DSpace at VNU: Automated capillary electrophoresis with on-line preconcentration by solid phase extraction using a sequential injection manifold and contactless conductivity detection

Stock solutions of ibuprofen, diclofenac, naproxen and bezafi-brateintheformofsodiumsalts1mmolL−1wereusedforthe dailypreparationofthestandards.Beforeuse, thecapillarywas preconditionedwit

jo u r n al h om ep 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

Thanh Duc Maia,b, Benjamin Bomastyka, Hong Anh Duongb, Hung Viet Phamb,∗∗, Peter C Hausera,∗

a University of Basel, Department of Chemistry, Spitalstrasse 51, 4056 Basel, Switzerland

b Centre for Environmental Technology and Sustainable Development (CETASD), Hanoi University of Science, Nguyen Trai Street 334, Hanoi, Viet Nam

a r t i c l e i n f o

Article history:

Received 1 February 2012

Received in revised form 15 March 2012

Accepted 22 March 2012

Available online 1 April 2012

Keywords:

Solid phase extraction

Capillary electrophoresis

Sequential injection analysis

Capacitively coupled contactless

conductivity detection

Drug residues

a b s t r a c t

© 2012 Elsevier B.V All rights reserved

Incapillaryelectrophoresis(CE),detectionsensitivitiesare

gen-erallynotasgoodasthoseinliquidchromatography,mainlydue

tothebynecessitysmallinjectionvolumesinthenano-literrange

ToimprovethesensitivityofCEseveralon-andoff-line

preconcen-trationstrategieshavebeenapplied,ofwhichthemostcommonly

employedareelectrokinetictechniques(isotachophoresis,

stack-ing,sweepinganddynamicpHjunctionmethods)andsolid-phase

extraction(SPE)methods.Foranoverviewoverthetwoapproaches

seeforexamplethesereviews[1–8].Theelectrophoreticmethods,

althoughinherentlybeingon-lineandhavingthegeneral

advan-tageofrelativesimplicity,havelimitationssuchasrestrictionof

theprimary sample volume, undesired discrimination between

analytesofdifferentelectrophoreticmobilities,theneedofalow

backgroundconductivityandthepossiblerequirementofa

preced-ingmatrixclean-upstepforcomplexsamples.TheSPE-methods,

ontheotherhand,allowthestrippingoflargesamplevolumes,

muchexceedingtheinternalvolumeoftheentireseparation

cap-illary,andthuscanachievehighpreconcentrationfactors.These

∗ Corresponding author Tel.: +41 61 267 1003; fax: +41 61 267 1013.

∗∗ Corresponding author Tel.: +84 4 3858 7964; fax: +84 4 3858 8152.

E-mail addresses: vietph@hn.vnn.vn (H.V Pham), Peter.Hauser@unibas.ch

(P.C Hauser).

techniquesalsoefficientlyremovecomponentsnotretainedonthe trappingmaterial,whichisusefulintheanalysisofsampleswith complexmatrices

Sofar,mostofthecoupledSPE–CEsystemsreportedhavebeen basedonliquiddeliverywithachromatographypumporaflow injection manifoldcoupled, often via a robotic arm, toa com-mercialelectrophoresisinstrumentemployingdetectionbasedon UV-absorption,laserinducedfluorescenceormass-spectrometry

[4–9].Thesesetupsarethusrelativelycomplexandexpensiveand notsuitedforon-sitedeploymentinautomatedmonitoringtasks

Anattractivealternativeapproachintheconstructionofa CE-instrumentistheuseofasimplesequential-injectionanalysis(SIA) setupbasedonasyringepumpand amulti-positionvalve.The SI-manifoldcanbeemployedasanalternativemeansforsample injectionand forflushing oftheseparation capillary and appli-cationsofthecombinationaresummarizedin[10].RecentlyMai

etal.[11]usedsuchaSIA–CEsystemforanunattendedon-siteand on-linemonitoringapplicationandalsoforpressure-assistedCE

[12,13].However,toourknowledge,theexplorationofthe poten-tialofa coupledSI-manifoldfor extendedsample pretreatment priortoanalysisbyCEhasonlybeenreportedonce.Horstkotteetal

[14]reportedamulti-syringeSPE-CEsetupforthedetermination

ofnitrophenolsbyUV-detection

Afurthersimplificationispossiblebyemployingcapacitively coupled contactlessconductivity detection (C4D) This is based

onasimple measuringcellconsistingofapairofshorttubular

0003-2670/$ – see front matter © 2012 Elsevier B.V All rights reserved.

