Journal of Chromatography A, 1000 (2003) 3–27www.elsevier.com/ locate/chroma

Review

B efore the injection—modern methods of sample preparation forseparation techniques

*Roger M. SmithDepartment of Chemistry, Loughborough University, Loughborough, Leics, LE11 3TU, UK

Abstract

The importance of sample preparation methods as the first stage in an analytical procedure is emphasised and examined.Examples are given of the extraction and concentration of analytes from solid, liquid and gas phase matrices, includingsolvent phase extractions, such as supercritical fluids and superheated water extraction, solid-phase extraction andsolid-phase microextraction, headspace analysis and vapour trapping. The potential role of selective extraction methods,including molecular imprinted phases and affinity columns, are considered. For problem samples alternative approaches,such as derivatisation are discussed, and potential new approaches minimising sample preparation are noted. 2003 Elsevier Science B.V. All rights reserved.

Keywords: Reviews; Sample preparation; Solvent extraction; Supercritical fluid extraction; Solid-phase extraction; Solid-phase microextraction; Molecular imprinting

Contents

1 . Introduction ............................................................................................................................................................................ 42 . The first theoretical plate? ........................................................................................................................................................ 53 . Problems with the old methods................................................................................................................................................. 5

3 .1. Sample preparation 100 years ago .................................................................................................................................... 53 .2. Sample preparation in early volumes of theJournal of Chromatography ................................................................... 6

4 . Filtration................................................................................................................................................................................. 65 . Extraction methods.................................................................................................................................................................. 6

5 .1. Unification ..................................................................................................................................................................... 66 . Analytes in solid samples......................................................................................................................................................... 7

6 .1. Enhanced solvent extraction methods ............................................................................................................................... 86 .1.1. Pressurised liquid extraction................................................................................................................................. 86 .1.2. Microwave and sonic wave assisted extraction ...................................................................................................... 86 .1.3. Supercritical fluid extraction ................................................................................................................................ 96 .1.4. Superheated water extraction................................................................................................................................ 10

6 .2. Problems with solid matrices ........................................................................................................................................... 106 .2.1. Biological matrices and matrix solid-phase dispersion............................................................................................ 106 .2.2. Insoluble solid matrices—pyrolysis ...................................................................................................................... 11

*Tel.: 144-1509-222-563; fax:144-1509-223-925.E-mail address: r.m.smith@lboro.ac.uk(R.M. Smith).

0021-9673/03/$ – see front matter 2003 Elsevier Science B.V. All rights reserved.doi:10.1016/S0021-9673(03)00511-9

4 R.M. Smith / J. Chromatogr. A 1000 (2003) 3–27

6 .2.3. Thermal desorption from solids ............................................................................................................................ 117 . Analytes in solution................................................................................................................................................................. 11

7 .1. Trapping the analytes ...................................................................................................................................................... 127 .1.1. Solid-phase extraction ......................................................................................................................................... 127 .1.2. Solid-phase microextraction ................................................................................................................................. 137 .1.3. Stir-bar extractions .............................................................................................................................................. 14

7 .2. Extraction of the analytes into a liquid phase..................................................................................................................... 167 .2.1. Membrane extraction........................................................................................................................................... 167 .2.2. Single drop extraction.......................................................................................................................................... 177 .2.3. Purge and trap..................................................................................................................................................... 17

8 . Analytes in the gas phase ......................................................................................................................................................... 188 .1. Trapping analytes from vapour samples ............................................................................................................................ 188 .2. Headspace analysis ......................................................................................................................................................... 18

9 . Direct combination of sample preparation and separation ........................................................................................................... 199 .1. Large volume injections in GC......................................................................................................................................... 199 .2. Coupled column systems LC–LC or GC–GC ................................................................................................................... 209 .3. Isotachophoresis in capillary electrophoresis ..................................................................................................................... 21

1 0. Selectivity enhancement......................................................................................................................................................... 211 0.1. Affinity methods ........................................................................................................................................................... 211 0.2. Molecular imprinting polymers ...................................................................................................................................... 221 0.3. Restricted-access media ................................................................................................................................................. 22

1 1. When separation alone is not enough—derivatisation to see the sample ..................................................................................... 231 1.1. Derivation to enhance volatilisation and separation .......................................................................................................... 231 1.2. Derivatisation to enhance thermal stability ...................................................................................................................... 231 1.3. Derivatisation to enhance detection................................................................................................................................. 23

1 2. Can sample preparation be avoided? ....................................................................................................................................... 241 3. Conclusions .......................................................................................................................................................................... 24References .................................................................................................................................................................................. 24

1 . Introduction considerable constraint on the throughput of anymethod and involve a significant additional workload

These days, when separation methods can provide for staff. A survey in 1991 claimed that samplehigh resolution of complex mixtures of almost every preparation can account for around two thirds (61%)matrix, from gases to biological macromolecules, of the effort of the typical analytical chemist andand detection limits down to femtograms or below, 92% of the respondents regarded sample preparationthe whole advanced analytical process still can be as moderately or very important[1]. However, awasted if an unsuitable sample preparation method more recent comment was that ‘‘ . . . in analyticalhas been employed before the sample reaches the chemistry laboratories, sample preparation is notchromatograph. Rather like the proverbial computer recognised as an important step in the whole ana-rule, garbage-in garbage-out (GIGO), poor sample lytical scheme and is often given to the less trainedtreatment or a badly prepared extract will invalidate chemist’’[2]. Although individual steps or samplethe whole assay and even the most powerful sepa- preparation methods have been reviewed in detail,ration method will not give a valid result. there are few general monographs or reviews on the

Yet sample preparation is often a neglected area, subject[3–6], probably also emphasising the broadwhich over the years has received much less atten- nature of the topic and the wide range of approachestion and research than the chromatographic sepa- that can be used.ration or detection stages. However, getting the The basic concept of a sample preparation methodsample preparation stages correct can be economical- is to convert a real matrix into a sample in a formatly valuable as well as analytically important. An that is suitable for analysis by a separation or otherinefficient or incomplete technique can represent a analytical technique. This can be achieved by em-

R.M. Smith / J. Chromatogr. A 1000 (2003) 3–27 5

ploying a wide range of techniques, many of which molecular mass compounds that have been de-have changed little over the last 100 years. They veloped over the century since chromatography washave a common list of aims: first reported, with selected recent samples. Within• The removal of potential interferents (for either the scope of a review, the coverage is necessarily

the separation or detection stages) from the representative rather than comprehensive, as effec-sample, thereby increasing the selectivity of the tively almost every assay of a real sample requiresmethod. some sample preparation and the potential examples

• To increase the concentration of the analyte and are endless. Frequently references are therefore givenhence the sensitivity of the assay. to more specialised reviews or monographs.

• If needed, to convert the analyte into a moresuitable form for detection or separation.

• To provide a robust and reproducible method that 2 . The first theoretical plate?is independent of variations in the sample matrix.With increasing demands on the analytical chemist There are close analogies between many sample

to provide accurate and valid analytical measure- preparation methods and analytical separations andments for regulatory requirements, poor manual frequently the sample preparation step can be consid-reproducibility during the sample preparation stage ered to be the first theoretical plate in the separationcan be a major cause of assay variability[7], hence a process. However, it is one with often relatively lowneed for automation and reduced manual sample discrimination but high capacity. It is still based, ashandling. However, robots and the automation of the are chromatographic separations, on a phase dis-laboratory bring their own problem of longer method tribution, charge interaction and/or size fractiona-development times and new skill requirements. tion. Frequently an inherent increase in the con-

Many of these ideas could apply to any analytical centration of the analyte can also be achievedprocess but we will concentrate here on preparations through a chromatographic focusing effect. The skillleading to assays by separation methods. In some of the analytical chemist has been in devising sampleways, this simplifies the requirements of the sample preparation methods to achieve the desired distribu-preparation process, as the final assay step often tion by manipulating the polarity or ionic state of thealready incorporates a powerful separation and dis- analyte, or by the appropriate selection of the phases.crimination technique.

Although many traditional sample preparationmethods are still in use the trends in recent years 3 . Problems with the old methodshave been towards:• The ability to use smaller initial sample sizes In looking at current sample preparation methods,

even for trace analyses. it is interesting to compare them with the methods• Greater specificity or greater selectivity in ex- used in the early days of chromatography and from

traction. the early volumes of theJournal of Chromatog-• Increased potential for automation or for on-line raphy.

methods reducing manual operations.• A more environmentally friendly approach (green 3 .1. Sample preparation 100 years ago

chemistry) with less waste and the use of smallvolumes or no organic solvents. In many ways, the extraction of natural productsThese goals are being achieved in a number of has changed little. In his original work Tswett[8]

different ways and are still the subject of active utilised a number of alternative solvent extractionsresearch and this has been recognised in the recent with alcohol–light petroleum, benzene, carbon tetra-addition of a new topic heading in theJournal of chloride or carbon disulfide to obtain the chlorophyllChromatography A on Sample Preparation. This pigments from plant material, after neutralisation ofreview surveys the wide range of sample preparation the leaves with MgO and CaCO . The need to obtain3

methods and combinations of methods for low a sample solution free of alcohol and water was

6 R.M. Smith / J. Chromatogr. A 1000 (2003) 3–27

recognised, as the presence of these solvents in the sample preparation, particularly in liquid chromatog-extract gave indistinct chromatograms. Thus from raphy (LC), where any insoluble material will blockthe earliest days of chromatography, the influence of the column or frits. The efficiency of filtration isthe sample preparation methods on the quality of the determined by the porosity of the filter and would beresulting chromatogram was identified, as was the typically 2mm or less for LC. Different types ofpotential of poor practice to destroy the advantages filters can be used include paper, glass fibre, andof the analytical technique. membrane filters[12,13]. In a recent development,

filters have been built into standard sized sample3 .2. Sample preparation in early volumes of the vials so that sample handling and solution transfer isJournal of Chromatography minimised, which can be important to avoid con-

tamination of the sample and reduce biohazards toThe coverage of volume 1 was very different from the operator[14].

