Chern. Anal., (Warsaw), 43, 787 (1998)

Flow Injection Determination of Pesticides

by C. Gomez Benito· and J. Martinez Calatayud2*

Review

lDepartamento deQufmica. Universidad Politecnica. Campus de Gandia. Gandia (Valencia) Spain2Departamento de Qufmica Analftica, Universidad de Valencia,

Edificio Seminario 46113 Moncada (Valencia) Spain

Key words: pesticides, flow injection analysis (FIA)

A general view about the determination of pesticides in the FIA methodology ispresented. The review is dealing with the different published analytical proceduresdivided into three groups: a) homogeneous systems (only solutions) in which thereported procedures are classified according to the used detector; b) heterogeneoussystems (solid-liquid or liquid-liquid) ; and c) a third group is presented with the papersdealing with the FIA assembly as the sample pretreatment for the injection into achromatograph. A table with the most relevant analytical information ofeach of reportedprocedure is also presented.

Przedstawiono przegl'ld zastosowan analizy przeplywowo-wstrzykowej (FIA) do ozna­czania pestycyd6w. Praca obejmuje r6ine post~powaniaanalityczne, kt6re podzielonona trzy grupy: a) uklady jednorodne (roztwory), w kt6rych post~powania analityczneklasyfikowano wedlug uiytego detektora; b) ukhidy niejednorodne (cialo stale":" cieczlub ciecz - ciecz); c) zastosowania FIA w celu wprowadzenia pr6bki do zloia chromato­graficznego. W tablicy przedstawiqnq najistotniejsze informacje, 0 kaidym post~powa­

niu analitycznym.

PESTICIDES. DEFINITION AND FUNDAMENTALS

A pesticideis any chemical used to fight a pest (an anim~l or vegetable parasitethreatening agricultural crops, cattle or human health). Such a broad definition, by

* COll"esponding author, e-mailjose.Martfnez@uv.es, Fax: 3463692684

788 C. Gomez B.enito and J.Martinez Calatayud

tacit convention, excludes drugs and veterinary products. Hence, a pesticide can bedefined as a chemical used to fight external (animal and vegetable) parasites thatattack crops.

The FAOIWHO define the word "pesticide" as any substance or mixture ofsubstances used to prevent and control growth of undesirable vegetable or animalspecies, including any single or mixed substance used for plant growth control, as adefoliant or as a desiccant. In an explanatory note, the term "pesticide" is said toencompass any substance used to control "pests" during production, storage, trans­port, marketing and processing of foods for human or animal consumption, as wellas any substance that can be administered to animals in order to prevent growth ofinsects and arachnids on their bodies. Therefore, the word "pesticide" is not used torefer to antibiotics, nor to any other chemicals given to animals for purposes otherthan those previously stated (e.g. to foster growth or improve reproductive behaviour)or used as fertilizers.

Based on the above definition, pesticides include the following groups: insec­ticides used against insects, acaricides against mites, nematocides against nematodes,molluscocides against snails and slugs, rodenticides against rodents, fungicides oranticryptogamic agents against fungal crop parasites, general and specific herbicidesagainst weeds, farming antibiotics against crop bacterioses and soil disinfectants,which fight nematodes, insects, pathogenic fungi and weeds present in cultured soils.

CLASSIFICATION

Pesticides can be classified according to various criteria, namely:(1) Chemical composition:

(a) Inorganic pesticides, which include sulphur, arsenic, tluorine, copper andchlorine containing compounds, and various other inorganic substances.

(b) Natural organic pesticides, typically coumarins, natural pyrethrins and plantextracts.

(c) Synthetic organic pesticides such as organochlorine and organophosphoruscompounds; carbamates anddithiocarbamates; organic cyanides; organic sulpho­cyanic esters; bipyridyl salts; aromatic nitro compounds; organic acid derivatives;urea, triazine and uracil derivatives as well as other heterocycles; synthetic pyre­throids, etc.(2) Use:

According to Spanish Technical Health Regulations, officially· enforced by theRoyal Decree 334911983, pesticides' can be used for phytosanitary, cattle-protective,food industrial, environmental, personal hygiene and household purposes.(3) Hazards:

Depending on the s~verity of the risks associated with their inhalation, ingestionand dermal contact, pesticides can be classified as:

(a) Scarcely hazardous, which pose no appreciable risks.(b) Noxious, which may pose moderately serious risks.

Determination a/pesticides 789

(c) Toxic, which can lead to severe damage (acute or chronic) or even death.(d) Very toxic, which pose 'extremely serious hazards of the acute or chroniC

type, or even result in death.(4) Formulation:

In their broadest sense, agricultural pesticides are applied to crops in two differentforms, viz. as pure chemicals or, more frequently, informulations (i.e. in mixturesintended to boost their efficiency). ,Pesticide formulations ,are either solid (wettable,soluble or sprinklingpowder, granules) or liquid (water-soluble or emulsion-formingproducts, or fluid suspensions).

TOXICITY AND ENVIRONMENTAL IMPLICATIONS

Because most pesticides are highly persistent, their extensive use has led to agradual build-up in the environment, which has started to feel its effects. Sincepesticides ate usually applied to crops, they are initially found mostly on plants andsoil, and in air. Subsequently, rain and irrigation water sweep most' of these com­pounds to waterways, aquifers, etc. The presence of pesticides in the aquatic envi­ronment decreases the quality' of Water, which acquires an odd colour and/or taste,and has various indirect toxic effects on many species.

Many countries have issued maximum allowed concentrations of pesticides overthe past few years. EEC directive 76/464 has compiled a black and a red list thatinclude pesticides and other toxic substances for ground water. European laws accepta maximum concentration of 0.1 J.lg 1-1 for any single pesticide and a maximumoverall pesticide concentration of0.5 J.lg 1-1 in drinking water. By contrast, USA lawshave established a specific allowable level for each individual pesticide and even forsome by-products, which are occasionally more toxic than the parent compounds.This appears more reasonable than the fixed levels established by the EEC sinceconcentrations above 0.1 J.lg 1-1 of some, pesticides are harmless and their analysisdoes not call for extremely sensitive methods [1].

Pesticide residues must be removed from the environment and their effects on itlessened. The so-called "persistent" or "refractory" pesticides cannot be degradednaturally (by biological processes, interaction with soil substances or the action ofsunlight), so they require more powerful methods such as treatment with activesludge, chemical oxidation,ozonization, photodegradation, etc. [2]. Pesticide build-upgradually leads to environmental deterioration reflected, for example, in the loss ofnatural resources such as genetic bio-diversity, drinking water, fertile soil, etc. Thisin turn decreases Ii ving standards. The combined effects of pesticides result in higheconomic losses through both the decay of resources (e.g. tourist resorts) and theneed to invest in repairing the damage.

