In-syringe-stirring: A novel approach for magnetic stirring-assisted dispersive liquid-liquid microextraction

ContentslistsavailableatSciVerseScienceDirect

Analytica Chimica Acta

j o ur na l ho me p a g e :w w w . e l s e v i e r . c o m / l o c a t e / a c a

In-syringe-stirring: A novel approach for magnetic stirring-assisted dispersive liquid–liquid microextraction

Burkhard Horstkotte

, Ruth Suárez

, Petr Solich

, Víctor Cerdà

aDepartmentofAnalyticalChemistry,FacultyofPharmacy,CharlesUniversity,Heyrovského1203,CZ-50005HradecKrálové,CzechRepublic

bLaboratoryofEnvironmentalAnalyticalChemistry–LQA²,UniversityoftheBalearicIslands,Cra.Valldemossakm7.5,07122PalmadeMallorca,Spain

h i g h l i g h t s

•Weproposeanewautomaticmag- netic stirring assisted dispersive liquid–liquidmicroextraction.

•Itallowstheextractionofaluminum fromseawaterandfreshwatersam- pleswithinlessthan4min.

•The method was applicableto the naturalsamples.

g r a p h i c a l a b s t r a c t

a r t i c l e i n f o

Articlehistory:

Received4February2013

Receivedinrevisedform16May2013 Accepted25May2013

Available online 3 June 2013

Keywords:

In-syringemagneticstirring-assisted liquid–liquidmicroextraction Single-dropextraction Aluminum

Seawater Lumogallion

a b s t r a c t

Forthefirsttime,theuseofamagneticstirrerwithinthesyringeofanautomatedsyringepumpand theresultingpossibleanalyticalapplicationsaredescribed.Asimpleinstrumentationfollowingroughly theonefromsequentialinjectionanalyzersystemsisusedincombinationwithanadaptor,whichis placedontothebarrelofaglasssyringe.Swirlingaroundthelongitudinalaxisofthesyringeandholding twostrongneodymiummagnets,itcausesarotatingmagneticfieldandservesasdriverforamagnetic stirringbarplacedinsideofthesyringe.

Inafirststudyitwasshownthatthisapproachleadstoasealedbutalsoautomaticallyadaptable reactionvessel,thesyringe,inwhichrapidandhomogeneousmixingofsamplewiththerequiredreagents withinshorttimecanbecarriedout.

Inasecondstudyin-a-syringemagneticstirring-assisteddispersiveliquid–liquid microextraction (MSA-DLLME)wasdemonstratedbytheapplicationoftheanalyzersystemtofluorimetricdetermination ofaluminuminseawatersamplesusinglumogallion.

Alinearworkingrangeupto1.1␮molL⁻¹andalimitofdetectionof6.1nmolL⁻¹werefound.Anaverage recoveryof106.0%wasachievedforcoastalseawaterswithareproducibilityof4.4%.Theprocedurelasted 210sincludingsyringecleaningandonly150␮Lofhexanoland4.1mLofsamplewererequired.

© 2013 Elsevier B.V. All rights reserved.

1. Introduction

Dispersiveliquid–liquidmicroextraction(DLLME)hasdrawna majorinterestfromscientistsfromdifferentanalyticaldisciplines sinceitsfirstdescriptionbyRezaeeetal.in2006[1].Thisismost

∗Correspondingauthor.Tel.:+34971173261;fax:+34971173462.

E-mailaddress:victor.cerda@uib.es(V.Cerdà).

likelyduetothepossibilityofhighextractionefficienciesandlarge enrichmentfactorswithasimpleandrapidprocedure.

DLLMEisbasedonthedispersionoftheextractionsolventinto finedroplets,whichmultipliesenormouslyitscontactsurfacewith theaqueoussampleandbythis,theextractionefficiencyforthe analyteofinterest.

Theoriginalmethodologyrequiresadispersionsolventasmajor componentoftheorganicphase,whichdissolvespreferablyinthe aqueousphaseattherapidinjectionofthesolventmixtureintothe aqueousphase.Thus,averysmallamountofextractionsolventis 0003-2670/$–seefrontmatter© 2013 Elsevier B.V. All rights reserved.

http://dx.doi.org/10.1016/j.aca.2013.05.049

effectivelydispersedintodroplets,whichafterwardsareforcedto coalescebyacentrifugationstep.Theorganicphaseisthentrans- ferredintothedetectororusedforchromatographicseparation.

However, the dispersion solvent assisted DLLME has a few inconveniences. The additional solvent leads to increase waste production. Themethod requires additional optimizationeffort (dispersionsolventquantityandkind)and,themostimportant, thedispersionsolventincreasesthesolubilityof theanalyte in theaqueousphase.Furthermore,thedistributionofthedispersion solventbetweenbothphasesandbythis,thefinalvolumeofthe organicphasedependsonthesamplesalinity.

Consequently, alternative DLLME methodologies have been developed, where extraction solvent dispersion is achieved by kinetic energy. Depending on the mode of achieving droplet formation or stabilization of the droplets in the aqueous phase, ultrasound-assisted DLLME [2], air-assisted DLLME [3], vortex-assisted DLLME [4], magnetic stirring-assisted DLLME (MSA-DLLME)[5],andsurfactantassistedDLLME[6]canbedistin- guishedamongothers.Fordetails,thereaderisreferredtorecent andextensivereviewarticlesonthistopic[7–10]).

Inspiteofthehighinterestinthedevelopmentandapplica- tionofDLLMEtechniques,thepotentialoftheirautomationusing analyticalflowtechniques(FT)[11,12]suchasSequentialInjec- tionAnalysis(SIA)[13–15]hasbeenwidelydisregarded.Direct couplingwith the intended detection technique, higher repro- ducibility,highersamplethroughput,andautomatedcleaningof theextractionvesselarepossible benefitsofFT-based automa- tion.Thesehavebeendemonstratedsuccessfully.However,only bythreedistinctautomationmodalitiessofar:

1.Extractioninflowbyconfluenceoftheaqueoussampleandan organicsolventmixturewithdropletcollectiononahydropho- bicmaterial and subsequent elutionto a detection flow cell requiringadditionalsolvent[16–18].

2.Useofanextractionvesselasabatchapproachofautomation ofthemanualDLLMEprotocol.Thesolutionhandlingisaccom- plishedbytwoseparateSIAsystems[19].

3.In-syringeDLLMEbyaspirationoftheorganicsolventmixture followedbysampleaspirationatveryhighflowratethatleadsto DLLME.Afterfloatingandcoalescingofthesolventdroplets,the organicphaseisexpelledintothedetectionflowcell[20–24].