that found today and sample preparation in 1958 had For some samples, such as environmental solu-advanced little from the methods utilised by Tswett. tions, the removal of relatively large solid materialIn the first few volumes of the journal, there were may be required as this may physically interfere withalmost no papers on gas or column liquid chromatog- extractions or later stages and an initial simpleraphy, the principal techniques being ion-exchange filtration will suffice. However, care must be takenseparations, electrophoresis, and paper chromatog- that there are no sample losses because of adsorptionraphy, with the most frequently examined analytes of analytes onto the solid material that is removed.being radiochemicals and inorganic samples. Alternatively centrifugation can be used to remove

By volume 500, in 1990, still relatively few papers insoluble material from solutions.referred specifically to sample preparation but it wasnoticeable that gas chromatography was now thedominant technique. However, a review of carbohy- 5 . Extraction methodsdrate analysis discussed recent derivatisation ad-vances[9] and another paper considered derivatisa- The oldest and most basic sample preparationtion for electron-capture detection with electrophoric method is extraction, in which the analyst aims toderivatives [10]. A fully automated method for separate the analyte of interest from a sample matrixnitrofuran in biological samples, using on-line using a solvent, with an optimum yield and selectivi-dialysis and column switching, showed a more ty, so that as few potential interfering species asmodern trend[11], although in a recent survey one possible are carried through to the analytical sepa-third of the respondents suggested that automation ration stage. Different extraction methods are used,was unnecessary in their laboratory mainly because including solvent extraction from solids and liquid–of a low sample load[6]. Interestingly the preface liquid extraction from solutions[15]. The solventsworried that advance in electronics would not permit may be organic liquids, supercritical fluids andthe journal to reach volume 1000. superheated liquids or the extraction liquid may be

In more recent years the importance of sample bonded to a support material, as in solid-phasepreparation has been reflected by special issues extractions (SPEs). Selectivity can be obtained byreporting related symposia and topics. These include altering the extraction temperature and pressure, bysolid-phase extraction (Vol. 885), preconcentration the choice of extraction solvent or liquid, and the useand sample enrichment techniques (Vol. 902), Ex- of pH and additives, such as ion-pair reagents.Tech 2001 (Vol. 963), and sample handling (Vol.975). Similar influences are reflected in other sepa- 5 .1. Unificationration science journals.

All extraction methods make use of the same basicset of concepts to concentrate the analyte selectively

4 . Filtration in one phase. Any analyte will be distributed be-tween two phases according to the distribution

Simple filtration can be an important part of constant, temperature, and the relative volumes of

R.M. Smith / J. Chromatogr. A 1000 (2003) 3–27 7

the phases. However, the extraction rates are based analyte being released from a matrix[17]. The initialon the migration kinetics and hence are governed by mild conditions, while selective, were simply nottemperature and the diffusion rates in the two phases. sufficiently strong to release the analyte from all theThese parameters are essentially those that are active matrix sites and give a quantitative yield.manipulated in chromatographic separations, and one These problems emphasise the need for extractioncan therefore consider the extractions as a form of methods to be tested with a range of real samples ofpre-assay chromatography. different types, not just with model systems (and in

In many of these methods, a balance must often be particular not just with spiked samples). Realisticobtained between the complete extraction of all the robustness studies should be undertaken before thesoluble organic components and the selective ex- extraction is used in an analytical method. If possibletraction of only the compounds of interest. This alternative independent extraction methods should beconflict has been a constant theme throughout sample used as a guide or the methods should be applied topreparation methods in analytical chemistry. Exhaus- samples of known composition, such as certifiedtive extraction techniques, such as Soxhlet extrac- standard reference materials.tions, are usually designed to give complete ex-tractions irrespective of the matrix. This is anessential feature of a method that can be applied to a6 . Analytes in solid samplesrange of samples, such a different soil types, butlimits selectivity. If the whole of a solid sample is readily soluble,

In contrast, when supercritical fluid extraction dissolution in a suitable solvent or water followed by(SFE) was first introduced, it was claimed to be liquid partitioning is usually the easiest method (seehighly selective compared to Soxhlet extraction but Section 7). However, most solid samples, such asin reality the carbon dioxide solvent was simply a soils, environmental solids, plant material, and poly-weaker eluent and hence more selective extraction mers, are largely insoluble and usually cannot bemedium. With standards and model matrices, there examined directly. In some cases, it is appropriate towere few problems but when the method was applied digest the sample in strong acid but in most casesto real samples, yields were found to depend on the this would destroy the analytes and is principally ofage of the sample[16] and type of soil being interest for the determination of inorganic elementsextracted[17]. The method might work for a simple or ions.matrix, such as sand, but real soil matrices with For most samples, it is necessary to extract thediffering interactions, moisture content and organic analyte of interest out of a residual matrix with 100%components often caused difficulties and incomplete efficiency but with also achieving as much specificityextractions. Interestingly, because compounds can be and selectivity as possible to simplify the subsequentmore tightly bound as a matrix ages, it has been separation steps. Typical methods use exhaustivesuggested[18] that the mild SFE extraction con- extraction in a Soxhlet system in which the solvent isditions might give a closer indication of the bioavail- continuously recycled through the sample for someability of the pollutant and thus be more environmen- hours. However, the analyte must be stable in thetally significant than more comprehensive extraction refluxing boiling solvent. Less efficient methodsmethods. included stirring the sample in hot or cold solvents

One further example is the problems that can arise for prolonged periods. All these processes were oftenif methods are not fully tested. In SFE there were quite slow and required the use of significantfrequent reports that an extraction was complete if a amounts of sample and large volumes of organicrepeat extraction under the same conditions yielded solvents to ensure complete extraction. The sub-no further analyte (for example, Ref.[19]), it was sequent work-up employed solvent evaporation andsubsequently found that the only reliable guide was concentration of the sample was slow and manuallythe extraction of a standard sample of known com- laborious. There was the added disadvantage that anyposition. It was often observed that more powerful impurities in the extraction solvent were also concen-extraction conditions (modifier additive, higher tem- trated.peratures or pressures) would result in additional The aims of most recent methods for the ex-

8 R.M. Smith / J. Chromatogr. A 1000 (2003) 3–27

traction of solids have been to reduce the amount of analyte trapped on glass beads or a cartridge, andsolvent and sample, reduce the time required, and subsequently extracted into a smaller solvent vol-enhanced the selectivity of extraction. The first two ume.aims have frequently been achieved but the last is The method has been applied to a number ofharder as in any extraction process there has to be a matrices, including marine particulate materials[22],balancing of selective and complete extraction. In pesticides in soils[23,24], medicinal plants[25,26].most cases, smaller samples are now used but this The many applications for soil[27] and environmen-does impose a restriction that the sample homo- tal samples[28] have been reviewed. Frequently thegeneity may limit reproducibility. There have been studies have compared PLE with conventional alter-two principal approaches, the use of conventional native methods, such as SFE[29,30], including asolvents in more efficient ways or the employment of comparison of methods for the extraction of en-alternative solvents, such as supercritical fluids. vironmental matrix standards[21]. In situ derivatisa-

tion of the sample can be used to enhance extract-6 .1. Enhanced solvent extraction methods ability [31]. Once the technique had been introduced,

the US Environmental Protection Agency (EPA)The extraction process can be speeded up by rapidly adopted it for the analysis of pesticides in

heating or agitating the sample (in pressurised liquid soils[32], as effectively it used the same solventextraction and microwave assisted extraction) or by systems as conventional liquid extraction. Manyusing an alternative solvent, which has a higher other EPA methods using PLE have since beendiffusion rate (as in supercritical fluid extraction and published. In contrast it has taken many years for thesuperheated water extractions). SFE method (Section 6.1.3) to be accepted.

The initial extraction can be often combined with6 .1.1. Pressurised liquid extraction a second sample preparation method, such as solid-

By employing a closed flow-though system, it is phase extraction or stir-bar extraction (see later), topossible to use conventional organic solvents at concentrate the analytes before analysiselevated temperatures above their atmospheric boil-ing points. This method, known as pressurised liquidextraction (PLE)[20,21], has been commercialised 6 .1.2. Microwave and sonic wave assistedin an automated or manual version as acceleratedextractionsolvent extraction (ASE). A restriction or backpres- For a number of years microwaves have beensure valve ensures that the solvent remains as a employed to assist the digestion of solid samples byliquid but has enhanced solvation power and lower focusing energy into the sample, resulting both inviscosities and hence a higher diffusion rates. Both heating and increased agitation[33]. This methodchanges increase the extraction rate. Both static and can also be used to enhance solvent extractionflow-through designs can be used. In the latter, fresh methods but the main disadvantage is that it uses asolvent is continuously introduced to the sample single extraction vessel and the sample vessel has toimproving the extraction but diluting the extract. been cooled, before the extract can be obtained.

As a consequence, extraction procedures, which Multiple samples can be extracted simultaneouslywould have taken many hours of Soxhlet refluxing, but it is difficult to employ the technique as a flowcan be carried out in minutes on a smaller sample, system and thus hard to automate.considerably speeding up the sample pre-treatment The method has been used to extract pesticidesand requiring a small fraction of the original solvent and herbicides from soil[34,35], fungal metabolitesvolume. An essential feature of the success of the [36] and essential oils from plant materials[37], andsystem is the ability to carry out multiple extractions polycyclic aromatic hydrocarbons (PAHs) in sedi-and hence move towards automation. The extracts ments[38]. Comparisons have been made with otherare generally much more concentrated than from extraction techniques, such as supercritical fluidconventional extractions. They could often be ana- extraction[39,40] or Soxhlet extraction[41,42] andlysed directly or the solvent could be cooled, and the the application to solid matrices have been reviewed

R.M. Smith / J. Chromatogr. A 1000 (2003) 3–27 9

[43]. Microwave extraction has also been combinedwith PLE [44] for the extraction of polymers.

Alternatively sonication can be used to enhanceextraction [45] and this has been applied for theextraction of organophosphorous pesticides.

6 .1.3. Supercritical fluid extractionOne area that stimulated an interest in enhanced

fluid extractions was SFE. This is a long establishedmethod, which has been used industrially for manyyears. However, it was not until an interest wasshown in supercritical fluids as a chromatographicmedium that it started to be seriously studied as anextraction technique on an analytical scale. It hassince been the subject of numerous books andreviews (for example, Refs.[46–50]).

