propionate and isobutyrate were attractive in trapping experiments. Based on these results, the authors concluded that at least in part of its range, the pheromone response of D. pini is not very specific with regard to the functional group, as long as this is an ester. When compared to traditional sampling methods, SPME was much faster, simpler and produced similar results. More recently, Chen et al.28 reported on the identification of a sex pheromone component of a cossid moth, Cossus insularis. Coupled gas chromatographic electroantennographic detection (GC EAD) analysis of SPME collections of volatiles released by live female moths showed that two compounds elicited EAG responses from the antennae of male moths. These compounds were identified as (E)-3-tetradecenyl acetate (E3-14:Ac) and (Z)-3-tetradecenyl acetate (Z3-14:Ac) by mass spectral analysis and retention index comparisons with synthetic standards. The ratio of E3-14:Ac and Z3-14:Ac was 95:5 in the effluvia of a female. In field bioassays, sticky traps baited with blends of E3-14:Ac and Z3-14:Ac showed that E3-14:Ac is an essential component of the pheromone.

428

Handbook of Solid Phase Microextraction

Interestingly, in a recent study performed by Colazza et al.86 to sample CHs from stink bugs (Nezara viridula), the authors compared the performance of in vitro solvent extraction with hexane versus direct-contact in vivo SPME. Subsequently, the authors found that the results for hexane extracts they obtained were incorrect because of contamination, while obtained SPME profiles were verified to be correct.87 This example indicates one important advantage of in vivo SPME the reduction in the number of steps during the sampling and sample preparation stages minimises the potential for (i) inadvertent losses of analytes due to sample manipulation steps and (ii) contamination of samples due to external sources, such as contaminated solvents, glassware and so on.

12.3.4 Sampling of Volatiles Emitted by Humans In recent years, an increase in research interest in studying volatile emissions produced by humans has been observed. In vivo SPME can play an important role in such studies and has been used to study human skin emissions, flavour release and volatile breath emissions (Table 12.4).

12.3.4.1 Sampling of Human Skin Emissions It is well known that VOCs synthesised by metabolism and emanations of human skin make up human odours. Hundreds of different kinds of substances appear in human odours, which can be classified according to their functional groups, such as carboxylic acids, alcohols, aldehydes, aliphatics, esters, ketones, amines, heterocyclics and so on. The alternant actions between both skin glands and excreting organic compounds achieve individual human odours. Special regulations of individual emission behaviours of human odours are interlinked to human hormonal control and bacterial populations localised at skin surfaces surviving by metabolising and transforming organic compounds absorbed from the external environment. Any changes of metabolism equilibrium cause alteration of human emanations. Therefore, the variety of individuals results in different characteristics of human odours, also called ‘fingerprint’ characteristics by analogy with the real dactylograms. Fingerprint characteristics of human odours are actually informative biomarkers and have been successfully used to train sleuthhounds, identify criminals and diagnose diseases. Zhang et al.37 have developed an efficient and non-invasive method for analysing volatile organic emanations from the skin of human arms consisting of an original sampling device (Figure 12.3), SPME and GC MS. The emanations were sampled by SPME connected with the active sampling device for 30 min and transferred into GC MS immediately for analysis. Finally, 35 compounds were identified according to various degrees of certainty. PCA was used to identify similar fingerprint characteristics obtained during different seasons. The top 10 emanations contributing to characteristics in different seasons were attempted to be described using comparisons based on common model strategy using a homemade

Analytes

Live Sample

Extraction Mode and Time

Extraction Phase

Calibration and/or Identification

Reference

VOCs (chloroform, 1,1,1trichloroethane, toluene and methyl tert-butyl ether)

Human skin exposed to VOCcontaminated water for 60 min

Headspace, 6 min

PDMS 100 μm

External calibration Data used for estimation of human dermal permeability coefficients in water

88

VOCs (metabolomics)

Human skin (on arms or abdomen)

Direct contact, 5 120 min

PDMS membrane 450 μm

Preliminary study (optimisation 46 of SPME parameters and evaluation of in vitro and in vivo reproducibility for selected subset of compounds)

VOCs (metabolomics)

Human skin (upper back and arms)

Headspace, 30 min

DVB/CAR/PDMS 50/30 μm

Qualitative data only, based on MS libraries, retention index and standard compounds

39

VOCs (metabolomics)

Human skin (on arms)

Direct contact by rolling over 10 cm2 area, 10 12 s

PDMS stir bar 24 μL

Quantitative profiling Embedded internal standards

43 45

Volatile organic emanations

Human skin (on arms)

Headspace, 30 min

PDMS/DVB 65 μm

Qualitative data only, based on MS libraries

37

Volatile organic emanations after fragrance application

Human skin (hand)

Direct contact and headspace 30 min

PDMS membrane 500 μm (technique referred to as sorptive tape extraction by the authors)

Preliminary study semiquantitative results describing changes in emitted volatiles over time

38

VOCs

Headspace 40 min (cell PDMS (thickness not Cell cultures (lung cancer specified by authors) and normal) and correlation culture) 60 min to human breath (breath)

External calibration

40

429

(Continued)

In Vivo Sampling with Solid-Phase Microextraction

Table 12.4 Use of In vivo SPME for Studies of Human-Related Emissions

430

Table 12.4 (Continued) Analytes

Live Sample

Extraction Mode and Time

Extraction Phase

Calibration and/or Identification

Reference

VOCs

Human breath (mechanically ventilated patients)

Headspace, 5 min

PDMS/Carboxen

External calibration

89

Flavour compounds

Human nose (breath)

Headspace, 8 60 s

PDMS 100 μm

External calibration curve based on standard compounds

90,91

Propofol (anaesthetic drug) 2-Pentylfuran

Human breath

Headspace, 5 min

CAR/PDMS/DVB

92

Human breath

Headspace, 48 h

DVB/CAR/PDMS

External calibration based on standard compounds External calibration based on standard compounds

Alkanes and aromatic hydrocarbons

Human breath

Headspace, 20 min

PDMS 100 μm

External calibration

41

VOCs (metabolomics)

Helicobacter pylori bacterial culture (in vivo) Human stomach tissue (in vitro)

Headspace, 15 min

Carboxen/PDMS

Metabolomics (normal versus disease) and identification of potential biomarkers based on MS libraries, retention index, and standard compounds

93

VOCs (metabolomics)

Human colon cell line

Headspace, 40 min

Carbowax/DVB 65 μm

Metabolomics (normal versus disease) and identification of potential biomarkers based on MS libraries, retention index, and standard compounds

94

42

Handbook of Solid Phase Microextraction

In Vivo Sampling with Solid-Phase Microextraction

431

chromatographic data processing system. The results suggested that the analysis based on fingerprint characteristics of human skin emanations could provide useful and important clues to reveal biomarkers among the mixture of human skin emanations. Such findings are supported by other, more extensive 200-subject studies.43 45 Riazanskaia et al.46 reported the utility of PDMS membranes for in vivo sampling using direct skin contact. The main objectives of the study were the optimisation of in vivo SPME parameters and the evaluation of analytical reproducibility (in vitro and in vivo) of a subset of observed compounds. Very rich volatile profiles containing several hundred peaks were obtained when PDMS membranes were analysed after thermal desorption by GC MS. In vivo reproducibility was evaluated using six identified compounds. The precision of five simultaneous samplings (using five individual PDMS membranes) was very good and ranged from 15 32% standard deviation of the amount extracted, indicating analytical utility of this method. The proposed methodology allows direct sampling of carcinoma lesions and can subsequently be employed for metabolomics studies to isolate the biomarkers of various skin cancer types and other skin disorders. Proper conditioning of membranes prior to use was necessary to eliminate sampling artefacts and the importance of storage time/conditions was established. Significant changes in the amounts of analytes detected were observed for samples stored for longer than 1 day and at temperatures above 4 C and attributed to microbial contamination. This is in contrast to study by Soini et al.43, who reported stability of Twister PDMS stir bars used to sample human skin was as long as 14 days when refrigerated. It is not clear what caused the difference observed between two studies. Riazanskaia et al. monitored stability of four compounds only, whereas Soini et al. do not report which compounds they used for their stability evaluation, making any kind of comparison difficult. Furthermore, Soini et al. do use two embedded internal standards, which may have compensated for some of the losses from the device. Clearly, additional investigations of long-term stability and best storage conditions of direct sampling skin devices are necessary. Gallagher et al.39 compared the performance of solvent extraction with in vivo HS-SPME for sampling of skin emissions in order to attempt to establish a baseline of ‘normal’ skin emissions and variations attributed to individuals, which could in future be used to search for biomarkers of human diseases. The authors were able to (tentatively) identify a total of 92 compounds: 58 were found in in vivo SPME samples and 49 in hexane extracts, indicating that the two techniques are complementary in nature. Higher-molecular-weight compounds predominated in hexane extracts, while SPME was more suitable for collection of lower-molecular-weight aldehydes and ketones. SPME was useful to extract five compounds that varied significantly according to age (Figure 12.18); among these, the authors propose dimethylsulphone, benzothiazole and nonanal as biomarkers of aging. Surprisingly, no similar correlations or possible biomarkers were obtained for gender differentiation. In the same study, the variation of skin emission with respect to location of sampling (arm versus back) was also investigated and found that the two body locations share a lot of common compounds

432

Handbook of Solid Phase Microextraction

Mean % of total VOCs older subjects

18.000 16.000

Back

14.000

Foream

12.000 10.000 8.000 6.000 4.000 2.000 0.000

Mean % of total VOCs younger subjects

18.000 16.000

Back

14.000

Foream

12.000 10.000 8.000

*

6.000 4.000 2.000

*

*

0.000 DMS Benz

C8

C9 C10 Compounds

Sal

Cinn

Iso

Figure 12.18 Mean percentages of selected compounds present in the back and forearm samples of younger (bottom graph) and older (top graph) human subjects collected by SPME. The compounds in the top graph are the same as those listed along the x-axis of the bottom graph: DMS, dimethylsulphone; Benz,benzothiazole; C8, octanal; C9, nonanal; C10, decanal; Sal, hexyl salicylate; Cinn, α-hexyl cinnamaldehyde; Iso, isopropyl palmitate. , compound in which locus was significant; V, compound for which age was significant. (Source: Figure reprinted with permission from Ref. 39.)

but also considerable differences (Figure 12.18) indicating that a consistent, welldefined location of sampling is a prerequisite for comparative studies.

12.3.4.2 Human Breath Analysis In recent years, with the development of more-sensitive analytical instrumentation, there is an increased interest in human breath analysis as a non-invasive means of obtaining a representative biological sample that can be used for applications such as monitoring of drug concentration, screening of toxicological exposure and for disease detection. In future, non-invasive sampling techniques such as breath analysis can provide benefits such as better patient acceptance and a reduction of the

In Vivo Sampling with Solid-Phase Microextraction

433

blood samples necessary (e.g. repeated blood sampling in critically ill patients can require transfusions to replace the blood volume lost). Miekisch et al.92 recently studied the utility of breath analysis by HS-SPME for monitoring concentrations of propofol (an intravenous anaesthetic). Clinical monitoring of this anaesthetic is very important to ensure the dosage administered is not too high (severe health consequences) or too low (intraoperative awakening of patient during surgical procedure). HS-SPME was found to meet all requirements for reliable quantitative analysis: excellent accuracy (97 103%), acceptable intraday precision (8 11% standard deviation) and linear range suitable for the analysis of expected therapeutic concentrations of this drug. Propofol concentrations determined by breath analysis (end tidal breath) were found to correlate well with arterial blood concentrations (Figure 12.19) but not central or peripheral venous blood concentrations. These experimental data suggest that HS-SPME could be useful as a clinical test for monitoring propofol. However, the lower correlation found in lung surgery patients (Figure 12.19) indicates that an impaired ratio of ventilation/ perfusion may significantly affect the correlation between blood and breath concentrations, thus complicating the reliable quantitative determination in all patients. A more fundamental study to study the relationship between analyte concentrations in blood and breath in mechanically ventilated patients was carried out by Schubert et al.89 The study was performed on volatile anaesthetic isoflurane as well as on common VOCs that are considered to be biomarkers of certain conditions according to literature: n-pentane (lipid peroxidation), isoprene (cholesterol biosynthesis) and acetone (dextrose metabolism, lipolysis). It was determined that analyte

Arterial propofol concentrations [µmol/L]

35 y = 53.148x + 10.506 r 2 = 0.7517

30 25 20 15 10

y = 45.363x + 1.9392 r 2 = 0.8492

5 0 0.0

0.1

0.2 0.3 0.4 Exhaled propofol concentrations [µmol/L]

0.5

0.6

Figure 12.19 Propofol concentrations in expired air and arterial blood from 16 patients given total intravenous anaesthesia (triangles) and from six patients undergoing lung surgery (circles). (Source: Figure reprinted with permission from Elsevier. r 2008.92)

434

Handbook of Solid Phase Microextraction

concentrations in exhaled air depended on venous blood concentrations and inspired concentrations. Breath monitoring was suitable only if inspired concentrations were ,5%, otherwise concentration profiles in breath were found to deviate significantly (and unpredictably) from blood concentration profiles. For example, for acetone and pentane, significant difference in venous blood concentrations could be found depending on whether or not patients had sepsis, but no significant difference was observed in corresponding breath samples because inhaled concentrations of these compounds were relatively high. This result was attributed to occurrence of dead space ventilation and shunt perfusion, which is a common problem with mechanical ventilation. More fundamental studies such as this one that clarify the relationship between blood and breath concentrations are necessary before breath analysis by SPME can be implemented into clinical practice. The potential of SPME for breath analysis with the objective of disease detection was examined in several studies.40 42 Yu et al.41 developed an HSSPME GC method for analysis of alkanes and aromatic hydrocarbons in human breath, with a potential application for the screening and early diagnosis of lung cancer. The proposed method had excellent precision (,10% standard deviation for all compounds evaluated) if extraction time, temperature and relative humidity of samples were well controlled. Method sensitivity (LOD in range of 1022 to 1 ng/mL) was sufficient to monitor physiological concentrations of these analytes in breath. Based on the analysis of 15 healthy and 15 subjects with lung cancer, a combination of compounds (benzene, styrene, propyl benzene and n-undecane) was proposed as a potential biomarker set for screening lung cancer, pending confirmation in larger studies. Chen et al.40 attempted to correlate the volatile biomarker compounds detected from cancer cells with their presence in breath samples of lung cancer patients in order to establish the basis for non-invasive breath diagnosis of lung cancer using an electronic nose. Four potential biomarker compounds were found to be common to all types and stages of cancer examined (29 lung cancer patients in total) using cultivated cell cultures. Eleven potential biomarkers were found by PCA analysis of breath samples obtained from healthy people versus lung cancer patients. However, relatively poor correlation between the compounds detected for cell cultures and potential biomarker compounds in breath was found and required further study. In contrast, using a similar experimental approach, Syhre et al.42 were able to isolate 2-pentylfuran as a potential biomarker of infection by Aspergillus fumigatus using cell cultures, and then subsequently confirm its presence in breath samples of individuals with cystic fibrosis who were infected with this pathogen (no 2-pentylfuran was detected in healthy control samples). This study indicates that breath analysis may be suitable for detection of pathogen infections in lungs at least, based on the results of the limited patient population studied.