isfullyelectronicandlessdemandinginconstructionandpower

consumptionthanthecommonopticaldetectionmethods

employ-ingUV-radiation.DiscussionsofapplicationsofC4DforCEcanbe

foundinrecentreviews[15–19]andfundamentaldetailsaregiven

in[20–24]andearlierworkcitedtherein.ThecombinationofSPE

withCE–C4DhasbeenreportedbyDingandRogers[25]forthe

determinationof haloacetic acidsin swimmingpool water, but

thepreconcentrationwascarriedoutoff-lineinamanualfashion

Herein,theimplementationofautomated SPEonaSIA–CE–C4D

systemisdescribed

2.1 Chemicalsandmaterials

Allchemicalswereofanalyticalorreagentgradeandpurchased

fromFluka(Buchs,Switzerland)orMerck(Darmstadt,Germany)

Stock solutions of ibuprofen, diclofenac, naproxen and

bezafi-brateintheformofsodiumsalts(1mmolL−1)wereusedforthe

dailypreparationofthestandards.Beforeuse, thecapillarywas

preconditionedwith1MNaOH for10min anddeionized water

for 10minprior tobeingflushed with theBGE solutionat the

appropriatepH(for1h).Deionizedwaterpurifiedusingasystem

fromMillipore(Bedford,MA,USA)wasusedforthepreparation

of allsolutions The water sampleswere filteredwith0.02␮m

PTFEmembranefilters(ChromafilO-20/15MS,Macherey-Nagel,

Oensingen,Switzerland),spikedwiththeselecteddrugresiduesas

needed,andultra-sonicatedfor2minfordegassing

2.2 Instrumentation

Adualpolarityhighvoltagepowersupply(SpellmanCZE2000,

Pulborough,UK)with±30kVmaximumoutputvoltageand

poly-imidecoatedfusedsilicacapillariesof365␮mODand25␮mID

(fromPolymicro,Phoenix,AZ,USA)wereusedforallexperiments

DetectionwascarriedoutwithaC4D-systembuiltin-house;details

canbefoundelsewhere[26].Ane-corder201dataacquisition

sys-tem(eDAQ,DenistoneEast,NSW,Australia)wasusedforrecording

thedetectorsignals.TheSIAsectionconsistedofasyringepump

(CavroXLP6000)fittedwitha5mLsyringeanda9-portchannel

selectionvalve(CavroSmartValve)(bothpurchasedfromTecan,

Crailsheim,Germany).Theisolationand3-gatevalvesusedwere

obtainedfrom NResearch (HP225T021 and HP225T031,

Gümli-gen,Switzerland).CommercialSPEcartridges(52602-U,Supelco,

Buchs,Switzerland)containing100mgofpackingmaterial

(octade-cylbondedsilicaparticlesof50␮mdiameter)werefittedintothe

systemwiththehelpoftwotubingadapters(57020-U,Supelco)

TheprogrammingpackageLabVIEW(version8.0forWindowsXP,

fromNationalInstruments,Austin,TX,USA)wasusedtowritethe

controlcode.Furtherdetailsontheinstrumentationcanbefound

in[11]

3.1 Systemdesignandoperation

A simplified diagram of the instrument is given in Fig 1

Several extensions and modifications have been made to the

earlierSIA–CE–C4Ddesign [11]inorder toincorporatethe

pre-concentrationprocedureintothefullyautomatedoperation.For

electrophoreticseparation withoutpreconcentration, the

previ-ouslyreportedsystem[11]reliesonacombination ofastepper

motor-driven2-way syringeand amulti-port selectorvalve for

deliveryofsolutions,andonaSI–CEinterfaceaswellasblocking

valvesforhydrodynamicinjectionandflushingofthecapillary.For operationwithSPEpreconcentration,theessentialchangeisthe inclusionofasecondholdingcoil(HC2)betweenthemulti-port valveandtheinterface.Thetwoholdingcoilsplaydifferentroles

inthetaskofliquidhandling.Theconventionalholdingcoil(HC1), situatedbetweenthepumpandthemulti-selectorvalve,is uti-lizedforaspirationofsampleandstandards(forseparationwithout preconcentration)oreluent(forelutionfromthetrappriorto sep-aration)whiletheothercoil(HC2)servesasareservoirtoholdthe solutionfollowingelutionfromthecartridgebeforeitispumped

totheCE-interfaceforhydrodynamicinjection.AY-shapetubing couplerisemployedtodivertthefluidtoHC2eitherfromthe multi-selectorvalve(fornon-preconcentrationoperations)orfromthe cartridge(fortheelutionstepintheprocencentrationprocedure) Theemploymentofa3-gatevalvepositionedafterthecartridge allowsthepassingofsolutioneithertowaste(duringloadingof sample,flushingandregenerationofthecartridge)ortoHC2during elution

Inordertoallowlargesamplevolumestobepassedthroughthe trappingcartridgerepetitiveloadingwasemployed.Inthismode, theentiresyringeisfilledwithsampleandthenemptiedthrough thepreconcentrationcolumn.Thesyringeisthenfilledagainwith sampleforthesubsequentloading.Thesestepsarerepeateduntil thedesiredsamplevolume(typicallymuchlargerthanthefixed volumeofthesyringe)iscompletelypassedthroughthecartridge Withthissetupthere isnoupperlimitfor theloadingvolume Thesamplealsoneedstobeacidifiedbeforeloadingontothetrap

