b917150a

PAPERwww.rsc.org/loc|LabonaChip

Electrochemicalsensinginpaper-basedmicrofluidicdevices†

ZhihongNie,aChristianA.Nijhuis,aJinlongGong,aXinChen,aAlexanderKumachev,bAndresW.Martinez,aMaxNarovlyanskyaandGeorgeM.Whitesides*aReceived19thAugust2009,Accepted4thNovember2009

FirstpublishedasanAdvanceArticleontheweb3rdDecember2009DOI:10.1039/b917150a

Thispaperdescribesthefabricationandtheperformanceofmicrofluidicpaper-based

electrochemicalsensingdevices(wecallthemicrofluidicpaper-basedelectrochemicaldevices,mPEDs).ThemPEDscomprisepaper-basedmicrofluidicchannelspatternedbyphotolithographyorwaxprinting,andelectrodesscreen-printedfromconductinginks(e.g.,carbonorAg/AgCl).WedemonstratedthatthemPEDsarecapableofquantifyingtheconcentrationsofvariousanalytes(e.g.,heavy-metalionsandglucose)inaqueoussolutions.Thislow-costanalyticaldeviceshouldbeusefulforapplicationsinpublichealth,environmentalmonitoring,andthedevelopingworld.

Downloaded on 08 September 2011Published on 03 December 2009 on http://pubs.rsc.org | doi:10.1039/B917150AIntroduction

Thisarticledescribesthefabricationandtheperformanceofpaper-basedelectrochemicalsensingdevices.WecallthesedevicesmPEDs(microfluidicpaper-basedelectrochemicaldevices).ThemPEDsarecapableofquantifyingtheconcentra-tionsofvariousanalytesinaqueoussolutions,includingbiologicalfluidssuchasurine,serumandblood.ThemPEDscomprisemicrofluidicchannels,fabricatedfrompatternedpaper(eitherchromatographypaperorapolyester–celluloseblend),andelectrodesthatwerescreen-printedonchromatographypaperorpolyesterfilm(Fig.1).Threeelectrodes(aworkingelectrode,acounterelectrode,andareferenceelectrode)weremadefromcarboninkorAg/AgClink.

ThisworkwascarriedoutindependentlyofasimilarandcomplementarystudybyDungchaietal.,1inwhichmPEDswerefabricatedandusedforelectrochemicaldetectionofglucose,lactate,anduricacidinbiologicalsamples.UsingmPEDs,wehavedemonstratedthedetectionandquantificationofglucoseinartificialurineusingachronoamperometricanalysisbasedonglucoseoxidase.Thedetectionofglucoseusingthedeviceshowedhighsensitivityandaccuracyoverthefullrangeofclinicallyrelevantconcentrationsofglucoseinurine.ThedetectionlimitofglucoseinthecurrentmPEDisabout0.22mM(correspondingto4mgmLÀ1),whichislowerthantheapprox-imately1.0mMspecifiedbyconventionalglucometers,and0.5mMobtainedbyacolorimetricdetectionmethodreportedpreviously.2,3Weestimatedthesensitivityoftheglucoseanalysistobe0.43mAmMÀ1mmÀ2.WefurtherdemonstratedtheuseofmPEDsintheselectiveanalysisofPb(II)inanaqueoussolution

DepartmentofChemistry&ChemicalBiology,HarvardUniversity,Cambridge,MA,02138,USA.E-mail:gwhitesides@gmwgroup.harvard.edu;Fax:+1016174959857;Tel:+1016174959430bDepartmentofChemistry,UniversityofToronto,80StGeorgeStreet,Toronto,Ontario,CanadaM5S3H6

†Electronicsupplementaryinformation(ESI)available:GeometryofthehydrodynamicmPEDfortheanalysisofmetals,estimationofdiffusioncoefficient,Cottrellplotfortheanalysisofglucose,chronoamperometricanalysisofglucoseinbloodplasmaandbovineblood,performanceofgoldelectrodesonthepaper-baseddevices.SeeDOI:10.1039/b917150a

acontainingamixtureofPb(II)andZn(II).Thecontinuouswickingofthesamplesolutionacrosstheelectrodesdramaticallyenhancedtheefficiencyofthedepositionofmetalsduringanodicstrippingvoltammetry,andimprovedthesensitivityandreli-abilityofthedetection.ThemeasurementofPb(II)showedalimitofdetectionof1.0ppb.ThisvalueismuchlowerthantheWHO(WorldHealthOrganization)guidelinevalue(<10ppb)forthesafelevelofleadindrinkingwater.4Wehaverecentlydevelopedmethodsforfabricatingmicro-fluidicdevicesfrompatternedpaper,anddemonstratedthat