12.3.4.3 Flavour Release Flavour release is an important issue in food science and has been extensively studied in recent years. Generally, the methodologies used for flavour studies were static or dynamic headspace. In vivo, the aroma stimulus depends on the concentration of

In Vivo Sampling with Solid-Phase Microextraction

435

aroma compounds in the nasopharynx, which, in turn, is affected by release rates of the compounds from the food in the mouth. In-mouth flavour release is known to be affected by food matrix and composition and by the mastication process. Pionnier et al.90 followed the kinetics of aroma compound release during model cheese consumption in order to clarify the relationships between flavour release and some oral parameters. Breathing, salivation, chewing and swallowing were monitored during the eating process of eight human subjects. The SPME GC MS method was useful for quantification and identification of diacetyl, heptan-2-one, ethyl hexanoate, heptan-2-ol, propanoic acid and butyric acid.91 Among five different fibre coatings evaluated for selectivity, sensitivity, stability and competition phenomena, PDMS coating was the most appropriate. A method based on short-time sampling (8 s) has been shown to be efficient in following the kinetics of release of these aromas over 3 min. Eight panellists were asked to eat 5 g of a model cheese in their own way, and nosespace experiments were carried out during that period. The use of SPME GC MS has been validated in comparison with atmospheric pressure ionisation mass spectrometry (API MS) as correlations between the two methodologies were observed for heptan-2-one and ethyl hexanoate. Differences between panellists could be observed in the kinetics of aroma release, but for a given panellist, the same pattern of release was observed whatever the aroma compound studied. Also, the authors studied the linear relationships between the concentration of volatile compounds in model cheese and the respective GC peak areas. No competition for the PDMS fibre was observed between the aroma compounds.

12.3.4.4 Other Studies Fan et al.88 used SPME in order to evaluate in vivo human dermal permeability coefficients of VOCs from water. Although in this particular study human skin was not sampled directly, this study shows that microextraction methods such as SPME permit various experimental designs, depending on the experimental objectives of the study in question. In this example, the authors exposed human skin into a custom-built apparatus that contained water with known environmentally relevant amounts of VOC pollutants. SPME was used to periodically sample the water (at 0, 30 and 60 min) in order to monitor the loss of volatiles. The loss of volatiles could be attributed to dermal uptake because other losses such as evaporation and adsorption were accounted for through the use of appropriate controls. The results of this in vivo study indicate that other models (in vitro models and animal models) used for predicting dermal coefficients may actually be seriously underestimating these coefficients. In other words, dermal absorption of VOCs from water may pose more considerable risk than previously thought and in vivo studies can contribute greatly in this area of research.

12.3.5 In Vivo Analysis of Drugs and Metabolites in Flowing Blood Analysis of drugs in biological samples and pharmaceutical products is growing in importance because of the need to understand therapeutic and toxic effects of drugs

436

Handbook of Solid Phase Microextraction

and the continuing development of more selective and effective drugs.67 Interest in the field of drug analysis is focusing on improving methodologies with regard to how quickly, accurately and sensitively the chemicals can be detected. The field is highly dependent on developing new analytical instruments and techniques.95 Knowledge of drug levels in body fluids such as serum, saliva and urine allows the optimisation of pharmacotherapy and provides the basis for studies on patient compliance, bioavailability, pharmacokinetics and genetics, organ function and the influences of co-medication.96 The quantitative and qualitative analysis of drugs and metabolites are extensively applied in pharmacokinetic studies. Variables such as time to maximal concentration in plasma, clearance and bioavailability have to be known before the approval of a new drug. Drugs that can be abused, illicit drugs and intoxications by drugs and poisons are often analysed in clinical and forensic toxicology. Biological materials and pharmaceutical products are very complex mixtures. They often contain proteins, salts, acids, bases and numerous organic compounds that may be similar to the analyte of interest. Furthermore, the analytes often exist at low concentration in these samples. Despite the recent advances in the development of highly efficient analytical instruments for the end-point determination of analytes in biological samples and pharmaceutical products, a sample pretreatment is usually necessary to extract and isolate the analytes of interest from complex matrices. It is often impractical to remove suitable samples for study from the living system, frequently because of their size. In pharmacokinetic studies with rodents, the limited blood volume results in a large number of animals being used to generate profiles with sufficient numbers of data points. For example, the recommendation not to exceed 20% of total blood volume in single experiment in mice (total blood volumeB1.85 mL per 25 g mouse; recommended total sampling volumeB350 μL), severely limits the number of blood samples that can be collected and makes handling of these small blood volumes difficult and prone to losses throughout the sample preparation procedure. If blood were not removed for analysis, smaller numbers of animals would be required and the data generated would be improved by reduced inter-animal variation. In any microextraction or membrane technique, compounds of interest are not exhaustively removed from the investigated system. On the contrary, conditions can be conceived where only a small proportion of the total compound is removed, thus avoiding disturbing the normal balance of the chemical components. In addition, as discussed in more detail in Chapter 11, the amount of analyte extracted by SPME is proportional to the free (unbound) concentration of the analyte. This is a unique advantage of SPME over other sample preparation methods because a direct assay of free concentration can be performed without the separation of phases.97 Experimentally, this translates into a decrease in sensitivity for methods based on SPME, especially for analytes that are highly bound to biopolymers. However, the analytical sensitivity of modern LC MS/MS instrumentation is generally sufficient to detect the small amounts of analytes extracted by SPME, thus opening a whole new field of possible applications of SPME (Table 12.5). The

Analytes and Application Type

Live Sample

Extraction Mode and Time

Extraction Phase

Calibration and/or Identification

Reference

PPY

External calibration (total concentration)

16

Pharmacokinetics of diazepam and metabolites

Whole blood Direct contact, (beagle vein) 30 min

Pharmacokinetics of antibiotics (linezolid)

Whole blood Direct contact, 5 min Hydrophilic PPY (pigs)

External calibration (total concentration)

98

Pharmacokinetics of diazepam and metabolites

Whole blood Direct contact, 30 s, (beagle vein) 2 min

Hydrophilic PPY PEG

External and kinetic calibration (total and free)

15

Pharmacokinetics of diazepam and metabolites

Whole blood Direct contact/ (rats) sampling interface, 2 min, 20 40 s

Hydrophilic PPY

External and kinetic calibration (total and free)

17

Pharmacokinetics of carbamazepine and carbamazepine-10,11-epoxide

Whole blood Direct contact/ (mice) sampling interface, 2 min

Supelco biocompatible 5 μm C18 Kinetic calibration (total and in vivo device, 45 μm thickness free)

47

Pharmacokinetics of fenoterol and Whole blood Direct contact using Culext sampler methoxyfenoterol (rats)

Supelco biocompatible 5 μm reversed-phase amide (RPA) in vivo device, 45 μm thickness

Kinetic calibration (total and free)

99

Global metabolite profiling

Whole blood Direct contact/ (mice) sampling interface, 2 min

Supelco biocompatible 5 μm mixed-mode (C18 1 benzenesulfonic acid) in vivo device, 45 μm thickness

Comparison of metabolite coverage and method precision to solvent precipitation and ultrafiltration

48

Neurotoxicity of toluene

Mouse brain Direct contact, 2 min StableFlex PDMS/DVB 85 μm

External calibration (not matrix matched)

100,101

In Vivo Sampling with Solid-Phase Microextraction

Table 12.5 Use of In Vivo SPME for Direct Extraction of Compounds from Plants and Animals

hippocampus 437

(Continued)

438

Table 12.5 (Continued) Analytes and Application Type

Live Sample

Extraction Mode and Time

Extraction Phase

Calibration and/or Identification

Reference

Fish

Direct contact with fish muscle (dorsalepaxial) 20 min

PDMS hollow tubing (165 μm)

Kinetic calibration (no need for isotopically labelled internal standard)

18

Relative bioavailability of pharmaceuticals in different fish tissue (atorvastatin, bisphenol A, atrazine, gemfibrozil, carbamazepine, naproxen, ibuprofen, diclofenac and fluoxetine)

Fish

Direct simultaneous contact with muscle and adipose fin, 10 min

Spatially resolved 1 mmsegmented PDMS hollow tubing (165 μm)

Kinetic calibration

50,102

Tissue distribution of triazine herbicides (atrazine, simazine and prometryn)

Plants

Direct extraction from stem, reed, shoots and onion, bulbs, 60 min

CW/TPR fibre

External calibration and standard addition

76

Allelopathy of 1,8-cineole, camphor, menthol, coumarin and carveol

Plants

Direct extraction from stem, 60 min

PDMS 100 μm fibre

External calibration (not matrix matched)

77

Direct insertion in leaf, 20 min

PDMS hollow tubing (165 μm)

Kinetic calibration and comparison with in vivo microdialysis

78,79

Pesticide distribution and Jade plant persistence (carbofuran, propoxur, carbaryl, aldicarb and promecarb)

Handbook of Solid Phase Microextraction

Bioaccumulation and toxicity of pharmaceuticals (diphenhydramine, diltiazem, carbamazepine, fluoxetine and norfluoxetine)

In Vivo Sampling with Solid-Phase Microextraction

439

studies performed to date indicate that direct extraction using in vivo SPME can be a useful tool for pharmacokinetic, toxicological and bioaccumulation studies, but of course, the development of other applications in biomedical and life science research can be expected with the recent availability of commercial in vivo SPME probes and further development of this technology. Lord et al.98 assessed the pharmacokinetic profile of linezolid in anesthetised pigs after a single intravenous dosage. Linezolid was selected because it is a newer oxazolidinone antibiotic showing great promise in the treatment infections by several drug-resistant bacteria. The SPME approach (with PPY coating) had a limit of detection of 1 μg/mL in whole blood and a linear dynamic range up to 30 μg/mL. Extraction time was 5 min, followed by another 10 20 min for analysis. Their results demonstrate the potential for fast analysis of circulating drug concentrations in a surgical or intensive-care situations, where catheter access to blood is present as a result of care requirements. A fast in vivo microextraction technique that has the potential to replace current methods of analysis based on blood draws has been reported recently.15 The SPME-based method was faster than conventional approaches, interfered minimally with the investigated system, minimised errors associated with sample preparation and limited personnel exposure to hazardous biological samples. Microextraction devices based on hypodermic needles and stainless steel wires coated with hydrophilic PPY or PEG were used for direct extraction of drugs from the flowing blood of beagle dogs over a period of 8 h. The drugs extracted on the probes were subsequently quantified by LC MS/MS. Two calibration strategies external and standard on the fibre were employed to correlate the amount extracted with the in vivo concentration. Sampling time was 2 min for external calibration and 30 s for standard with the fibre approach, which is a great improvement when compared to the 30 min extraction time previously reported.16 Diazepam and its metabolites were successfully monitored over the course of a full pharmacokinetic study, repeated three times on three beagles. The microextraction technique was validated by comparison with conventional plasma analysis, and a correlation factor of 0.99 was obtained (see Figure 3.18A in Chapter 3). In addition to total concentrations, the method was very useful for determining free drug concentrations (see Figure 3.18B). The results demonstrate the unique advantages of in vivo SPME and highlight its utility as a valuable new tool for fast clinical analysis. A detailed stepby-step protocol for performing in vivo SPME studies on beagles or other large animals has been published recently and can also be easily adopted to other species with minor modifications.103 More recently, in vivo SPME was applied to rodent pharmacokinetic study of diazepam and its metabolites.17 The main objective of this study was to develop an appropriate procedure for sampling of rodents, which are the most commonly studied laboratory animals. For rodents such as rats, direct insertion of SPME probe in blood vessel is not feasible due to the small size of blood vessels and potential for obstruction of blood flow. To address this issue, a new sampling interface was designed, as shown in Figure 12.5. The interface was inserted surgically into carotid artery and blood flow was established through the interface. Repeated

440

Handbook of Solid Phase Microextraction

sampling of the same rat was possible by insertion of SPME probe through the septum of the interface. This approach allowed the researchers to obtain more data points for each animal, and an entire pharmacokinetic time profile could be constructed from a single animal. Furthermore, sampling can be performed simultaneously at multiple sites in one animal without the risk of exsanguination. Such an approach can be used not only for drugs but also for any other endogenous or exogenous compounds. Follow-up studies further extended the utility of this interface sampling approach to mice.47,48 One important aspect of in vivo sampling using SPME that should be emphasised is that the studies can be performed without anaesthesia on freely moving animals, similar to microdialysis. Depending on the exact experimental setup used and the species under study, some handling of the animal may be required (e.g. in the above-described rodent workflow, handling of the rodent is necessary during SPME probe insertion and removal), which may cause some stress on the animal and thus affect concentrations of some of the metabolites. This is an important issue that requires further investigation in future studies on metabolomic scale. An alternative approach currently pursued in Pawliszyn’s laboratory is to minimise handling of the animals during in vivo SPME sampling even further, through the use of commercial automated blood sampler units, which can be modified in order to permit insertion/removal of SPME probes. This approach is described in a recent study of pharmacokinetics of fenoterol and methoxyfenoterol in rats.99 It completely eliminates any animal handling during in vivo SPME sampling and facilitates simultaneous studies on large cohorts of animals, but the main disadvantage is the high cost of these sampling units.