Inordertoachievecompletemixingwiththeacid,whichwould

bedifficulttoassureinthisapproach ifaspiratedconsecutively, theauxiliarywasmergedinfromaseparatestream.Thismaybe accomplishedusingaseparatesyringepump[14],butamore sim-pleapproachwasusedherewhichisbasedonasplitinlettoallow simultaneous aspirationoftwo separate solutions.A graduated needlevalvewasemployedtoadjustthemixingratiotothedesired value

Allelectronicaswellasthefluidicpartswereassembledinto

astandard19in.rack.Theelectronicpartswerearrangedintwo chassis.Oneoftheseholdsthepowersupplieswiththedifferent requisiteDC-voltagesforthesyringepumps,thevalvesandthehigh voltageunit,andthesecondthecontrolandinterfaceelectronics forthedifferentmodules.Switches,controlsaswellasdisplaysfor thevoltageandcurrentofthehighvoltageunitareaccessibleonthe frontpanels.Theserack-mountedcasescanbeeasilywithdrawnfor modifications.Allfluidiccomponents,includingthepump,valves, holdingcoils,connecting tubingandliquid containers,arefixed ontoapanelsituatedabovethetwoelectronicrackinserts The detailsof a typicalsequence of operations are given in

Table1.Allstepsarefullyautomatedandcanbeperformed unat-tended as controlled by the software program running onthe personalcomputer.Notethatthemovementofthesyringepump

isdeterminedbysettingthedesiredvolumetobeaspiratedor dis-pensedandtheflowrate.Theprotocolstartswithrinsingofthe cartridge(step1)priortorepeatedloadingofsampleontothe pre-concentrationcolumn(step2).Theloadedcartridgeisthenrinsed againwithwater(step3).Inpreparationfortheelectrophoretic separationaflushingoftheSI–CEinterface(step4)isthencarried outbypumpingthebufferthroughtheinterfaceonsimultaneous openingofbothstop-valves(designatedasV1andV2inFig.1) Thecapillaryitself isthenflushed(step5)byslowlyadvancing thesyringepumpwhile bothV1 and V2 areclosedin orderto push all of the dispensedfluid through the separation tubing Elutionof thetrappedanalyteisthen implemented(step6)by passingeluentthroughthecartridgewhiletheswitchingvalveis

atposition2.Oncetheelutedsolutionhasbeencollectedinthe holdingcoilHC2,hydrodynamicinjectiontakesplace(step7), fol-lowedbyelectrophoreticseparation(steps8,9).Thehydrodynamic

C4D Capillary: 25 µm, 60 cm

Grounded interface

Pt

Electrolyte

Solution

1 M NaOH

+/-

HV

Pt HV-interface

Safety case W

W

HC1 (500 µL)

5000 µL Syringe

Needle-Valve 2

Sample

V1

V2

Safety switch

HCl 0.1 M

DI Water

W

2 1

3

7 8 9

Eluent

W

Cartridge (50 µm C18 particles)

HC2 (500 µL) Y- Coupler

Switching Valve

W Regenerating solution

Needle-Valve 1 HCl 0.1 M

Selection Valve

Fig 1. Schematic drawing of the SIA–CE–C 4 D-system for automated electrophoretic separation with on-line SPE preconcentration C 4 D: contactless conductivity detector; HV: high-voltage power supply; Pt: platinum electrode; W: waste; V1, V2: stop valves; HC: holding coil.

split-injectioniscarriedoutbypumpingasampleplugpastthe

capillaryinletintheSIA–CEinterfacewhilepressurizingthe

man-ifoldbyclosingonlyV2.Thesplittingratioisverylarge,asonly

a smallvolume in thenanoliter range canbeinjected intothe

capillary,butnotknownasthesetupisdoneempiricallyby

chang-ingthepositioningoftheneedlevalveuntilagoodcompromise

betweensensitivityandpeakresolutionisobtained.Moredetails

onthis procedure canbe foundin a previouspublication [11]

Separationisimplementedbyapplicationofthehighvoltageof

appropriatepolarityfromthedetectorend,withtheinjectionend

beinggrounded.Rinsingofthemanifolds,interfaceandcapillary

withbuffer(steps 10–12)is then carriedout oncompletion of

theelectrophoreticseparation Finally,thecartridgeis

regener-atedandflushedthoroughly(steps13–16)toreadyitforthenext

preconcentrationoperation.Notethataftereachsolutiondelivery, thesyringeisalsorinsedwithdeionizedwater (steps3,14,16, requiring36s)beforecontinuingwithdeliveryofanothersolution Separationswithorwithoutpreconcentrationcanbeselectedfrom thecomputer.Forseparationwithoutpreconcentration,the sam-pleistransferreddirectlytotheSI–CEinterfaceinsteadofbeing loadedontothecartridgeandsteps1,2,3,4,6,13,14,15and16 areomittedfromtheprotocol