Uptodate,therearehardlyanyworksonFT-basedautoma- tionofDLLME,inwhichthedispersionsolventwasomittedleaving aloneanautomationapproachofair-assistedDLLME[25].Tothe bestofourknowledge,thepresentworkreportsthefirstFT-based automationofMSA-DLLME.Itisbasedonanovelapproachusing amagneticstirringbarwithinthesyringepumpofaSIAsystem.

Hence,asealedbutadaptablereactionvesselisobtained,inwhich allsolutionscanbeaspiratedwithhighprecisionandmixedhomo- geneouslyandnearlyinstantaneous.Ifairandanextractionsolvent lighterthanwaterareused,vortexformationwillallowthecontact oftheextractionsolventwiththeturningstirringbarandhereby, thedispersionofthesolventintofinedroplets.Stoppingthestir- ringallowsthendropletfloatation,coalescence,andexpulsionof theextractionsolventintoadetectionflowcell.

Thesystem wasused for theextractionof aluminum (Al³⁺) aslumogallion(LMG)complexfromseawatersamples.Thisalso allowed a critical comparison with a similar application but based on in-syringe dispersion solvent-assisted DLLME, which wasreportedrecently[23].In both works,LMGwaschosen as averyselectivefluorescencereagentforaluminum[26].Incon- trasttotheoften-usedmorin,theLMG-Alcomplexisextractable intomoderatelyhydrophobicorganicsolvents. It furthershows low interference from sample matrix or other cations and has

thereforebeensuccessfullyusedinoceanographicresearchover aboutthreedecades[27–32].

Althoughaluminum is a non-essential element, its determi- nationinseawaterisofinterestasconcentrationdataallowthe calculationofatmosphericdepositionofdustparticlesontheocean surfaceduetoitspresenceinnumerousminerals.Then,thesecalcu- lationsallowtheestimationoftheentryofessentialtracenutrients suchasironoriginatingfromthedissolutionofthedustparticles andwhicharelimitationfactorsforgrowthofalgae.

Herein, in-syringe MSA-DLLME is presented. The improve- mentof existinganalyzersystems foraluminum aswellasthe demonstrationandapplicationofanovelextractiontechniquewas intendedwiththecriticaldiscussionofitsshortcomingsandpoten- tialbenefitsforfutureworks.

2. Methodsandmaterials

2.1. Reagents

All chemicals were of reagent grade for analysis and ultra- purewater(resistivity>18McmMilliporeIbericaS.A.U.,Madrid, Spain)wasusedthroughout.Allglasswareandpolyethylenebot- tleswerepreviouslysoakedin 10%(v/v)HNO₃ and rinsedwith ultrapurewaterpriortouse.Allworkingsolutionswerestoredin polyethylenebottlesat4^◦Cinthedarkwhennotused.

Analuminumstock solutionof13.5mgL⁻¹ waspreparedby dilutingacommercial1000mgL⁻¹Al(NO₃)₃·9H₂Oatomicabsorp- tion standard (Scharlab, Barcelona, Spain) in 0.5molL⁻¹ HNO3. Syntheticseawater(SSW)preparedaccordingtostandardrecipeas givenelsewhere[33]wasusedformostoptimizationexperiments andforstandardpreparation.Toeliminatealuminumcontamina- tionoftheSSW,theformedAl(OH)3 attheslightlyalkalinepH oftheSSW(pH8)wasremovedbyfiltrationthrougha0.45␮m membranefilter.

AcidificationwasdonetoavoidAl³⁺hydrolysisandlossofAl³⁺

availabilityforthecomplexformationwithLMG.Adjustmenttoa lowerpHwasimpracticalduetothelaterrequiredadjustmentto theoptimalreactionpHof5.0.

A reagentsolution of 1.5mmolL⁻¹ lumogallion(4-chloro-6- (2,4-dihydroxyphenylazo)-1phenol-2-sulfonicacid)andabuffer solution of 5molL⁻¹ of ammonium acetate (NH4Ac) buffer, adjusted withglacial acetic acidtopH 5.4, wereprepared. For extraction,n-hexanolwasusedthroughout.

Formeasurement ofaluminum inseawater,theinterference of fluoride anion hasto betaken into consideration since alu- minumfluoride formation competes withthe formation ofthe LMG-complex.Inapreviouswork,thisinterferencewasconsider- ablyreducedattheadditionofberyllium.Therefore,25mmolL⁻¹of berylliumnitratewereaddedtotheLMGreagentsolutiontoyield afinalconcentrationof350␮molL⁻¹Be²⁺,whichhadbeenfound tobetheoptimalvalueinourpreviouswork[23].

A1mgL⁻¹ rhodamineBsolutionwasusedfor studiesofin- syringehomogenizationbystirred-assistedmixing.

2.2. Samplecollectionandpreparation

Coastalseawatersampleswerecollectedatdifferentbaysofthe islandMallorcatoevaluatethemethodsapplicabilitytosurface seawateranalysis.ThesampleswereacidifiedtopH3atthetime ofcollection.Thesamplesweremeasuredwiththeproposedana- lyzersystemwithoutanyotherprevioustreatmentbutallowing onlythegrouseparticlestosediment.Likewise,twopondwater sampleswerecollectedondifferentplacesontheMallorcaIsland, acidifiedtopH3,andmeasuredunderthesameconditions.After acidificationandbeforemeasurement,thesampleswereallowed

Fig.1. (A)Analyzermanifoldwithselectionvalve(SV),syringepump(S),solenoid 3-wayheadvalve(V),detectionflowcell(D),heatingdeviceintegratedintotheHC (B)andthemagneticstirringbardriver(H)placedontothesyringebarrel.Amotor (M)isusedtodriveitviaarubberband(G).PTFEtubing(0.8mmi.d.)of15cm(A, C),10cm(E),and40cm(F):10cm.(B)Themagneticstirringbardriverplacedonto thesyringeglassbarrelshownindetailconsistingoftwonylonrings,twolongiron screwsandtwoneodymiummagnets.

tostandforatleast3h,bothforsedimentationbutalsotoensure thedissolutionofaluminumhydroxides.

2.3. Manifoldconfiguration

TheMSA-DLLMEmanifoldisdepictedinFig.1Awithalltub- ingdimensionsindicated.Polytetrafluoroethylene(PTFE)tubingof 0.8mminnerdiameter(id)wasusedfortheentiremanifold.

Thecomputercontrolledflowsetupcompriseda5000-stepmul- tisyringepump(CrisonSL,Alella,Barcelona)andtherotary8-port selectionvalve(SV,CrisonSL,Alella,Barcelona)forliquidhandling anddistribution.Themultisyringepumpwasequippedwithasole glasssyringe(S)of5mLpurchased fromHamiltonBonaduzAG (Bonaduz,GR,Switzerland).Athree-waysolenoidheadvalve(V) on-topofthesyringeenabledtheconnectiontoeitherthecentral portoftheSV(positionON,activated)ortothedetectioncelland downstreamlocatedwasteforquantificationoftheextractedana- lyteaswellasfordischargeduringsyringecleaning(positionOFF, deactivated).