12.3.6 In Vivo Tissue Analysis Within the past 2 years, SPME was applied for the first time for in vivo sampling of analytes from soft tissues in freely moving animals. This represents a very important innovation because the analysis of drugs, metabolites and endogenous compounds in soft tissues usually requires cumbersome and very invasive biopsies. The development of SPME methods for direct extraction from the interstitial fluid of soft tissues in freely moving animals can reduce/eliminate the need to sacrifice experimental animals and eventually greatly improve patient care. Nakajima et al.101 were the first to report the use of SPME for in vivo sampling of toluene concentrations in brain in freely moving mice. In vivo SPME sampling was accomplished by inserting the SPME fibre through a surgically inserted guide cannula into hippocampus. A short sampling time (2 min) was sufficient to extract detectable concentrations of toluene. Such short extraction time enabled the monitoring of toluene changes in brain with time. The authors found that the amount of toluene in the brain depended on the concentration of inhaled toluene and that the levels of toluene would return to pre-exposure levels within 60 min of exposure. For this particular application, in vivo microdialysis was not a feasible alternative because in vivo microdialysis is not suitable for measurement of non-polar analytes such as toluene because of possible adsorptive losses on the microdialysis

In Vivo Sampling with Solid-Phase Microextraction

441

membrane. Furthermore, the use of in vivo SPME enabled the experiments to be performed without anaesthesia during sampling and in freely moving animals, thus eliminating the influence of anaesthesia from the results. These advantages are particularly important for the field of neurotoxicology and indicate that SPME could become a very useful tool in this field. In a subsequent study, same authors studied the toxicological effects of acute toluene exposure using in vivo microdialysis to monitor amino acid neurotransmitter changes after toluene exposure. In order to investigate how changes in neurotransmitter levels correlate with toluene concentrations in the brain, they also monitored toluene concentrations in the brain using in vivo SPME sampling. It is important to mention that in this particular study, the calibration method selected by the authors cannot be expected to provide truly quantitative results, as the calibration standards employed were not matched to the sample matrix under study. For an application like this, the use of kinetic calibration would be expected to provide more accurate quantitative results. However, despite this, the results obtained were able to provide a useful time profile regarding maximum brain levels, and this was the first such study reported in literature. In a more recent study, Zhou et al.18 used in vivo SPME to directly sample muscle of freely moving fish (both in laboratory tanks as well as wild fish in river). The proposed method enabled the monitoring of both free and total concentration of pharmaceuticals in living fish tissue with good detection limits: diltiziem 0.2 pg/g, 3.4 pg/g diphenhydramine, 50.4 pg/g carbamazepine, 8.3 pg/g fluoxetine and 1.6 pg/g norfluoxetine. In fact, diltiziem and diphenhydramine were found in high concentrations in wild fish sampled near a treatment plant. The method is suitable to study bioaccumulation of various compounds in fish, to investigate toxicity as well as to investigate whether or not bioaccumulation of various compounds can cause up-regulation of metabolism over time. It was also shown that even lab-made SPME devices can be sufficiently robust for field sampling, in contrast to in vivo microdialysis, where microdialysis probe may be susceptible to damage during field applications. This fish sampling technique was further refined in follow-up studies in order to introduce the use of spatially resolved SPME fibres and to reduce sampling time to 10 min.102 Five (atrazine, gemfibrozil, carbamazepine, ibuprofen and fluoxetine) of the nine study compounds were found to bioconcentrate in adipose and muscle tissue under controlled laboratory conditions after 8-day exposure. These studies illustrate how in vivo SPME can be a very effective tool in toxicology for studying biochemical changes after exposure to a toxic chemical of interest and to study bioaccumulation. The availability of this sampling method also opens up numerous new possibilities to sample different types of tissues to verify the distribution of various analytes and can become a powerful tool for various toxicological, biomedical and pharmaceutical applications in the future. One of the main disadvantages of using in vivo SPME to monitor the bioconcentration of pharmaceuticals and other pollutant in semisolid tissue samples is the effort required to determine the distribution coefficients (Kfs) between the fibre coating and tissue of interest, which is necessary in order to calculate total tissue

442

Handbook of Solid Phase Microextraction

concentrations.102 A feasible solution lies in the development of a database containing Kfs values for common fibre coatings and common analytes of interest, which would enable a very rapid measurement and enable a high sample throughput for this type of studies.102 A second important area requiring further research is the development of strategies to increase the sensitivity of in vivo SPME further, particularly if using spatially resolved SPME with very short lengths of coatings. Although the utility of in vivo SPME was already demonstrated in field sampling near treatment plant where pollutant concentrations tend to be elevated,18 further improvements in sensitivity would enable the use of in vivo SPME in environmentally relevant field conditions in cases of very low pollutant concentrations and also further extend the range of pollutants that can be monitored with this technique.

12.3.7 Metabolomics Studies Using SPME Metabolomics is an emerging field with the main objective of the analysis of all endogenous metabolites present in a particular living system. As such, metabolomics data are complementary to proteomics, genomics and transcriptomics data and can provide better understanding of dynamic processes occurring in living systems. The potential applications of metabolomics are extremely varied and can range from systems biology, nutrition, pharmaceutical (drug discovery, efficacy, toxicity and personalised medicine), disease models (understanding of disease mechanism, diagnostics and prognostics) and plant biotechnology (improvement of quality and plant interactions with environment), just to name a few.104 Clearly, no single analytical technology can successfully capture all the diverse components of metabolome, so a combination of technologies is used, with nuclear magnetic resonance and MS as the primary choices for this type of untargeted global analysis. SPME can be a very useful sample preparation tool for such studies and has been used extensively to capture volatile metabolites from various living systems. In fact, some of the studies described in previous sections can also be classified as metabolomics studies because they provide insight into metabolites present in the particular living system. For example, Demyttenaere et al.35 successfully used HS-SPME and stir-bar sorptive extraction to sample volatile emissions of five strains of mycotoxic fungi in order to develop a monitoring method for the differentiation of different toxigenic strains of Fusarium. Based on the results of this study, they proposed the use of trichodiene as a volatile biomarker for the detection of infestation of maize by trichothecene-producing fungi. Furthermore, Hurd et al.29 were able to isolate and identify for the first time a putative sex pheromone (1:3 mixture of tetradecanal:pentadecanal) in a praying mantid. They confirmed their findings by performing behavioural tests where the attraction of males to a synthetic mixture of identified pheromones was monitored. The results of this study were unexpected because these two aldehydes were previously reported only as minor components of pheromones of other insects. Soini et al.43 developed a very rapid (10 12 s) sampling methodology for metabolomics studies of human skin emissions. In the proposed approach, a commercially available Twister PDMS stir bar (24-μL volume of extraction phase) is rolled

In Vivo Sampling with Solid-Phase Microextraction

443

on the specific area (10 cm2) of human skin. The authors embedded two internal standards (7-tridecanone and 13C-benzyl alcohol) in the PDMS devices prior to sampling in order to compensate for small variations in sampling and instrumental responses over time. In fact, when results for total of 960 in vivo samples analysed over a 3-month period were evaluated, long-term reproducibility of 14.3% and 14.7% was obtained for two internal standards. The authors also evaluated in vivo reproducibility for multiple same-day samplings (3 25% standard deviation) and samplings performed on different consecutive days (4 30% standard deviation of peak area). This methodology was subsequently applied to metabolomics studies of human sweat involving 200 subjects in order to determine whether human sweat samples can serve as individual chemical fingerprints and whether correlation between human axillary odour and skin microbial profiles exists.44,45 Sweat was found to be a very rich matrix containing hundreds of endogenous and exogenous compounds. The presence of exogenous compounds complicated data interpretation, and although this initial study indicates the potential of human sweat as a characteristic fingerprint, much larger studies are still necessary to examine how stable individual sweat fingerprint is over the long term (especially with aging). Overall, the authors found the use of PDMS Twister bars for skin sampling particularly useful because of (i) very high throughput, (ii) compatibility with automated thermal desorption followed by GC MS analysis, (iii) sufficient sample stability to permit transportation of the devices from field to laboratory and (iv) excellent analytical reproducibility. This last point is particularly important because there is some bias in analytical community that microextraction-based methods such as SPME cannot be used for quantitative analysis due to poor reproducibility. Largescale studies such as these (960 samples analysed over several months) show that with proper experimental design and technique, microextraction methods can have as good reproducibility as traditional sample preparation methods. Zimmermann et al.94 used in vivo SPME to examine metabolic pathways of colon cancer cells. Pronounced differences in metabolic pathways of normal versus colon cancer cells were observed, especially in carbohydrate and lipid metabolism. Methyl dodecanoate, decan-1-ol, heptan-1-ol, 3-methylbutan-1-ol, undecan-2-ol, nonan-2-one and pentadecan-2-one were detected for the first time as part of human metabolic cycles, and they require future isolation of corresponding enzymes and correlation to appropriate metabolic pathways. This study illustrates the great potential of SPME to isolate new and previously unknown metabolites in the field of metabolomics. In fact, the authors selected SPME as the sample preparation method of choice for this study because it causes minimal disturbance to the system under study because it does not require removal of culture headspace. The potential of SPME for identifying potential disease biomarkers in humans was recently demonstrated by in vitro studies of whole blood samples obtained from healthy and individuals with liver cancer.105 A volatile metabolic profile was obtained in combination with GC MS, and statistically significant differences in concentrations of 19 volatile compounds were found using the chi-squared test. Among these, the authors suggest that hexanal, 1-octen-3-ol and octane may have clinical diagnostic value. In the future, such studies may be extended to in vivo.

444

Handbook of Solid Phase Microextraction

For example, Buszewski et al.93 combined in vivo and in vitro SPME approaches to identify potential biomarkers of infection by Helicobacter pylori that could serve as an early diagnosis tool for gastric cancer. The VOC profiles excreted by bacterial cultures in vivo were compared to VOC profiles obtained from stomach tissue samples obtained by excision of cancerous and healthy stomach tissue during gastric cancer surgery. Carbon disulphide and 1-propanol levels were found to be elevated in cancerous stomach tissue and were also found to be directly produced by bacterial cultures. Very recently, in vivo SPME was applied for the first time for metabolomics using direct extraction mode.48 This study had two main objectives: (i) to apply in vivo SPME to study the effects of a single dose of carbamazepine in mice and (ii) to compare the performance of in vivo SPME to blood withdrawal methods followed by ultrafiltration and solvent precipitation. This work demonstrates for the first time that direct in vivo extraction of hundreds or thousands of metabolites is possible with good temporal resolution (2-min sampling), thus making SPME suitable for global LC MS metabolomics studies. Figure 12.20A shows that in vivo SPME can be used to monitor intra- and inter-animal variability in concentrations of various metabolites, with inter-individual variations generally more prominent than the variations within the same individual except for adenosine whose concentrations varied widely on the timescale of the experiment (five consecutive 2-min samplings). Figure 12.20B summarises the main findings for the comparison of in vivo SPME to the methods based on blood withdrawal (ex vivo SPME, ultrafiltration and solvent precipitation) after analysis using the reversedphase LC MS method with a pentaflurophenyl column. The highest metabolite coverage was observed with solvent precipitation (3,609 features), whereas SPME coverage was the lowest (1,868 features) due to lower analytical sensitivity caused by the non-exhaustive nature of the extraction. However, when the metabolite coverage was examined in more detail, it was found that in vivo SPME was capable to detect 70 (positive reversed-phase LC MS) and 85 (negative reversedphase LC MS) unique features. These features were not detected after blood from the same mice was withdrawn and resulting plasma samples subjected to SPME, ultrafiltration or solvent precipitation, and they are likely to represent unstable metabolites or metabolites with fast turnover rates. One of these metabolites was positively identified as β-NAD by comparison with authentic standards, while the characterisation of the other metabolites is currently underway. The preliminary database searching indicates that these species may include glutathione conjugates, thionenes, carotenes and so on. In addition to these unique metabolites not observed by any other method, in vivo SPME had the advantage of more accurate quantitation of certain energy metabolites, such as adenosine monophosphate and glutathione species. The results for the glutathione ratio obtained by all methods are illustrated in Figure 12.21 and show that incorrect ratios from biological perspective were obtained using methods based on blood withdrawal because these methods do not incorporate the metabolism quenching step and permit in-solution oxidation of glutathione. These results are extremely significant because current metabolomics literature does not discuss that metabolite profiles obtained with

(A) 200 % RSD

150 100 50

3,500

60

3,000

50

2,500 40 2,000 30 1,500 20

1,000

Retinol

β-Carotene

Progesterone

CBZ D10

70

10

500

0 PM

Metabolite coverage (No. of features detected)

β-Estradiol

Tryptophan

Phenylalanine

Adenosine

Phenylalanine D5(IS)

Adenine

Glutathione oxidized (2+)

Nicotinamide

Lysine

4,000

0

(C)

QC

UF

In vivo SPME (2 min)

Ex vivo SPME (5 min)

4,000

30

3,500

25

3,000 20

2,500

15

2,000 1,500

10

1,000 5

500

Median RSD (n=7 extractions of pooled human plasma sample)

Metabolite coverage (No. of features detected)

(B)

Same animal (n=5 samplings)

Median RSD (n=4 mice)

Inter-animal (n=8 mice)

Choline

Uridine

Histidine

Glutathione reduced

β-NAD

Glutamic acid

AMP

Sucrose

Glucose D7 (IS)

Glucose+fructose

0

0

0 PM

UF

Ex vivo SPME (5 min)

Figure 12.20 Results of in vivo SPME study for sampling of circulating mouse blood.48 (A) Comparison of intra-animal (n 5 5 samplings of the same animal) and inter-animal variability (n 5 8 animals) in free concentrations of identified metabolites. (B) Comparison of in vivo SPME performance versus ultrafiltration (UF), solvent precipitation (PM) and ex vivo SPME expressed as metabolite coverage (number of features detected) and method precision (median standard deviation of signal intensity of all features detected by each method). (C) Comparison of ex vivo SPME performance versus UF and PM for repeated extraction (n 5 7) of pooled human plasma sample expressed as metabolite coverage (number of features detected) and method precision (median standard deviation of signal intensity of all features detected by each method). (Source: Reprinted with permission from Wiley-VCH. r 2011.)