3.2 Determinationofpharmaceuticals

Inordertodemonstratethesystemamethodforthe determi-nationofdrugresidueswasimplemented.Manypharmaceuticals are classified as environmental contaminants due to their low

Table 1

Typical operation sequence with preconcentration.

selection valve

Volume dispensed (␮L)

Flow rate (␮L s−1)

Position of V1 Position of V2 Position of

switching valve

* Steps 2a and 2b are repeated n times until the desired sample volume has been loaded.

**

leadtoadverseeffectonhumanbeingsandecosystems[27–30]

ThedeterminationofdrugresiduesinwaterwithCEusingUVor

mass-spectrometricdetectionhasbeenreportedrepeatedlywith

andwithout usingelectrokineticstacking preconcentration,see

forexample[31–34].Queketal.[35]alsoreportedthe

determina-tionof13pharmaceuticalsdeemedpotentialpollutantsbyCE–C4D

Nopreconcentrationmethodswereemployedanddetection

lim-its of typically 1␮M were obtained In this study,four widely

usedpharmaceuticals,namelyibuprofen,diclofenac,naproxenand

bezafibrate,wereselectedasexemplaryanalytes

3.2.1 Separationbuffer

Fortheelectrophoreticseparationthebackgroundelectrolyte

(BGE)based ontheonereportedbyQueket al.[35] consisting

ofaTris/lacticacidbufferatpH8andcontaining

hydroxypropyl-␤-cyclodextrin(HP-␤-CD)was employed.ThepHvalue assures

deprotonation and thus anionic ionization of the compounds

CyclodextrinHP-␤-CDwasaddedintotheBGE asa complexing

reagentforseparationofibuprofenanddiclofenac.Itdoesnot

pos-sessacharge,andthusdoesnotcontributetotheconductivityof

theBGE.Asdiscussedbelow,thereleaseoftheanalytesfromthe

trappingmaterialrequiresaneluentwithapHwhichleadsto

ion-izationaswellastheinclusionofasubstantialfractionofanorganic

solvent.ForthisreasontheinjectionofstandardsinaTris/lactic

acidbufferwithanacetonitrilecontentof37.5%(v/v)wastested

Asshowninelectropherogram(b)ofFig.2,broadpeakswithlow

sensitivityandpoorresolutionwereobtained,whereasthe

injec-tionofapurelyaqueoussolution(nobuffer)ofthefourspeciesled

tothegoodseparationofelectropherogram(a).Itwasassumedthat

thesharperpeaksofthistoptracewereduetoatransientstacking

effectduetothelowconductivityofthesampleplug.Inorderto

induceasimilarperformancewiththebackgroundofthegiven

elu-entthecompositionoftheBGEwasadjusted,i.e.theconductivityof

thelatterwasincreasedbyincreasingthebufferconcentration.As

seeninelectropherogram(c)ofFig.2,indeedagoodperformance

couldalsobeobtainedwiththeseconditions.Buffersofevenhigher

concentrationswerealsotried,butledtonoisybaselines,

presum-ablyduetoJouleheating.Thusthereisalimittothisapproach.The

performancedataforthedirectdeterminationwithout

preconcen-trationofthefouranalgesicsontheCEsystemaregiveninTable2

Thedetectionlimitsachievedareintherangefromof0.8to1.5␮M

Calibrationswerecarriedoutfortherangefrom2.5␮Mto100␮M

(6concentrations)andlinearitywasobtained

3.2.2 Sampleloading

The four investigated drugs are carboxylates, and thus are

negativelychargedinwaterofneutralpH.Foreffectiveadsorption

ontoC18 sorbentparticles,thesample needstobeacidified in

ordertorenderthecompoundsintheirneutralform.Theaddition

of0.1MHClwasfoundtobesuitableforsampleacidificationsince

aminorfraction(lessthan1%,v/v)ofthisstronginorganicacid

solutionis sufficienttoadjustthepHof thesamplesbelowthe

pKaofthesecarboxylates(<4).TheamountofHClsolutionadded

450 400

350 300

250

Migration time (s)

0.2 V

(a)

(b)

(c)