PeripheralportsofSVwereconnectedtoreservoirsofwaste(1), water(2),sample(3),buffer(4),lumogallionreagent(5),n-hexanol (6),air(7),andacetonitrile(8).Waterandacetonitrilewereused forcleaningofthedetectionflowcellorthesyringe,whichwas routinelydonedaily.

The connectionbetweenthe centralport of the SVand the syringeheadvalvewasdonebyashortholdingcoil(HC)consist- ingoftwoPTFEtubesof15cminlengthholdingapriordescribed heatingdevice[23]inthemiddle.Heatingwasdonetofavorthe slow reaction betweenLMG and Al³⁺. Briefly, it consisted of a 12cm long, 1.5mmid glasstube insertedintoa brass support, whichwasheatedusinga commercialhalogenlightbulb(12V, 20W).Temperaturecontrolwithahysteresisof<1Kwasachieved usinga control circuitfromCEBEK FadiselSL(Barcelona,Spain Ref.I-81).

2.4. Magneticstirringbardriver

Theentireanalyticalprocedurewascarriedoutinthesyringe includingsamplemixingwithreagentsandextraction.Toachieve

homogeneous and rapid mixing without an additional mixing chamberasgenerallydone[21,23,24],amagneticMicrostirring bar(10mmlength,3mmdiameter)wasusedwithinthesyringe.

Thisarrangementwasdonetothebestofourknowledgeforthe veryfirsttime.Thetoppositionofthesyringepistonwasadjusted insuchaway,thatagaplessabout0.5mmwasleftatemptying thesyringetoavoidanydamage.

Todrivethestirringbarinthesyringe,acommercialmagnetic laboratorystirrerwasimpractical.Therefore,arotatingmagnetic field wasachieved bythe useof a speciallydeveloped magnet driver,showninFig.1B.Tworingsmadeofnylonwereusedas bearings,whichcouldbeplacedeasilyontothesyringe,withthe bottomringslidingontheflangeofthesyringebarrel.TwoM4steel screwsof80mminlengthwereusedasspacersandconnection betweenbothnylonrings.Theso-obtainedassemblycouldfreely rotatearoundthesyringelongitudinalaxis.

Byplacingtwoneodymiummagnets(5mm×4mmo.d.)ontop ofthescrews,thescrewsweremagnetizedandthus,amagnetic fieldinthesyringealongitswholelengthwasobtained.Thismag- neticforcewassufficienttoattractand,atturningthedeviceto forcetherotationofthestirringbarinsidethesyringeindepen- dentlyfromthepositionofthesyringepiston.

Thebottom ringshowedfurtheragroove fortheplacement of a rubber band, which allowed propelling the driver with a low-costDCmotor.TheDC motorwasactivatedusinga home- maderelayandregulationcircuitboardbyanauxiliarysupplyport of themultisyringemodule.The circuitis givenin Supplement material1.

2.5. Detectionequipment

Aspeciallymadedetectioncellwasusedforfluorescencemea- surements.Adetaileddescriptionofthecelldesigncanbefound elsewhere[23].Shortly,itcomprisedaglasstubeof3mmidisused asdetectioncellflowchannel.AbrightgreenLEDwithanemis- sionwavelengthof500nm,poweredbyamobilephonecharger, wasusedasexcitationlightsourceandalignedwiththeglasstube.

Aphotomultipliertube(PMT)fromHamamatsuPhototonicsK.K.

(Hamamatsu,Japan,Ref.:HS5784-04)wasusedfordetectionof fluorescenceemissionandwasmountedinperpendicularposition ontotheglasstube.

Aninterferenceband-passfilterof500±10nm(Ref.:NT62-091) and a long-passglass filterof 580nm cut-off wavelength(Ref.:

NT66-042)fromEdmundOptics(Barrington,NJ,USA)wereplaced betweenLEDandglasstubeandglasstubeandPMT,respectively.

SpectraoftheusedLEDandfilterscanbefoundinapreviouswork [23].

Inaddition,apolycarbonatecollectorlens(F22mm,Ø22mm) wasplacedontothePMTtoachievehighersensitivity.Acontrol unitfromSciwareSystems,S.L.(PalmadeMallorca,Spain)wasused forPMTsupplyanddatareadout.Againof18%waschosenforthe PMT.

2.6. Softwarecontrolanddatahandling

ThesoftwareAutoAnalysis5.0(SciwareSystems,S.L.,Palmade Mallorca,Spain)wasusedforoperationalcontroloftheflowinstru- mentationaswellasdataacquisitionfromthedetectionequipment anddataevaluation.

Theprogram,writteninDelphiandC++,allowsthedefinition andexecutionofinstructionprotocols,includingtheuseofvari- ables,loops,waitingsteps,andproceduresonwindowsbaseduser surface.Detaileddescriptionsofthesoftwarestructureandfeatures aregivenelsewhere[34].

Fig.2.Schemesofthebothoperationschemestestedinthiswork.(A)MixingandhomogenizationofrhodamineBsolutionandwaterand(B)MSA-DLLMEofLMG-Al complexwithn-hexanol.

2.7. Analyticalprotocolsandmethods

The operation methods for testing in-syringe dilution and homogenizationaswellasMSA-DLLMEaregivenschematicallyin Fig.2.TheoperationmethodforMSA-DLLMEisfurtheravailablein Supplementmaterial2.

Allanalyticalproceduresrequiredthecleaningofthesyringe duetotheunavoidabledeadvolumecausedbythestirringbar.

Itwasgiven bythesyringeinner diameterof10.5mm andthe heightofthemagneticstirringbarof3mmminusitspropervolu- metricdisplacementofabout70␮L.However,thecleaningcould beperformed efficiently because the stirring allowedinstanta- neoushomogenizationofthedeadvolumeinthesyringewiththe cleaningsolution.So,three-foldaspirationof0.8mLofwater(Vin positionON,stirringactivated)anddischargetowaste(Vinposi- tionOFF)wassufficientandallowedsyringecleaninginlessthan 30s.Inaddition,proceduresforcleaningofsupplytubesontheSV andthedetectioncellwereestablished.

In-syringe dilution and homogenization was studied using 1mgL⁻¹rhodamineBsolutionandfluorimetricdetection.Subse- quently,1mLofrhodaminesolution,3mLofultrapurewater,and 200␮Lofairwereaspiratedintothesyringeomittingstirring.Then, thesyringecontentwasmixedbyactivationofthestirringfora definedtime.Afterwards,thestirringwasstoppedandthesyringe contentwasdispensedthroughthedetectioncellfortheevaluation oftheachievedhomogenization.