Almost all practical work has employed carbondioxide as the supercritical fluid as potential alter-native solvents, such as nitrous oxide proved dan-gerous because of their oxidising power[51] andmore exotic solvents like xenon were ruled out bytheir cost. In many ways carbon dioxide is an idealsolvent as it combines low viscosity and a highdiffusion rate with a high volatility. The solvationstrength can be increased by increasing the pressureand extractions can be carried out at relatively lowtemperatures. The high volatility means that thesample is readily concentrated by simply reducingthe pressure and allowing the supercritical fluid toevaporate.

The principal problem is the relatively low polari-ty of the carbon dioxide, ideal for PAHs and halo-genated pesticides, or lipids and fats, but unsuitablefor most pharmaceuticals and drug samples. It hasbeen quite a popular method for solid matrices,including powdered plant materials, herbalmedicines, some foods, and polymers[52] but thereare problems with liquids, such a biological fluids,which need immobilising on a solid support material.Although one advantage was claimed to be the mildextraction conditions, which would enable the ex-traction of thermally unstable compounds, there arefew examples, such as the extraction of fire re-tardants from plastic foams[53]. Often the extrac-

Fig. 1. Comparison of the gas chromatograms of extracts oftions were compared with alternative methods offeverfew obtained by different extraction methods. (A) SFE; (B)

sample preparation (Fig. 1) [54]. The addition of parthenolide standard; (C) steam distillation; (D) headspacemodifiers, such as methanol, to the carbon dioxide analysis; (E) solvent extraction. Peaks: 3, camphor; 5 chrysan-enables more polar analytes to be extracted and thenyl acetate; 12, dihydroparthenolide; 14, parthenolide[54].

10 R.M. Smith / J. Chromatogr. A 1000 (2003) 3–27

increases the scope of the method[55,56]. The high 6 .2. Problems with solid matricespressures required have caused some problems indeveloping automated systems but commercial sys- 6 .2.1. Biological matrices and matrix solid-phasetems are now available. dispersion

Most of the previous methods cannot be applied to6 .1.4. Superheated water extraction biological samples, such as meat or fish tissues and

Because the polarity of water decreases markedly undried plant material, because they rely on a non-as the temperature is increased, superheated water polar solvent and this cannot penetrate the largely(sometimes termed subcritical or pressurised hot aqueous matrix. Sometimes more polar water misc-water) at 100–2008C, under a relatively low pres- ible solvents can be employed for plant material butsure, can act as a medium to non-polar solvent and is this approach cannot be used with fatty tissues. Onean efficient extraction solvent for many analytes successful approach for pesticide analysis has been[57]. Typical applications of superheated water ex- to disperse the solid tissue, such as liver or kidneys,traction (SHWE) have included PAHs and poly- by macerating with a dispersion matrix—typicallychlorinated biphenyls (PCBs)[58] or pesticides[59] thin-layer chromatography (TLC) grade octa-from soils, and natural products[60] from plant decylsilyl (ODS)-bonded phase silica. This matrixmaterial. solid-phase dispersion provides a porous structure

So far the equipment has usually been laboratory- and enables the solvent to penetrate and extract themade but PLE systems can also be employed at a analytes. It also appears to partially carry out thehigher temperature than normal extractions[61]. The initial extraction from the aqueous sample phase.conditions are usually lower than the critical point of Sequential eluent then enables the analytes of interestwater at 3748C and 218 bar, because under those to be released. The ODS phase has the advantage ofconditions the high temperature causes sample de- retaining lipids so they do not interfere with thecomposition. At lower temperatures, the pressure has subsequent assays.little effect on the density of water and is not a However, the method is fairly labour intensivecritical operating parameter unlike in SFE. As with requiring the tissue to be ground up with the matrixother liquid extraction methods, superheated water and packed into an SFE type tube for extraction. Itsextractions are most suitable for powdered samples. application in food analysis has been reviewedA number of linked methods have also been de- [66,67], including drugs in fish[68], sulfonamides inscribed, including SHWE–gas–liquid chromatog- bovine and porcine muscle[69], and clenbuternolraphy (GLC)[62], SHWE–LC–gas chromatography from bovine liver[70]. Other dispersion and de-(GC) (Fig. 2) [63,64] and SHWE–superheated water siccant agents can also be used including sodiumchromatography[65]. sulfate and hydromatrix (particularly for SFE)[71].

Fig. 2. PHWE–LC–GC apparatus. 15N ; 2a,2b5pumps; 35elution and LC solvent; 45water; 55oven; 65preheating coil; 75extraction2

vessel; 85cooling coil; 95trapping column; 105restrictor; 115LC column; 125precolumns; 135analytical column; 145SVE; 155detector; V15extraction valve; V2–V45multiport valves[63].

R.M. Smith / J. Chromatogr. A 1000 (2003) 3–27 11

6 .2.2. Insoluble solid matrices—pyrolysis 6 .2.3. Thermal desorption from solidsThe pyrolysis of samples to form characteristic Volatile analytes in solid matrices can be released

fragments, which can be separated and analysed by for analysis by thermal desorption, for example theGC [72,73], has been used for many years for the analysis of chlorinated components in soils[80], oranalysis of insoluble matrices, such as polymers volatile constituents of oak wood[81].[74,75] plastics, automotive paints[76] and somedrugs. Some recent examples have examined Egyp-tian mummies (Fig. 3) [77] and have combined 7 . Analytes in solutionpyrolysis with in situ silylation to give trimethylsilyl(TMS) derivatives of resin acids from Manila copal The traditional method to obtain analytes from[78]. A novel application was the use of a thermal liquid samples has been either by partitioning into anprobe on a scanning probe microscope to select and immiscible solvent, trapping the analyte onto apyrolyse a small area on the surface of a polymer or column or solid-phase matrix of some sort, or as aplant material followed by GC–mass spectrometry last resort evaporation of the sample to dryness and(MS) [79]. selective solvation of the analytes. The most com-

mon method for an aqueous matrix was to use aseparating funnel and extract any organic compounds

into a non-polar solvent. The method would typicallyuse large volumes of organic solvent (100–250 ml)from a similar volume of sample and the extractionwould have to be repeated 2–3 times to achieve ahigh recovery. After drying, the solvent would beconcentrated by evaporation. The resulting samplewould frequently require a further clean-up stage.With some samples, the initial solvent extraction stepresults in the formation of an emulsion and theextraction process could become prolonged.

Overall the process was slow, required consider-able manpower and was hence costly. It generated alarge volume of organic waste, which was environ-mentally unfriendly, and its disposal is becomingincreasing difficult (and costly). The repetitive manu-al operations often lead to errors and could be aboring task for the operator, although crucial toobtaining reliable results. There has also been arecognition that the use of large volumes of solventposes hazards to the health of the laboratory workerand can have a direct impact on the environment.The final blow to the method came with the Montrealprotocol, which limited the widely used chlorinatedsolvents because of their effect on the ozone layer.There has hence been a considerable interest in thereduction of solvent usage and/or alternatives to

Fig. 3. Total ion current (TIC) of the pyrolysis profile of (a) chlorinated solvents, and in methods capable ofHoremkensi, resin-like material) and (b) Khnum Nakht, bandage/ automation.resin / tissue after thermal desorption. Note:j5alkenes; d5

Two groups of methods have been developed,alkanes,.5alicyclic hydrocarbons;m5aromatic hydrocarbons;those which trap the sample out of solution onto as52-alkanones;h53-alkanones;n5cyclic ketones;,5nitriles;

�5amides; *5steroids[77]. small volume of an immobilised phase, such as SPE

12 R.M. Smith / J. Chromatogr. A 1000 (2003) 3–27

and solid-phase microextraction (SPME) and related early problems, the retention properties of the car-methods, and those which transfer the analytes to a tridges can now be expected to be consistent betweensmaller volume of a second solvent, such as mem- batches and the flow-rates and trapping efficiencybrane extractions. Both methods are compatible with will be reproducible. However, as with high-per-automation. In addition to the direct extraction, these formance liquid chromatography (HPLC) columns,methods can also be used to concentrate the analytes nominally equivalent (for example, ODS phases)from extraction solutions of solid samples (see from different manufacturers may have differentprevious sections). Often the methods are directly bonding chemistries and carbon loadings and so canintegrated with the separation stages to further behave differently. It took some time for SPE to bereduce sample handling. widely adopted and for robust methods to be de-

veloped. For example, there was a need to under-7 .1. Trapping the analytes stand the requirements of preconditioning and the

importance of consistent flow control.These methods extract the analyte by trapping it Although the cartridges are single-use and dispos-

onto an immobilised phase, the analyte is then able and thus represent a significant consumablewashed off with a minimal small volume of solvent cost, this has been claimed to be much lower thenor eluted thermally. They are usually considerably the cost of chemicals and manpower needed for thefaster and use significantly smaller volumes of corresponding traditional solvent extraction methods.solvent and sample than traditional extraction meth- Other formats have also been developed for solid-ods phase extraction, including flat disks with the station-

ary phase particles supported on a mesh, enabling7 .1.1. Solid-phase extraction very large volumes to be rapidly extracted[87].