Fig.1(a)Schematicofapaper-basedelectrochemicalsensingdevice.Thesensorcomprisesthreeelectrodesprintedonapieceofpapersubstrate(orplastic)andapaperchannel.Thepaperchannelwasinconformalcontactwiththeelectrodes,andwasheldinplacebydouble-sidedadhesivetapesurroundingtheelectrodes.Aphotographofapaper-basedelectrochemicalsensingdevicefortheanalysisofglucose(b),andahydrodynamicpaper-basedelectrochemicalsensingdeviceforthemeasurementofheavy-metalions(c).Thedeviceconsistsoftwoprintedcarbonelectrodesastheworkingandcounterelectrodes,andaprintedAg/AgClelectrodeasthepseudo-referenceelectrode.ThepaperchannelwasfabricatedbypatterningSU-8asahydrophobicbarrierforaqueoussolution.Thepaperchannelin(b)wascoloredwithredinktoenhanceimaging.Thedashedlinein(c)indicatestheedgeofthepaperchannel.Thescalebaris4mm.

ThisjournalisªTheRoyalSocietyofChemistry2010LabChip,2010,10,477–483|477

View Online

Downloaded on 08 September 2011Published on 03 December 2009 on http://pubs.rsc.org | doi:10.1039/B917150Apaper-basedmicrofluidicdevicescanbeusedforcolorimetricdetectionofglucoseandproteininartificialurine.2,5–7Low-cost,paper-baseddiagnosticandanalyticaldevicesareattractiveforuseindevelopingcountries,inthefield,orinhomehealth-caresettings.8,9Althoughcolorimetricassaysarethebestchoiceforsomeapplications,electrochemicalsensingprovidesamoreversatileandquantitativemethodologyforothers.10ThemPEDsdescribedinthisarticleprovideasimple,fast,andlow-cost($2centsperdevice)platformforthequantitativedetectionofanalytes(biologicalandinorganicspecies)inaqueoussolutions,includingbodyfluids.Thistypeofdetectionhasfiveadvantages:(i)itislight-weight,portable,single-useanddisposable;(ii)itisflexibleandfoldable;(iii)ithasexcellentreproducibilitywithhighsensitivityandaccuracy;(iv)itdoesnotrequireprofessionalmedicalpersonnelorcomplicatedinstruments;and(v)itallowsfortheintegrationofhigh-densitydetectionsystemsintoasmalldevice.

(103mgLÀ1),andbismuth(III)(103mgLÀ1)wereobtainedfromAldrich,USA,anddilutedasrequired.0.1Macetatebuffer(pH4.5)wasusedasasupportingelectrolyte.Apparatus

Allchronoamperometricmeasurementswereperformedwithabipotentiostat(PINEInstrumentCompany,ModelAFCBP1).AmodularelectrochemicalsystemAUTOLABequippedwithPGSTAT12wasusedincombinationwithGPESsoftware(EcoChemie)fortheanodicstrippingvoltammetricmeasure-mentsofheavy-metalions.Fabricationofthedevices

Electrodes.Wefabricatedathree-electrodesensorbyscreen-printingcarbonink(orAg/AgClinkforareferenceelectrode)onapieceofpaperorpolyesterfilm.WegeneratedastencilforprintingbydesigningpatternsofelectrodesusingAdobeFree-handÒ,followedbycuttingthepatternintodouble-sidedadhe-sivetapeusingalaser-cutter(VersaLASERVLS3.50,UniversalLaserSystemsInc.).Wetapedthestencilontopofapaperorplasticsubstrate,andfilledtheopeningsofthestencilwithink.Webakedtheelectrodesonahotplateat100󰀄Cfor30min.Aftertheinkdried,weremovedtheprotectivebackinglayerofthetape,andlefttheadhesivelayeronthesubstratefortheassemblyofapaperchannelonthetopofelectrodes.Thethicknessoftheelectrodeswasapproximately100mm.Atypicalworkingandcounterelectrodehaddimensionsof1.5cminlengthand4mminwidth,andatypicalreferenceelectrodehaddimensionsof1.5cminlengthand3mminwidth.

Microfluidicchannels.Paper-basedmicrofluidicchannelswerefabricatedbypatterningchromatographypaper(Whatman1Chr)orpolyester–celluloseblendpaper(VWRÒSpec-Wip)byphotolithography2,5–7orwaxprinting.13Briefly,wesoakedapieceofpaperwithSU-82010photoresist,bakeditat95󰀄Cfor5mintoremovesolvents,andphotoexposedittoUVlightfor10sthroughaphotomask.Theunpolymerizedphotoresistwasremovedbysoakingthepaperinacetoneandwashingthreetimeswithisopropanol.Thepolymerspatternedonthepaperformhydrophobicbarrierstoconfineliquidsinthemicro-channel.Thepapermicrofluidicchannelhaddimensionsof4mminwidthand100mminheight(determinedbythethicknessofthepaper),Fig.1.Thepaper-basedchannelswereassembledontotheelectrodesusingdouble-sidedadhesivetapes.Thecontactareabetweenapaperchannelandaworkingelectrodewas4mmby4mm.