MSA-DLLME wasstarted by the aspirationof 240␮L buffer, 60␮LofLMGreagent,and4.1mLofsampleintothesyringe.Sam- pleaspirationwasdoneata reducedflowrateof4mLmin⁻¹ to increasetheheattransferfromtheheatingdevicetothesample andduringrepeatedactivationofthein-syringestirring.Then,the stirringwasdeactivatedandduringareactiontimeof15s,150␮L ofn-hexanolwereaspiratedintotheHCtoheatingitup.

Afterwards,thestirringwasstartedagainand400␮Lofairwere aspiratedsothatthen-hexanolintheHCandalsopartoftheair couldenterthesyringe.Theairallowedtheformationofavortexin thesyringe(seeSection3.2.3).Atcontactoftheorganicphasewith thestirringbar,itwasdispersedintosmalldroplets.Thestirring waskeptactivatedfor40stoperformMSA-DLLME.Thestirring speedwas2000min⁻¹.

Afterwards,thestirringwasstopped,whichallowedtheformed n-hexanoldropletstofloatandcoalesceduring30satthebrimof theconcaveliquidmeniscusformedbytheaqueousphaseinthe syringe.Toimprovedropletaggregation,theliquidsurfacewasput inmovementbyashortmovementofthepiston(approx.1mm) bytheinstructionofcompletefillingjustbeforethenextstep.The

methodwasfinalizedbypushingtheorganicsolvent,enrichedwith theLMG-Alcomplex,slowlythroughthedetectioncelltowaste undercontinuousdataevaluation.Finally,theremainingliquidwas rapidlydischargedfromthesyringetowaste.

3. Resultsanddiscussion

3.1. Studyofin-syringemixing

To evaluate the potential and characteristics of in-syringe magneticstirring,therequiredmixingtimeforcompletehomoge- nizationwasstudiedfor1,3,7,12,and18s.Theexperimentwas donewithanaqueousdyesolutionaswellaswithadyesolution preparedina20%(w/V)glycerolmixture.Thelatersolutionwas usedtosimulateasampleofapproximatelytwicetheviscosityof water(about12%higherafterhomogenization).Theexperiment wascarriedoutintriplicateinordertoevaluatethereproducibility ofthemixingprocess.Theoperationschemeofthisexperimentis representedinFig.2A.Theexperimentalconditionsandtheaverage measurementsateachmomentduringsyringecontentexpulsion andrespectivestandarddeviationsarerepresentedinFig.3.

Fora mixingtimeof1s, thedifferencebetweenusingaque- ousor20%(w/V)glyceroldyesolutionwaseasilydiscerniblewhile for3s,thebehaviorwassimilar.Besides,themixingpatternafter 3scouldbedescribedasreproducibleasthestandarddeviation decreasedconsiderably.After7sofstirring,thedyegradientinthe syringewaslessthan5%andafter12s,completehomogenization wasachievedforbothdyesolutions.

Itshouldbepointedoutthat laterexperimentsshowedthat homogenizationcanbeachievedevenfasterifasmallvolumeofair isaspiratedintothesyringepermittingtheformationofavortex orevidently,iftheaspirationisalreadyinitiatedduringaspiration ofthesolutions.Inconclusion,in-syringestirringpermitshomo- geneousmixingoflargevolumeswithinafewsecondsand,within thestudiedrange,independentlyfromtheviscosityofthesample, whichisincontrasttonon-segmentedFT.

3.2. In-syringemagneticstirringassistedDLLME 3.2.1. Preliminaryremarks

Themainchallengewasthestudyandoptimizationofthephysi- calparametersrelatedtotheextraction.Forthesefirstexperiments, thesamereagentasoptimizedduringapreviousworkwasused, whichseemedjustifiedsinceitwasbasedonthesamechemical reactionanddetectiontechniquebutonanotherdifferentextrac- tiontechnique[23].

Fig.3.Resultsfromthestudyontherequiredtimeforhomogenizationof1mLof1mgL⁻¹rhodamineBsolutionwith3mLofultrapurewater.(A)Dyesolutionprepared withwaterand(B)dyesolutionpreparedwith20%(w/V)glycerol–water.

Likewise,n-hexanolwasusedfortheextractionoftheLMG-Al complexsinceithasfulfilledallrequirementsinthepreviouswork.

Thesewerealowerdensitythanwater,alowsolubilityinwater, andthebestextractioncapacityfoundofalltestedsolventsforthe LMG-Alcomplex[23,29].

Therefore,onlyn-octanolwastestedasalternativeextraction solventbutdiscardedduetotheobservedstickingtothehydropho- bicsurfacesofthepistonheadandstirringbar,slowcoalescenceof thedropletsafterDLLME,andlowersignalreproducibility.Other typicalextraction solvents of lower polarityor lighter alcohols showedtobeimpractical.

3.2.2. Voltageofmotorandstirringrotationspeed

Thestirringduringtheaspirationofn-hexanolintothesyringe causesthedisruptionofthesolventintofinedropletsand thus favors the mass transfer of theLMG-Al complex over a larger boundary surface increasesthe extractionefficiency. The influ- enceofthevoltageofthemotorusedfordrivingthestirringbar driverwasstudiedintherangeof3.6(ca.1500rpm)to6.3V(ca.

2600rpm).TheresultsaregiveninSupplementmaterial3A.The signalheightescalatedbyfactorof2betweenvoltagesof4.6Vand 5.2V.Forlowervoltagesthan4.6Vorhighervoltagethan5.2V,the effectofvoltagechangeonthesignalheightwaslittle.Avoltageof 5Vcorrespondedtoarevolutionspeedofthestirrerinthesyringe ofabout2000rpm.Onlyforhighervoltagesthan5V,i.e.higher rotationspeedsthan2000rpm,apronouncedvortexwasformed inthesyringeandonlythen,then-hexanolwasdrawnefficiently intothevortexanddisruptedintosmalldroplets.Therefore,avolt- ageof5.5Vwaschosenforallfurtherexperimentscorresponding toarotationspeedofapproximately2300rpm.

3.2.3. Volumeofair

Aftertheaspirationofn-hexanol,avolumeofairwasaspirated intotheHCtoaspiratethen-hexanolvolumecompletelyintothe syringe.Vortexformationcausingn-hexanoldispersionaswellas dropletrecombinationafterMSA-DLLMEwasimprovedifthevol- umeofthisairwaslargerthantheHCinnervolume(approximately 200␮L),i.e.whentheheadspaceofthesyringewaspartlyfilled withair.