Handbook of Solid Phase Microextraction

10 9 8 7 6 5 4 3 2 1 0

0.01 Glutathione ratio (GSH/GSSG)

Glutathione ratio (GSH/GSSG)

446

0.008 0.006 0.004 0.002 0 PM

In vivo SPME

Ex vivo SPME Solvent precipitation (PM)

UF

Ultrafiltration (UF)

Figure 12.21 Box plots of the glutathione ratio (reduced/oxidised forms) obtained using ex vivo SPME sampling in circulating mouse blood and after blood withdrawal using SPME, solvent precipitation (PM) and ultrafiltration.74 (Source: Reprinted with permission from Wiley-VCH. r 2011.)

traditional workflow may not be very accurate for some metabolites and may lead to erroneous conclusions during the data interpretation stage. Furthermore, Figure 12.20B also shows method precision achievable by all methods tested and shows that the results for SPME are in line with other methods, with median standard deviation ranging from 46% to 65% for all methods. To further demonstrate that this high standard deviation is reflective of true biological variability rather than methodological shortcomings, Figure 12.20C shows the results for the repeated in vitro extraction of pooled human plasma samples where median standard deviation ranged from 12% to 20% depending on the method employed, and similar total number of metabolites was observed. In conclusion, Figure 12.20 demonstrates that in vivo SPME is suitable for global metabolomics applications with method precision similar to that of traditional methods, while opening up new and exciting possibilities to capture unstable and possibly previously unobserved metabolites. Finally, the comparison of ion maps obtained using all techniques clearly demonstrated that the use of ultrafiltration results in a significant loss of hydrophobic metabolites via adsorption to the membrane, as indicated by the very few metabolites observed with retention times .10 min in both positive and negative electrospray ionization (ESI) modes. These results are consistent both for in vivo sampling of circulating blood of mice48 as well as for in vitro analysis of pooled human plasma sample66 and are in agreement with previous NMR metabolomic studies, where an additional method for extraction of hydrophobic species from the ultrafiltration membrane was proposed by the authors to overcome this difficulty.106 However, this significantly reduces sample throughput and introduces new sample preparation steps, which can cause inadvertent analyte loss or degradation. In contrast, SPME is able to provide a balanced coverage of both polar and hydrophobic species in a single step.

Extraction Method

Analytical Sensitivity

Ionisation Suppression

Free (Unbound) Concentration

Metabolite Coverage

Metabolism Quenching and Ability to Capture Short-lived and Unstable Metabolites

SPME

Low

Low

Yes

Medium

Yes (in vivo)

Solvent precipitation

High

High

No

High

No

Ultrafiltration

High

High

Yes

High for polar metabolites, very poor for hydrophobic metabolites

No

In Vivo Sampling with Solid-Phase Microextraction

Table 12.6 Summary of the Performance Characteristics for Sample Preparation Methods Employed in Global Metabolomics Studies by LC MS

447

448

Handbook of Solid Phase Microextraction

The main performance characteristics of SPME versus ultrafiltration and solvent precipitation for global metabolomics applications are summarised in Table 12.6 to help the user select the most promising method for a given research question. A combination of free and total method (e.g. in vivo SPME and solvent precipitation) could provide useful complementary information that may improve the interpretation of metabolomic data in future. In vivo SPME, in combination with LC-MS/MS quantification, has also been recently applied successfully to measure neurotransmitters in brain of freely moving rats. In this targeted metabolomics study, biocompatible SPME coatings were used for simultaneous extraction and measurement of monoamine and amino acid neurotransmitters without the need for further derivatisation. The technique offers an alternative sampling method that can be used to monitor multiple neurotransmitters that may be involved in neuronal activity alterations. The in vivo SPME technique was validated by the simultaneous use of microdialysis to monitor dopamine and serotonin changes in the brain with fluoxetine dosing. In addition, in the same set of experiments, the in vivo SPME probes were used to further monitor changes in aminobutyric acid and glutamic acid. Both left and right hemispheres of the brain were monitored at the same time. SPME and microdialysis probes were each placed in the left and right brain hemispheres, respectively. Both SPME and microdialysis detected significant changes in serotonin after fluoxetine administration to the rats, whereas there was no significant increase recorded for dopamine. Control rats, which were not administered with any drug, did not show any changes in the basal levels of all for neurotransmitters monitored. Owing to the successful application of in vivo SPME to measure these endogenous metabolites in the brain neurotransmission, the sampling technique was applied to rats made to undergo deep brain stimulation in the ventromedial prefrontal cortex. This was done to study any involvement of neurotransmitters in the therapeutic effects of deep brain stimulation. In this study, only SPME probes were used to monitor the neurotransmitters from both brain hemispheres. Results show increased levels of serotonin during deep brain stimulation of ventromedial prefrontal cortex of the rat, which corroborated independent studies conducted using microdialysis.

12.4

Conclusions

SPME is a simple, solvent-free and reliable microextraction technique that has continued to revolutionise sampling and sample preparation ever since its discovery two decades ago. The small dimensions of SPME devices and their solvent-free feature enable convenient in vivo sampling. SPME offers significant advantages over current methods: it is much faster, no biological sample processing is required, supplementary parameters are revealed (free concentrations, binding constants) and the analysis of extracted compounds can be performed with highly specific instruments, such as GC MS or LC MS/MS. SPME has been successfully applied to a wide range of analytical applications involving sampling of living organisms. In

In Vivo Sampling with Solid-Phase Microextraction

449

some cases, these studies would not be practical with other sample preparation methods because they would cause severe damage to the live organisms or would demand their sacrifice. Methodologies similar to those applied in beagles for monitoring free and total circulating drug concentrations could be directly transferred to human subjects, as biocompatible materials were used and the dimensions of the SPME devices were appropriate for human veins. Still, more tests are required to ensure method safety and regulatory approval. The development of biocompatible extraction phases for SPME has led to significant advances in bioanalysis: all sample preparation steps can be combined into a single one, even for complex biological samples such as whole blood or plasma. Furthermore, biocompatible devices permit direct extraction of target analytes from the flowing blood and soft tissues of living organisms. Future research in this area should focus on further automation of sampling and analysis, further miniaturisation of SPME devices and utilisation of highly specific extraction phases. Direct extraction of target analytes from complex biological samples often results in co-extraction of interference compounds. In the case of solid (porous) coatings, co-extraction is an important issue because the interferant may displace the analyte from the extraction phase. Besides, analysis of a complex extract requires a suitable separation method, such as GC or LC, which increases the total processing time. An obvious choice is the application of extraction phases that are specific for the target compound. SPME devices based on molecularly imprinted polymers (MIPs) or antibodies would possess unsurpassed sensitivity and specificity. Such devices could be interfaced directly with mass spectrometers, resulting in very fast analytical methods. When the concentration of the target compounds changes quickly with time, SPME fibres with short equilibration time in static conditions should be used. Because such fibres usually extract a low amount of target compound, an additional strain is imposed on the detection and quantification method. In this case, a good solution is the application of one of the kinetic calibration methods, so there is no need to wait for equilibrium to be established. The emerging field of metabolomics is another important application area where in vivo SPME can make a significant contribution. An increase in interest in this type of studies for microorganisms, plants, animals and humans can be expected in the future. The amount of analyte extracted by SPME is proportional to free (unbound) concentration of metabolite in the biological sample. Only unbound metabolites are biologically active, so using SPME as the sample preparation method may be helpful in metabolomics studies aiming to understand biological processes, although this aspect still remains to be fully explored. The second major contribution of SPME in this area is the ability to capture stable and/or short-lived species and the metabolites with fast turnover rates (such as energy metabolites) that are not observed by other methods. The ability of in vivo SPME to capture an unstable portion of metabolome more accurately is due to the incorporation of metabolism quenching step directly during the sampling by the virtue of the fact that proteins are not co-extracted into the coating. More extensive characterisation of the metabolites uniquely captured by in vivo SPME is still needed and currently

450

Handbook of Solid Phase Microextraction

underway, but early studies indicate that in vivo SPME is particularly powerful for sampling of nucleosides, carotenes and glutathione species. This makes proposed technology particularly interesting from the point of view of studying reactive drug metabolites and antiviral nucleoside analogue drugs as well as for both targeted and global metabolomic studies focused on these biological pathways. Further work in the area of reactive drug metabolites could explore the development of new biocompatible SPME coatings using glutathione or methoxylamine trapping agents, for example.64,107,108 In comparison to the gold standard technique of microdialysis, in vivo SPME offers unique advantages, including (i) better compatibility with LC MS analysis due to reduced potential for ionisation suppression, (ii) more balanced simultaneous extraction of both hydrophobic and hydrophilic analytes, (iii) less damage to living systems due to reduced dimensions of SPME probes and (iv) improved spatial resolution using specially designed SPME devices.74 In vivo SPME is typically not applicable for quantitative studies of very rapid processes occurring on time scales of seconds and is also not generally applicable for processes with large concentration changes per unit time. Thus, the main advantage of in vivo microdialysis compared to SPME is improved temporal resolution because microdialysis allows the monitoring of analyte concentrations in almost real time, the ability to detect rapidly changing concentrations and improved analytical sensitivity.

References 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12.

13. 14. 15. 16.

F Augusto & A Luiz Pires Valente, TrAC, Trends Anal Chem 21 (2002) 428 DR Walt, Science 308 (2005) 217 W Tseng, Y Sun, C Lee, M Yang & Y Huang, Anal Sci 21 (2005) 225 M Brunner, A Schmiedberger, R Schmid, D Jager, E Piegler, HG Eichler & M Muller, Br J Clin Pharmacol 46 (1998) 431 E Benfeldt, J Serup & T Menne, Acta Derm Venereol 79 (1999) 338 I Dostalova, K Pacak & J Nedvidkova, Int J Biol Macromol 32 (2003) 205 RT Kennedy, JE Thompson & TW Vickroy, J Neurosci Methods 114 (2002) 39 MJ Cano-Cebrian, T Zornoza, A Polache & L Granero, Curr Drug Metab 6 (2006) 83 G Leegsma-Vogt, E Janle, SR Ash, K Venema & J Korf, Life Sci 73 (2003) 2005 C Jianrong, M Yuqing, H Nongyue, W Xiaohua & L Sijiao, Biotechnol Adv 22 (2004) 505 G Jia, H Wang, L Yan, X Wang, R Pei, T Yan, Y Zhao & X Guo, Environ Sci Technol 39 (2005) 1378 AA Shvedova, AR Murray, VJ Johnson, O Gorelik, S Arepalli, A Hubbs, RR Mercer, P Baron, AD Maynard, VE Kagan, AI Potapovich, V Castranova and E Kisin, 229th ACS National Meeting, San Diego, CA, March 13 17, 2005, IEC-014. J Pawliszyn, Aust J Chem 56 (2003) 155 RM Crewe, RFA Moritz & MG Lattorff, Chemoecology 14 (2004) 77 FM Musteata, ML Musteata & J Pawliszyn, Clin Chem 52 (2006) 708 HL Lord, RP Grant, M Walles, B Incledon, B Fahie & JB Pawliszyn, Anal Chem 75 (2003) 5103

In Vivo Sampling with Solid-Phase Microextraction

451

17. FM Musteata, I de Lannoy, B Gien & J Pawliszyn, J Pharm Biomed Anal 47 (2008) 907 18. SN Zhou, KD Oakes, MR Servos & J Pawliszyn, Environ Sci Technol 42 (2008) 6073 19. I Saı¨d, G Costagliola, I Leoncini & C Rivault, J Insect Phys 51 (2005) 995 20. J Tentschert, HJ Bestmann & J Heinze, Chemoecology 12 (2002) 15 21. C Peeters, T Monnin & C Malosse, Proc R Soc B Biol Sci 266 (1999) 1323 22. T Monnin, C Malusse & C Peeters, J Chem Ecol 24 (1998) 473 23. J Tentschert, K Kolmer, B Ho¨lldobler, H Bestmann, J Delabie & J Heinze, Naturwissenschaften 88 (2001) 175 24. DC Gilley, G DeGrandi-Hoffman & JE Hooper, J Insect Phys 52 (2006) 520 25. D Djozan, T Baheri, R Farshbaf & S Azhari, Anal Chim Acta 554 (2005) 197 ¨ strand, G Bergstro¨m, A Wassgren, M Auger-Rozenberg, C Geri, 26. O Anderbrant, F O E Hedenstro¨m, H Ho¨gberg, A Herz & W Heitland, Chemoecology 15 (2005) 147 27. D Rochat, P Ramirez-Lucas, C Malosse, R Aldana, T Kakul & J Morin, J Chromatogr A 885 (2000) 433 28. X Chen, K Nakamuta, T Nakanishi, T Nakashima, M Tokoro, F Mochizuki & T Fukumoto, J Chem Ecol 32 (2006) 669 29. LE Hurd, FR Prete, TH Jones, TB Singh, JE Co & RT Portman, J Chem Ecol 30 (2004) 155 30. KL Prudic, S Khera, A Solyom & BN Timmermann, J Chem Ecol 33 (2007) 1149 31. L Cai, JA Koziel & ME O’Neal, J Chromatogr A 1147 (2007) 66 32. V Witte, L Abrell, AB Attygalle, X Wu & J Meinwald, Chemoecology 17 (2007) 63 33. JC Verdonk, CHR De Vos, HA Verhoeven, MA Haring, AJ Van Tunen & RC Schuurink, Phytochemistry 62 (2003) 997 34. JJ Beck, L Smith & GB Merrill, J Agric Food Chem 56 (2008) 2759 35. JCR Demyttenaere, RM Morin˜a, N De Kimpe & P Sandra, J Chromatogr A 1027 (2004) 147 36. JCR Demyttenaere, J Vanoverschelde & N De Kimpe, J Chromatogr A 1027 (2004) 137 37. Z Zhang, J Cai, G Ruan & G Li, J Chromatogr B 822 (2005) 244 38. C Bicchi, C Cordero, E Liberto, P Rubiolo, B Sgorbini & P Sandra, J Chromatogr A 1148 (2007) 137 39. M Gallagher, CJ Wysocki, JJ Leyden, AI Spielman, X Sun & G Preti, Br J Derm 159 (2008) 780 40. X Chen, F Xu, Y Wang, Y Pan, D Lu, P Wang, K Ying, E Chen & W Zhang, Cancer 110 (2007) 835 41. H Yu, L Xu & P Wang, J Chromatogr B 826 (2005) 69 42. M Syhre, JM Scotter & ST Chambers, Med Mycol 46 (2008) 209 43. HA Soini, KE Bruce, I Klouckova, RG Brereton, DJ Penn & MV Novotny, Anal Chem 78 (2006) 7161 44. Y Xu, F Gong, SJ Dixon, RG Brereton, HA Soini, MV Novotny, E Oberzaucher, K Grammer & DJ Penn, Anal Chem 79 (2007) 5633 45. Y Xu, SJ Dixon, RG Brereton, HA Soini, MV Novotny, K Trebesius, I Bergmaier, E Oberzaucher, K Grammer & DJ Penn, Metabolomics 3 (2007) 427 46. S Riazanskaia, G Blackburn, M Harker, D Taylor & CLP Thomas, Analyst 133 (2008) 1020 47. D Vuckovic, I de Lannoy, B Gien, Y Yang, FM Musteata, R Shirey, L Sidisky and J Pawliszyn, J Chromatogr A 1218 (2011) 3367