Chronoamperometricmeasurements

Chronoamperometricexperimentswereperformedusinga500mVsteppotential(vs.acarbonpseudo-referenceelectrode)togeneratethecalibrationcurve;theseexperimentsusedglucosewithconcentrationsrangedfrom0to22.2mM(correspondingto400mgdLÀ1).Solutions(glucoseoxidase250UmLÀ1,K3[Fe(CN)6]600mM,KCl1.0MinpH7.0PBSbuffer)con-tainingglucosewithdifferentconcentrationsweremeasured(eachsamplewasexaminedeighttimes).Wespottedthesolutionofenzymeontopofthepapermicrochannel.Thesolutionof

ThisjournalisªTheRoyalSocietyofChemistry2010

Experimentaldesign

Choiceofmaterials

Weusedpaperasthesubstrateforelectrochemicaldetection.WeusedconductivecarboninkandAg/AgClinktofabricatetheelectrodesusedinthework.11Theconductiveinkelectrodeshavefouradvantages:(i)theyareinexpensive;(ii)theyarerobust;(iii)theyareeasytofabricate;and(iv)theyarewidelyusedinbothindustrialanduniversityresearch.12WeevaluatedthequalityofthemPED-basedelectrochemicalcellsbycyclicvoltammetricmeasurements,usingferrocenecarboxylicacidasastandard,becauseinbulksolutiontheelectrochemicalreactionisreversibleunderambientconditionsatroomtemperature.Fabricationanddesignofthedevice

Wefabricatedathree-electrodesensorbyscreen-printingcarbonink(orAg/AgClink)ontoapieceofpaperorpolyesterfilm.Theprintingprocessisfast,simple,andinexpensive,andsuitableformassproduction.Paper-basedmicrofluidicchannelswerefabri-catedbypatterningchromatography(Whatman1Chr)orpolyester–celluloseblendpaper(VWRÒSpec-Wip)byphotoli-thography2,5–7orwaxprinting.13Wechosethesepapersinthisworkbecausetheydonotdeformonthesurfaceofelectrodeswhentheyarewettedbyfluids.Inordertoachieveconformalcontactbetweenpaperchannelsandelectrodes,thepatternedpaperwastapedontothesubstratethatsupportedtheelectrodesusingdouble-sideadhesivetapes(Fig.1).

Experimental

Chemicalreagents

Carbonink(E3456)andAg/AgClink(AGCL-675C)werepurchasedfromErconInc(Wareham,MA)andConductiveCompound(Hudson,NH),respectively.Glucoseoxidase(136300UmgÀ1,Aspergillusniger),glucose,andpotassiumferricyanidewerepurchasedfromAldrichandusedasreceived.Stocksolutionsofb-D-glucosewerepreparedinaPBSbuffer(pH7.0)andallowedtomutarotateovernightbeforeuse.AtomicabsorptionstandardsolutionsofPb(II)(104mgLÀ1),Zn(II)

478|LabChip,2010,10,477–483

View Online

Downloaded on 08 September 2011Published on 03 December 2009 on http://pubs.rsc.org | doi:10.1039/B917150Aenzymewasdistributedevenlyinthepaperchannelduetothecapillarywicking.Afterthesolutiondried,theenzymewasuniformlyabsorbedinthepaper.Whenasolutionofanalyteswasaddedintothemicrochannel,thesolutiondistributedandmixedwellwiththepre-loadedenzyme.Inanothersetofexperiments,wepremixedtheenzymesolutionwithglucosesamplesbeforethechronoamperometricmeasurements.Wedidnotobserveobviousdifferencebetweenthetwomethods.Papercanalsobeeasilymodifiedtoimmobilizeenzymesifnecessary.Allmeasurementswereconductedatroomtemperatureunderambientconditions.Weusedacarbonworkingelectrode(onaplasticsubstrate)withasurfaceareaof16mm2incontactwiththefluidforthedetectionofglucoseinthebulksolution.Weaveragedtheeightmeasurementsofcurrentreadoutandcalcu-latedthecorrespondingstandarddeviation.Anodicstrippingvoltammetry