Thefirst observationisdue tothe factthat a vortexcanbe formedonlyonanopenliquidsurfaceinthesyringeandthework requiredfortheformationofthevortexissmallerifthesurface islarger.Withoutairinthesyringe,theextractionmodewould mimicsingledropextraction.However,preliminaryexperiments showedthat this mode requires considerablylonger extraction

timestoyieldsimilarefficiencies,i.e.severalminutes.Thisisinwell accordance totheextractiontimesreported fornon-automated singledropextractions[35].Thesecondobservationisduetothe formedmeniscusoftheaqueousphaseandresultingprominence oftheliquidsurfaceonthesyringewalls.Consequently,thefloat- ingn-hexanoldropletsareforcedtoaccumulateinthislimitedarea achievingfasterdropletcoalescence.

Theinfluenceoftheairvolumeonthesignalheightwasstud- iedinarangeof300–500␮Lwithconditionsandresultsgivenin Supplementmaterial3B.Itwasobserved,thatthesignalincreased about12%withalargervolumeofairandalsothereproducibility ofmeasurementimprovedslightlyfrom6.6%to2.3%RSD.Onthe otherhand,aplusofairinthesyringealsoreducedtheusableliq- uidvolumeinthesyringeandbythistheamountofsample,which couldbeusedfortheextractionprocedure.Takingthisintoaccount, theeffectivesignalincreaseisonly7%.Thus,avolumeof400␮Lair wasfinallychosen.

3.2.4. Reactionandextractiontimes

ThereactiontimetRandtheextractiontimetEwereoptimized following acentralcomposite experimentaldesign.The studied range,experimentalconditions,andresultsaregivenasSupple- mentmaterial4 and5.Here,tRis definedasthetime between in-syringemixingofsampleandreagentsandtheadditionofn- hexanol.Theparametert_E is definedasthestirringtimein the presenceofn-hexanol.

Twoapproachesweretested.Inthefirst,n-hexanolandairare aspiratedaftert_R,i.e.then-hexanolisonlyinshortcontacttime withtheheatingdeviceandentersthesyringenearlyatambient temperature.Inthesecond,n-hexanolisaspiratedduringtRinto theHCwhiletheairisaspiratedaftert_R,i.e.thesolventisallowed toheatupduringtRintheHC.

Fromtheresultsitbecomesclearthatn-hexanolheatingwas highlyfavorable.Thiswasduetothelowerviscosityofhexanol athighertemperatureandthusdisruptionintosmallerdroplets andimprovedextractionefficiency.Usingthefirstapproach,i.e.

n-hexanol at ambient temperature, the best results were pre- dictedforaminimalt_R,mostlikelybecausethemixtureofsample andreagentsinthesyringeisatitsmaximaltemperature,which decreasesduringt_R.

Using the second approach, i.e. pre-heated hexanol, signal heights were doubled. Longer tR were advantageous due to proceededreactionyieldwithapredictedmaximumat20s.Simul- taneously, shorter t_E were required than in the first approach where maximizationof tE waspredictedasoptimalinorder to

Fig.4. Influenceofaspirationflowrate(A),thevolumeofn-hexanol(B),theconcentrationoflumogallion(C),andNH4Acbuffer(D)inthefinalmixture.Conditions 500nmolL⁻¹aluminumultrapurestandard(triangles),SSWstandard(squares)andSSWblank(diamonds),motorvoltage5.5V,400␮Lair,reactiontemperature45^◦C, 15sreactiontime,40sextractiontime.Further:(A)with150␮Ln-hexanol,120␮Lof365␮molL⁻¹LMGand5molL⁻¹NH4AcpH5.1.(B)as(A)withaspirationflowrate 4mLmin⁻¹,(C)as(B)with150␮Ln-hexanol,(D)as(C)with60␮L1.5mmolL⁻¹LMGreagent.

compensatetheshorterreactiontime.Onthebaseofthefound resultsandoptimization,20soft_Randt_Eof40sapplyingthesecond approachofheatedn-hexanolwereusedfurther.

3.2.5. Volumeoftheextractionsolvent

Thevolumeofn-hexanolwasstudiedintherangeof70–180␮L forblankSSWandfortwo500nmolL⁻¹Al³⁺standards,onepre- paredwithSSW,theotheronepreparedwithultrapurewater.The resultsandexperimentalconditionsaregiveninFig.4A.

ItwasfoundthatfortheSSWblankandSSWstandard,thesig- nalsincreasedupto130␮Lreachingastablelevelbeyond.Asimilar behaviorwasobservedforthesignalobtainedwiththeultrapure standard.Here,thesamesignallevelwasreachedbutanabout 20␮Llargervolumeofn-hexanolwasrequiredtoachievecompa- rableresults,beingtheresultofthehighersolubilityofn-hexanol inultrapurewatercomparedtosaltwater.

Visualinspectionandtheobtainedresultsallowedthefollow- ingconclusions.First,a smallportionof then-hexanolwasnot disruptedintodropletsbutremainsfloatingatthesurfaceexplain- ingwhytheDLLMEwaslessefficientatn-hexanolvolumeslower than100␮L.Second,theionicstrengthofthesampledidnotaffect thesignalheightaslongasthesolubilityofn-hexanolinthesam- ple[<6␮LmL⁻¹]istakenintoaccount.Thesignalheightsobtained withultrapureandSSWstandardswith150␮Lofn-hexanoldid notdiffersignificantly(3%found),i.e.thedependencyofthesignal heightontheionstrengthofthesamplesolutionwasminimal.As consequence,avolumeof150␮Lofn-hexanolwaschosenforall furtherwork.Increasingthen-hexanolvolumebeyondthestudied rangewouldprobablyhaveledtodecreasingsignalheightsdueto dilutionoftheextractedLMG-Alcomplexinthesolvent.

3.2.6. Flowrateforsampleaspiration

Theflowrateforsampleaspirationintothesyringewasofhigh interestsinceitdeterminedthecontacttimeofthesamplewith theheatingdeviceintegratedintheHC.Alowerflowratewould favorafasterreactionrateandhigheryieldasreportedelsewhere [23,28,29]butprolongthemethodexecutiontime.Ahighersample temperaturecouldfurtherimprovetheextractionefficiencydueto lowerviscosity,whiletheantagonisticeffectofahighersolubility ofn-hexanolandtheLMG-Alcomplexintheaqueousphasehasto beconsidered.

Theeffectoftheflowrateonthesignalheightwasstudiedin therangeof2–5mLmin⁻¹.Theexperimentalconditionsandresults arerepresentedinFig.4B. Itwasfoundthatthesignalsofboth SSWblankand standardfollowedanexponentialdecreasewith higherflowrates.Asacompromisebetweensensitivityandtime ofanalysis,aflowrateof4mLmin⁻¹waschosen.

Attheadjustedtemperatureoftheheatingdeviceof65^◦C,the sampleenteredthesyringeatthechosenflowratewithabout45^◦C.

Ahigherheatingdevicetemperaturewasdiscardedtoavoidbubble formation.