The persistence and toxicity of pesticides calls for their strict, exhaustive controlby useof sensitive methods affording expeditious processing of large numbers ofsamples. This paper reviews reported FIA analytical methods for the determinationof pesticides. Applications are dealt with in two groups according to the type of

790 C. Gomez Benito and J. Martinez Calatayud

chemical system used: homogeneous (single-phase) or heterogeneous (multi-phase).The former are in turn exposed according to the type of detector used.

HOMOGENEOUS SYSTEMS

Detectors based on radiant energy-matter interactions

UV-Visible" detection. The pesticide paraquat was determined spectrophotome­trically by reduction to a blue semi-quinoid radical of Arnax =600' nm using twodifferent methods. In one [3], the pesticide was reduced with alkaline sodiumdithionite. The sample was injected into a water stream that was subsequently mergedwith one of 0.1 mol I-I NaOH containing a 1% concentration of the reductant. Thedetermination of paraquat in mixed formulations is interfered by the presence ofdiquat, which gives a greenish radical of Arnax = 605 nm. In order to avoid its adverseeffect, diquat is removed from the sample by precipitation with sodium hydroxidebefore injection. Determining paraquat in waters poses nospecial problem. In foods(potatoes), it requires preconcentration on a column packed with.a cation-exchangeresin and elution with an NaCl saturated solution.

The, other method for paraquat [4] includes reduction of the pesticide with asolution of ascorbic acid partly oxidized to dehydroascorbic acid with potassiumiodate. According to these authors, this solution remains stable for longer (up to 5days) than one of ascorbic acid Qr sodium dithionite. The idea of pre-oxidizing theascorbic acid arose from the finding that the absorbance increased as the solutionaged. The authors checked empirically that dehydroascorbic acid plays a importantrole in the development of the blue colour.

This reaction has also been used with p~aminophenol [5-9]. In an alkalinemedium containing an oxidant, benzoquinoneimine is produced that can be coupledto a pesticide (e.g. to m-aminophenol, the primary metabolite for formethanate) byelectrophilic attack on C4 to give the corresponding blue diamine. This reaction hasbeen used in FIA spectrophotometric determinations for other types of compound(e.g. drugs).

The organophosphorus pesticide chlorpyrifos [10] can be determined by spec­trophotometric monitoring of its decomposition in an alkaline medium, where theanalyte undergoes gradual hydrolysis and a concomitant bathochromic shift in theabsorption maximum compared to the spectrum of the starting compound. Thehydrolysis reaction is started by injecting the sample into a carrier stream containingsodium hydroxide, hydrogen peroxide and ethanol. The presence of H20 2 appears toincrease the rate at which the pesticide is degraded, the process being monitored viathe increase in the absorption at 328 nm (new band) and the decrease in thechlorpyrifos band at 290 nm, using the stopped flow technique. The other factorinfluencing the decompqsition rate is temperature (45°C was found to be optimal).The calibration curve is obtained by plotting the rate of the first-order reactioninvolved against the pesticide concentration. Using kinetic methodology, the methodallows the determination of chlorpyrifos in the presence of dimethoate and endosul­fan with the need for no prior separation.

Determination ofpesticides 791

IR absorption detection. Infrared spectrophotometry is one of the less frequent­ly used detection techniques in FlA. Its inherent shortcomings (e.g. the need to useKBr pellets and its scant application in solution) have somehow deterred usage incontinuous-flow assemblies. Its application to organic compounds entails using alsoorganic solvents and hence solvent-proof rather than conventional pump tubing.The alterations required to use this type of detector in FIA include a novel type offlow-cell [11], the "circle cell", which causes light to be reflected many timesbetween the sample and a ZnSecrystal.

The earliest FIA-IR determination was reported by Curran and Collier in 1985[12]. They quantified phenyl isocyanate by using a CCl4 stream and a stainless steelpiston pump.

More recent FIA-IR determinations include those of organic drugs and thepesticide carbaryl [13, 14]. A pesticide solution in dichloromethane was injected intoa carrier stream containing the same solvent. Readings were made directly at 1747 cm-I ,

in both continuous manner or in the stopped flow mode. The flow-cell used had alight path length of 1.0 mm. This method offers the' typical advantages of IRmeasurements. The sensitivity of the method was increased by inserting a conven­tional cartridge to precQncentrate the sample by solid-liquidextractioIl" followed byelution with the dichloromethane carrier.

Fluorimetric detection. One highly sensitive method for determining diquat isbased on its reduction with alkaline sodium dithionite to a highly emis.sive greenishcation radical (see Reaction.!) [15]. The radical is more extensively conjugated thanthe non-reduced molecule and exhibits an absorption band that is much stronger andappears at a higher wavelength than that for diquat (497 vs 346 nm). The method isalso selective enough for determining the herbicide in a wide variety' of sampleswithout a prior separation. Thus, paraquat has been quantified in commerciallyavailable pesticide formulations, waters, foods (potatoes), soil,flowers, blood serum,and synthetic,and true urine samples. Use of a single-channel, FIA assembly (with asodium dithionite carrier, buffered with borax at pH 8) was precluded by theinstability of the reductant solution. Rather a two-channel manifold should be used,where the sample is injected into a borax carrier (pH 8) and subsequently mergedwith 0.2% reductant solution.

diquat

Reaction l.Reduction of diquat

stable green radical

The well-known increased quantum yield of various molecules in organizedmedia (micelles) was used for the determination of carbendazim in water samples[16]. This compound is the main metabolite of benomyl and results from its hydro­lysis, which takes place in both acid and basic media. Carbendazim is highly stablein an acid or neutral medium, but is further degraded to 2-aminobenzimidazole in

792 C. Gomez Benito and J. Martinez Calatayud

alkaline solution. Water samples containing the pesticide wen~ injected into a buf­fered carrier that was merged with a surfactant solution. The two surfactants studied,sodium dodecyl sulphate andcetyltrimethylammonium bromide (both in 0.1 mol 1-1HCl), increased the emission intensity by a factor of 2.6 and 11.6, respectively.