Strippingvoltammetricmeasurementswereperformedbyinsitudepositionofthebismuth(500mgLÀ1)andthetargetmetalionswithconcentrationsrangedfrom0to100ppb(mgLÀ1).Non-deaeratedsolutionswereusedforallmeasurements,andeachsamplewasmeasuredeighttimes.Allmeasurementswerecarriedoutusingthesquare-waveanodicstrippingvoltammetry(SWASV)withafrequencyof20Hz,apotentialstepof5mV,andanamplitudeof25mV.SWASVexperimentscomprisedanelectrochemicaldepositionstepatÀ1.2Vfor120s,anequili-brationperiodof30s,andasquare-wavevoltammetricstrippingscanusuallyfromÀ1.2toÀ0.5V.Beforeeachmeasurementapre-conditioningstep(forcleaningoftheelectrode)atapotentialof+0.5Vwasappliedfor60s.Detectionmethodandprinciple

Glucose.WedetectedglucoseinthemPEDsbychro-noamperometricmethod.Thereactionsfortheglucosedetectionwere:

ÂÃ

Glucoseþ2K3FeðCNÞ6þH2O

ÂÃglucoseoxidase

(1)󰀄󰀄󰀄󰀄󰀄󰀄󰀄󰀄!Gluconicacidþ2K4FeðCNÞ6

anodicelectrode

Fig.2(a)Cyclicvoltammogramsof2.0mMferrocenecarboxylicacidin

0.5MKClaqueoussolution(pH¼7.0)inamPEDatvariousscanrates(ascendingalongy-axis):50,100,200,300,400,and500mVsÀ1.Weuseda4mmby4mmcarbon-printedelectrodeastheworkingelectrode,andaprintedAg/AgClelectrodeasthereferenceelectrode.(b)Theplotofanodicpeakcurrentvs.thesquarerootofthescanrate(n1/2)forCVexperimentsconductedonapaperdevice(C)andinabulksolution(B).Thesolidlinesrepresentalinearfitto(C)withregressionequation:y¼À3.6+3.5x(R2¼0.998,n¼8),andalinearfitto(B)withregressionequation:y¼À7.1+5.6x(R2¼0.999,n¼8).

Resultsanddiscussion

Electrochemicalcharacterizationofpaper-basedelectrochemicaldevices

Weusedferrocenecarboxylicacidasamodelredox-activecompoundtocharacterizeelectrochemicalbehaviorofmPEDs(Fig.2a).ThepeakshapeoftheCVsshowsatypicalreversible(Nernstian)electrochemicalreactioninwhichtherateofreac-tionisgovernedbythediffusionoftheelectroactivespeciestothesurfaceofaplanarelectrode.14Thedifferenceinpotentialbetweenthepeaksofthereduction(Epc)andoxidation(Epa)curvesis0.068V(avaluethatisclosetothetheoreticalvalueof0.059Vfortheferroceneredoxpair)forallscanratesbetween50and500mVsÀ1,andthepeakcurrentratio(ipa/ipc)isequalto1.0.14Thisreversiblebehaviorindicatesthatnosidereactionstakeplace,andthat,asexpected,thekineticsofelectrontransferissufficientlyrapidtomaintainthesurfaceconcentra-tionsofredox-activespeciesatthevaluesrequiredbytheNernstequation.

LabChip,2010,10,477–483|479

À3ÀÀ

2FeðCNÞ46󰀄󰀄󰀄󰀄󰀄󰀄󰀄󰀄!2FeðCNÞ6þ2e

(2)

À4ÀÀ2FeðCNÞ36þ2e󰀄󰀄󰀄󰀄󰀄󰀄󰀄󰀄󰀄!2FeðCNÞ6

cathodicelectrode

(3)

Inthefirststep,glucoseoxidasecatalyzedtheoxidationof

glucosetogluconicacidwithconcomitantreductionofFe(III)toFe(II)(eqn(1));theFe(CN)64Àionsgeneratedweredetectedchronoamperometrically(eqn(2)).Thecorrespondingcathodicreactionwasdescribedineqn(3).

Metals.Weusedsquare-waveanodicstrippingvoltammetryforthemeasurementsofheavy-metalionsinthemPEDs.Ourmeasurementsoftracemetalsreliedonthesimultaneous(insitu)platingofbismuthandtargetmetalsontoscreen-printedcarbonelectrodes,whichformedalloysfollowedbyanodicstrippingofmetalsfromtheelectrode.

ThisjournalisªTheRoyalSocietyofChemistry2010

View Online

Fig.2bshowsthattheanodicpeakcurrent,ip,islinearlyproportionaltothesquarerootofthescanrate(n1/2)inboththebulksolutionandthemPED.Thevalueofdiffusioncoefficientevaluatedbyanalyzingtheslopeobservedinbulksolutionwas4.3Â10À6cm2sÀ1(seeESI†),whichisfairlyclosetothereportedvalueof5.7Â10À6cm2sÀ1.15Thecurrentreadout(Fig.2b)measuredusingthepaperdeviceisabout30%lowerthanthatmeasuredinbulksolution.Wepresumethatthisdifferenceisduetothefactthat30–40%ofthevolumeindiffusionalcontactwiththeelectrodesisoccupiedbythecellulosefiberofthepaper.Theseresultscontainingtheredox-activespeciesdonotslowtherateofmass-limitedchargetransferrelativetothatinsolution.