452

Handbook of Solid Phase Microextraction

48. D Vuckovic, I de Lannoy, B Gien, B Shirey, L Sidisky, S Dutta and J Pawliszyn, Angew Chem Int Ed 50 (2011) 5344 49. D Vuckovic, X Zhang, E Cudjoe & J Pawliszyn, J Chromatogr A 1217 (2010) 4041 50. X. Zhang, J. Cai, K.D. Oakes, F. Breton, M.R. Servos & J Pawliszyn, Anal Chem 81 (2009) 7349 51. G Ouyang, D Vuckovic and J Pawliszyn, Chem Rev 111 (2011) 2784 52. FM Musteata, M Walles & J Pawliszyn, Anal Chim Acta 537 (2005) 231 53. ML Musteata, FM Musteata & J Pawliszyn, Anal Chem 79 (2007) 6903 54. J Chanard, S Lavaud, C Randoux & P Rieu, Nephrol Dial Transplant 18 (2003) 252 55. A Es-haghi, X Zhang, FM Musteata, H Bagheri & J Pawliszyn, Analyst 132 (2007) 672 56. X Zhang, A Es-Haghi, FM Musteata, G Ouyang & J Pawliszyn, Anal Chem 79 (2007) 4507 57. Y Wang, S Nacson & J Pawliszyn, Anal Chim Acta 582 (2007) 50 58. JK Schubert, W Miekisch, P Fuchs, N Scherzer, H Lord, J Pawliszyn & RG Mundkowski, Clin Chim Acta 386 (2007) 57 59. J Wu & J Pawliszyn, Anal Chim Acta 520 (2004) 257 60. WM Mullett & J Pawliszyn, J Pharm Biomed Anal 26 (2001) 899 61. M Walles, WM Mullett & J Pawliszyn, J Chromatogr A 1025 (2004) 85 62. WM Mullett & J Pawliszyn, Anal Chem 74 (2002) 1081 63. J Lambert, WM Mullett, E Kwong & D Lubda, J Chromatogr A 1075 (2005) 64. D Vuckovic, Solid-Phase Microextraction as Sample Preparation Method for Metabolomics, Ph.D. Thesis, University of Waterloo, Waterloo, ON, Canada (2010) 65. D Vuckovic, R Shirey, Y Chen, L Sidisky, C Aurand, K Stenerson & J Pawliszyn, Anal Chim Acta 638 (2009) 175 66. D Vuckovic & J Pawliszyn, Anal Chem 83 (2011) 1944 67. J Pawliszyn, Fundamental and New Directions in Sample Preparation (2002) 68. Y Chen & J Pawliszyn, Anal Chem 76 (2004) 5807 69. J Yeung, D Vuckovic & J Pawliszyn, Anal Chim Acta 665 (2010) 160 70. J O’Reilly, Q Wang, L Setkova, JP Hutchinson, Y Chen, HL Lord, CN Linton & J Pawliszyn, J Sep Sci 28 (2005) 2010 71. G Theodoridis, EHM Koster & GJ De Jong, J Chromatogr B 745 (2000) 49 72. M Abdel-Rehim, M Bielenstein & T Arvidsson, J Microcolumn Sep 12 (2000) 308 73. FM Musteata & J Pawliszyn, J Biochem Biophys Methods 70 (2007) 181 74. D Vuckovic, S Risticevic and J Pawliszyn, Angew Chem Int Ed 50 (2011) 5618 75. EE Stashenko & JR Martı´nez, J Sep Sci 31 (2008) 2022 76. HL Lord, M Mo¨der, P Popp & JB Pawliszyn, Analyst 129 (2004) 107 77. RX Loi, MC Solar & JD Weidenhamer, J Chem Ecol 34 (2008) 70 78. SN Zhou, W Zhao & J Pawliszyn, Anal Chem 80 (2008) 481 79. SN Zhou, G Ouyang & J Pawliszyn, J Chromatogr A 1196 1197 (2008) 46 80. A Mallouchos, M Komaitis, A Koutinas & M Kanellaki, J Agric Food Chem 50 (2002) 3840 81. A Halasz & J Hawari, J Chromatogr Sci 44 (2006) 379 82. T MacPherson, CW Greer, E Zhou, AM Jones, G Wisse, PCK Lau, B Sankey, MJ Grossman & J Hawari, Environ Sci Tech 32 (1998) 421 83. LG Whyte, J Hawari, E Zhou, L Bourbonniere, WE Inniss & CW Greer, Appl Environ Microbiol 64 (1998) 2578 84. J Hawari, A Halasz, L Paquet, E Zhou, B Spencer, G Ampleman & S Thiboutot, Appl Environ Microbiol 64 (1998) 2200

In Vivo Sampling with Solid-Phase Microextraction

453

85. M Mirata, M Wu¨st, A Mosandl & J Schrader, J Agric Food Chem 56 (2008) 3287 86. S Colazza, G Aquila, C De Pasquale, E Peri & JG Millar, J Chem Ecol 33 (2007) 1405 87. S Colazza, G Aquila, C De Pasquale, E Peri & JG Millar, J Chem Ecol 33 (2007) 2345 88. VS Fan, RE Savage Jr. & TJ Buckley, Toxicol Mech Methods 17 (2007) 295 89. JK Schubert, W Miekisch, T Birken, K Geiger & GFE Noldge-Schomburg, Biomarkers 10 (2005) 138 90. E Pionnier, C Chabanet, L Mioche, J Le Que´re´ & C Salles, J Agric Food Chem 52 (2004) 557 91. E Pionnier, E Se´mon, C Chabanet & C Salles, Sci Des Aliments 25 (2005) 193 92. W Miekisch, P Fuchs, S Kamysek, C Neumann & JK Schubert, Clin Chim Acta 395 (2008) 32 93. B Buszewski, A Ulanowska, T Ligor, M Jackowski, E Kłodzi´nska & J Szeliga, J Chromatogr B 868 (2008) 88 94. D Zimmermann, M Hartmann, MP Moyer, J Nolte & JI Baumbach, Metabolomics 3 (2007) 13 95. T Kumazawa, X Lee, K Sato & O Suzuki, Anal Chim Acta 492 (2003) 49 96. H Kataoka, TrAC, Trends Anal Chem 22 (2003) 232 97. FM Musteata & J Pawliszyn, J Proteome Res 4 (2005) 789 98. HL Lord, R Mundkowski, W Miekisch, J Schubert and J Pawliszyn, HTC-9, Ninth International Symposium on Hyphenated Techniques in Chromatography, P110 (2006) 99. CY Yeung, In vivo Calibration Methods of SPME and Application to Pharmacokinetic Studies, M.Sc. Thesis, University of Waterloo, Waterloo, ON, Canada (2009) 100. T Win-Shwe, D Mitsushima, D Nakajima, S Ahmed, S Yamamoto, S Tsukahara, M Kakeyama, S Goto & H Fujimaki, Toxicol Lett 168 (2007) 75 101. D Nakajima, T-T Win-Shwe, M Kakeyama, H Fujimaki & S Goto, NeuroToxicology 27 (2006) 615 102. X Zhang, KD Oakes, S Cui, L Bragg, MR Servos & J Pawliszyn, Environ Sci Technol 44 (2010) 3417 103. HL Lord, X Zhang, FM Musteata, D Vuckovic and J Pawliszyn, Nat Protoc 6 (2011) 896 104. EJ Want, A Nordstro¨m, H Morita & G Siuzdak, J Proteome Res 6 (2007) 459 105. R Xue, L Dong, S Zhang, C Deng, T Liu, J Wang & X Shen, Rapid Commun Mass Spectrom 22 (2008) 1181 106. S Tiziani, A Emwas, A Lodi, C Ludwig, CM Bunce, MR Viant & UL Gu¨nther, Anal Biochem 377 (2008) 16 107. A Tolonen, M Turpeinen & O Pelkonen, Drug Discov Today 14 (2009) 120 108. S Ma, SK Chowdhury & KB Alton, Curr Drug Metab 7 (2006) 503

13 Solid-Phase Microextraction Protocols

Dajana Vuckovica,b, Sanja Risticevica and Janusz Pawliszyna a

Department of Chemistry, University of Waterloo, Waterloo, Ontario, Canada Donnelly Centre for Cellular and Biomolecular Research, University of Toronto, Toronto, Ontario, Canada

b

13.1

Protocol for Automated High-Throughput SPME-LC using the Concept 96 Robotic Sample Preparation Station

This protocol describes a generalised sample preparation procedure using automated solid-phase microextraction (SPME) or thin-film microextraction (TFME) and the Concept 96 robotic station (PAS Technology) for the extraction of target compounds from a complex biofluid, tissue homogenate or any other complex matrix of interest.1 More generally, this procedure can also be applied for the development of any high-throughput method for the analysis of non-volatile target analytes using Concept 96. This sample preparation technique is ideally suited for hyphenation with liquid chromatography (LC) due to the non-volatile nature of the analytes. Any detection system can be used, but mass spectrometric (MS) detection is recommended because of high sensitivity and selectivity. The design and main features of the Concept 96 station are described in Section 5.2.2. Thin-film geometry provides enhanced sensitivity and extraction rates, and this design is used in the commercial device.2 Therefore, for the remainder of this protocol, the technique will be referred to as TFME. The main steps of an automated TFME procedure are: (i) preconditioning, (ii) extraction, (iii) wash and (iv) desorption, as shown in Figure 13.1. Prior to the use of this procedure, SPME/TFME method development should be carried out according to the procedures described in Chapter 7 and/or the general method development protocol laid out by Risticevic et al.3 The main parameters that should be optimised for this application are: (i) type of coating, (ii) coating conditioning procedure, (iii) extraction time, (iv) desorption solvent and desorption time and (v) fast wash procedure step. Depending on the particular application, the user may also want to perform sample pH adjustment (incorporation of such step is recommended if the analyte is ionisable) and/or evaporation/reconstitution step in order to enhance sensitivity. Figure 13.2 summarises the parameters which require optimisation and offers practical suggestions during such optimisation. Handbook of Solid Phase Microextraction. DOI: 10.1016/B978-0-12-416017-0.00013-9 © 2012 Elsevier Inc. All rights reserved.

456

Handbook of Solid Phase Microextraction

(Optional) Evaporation/ reconstitution

Preconditioning

Extraction

Solvent desorption

Wash

Figure 13.1 Main steps of automated high-throughput TFME using Concept 96.

One important aspect for the development of any SPME-LC method (regardless of degree of automation) is the verification of desorption efficiency. This is performed by evaluation of carry-over, which is defined as the amount of analyte remaining in the coating after the completion of a full desorption step. Figure 13.3 describes a step-by-step procedure to use during the evaluation of carry-over in SPME-LC applications.1 Furthermore, as discussed in Chapters 6 and 11, SPME can be used to determine either free or total concentration of analyte in presence of binding matrix. This feature of SPME is particularly important in bioanalysis because only the free drug concentration is responsible for pharmacological activity. Different calibration strategies (matrix-free, matrix-matched or both) can be employed when developing methods using Concept 96 depending on the particular application. The main steps of these procedures are summarised in Figure 13.4. Another important aspect of validation of any method is the determination of absolute recovery, so this procedure is included in Figure 13.4 as well. It is important to emphasise that in most cases, the absolute recovery of a microextraction method such as SPME or TFME will not be exhaustive (usually defined as greater than 80%), but much lower. However, this does not preclude the use of microextraction methods in regulated bioanalysis, so long as it is demonstrated that the absolute recovery is consistent across all concentration levels.4 Therefore, the procedure for the determination of absolute recovery should be carried out on a minimum of three different analyte concentrations (low, medium and high), and it should be verified that the absolute recovery is consistent and reproducible at all levels tested.

13.1.1 Concept 96 Procedure1 Step 1: Place fibres or thin films with appropriate coating on PAS Concept 96 autosampler by attaching the 96 thin film device on the appropriate arm. Step 2: Initialise the autosampler by running the Start-up script in the Concept software. Check that the thin films are positioned exactly in the centre of the wells at all agitator positions. If not, adjust x and y positions as required using the Concept software. Once the positions are properly set, save the positions in the Concept software. Verify the vertical position

Solid-Phase Microextraction Protocols

Type of coating

Extraction time

Desorption solvent

Desorption time

457

Match coating polarity to analyte polarity

Keep extraction time between 5 and 120 min if possible

Use desorption solvent compatible with LC method Verify chromatographic peak shape Use sufficient solvent volume to immerse the entire coating Verify carry-over

Wash procedure

Keep very short (<1 min) Verify that no loss of analyte occurs

Conditioning procedure

Do not let coating dry out between preconditioning and extraction

(optional) Ionic strength

(optional) pH

(optional) Sample volume (optional) Evaporation time (optional) Reconstitution solvent

Add salt and/or adjust pH to increase extraction efficiency Add salt and/or adjust pH to normalise the variability in sample pH and/or salt content Select sample volume sufficient to immerse entire length of coating Verify that no well cross-talk occurs, if applicable Add evaporation/reconstitution step to increase analytical sensitivity Use reconstitution solvent compatible with LC method Verify chromatographic peak shape Verify reconstitution efficiency

Figure 13.2 Method development recommendations for automated SPME/TFME. of the fibres or thin films in the well to ensure that the extraction phase does not touch the bottom of the wells (place the bottom edge of thin films 2
3 mm from the bottom). Step 3: Dispense 1 mL of conditioning solvent into each well of a 96-well plate (using manual pipetting, an automated dispenser or a robotic liquid-handling station). Lower the thin-film device into the plate containing conditioning solvent. Agitate (850 rpm) for a

458

Handbook of Solid Phase Microextraction

Step 1

Step 2 Step 3 Step 4 Step 5

Perform SPME procedure using standard with high analyte concentration (concentration at the upper limit of linear dynamic range) using optimized extraction time and coating. Perform desorption using selected desorption solvent and time. Perform second desorption using fresh portion of same desorption solvent used in Step 2 for the same amount of time used in Step 2. Analyze the samples from Steps 2 and 3 and determine the amount of analyte(s) present. Calculate % carry-over using the formula amount of analyte remaining in the coating

%Carry-over =

Step 6

× %100 total amount of analyte extracted in Steps 2 and 3

where the amount remaining in the coating is the amount found in Step 3, while the total amount extracted is the sum of the amounts obtained in Steps 2 and 3. For quantitative analysis, it is recommended to find desorption conditions that will keep carry-over <2%. Any carry-over should be removed completely prior to subsequent extraction.

Figure 13.3 Procedure for the evaluation of carry-over for methods relying on solvent desorption.

Calibration strategies

Matrix-free

Matrix-matched

To obtain free concentration

To obtain total concentration

1. 1.

2. 3. 4. 5.

6.

Prepare set of calibration standards which contain NO binding matrix. Add IS. Perform entire SPME procedure. Inject in LC–MS/MS. Plot standard concentration versus area ratio (analyte/IS). Perform linear regression (non-weighted or weighted).