The introduction of the disposable pre-packaged Recent use of high flow-rates through extractionSPE cartridge had a major effect on methods for the cartridges has been claimed to give improved ex-examination of analytes in solution[82–86]. Al- traction [88] but such ‘‘turbulent flow extractions’’though the concept of using a short column for seem little different to conventional extractions.sample clean-up has been employed for many years, The scope of SPE is considerable, with a wideusually hand-packed normal-phase materials were range of reported permutations of cartridge materialused, such as silica or Fluorisil. Their principal role and eluents /sample matrices. Numerous methodswas the retention of unwanted components from the have been developed and reported and libraries ofsample, such as tars and polar or involatile com- applications are available on manufacturers’ websitespounds, in the clean-up of pesticide residues and and in the literature.environmental samples. The SPE cartridge intro- One of the principal applications of SPE has beenduced two important features, standardisation and in the extraction of drugs and their metabolites fromhence greater reproducibility, and a much wider body fluids. The disposable cartridges reduce therange of phases, importantly including reversed- handling of body fluids, such as urine and blood, andphase and ion-exchange materials enabling aqueous hence the biohazard to the operator is minimised.solutions to be treated and additional trapping mech- When large numbers of related assays are required asanisms to be utilised. in toxicology studies the process can be further

A wide range of phases means that either polarity, automated using a robot[89,90] or an intelligenthydrophobicity or ionisation can be used as trapping autosampler[91,92] almost completely eliminatingmechanisms and the sample matrix may now be sample handling. Extraction onto sample disks hasnon-polar or aqueous. Once trapped, the analyte can been developed as a method for the determination ofbe released into a small volume of an extraction organochlorine pollutants in body fluids[93].solvent by altering the polarity or pH. In some The second widespread application of SFE hasexamples, impurities are trapped and the analyte of been for environmental samples, such as river watersinterest passed through the cartridge, but it is usually and sewage outflow, where large volumes of verythen concentrated on a second cartridge. After some dilute solutions have to be extracted[94]. With

R.M. Smith / J. Chromatogr. A 1000 (2003) 3–27 13

conventional solvent extraction, large volumes of trapped on a short cartridge, then eluted thermallysample solution had to be manipulated to obtain directly onto a superheated water chromatographicsufficient analytes for assay. With SPE cartridges, the separation[65].sample is simply pumped through the SPE bed andthe analytes are then eluted with a small volume of 7 .1.2. Solid-phase microextractionorganic solvent. Typical examples are the assays of In solid-phase extraction, it is still necessary totrace levels of PAHs from river water or non-polar extract the sample from the column, usually with anpesticides. A limit of the degree of concentration is organic solvent, before it can be injected into aimposed by the breakthrough volume of the cartridge separation method. This last step and the need for an(when even the weak aqueous eluent effectively organic solvent were eliminated in the ingeniousstarts to elute the sample) or the overloading of the SPME method, which was invented by Pawliszyncartridge by other sample components. The large and co-workers[100–102].They used a fibre coatedsample volumes required are aided by the use of the with a stationary phase as the extraction medium.disk format, such as the extraction of estrogen from After carrying out an extraction from a samplesewage and river waters[95]. solution, the fibre could be placed in the injection

The extraction of the concentrated analytes from port of a gas chromatograph so that the analytes werethe cartridge can either use a solvent or the elution thermally desorbed directly into the carrier gascan be accelerated by heating, effectively combining stream. The method has been automated and com-SPE and PLE. The eluted sample can be linked mercial systems are available that will both extract,directly to GC (Fig. 4) [94,96] or to an LC sepa- agitate the sample and inject into a GC system.ration [97,98]. In recent work, the cartridge can also Assay by HPLC can also be employed but thebe eluted with superheated water[62,99] for off-line sample is extracted directly into the eluent streamanalysis by HPLC or to on-line gas chromatography rather than thermally desorbed (Fig. 5) [102]. A[63]. A further method has been described in which number of different fibre coatings are available,the solution from a superheated water extraction is which offer a range of analyte solubilities and

Fig. 4. Scheme of an on-line SPE–GC system consisting of three switching valves, two pumps and a GC system equipped with an SVE, anda mass-selective detector[96].

14 R.M. Smith / J. Chromatogr. A 1000 (2003) 3–27

different alcoholic drinks[106]. For some routineapplications, non-equilibrium conditions can be usedas long as the extraction conditions are reproducible.

The main advantages of the system are that nosolvent is required to elute the sample from the fibreand there is a direct transfer from the sample solutionto the separation method. Unless the matrix is verycomplex or involatile, the fibre can be reused numer-ous times as the thermal elution step also cleans thefibre. The disadvantages are that the fibre is fragileeven though it is shielded when out of the sampleand it can be damaged by a build-up of involatilematerials from the samples. The extraction processcan be relatively slow because it relies on sufficientstirring or diffusion to bring the analytes into thelocation of the fibre and good reproducibility re-quires that an equilibrium is established. The fibrecan be also used to assay the headspace above thesample (see Section 8.2) and this method is preferredfor volatile analytes as the fibre avoids contact withthe matrix solution.

Fig. 5. Isocratic separation of a four-PAH mixture by (a) 1ml loop The scope of SPME–GLC can be expanded forinjection and (b) fibre injection, 7mm PDMS extraction for 30 some involatile analytes by on-fibre derivatisation tomin from 100 ppb of each compound spiked into water. Peaks: (1)

enhance either separation[107] or detection, forfluoranthrene, (2) pyrene, (3) benz[a]anthracene and (4) ben-example the reaction of chlorophenol with penta-zo[a]pyrene[102].fluorobenzoyl chloride to give increased responsefrom the electron-capture detector[108].

porosities, including the non-polar polydimethyl Although conventional SPME uses a coated fibre,siloxane (PDMS), semi-polar PDMS–divinylbenzene which is immersed in the sample solution, an inter-and polar polyacrylate, and Carbowax–divinylben- esting variant employs an internally coated capillaryzene liquid like phases and the coated porous particle through which the sample flows or into which thephase PDMS–Carboxen, They are available in in- sample is sucked up repeatedly[109,110]. Thecreasing thicknesses from 7 to 100mm, which extraction components are then eluent by a solvent.increases the partitioning ratio and hence improves In recent developments, a restricted access coatedsensitivity but increases equilibration times. tube using an alkyl diol-coated silica material (Fig.

The theory and practice of the method has been 6) has been used to selectively trap drugs fromexamined in considerable detail in recent years[103] serum without suffering protein fouling of the sur-and numerous applications has been reported and face[111,112].reviewed [104,105]. The basic theory is that of aphase distribution and the amount extracted depends7 .1.3. Stir-bar extractionson the partition coefficient between the sample Because the SPME fibre has a relatively smallsolution and the fibre. However, the fibre volume is volume of bound stationary phase, the extraction issmall so that the target analyte is often not complete- frequently incomplete. Even with a favourable dis-ly extracted. However, a representative sample is tribution constant, the phase ratio between the fibreobtained that can be compared with the extraction of and sample solution are often unfavourable, so thata standard solution. The yield can be susceptible to the partitioning can still leave a significant amount ofmatrix effects, if these alter the distribution constant, the analyte in the sample phase. This problemsuch as changes in the ethanol content between prompted the development of the stir-bar extraction

R.M. Smith / J. Chromatogr. A 1000 (2003) 3–27 15

Fig. 6. In-tube alkyl diol silica restricted access SPME system in (A) load position for extraction from serum and (B) injection position(elution onto analytical column)[111].

system (marketed commercially as the Twister), tubing. The surface area of the stirrer bar is higherwhich uses a magnetic stirrer bar or flea coated with than a fibre and the volume of the adsorbent layer isa bonded adsorbent layer (such as a polymethyl much larger so that there is a higher phase ratio thandimethyl siloxane)[113]. Alternatively a magnetic in SPME and hence a higher extraction yield.stirrer can be inserted into a short length of PDMS The stir-bar is simply rotated in the sample,

16 R.M. Smith / J. Chromatogr. A 1000 (2003) 3–27

removed and extracted thermally for gas chromatog- 7 .2. Extraction of the analytes into a liquid phaseraphy[113] (using a thermal desorption unit) or intoa solvent for liquid chromatography[114,115].It has Rather than distribute the sample between a pair ofproved very good for complex and semi-solid ma- immiscible (usually polar and non-polar) solvents intrices, such as yoghurt or beer, and pesticides in wine a traditional separating funnel, three alternative[116]. More unusual applications, included the assay liquid–liquid extraction methods have been reported,of PCBs in human sperm (Fig. 7) [117]. The main which give a more concentrated extract ready fordifficulty is that it is hard to automate the removal of direct chromatographic examination. However, truethe stir-bar from the sample matrix, rinse it, and liquid–liquid counter-current methods, in which twoextract. immiscible liquids flow through a tube in opposite

As with a number of related methods, it can also directions are now fairly rarely used, largely becausebe used to concentrate the analytes in an extract from of the time taken to set up and the difficulty ofan alternative extraction process, for example it has obtaining two truly immiscible liquids.be used to concentrate the analytes from a PLEsolution to determine the pesticides in strawberries 7 .2.1. Membrane extraction[118]. A membrane can act as a selective filter, either

Fig. 7. GC in the selected ion monitoring (SIM) mode of seven PCBs extracted using a stir-bar from human sperm at 10 ppt (A) and 1 ppt(B) [117].

R.M. Smith / J. Chromatogr. A 1000 (2003) 3–27 17

just limiting diffusion between two solutions or as anactive membrane in which the chemical structure ofthe membrane determines the selectivity of sampletransfer[119–122].In most cases, the driving forcefor the movement of the analyte across the mem-brane is a concentration gradient. This can beenhanced by effectively removing the analyte fromthe receiving phase by either ionisation using buf-fers, complexation, or derivatisation, so that the freesolution concentration of the analyte species isreduced. By altering the flow-rate of the solutionspassing either side of the membrane, a low con-

Fig. 8. Different membrane modules for flow systems. (a) Flatcentration in a large volume can be converted into amembrane module with spiral channel; (b) flat membrane modulehigher concentration in a smaller volume (Fig. 8).with 10 ml channel volume; (c) hollow fibre module with 1.3ml

The extraction can be also carried out to transfer a acceptor channel[122].volatile analyte from a liquid to a gas phase by usinghollow fibre membranes, linked directly to a GCsystem (Fig. 9) [123]. Recently a microporousmembrane has been incorporated into a superheatedwater extraction to concentrate a sample of PAHs

from soil before GC analysis[124].Dialysis methods and microdialysis[125] are

closely related to membrane separation, with acontrolled pore structure providing a separationdiffusion process based on molecular size. In vivomicrodialysis with the end of the microdialysis probeplaced in living tissue enables real time measure-ments of body chemicals in test animals[126]. Themembrane or dialysis method can be directly con-nected to the sample loop of a HPLC injection portso that the dialysate can be directly injected[127].

7 .2.2. Single drop extractionIn a recently developed microscale method, rather

than using an immobilised phase, a single liquid dropis utilised as the collection phase[128,129]. Al-though elegant the method appears to require highmanual dexterity. It requires a collection phase witha sufficiently high surface-tension to form a distinctdrop, which can be exposed to the analyte solution(Fig. 10). It has been used for pollutants and canreadily be linked to GC.