1

4 3 2

1

4 3 2

Fig 2.CE separations of pharmaceuticals in different matrices (a) Matrix: water; BGE: 9 mM Tris/5 mM lactic acid/1 mM HP-␤-CD; (b) matrix: 9 mM Tris/5 mM lactic acid in water (62.5%, v/v) mixed with acetonitrile (37.5%, v/v); BGE: 9 mM Tris/5 mM lactic acid/1 mM HP-␤-CD; (c) matrix: 9 mM Tris/5 mM lactic acid in water (62.5%, v/v) mixed with acetonitrile (37.5%, v/v); BGE: 36 mM Tris/20 mM lactic acid/1 mM HP-␤-CD CE conditions: hydrodynamic split-injection with a dispensed sample vol-ume of 0.5 mL and a setting of 0.11 on the micrometer screw of the needle valve,

E = 400 V cm−1, feedback resistor for C 4 D = 1 M, capillary of 25 ␮m inner diame-ter and 41 cm effective length (1) Ibuprofen; (2) bezafibrate; (c) diclofenac; (4) naproxen.

can beconsidered negligible and therefore doesnot affectthe enrichmentfactor.Notethat higherconcentrationsof HClwere avoided to preventcorrosion of the graduated needle valve If higherconcentrationsofacidshouldberequiredinapplications where samples witha high buffer capacity occur then it may thereforenotbepossibletoemploythissimpleapproachandthe useofasecondpumpforfeedingintheacidmaybenecessary ForloadingsampleontotheSPEcartridgeinprinciplethe high-estflow ratewhichallowscompleteanalytetrappingshouldbe employedinordertoobtaintheshortestpossibleanalysistimes Theconfigurationofthestepper-motordrivenpumpinprinciple allowsavariationoftheflowratefromsome␮Ls−1tomorethan

1000␮Ls−1,dependingonthesyringecapacity.However,inorder

tominimizediffusionanddilutionduringthedeliveryofsolutions narrowtubingsof0.02in.internaldiameterwereemployedwhich limitedthemaximumpossibleflowrateto400␮Ls−1duetothe creationofbackpressures.However,inpracticeitwasobservedthat forflowrateshigherthanabout150␮Ls−1theoccasional forma-tionofbubblesinthemanifoldduetooutgassingwasobserved.In ordertoavoidthiseffecttheflowratewasthussetto140␮Ls−1 Efficientretentionoftheanalyteswasfoundforthisflowrateas thesignalswerenotdeterioratedcomparedtolowerflowrates

Table 2

Determination of standards of 4 drug compounds.

coefficient * (r)

LOD ** (␮M) Reproducibility of

migration time RSD% (n = 4)

Reproducibility of peak area RSD% (n = 4)

* 6 concentrations, 2.5–100 ␮M.

** Based on peak heights corresponding to 3 times the baseline noise.

Conditions: capillary 25 ␮m ID, l = 41 cm; E = 400 V cm −1 ; BGE: 36 mM Tris/20 mM lactic acid/1 mM HP-␤-CD.

80

60

40

20

0

40 30

20 10

0

Diclofenac Ibuprofen Naproxen Bezafibrate

CH3CN in the buffer (%)

Fig 3. Effect of acetonitrile concentration on efficiency of elution from the trap.

The eluent was prepared by addition of acetonitrile to an aqueous solution of 9 mM

Tris and 5 mM lactic acid Peak areas were obtained from electropherograms with

the following CE conditions: E = 400 V cm −1 , feedback resistance for C 4 D = 1 M,

capillary of 25 ␮m inner diameter and 41 cm effective length, BGE composed of

36 mM Tris/5 mM lactic acid/1 mM HP-␤-CD, hydrodynamic split-injection with a

dispensed sample volume of 0.5 mL and a setting of 0.11 on the micrometer screw

of the needle valve.