Downloaded on 08 September 2011Published on 03 December 2009 on http://pubs.rsc.org | doi:10.1039/B917150AChronoamperometricanalysisofglucoseinurine

Wedemonstratedtheuseofourdeviceintheanalysisofaclini-callyrelevantanalyteinasimulantbodyfluidofglucoseinartificialurineusingchronoamperometry.Chronoamperometryoffersabettersignal-to-noiseratiothanotherelectrochemicaltechniquesinthiskindofexperiment,andtheuseofathinslaboffluidsmechanicallyclampedtotheelectrodesismoreresistanttovibrationthananalysisinalargervolumeofsolution.14Thechronoamperometricmeasurementofcurrent—reflectingchargetransferto/fromtheredox-activespeciesasafunctionoftimeatconstantappliedvoltages—beginswithaninitiallylargecapac-itivecurrent.Uponthedecayoftheinitialcapacitivecurrentwithin1–2s,Faradaiccurrent(thecurrentthatisproportionaltotheconcentrationoftheanalyte)dominates.Thecurrent,i,decaysastÀ1/2asdescribedbytheCottrellequation14(eqn(4))

nFAD1=2Ci¼

p1=2t1=2

(4)

wherenisthenumberofelectrons,tisthetime,FisFaraday’sconstant,Aistheareaoftheelectrode,Disthediffusioncoef-ficientofanalytes,andCistheinitialconcentrationofthereactants.

ThemPEDconfinesfluidsinthepaperchannel,inhibitstheconvectivemovementoffluids,andthusfacilitatesthechro-noamperometricmeasurementsbyminimizingthedisturbancesofthestationaryboundarylayerinthevicinityofelectrodesduetovibration,thermalordensity-basedconvection,andotherdisturbingsources.Fig.3ashowsarepresentativechro-noamperometricresponseofthemeasurementsofglucoseusingamPED.Overtherangeofconcentrationsofglucoseexamined(0–22.2mM),alltheresponsecurvesreachedasteadystate2safterthesteppotential(alsoseetheCottrellplotinESI†).Fig.3bshowsacalibrationcurveforthedetectionofglucose.Whentheconcentrationofglucoseisintherangeof0–22.2mM,thecurrentislinearlyproportionaltotheglucoseconcentrationintheartificialurine.

Wetestedtheinterferenceofthesensingdevicewithbovineserumalbumin(BSA)asatypicalglobularprotein;serumalbuminsarepresentinhighestconcentrationsinserum,andthusrelevanttobioanalysis.Wefoundthatthepresenceof40mMBSAdidnotinterferewiththemeasurementofglucose.Thisselectivityisduetothespecificityofenzymaticoxidationofglucose;theBSAapparentlydoesnotfoultheelectrodes.ComparingtheresultsofthedetectionofglucoseinthemPEDstothoseinbulksolutions(Fig.3b),wenoticedthatthetwomethods

480|LabChip,2010,10,477–483

showedcomparablesensitivityanddetectionlimits;thepapermatrixdoesnotinterferewiththedetection.ThepapermatrixinthemPEDshas,however,severaladvantages:(i)itstabilizesthegeometryoftheelectrode;(ii)itreducestheeffectofconvectionofliquidsduetorandommotion,vibrationandheating;(iii)itminimizesthetotalvolumeofsolutionrequiredforanalysis.Thenormallevelofglucoseinurineis0.1–0.8mMand3.5–5.3mMinwholeblood.3Ourdeviceshouldthereforebecapableofmeasuringglucoseinotherbiologicalfluidssuchasserumandblood(seemeasurementofglucoseinbloodplasmaandwholebloodinESI†).ThedetectionlimitofglucoseinthecurrentmPEDisabout0.22mM(correspondingto4mgmLÀ1).Thisvalueisbelowtheapproximately1.0mMclaimedinspecifica-tionsofconventionalglucometers,and0.5mMobtainedbycolorimetricdetectionmethodreportedpreviously.2,3Weesti-matethesensitivityoftheglucoseanalysistobe0.43mAmMÀ1mmÀ2.Inprinciple,otherspeciesinrealurineandbloodmay