7 .2.3. Purge and trapPurge and trap systems in which a volatile analyte

is expelled from a solution by flushing it out with agas [130] and then trapping the components of Fig. 9. Different configurations of hollow fibre membrane ex-interest in a cryogenic trap, solvent or solid-phase traction modules for volatile organic compounds[123].

18 R.M. Smith / J. Chromatogr. A 1000 (2003) 3–27

recent years alternative trapping methods have beenused and these are still developing.

Gaseous samples are of interest directly as ameasurement of the environment, for example inworkplace exposure to solvents, or as the products ofa chemical process or combustion. The vapour abovea sample is also of analytical interest as the con-centration of volatile analytes in the vapour phasecan be directly related to their concentration in thematrix.

8 .1. Trapping analytes from vapour samples

A number of methods have been used to trap andconcentrate components from gases. Some of themore efficient methods have effectively passed thegas over a cold adsorption tube packed with a formof GC stationary phase, including adsorptive materi-als, such as porous carbon, or sorptive polymers,such as Tenax, polystyrene–divinyl benzene orFig. 10. Schematic of a single drop microextraction apparatus

[129]. PDMS [132]. The gas may be pumped for a specifictime or can be allowed to diffuse into the trap in

trap (see also the next section) have been useful for long-term workplace exposure studies. The trappedlow levels of analytes in environmental solutions. components are then usually desorbed thermally andFor example it can be used to examine sulfur- passed directly into a gas chromatograph for sepa-containing analytes in beer, coffee and water[131]. ration and quantification. A typical recent example is

the indoor air monitoring of monoterpenes[133].Alternatively, the adsorption tube can be eluted using

8 . Analytes in the gas phase a volatile solvent. Typically carbon disulfide is usedbecause of its high volatility and lack of response in

It might seem that little sample preparation of a flame ionisation detector. However, it is a hazard-gases should be needed as they can be analysed ous chemical and this method is difficult to auto-directly by gas chromatography. The whole sample mate, whereas automated thermal desorption (ATD)is volatile and thus will leave no residues. However, systems are commercially available, although largethe analytes of interest are often at low concentration sample numbers are needed to justify the investment.near the limit of detection and the high diffusionrates in gases mean that the integrity of the sample is 8 .2. Headspace analysishard to maintain from the collection point to theanalyser. There has therefore been considerable If the components of interest in a solid or in-interest in concentrating, focusing, or trapping out volatile matrix are volatile, a well established meth-the analytes of interest to increase sensitivity and od[134–136] is to assay them by examining theirtransportability. concentration in the headspace gas above the matrix,

Early methods tried to trap out the analytes using a either by taking a direct gaseous sample or trappingcold trap or solvent trap from a flowing stream. the volatile material on an SPME fibre (see below).However, misting rather than condensation can occur The sample is usually heated to increase the vapouror the flowing gas bubbling through a trap can phase concentration and both manual and automatedpartially desolvate volatile components, causing low systems are available, the latter giving higher repro-yields and under-estimating real concentrations. In ducibility.

R.M. Smith / J. Chromatogr. A 1000 (2003) 3–27 19

Either a sample can be taken directly from the applied to organic pollutants[146,147], arson sam-headspace (static headspace analysis) or the gas ples[148], packaging materials (Fig. 11) [149].above the matrix can be flushed from the sample Although a sealed system might seem necessary,vessel and trapped as in the previous section (dy- open-capped vials in which there is a narrow re-namic headspace analysis). The latter effectively stricted inlet have also been used and are easier toflushes the full headspace gas and concentrates the handle in automated systems[150]. Another recentsample and thus is inherently more sensitive. The innovation has been to use microwaves to assist thetime of extraction and the degree of sample agitation evaporation into the headspace coupled with SPMEare important, as these will influence the rate of [151]. Gas-phase membrane extraction has also beenrelease of the analyte from the matrix. The dynamic used to trap analytes from the headspace of samplesmethod is very similar to purge and trap except that [152].the incoming gas flow is not passed over not througha liquid matrix.

Typical analytes and matrices are solvents in body 9 . Direct combination of sample preparationfluids (in particular ethanol in blood as a test for and separationdrunken drivers), solvents in matrices, such as poly-mers or paints, and plastic monomers in food pack- To reduce the manual stages involved in sampleaging plastics. There are also numerous applications preparation, analysts have spent considerable effortto food samples, such as tomatoes[137], the sulfur to link extraction or sample clean-up steps directly tocomponents of beer[106], fatty acid esters in rum the separation methods. These linkages can be[138] and spice samples[139], such as coriander relatively simple, like thermal desorption into gas[140]. chromatographs, to automated sample stations like

The principal difficulty is accurate quantitation, AASP[153] and ASTED[91] in which a sample canalthough this is aided by automation, and standards be extracted onto a SPE cartridge after addition of anneed to be prepared by the method of standard internal standard and the extract eluted and injectedadditions or matrix spiking. Because the assay is into a HPLC system. More complex sequences canbased on the distribution of the analyte between the be a carried out by robotic arms. However, thesegaseous and matrix phases, the concentration in the require more careful and extended setting-up andvapour phase can be altered by the solubility of the verification procedures and the time and effort spentanalyte in the matrix phase. For example, with at this stage must be balanced by a saving over analcoholic beverages the concentration will vary with extended series of analyses[89,154].the ethanol content of the drinks[141,142].Desirab- Virtually every possible combination and multiplely a similarly volatile internal standard should be combinations have been explored; including super-used. Quantitation can also be obtained by sequential critical fluid extraction to supercritical fluid chroma-extraction[143,144]and back-calculation. tography[155], SFE to LC[156], PLE–SPE–HPLC

Rather than extracting the vapour or flushing it [157]. As an excess of solvent is usually employed infrom the analysis bottle, the headspace can be an extraction frequently some type of focusing of thetrapped on a SPME fibre[145]. However, the analyst sample is usually required at the injection point ofneeds to be aware that the distribution is between the the separation method, such a low temperatures infibre and matrix. Thus raising the temperature re- GC.duces the deposition onto the fibre (because itincreases the vapour concentration above the fibre as9 .1. Large volume injections in GCwell as above the sample), even though it increasesthe concentration in the headspace. Thus SPME The amount of liquid that can be injected directlysampling can give a very different selectivity to into a gas chromatographic capillary column withoutdirect headspace analysis. The headspace sample will causing band spreading can be very limited. A fairlyfavour the volatile analytes but the fibre will favour recent development has been methods, which enablethe less volatile components. This approach has been quite large samples to be injected. By the addition of

20 R.M. Smith / J. Chromatogr. A 1000 (2003) 3–27

Fig. 11. GC–flame ionisation detection (FID) chromatograms from a packaging material with an unacceptable odour obtained by (a)headspace analysis (b) headspace SPME analysis. Reproduced from Ref.[149].

a vent after a pre-column, large amounts of solvent concept to determine the hydrolysis products ofcan be vaporised prior to the main analytical col- sulfur mustards[161] and triazines after membraneumns but leaving a film on the pre-column wall extraction[162].which solvates the analyte[158–160].As the evapo-ration ends, the vent is closed and the residual 9 .2. Coupled column systems LC–LC or GC–GCsample is chromatographed. The technique has beenused to inject 100–200ml or up to 500ml of aqueous Coupled-column separations or multidimensionalenvironmental samples. Examples have used the chromatography can be considered as a form of

R.M. Smith / J. Chromatogr. A 1000 (2003) 3–27 21

sample preparation, as one column is used to derive phoresis, utilising differences in the migration ratesfractions for the second column. Most of the con- of a pusher solution so that the analyte is focused tocepts have been well developed and reported as a single point before the electromigration techniquecoupled or multidimensional chromatography[163]. occurs[167].Related methods include column-switching tech-niques, such as heart-cut, in which a fraction fromone column is transferred to a second column for an 1 0. Selectivity enhancementadditional separation and back-flushing, in whichmore highly retained materials are washed back from In most of the methods described so far, thea column system through the inlet. These methods discrimination between analytes has been based onare more commonly used in GC than LC as in the differences in their physical properties, which islatter case the reversal of the flow is harder and more exploited as solubility, partitioning or volatilitylikely to disturb the bed of the column. The complete differences enabling discrimination. A further dis-combination is two-dimensional chromatography in tinction is also possible in which discrimination canwhich fractions from the first column are continuous- be obtained by a specific structural difference inly passed to a second column to give a very high interaction, either utilising or mimicking a biologicalsample capacity. These can include GC3GC difference.[164,165], which can generate very high resolution(Fig. 12). 1 0.1. Affinity methods

9 .3. Isotachophoresis in capillary electrophoresis Affinity chromatography is a long employed tech-nique that uses the very specific interactions that

In capillary electrophoresis, dilute samples can be occur between analytes and biological systems tofocused within the separation capillary by isotacho- specifically retain or trap compounds because the

Fig. 12. GC3GC analysis of cracked gasoline using column 1, 10 m DB-1 and column 2, 0.5 m OV1701 with an oven temperatureprogramme of 28C/min from 30 to 2008C [166].