3.2.3 Analyteelution

Foragoodpreconcentrationfactorandlowpossiblesample

vol-umes,theeluentvolumemustbekeptlow.Preliminarytrialswith

purpose-mademicroSPEcartridgestominimizetheeluentvolume

werenotsuccessfulduetohighbackpressurecreatedwhenpassing

solutionsthrough.Commercialcartridgesofthesmallestavailable

volume(containing100mgsorbentparticles)with2endcapped

fritsofappropriateboreholes(20␮mdiameter)werethenselected

forpreconcentration.Accordingly,theeluentvolumemustatleast

matchtheinternalvolumeofthecartridgewhichwas

approxi-mately250␮Lfortheproductused.Aneluentvolumeof500␮L

wasthusemployedwithalowflowrateof25␮Ls−1.The

composi-tionoftheeluentsolutionmustfulfilltworequirementsinorderto

achievecompleteelutionofthefourprotonateddrugcompounds

retainedontheC18cartridge.Firstly,thepHoftheeluentshould

behigherthanthepKasoftheretainedcompound(s)inorderto

rendertheminthechargedformtofacilitatedesorption.Secondly,

asuitableorganicsolventmiscibleinwatermustbeaddedinto

theeluenttoacertainpercentage.Inaddition,becausepartofthe

eluentcontainingtheanalytesisinjectedintotheseparation

capil-larytheeluentshouldcontainonlythecomponentspresentinthe

BGEusedintheseparationinordertoavoidlargesystempeaks

Accordingly,a solutioncomposedof9mM Trisand5mMlactic

acidwasemployedforelution.ThiseluenthasapHof8,which

isexactlythepHoftheBGE.Differentorganicsolventsfrequently

usedinHPLC,namelymethanol,acetonitrileandtetrahydrofuran,

werethentestedfortheTris/lacticacideluent.Notethattheorganic

proportionmustbekeptaslowaspossibletominimizethe

pos-sibilityofbubbleformationcausedbyevaporationoftheorganic

solventduringtheseparation.Methanol,duetoitsweakelution

strengthwasfoundtorequireaveryhighconcentration(>50%,v/v)

forgoodelution.Tetrahydrofuran,whichpossessesthehighest

elu-tionstrengthamongallcommonlyusedorganicsolventsmiscible

withwater,thoughofferingveryefficientdesorptionevenatlow

concentrations(lessthan25%,v/v),wasfoundtodisturbthe

resolu-tionofthepeaksofnaproxenandibuprofenintheelectrophoretic

separation.Acetonitrile,withintermediateelutionstrength,was

thereforechosenastheorganicadditive.Theeffectofthe

concen-trationofacetonitrileonelutionefficiencyisshowninFig.3.The

measurementswererepeatedtwice,andtheresultswerewithin

0.5

0.4

0.3

0.2

0.1

0.0

50 40

30 20

10

0.5

0.4

0.3

0.2

0.1

0.0

50 40

30 20

10

Run Number

(A)

(B)

Ibuprofen

Bezafibrate

Diclofenac

Naproxen

Fig 4. Reproducibility of peak areas for the preconcentrated pharmaceuticals dur-ing continuous operation of 10 h per day for 3 successive days For each run 15 mL solution of standard mixtures of 0.5 ␮M in deionized water was loaded onto a C18 cartridge followed by an elution with 0.5 mL eluent (9 mM Tris/5 mM lactic acid (62.5%, v/v) + CH 3 CN (37.5%, v/v)) Peak areas were obtained for a preconcentration factor of 30 Other CE conditions as for Fig 3

±1.5%.Thefourretaineddrugsareelutedwithdifferentefficiencies, withdiclofenacbeingthecompoundwhichismosteasilydesorbed Completeelutionoccursfor acetonitrileconcentrationsof37.5% (v/v)orhigherinaneluentcontaining9mMTrisand5mMlactic acid.Notethatnosignificantelutionwasobservedwhenonly ace-tonitrileinwaterwasemployedastheeluentevenatacetonitrile concentrationshigherthan50%(v/v)

Followingtheanalyte elutionthe cartridgewasregenerated

by first passing a relatively largevolume of 5mLof the 9mM Tris/5mMlacticacidbuffermixedwithacetonitrile(50%,v/v)in ordertoassureefficientremovalofanyorganicspeciesthatmay stillberetainedonthecartridgeafterpreconcentration.Thetrap wassubsequentlyrinsedwith0.5mLof0.1MHCltodissolveany precipitatesthatmayhaveformedonthesurfaceofthesorbent particlesduringtheprecedingoperationsathigherpH.Finally,the cartridgewasrinsedwithdeionizedwater(5mL)beforethenext trappingsequence

3.3 Performanceandsampleanalysis

In order toevaluatethe potentialfor unattended operation, thesystemwassetupforatestrunforaperiodof10hperday over3continuousdays,inwhichrepeatedpreconcentrationsfrom

15mL of sample solution with an enrichmentfactor of 30 (as followsfromthevolumeratiosforefficienttrappingandelution) andCEmeasurementsofthepharmaceuticals(0.5␮Mpreparedin

0.25

0.20

0.15

0.10

0.05

0.00

6 5

4 3

2 1

0.30

0.25

0.20

0.15

0.10

0.05

0.00

6 5

4 3

2 1

Run Number

(A)

(B)

Ibuprofen

Bezafibrate

Diclofenac

Naproxen

Fig 5.Reproducibility of peak areas for the preconcentration and CE determination

of the pharmaceuticals spiked in tap water For each operation, 375 mL solution of

standard mixtures of 0.01 ␮M prepared in filtered tap water was loaded onto a C18

cartridge followed by an elution with 0.5 mL eluent (9 mM of Tris/5 mM lactic acid