Chemiluminescence detection. There are very few references to FIA chemi­luminescence (CL) methods. In one [17], the emission arises from oxidation ofbis(2,4,6-trichlorophenyl)oxalate by hydrogen peroxide. The oxidation producttransfers its excited state to an "effective" fluorophore (the reaction product formedbetween the pesticide and a fluorescent reagent). Of the various reagents tested, morinexhibited a high CL intensity, which, however, was accompanied by also highbackground noise. The flavones tested behaved similarly to morin. Quinoline deri­vatives proved scarcely fluorescent and highly chemiluminescent, so they wereselected. The experimental procedure involves injecting the sample into a carrier-re­agent stream containing the fluorescent agent (quinoline), followed by merging witha buffer carrier containing imidazole nitrate at a roughly neutral pH and then with astream of hydrogen peroxide and bis(2,4,6-trichloro-phenyl)oxalate.

PESTICIDE + QUINOLINE ---- PESTICIDE - QUINOLINE

(Fluorophore)

CI-<o&lt~o>-CI + H202 - ......r~-~12 HO~O>-C,.ICI CI l~ ~J CIbis(2,4,6-trichlorophenyl)oxalate(TCPO) 1,2-dioxetane-dione

~~ ~J,C-CI I00

+ Fluorophore ----- Fluorophore* (excited state) + 2eOi

Fluorophore* ------ Light + Fluorophore

Reaction 2. Chemiluminescent procedure transferring the excited state to an effective fluorophore

A method for the determination ofparaoxon arid aldicarb relies on their inhibitoryeffect on acetylcholinesterase [18]. The choline formed from the enzymeis oxidizedby choline oxidase and the hydrogen peroxide released is determined by the luminolreaction with peroxidase. The chemical process involved is as foll<?ws:

Determination ofpesticides

acetylcholinesteraseAcetylcholine • acetate + choline

choline oxidaseCholine • betaine + H202

peroxidaseLuminol + 2H20 2 + OH- •

aminophthalate anion + N2 + 3H20 + light

Reaction 3.Chemiluminescent basis for the determination of paraoxon and aldicarb

793

(1)

(2)

(3)

The resulting decrease in chemiluminescence intensity is related to the amountof pesticide in the sample.

Atomic absorption detection. Organoarsenicalpesticides are usually deter­mined by oxidation to arsenate ion using a mixed chemical-photochemical reaction[19]. The sample isirijected into a distilled water carrier and merged with a streamof potassium sulphate; the resulting solution is irradiated with UV light from amercury lamp. The following steps involve the conversion of arsenate ion into arsineand its subsequent volatilization with the aid of a helium stream that drives it to thedetector (an atomic absorption spectrometer). According tothe authors of the method,the conversion of the organoarsenical into arsenate involves the decomposition ofpersulphate by the UV light; under these conditions, the anion gives highly reactivehydroxyl radicals that can decompose organic compounds to carbon dioxide. Theproponents claim that the light-induced decomposition of organoarsenicals to arsen­ate, water and carbon dioxide is also mediated by hydroxyl radicals.

On-line liquid-liquid separation was used for the FIA-AAS determination ofdimethoxydithiophosphate [20] (see under liquid-liquid systems).

Detectors based on electrical current-matter interactions

This group includes several amperometric determinations proposed by the sameresearch group. The earliest was reported in 1988 and involved the determination ofthe pesticides fenthion and·fenitrothion [21]. The only channel of the FIAmanifoldwas used to circulate a methanol-water 'carrier stream buffered with acetic-acetatesolution that wasdeaerated by passage through a zinc column and gaseous nitrogen;the stream also contained the sample. The electrode, glassy carbon, was ground withsandpaper prior to use. The determination was cathodic (at -0.9 V) for fenthion andanodic (at +1.3 V) for fenitrothion. Polarizing the working electrode took 30 min forthe anodic process and 15 min for the cathodic process at the potentials used.Hydrodynamic voltammograms exhibited no plateau for fenthion, so the potentialleading to the most favourable signal-to-noise ratio was selected. The limiting current

794 C. Gomez Benito and J. Martinez Calatayud

for fenitrothion reached a plateau at E = -0.9 V, above which background noise roseconsiderably. Also, below -'-"1.0 V the pesticide signal was not significantly greater.

Paraoxon (0,0-diethyl:"0-4-nitrophenylphosphate) and parathion (0, O-diethyl­-0-4-nitrophenylphosphorothioate) have also been determined electrochemically.They are used jointly in some pesticide formulations, so an effective quantitationmethod for both does not require prior separation. FIA amperometric method [22]determines paraoxon in the presence of parathion after alkaline hydrolysis of theformer and subsequent oxidation of the p-nitrophenol formed at a glassy carbonelectrode. Parathion is resolved on the basis of its different rate of hydrolysis, Thehydrolytic reaction is conducted on-line. For this purpose, the sample, in a water-alcoholsolution, is injected into a 0.40 mol 1-1 NaOH stream flowing at a rate of 1.5 ml min-1

along a 3 m line; the reaction yield is greater with a knotted reactor than with a coiled one.The resulting solution is merged with a 0.5 mol 1-1 acetic acid solution in such a waythat partly hydrolysed paraoxon reaches the detector at pH 5 and the p-nitrophenolproduced is oxidized at the glassy carbon electrode. This electrode is used inconjunction with an Ag!AgCl reference electrode and a gold counter-electrode.p-Nitrophenol is oxidized rather than reduced in order to avoid the interference of .non-hydrolyzed parathion, which cannot be oxidized electrochemically, and that ofdissolved oxygen, which thus need not be previously removed.

Parathion has also been determined in the presence of paraoxon by previouslyhydrolysing the analyte and oxidizing thep-nitrophenol produced at E = 1.2 V [23].The selective hydrolysis of the analyte is carried out outside the manifold because itis not rapid enough for FIA implementation. It is performed in the presence of Pd(II)or Hg(Il), which have a high affinity for the thiophosphate group and thus favour thehydrolysis of parathion over that of paraoxon. The determination can be carried outin the presence of p-nitrophenol by subtracting theFIAsignals obtained with andwithout the prior hydrolysis.

A FIA coulometric assembly was used for the oxidation of carbamate and ureapesticides [24]. The ensuing method uses only a few nanolitres of samples andfeatures a relative standard deviation below 1%. The flow-cell accommodates a Ptwire as the working electrode, Ag/AgCLas the working electrode and another Ptwireas the auxiliary electrode. The integration time is not critical since it is long enoughfor the whole sample to pass through the cell; however, it should not be so long thatbackground current may introduce some error. According to the authors, the coulometricefficiency of the method is close to 100% at low flow-rate and very high potential.