22 R.M. Smith / J. Chromatogr. A 1000 (2003) 3–27

column coating recognises a particular structural been examined to provide chiral selectivity althoughshape or interaction[168–170]. The most specific the discrimination is relative rather than absolute.form is immunoaffinity chromatography, which em- Because the specificity of the interaction is oftenploys an antibody of the analyte to interact and dependent on a hydrogen-bonding interaction thespecifically retain it from a solution. The interaction MIPs are often restricted to use with normal-phaseis then broken by solvent or pH changes. Apart from solvents as aqueous solutions preferentially bond andvery closely related analytes, the method is highly deactivate the interaction sites.specific. For example, a test for one barbiturate Recent examples of the use of MIPs have includedmight trap other barbiturates to different extents phases to trap caffeine[175], which also show some[171]. However, the need for the antibody mean that selectivity toward theophylline and theobromine,few commercial columns are available and it is salicylic acid[176], cholesterol[177] and quercetintherefore difficult to obtain columns for specific (Fig. 13) [178].assays. However, if the number of assays requiredcan justify the method it can provide a very simpleand efficient clean up. 1 0.3. Restricted-access media

1 0.2. Molecular imprinting polymers One concept that was examined with some suc-cess, was developed originally by Hageston and

Attempts have been investigated to mimic the Pinkerton[179], who to designed a HPLC columnselectivity of interaction in affinity separations by whose packing had a hydrophilic external surfacemaking a synthetic polymer, which contains im- and a hydrophobic internal surface, which acted as aprinted cavities generated by a template molecule. reversed-phase material. These restricted-accessThese molecular imprinted polymers (MIPs) have phases could be effectively used as an on-columnbeen used for both separations and sample clean-up sample preparation media, which excluded biopoly-as SPE cartridges[172–174].However, the degree of mers, which were rapidly eluted, but retained smallerselectivity has been questioned and often they func- analytes for separation[180,181].More recently thetion as group-selective systems for compounds re- same types of materials have been used in SPElated to the original template. This may have advan- cartridges and in-line traps designed for repeated use,tages in areas, such as pesticide analysis, when only in which the external biocompatible outer layer isa group separation is required. Attempts have also based on aa -acid glycoprotein [111,182]. The1

Fig. 13. HPLC separation of merlot (2) before MIPS extraction and (1) fraction eluted from MIPS cartridge with acetonitrile at 265 nm on aKromasil C column[178].18

R.M. Smith / J. Chromatogr. A 1000 (2003) 3–27 23

phase materials can be polymeric[183] or based on which still has the advantage of high efficiency andsilica. easy linkage to mass spectrometry needed for many

studies, such as drug screening. The principal re-actions are the formation of trimethylsilyl ethers

1 1. When separation alone is not enough— from sugars, steroids and alkaloids, the methylationderivatisation to see the sample of fatty acids and transesterification of lipids, and the

acylation of amines.The above methods have generally tried to convert Some early methods for chiral separations used

a sample into a form for direct analysis, however, derivatisation to create diastereoisomeric mixturesbecause of analytical and detection limitations, many enabling separation on achiral column but so manysamples are incompatible with the separation meth- chiral separation columns are now available that thisods. The derivatisation can either be as part of method has fallen into disuse. There also concernsample preparation (pre-column) or as an aid to that the reaction could itself be stereoselective anddetection (post-column) although often the two roles hence the results would not reflect the originalare combined and pre-column reagent are selected to enantiomeric ratioalso enhance detection. The original methods weredriven by the inability of GLC to handle directly

1 1.2. Derivatisation to enhance thermal stabilitymany of the involatile or polar analytes found inbiochemistry, such as carbohydrates, lipids, fatty

Although often mentioned in texts this concept isacids and sterols. Frequently these analytes were also

rarely applied in practice. It was principally a GCaliphatic and although could eventually be examined

concern but most affected compounds can now beby HPLC, they had detection problems as they

examined by LC.contained only weak chromophores, such as manyamino acids and sugars.

A very large number of reactions have been 1 1.3. Derivatisation to enhance detectionreported but in reality only a few have been used inroutine analyses. Even though many textbooks and Particularly in HPLC, some analytes are moremonographs have reported compilations of derivati- difficult to detect and pre-column reagents weresation techniques as part of sample preparation[184– selected, which introduced chromophores or fluoro-186], this is an approach that most analytical chem- phores to enhance detectability and often also re-ists will avoid for a number of reasons. The problem duced interaction problems on the column by reduc-is that derivatisation adds an additional step to the ing the ability of the analysts to ionise. However, insample preparation procedure. As well as the extra more recent years the use of less active stationarycosts involved, care must be taken to ensure that the phases, and the introduction of ion-pair separationsreaction is working by introducing derivatisable (and even ion chromatography) and more universalstandards. The additional manual or reagent addition detection, with the mass evaporative and the nowstages introduce additional uncertainty into quantita- increasing spread of mass spectroscopic detectors,tion. Despite their limited role many research groups has changed the situation considerably. Consequent-still study derivatisation reactions but often propose ly, few routine methods would now use derivatisa-methods that in reality offer little advance on exist- tion unless the limits of detection were being ex-ing methods and frequently employ reagents that amined. In many cases laboratories will examinehave to be specifically synthesised. almost any alternative to avoid derivatisation.

Derivatisation is still used for a few samples, such1 1.1. Derivation to enhance volatilisation and as amino acid separations, or in fields, such asseparation capillary electrophoresis, capillary electrochromatog-

raphy and microbore LC, where detection is aThe main application of derivatisation is to in- problem because of the limited cross-column path

crease the volatility of analytes for GLC analysis, length for spectroscopic detection. It is also often

24 R.M. Smith / J. Chromatogr. A 1000 (2003) 3–27

also applied in some lab-on-a-chip applications may also increase discrimination by using improvedwhere sample mass is limiting. mass discrimination as a form of resolution avoiding

the need for clean up but expensiveHowever, the view was expressed at a recent

1 2. Can sample preparation be avoided? meeting that one effect of the use of LC–MS hadbeen the disappearance of a thorough knowledge of

Because of the extra work in the inclusion of a sample preparation[189] and it was felt that to getsample preparation stage methods in an assay, there the full advantages of LC–MS, extensive work-up ofis considerable interest in simplifying method or in the sample could still be needed.finding ways to combine the preparation and assay ina single stage. Many examples have already beenindicated; the reduction of solvent extraction by 1 3. Conclusionsusing SPE and SPME, and reduced use of deri-vatisation. More specific extraction can help so there As can be seen sample preparation is still evolvingis a continuing interest in MIPs. However, some and may still be required as even highly discriminat-sample preparation is often still required to overcome ory detector methods may suffer interferences. Gen-the influences that differences in the sample matrix erally extraction methods are becoming more selec-might have on the analytical step. A further problem tive and more readily combined directly with sepa-can arise if residues of the matrix are left in the ration methods. Temperature, alternative solvents,injection or separation system, as they can affect and smaller sample sizes are reducing the use oflater separations. For example, a build-up of lipids organic solvents but care is still needed that with realon a reversed-phase column can change the sepa- samples that the amount taken for the assay isration characteristics. representative of the total sample.

An alternative approach has been to make thedetection process more selective so that interferingspecies are simply not detected. Here single and R eferencesmultiple ion monitoring in GC–MS are a help andLC–MS–MS methods can greatly increase specifi- [1] R .E. Majors, LC?GC Int. 4 (1991) 10.

[2] M .-C. Hennion, C. Cau-Dit-Coumes, V. Pichon, J. Chroma-city. Effectively these methods filter the ions pro-togr. A 823 (1998) 147.duced by the MS process to separate the characteris-

[3] W .G. Jennings, A. Rapp, Sample Preparation for Gas Chro-tic ions for the analyte and to ignore those from the matographic Analysis, Huthig, Heidelberg, 1983.matrix or from other components, hence reducing the [4] S .C. Moldoveanu, V. David, Sample Preparation in Chroma-background signal. tography, Elsevier, Amsterdam, 2002.

[5] M . Szumski, B. Buszewski, Crit. Rev. Anal. Chem. 32However, some sample preparation still may(2002) 1.needed, otherwise interferences and signal suppres-

[6] R .E. Majors, LC?GC Eur. 16 (2003) 71.sion for LC–MS can occur, even if when there is no [7] V .R. Meyer, LC?GC Eur. 15 (2002) 398.obvious sample signal[187,188]. The sensitivity of [8] M . Tswett, Ber. Dtsch. Bot. Ges. 24 (1906) 384, TranslatedMS can also be dependent on the composition of the by H.H. Stain, J. Sherma, J. Chem. Educ. 44 (1967) 238.

[9] S .C. Churms, J. Chromatogr. 500 (1990) 555.mobile phase. LC–MS–MS is generally more sensi-[10] J . Bakthavachalam, R.S. Annan, F.A. Beland, P.Vouros, R.W.tive for readily ionisable analytes. This has the

Geise, J. Chromatogr. 500 (1990) 373.advantage that smaller sample volumes of biological [11] M .M.L. Aerts, W.M.J. Beek, U.A.Th. Brinkman, J. Chroma-fluids, such a blood samples, are required. Although togr. 500 (1990) 453.it was thought that very short separation runs could [12] R . Lombardi, LC?GC Int. Suppl. 11 (September) (1998) 28.

[13] N .C. van de Merbel, J. Chromatogr. A 856 (1999) 55.be used it has now been found that more convention-[14] B . Johnson, Scientist 13 (1999) 17.al separations may be needed as these will separate[15] A .J. Handley (Ed.), Extraction Methods in Organic Analysis,

analytes from endogenous interferents. High res- Sheffield Academic Press, Sheffield, 1999.olution LC–MS and other approaches such as ma- [16] V . Camel, A. Tambute, M. Caude, J. Chromatogr. A 693trix-assisted laser desorption ionisation time-of-flight (1995) 101.

R.M. Smith / J. Chromatogr. A 1000 (2003) 3–27 25

[17] M .D. Burford, S.B. Hawthorne, D.J. Miller, Anal. Chem. 65 [47] S .A. Westwood (Ed.), Supercritical Fluid Extraction and its(1993) 1497. Use in Chromatographic Sample Preparation, Blackie, Lon-

[18] S .B. Hawthorne, C.B. Grabanski, E. Martin, D.J. Miller, J. don, 1993.´Chromatogr. A 892 (2000) 421. [48] M .D. Luque de Castro, M. Valcarcel, M.T. Tena, Analytical

[19] J .M. Levy, R.A. Cavalier, T.N. Bosch, A.F. Rynaski, W.E. Supercritical Fluid Extraction, Springer, Berlin, 1994.Huhak, J. Chromatogr. Sci. 27 (1989) 341. [49] B . Wenclawiak (Ed.), Analysis With Supercritical Fluids:

[20] B .E. Richter, B.A. Jones, J.L. Ezzell, N.L. Porter, N. Extraction and Chromatography, Springer, Berlin, 1992.Avdalovic, C. Pohl, Anal. Chem. 68 (1996) 1033. [50] R .M. Smith, J. Chromatogr. A 856 (1999) 83.