(62.5%, v/v) + CH 3 CN (37.5%, v/v)) Peak areas were obtained for an enrichment factor

of 750 Other CE conditions as for Fig 3

deionizedwater)werecarriedout.Thepreconcentrationprocess

forthisenrichmentfactorrequires4min,whereasittakesabout

25minfortheentireprotocolincludingpreconcentration,analysis

andallflushingoperationstobecompleted.Theresultsforpeak

areasareshowninFig.4.Themaximumdeviationisabout±8%,

whichisdeemedacceptableconsideringthatthisdeviationisdue

totheaccumulationoferrorsofalloperations,i.e.sampleloading,

elution,injectionand separation.Without thepreconcentration

procedure,deviationsofupto±4%wereobservedforautomated

CEseparations [11] No bias of peak areas was recorded after

morethan50continuousrunsin3days,whichdemonstratesthe

suitabilityofthesystemforunattendedoperation

Theoperationalperformancewasfurtherevaluatedby

carry-ingoutpreconcentrationsundermoresevereconditions,inwhich

solutionsofthepharmaceuticalswereprepareddirectlyinatap

watermatrixinsteadofdeionizedwater.Theloadingofalarge

vol-umeof375mLoftapwaterspikedwithpharmaceuticals(10nM)

anddesorptionwith0.5mLeluentwasrepeatedseveraltimeswith

thesamecartridge.Theenrichmentfactorinthiscaseis750,which

isverylargeforSPEpreconcentration.Theresultsforpeakareas

areshowninFig.5.Themeasurementswererepeatedtwiceand

theresultscouldbereproducedwithin±4%.It isapparentthat

after3successivepreconcentrations,thepeakareasofall

pharma-ceuticalsdecreasedsignificantly.Thisisthoughttobeduetothe

sorbentmaterialbecomingpermanentlysaturatedwithstrongly

boundspeciespresentinthematrixinminorquantitieswhichare

notreleasedintheelutionstep.Thus,fortheseconditionsatotal

loadingvolumeof1000mLshouldnotbeexceededbefore

replac-ingthetrap.Furthertestswerecarriedoutbyspikingwaterfrom

theriverRhine(Basel,Switzerland)withthepharmaceuticalsas

100 80 60 40 20 0

200 160

120 80

40 Loading volume (mL)

Recovery (%) DiclofenacIbuprofen

Naproxen Bezafibrate

Fig 6.Recoveries of the four pharmaceuticals spiked in Rhein river water at differ-ent preconcentration factors The spiked concentrations were varied from 25 nM to

250 nM depending on the loading volume in order to keep the concentrations after enrichments fixed at 10 ␮M Other conditions as for Fig 3

itwasexpectedthat this effectwouldbemore pronouncedfor themorecomplexmatrix.Thiswasindeedthecase.Theresults aregiveninFig.6,wheretherecoveryindependenceofthe load-ingvolumeisshown.Themeasurementswererepeatedtwiceand theresultscouldbereproducedwithin±4.5%.Forequivalent com-parisonofrecovery,theconcentrationinthedonorsolutionwas varied(25–250nM)accordingtothevolumepassedthroughthe trap(from200mLto20mL)sothatthenominalconcentrationin theelutingsolutionforcompleteextractionwasfixedat10␮M.The dataofFig.6demonstratesthatasignificantreductioninrecovery forriverwaterspikedwiththepharmaceuticalsoccurredwhena samplevolumeof50mLwasexceeded.Thisindicatesageneral lim-itationofthemethodwhichhastobecarefullyassessedforatask

athand

ThesystememployingtheoptimizedSPE–CE–C4Dconditions wasthenusedtoanalyzeawatersampletakenfromtheoutlet

ofawastewatertreatmentplanofahospitalinHanoi,Vietnam Electropherogramsforthesamplewithandwithoutenrichment areshowninFig.7.Ibuprofen,whichisbarelydiscernedwithout preconcentration,canbeclearlyobservedafterenrichment.Some otherminorpeaks alsoappear intheelectropherogram forthe enrichmentfactorof40,butnoeffortwasmadetoidentifythese species

400 380 360 340 320 300 280 260

Migration time (s)

200 mV

Without preconcentration

Enrichment factor = 20

Enrichment factor = 40

Fig 7. Analysis of a water sample taken from a wastewater treatment plan of a

4 Conclusions

TheautomatedSPEtechniquecouldbereadilyimplementedon

theSIA–CE–C4Dsystemandprovidesarobustandstraightforward

meanstoextendthedetectionlimitofCE–C4D.Enrichmentfactors

ofupto750couldbedemonstrated.Forthepharmaceuticalstested

alowerlimitofquantification(LOQ)ofabout5␮Mispossibleby

directCE–C4Danalysiswithoutpreconcentration.Consequently,

withpreconcentrationLOQsinthelownM-rangearepossibleif

thesamplevolumeissufficient.Inpracticethesamplematrixmay

alsoimposealimittothevolumeofsamplewhichcanbepassed

throughthetrap,andthusthehighestpreconcentrationfactorthat

canbeachieved

Duetotheemploymentofelectrophoreticseparationand

con-ductivitydetection theconstruction ofthe entireinstrument is

relativelysimple.Thecomponentsforthestandardized19-in

sys-temarewidelyavailablecommercially,thefluidicsectionisbased

onstandardparts,andonlyafewspecialcomponentshadtobe

manufacturedforpurpose.Thecompactall-in-one designofthe

overallinstrumentallowseasytransportanddeploymentfor

auto-matedon-sitemonitoringapplicationsisreadilypossible

Acknowledgments

Theauthors would liketothank theSwiss NationalScience

Foundation(GrantNo.200020-137676/1)andtheNational

Foun-dation for Science and Technology Development (NAFOSTED),

Vietnam(GrantNo.104.07-2010.45)forfunding

References

[1] R Ramautar, G.J de Jong, G.W Somsen, Electrophoresis 33 (2012) 243–250.