ZOrn et at. [25] developed an membrane-immobilized urease electrode. Ion-sensitivefield-effect transistors (ISFETs) are effective alternatives to the glass electrode as

. theycanact as enzyme-sensitive transducers (ENFET) when the biochemical reactioninvolved produces or consumes hydrogen ions. A membrane containing urease wasused to extend application of ISFETs to urea an~ pesticides. Urease catalyses thehydrolysis of urea to ammonium ion and carbon dioxide, which is monitored via pH.The membrane was prepared photolithographically. Urease activity is known to beinhibited by metal ions. This principle was recently exploited for determiningcarbofuran with a detection limit of 0.1 Jig 1-1 by use of the enzyme immobilized ona membrane.

Determination ofpesticides

HETEROGENEOUS SYSTEMS

795

Liquid-liquid systems

Dimethoxydithiophosphate (DDTP) can be determined indirectly by formationof a complex with Cu(II) that is subsequently extracted into chloroform. The amountof copper in the complex is determined in an ammonia-ammonium buffer at pH·1 0following back-extraction [20]. The method is based on an earlier, batch method that,according to the authors, was somewhat imprecise owing to the low stability of thecomplex, formed between copper and the pesticide, Cu(DDTP)2' For this reason, theyused extraction with an organic solvent followed by mineralization of the extract orback-extraction. In the manifold, the organic solvent (containing the analyte extractedfrom the sample) was merged with an aqueous stream (back-extractant) and reachedthe phase separator, where the aqueous phase, free from the organic one, was injectedwith the aid of a conventional valve into a distilled water carrier that drove it to thedetector.

buffer

organicsample

P1

Figure 1 PIA determination including liquid-liquid extraction. S :....phase segmenter; PS - phaseseparator; P1P2 -peristaltic pumps; W - waste; Ee - extraction coil; AAS ~ atomic absorptionspectrometer; R - recorder (or computer) .

The aqueous and organic phases are separated by passage through a hydrophilicmembrane held between twoteflonsheets. The results thus' obtained are comparedwith those provided by batch back-extraction and with the mineralization· of theorganic extracts. This methodology was applied to the determination of malathion ina commercially available formulation.

Solid-liquid systems

Enzymes are the most widely immobilized reagents, and also the best do­cumented in this respect. Their high selectivity and their catalytic action makes themideal for use in solid-phase reactors.

There are hundreds of reported FIA applications ofenzymes, particularlY deter­minations of substrates where the enzyme acts as a catalyst, whether in solution orin immobilized form - the more frequent by far. Less often, the analyte is the enzymeand its activity determined. Also, the inhibition of enzyme activity can be used

. for analyticalpurposes; this strategy, however;has scarcely been explored in FlA.

796 C. Gomez Benito and J. Martinez Calatayud

In an application of the latter type [26], a known amount of enzyme was incubatedwith the sample solution. The resulting activity was inversely proportional to theamount of inhibitor present in the sample. For example, the inhibition of the activityof acetylcholinesterase can be monitored photometrically according to the followingreaction (Reaction 4a): or electrochemically according to the following mechanism(Reaction 4b):

AcetylthiocholineACHE- thiocholine + acetate

Thiocholine + dithiobisnitrobenzoatethionitrobenzoate (yellow dye)

ACHEAcetylcholine - choline + acetate

CHOcholine + 202 + H20 - 2H20 2 + betaine

ACHE = Acetylcholinestenlse

CHO = Choline oxidase

Reaction 4.Determination based on the inhibitory effect of the activity of acetylcholinesterase

(a)

(b)

The measured parameter is the residual activity of the immobilized enzyme(following reaction with the sample), which entails immobilization and release of theenzyme for each new sample processed. For this purpose, acetylcholinesterase isimmobilized on magnetic particles via its amino groups by using a magnetic reactorconsisting of a permanent magnet that can be electrically offset by connection to theintegrated circuit. This method was used to determine carbofuran and malaoxon inwater samples.

Paraoxon was determined by immobilizing acetylcholinesterase on controlled­pore glass (CPG) beads [27]. The method relies on the inhibition of the enzymatichydrolysis of a-naphthyl acetate and the subsequent reaction of the a-naphtholformed with p-nitrobenzene-diazonium fluoroborate. For this purpose, the pesticideand substrate are made to reach the detector simultaneously. The FIA manifold usedincludes two injection valves for the same channel; one is used to inject the sampleinto a 0.05 mol I-I phosphate carrier containing NaCI and MgClz, while theother isemployed to inject the substrate solution into the sample stream lOs later. The samplethen travels through the reactor containing the immobilized enzyme and merges withthe reagent channel to form a product that is monitored spectrophotometrically at 500 nm.

In a sequel to the previous application, reaction and detection were integrated ina spectrophotometric flow-cell. In the next method [28], carbofuran, propoxur andcarbaryl were determined with no prior separation by hydrolysis to their corresponding

Determination ofpesticides 797

phenols (carbofuran-phenol, 2-isopropoxyphenol and I-naphthol). They werecoupled with sulfanilic acid to give coloured products that were determined jointlyon a diode array spectrophotometric detector using nine different wavelengths. Twodifferent manifolds were tested. In one, the analytical reaction was monitored whilethe sample solution was passed through the detector (i.e. in the usual way). In theother, the flow-cell was packed with CIS bonded phase beads to retain the reactionproducts, which were then eluted, thus flushing the cell and making it ready for anew sample. The chief conclusion drawn was that, of the two operating modes used(conventional FIA andintegrated reaction detection), the integrated one was signifi­cantly more sensitive.

In the other method of the sequel [29], paraquat was determined by integratingretention of the analyte on a Dowex-type cation-exchange resin and reaction in asolid-phase reactor that was placed in the flow-cell. The reaction was monitored viathe increase of absorbance. Finally, the reaction product was eluted by injecting anappropriate solvent. The analytical reaction employed was the classical reduction ofparaquat with alkaline sodium dithionite to give a blue radical. Ammonium chlorideproved to be the best reagent for eluting the reaction product retained in the flow-cell.In fact, ammonium ion was found to excel potassium, calcium, magnesium andsodium ions for this purpose. Because these ions can elute paraquat, they can alsopose some interferences. As a preventive measure, samples were treated with EDTAat pH 9 to avoid them. The method was used for'the determination of paraquat inwaters, as well as to study the pesticide adsorption in soils (clays and humic acids).For the latter purpose, a known amount of soil was placed in a beaker and suppliedwith water and paraquat. Subsequently, periodic measurements of the paraquatcontent in the supernatant solution were made. The temporal variation of paraquatadsorption by the soils always fitted an exponential curve; adsorption peaked at about6 h and was very high (ca 90% or even higher in some soils). The organic mattercontent of the soil had a marked effect on the pesticide adsorption.