[21] M .M. Schantz, J.J. Nichols, S.A. Wise, Anal. Chem. 69 [51] D .E. Raynie, Anal. Chem. 65 (1993) 3127.(1997) 4210. [52] M . Ude, M. Ashraf-Khorassani, L.T. Taylor, Chromato-

[22] O .P. Heemken, N. Theobald, B.W. Wenclawiak, Anal. Chem. graphia 55 (2002) 743.69 (1997) 2171. [53] G .A. MacKay, R.M. Smith, Analyst 118 (1993) 741.

[23] W .C. Brumley, E. Latorre, V. Kelliher, A. Marcus, D.E. [54] R .M. Smith, M.D. Burford, J. Chromatogr. Sci. 32 (1994)Knowles, J. Liq. Chromatogr. 21 (1998) 1199. 265.

[24] L . Wennrich, P. Popp, M. Moder, Anal. Chem. 72 (2000) [55] J .M. Levy, E. Storozynsky, M. Ashraf-Khorassani, in: Super-546. critical Fluid Technology, ACS Symposium Series, No. 488,

[25] B . Benthin, H. Danz, M. Hamburger, J. Chromatogr. A 837 1992, p. 336.(1999) 211. [56] Q .Y. Lang, C.M. Wai, Talanta 53 (2001) 771.

[26] E .-S. Ong, S.-O. Woo, Y.-L. Yong, J. Chromatogr. A 904 [57] R .M. Smith, J. Chromatogr. A 975 (2002) 31.(2000) 57. [58] B . van Bavel, K. Hartonen, C. Rappe, M.-L. Riekkola,

[27] H . Giergielewicz-Mozajska, L. Dabrowski, J. Namiesnik, Analyst 124 (1999) 1351.Crit. Rev. Anal. Chem. 31 (2001) 149. [59] C . Crescenzi, G. D’Ascenzo, A. Di Corcia, M. Nazzari, S.

[28] V . Camel, Trends Anal. Chem. 19 (2000) 229. Marchese, R. Samperi, Anal. Chem. 71 (1999) 2157.¨ ´ ´[29] A . Kreisselmeier, H.W. Durbeck, J. Chromatogr. A 775 [60] A . Kubatova, A.J.M. Lagadec, D.J. Miller, S.B. Hawthorne,

(1997) 187. Flavour Frag. J. 16 (2001) 64.[30] E . Conte, R. Milani, G. Morali, F. Abballe, J. Chromatogr. A [61] L . Wennrich, B. Popp, J. Breuste, Chromatographia 53

765 (1997) 121. (2001) S380.¨[31] J . Porschmann, J. Plugge, R. Toth, J. Chromatogr. A 909 [62] B . Li, Y. Yang, Y. Gan, C.D. Eaton, P. He, A.D. Jones, J.

(2001) 95. Chromatogr. A 873 (2000) 175.¨ ¨[32] P ressurised Fluid Extraction (PFE), EPA Method 3545, US [63] K . Kuosmanen, T. Hyotylainen, K. Hartonen, M.-L. Riek-

Environmental Protection Agency, Washington, DC, 1996. kola, J. Chromatogr. A 943 (2002) 113.¨ ¨[33] V . Lopez-Avila, in: D. Barcelo (Ed.), Sample Handling and [64] T . Hyotylainen, T. Andersson, K. Hartonen, K. Kuosmanen,

Trace Analysis of Pollutants, Elsevier, Amsterdam, 1999. M.-L. Riekkola, Anal. Chem. 72 (2000) 3070.[34] J .R. Dean, I.J. Barnabas, I.A. Fowlis, Anal. Proc. 32 (1995) [65] R . Tajuddin, R.M. Smith, Analyst 127 (2002) 883.

305. [66] S .A. Barker, J. Chromatogr. A 880 (2000) 63.´ ´[35] O . Lopez de Sabando, Z. Gomez de Balugera, M.A. [67] S .A. Barker, J. Chromatogr. A 885 (2000) 115.

´Goicolea, E. Rodrıguez, M.C. Sampedro, R.J. Barrio, Chro- [68] C .C. Walker, H.M. Lott, S.A. Barker, J. Chromatogr. 642matographia 55 (2002) 667. (1993) 225.

[36] J .C. Young, J. Agric. Food Chem. 43 (1995) 2904. [69] L .V. Walker, J.R. Walsh, J.J. Webber, J. Chromatogr. 595[37] M .D. Luque de Castro, M.M. Jimenez-Carmona, V. Fernan- (1992) 179.

dez-Perez, Trends Anal. Chem. 18 (1999) 708. [70] C . Crescenzi, S. Bayoudh, P.A.G. Cormack, T. Klein, K.¨[38] M . Ericsson, A. Colmsjo, J. Chromatogr. A 964 (2002) 11. Ensing, Anal. Chem. 73 (2001) 2171.

[39] S .P. Frost, J.R. Dean, K.P. Evans, K. Harradine, C. Cary, [71] K .I. Eller, S.J. Lehotay, Analyst 122 (1997) 429.M.H.I. Comber, Analyst 122 (1997) 895. [72] T .P. Wampler, J. Chromatogr. A 842 (1999) 207.

[40] B . Enders, G. Schwedt, J. Prakt. Chem.-Chem.-Ztg. 339 [73] W .J. Irwin, Analytical Pyrolysis, Marcel Dekker, New York,(1997) 250. 1982.

[41] E . Eljarrat, J. Caixach, J. Rivera, Chemosphere 36 (1998) [74] S .A. Leibman, E.J. Levy, Pyrolysis and GC in Polymer2359. Analysis, Marcel Dekker, New York, 1984.

[42] A . Egizabal, O. Zuloaga, N. Etxebarria, L.A. Fernandez, J.M. [75] F .C.Y. Wang, A.D. Burleson, J. Chromatogr. A 833 (1999)Madariaga, Analyst 123 (1998) 1679. 111.

[43] V . Camel, Analyst 126 (2001) 1182. [76] B .K. Kochanowski, S.L. Morgan, J. Chromatogr. Sci. 38[44] H .J. Vandenburg, A.A. Clifford, K.D. Bartle, J. Carroll, I.D. (2000) 100.

Newton, Analyst 124 (1999) 397. [77] S .A. Buckley, A.W. Stott, R.P. Evershed, Analyst 124 (1999)[45] C . Sanchez, M. Ericsson, H. Carlsson, A. Colmsjo, E. 443.

Dyremark, J. Chromatogr. A 957 (2002) 227. [78] G . Chiavari, D. Fabbri, S. Prati, Chromatographia 55 (2002)[46] M .A. McHugh, V.J. Krukonis, Supercritical Fluid Extraction: 611.

Principles and Practice, 2nd ed., Butterworths, London, [79] D .M. Price, M. Reading, R.M. Smith, H.M. Pollock, A.1994. Hammiche, J. Therm. Anal. Cal. 64 (2001) 309.

26 R.M. Smith / J. Chromatogr. A 1000 (2003) 3–27

[80] C .M. Cuppett, P.M. Findeis, J.C. Klotz, L.A. Woods, T.G. [112] W .M. Mullett, J. Pawliszyn, Anal. Chem. 74 (2002) 1081.Strein, LC?GC N. Am. 17 (1999) 532. [113] E . Baltussen, P. Sandra, F. David, C. Cramers, J. Microcol.

´[81] M .S. Perez-Coello, J. Sanz, M.D. Cabezudo, J. Chromatogr. Sep. 11 (1999) 737.A 778 (1997) 427. [114] P . Popp, C. Bauer, L. Wennrich, Anal. Chim. Acta 436

ˇ[82] I . Liska, J. Chromatogr. A 885 (2000) 3. (2001) 1.[83] C .F. Poole, A.D. Gunatillekam, R. Sethuraman, J. Chroma- [115] B . Kolahgar, A. Hoffmann, A.C. Heiden, J. Chromatogr. A

togr. A 885 (2000) 17. 963 (2002) 225.[84] E .M. Thurman, M.S. Mills, Solid-Phase Extraction: Princi- [116] P . Sandra, B. Tienpont, J. Vercammen, A. Tredoux, T.

ples and Practice, Wiley, New York, 1998. Sandra, F. David, J. Chromatogr. A 928 (2001) 117.[85] J .S. Fritz, Analytical Solid-Phase Extraction, Wiley–VCH, [117] T . Benijts, J. Vercammen, R. Dams, H.P. Tuan, W. Lambert,

New York, 1999. P. Sandra, J. Chromatogr. B 755 (2001) 137.[86] N .J.K. Simpson (Ed.), Solid-Phase Extraction. Principles, [118] L . Wennrich, P. Popp, G. Koller, J. Breuste, J. AOAC Int.

Techniques and Applications, Marcel Dekker, New York, 84 (2001) 1194.2000. [119] R .D. Noble, S.A. Stern, Membrane Separations Technolo-

[87] J .S. Fritz, J.J. Masso, J. Chromatogr. A 909 (2001) 79. gy. Principles and Applications, Elsevier, Amsterdam,1995.[88] A . Asperger, J. Efer, T. Koal, W. Engelwald, J. Chromatogr.

¨A 960 (2002) 109. [120] J .A. Jonsson, L. Mathiasson, Adv. Chromatogr. 41 (2001)53.[89] G .D. Owens, R.J. Eckstein, Anal. Chem. 54 (1982) 2347.

[121] M . Gilar, E.S.P. Bouvier, B.J. Compton, J. Chromatogr. A[90] G .J. Schmidt, M.W. Dong, M. Salit, J. Chromatogr. Sci. 25909 (2001) 111.(1987) 453.

¨[122] J .A. Jonsson, L. Mathiasson, J. Sep. Sci. 24 (2001) 495.[91] D . Turnell, J.R.H. Cooper, B. Green, G. Hughes, D.R.Wright, Clin. Chem. 34 (1988) 1816. [123] M .J. Yang, S. Harms, Y.Z. Luo, J. Pawliszyn, Anal. Chem.

66 (1994) 1339.[92] D .R. Lachno, N. Patel, M.L. Rose, M.H. Yacoub, J. Chroma-¨ ¨togr. 525 (1990) 123. [124] T . Hyotylainen, M.L. Riekkola, LC?GC Eur. 15 (2002) 298.

[93] A . Covaci, P. Schepens, Anal. Lett. 34 (2001) 1449. [125] N . Torto, J. Mwatseteza, T. Laurell, LC?GC Eur. 14 (2001)536.[94] J .J. Vreuls, A.J.H. Louter, U.A.Th. Brinkman, J. Chromatogr.