2. 3. 4. 5. 6.

7.

Prepare set of calibration standards using appropriate blank biological matrix. Allow standards to stand to allow binding to take place. Add IS. Perform entire SPME procedure. Inject in LC–MS/MS. Plot standard concentration versus area ratio (analyte/IS). Perform linear regression (non-weighted or weighted).

Liquid injection To evaluate absolute recovery

1.

2. 3. 4.

Prepare set of calibration standards in desorption solvent. Add IS. Inject directly in LC–MS/MS. Calculate absolute recovery.

%Recovery =

matrix − matched SPME x100% direct injection

Figure 13.4 Detailed procedures for SPME calibration in bioanalytical applications. pre-determined amount of time. Fibre preconditioning may be omitted for some coatings (e.g. carbon-tape coating does not require any fibre preconditioning and can be used as is5). Step 4: Dispense biofluid samples and calibration standards into the wells of a 96-well plate. For optimum sensitivity using 1.2 mL deep-well plates, 800
1,000 µL of biofluid is recommended. Smaller volumes may be used, but ensure that the sample volume selected allows the immersion of the entire length of coating.

Solid-Phase Microextraction Protocols

459

Step 5: Add the appropriate solution to adjust pH/ionic strength if necessary (determined during method development). Add internal standard (IS) solution (5 µL volume of IS is recommended) to all samples and calibration standards. The same volume of IS should be added to all samples and standards. IS can be dissolved in any appropriate organic or aqueous solvent, but the volume of added standard should be kept as small as possible. Mix the contents of the wells for 5 min at 850 rpm. Step 6: Load the appropriate SPME method using Concept software. An example of the screen used for input of SPME method parameters is shown in Figure 13.5. Input appropriate times is given in seconds and agitator frequency in rpm.

Figure 13.5 (A) Main Concept screen for input of SPME method parameters and (B) screen for input of additional agitator parameters, including extraction temperature.

460

Handbook of Solid Phase Microextraction

Step 7: Pipet wash solvent in a new 96-well plate and place in the wash position of a PAS autosampler. Keep wash time short to ensure no desorption of the analyte of interest. Step 8: Pipet desorption solvent (typically 800
1,000 µl) in a new 96-well plate and place in the desorption position of a PAS autosampler. Step 9: If evaporation/preconcentration step is used, place fresh reconstitution solvent in the appropriate vial. Step 10: Activate the SPME method through Concept software. Step 11: After the method is completed, transfer the completed multi-well plate for injection into LC-MS/MS. This can be performed either manually or using a plate feeder to perform this step. Any commercial high-performance liquid chromatography (HPLC) autosampler can be used for sample injection so long as it is capable of handling commercial 96-well plates. Seal the plate containing final sample extracts using aluminium foil or other appropriate method in order to minimise solvent evaporation. Sample plates waiting for injection can also be refrigerated at 4 C to prevent solvent evaporation and analyte degradation. Step 12: Inject standards and samples using appropriate LC-MS/MS method using appropriate standard operating procedure. Integrate all analyte and IS peaks using instrument software. Calculate area ratios (peak area of analyte/peak area of IS). Check the chromatograms for blank injections (blank desorption solvent, blank fibre, and so on). Step 13: Construct calibration curves (matrix-matched, matrix-free or both, as described in Figure 13.4) by plotting standard concentration versus area ratios. Use these calibration curves to determine the amount of analyte in each sample.

13.1.2 Important Considerations and Troubleshooting Strategies Device setup G Do not overfill wells with sample. G Ensure that the devices are centred in the wells and that they do not touch the bottom of the wells. G Ensure that Concept 96 is placed on a level surface. G Use the highest agitation speed that does not result in sample spilling. The speed of 850 rpm is recommended for 800
1,000 µL sample volumes when 1.2-mL-deep wells are used. G For manual dispensing of volumes less than 10 µL, it is recommended to use a positive displacement pipette instead of an air displacement pipette to ensure acceptable accuracy and precision. The positive-displacement type of pipette is known to introduce an uncertainty of approximately 1%, while a typical air-displacement pipette can introduce much larger uncertainty (approximately 15%) when pipetting organic solvents.6 G A simple way to evaluate the overall performance of Concept 96 (coating extraction efficiency, uniformity of agitation, and so on) is to perform a control extraction periodically using all 96 thin films. The amount extracted can then be plotted against the well position, as illustrated in Figure 13.6, and any outliers can be easily detected.1 SPME recommendations G Do not allow coating to dry between preconditioning and extraction steps in order to obtain the best extraction efficiency.

Solid-Phase Microextraction Protocols

461

Evaluation of inter-fibre reproducibility and/or success of coating procedure

Amount extracted (ng)

40 35 30 25 20 15 10 0

12

24

36

Individual results

48 Fibre UCL

60

LCL

72

84

96

Mean

Figure 13.6 Control chart used for the evaluation of Concept 96 performance and/or the success of the coating procedure.1

G

G

G

G G

G

G

G

G G

G

Use a fresh portion of conditioning solvent each time the conditioning step is performed to eliminate any cross-contamination. Verify that no analyte is lost during the wash step by collecting the wash solution and analysing for the presence of analyte. This test should be carried out using a high-concentration standard to facilitate the detection of analyte loss (if any analyte is desorbed during the wash step). We typically employ a 30-s wash step in purified water with the autosampler programmed to move the fibres up and down three times into and out of the wash solution, with no additional agitation employed. The use of purified water as a wash solvent works well for non-polar to semipolar analytes. For polar analytes, pure organic solvent can be employed in this step because the use of purified water can result in significant loss of analyte. Eliminate intrinsic sample pH and ionic strength variations to improve method precision and accuracy for ionisable analytes. Use high-purity solvents (HPLC grade or MS grade) for desorption. Choose the desorption conditions that minimise or eliminate carry-over. If coatings are reusable and any carry-over is detected, this carry-over must be removed using the appropriate cleaning procedure prior to subsequent analysis. Use the appropriate calibration strategy (IS, fibre constant or standard-on-fibre) in order to compensate for any inter-fibre variability in the amount extracted.1 When preparing calibration standards, minimise the proportion of organic solvent and keep constant for all the standards. The presence of the organic modifier can affect the partitioning equilibrium between analyte and the coating, so it is recommended to keep this concentration below 1% organic. Choose the concentrations of the standards that (i) span the entire range of expected sample concentrations and (ii) do not exceed the linear dynamic range of the instrument employed. Always pipette the same volumes of sample and calibrant. Perform the calibration in duplicate and perform SPME on calibrants and unknown samples simultaneously for the most reliable quantitative results. Ensure the temperature of samples and standards is the same prior to extraction.

462 G

G

G

Handbook of Solid Phase Microextraction

For matrix-matched calibration, after the biological fluid is spiked with analyte standard, incubate it for a preset time in order to establish equilibrium between the analyte and the matrix. Agitation can be employed in order to speed this process. Typically, 60 min is sufficient to establish equilibrium, but this should be verified experimentally during method development. For example, carbamazepine was found to require longer incubation times.7 If the r2 obtained for linear regression analysis is poor (, 0.99), examine the data for deviations from linearity at high concentrations. This is usually caused by exceeding the linear range of the instrument. To extend the linear range, such high concentration samples can be diluted as appropriate and re-analysed. Also, verify that the proportion of organic solvent is the same in all standards. Weighted linear regression usually provides better accuracy for low concentration standards when using electrospray liquid chromatography-mass spectrometry (LC-MS) methods. Verify that the absolute recovery is consistent and reproducible across all concentration levels using the procedure described in Figure 13.4.

LC-MS recommendations Place a 0.45-µm filter (with small internal volume to minimise band broadening) before the entrance to detector to ensure optimum nebuliser performance if using MS detection. G Remember to evaluate the relative and absolute matrix effects if using electrospray MS detection.8 For the evaluation of relative matrix effects, the precision of calibration line slopes expressed as %RSD should not exceed 3
4% for the method to be considered free of relative matrix effects. For the evaluation of absolute matrix effects, two sets of standards should be prepared: (i) prepare Standard 1 by dissolving the appropriate concentration of analyte directly in the desorption solvent and (ii) prepare Standard 2 by dissolving the same concentration of analyte in a blank extract that was obtained after a blank sample of biological fluid has undergone the entire TFME procedure (post-extraction spike). The two sets of standards are subsequently analysed by LC-MS/MS. The comparison of signal intensities for the two standards (the ratio of the signals for Standard 1 and Standard 2) provides insight regarding the presence of absolute matrix effects (and the magnitude of the effect). Example results for such determination are shown in Table 7.9. Because SPME/TFME rely on a limited amount of the extraction phase [in contrast to solid-phase extraction (SPE)], co-extraction of unwanted interferences is less problematic, so significant matrix effects should not be observed for most SPME/TFME methods. However, it is still important to verify matrix suppression experimentally. For example, in one SPME-LC-MS method, a blank interference with a high signal intensity eluted at the same retention time as one of the analytes of interest and caused significant matrix suppression.7 This issue was resolved successfully by the redevelopment of the LC method in order to separate this blank interference from the analyte. G

13.1.3 Typical Automated SPME/TFME Performance The typical time required to perform an entire automated SPME/TFME procedure can range between 1
3 h (no evaporation/reconstitution step) and 3
6 h (with evaporation/reconstitution step) for 96 samples.5,9 It is possible to shorten these times further if sufficient analytical sensitivity is achieved, so that very short extraction times can be employed. The main limitations of automated SPME

Solid-Phase Microextraction Protocols

463

described in this protocol are: (i) it is applicable only to non-volatile analytes due to possible evaporative losses and (ii) it is necessary to employ analytical methods with sufficient sensitivity. The requirement of high instrumental analytical sensitivity is especially important for bioanalytical applications of SPME because the amount of analyte extracted by SPME can be very small depending on the degree of binding of analyte to biomolecules. SPME is an equilibrium technique where the amount of analyte extracted is proportional to free (unbound) analyte concentration and not total analyte concentration (see Chapter 11). This means that for highly bound drugs, where the free drug concentration in blood, serum or plasma is low and the amount extracted by SPME is also low, very sensitive analytical instrumentation such as LC-MS/MS may be required for the analysis.

13.2

Protocol for Automation of Ligand-Receptor Binding Studies Using Concept 96

The fundamental principles behind the use of SPME for studying ligand-receptor binding have been described in detail in Chapter 11. Very recently, this type of application was automated using Concept 96.10 The objective of this protocol is to provide a procedure to perform such automated ligand-receptor binding studies. The experimental design described herein uses the method of multiple standard solutions where ligand is spiked directly into the receptor solution11 because this type of design allows an increase in throughput by performing assay in parallel using the multi-well plate format and Concept 96. For enhanced throughput, this procedure describes how to perform single-point calibration, which was shown to be sufficient and yield equivalent results to multiple-point calibration.10 The 96well format allows the simultaneous preparation of up to 96 samples, while most binding studies use 8
12 experimental points to construct a binding curve. This means that multiple binding studies can be performed simultaneously. The main limitations of the procedure discussed here are: (i) only a small amount of ligand is extracted, which necessitates the use of sensitive analytical instrumentation for analysis; (ii) it is necessary to keep extraction conditions (pH, ionic strength and temperature) well controlled to obtain good method precision and (iii) it is applicable only to non-volatile analytes. In the future, the technique can be further miniaturised and adopted for use with microwell plates in order to reduce sample volume requirements. The overview of the main steps required to perform ligand-receptor binding study using TFME/SPME is shown in Figure 13.7. As demonstrated in the figure, the equilibration time for the ligand-receptor binding is determined experimentally.10,12 This is consistent with the procedures employed in equilibrium dialysis, which are considered to be the gold standard for binding studies. In contrast, some of the newer methods assume either short or long equilibration times, which may inadvertently introduce errors. Second, it is important to emphasise that it is not necessary to perform the extraction under conditions of negligible depletion

464

Handbook of Solid Phase Microextraction

Partial SPME method development

Equilibrium time determination in the presence and absence of receptor molecules

Automated SPME using one standard solution Fibre constant calibration

Binding study

Determine fc

Prepare standard solutions containing varying amounts of ligand and constant amount of receptor Automated SPME

Instrument response calibration

Determination of free concentrations and calibration

Determine m

Determine free ligand concentration using fc and m

Determination of binding parameters

Figure 13.7 Overview of the main steps to employ during ligand-receptor binding studies by SPME/TFME.1

(defined as experimental SPME conditions where the amount of ligand extracted by the fibre is negligible in comparison to the amount of ligand in the system under study) so long as appropriate calculations are used.11

13.2.1 Procedure for Ligand-Receptor Binding Studies Using Concept 961 Step 1: Perform partial SPME development to find a suitable (i) SPME/TFME coating, (ii) coating conditioning procedure, (iii) wash procedure and (iv) desorption conditions (see Section 13.1.1 for details of method development strategies). Step 2: Construct extraction time profiles for the ligand in the presence and absence of receptor molecules. These profiles can be obtained simultaneously by placing standard

Solid-Phase Microextraction Protocols

465

solutions 1 and 2 into appropriate wells. Prepare standard 1, which contains only ligand, and standard 2, which contains both ligand and receptor. For single-receptor studies, both standards 1 and 2 are prepared using the physiological buffer that will be employed in the binding study. For plasma protein binding studies or other biofluid binding studies, standard 2 is prepared by dissolving ligand in biofluid of interest. (i) Perform the automated SPME-LC-MS/MS procedure. Use the same extraction temperature and agitation speed and the optimised wash and desorption conditions for all experiments (as determined in Step 1). The extraction temperature may be set at 37 C to mimic the physiological conditions or any other appropriate temperature. Vary the extraction time in the SPME method. Recommended time points are 5, 15, 30, 45, 60 and 90 min if using thin coatings based on silica sorbents (C18 or C16 amide). Depending on the characteristics of analyte and coating, longer extraction times may be necessary. (ii) Plot the amount of ligand extracted versus extraction time. (iii) Determine the time required to reach equilibrium with and without the receptor present. Equilibrium time is defined as the time when no further increase in the amount extracted is observed (within experimental error). Select the extraction time for the binding study to be the longer of the two equilibrium times determined in this step. This ensures that both equilibria are established in solution. Step 3: Perform fibre-constant (fc) calibration for all fibres that will be used in the binding study. Omit this step for coatings with excellent inter-fibre reproducibility (,2% for example). (i) Prepare one calibration standard by spiking the appropriate concentration of stock analyte standard in physiological buffer [phosphate buffered saline (PBS), pH 7.4]. The biological buffer employed should not contain receptors. The concentration of analyte spiked is equivalent to the free (unbound) concentration (Cstd). (ii) Pipette the appropriate volume of the above calibration standard in appropriate wells of 96-well plate. (iii) Perform automated TFME-LC-MS/MS exactly as described in Section 13.1.1. Omit the addition of IS prior to the extraction. IS can be spiked directly into the desorption solvent and used to correct for any slight variations in the injection volume and instrumental response. Remember to set SPME extraction time equal to the time required to reach equilibrium as determined in Step 2 of this procedure. (iv) Determine the fc for each individual fibre used in the binding study. For single-point calibration, calculate fc using the following formula:

fc 5

n Cstd

ð13:1Þ

where n is the number of moles of ligand extracted from the standard solution (obtained in Step 3.3), Cstd is the molar concentration of the standard solution used for this determination (Step 3.1) and fc is the fibre constant, which represents the product between the fibre volume and Kes (equilibrium distribution constant between sample and extraction phase) for absorptive coatings, or the product of KA (adsorption equilibrium distribution constant between sample and extraction phase) and the surface area for solid coatings.11 Step 4: Perform a binding study using the desired number of ligand-receptor solutions (8
12 points is typically sufficient). (i) Prepare and dispense into a multi-well plate a set of sample solutions containing varying amounts of ligand and the same concentration of receptor.