3.2.7. Concentrationoflumogallion

Afteroptimizationofthephysicalparameters,thereagentcom- position,chosenfromthepreviouswork[23],wasre-evaluated.

First,theinfluenceofLMGquantityintheaqueousphasebefore extractionwasstudiedintherangeof3.8–75␮molL⁻¹finalcon- centrationusinga500nmolL⁻¹SSWstandardandSSWblank.The resultsandexperimentalconditionsaregiveninFig.4C.

It was observed that both the standard and blank signals increasedrapidlyup toa maximumat 15␮molL⁻¹ witha less

Table1

Resultsfromtheanalysisofcostalseawatersamplesandnaturalwatersamplesandwithonespikedconcentration.ConditionsasinFig.4D.

Type Addedconcentration[nmolL⁻¹] Signal(n=3) Foundconcentration[nmolL⁻¹] Recovery

Seawater1 0 504.2±73.1 104.5

50 598.1±20.2 153.1 97.3%

Seawater2 0 1070.1±51.5 398.8

100 1281.5±105.1 508.7 109.9%

Seawater3 0 496.0±15.9 100.5

100 701.7±27.3 207.4 106.8%

Seawater4 0 378.2±11.5 39.3

200 838.0±39.1 278.2 119.4%

Seawater5 0 475.3±1.5 89.7

100 674.9±8.7 193.4 103.7%

Pondwater1 0 581.1±42.8 144.7

200 962.6±59.2 342.9 99.1%

Pondwater2 0 400.3±9.7 50.8

200 838.8±39.4 262.9 106.1%

pronounceddecreaseforhigherconcentrations.Thisbehaviorwas mostlikelyduetothemoderatesolubilityoftheLMG-Alcomplexin waterandthusretentionofaluminumintheaqueousphaseathigh LMGconcentration.Therefore,aLMGconcentrationof15␮molL⁻¹ waschosenyieldingthemaximalsignals.

3.2.8. Concentrationofbuffer

NH₄AcbufferwasusedtoadjusttheoptimalreactionpHand toincreasetheionicstrengthoftheaqueousphasetoimprovethe extractionefficiencyandtodecreaseitssolubilityforn-hexanol.A largervolumeofbufferfavorsthemethod’srobustnessinrespectof thepHandionicstrengthofthepropersample.However,alarger volumeofbufferalsoimpliesasmallervolumeavailableforsample inthesyringeandso,alowerpossiblepreconcentrationfactor.

Theeffectofthefinalbufferconcentrationintheaqueousphase wasstudiedintherangeof60–250mmolL⁻¹.TheNH4Acbuffer solutionwaspreparedhighlyconcentrated(5molL⁻¹).Atthiscon- centration,themeasurementofthepHvaluewithacommercial pHmeterisnotreliable,sothatthebufferwasadjustedtopH5.4, whichyieldedthereportedoptimalreactionpHvalueof5.0[23,29]

ata50-folddilutionwithultrapurewater.Resultsandexperimental conditionsaregiveninFig.4D.

Whiletheblanksignalincreasedwiththebufferconcentration upto150mmolL⁻¹ and remainedstable beyond, aclearsignal maximumwasfoundat150mmolL⁻¹ NH₄Acfora500nmolL⁻¹ acidifiedSSWstandard.Therefore,thisconcentrationwaschosen asoptimal.

3.2.9. Phaseseparationtime

Thetimeofphaseseparationbydropletflotationandaggrega- tionwastestedfor20,30,and40susinga500nmolL⁻¹aluminum SSWstandard.Averagesignalheightsof3subsequentextractions yielded1771±82,1915±33,and1949±13unitsfor20,30,and 40s,respectively.Asexpected,thesignalheightincreasedandthe signalreproducibilityimprovedwithlongertimesbut lesspro- nouncedfrom30 to40sthanfrom20to30s. Tominimizethe executiontime,aphaseseparationtimeof30swaschosenforall furtherwork.

3.2.10. Realsampleanalysisandanalyticalperformance

Calibrationwasdonewithstandardsupfromto1.9␮molL⁻¹ andfoundtobelinearuptoatleast1.1␮molL⁻¹.At1.4␮molL⁻¹, the deviationfrom the extrapolated calibrationcurve was still onlyabout 6%.Thecalibrationcurve followedthe functionsig- nalheight=1.925[Lnmol⁻¹]·c[nmolL⁻¹]+302.6,r²=0.998.The limitofdetectionandthelimitofquantificationwerecalculated fromthetripleand 10-foldstandard deviationoftheblankand thecalibrationcurveslopeyielding6.1nmolL⁻¹and20.2nmolL⁻¹, respectively.

Thepreconcentrationfactorwascalculatedfromtheusedsam- plevolumeandthefinalvolumeofhexanol(ca.125␮L)tobeabout 33.Theextractionefficiencycanbeestimatedtobehigherthan95%

sincethebaselinefoundfortheaqueousphaseaftertheextraction wasnegligible.

ToevaluatetheapplicabilityoftheproposedautomatedMSA- DLLME method,five coastalsurface seawater samplesand two pond-watersamplesweremeasuredwiththedescribedanalyzer system.Allsampleswerefurtherspikedwithaluminumstandards inasimilarconcentrationrangeasthenaturalconcentration.The samecalibrationpreparedwithsyntheticseawaterwaterwasused fortheevaluationofallsamples.Theresultsaresummarizedin Table1.

The sampleconcentrations were allwithin thelinear work- ing range. An average recovery of 106.0%±7.3% was achieved, 102.6%±5.4%omittingseawatersample4,forwhichtherecovery valuesexceededtheacceptablerangeforunknownreason.

Themethodshowedtobeapplicabletothedeterminationof aluminuminvariouswatersamples.Thepositivedeviationofmost recoveryvalueswasmostlikelyduetothefactthatspikingwas donedirectlybeforeanalysiswithoutagingtime.Sinceinthiswork, similarconditionsofthechemicalparametershavebeenappliedas inourpreviouswork,andthereactioniswell-knowntobehighly selective,astudyofsinglecomponentsaspotentialinterferences wasnotrepeated.

An average repeatability of 3.3% (n=4) was found for cal- ibrations standards while for real samples, the average value wasslightly higherwith 4.4%(n=4).The entire analytical pro- cedure including initialsyringe cleaning took 210s, allowing a measuringfrequencyof17h⁻¹.IncontrasttoMSA-DLLMEbased on manual operation [4], the proposed system achieved effi- cient extraction in 40s instead of several minutes and under fully automated conditions. Thus, the analytical performance was adequatefor thedetermination of aluminum in all tested matrices.

Incomparisonwithaformerworkusingthesamereactionand instrumentationbutbasedondispersionsolventassistedDLLME [23],about25%lowerLODandLOQvalues,a20%shortertimeof analysisandan8%highersensitivitywereachievedbysimulta- neousreduction oforganicsolventsfrom950␮L to150␮L.The repeatabilityandlinearworkingrangewerecomparable.