A 856 (1999) 279. [126] T . Niwa, W. Maruyama, D. Nakahara, N. Takeda, H.Yoshizumi, A. Tatematsu, A. Takahashi, P. Dostert, M.[95] A . Gentili, D. Perret, S. Marchese, R. Mastropasqua, R.Naoi, T. Nagatsu, J. Chromatogr. 578 (1992) 109.Curini, A. Di Corcia, Chromatographia 56 (2002) 25.

[127] W .C. Tseng, M.H. Yang, T.P. Chen, Y.L. Huang, Analyst[96] A .J.H. Louter, C.A. van Beekvelt, P. Cid Montanes, J.127 (2002) 560.Slobodnik, J.J. Vreuls, U.A.Th. Brinkman, J. Chromatogr. A

725 (1996) 67. [128] T . Ligor, B. Buszewski, Chromatographia 51 (2000) S279.[97] D . Barcelo, M.C. Hennion, Trace Determination of Pes- [129] B . Buszewski, T. Ligor, LC?GC Eur. 15 (2002) 92.

ticides and Their Degradation Product in Water, Elsevier, [130] G . Matz, G. Kibelka, J. Dahl, F. Lennemann, J. Chroma-Amsterdam, 1997. togr. A 830 (1999) 365.

[98] A .C. Hogenboom, W.M.A. Niessen, U.A.Th. Brinkman, J. [131] C . Gerbersmann, R. Lobinski, F.C. Adams, Anal. Chim.Sep. Sci. 24 (2001) 331. Acta 316 (1995) 93.

[99] J .R. Bone, R.M. Smith, Anal. Commun. 36 (1999) 375. [132] E . Baltussen, C.A. Cramers, P.J.F. Sandra, Anal. Bioanal.[100] H . Lord, J. Pawliszyn, J. Chromatogr. A 885 (2000) 153. Chem. 373 (2002) 3.

¨[101] Z .Y. Zhang, M.J. Yang, J. Pawliszyn, Anal. Chem. 66 [133] J . Hollender, F. Sandner, M. Moller, W. Dott, J. Chroma-(1994) A844. togr. A 962 (2002) 175.

[102] J . Chen, J.B. Pawliszyn, Anal. Chem. 67 (1995) 2530. [134] B . Kolb, L.S. Ettre, Static Headspace-Gas Chromatography:Theory and Practice, Wiley, New York, 1997.[103] J . Pawilszyn, Solid-Phase Microextraction—Theory and

Practice, Wiley, New York, 1997. [135] B .V. Ioffe, A.G. Vitenberg, I.A. Maratov, Head-SpaceAnalysis and Related Methods in Gas Chromatography,[104] J . Pawilszyn (Ed.), Applications of Solid-Phase Microex-Wiley, New York, 1984.traction, Royal Society of Chemistry, Cambridge, 1999.

[105] S .A. Sheppers Wercinski, Solid-Phase Microextraction: A [136] N .H. Nicholas, H. Snow, Trends Anal. Chem. 21 (2002)Practical Guide, Marcel Dekker, New York, 1999. 608.

[106] P .G. Hill, R.M. Smith, J. Chromatogr. A 872 (2000) 203. [137] R .S.T. Linforth, I. Savary, B. Pattenden, A.J. Taylor, J. Sci.[107] L . Pan, J. Pawliszyn, Anal. Chem. 69 (1997) 196. Food Agric. 65 (1994) 241.

´[108] F . Bianchi, M. Careri, C. Mucchino, M. Musci, Chromato- [138] J . Pino, M.P. Marti, M. Mestres, J. Perez, O. Busto, J.graphia 55 (2002) 595. Guasch, J. Chromatogr. A 954 (2002) 51.

´ ´[109] Y . Gou, J. Pawliszyn, Anal. Chem. 72 (2000) 2774. [139] M .C. Dıaz-Maroto, M.S. Perez-Coello, M.D. Cabezudo,[110] B .C.D. Tan, P.J. Marriott, H.K. Lee, P.D. Morrison, Analyst Chromatographia 55 (2002) 723.

124 (1999) 651. [140] A .S.R. Machado, E.G. de Azevedo, M. Nunes da Ponte,[111] W .M. Mullett, K. Levsen, D. Lubda, J. Pawlisyzn, J. R.M.A. Sardinha, J. Essent. Oil Res. 5 (1993) 645.

Chromatogr. A 963 (2002) 325. [141] M . Nedjma, A. Maujean, J. Chromatogr. A 704 (1995) 495.

R.M. Smith / J. Chromatogr. A 1000 (2003) 3–27 27

[142] M .E. Miller, J.D. Stuart, Anal. Chem. 71 (1999) 23. [167] L .A. Holland, N.P. Chetwyn, M.D. Perkins, S.M. Lunte,[143] B . Kolb, Chromatographia 15 (1982) 587. Pharm. Res. 14 (1997) 372.[144] B . Kolb, L.S. Ettre, Chromatographia 32 (1991) 505. [168] P . Bailon, G.K. Ehrlich, W.J. Fung, W. Berthold (Eds.),[145] J . Ai, Anal. Chem. 70 (1998) 4822. Affinity Chromatography, Humana, Totowa, NJ, 2000.[146] M . Llompart, K. Li, M. Fingas, J. Chromatogr. A 824 [169] P . Matejtschuk, Affinity Separations: A Practical Approach,

(1998) 53. Oxford University Press, Oxford, 1997.[147] B . Zygmunt, A. Jastrzebska, J. Namiesnik, Crit. Rev. Anal. [170] T .M. Phillips, B.F. Dickens, Affinity and Immunoaffinity

Chem. 31 (2001) 1. Purification Techniques, Eaton Press, Natick, MA, 2000.¨[148] A . Steffen, J. Pawliszyn, Anal. Commun. 33 (1996) 129. [171] J . Dalluge, T. Hankemeier, R.J.J. Vreuls, U.A.Th. Brink-

´[149] O . Ezquerro, B. Pons, M.T. Tena, J. Chromatogr. A 963 man, J. Chromatogr. A 830 (1999) 377.(2002) 381. [172] V .T. Remcho, Z.J. Tan, Anal. Chem. 71 (1999) 248A.

ˇ´ ´ ´[150] E . Matisova, M. Medved’ova, J. Vraniakova, P. Simon, J. [173] L .I. Andersson, J. Chromatogr. B 739 (2000) 163.´ ´Chromatogr. A 960 (2002) 159. [174] N . Masque, R.M. Marce, F. Borrull, Trends Anal. Chem. 20

[151] M .-C. Wei, J.F. Jen, Chromatographia 55 (2002) 701. (2001) 477.[152] M .J. Yang, M. Adams, J. Pawliszyn, Anal. Chem. 68 (1996) [175] G . Theodoridis, P. Manesiotis, J. Chromatogr. A 948 (2002)

2782. 163.[153] L . Yago, Am. Lab. 17 (October) (1985) 17. [176] T . Zhang, F. Liu, J. Wang, N. Li, K. Li, Chromatographia[154] A .N. Papas, M.Y. Alpert, S.M. Marchese, J.W. Fitzgerald, 55 (2002) 447.

M.F. Delaney, Anal. Chem. 57 (1985) 1408. [177] C .C. Hwang, W.C. Lee, J. Chromatogr. A 962 (2002) 69.[155] K . Suto, S. Kakinuma, Y. Ito, K. Sagara, H. Iwasaki, H. [178] A . Molinelli, R. Weiss, B. Mizaikoff, J. Agric. Food Chem.

Itokawa, J. Chromatogr. A 810 (1998) 252. 50 (2002) 1804.´ ´[156] C . Mardones, A. Rıos, M.Valcarcel, Anal. Chem. 72 (2000) [179] I .H. Hageston, T.C. Pinkerton, Anal. Chem. 57 (1985)

5736. 1757.[157] M . Papagiannopoulos, B. Zimmermann, A. Mellenthin, M. [180] J .A. Perry, L.J. Glunz, T.J. Szczerba, R.D. Rateike, LC?GC

Krappe, G. Maio, R. Galensa, J. Chromatogr. A 958 (2002) Int. 3 (1990) 10.9. [181] J . Haginaka, Trends Anal. Chem. 10 (1991) 17.

[158] K . Grob, M. Biedermann, J. Chromatogr. A 750 (1996) 11. [182] J . Hermansson, A. Grahn, I. Hermansson, J. Chromatogr. A[159] J . Teske, Trends Anal. Chem. 21 (2002) 584. 797 (1998) 251.[160] J . Beltran, F.J. Lopez, M. Forcada, F. Hernandez, Anal. [183] M . Beth, K.K. Unger, M.P. Tsyurupa, V.A. Davankov,

Chim. Acta 356 (1997) 125. Chromatographia 36 (1993) 351.[161] E .W.J. Hooijschuur, C.E. Kientz, A.G. Hulst, U.A.Th. [184] D .R. Knapp, Handbook of Analytical Derivatization Re-

Brinkman, Anal. Chem. 72 (2000) 1199. actions, Wiley, New York, 1979.[162] B . Hauser, P. Popp, E. Kleine-Benne, J. Chromatogr. A 963 [185] G . Lunn, L.C. Hellwig, Handbook of Derivatization Re-

(2002) 27. actions for HPLC, Wiley, New York, 1998.[163] M . Mondello, A.C. Lewis, K.D. Bartle, Multidimensional [186] T . Toyo’oka, Modern Derivatisation Methods for Sepa-

Chromatography, Wiley, New York, 2001. ration Science, Wiley, New York, 2002.[164] R . Ong, P. Marriott, P. Morrison, P. Haglund, J. Chroma- [187] F . Klink, LC?GC Eur. 13 (2000) 396.

togr. A 962 (2002) 135. [188] N .C. van de Merbel, Chromatographia 55 (2002) S53.[165] P . Marriott, R. Shellie, Trends Anal. Chem. 21 (2002) 573. [189] D iscussion forum, Chromatographia 55 (2002) S199.[166] P .J. Schoenmakers, J.L.M.M. Oomen, J. Blomberg, W.

Genuit, G. van Velzen, J. Chromatogr. A 892 (2000) 29.