Table I provides a global picture of FIA methods for the quantitation of pes­ticides; applications are grouped according to the particular pesticide determined.Also, Table I shows the chemical methods and detector types used, as well as theanalytical figures of merit of each application.

FIA assemblies for the preparation and insertionof samples into a .chromatograph

FIA..HPLC. FIA methodology arose and was initially developed as a convenient,expeditious means for transferring samples to a detector, as an interface betweensample preparation and analytical instrumental measurement. Therefore a naturaltrend is reflected in some applications to pesticides.

One determination for the organophosphorus pesticides azinphosmethyl andfenthion [43] involves on-line extraction in an FIA manifold coupled to an HPLCinstrument equipped with a spectrophotometric detector. Both pesticides are separ­ated with n-heptane, which effects partial extraction. The sample is inserted into the

Tab

le1.

Ana

lyti

cal

surv

eyo

fPIA

met

hods

for

dete

rmin

atio

nof

pest

icid

es

Ana

lyte

Che

mic

alM

etho

dD

etec

tor

Lin

eari

tyra

nge

(det

ecti

onR

epro

duci

bili

ty(%

)S

ampl

eR

efer

ence

lim

it)

(S~ple

thro

ughp

ut,h

-l )

Ald

icar

b(a)

inhi

biti

ono

fC

L(4

J.lg

rl)(

a)..

6.6(

a)so

il18

Para

oxon

(b)

acet

ylch

olin

este

rase

(0.7

5J.l

gr1

ib )3.

7(b)

vege

tabl

es(i

mm

obil

ised

)(1

5)

Atr

azin

ean

tibo

dies

F3

20

-40

4nm

0.02

-0.3

J.lg

rl

wat

er37

imm

obil

ised

Atr

azin

eau

tom

ated

quas

i-U

V-V

IS30

ngr

l ..:..l

J.lg

rl

3-2

039

cont

inuo

usim

mun

oana

lysi

s

Atr

azin

ere

acti

onW

ithtw

oF

32

0-4

04

nm1

and

30ng

rl

2.9-

8.3

wat

er42

anti

bodi

esso

ilim

mob

ilis

ed

Azi

npho

smet

hyl-

Aal

kali

nehy

drol

ysis

ampe

rom

etri

cup

96.6

J.lm

olr

l2.

234

(0.4

1J.l

mol

rl )

Car

bam

ates

inhi

biti

ono

fU

V-V

IS(0

.5-2

75pp

b)30

Org

anop

hosp

horu

sbu

tylc

holi

nest

eras

e(i

mm

obil

ised

)

Car

bam

ates

inhi

biti

ono

fpo

tent

iom

etri

c(0

5-2

75

ppb)

(4)

wat

er31

Org

anop

hosp

horu

sac

etyl

chol

lines

tera

seel

ectr

ode

pH(i

mm

obil

ised

)

Car

bary

l(a)

form

atio

nin

doph

enol

UV

-VIS

596

nm1

.2-6

mg

rl

(a)w

ater

s6

l-:n

apht

oICb

)(2

6.5

ngm

rl)C

b)(m

etab

olit

e)2

-8m

grl

(a)0.

3Cb

)(1

1.8

ngm

rl)(

b)(1

10)

Car

bary

lw

itho

utre

acti

on,

FT

IRco

ntin

uous

-flo

w0.

8co

mm

erci

alfo

rmul

atio

n13

sam

ple

inC

H2C

l 20.

25-1

0m

gm

r!(5

3)st

oppe

d-fl

ow0.

30.

125-

10m

gm

rl(9

0)(1

1J.

lgm

rl )

-l

\0 00 51 ~ ~' A O::

l~ S· § ~ :--.

. ~ ~. N ~ is" ~ "< t

Tab

le1

(con

tinu

atio

n)

Par

athi

onof

f-li

nehy

drol

ysis

,am

pero

met

ric

1.96

xlO

-7 -7.1

4x

lO-6

mo

l.-1

2.3

23de

term

inat

ion

of

glas

syca

rbon

(3.4

xlO

-8m

ol.-

1)

p-ni

trop

heno

lel

ectr

ode

Pes

tici

dem

embr

ane

wit

hF

(atr

azin

eO.0

2~g

.-1

)(4

)w

ater

40

pest

icid

esp

ecif

ic(p

ropa

zine

0.03

~g

1-1)

antib

ody,

auto

mat

edqu

asi-

cont

inuo

us

Pro

poxu

rfo

rmat

ion

UV

-VIS

600

nm(0

.1m

g.-

1)0.

1na

tura

lw

ater

s7

indo

phen

ol,o

n-li

ne(8

0)hy

drol

ysis

Zir

amco

mpl

exfo

rmat

ion

AA

S32

4.7

nm6.

0-12

~molr1

2.6

35w

ith

Cu,

(0.1

2~molr1 )

liqu

idex

trac

tion

wit

hM

IBK

1S

DS

:so

dium

dode

cyls

ulph

ate.

2C

TA

B:

cety

ltri

met

hyla

mm

oniu

mbr

omid

e.3

Rep

orte

dre

sult

sar

eco

rres

pond

ing

toac

etyl

chol

ines

tera

se.

4D

DT

P:

Dim

etho

xydi

thio

phos

phat

e.A

cert

ain

num

ber

ofo

rgan

opho

spho

rus

pest

icid

esbe

long

toth

eD

DT

Pfa

mil

y(l

ike

Mal

athi

on).

5O

rgan

otin

com

poun

ds:d

i-n-

buty

ltin

dich

lori

de(D

BT

C);

diph

enyl

tin

dich

lori

de(D

PT

C);

tri-

n-bu

tylt

inch

lori

de(T

BT

C);

trip

heny

ltin

chlo

ride

(TP

TC

).(a

),(b

),(c

)-

see

mis

cell

aneo

usap

plic

atio

ns.

~ ~ ~ I: ~ ~ ~ ~: ~ "" -J \0 \0

Tab

le1

(con

tinu

atio

n)

Dic

hlor

vos

inhi

biti

ono

fpo

tent

iom

etri

c(1

.5Jl

mol

r1 )

3.8

41ac

etyl

chol

ines

tera

seim

mob

ilis

edin

ael

ectr

ode

Diq

uat

redu

ctio

nw

ith

F4

28

-49

7nm

inba

tch

3-9

0Jl

gr

10.