466

Handbook of Solid Phase Microextraction

(ii) Perform automated TFME-LC-MS/MS exactly as described in Section 13.1.1. Omit the addition of IS prior to the extraction. Spike the IS directly into the desorption solvent in order to correct for any slight variations in the injection volume and instrumental response. Remember to set SPME extraction time equal to the time required to reach equilibrium as determined in Step 2 of this procedure. Step 5: Determine the absolute amount of ligand (m) extracted by each thin film using a calibration curve (weighted or non-weighted linear regression) obtained by the direct injection of calibration standards dissolved in desorption solvent in LC-MS/MS. Determine the free concentration (Cf) of ligand in each solution using the following equation:

Cf 5

m fc

ð13:2Þ

where m is the number of moles of ligand extracted from the sample solution and fc is the fibre constant determined in Step 3 for each individual fibre. Step 6: Calculate the total concentration of ligand after SPME using the formula

Ct 5

m0 2 m Vs

ð13:3Þ

where Ct is the total concentration of ligand after SPME, Vs is the sample volume, m0 is the number of moles of ligand spiked in Step 4(i), and m is the number of moles of ligand extracted from the sample solution and determined in Step 5. Step 7: Calculate the binding ratio (B), which is defined as the ratio of the amount of bound ligand (after SPME) to the amount of receptor present in the well using the equation

B5

Cb Creceptor

5

Ct 2 Cf Creceptor

ð13:4Þ

where Cb is the concentration of bound ligand, Ct is the total concentration of ligand after SPME (Step 6), Cf is the free concentration of ligand after SPME (Step 5) and Creceptor is the concentration of receptor used in the binding study (Step 4). Step 8: Plot B versus free ligand concentration. Fit the results to appropriate binding model and extract binding parameters using appropriate software.

13.2.2 Important Considerations and Troubleshooting Strategies G

G

G

G

Refer to Section 13.1.2 for general troubleshooting suggestions for automated TFME methods using Concept 96. Use this protocol for non-volatile ligands only. For volatile and semivolatile ligands, use closed vials with a limited, well-defined headspace instead of the Concept 96 system. For non-negligible SPME of highly bound ligands, choose a coating with a high affinity for the target ligand. Add IS to the desorption solvent prior to injection in order to compensate for any instrument variability in response and injection volume. Avoid or minimise the use of organic

Solid-Phase Microextraction Protocols

G

G

G

G

467

solvents when preparing ligand stock solution because the presence of organic solvent can affect partitioning process. For a single-point fibre constant calibration (Step 3), the concentration of the standard should be chosen to fall in the middle of the linear range of the instrument for the most reliable results. For binding studies in plasma or whole blood, keep the pH of biofluid at physiological level and constant for all samples. Plasma pH can change during storage and incubation, and this can be addressed by incubating with 10% CO2 gas or by dilution with appropriate buffers, as shown in Table 11.2.13,14 Therefore, pH modification such as 1:10 dilution with an isotonic PBS is recommended for use with the automated system described in this protocol because it is simpler to implement than CO2 incubation. If this protocol is used for the determination of overall ligand binding in a biofluid (e.g. plasma or whole blood), determine the extraction time profile in the appropriate biofluid because the viscosity of solution affects the time required to reach equilibrium. Always pipette the same volumes of sample and calibrant.

13.3

In Vivo SPME Protocol for Direct Monitoring of Circulating Intravenous Blood Concentrations

This protocol describes briefly the main steps of an in vivo SPME procedure that can be used to monitor intravenous concentrations of drugs and metabolites in beagles or other large animals without needing to withdraw a blood sample for analysis.15 For more detailed procedural information, the reader is referred to the original protocol by Lord et al.15 Chapter 12 of this book is dedicated to in vivo SPME and various applications of this technique. With minor modifications, this protocol can also be applied to in vivo sampling of fish tissue16 or for repeated sampling of rats using a sampling interface.17 The use of in vivo SPME for direct extraction of analytes from blood or tissue eliminates the need for blood withdrawal and offers numerous advantages, such as (i) reduced animal use; (ii) a reduced number of sample preparation steps, which can in turn eliminate inadvertent analyte losses during sample processing; and (iii) a non-destructive and minimal disturbance to living systems.16,18
21 The main steps of method development for in vivo SPME are described in Figure 13.8. The general method development discussion presented in Chapter 7 is also applicable, with a few main differences: (i) the coating used should be biocompatible in order to prevent adverse/toxic reactions in the animal and to prevent adhesion of biomolecules to the surface, which can affect the uptake of analyte (Section 3.4.3 discusses biocompatible coatings); (ii) inter-fibre reproducibility becomes a critical parameter (and should be evaluated during method development experiments) to achieve good method accuracy and precision because single individual SPME fibres are employed for sampling of each time point and animal; (iii) matrix modification experiments can be omitted because it is not feasible to apply these approaches in vivo and (iv) careful attention should be paid during calibration. Figure 13.9 shows the two calibration methods most commonly

468

Handbook of Solid Phase Microextraction

Type of coating

Choose biocompatible coating with good inter-fibre reproducibility

Extraction time

Determine time required to reach equilibrium using either (i) static conditions or (ii) flow-through system using expected in vivo blood flow rate

Calibration strategy Desorption solvent composition Desorption solvent volume Desorption time

Rinse procedure

Equilibrium external calibration or kinetic calibration Select desorption strategy: manual interface or offline Use desorption solvent compatible with LC method Verify chromatographic peak shape Use sufficient solvent volume to immerse the entire coating (offline desorption) Verify carry-over Keep very short (<1 min) Perform immediately after extraction Verify that no loss of analyte occurs

Conditioning procedure

Do not let coating dry out between preconditioning and in vivo SPME Perform under sterile conditions Combine with standard loading if using kinetic calibration

Sterilisation

Select sterilization procedure Verify effect of sterilization on extraction efficiency of coating

Probe storage

Perform stability testing to determine how long probes can be stored after in vivo sampling

LOD, LOQ and linear range

Verify that overall method sensitivity is sufficient for selected application

(optional) Evaporation time (optional) Reconstitution solvent

Applicable only if analyte is non-volatile Add evaporation/reconstitution step to increase analytical sensitivity Use reconstitution solvent compatible with LC method Verify chromatographic peak shape Verify reconstitution efficiency

Figure 13.8 Method development strategies for in vivo SPME.

employed in conjunction with in vivo SPME. Refer to Chapter 6 for a detailed, theoretical discussion (including the derivation of the equation shown in Figure 13.9), while applications of these methods to in vivo studies are described elsewhere.17
25

Solid-Phase Microextraction Protocols

469

Equilibrium

Kinetic

1. 1.

2.

3. 4. 5.

6.

Prepare set of calibration standards which contain NO binding matrix (FREE concentration) and/or set of calibration standards in binding matrix (TOTAL CONCENTRATION). Conduct in vitro SPME sampling in the above standard solutions at EQUILIBRIUM using extraction conditions which mimic closely in vivo system under investigation (temperature, pH, etc.). Inject in LC–MS/MS. Plot standard concentration versus area ratio (analyte/IS). Perform linear regression (non-weighted or weighted). The slope of the line is equal to the product of Kfs and Vf. Use calibration curve from Step 5 to determine C0.

2. 3. 4. 5.

6.

7.

Prepare loading standard (deuterated IS) using sterile methanol/water or 100% water. Load the fibres using direct immersion for preset amount of time. Perform entire in vivo SPME procedure as described in Section 13.3.1. Inject in LC–MS/MS and determine Q and n simultaneously. OPTION 1: Determine q 0 by injecting standards from Step 2. q0 is the difference between initial amount of standard (Step 1) and the amount of standard remaining in the solution after loading. OPTION 2: q0 can be determined by loading an excess of fibers in Step 2. Desorb these excess fibres and analyse (no in vivo SPME performed) to determine q0. Determine product of Kfs and Vf by performing in vitro external calibration at EQUILIBRIUM from biofluid (TOTAL concentration) or from buffer (FREE concentration) or both. nq0 1 Determine C0. C0 =

q0 − Q k fsV f

Figure 13.9 Common calibration procedures for in vivo SPME.

13.3.1 In Vivo SPME Sampling Procedure Using In-dwelling Catheter15 Step 1: Obtain in vivo SPME probes from Supelco (see Figure 3.17) or prepare appropriate coatings and devices according to the procedures described in literature.15,18
21,24,26 Step 2: Sterilise in vivo probes by autoclaving (121 C, 13.5 psi for 20 min in PBS on a liquid cycle) or another appropriate sterilisation procedure.15 Step 3: Precondition the probes using preconditioning solvent prepared under sterile conditions (using sterile water and filtered using a 0.25-µm sterile syringe filter), if a preconditioning step is required for optimal performance of the coating. If using kinetic calibration, load the coating with deuterated IS at the same time (Figure 13.9). Step 4: Perform in vivo SPME sampling as outlined in Figure 13.10B. (i) Prepare the sampling site by clipping hair and using appropriate topical anaesthetic. (ii) Place a standard medical catheter in the target vein (performed by a veterinary technician or biologist). Select a vein that is sufficiently long and straight to accommodate the catheter plus the SPME probe without having to negotiate curves in the vein (e.g. cephalic). Attach a rubber intravenous injection bulb (PRN) to the end of the catheter in order to enable SPME sampling. (iii) Pierce PRN adapter with sterile hypodermic needle and insert sterile in vivo SPME probes through the catheter. (iv) Expose the coating to the flowing blood for a predetermined amount of time. Start the timer immediately after exposure. (v) Remove the probe once the extraction time has elapsed. First, retract the coating, and then remove the hypodermic needle.

470

Handbook of Solid Phase Microextraction

(A)

(B) Sterilisation of assembly Prepare sampling site

Preconditioning and standard loading

Place in-dwelling catheter in vein

In vivo sampling

Insert SPME probes using hypodermic needle and PRN adapter

Expose coating to flowing blood

Wash (30 s)

Start the timer

Desorption Remove the probes after elapsed sampling time

(optional) Evaporation/reconstitution Rinse the probes for 10–30 s with purified water

Flush catheter with saline or heparinized saline solution

LC–MS/MS

Figure 13.10 (A) Overview of complete procedure for in vivo SPME sampling. (B) Main steps of in vivo SPME sampling (direct extraction).

(vi) Rinse the probe immediately and store in a clean, labelled tube or vial. Flush the catheter with saline or a heparinised saline solution to prevent any clotting. (vii) Once all the sampling points have been completed, remove the catheter and clean and bandage the site. Step 5: Desorb the probes using either (i) the manual interface or (ii) offline desorption, as determined during method development. See Chapter 7 for a description of typical desorption conditions for both methods and the advantages and disadvantages of the two approaches. Step 6: Inject standards and samples using the appropriate LC-MS/MS method using the appropriate standard operating procedure. Integrate all analyte and IS peaks using instrument software. Calculate area ratios (peak area of analyte/peak area of IS). Check the chromatograms for blank injections (blank desorption solvent, blank fibre, and so on). Perform LC-MS/MS analysis and perform appropriate calibration. Two common calibration procedures are described in Figure 13.9.

Solid-Phase Microextraction Protocols

471

13.3.2 Important Considerations and Troubleshooting Strategies SPME recommendations G Supelco commercial probes can require equilibration times greater than 60 min because of the coating thickness (45 µm).7 For most in vivo studies (especially pharmacokinetic studies), much shorter sampling times are needed ( , 5 min). Therefore, to use these probes with such short sampling times, it is necessary to perform kinetic calibration, as described in Figure 13.9. G Do not reuse commercial Supelco probes in vivo. G Ensure extraction timing is accurate and reproducible if using pre-equilibrium SPME with kinetic calibration. G Use fibres from the same batch to complete all sampling as inter-fiber reproducibility between batches may not be satisfactory. Calibration recommendations Match the biological fluid used to prepare calibration standards for matrix-matched (equilibrium external) calibration as closely as possible to exact in vivo conditions. G Use the same anticoagulant (and the same amount of anticoagulant) for sampling and for calibration. Use EDTA or citric acid as anticoagulants instead of heparin because of better stability. Ensure analyte of interest is chromatographically resolved from the anticoagulant peak to avoid problems with ionization suppression.15 G Store plasma frozen until use. Store whole blood refrigerated and use it within a month. G Do not use plasma instead of whole blood for the equilibrium external calibration method because they do not provide equivalent results depending on the exact analyte partitioning ratio between whole blood and plasma, as discussed in detail elsewhere.15 G For in vivo SPME, the deionised water for preparing the loading solution should be autoclaved, the fibres should be sterilised and the loading should be conducted in sterile facilities. G Verify that the signal for samples and standards does not exceed the linear dynamic range of the instrument. If the intensity of some signals exceeds this threshold, dilute the samples/standards and re-analyse. For intravenous dosing, this is most likely to occur during the early part of a pharmacokinetic study. G Avoid excessive agitation or other conditions that can cause haemolysis during preparation of the standard solutions. G If deuterated IS is not available for use with kinetic calibration, an analogue can be used in some cases, but the equivalence of time constant a for analyte and proposed IS should be established experimentally. Alternatively, new calibration methods (diffusion-based) and dominant desorption can be used because these do not require availability of an IS (see Chapter 6). G

In vivo sampling recommendations Use sterile instruments, reagents and techniques for all steps during animal sampling. G Good placement and immobilisation of the catheter is critical for the success of the study and should be performed by a veterinary technician or biologist. G Do not dose the drug via the catheter used for in vivo sampling in order to avoid accidental contamination and erroneous results. G Ensure no clotting occurs in the catheter at any point during sampling. This is facilitated by flushing saline or heparinised saline through the catheter immediately after inserting G

472

G

Handbook of Solid Phase Microextraction

the probes, immediately after removing the probes, and at regular intervals (every 30
60 min) if sampling points are spaced very far apart. Rinse probes immediately after extraction. If SPME probe after in vivo sampling and rinse step is red in appearance, the probe is defective. Discard this probe and repeat the sampling with a fresh probe.