Another improvement over the former work was a higher robustnessinrespectofthesamplesalinityduetotheomission ofthedispersionsolventundertheoptimizedconditions,i.e.an- hexanolvolumeof150␮LasdiscussedinSection3.2.5.Thiswas alsodemonstratedbythefactthatthesamecalibrationwithstan- dardspreparedwithSSWwasadequateforbothfreshwaterand seawatersamplesandtherecoveryvaluesfoundbothsampletypes werecomparable.

Table2

ComparisonwithpriormethodsforthedeterminationaluminumusingsolventemulsificationorDLLME,respectively.

Extractionsolvent[␮L] RSD[%] LOD[ppb] ULR[ppb] Sample[mL] Time[min] Extraction Detection Ref.

600 1.7 0.05 −15 25 <10 IL-DLLME FL [40]

132 4.5 0.8 −250 20 >8 DLLME-SFO ICP-OES [44]

48 2.6–5.3 0.6–0.9 −1000 10 >11 US-DLLME ICP-OES [42]

75 3.2 1.7 n.g. 10 >15 USILDLLME UV–vis [43]

98 1.87 0.13 −1000 10 >10 US-DLLME ICP-OES [41]

950 <5 0.22 −27 3.9 4.4 In-syringeDSA-DLLME FL [23]

150 3.3–4.4 0.16 −33 4.1 3.5 In-syringeMSA-DLLME FL Thiswork

Abbreviations:DSA,dispersionsolventassisted;DLLME,dispersiveliquid–liquidmixroextraction;FL,fluorescence;IL,ionic-liquidbased;IS-MSA,in-syringemagnetic stirringassisted;n.g.,notgiven;SE,surfactantenhanced;SFO,solidificationoforganicdrop;US,ultrasoundassisted;ULR,upperlinearworkingrangelimit;UV–vis, spectrophotometry.

Theseimprovementsareconsideredtoberelatedtothepossibil- itytoperformthewholeprocedureincludingreagentandsample mixinginthesyringe,i.e.faster,andduetotheomissionofdisper- sionsolvent,whichpermittedtheuseofalargervolumeofsample (4.1mL),andahighandreliableextractionefficiency,whichisnot affectedbythemixtureofdispersionsolventandaqueousphase.

3.2.11. Comparisonandoutlookonfurtherpotentialand applications

Acomparisonwithpriorreportedmethodsbasedonemulsifi- cationofextractionsolventandsampleordispersiveliquid–liquid microextractionfor the determination of aluminum is given in Table2.Apartfromthisandourformerwork[23],allotherworks reportedmanualprocedures.Theproposedworkwasfoundcom- parableorbetterin respectofmostcharacteristics totheprior reported applications whereas the preconcentration factor and thelinearworkingrangeweresmaller.Nevertheless,theworking rangecouldbeextendedbyin-syringedilutionofthesamplewith ultrapurewater.Inrespecttothetimeofanalysis,thepresented workwasclearlysuperior.However,ithastobetakenintoaccount thatmanualproceduresallowtreatingseveralsamplesinparallel andconsequentlyallowincreasingtheeffectivesamplefrequency.

Thepresentedworkisthefirstapplicationofin-syringeMSA- DLLME and was done with the intention to demonstrate the potentialofin-syringestirringincomparisonwithdispersionsol- vent assisted DLLME [23].The usefulness of in-syringe stirring when dealing with samples of distinct viscosities was further proven.

Asadisadvantage,thedeadvolumeproducedbythestirringbar hastobeaddressed.However,syringecleaningcanbeperformed fastand,duetothestirring,veryefficient.Itshouldbepointedout, thatnomemoryeffectwasobservedandcleaningwithwaterwas sufficienttoavoidcross-overofanalyteatthechangeofsampleor standardsolution.

Ontheotherside,theuseofexternalextractionchamberselse- whereproposed[19] impliesevenmoretimefor cleaningsince theentireextractionchamberneedstobefilledwiththecleaning solutionandthen emptied.Ifonlyonepumpisused,thistakes anadditionalsteptore-aspiratethecleaningsolutionfromthe extractionchamberbeforedischargingittowaste.Usingin-syringe extraction,theinnerwallsofthe“extractionchamber”(thesyringe) arewipedbythesyringepistonandtherequiredvolumetoclean isreducedtothesmalldeadvolume,sothatnottheentiresyringe hastobecleaned.

Temperature turned out to be one of the most important parameters of this work due to its influence on both the liq- uid viscosity and the reaction rate. A more efficient heating devicecouldthereforeimprovetheachievedperformance.Another improvementmightbepossiblebyfluorimetricmeasurementof theorganicphase directlyin thesyringe followingtherecently proposedmethodologyof Lab-In-A-Syringe[22].Suchcombina- tioncouldalsoenablein-syringe titrations withtheinteresting

characteristicthatthetitrationvesselcanbeadaptedinvolume withouttheinterferenceofair.

Theproposedinstrumentationcouldfurtherbeusedtocarry outclassicalanalyticalprotocols, i.e.thestep-wiseadditionand mixingofsmallvolumesofreagentstoalargevolumeofsample suchasdonein“batchautomation”[36–38].Finally,theproposed systemcouldbecoupledtoliquidorgaschromatographytocarry outsampleclean-upandanalytepre-concentration.

Incomparison withformerin-syringe automationinsyringe extractionbutposteriorderivatizationinareactionvial[39].Here, both,thereactionandtheextractionwerecarriedoutin-syringe, sincethestirringactionenabledcompletemixingofallsolutions.

Nevertheless,fullyautomationcouldonlybeachieved,ifstandard preparationandsampleprovisionusinganautosamplerwouldbe enabled.

4. Conclusions

Astirringbarplacedintothesyringeofacomputercontrolled syringepumpwasusedforthefirsttimefor magneticstirring- assisted dispersive liquid–liquid microextraction (DLLME). The optimizedmethodenabledefficientDLLMEwithinacomparably shorttimeandisbasedonthedisruptionoftheextractionsolvent bythekineticenergyoftheswirlingstirringbar.Betterorsimilar analyticalperformancethaninpreviousworksbasedonDLLME wasachieved and themethod’s applicabilitytothe determina- tionofaluminuminsurfaceseawaterandfreshwatersampleswas proven.Dependencyoftheanalyticalperformanceonthesample salinityandviscositywasdemonstratedtobewidelyovercome.In- syringestirringcanenablenovelprotocolsforsamplepreparation, analytepre-concentration,andcomplexanalyticalapplications.