82co

mm

erci

alfo

rmul

atio

n15

dith

ioni

te(0

.4Jl

g.-

1)

wat

ers

inP

IA18

-400

0Jl

gr

10.

51po

tato

es(7

0)fl

ower

sso

ilur

ine

and

seru

m

Diq

uat(

a)re

duct

ion

wit

hF

428-

497

nm(a

)0

.06

-10

Jlg

mr1

(a)(7

0)w

ater

32Pa

raqu

at(b

)di

thio

nite

UV

-VIS

605

nm(b)

0.2

-20

Jlg

mr1

(b)

food

biol

ogic

alfl

uids

Diq

uat

redu

ctio

nw

ith

UV

-VIS

(0.9

ngm

r1fo

r10

0m

l6.

3w

ater

36di

thio

nite

,sa

mpl

e)fl

ow-t

hrou

ghse

nsor

Eth

iofe

ncar

bfo

rmat

ion

indo

phen

ol,

UV

-VIS

638

nm

(33

Jlg

r1 )

0.4

.na

tura

lw

ater

s8

stop

ped-

flow

(6m

in)

i

Fent

hion

(a)

cath

odic

(a)

ampe

rom

etri

c1.

6xlO

-7-1

.6x

lO-s

mol

r1

(a)

.1.6

7(a

)21

Feni

trot

hion

(b)

anod

ic(b

)gl

assy

carb

on(1

.5xl

<r7

mol

r1 )

elec

trod

e8x

lO-6

-8xl

O-S

mo

lr1

(b)

0.87

(b)

(5.2

xlO

-7m

olr

1 )(1

37)

Form

etan

ate(

a)fo

rmat

ion

indo

phen

olU

V-V

IS57

6n

mD

-7.7

7xlO

-sm

ol1-

1(a)

0.5(

a)w

ater

s5

m-a

min

ophe

nol(

b)(1

.9x1

O-7

mol

r1 )

(a)

(met

abol

ite)

D-7

.33x

lO-s

mol

r1

(b)

1.8(

b)(5

xlO

-7m

olr

1 )(b

)

For

met

anat

ehy

drol

ysis

off-

line

UV

-VIS

576

nm0.

025

mg

r1

0.5

wat

er9

(mic

row

ave)

,rea

ctio

np-

amin

ophe

nol

00 8 !J ~ ~\ ~ ~ S· ~ ;.... ~ ;::;. ~. ~ s l

Org

anoa

rsen

ical

sox

idat

ion

toar

sena

teA

AS

0.1-

2.5

mg

.-1

As(

V)

19(K

2S20

S+

UV

)an

dre

duct

ion

toar

sine

Org

anot

in5

chem

ilum

ines

cenc

eD

BT

CD

BT

CD

BT

C17

Com

poun

ds(D

BT

C;

reac

tion

ofh

is(2

,4,6

-C

L35

2-5o

onm

(0.5

-50)

xlO

-6m

ol.-

14.

4D

PT

C;T

BT

Can

dtr

ichl

orop

heny

l)ox

alat

(0.5

xlO

-6m

ol.

-I )

TP

TC

)w

ith

H20

2is

appl

ied

tode

tect

ion

of

fluo

resc

ento

rgan

otin

-qu

inol

ine

com

plex

es

Par

aoxo

non

-lin

ehy

drol

ysis

ampe

rom

etri

cup

9~75xlO-s

mo

l.-I

2.8

22w

ithN

aOH

,det

erm

i-gl

assy

carb

on(7

.0xl

O-7

mol

.-1

)

natio

no

fp-n

itro

phen

olel

ectr

ode

.Par

aoxo

nre

acto

rw

ith

UV

-VIS

500

omco

ntin

uous

-flo

w1.

427

acet

ylch

olin

este

rase

,5x

lO-7 -1

.5xl

O-s

mo

l.-I

(60)

reac

tion

ofn

apht

ol(4

.Oxl

O-7

mo

l.-I )

form

edst

oppe

d-fl

ow0.

9lx

l0-8 -4

xlO

-7m

ol.

-I(3

0)(8

.0xl

O-9

mo

l.-I )

Par

aqua

tre

duct

ion

wit

hU

V·N

IS60

5nm

0.1

-30

mg

rlpo

tabl

ew

ater

3di

thio

nite

(80)

pota

toes

com

mer

cial

form

ulat

ion

Par

aqua

tre

duct

ion

wit

hU

V-V

IS60

0nm

0.1-

100

mg

.-I

1-2.

5po

tabl

ew

ater

4de

hydr

oasc

orbi

cac

id(2

0ng

mrl )

(asc

orbi

cllc

.oxi

dise

d(1

20)

wit

hio

date

at6

0°C

)

Par

aqua

tre

duct

ion

wit

huv

-VIS

605

nm0.

4-5.

0ng

mrl

for

250

ml

7.9

wat

er29

dith

ioni

te,

sam

ple

(0.9

)so

ilfl

ow-t

hrou

ghse

nsor

(40-

120

ngm

rlfo

r10

ml

5.3

sam

ple

(12)

~ ~ ~ ~. §. ~ l: ~ ~ 00 o -

Car

bary

lw

ithou

trea

ctio

n,FT

IR(1

S~grl)

8.6

wat

er14

sam

ple

inC

H2C

l 2,

solid

-liq

uid

extr

act

Car

bary

lre

actio

nw

ithU

V-V

ISSI

Snm

0.1-

40pp

m3.

8w

ater

33N

aN0 2

+su

lpha

nilic

O.S

ppm

form

ulat

ion

acid

Car

bend

azim

mic

elar

med

ium

Fw

ater

s16

SDS1

(a)28

3-36

8nm

(a)(3

2~g

r1 )

(a)2.

2(a)

CTA

B2

(b)30

6-32

7nm

(b)(4

~g

r1)(

b)O

.S(b

)(1

60)

Car

bofu

ran

inhi

bitio

nof

UV

-VIS

40S

om(O

.S~g

r1 )(3

)w

ater

26M

alao

xon3

enzy

mat

icel

ectr

oche

mic

alac

tivity

Car

bofu

ran(

a)hy

drol

ysis

,U

V-V

ISco

ntin

uous

-flo

w3.

61(a)

3.S(

b)2.

26(c

)w

ater

28Pr

opox

ur(b

)di

azot

ized

1-4

0m

gr1

(60)

Car

bary

l(C)

flow

-thr

ough

sens

or2.