ThisjournalisªTheRoyalSocietyofChemistry2010

View Online

interferewhenweuse500mVpotentialsfortheanalysisofglucose.Thepotentialcanbereducedtoaround300mVforthemeasurementsinbiologicalfluids,thankstotheenzymaticselectivityofglucoseoxidase.Thepaper-baseddevicealsohasthepotentialtobeintegratedwithvariousseparationtechniquessuchaspaperchromatographytominimizeinterferences.Anodicstrippingvoltammetricanalysisofheavy-metalionsHeavy-metalionssuchasmercury,lead,andcadmiumaretoxic,non-biodegradable,andtendtoaccumulateinplantsandanimals.16,17Thepollutionofheavy-metalionsinsoilandwaterpresentsaglobalissue,andposesaseverethreattoboththeecosystemandhumans.Square-waveanodicstrippingvoltam-metry(SWASV)isanASVmethodfrequentlyusedforthemeasurementoftraceheavymetalsbecauseitgreatlyreducesthebackgroundnoisecomingfromthechargingcurrentduringthepotentialscan.18,19ConventionalASVmeasurementsofheavy-metalionsareusuallyperformedeitherbydippingelec-trodesinasamplesolutionundercontrolledstirringconditionorbyplacingasampledropletontotheelectrodes.20–23Theformerapproachisnotpracticalinfieldmeasurementsduetothediffi-cultyofsynchronizingthestirringandASVprocedures.20Thelatteroneshowslimitedsensitivityofmeasurementbecausepre-accumulationofanalytesinstrippingislimitedbydiffusion.Additionally,inthiscase,newelectrodesareusuallyrequiredforeachmeasurementsinceitisdifficulttoremovetheresidueofdepositedmetalsinastagnantdropofsolutionbeforethenextcycleofASV.

WehavedemonstratedtheuseofmPEDsintheselectivemeasurementofPb(II)inanaqueousmixtureofPb(II)andZn(II)usingSWASV.WemodifiedthedesignofthemPEDbyintro-ducingapadofcelluloseblottingpaperasasinkintheoutletofthepaperchannel(Fig.1c)(seeESI†).ThemPEDallowsthecontinuouswickingoffluidstopassacrosstheelectrodes,andfacilitatestheplatingofmetals,aswellasthecleaningofelec-trodes.Bytuningthesizeofthecelluloseblottingpaper,weoptimizedthewickingtimeoffluidsinthepaperchannelofmPEDssothattheflowstoppedbeforethesystementeredtheequilibrationstepintheprocessofSWASV.24Fig.4displaysrepresentativestrippingvoltammogramsforthemeasurementof25ppb(mgLÀ1)Pb(II)inacetatebuffersolutioninthepresenceofZn(II).Thevoltammogramsinthehydrody-namicmPEDs,inwhichthefluidofthesamplesolutioncontin-uouslywickedinthepapermicrochannel,showedawell-defined,sharppeakforPb(II)atca.780mVvs.theAg/AgClreferenceelectrode(CandDinFig.4).Incontrast,underthesameSWASVconditions,astagnantsolutionofanalytes,bothinthemPEDs(withoutapadofblottingpaperasasink)andinanexperimentthatplacedadropletofsamplesolutionontheelectrodes,resultedinamuchweakersignal(AinFig.4)orapoorlydefinedresponse(BinFig.4).ThehydrodynamicmPEDsthusexhibitedamuchhighersensitivitybyafactoroffivethanthestagnantsystems.Weascribetheenhancedsensitivityofthedynamicsystemtothehighefficiencyoftheaccumulationofmetalsontheelectrodesbyconvectionofflowingfluidsintheporousmatrixofpaperovertheelectrodes,andtothelargevolume($800mL)ofsamplethatflowsacrossthesurfaceoftheelectrodes.

ThisjournalisªTheRoyalSocietyofChemistry2010

Downloaded on 08 September 2011Published on 03 December 2009 on http://pubs.rsc.org | doi:10.1039/B917150AInthehydrodynamicmPEDs,thepeakcurrentoftheanalysisofPb(II)dramaticallyincreasedwithincreasingthedepositiontime(Fig.4).Thepeakcurrentincreasedfrom3.9mAto10.3mAwiththeincreaseofdepositiontimefrom120sto360s.Thisincreasewasnotobviousinthestagnantsystems,sincethedepositionefficiencydecaysquicklywithtimeduetothemass-transfer-limitedreactioninthevicinityofthesurfaceofelec-trodes.Moreover,wefoundthatthestagnantmPEDsshowedamorepoorlydefinedsignal,comparedtothesystemwithadropofsamplesolutiondirectlyplacedontotheelectrodes(Fig.4).WepresumethatthisisbecausethecellulosematrixinthestagnantmPEDsinhibitsconvectivemovementofthesolutions,andthusaffectedthestrippingbehaviorofPb(II).