[2] M.C Breadmore, M Dawod, J.P Quirino, Electrophoresis 32 (2011) 127–148.

[3] S.L Simpson, J.P Quirino, S Terabe, J Chromatogr A 1184 (2008) 504–541.

[4] R Ramautar, G.W Somsen, G.J de Jong, Electrophoresis 31 (2010) 44–54 [5] P Puig, F Borrull, M Calull, C Aguilar, Anal Chim Acta 616 (2008) 1–18 [6] P Puig, F Borrull, M Calull, C Aguilar, TrAC Trends Anal Chem 26 (2007) 664–678.

[7] F.W.A Tempels, W.J.M Underberg, G.W Somsen, G.J de Jong, Electrophoresis

29 (2008) 108–128.

[8] L Saavedra, C Barbas, J Biochem Biophys Methods 70 (2007) 289–297 [9] J.R Veraart, H Lingeman, U.A.T Brinkman, J Chromatogr A 856 (1999) 483–514.

[10] P Kubá ˇ n, B Karlberg, Anal Chim Acta 648 (2009) 129–145.

[11] T.D Mai, S Schmid, B Müller, P.C Hauser, Anal Chim Acta 665 (2010) 1–6 [12] T.D Mai, P.C Hauser, Talanta 84 (2011) 1228–1233.

[13] T.D Mai, P.C Hauser, Electrophoresis 32 (2011) 3000–3007.

[14] B Horstkotte, O Elsholz, V.C Martín, Talanta 76 (2008) 72–79.

[15] P Kubá ˇ n, P.C Hauser, Electrophoresis 32 (2011) 30–42.

[16] A.A Elbashir, H.Y Aboul-Enein, Biomed Chromatogr 24 (2010) 1038–1044 [17] M Trojanowicz, Anal Chim Acta 653 (2009) 36–58.

[18] P Kubá ˇ n, P.C Hauser, Anal Chim Acta 607 (2008) 15–29.

[19] P Kubá ˇ n, P.C Hauser, Electroanalysis 16 (2004) 2009–2021.

[20] P Kubá ˇ n, P.C Hauser, Electrophoresis 30 (2009) 176–188.

[21] J.G.A Brito-Neto, J.A.F da Silva, L Blanes, C.L do Lago, Electroanalysis 17 (2005) 1207–1214.

[22] J.G.A Brito-Neto, J.A.F da Silva, L Blanes, C.L do Lago, Electroanalysis 17 (2005) 1198–1206.

[23] P Kubá ˇ n, P.C Hauser, Electrophoresis 25 (2004) 3398–3405.

[24] P Kubá ˇ n, P.C Hauser, Electrophoresis 25 (2004) 3387–3397.

[25] Y.S Ding, K Rogers, Electrophoresis 31 (2010) 2602–2607.

[26] L Zhang, S.S Khaloo, P Kubá ˇ n, P.C Hauser, Meas Sci Technol 17 (2006) 3317–3322.

[27] M.J.M Wells, K.Y Bell, K.A Traexler, M.L Pellegrin, A Morse, Water Environ Res 82 (2007) 2095–2170.

[28] D.W Kolpin, E.T Furlong, M.T Meyer, E.M Thurman, S.D Zaugg, L.B Barber, H.T Buxton, Environ Sci Technol 36 (2002) 1202–1211.

[29] T.A Ternes, Water Res 32 (1998) 3245–3260.

[30] R Hirsch, T Ternes, K Haberer, K.L Kratz, Sci Total Environ 225 (1999) 109–118.

[31] W Ahrer, E Scherwenk, W Buchberger, J Chromatogr A 910 (2001) 69–78 [32] V.J Drover, C.S Bottaro, J Sep Sci 31 (2008) 3740–3748.

[33] A Macià, F Borrull, M Calull, C Aguilar, J Chromatogr A 1117 (2006) 234–245.

[34] A Macià, F Borrull, C Aguilar, M Calull, Electrophoresis 25 (2004) 428–436 [35] N.M Quek, W.S Law, H.F Lau, J.H Zhao, P.C Hauser, S.F.Y Li, Electrophoresis

29 (2008) 3701–3709.