Acknowledgements

TheauthorsacknowledgethefinancialsupportfromtheSpanish MinistryofScienceandInnovationthroughtheprojectCTQ2010- 15541andfromtheConselleriad’Economia,Hisenda,eInnovació oftheGovernmentoftheBalearicIslandsthroughtheallowanceto competitivegroups(43/2011).B.Horstkottewasfurthersupported byapostdoctoralfellowshipoftheprojectCZ.1.07/2.3.00/30.0022 supportedbytheEducationforCompetitivenessOperationalPro- gram(ECOP)andco-financedbytheEuropeanSocialFundandthe statebudgetoftheCzechRepublic.

AppendixA. Supplementarydata

Supplementarydataassociatedwiththisarticlecanbefound,in theonlineversion,athttp://dx.doi.org/10.1016/j.aca.2013.05.049.

References

[1]M.Rezaee,Y.Assadi,M.-R.M.Hosseini,E.Aghaee,F.Ahmadi,S.Berijani,J.

Chromatogr.A1116(2006)1.

[2]A.Saleh,Y.Yamini,M.Faraji,M.Rezaee,M.Ghambarian,J.Chromatogr.A1216 (2009)6673.

[3]M.A.Farajzadeh,M.R.A.Mogaddam,Anal.Chim.Acta728(2012)31.

[4]Y.Zhang,H.K.Lee,J.Chromatogr.A1249(2012)25.

[5]P.-P.Zhang,Z.-G.Shi,Q.-W.Yu,Y.-Q.Feng,Talanta83(2011)1711.

[6]Q.Wu,Q.Chang,C.Wu,H.Rao,X.Zeng,C.Wang,Z.Wang,J.Chromatogr.A 1217(2010)1773.

[7]M.Rezaee,Y.Yamini,M.Faraji,J.Chromatogr.A1217(2010)2342.

[8]A.Sarafraz-Yazdi,A.Amiri,TrendsAnal.Chem.29(2010)1.

[9]A.V.Herrera-Herrera,M.Asensio-Ramos,J.Hernández-Borges,M.A.Rodríguez- Delgado,TrendsAnal.Chem.29(2010)728.

[10]A.N.Anthemidis,K.-I.G.Ioannou,Talanta80(2009)413.

[11]M.Trojanowicz,AdvancesinFlowAnalysis,Wiley,2008.

[12]S.D.Kolev,I.D.McKelvie,Wilson&Wilson’sComprehensiveAnalyticalChem- istry:AdvancesinFlowInjectionAnalysisandRelatedTechniques,Elsevier ScienceLimited,Amsterdam,Oxford,2008.

[13]J.Ruzicka,G.D.Marshall,Anal.Chim.Acta237(1990)329.

[14]A.Economou,TrendsAnal.Chem.24(2005)416.

[15]R.B.R.Mesquita,A.O.S.S.Rangel,Anal.Chim.Acta648(2009)7.

[16]A.N.Anthemidis,K.-I.G.Ioannou,Talanta79(2009)86.

[17]A.N.Anthemidis,K.-I.G.Ioannou,Anal.Chim.Acta668(2010)35.

[18]A.N.Anthemidis,K.-I.G.Ioannou,Talanta84(2011)1215.

[19]V.Andruch,C.C.Acebal,J. ˇSkrlíková,H.Sklenáøová,P.Solich,I.S.Balogh,F.Billes, L.Kocúrová,Microchem.J.100(2012)77.

[20]F. Maya, J.M. Estela, V. Cerdà, Anal. Bioanal. Chem. 402 (2012) 1383.

[21]B.Horstkotte,M.Alexoviè,F.Maya,C.M.Duarte,V.Andruch,V.Cerdà,Talanta 99(2012)349.

[22]F.Maya,B.Horstkotte,J.M.Estela,V.Cerdà,Anal.Bioanal.Chem.404(2012) 909.

[23]R.Suárez,B.Horstkotte,C.M.Duarte,V.Cerdà,Anal.Chem.84(2012)9462.

[24]B.Horstkotte,F.Maya,C.M.Duarte,V.Cerdà,Mikrochim.Acta179(2012)91.

[25]J. ˇSkrlíková,V.Andruch,H.Sklenáøová,P.Chocholouˇs,P.Solich,I.S.Balogh, Anal.Methods2(2010)1134.

[26]D.J.Hydes,P.S.Liss,Analyst101(1976)922.

[27]A.G.Howard,A.J.Coxhead,I.A.Potter,A.P.Watt,Analyst111(1986)1379.

[28]J.A.Resing,C.I.Measures,Anal.Chem.66(1994)4105.

[29]J.Zhang,H.Xu,J.L.Ren,Anal.Chim.Acta405(2000)31.

[30]J.L.Ren,J.Zhang,J.Q.Luo,X.K.Pei,Z.X.Jiang,Analyst126(2001)698.

[31]C.Y.Zhou,J.Wu,H.Chi,M.K.Wong,L.L.Koh,Y.C.Wee,Talanta42(1995)415.

[32]M.P.Mánuel-Vez,C.Moreno,D.J.González,M.García-Vargas,Anal.Chim.Acta 355(1997)157.

[33]APHA,AWWA,WEF,L.S.Clesceri,A.D.Eaton,A.E.Greenberg,StandardMeth- odsfortheExaminationofWaterandWastewater,AmericanPublicHealth Association,Washington,DC,1998.

[34]E.Becerra,A.Cladera,V.Cerdà,Lab.Rob.Autom.11(1999)131.

[35]M.A.Jeannot,F.F.Cantwell,Anal.Chem.68(1996)2236.

[36]F.A.S.Cunha,R.A.Sousa,D.P.Harding,S.Cadore,L.F.Almeida,M.C.U.Araújo, Anal.Chim.Acta727(2012)34.

[37]L.F.Almeida,M.G.R.Vale,M.B.Dessuy,M.M.Silva,R.S.Lima,V.B.Santos,P.H.D.

Diniz,M.C.U.Araújo,Talanta73(2007)906.

[38]S.S.Monte-Filho,M.B.Lima,S.I.E.Andrade,D.P.Harding,Y.N.M.Fagundes,S.R.B.

Santos,S.G.Lemos,M.C.U.Araújo,Talanta86(2011)208.

[39]J.Lee,H.K.Lee,Anal.Chem.83(2011)6856.

[40]H.Abdolmohammad-Zadeh,G.H.Sadeghi,Talanta81(2010)778.

[41]H.Sereshti,Y.EntezariHeravi,S.Samadi,Talanta97(2012)235.

[42]H.Sereshti,S.Bakhtiari,V.Khojeh,Anal.Methods3(2011)2936.

[43]J.B.Ghasemi,E.Zolfonoun,Environ.Monit.Assess.184(2012)3971.

[44]M.Rezaee,Y.Yamini,A.Khanchi,M.Faraji,A.Saleh,J.Hazard.Mater.178(2010) 766.