ThestrippingvoltammogramsfortheanalysisofPb(II)inthehydrodynamicmPEDsshowedwell-definedpeaksandastrongsignaloverawiderangeofconcentrationsofPb(II);thislevelofperformanceoffersconvenientquantificationoflowppblevelsoflead(Fig.5a).ThepeakintensityincreasesproportionallywiththeconcentrationofPb(II),whichyieldsahighlylinearcalibrationplotwithaslopeof0.17mAppbÀ1forlead(corre-lationcoefficient,0.996)(Fig.5b).Thelimitofdetectionofleadwasestimatedfromthesignal-to-noisecharacteristicsofthedatatobeapproximately1.0ppb(mgLÀ1);thisvalueisevenlowerthan2.5ppbobtainedinconventionalsystemswith

LabChip,2010,10,477–483|481

View Online

Downloaded on 08 September 2011Published on 03 December 2009 on http://pubs.rsc.org | doi:10.1039/B917150AWecomparedtheperformanceofthehydrodynamicmPEDsfortheanalysisofleadwiththestagnantsystem,inwhichadropofsamplesolutionwasplacedonelectrodes(Fig.5b).Thestagnantsystemexhibitsamuchlowersensitivityof0.05mAppbÀ1forlead(correlationcoefficient,0.978),andahigherlimitofdetectionof4.3ppb,thanthedynamicmeasurement.Unlikethestagnantsystem,smallperturbations(e.g.,vibration,heating)didnotinterferewiththeanalysisofleadinthehydrodynamicmPEDsduetothestabilizationoftheflowofthesamplesolutionbythepapermatrix;thisstabilizationresultsinahighreliabilityandreproducibilityofthemeasurements.Thedeviceisreusablebysimplyreplacingthepadofblottingpaper,sincethecontin-uouswickingremovesdissolvedanalytesbeforethenextcycleofthedepositionofmetals.

Conclusions

Paperprovidesausefulmatrixforsmallelectrochemicaldevices.Itprovidesathin,mechanicallystabilizedfilmofwaterorotherfluidsthatdeliveranalytestothesurfaceoftheelectrodes.Thefactthatthecellulosematrixismechanicallyrigid,andthattheelectrodesareinphysicalcontactwiththepaperslabcontainingtheanalytesinhibitsconvectivemovementofthesolutionsafterthecompletewettingofthepaperchannelwithfluids,andthusincreasestheaccuracyoftime-dependentmeasurements(chro-noamperometryandchronopotentiometry).Thecontinuouscapillarywickingoffluidsalongthepaperchannelandacrosstheelectrodethatiseasilyaccomplishedbyintroducingacelluloseblottingpaperthatactsasasinkprovidesanexcellentplatformforthemeasurementsrequiredforconvectionandhydrodynamicconditions(anodicstrippingvoltammetry,hydrodynamicamperometryandvoltammetry).

Electrochemistrysubstantiallyincreasesthescopeofoptionsfordetectionusingpaperandfiber-basedsystems.1,27,28Wehavedemonstratedthechronoamperometricanalysisofglucose,andthesquare-waveanodicstrippingvoltammetrymeasurementsofheavy-metalionsusingmPEDs.Theresultinganalysisofglucoseissimilartothatinbulksolution(thecellulosematrixoccupies30–40%ofthevolumeofthespacebetweentheelectrodes,andlimitingdifferentialcurrentsareaccordingly$30%loweronpaperthanunstirredbulksolution).Themeasurementsofheavy-metalionsinthehydrodynamicmPEDshavehighersensitivityandlowerlimitofdetection(ca.1.0ppb)thanconventionalsystemswithcontrolledstirring(ca.2.5ppb).21,29ThemPEDsthushavethreeimportantadvantagesoverconventionalsystemswithcontrolledstirring:(i)theyaresimple,low-cost,andportable,(ii)theydonotrequireextrainstruments,suchasstirringplates,and(iii)theyarereusable,andareeasilydispos-ableafteruse.

Theuseofconductiveinkstoprintelectrodesiswellsuitedformassproductionusingscreen-printingtechnology.Thismeth-odologyissuitableforfurtherminiaturization,andtoaccom-modatesmallerlateralspacingbetweenelectrodes,andthus(withdevelopment)tohigherdensityofdetectors.Theresultingpaper-basedelectrochemicalsensingdevicescombinetheadvantagesofpaper-basedmicrofluidics—low-cost,smallsamplevolumes,capillarywickingoffluids,facilemultiplexedassays—withtheadvantagesofelectrochemicaldetection—highsensitivity,quantitativeanalysis,rapidresponseandapplicability

ThisjournalisªTheRoyalSocietyofChemistry2010

controlledstirring.20–22,25,26Thisvalueisalsomuchlowerthanthe10ppb(mgLÀ1)WHOguidelinevalueforleadconcentra-tionindrinkingwater.4Webelievethatevenlowerconcentra-tionsofleadcouldbedetectediflongerdepositionperiodswereused.ThesensitivemeasurementofPb(II)inthehydrodynamicmPEDsishighlyreproducible,asindicatedbythelowrelativestandarddeviation.