88(a

)3.4

(b)

1.82

(c)

0.O

S-2

mg

r1(4

0)

Chl

orpy

rifo

ski

netic

dete

rmin

atio

n,U

V-V

IS(0

-:-2

x10

-4m

olr

1 )1.

8co

mm

erci

alfo

rmul

atio

n10

degr

adat

ion-

reac

tion

(6.6

x10

-6m

olr1 )

inm

ediu

mN

aOH

+H

202

o-C

reso

l(a)

hydr

olis

isw

ithN

aOH

,U

V-V

ISup

4.8~g

mrl

(a)0.

7(a)

wat

er38

m-C

reso

l(b)

reac

tion.

with

KI0

4+61

4nm

(a)(8

0~grl)

(a).

.O.

S(b)

p-am

inop

heno

l63

2nm

(b)

Up9.

6~g

mr1

(b)(8

0)(1

40~g

r1 )

(b)

DD

Tp4

com

plex

form

atio

nin

AA

S32

4.8

nm8.

S-1

7mg

r1

1.6

com

mer

cial

form

ulat

ion

20C

HC

I 3•ba

ck-e

xtra

c-(0

.39

mg

r1 )tio

nin

toN

H3I

NH

San

din

toth

eFI

Am

anif

old

Car

bam

ate

and

urea

coul

omet

ric

24de

riva

tives

~ N f')

G) ~' ~ S· ~ ;.... ~ ::t ~. ~ is'" S l

Determination ofpesticides 803

FIA system in a continuous manner and, once extracted, the organic phase is injectedin small portions at preset intervals. The FIA manifold extracts the pesticides and thechromatograph performs the separation and individual detection of the pesticides inthe extract. The method was developed for determining pesticides in waste water.

On-line separation as a preliminary step in the insertion of extract into a chroma­tographic column was performed in the determination of chlorfenvinphos, methylparathion and their respective degradation products (2,4,5-trichlorophenol and4-nitrophenol) [44]. The sample was also injected in a continuous mode, using noinjection valve, and merged with the organic extractant.

HPLC-FIA. The on-line preparation of samples of mono- di- and tributylphe­nyltin [45] involves retaining the analytes by adsorption on bonded silica withoctadecyl functional groups and isolating the analytes from the sample matrix.This separation-preconcentration step, which is followed by in situ derivatization ofthe retained pesticides by ethylation with 3% sodium tetraethylborate in water. Oncederivatized, the analytes are eluted with methanol and driven to the gas chromato­graph for separation and. subsequent microwave-induced plasma atomic emissiondetection. The method was applied to river samples.

The other natural trend. in FIA is the development of detection methods for usewith post-column derivatization in HPLC. Because the weakest part of the chroma­tograph - in terms of research developments - is the detector, this could make asuitable target for FIA results. However, this has not been the case so far. Strictlyspeaking, this is not a HPLC-FIA coupling since the latter uses no injection valveand the sample (viz. the effluents from the chromatographic column) is inserted in acontinuous manner. The organophosphorus pesticides paraoxon and di-isopropylflu­orophosphate, and the carbomates namely isopropyl-N-(3-chloromethy1) carbamateand N-phenyl carbamate were determined in binary mixtures by injecting twopesticides into a chromatographic column (reversed-phase chromatography withtetrahydro-furan-water as the mobile phase). The effluent was merged with a streamof the substrate (a-naphthyl acetate) that was circulated along one ofthe channels in theunsegmented continuous,..flow manifold. It passed through a solid-phase reactor con­taining acetylcholinesterase immobilized on controlled pore glass (CPG) beads [46].Beyond the reactor, the mixture was merged with a p-nitrobenzenediazonium fluo­roborate solution and driven to the spectrophotometric detector for measurement at500 nm. Detection relied on inhibition of the enzyme catalytic action.

Misce!laneonsapplications

This section deals with special methods that depart somehow from those de­scribed in the preceding sections.

Eight phenoxyacid herbicides (viz. MCPA; MCPP; MCPB; 2,4-D; 2.4-DP;2,4-DB; 2,4,5-T and 2,4,5-TP) plus bentazone were determined in water sampleswith thermospray tandem mass spectrometry [47]. Experimental parameters wereoptimized in two steps. In the first one, the analytes were continuously introducedinto the system and the influence of the vaporizer temperature, discharge voltageandrepeJIer voltage was studied under a wide range ofexperimental conditions.

804 C. G6mez Benito and J. Martinez Calatayud

From the different solvent mixtures and source block temperatures (vaporizer tem'­peratures) tested, 90+ 10 (v/v) ammonium acetate-acetonitrile was found to be thebest option. The procedure was fully automated and afforded detection limits as lowas 1 mg I-I; also, the calibration graph was linear up to 50 mg 1-1.

The determination of water pollutants by immunoassay is a novel research trendaimed at gradually replacing chromatographic procedures. Special emphasis in thisrespect has been placed on pesticides and, chiefly, herbicides. The reason behind thegrowing use of immunoassays is the need to detect concentrations below the micro­gram-per-litre level. Wortberg et at. combined the principles of immunoaffinitychromatography [48] with those ofFlA by immobilizing fluorescein-Iabelledhaptenson oxirane acrylic beads that constituted the affinity portion of the assembly. If thehaptens are saturated with antibodies of the Eu(lll) chelate-labelled type, then someantibodies are displaced by the analyte to an extentdependenJ on the affinity of theantibodies. The FlA manifold used is very simple: the" sample is injected into a streamresulting from the merging of the reagent and a buffer solution. Subsequently, thesample is passed through an affinity column containing the hapten derivative immo­bilized on oxirane acrylic beads and driven to a fluorimeter. The FlA cycle involvesthree steps, namely:

(a) packing of the affinity column" with the labelled antibodies and subsequentflushing;

(b) injection of the sample and halting of the flow for 20-40 min;(c) fluorimetric detection. Taking into account that loading the antibodies in the

first step can take about 15 min, the overall cycle requires 45-70 min.The method was applied to s-triar.ine herbicides. One of the methods for the

determination of herbicides relies on the photobleachingof the bacteriochlorophylldimer (P), which as a result is oxidized [4.9]. The process can be monitored via thedecrease in the absorbance at 860 nm, where the starting dimer absorbs, but itsoxidation product does not. In addition to macromolecular complexes present in thesolution, electrons can recombine directly with the oxidation product. The recombi'­nation rate varies markedly depending on the actual process that takes place. Thus,recombination with the different intermediate states can take from 1 to 100 s(lifetime), depending on the particular bonded groups or substances (ubiquinones orherbicides). The recombination can be triggered by flash excitation and monitoredvia temporal changes in the absorbance. The irradiation must be controlled in such away as to maximize differences between the bleaching levels obtained from thereplacement with herbicides or ubiquinone. This prihciplewas used to developbiosensors with reaction sites holding Rodobacter sphaeroides (from purple bac­teria), which allowed the determination of various herbicides in the batch mode [50].Subsequently, the same strategy was used in a continuous FlA assembly that wasirradiated with a He-Ne laser source located between two photodiodes whichmeasured the incoupling light.