13.4

Protocol for Setting up Automated SPME-GC Methods

A recent protocol described by Risticevic et al. is universally applicable for setting up a variety of automated SPME-GC methods using commercially available autosamplers.27 This protocol can be used for fast and automated SPME method development, as well as to carry out automated and high-throughput SPME-GC analyses. Example applications include both targeted and global analyses in fields ranging from environmental analysis, food chemistry, bioanalysis, industrial hygiene and metabolomics.

13.4.1 Development of Automated SPME Methods with the CombiPALt Autosampler 27 The CombiPAL autosampler provides the standard SPME procedures in the Cycle Composer, which allows the user to modify the parameters according to the application of interest. Alternatively, the Macro Editor software can be used to develop custom, sophisticated programs, as described in the procedure below.27 Step 1: On the PAL Cycle Composer screen, select the Macro Editor tab. Select atoms for the macro atom sequence from the pull-down menu entitled ‘Selected atom’. Table 13.1 describes the available atoms and their function. Step 2: Insert the atoms in the macro atom sequence. Step 3: Specify the parameters for each atom. Consult Table 13.1 for a description of parameters requiring user input. Once all desired atoms have been inserted and customised, name the new custom macro under ‘Macro name’. Step 4: Switch to the Method Editor tab. Insert the created macro into the method macro sequence. Name the method under ‘Method name’. Step 5: Switch to the Sample List screen. Specify the method name, tray type and vial number in the sample list. Step 6: Name the created sample list and start it.

13.4.2 Recommended Procedure for Automated Loading of IS Using the CombiPAL Autosampler 27 The flexibility and versatility of the CombiPAL autosampler allows the automation of the in-fibre standardisation technique initially proposed by Chen et al. 22,28 and

Solid-Phase Microextraction Protocols

473

Table 13.1 Summary of Atoms and Their Functions Name of Atom in Macro Atom Sequence

Function of Atom

User Step

SET TEMP

Temperature control in the agitator Transport of the vial from tray to agitator and from agitator to tray Activation/deactivation of agitation processes; agitation speed control

Input the sample temperature

TRANSP VIAL

SET AGI

WAIT TIMER

Incubation and extraction timing procedures

MOVE REL

Movement of the needle/fibre inside the vial to ensure the highest linear velocity around the fibre coating during the extraction Waiting for instrument-ready signal for injection/desorption Desorption and fibre cleaning timing procedures

WAIT SYNC SIGNAL WAIT

DISPENSE SYR

ASPIRATE SYR

Specify tray positions of sample vials Input agitation speed Input agitation on and off times Input times for timed incubation and extraction SPME processes Specify the distance the needle moves from the centre of the vial and back to the centre of the vial Minimise time between extraction and analysis Input times for timed desorption and fibre cleaning SPME processes

Fibre coating exposure in the vial, injection port and fibre conditioning station Fibre coating retraction inside the needle after extraction, desorption and fibre cleaning procedures

Table adapted from Ref. [27].

subsequently used in various studies29,30 (see Chapter 6 for a theoretical description of this technique). The loading of IS from vacuum pump oil allows the use of a single standard solution for more than 100 loadings, which makes it ideally suited for use in automated procedures such as the one described below.28 Step 1: Add 4.0 g of vacuum pump oil into a 20-mL vial. Step 2: Keep the vial in a heat block at 120 C with a flow of nitrogen at 5
10 mL/min over the vacuum pump oil for at least 12 h.

474

Handbook of Solid Phase Microextraction

Step 3: Cool the vacuum pump oil to 45 C in chamber #6 of an agitator of the CombiPAL autosampler (operated via the PAL Cycle Composer with the Macro Editor software). Step 4: Add the appropriate amount of IS dissolved in an organic solvent (such as methanol) into the vacuum pump oil and cap the vial. A high concentration of IS should be used so that many samples can be processed using a single vial without causing IS depletion. Spike the standard without piercing the vial in order to prevent any analyte losses due to a pierced septum.28 Therefore, perform the spiking below the level of solution (as recommended for volatile compounds) prior to vial capping. Step 5: Agitate the vacuum pump oil at a rate of 500 rpm for a minimum of 4 h prior to use. Step 6: Install appropriate SPME fibre on the autosampler. Step 7: Write the method in the Cycle Composer software of the CombiPAL autosampler, which consists of the following steps according to the procedure described in Section 13.4.1: (i) Transfer the sample vial from the tray to chamber #1 of the agitator to start incubation. (ii) Clean the fibre coating in the fibre conditioning station using the appropriate conditions. (iii) Move the fibre into the IS vial in chamber #6 of the agitator. (iv) Expose the coating for a predetermined amount of time (such as 30 s) to the vial headspace in order to load the IS. Use agitation during loading. (v) Withdraw the fibre coating inside the needle. (vi) Move the fibre into the sample vial in chamber #1 of the agitator. (vii) Expose the fibre to the headspace above the sample for the preset amount of time using the appropriate agitation speed. (viii) Withdraw the fibre inside the needle. (ix) Move the fibre into a GC injector. (x) Expose the fibre in the GC injector for the predetermined desorption time using the appropriate desorption temperature. (xi) Perform post-injection cleaning of the fibre in the fibre conditioning station using the appropriate conditions.

13.4.3 Development of Automated SPME Methods with the MPS 2, TriPlus and Concept Autosamplers27 The development of automated methods involving a fibre-SPME device with the MPS 2, TriPlus and Concept autosamplers is based on the user input of the appropriate SPME parameters in the dedicated SPME software. Figure 13.11 shows an example of the screen used for the input of SPME parameters for the Gerstel MPS 2 autosampler. Table 13.2 summarises the features of the common commercial SPME autosamplers. An additional parameter corresponding to the speed of needle penetration is available in the TriPlus autosampler. The appropriate needle speed (1
250 mm/s) should be selected on the basis of septum type: (i) use slow speed for the penetration of hard septa and (ii) use a faster speed for the penetration of elastic septa.

Solid-Phase Microextraction Protocols

475

Figure 13.11 Example of SPME screen for the input of SPME parameters for the Gerstel MPS 2 autosampler.

13.4.4 Important Considerations and Troubleshooting Strategies SPME performance G Consult Section 7.3.3 for an extensive list of the likely causes of unacceptable precision and suggestions on how to resolve these issues. G Always follow the manufacturer’s recommendations to condition new SPME fibres prior to their first use. It is also recommended to keep the fibres at the desorption temperature (in the auxiliary injection port or in the fibre conditioning station of the autosampler) for an additional 30 min prior to each sample sequence to eliminate high background in the chromatogram.27 G Some fibre coatings do not exhibit stable and reproducible performance at the very beginning. This is particularly true for metal assemblies. To address this issue, condition new fibre coatings prior to first use by preparing 5
10 test samples and do not use those results. Subsequent fibre performance should be reproducible and stable for the lifetime of the fibre. G It is recommended to run periodic control samples (e.g. every 15 injections or so) to monitor for any potential retention time shifts and/or the fibre’s extraction performance. If a particular application requires the use of retention indices to aid in analyte identification, then the loading of retention index standards can be performed27,29 as described in Section 13.4.3, with a few modifications in Steps 1
5, namely (i) the retention index standard can be prepared by spiking C8
C20 and C21
C40 retention index standards (dissolved in hexane) into 1 mL of deionised water placed in a 10-mL vial and (ii) after placing the retention index standard vial in chamber #6, it is sufficient to incubate the solution for about 2 min at 45 C and 500 rpm. Then Steps 6 and 7 can be completed on a control sample in order to spike it with the retention index standard.

476

Table 13.2 Summary of Features of Commercially Available SPME
GC Autosamplers CombiPAL

MPS 2

TriPlus

Concept

Extraction temperature range Type of agitation mechanism Extraction agitation speed range Agitator tray during extraction process Temperature range of fibre conditioning station Ability of autosampler to start the preparation of next sample while previous sample is analysed Specification of needle/ penetration depth

30
200 C

40
150 C

Orbital and magnetic stirring 250
750 rpm

30
200 C 10
200 C (Peltier version) Orbital and magnetic stirring 250
750 rpm

Unadjustable

Room temperature
100 C Orbital and magnetic stirring 100
1,000 rpm

Open

Open

Closed

Open

30
350 C

30
350 C

40
350 C

Unavailable

Yes

Yes

Yes

Yes

Include fibre coating

Include fibre coating

Depth of the needle only

Either needle or coating

Source: Adapted from Ref. [27].

Rocking

Handbook of Solid Phase Microextraction

Parameter

Solid-Phase Microextraction Protocols

477

Autosampler performance G If a sequence list stops prior to completion, inspect the autosampler connection and reestablish if necessary. G When using septumless sealing systems, such as high-pressure Merlin Microsealt and the septumless head (SLH) on the Gerstel Programmed Temperature Vaporizer (PTV), monitor the first injection to verify for potential pressure problems during the injection.27 Such problems could be caused by undertightening or overtightening of the injection nut of the septumless system. If the problem persists, it can be resolved by installing a new septum. The typical lifetime of the Merlin Microseal can range from 1,000 to over 10,000 injections, depending on the types of samples and analysis conditions. G If pressure or leak problems are encountered with SLH on the Gerstel PTV, replace the needle guide. It is also important to note that the same type of needle guide can be used for both liquid and SPME injections.27 However, if the needle guide was first used for SPME prior to its use for liquid injection, the enlargement of the needle guide can occur because of the wider 23- or 24-gauge SPME needle, as compared to the traditional 26gauge needle for standard liquid injection. To address this problem, tighten the nut of the SLH as much as possible to avoid the pressure problems during the liquid injection runs. Alternatively, use the standard 23-gauge syringes suitable for the MPS 2 autosampler for all injection types. G When using metal fibres, a slight misalignment between the needle position and the hole in the injection nut can cause the needle to catch and bend. This problem can be eliminated by performing frequent alignment checks and regular autosampler maintenance. This maintenance should include the change of the driving belt in the autosampler at least once a year to ensure accurate positioning of the needle.

References 1. 2. 3. 4.

D Vuckovic, E Cudjoe, FM Musteata & J Pawliszyn, Nat Protoc 5 (2010) 140 E Cudjoe, D Vuckovic, D Hein & J Pawliszyn, Anal Chem 81 (2009) 4226 S Risticevic, H Lord, T Go´recki, CL Arthur & J Pawliszyn, Nat Protoc 5 (2010) 122 Food and Drug Administration, Guidance for Industry, Bioanalytical Method Validation (2001) http://www.fda.gov/downloads/Drugs/GuidanceComplianceRegulatoryInformation/Guidances/ucm070107.pdf,

date accessed October 24, 2011 5. R Vatinno, D Vuckovic, CG Zambonin & J Pawliszyn, J Chromatogr A 1201 (2008) 215 6. J O’Reilly, Q Wang, L Setkova, JP Hutchinson, Y Chen, HL Lord, CN Linton & J Pawliszyn, J Sep Sci 28 (2005) 2010 7. D Vuckovic, B Shirey, Y Chen, L Sidisky, C Aurand, K Stenerson & J Pawliszyn, Anal Chim Acta 638 (2009) 175 8. RN Xu, L Fan, MJ Rieser & TA El-Shourbagy, J Pharm Biomed Anal 44 (2007) 342 9. D Vuckovic, E Cudjoe, D Hein & J Pawliszyn, Anal Chem 80 (2008) 6870 10. D Vuckovic & J Pawliszyn, J Pharm Biomed Anal (2008) 50 (2009) 550 11. FM Musteata & J Pawliszyn, J Proteome Res 4 (2005) 789 12. FS Mirnaghi, Y Chen, LM Sidisky & J Pawliszyn, Anal Chem 83 (2011) 6018 13. FM Musteata, J Pawliszyn, MG Qian, J Wu & GT Miwa, J Pharm Sci 95 (2006) 1712 14. H Wan & M Rehngren, J Chromatogr A 1102 (2006) 125 15. HL Lord, X Zhang, FM Musteata, D Vuckovic & J Pawliszyn, Nat Protoc 6 (2011) 896

478

Handbook of Solid Phase Microextraction

16. SN Zhou, KD Oakes, MR Servos & J Pawliszyn, Env Sci Tech 42 (2008) 6073 17. FM Musteata, I de Lannoy, B Gien & J Pawliszyn, J Pharm Biomed Anal 47 (2008) 907 18. HL Lord, RP Grant, M Walles, B Incledon, B Fahie & JB Pawliszyn, Anal Chem 75 (2003) 5103 19. FM Musteata, ML Musteata & J Pawliszyn, Clin Chem 52 (2006) 708 20. X Zhang, A Es-Haghi, FM Musteata, G Ouyang & J Pawliszyn, Anal Chem 79 (2007) 4507 21. A Es-haghi, X Zhang, FM Musteata, H Bagheri & J Pawliszyn, Analyst 132 (2007) 672 22. Y Chen, J O’Reilly, Y Wang & J Pawliszyn, Analyst 129 (2004) 702 23. Y Chen & J Pawliszyn, Anal Chem 76 (2004) 5807 24. FM Musteata & J Pawliszyn, J Biochem Biophys Methods 70 (2007) 181 25. FM Musteata & J Pawliszyn, TrAC Trends Anal Chem 26 (2007) 36 26. ML Musteata, FM Musteata & J Pawliszyn, Anal Chem 79 (2007) 6903 27. S Risticevic, Y Chen, L Kudlejova, R Vatinno, B Baltensperger, JR Stuff, D Hein & J Pawliszyn, Nat Protoc 5 (2010) 162 28. Y Wang, J O’Reilly, Y Chen & J Pawliszyn, J Chromatogr A 1072 (2005) 13 29. L Setkova, S Risticevic & J Pawliszyn, J Chromatogr A 1147 (2007) 213 30. SN Zhou, X Zhang, G Ouyang, A Es-Haghi & J Pawliszyn, Anal Chem 79 (2007) 1221