REFERENCES

1. Barcel6 D., J. Chromatogr., 643, 117 (1993).2. Chamarro E. and Espuglas S.,Qufm.lnd., 42, 19 (1995).3. Chico Guijarro E., Yafiez Sedefio P. and Polo Dfez L.M., Anal. Chim. Acta 199,203 (1987).

Determination ofpesticides 805

4. Jain A, Venna KK. and Townshend A, Anal. Chim. Acta, 284, 275 (1993).5. Khalaf K.D., Sancen6n J. and de la Guardia M., Talanta, 40, 1173 (1993).6. Khalaf KD., Morales Rubio A and de la Guardia M., Anal. Chim. Acta, 280, 231 (1993).7. KhalafKD., Sancen6n J. and de la Guardia M:,Anal. Chim. Acta, 266, 119 (1992).8. Khalaf KD., Sancen6n J. and de la Guardia M., Fresenius J. Anal. Chem., 347,52 (1993).9. Khalaf K D., Morales Rubio A and de la Guardia M., Anal. Chim. Acta, 281,249 (1993).

10. Espinosa Mansilla A, Salinas F. and Zamora A, Talanta, 41, 651 (1994).11. McKittrick P. T., Danielson N. D. and Katon J. E., J. Liq. Chromatogr., 14,377 (1991).12. Curran D. J. and Collier W. G., Anal. Chim. Acta, 177,259 (1985)13. Gallignani M., Garrigues S., Martinez Vado A. and de la Guardia M., Analyst, 118, 1043 (1993).14. Garrigues S., Vidal T., Gallignam M. and de la GuardiaM., Analyst, 119,659 (1994).15. Perez Ruiz T., Martinez Lozano C. and Tomas V., Anal. Chim. Acta, 244,99 (1991).16~ Sancen6n J. and de la Guardia M., Anal. Chim. Acta, 287,49 (1994).17. Fujimaki T., Tani T. and Watanabe S.,Anal. Chim. Acta, 282,175 (1993).18. Roda A, Rauch P., Ferri E., Girotti S., Ghini S., Carrea G. and Bovara R,Anal, Chim. Acta, 294, 35 (1994)19. Atallah RH. and Kalm~n D.A., Talanta, 38,167 (1991).20. Jimenez de Bias 0., Pereda de PazJ.L. and Hernandez Mendez J., Talanta, 38, 857 (1991).21. Hernandez Mendez J., Carabias MartInez R, Becerro Dominguez F. and Jimenez Jimenez J.I., Anal.

Chim. Acta, 209, 205 (1988).22. Hernandez Mendez J., Carabias Martinez R, Rodriguez Gonzalo E. and Perez Trancon J., Analusis, 18,

486 (1990).23. Hernandez Mendez J., Carabias MartInez R, Rodrfguez Gonzalo E. and Perez Tranc6n J., Anal. Chim.

Acta, 228,317 (1990).24. Gordon von Nehring Q., West Hightower 1. and Anderson J.L., Anal. Chem., 58,2777 (1986).25. Zurn A, Rabolt B., Grafe M. and Muller H., Fresenius J. Anal. Chem., 349, 666 (1994).26. Kindervater R, Kunnecke W. and Schmid RD., Anal. Chim. Acta, 234, 113 (1990).27. Leon Gonzalez M.E. and Townshend A, Anal. Chim. Acta, 236, 267 (1990).28. Fernandez Band B., Linares P., Luque de Castro M.D. and Varcarcel M., Anal. Chem., 63, 1672 (1991).29. Agudo M., Rios A and Varcarcel M., Anal. Chim. Acta, 281, 103 (1993).30. Trau-Minh C., Anal. Proc., 30, 73 (1993).31. Kumaran S. and Trau-Minh C., Anal. Biochem., 200, 187 (1992).32. Perez Ruiz T., MartInez Lozano C. and Tomas V., lnt. J. Environ., Anal. Chem., 44, 243 (1991).33. Yanez Sedeno P., Nova Nova C. and Polo Diez L.M., Microchem. J., 38,370 (1988).34. Hernandez Mendez J., Carabias Martinez R, RodrIguez Gonzalo E. and Perez Tranc6n J., Electroana-

lysis 2, 487 (1990).35. Jimenez de BlasO., Pereda de la Paz J. and Hernandez Mendez 1., J. Anal. At. Spectrom., 5,693 (1990).36. Agudo M., Rios A and Varcarcel M., Quim. Anal., 12, 100 (1993).37. Kraemer P. and Schmid R, Biosens. Bioelectron., 6, 239(1991).38. Khalaf KD., Hasan B.A., Morales Rubio A. and de la Guardia M., Mikrochim. Acta, 112,99 (1993).39. Wittman C. and Schmid RD., Sens. Actuators B., 815, 119 (1993).40. Kraemer P.M. and Schmid RD., Pestic. Sci., 32, 451 (1991).41. Ivnitskii D.M. and Rishpon 1., Biosens. Bioelectron., 9, 569 (1994).42. Wittmann C. and Schmid RD., J. Agric. Food. Chem., 42, 1041 (1994).43. Farran A. and de PabloJ., Anal. Chim. Acta, 212,123 (1988).44. Farran A, Cortina 1.L., de Pablo J. and Barcel6 D., Anal. Chim. Acta, 234,119 (1990).45.Szpunar-Lobinska J., Ceulemans M., Lobinski R and Adams F.C.,Anal. Chim. Acta, 278,99 (1993).46. Leon Gonzalez M.E. and Townshend A., J: Chrom., 539,47 (1991).47. Geerdink RB. and Kienhuis P.G., J. Chrom., 647, 329 (1993).48. Wortberg M., Middendorf C., Katerkamp A., Rump T., Krause J. and Cammann K, Anal. Chim. Acta,

289,177 (1994).49. Fockers R,. Bier F. S and Schmid R D., Anal. ChimActa, 274,185 (1993)50. Fockers R, Bier F. S. and Schmid R D., Anal. Chim Acta, 280, 53 (1993).

ReceivedJune 1997Accepted March 1998