482|LabChip,2010,10,477–483

View Online

toawiderangeofanalytes.Fortheuseinthefield,itwouldbenecessarytodevelopaportableelectrochemicalreader.ThemPEDsdescribedinthepapercouldeasilybeadaptedtoaportablereader.Thedevelopmentofthiskindofdevicesisunderwayinthislab.

Acknowledgements

ThisworkwassupportedbytheNano/Micro-electromechanicalSystemsScienceandTechnologyMicro/NanoFluidicsFunda-mentalsFocusCenter(DefenseAdvancedResearchProjectsAgency),theNationalInstitutesofEnvironmentalHealthandSafety(016665),theBillandMelindaGatesFoundation(51308),andapostdoctoralfellowshipfromtheNaturalScienceandEngi-neeringResearchCouncilofCanada(toZ.N.).TheNetherlandsOrganizationforScientificResearch(NWO)iskindlyacknowl-edgedfortheRubicongrant(C.A.N.)supportingthisresearch.

Downloaded on 08 September 2011Published on 03 December 2009 on http://pubs.rsc.org | doi:10.1039/B917150AReferences

1W.Dungchai,O.ChailapakulandC.S.Henry,Anal.Chem.,2009,81,5821–5826.

2A.W.Martinez,S.T.Phillips,E.Carrilho,S.W.Thomas,H.SindiandG.M.Whitesides,Anal.Chem.,2008,80,3699–3707.

3J.Hones,P.MullerandN.Surridge,DiabetesTechnol.Ther.,2008,10,S10–S26.

4WorldHealthOrganization.GuidelinesforDrinking-WaterQuality,WHO,WHOPress,Switzerland,2006.

5A.W.Martinez,S.T.Phillips,M.J.ButteandG.M.Whitesides,Angew.Chem.,Int.Ed.,2007,46,1318–1320.

6A.W.Martinez,S.T.PhillipsandG.M.Whitesides,Proc.Natl.Acad.Sci.U.S.A.,2008,105,19606–19611.

7A.W.Martinez,S.T.Phillips,B.J.Wiley,M.GuptaandG.M.Whitesides,LabChip,2008,8,2146–2150.

8D.Mabey,R.W.Peeling,A.UstianowskiandM.D.Perkins,Nat.Rev.Microbiol.,2004,2,231–240.

9P.Yager,G.J.DomingoandJ.Gerdes,Annu.Rev.Biomed.Eng.,2008,10,107–144.

10Y.Wang,H.Xu,J.M.ZhangandG.Li,Sensors,2008,8,2043–2081.

11Othertypesofelectrodes,forexample,goldstripescoatedonplasticsarealsosuitableforthepaper-basedelectrochemicaldevice,andmaybethebestusedformorespecificapplications.Paperdeviceswithelectrodesmadefromgoldstripesshowedexcellentperformanceincyclicvoltammetry,butwerelessrobustthanthecarbon-basedelectrodes.Thisproblemis,however,simplyduetotheadhesion,whichcouldbefixedbyimprovedengineering.

12O.D.Renedo,M.A.Alonso-LomilloandM.J.A.Martinez,Talanta,2007,73,202–219.

13E.Carrilho,A.W.MartinezandG.M.Whitesides,Anal.Chem.,2009,81,7091–7095.

14A.J.BardandL.R.Faulkner,ElectrochemicalMethods:FundamentalsandApplications,Wiley,NewYork,2000.

15P.N.BartlettandK.F.E.Pratt,J.Electroanal.Chem.,1995,397,53–60.

16M.Patra,N.Bhowmik,B.BandopadhyayandA.Sharma,Environ.Exp.Bot.,2004,52,199–223.

17S.K.Porter,K.G.Scheckel,C.A.ImpellitteriandJ.A.Ryan,Crit.Rev.Environ.Sci.Technol.,2004,34,495–604.

18B.K.JenaandC.R.Raj,Anal.Chem.,2008,80,4836–4844.

19W.Yantasee,Y.Lin,K.Hongsirikarn,G.E.Fryxell,R.AddlemanandC.Timchalk,Environ.HealthPerspect.,2007,115,1683–1690.

20J.Wang,J.M.Lu,U.A.Kirgoz,S.B.HocevarandB.Ogorevc,Anal.Chim.Acta,2001,434,29–34.

21M.A.G.Rico,M.Olivares-MarinandE.P.Gil,Electroanalysis,2008,20,2608–2613.

22K.C.HoneychurchandJ.P.Hart,TrAC,TrendsAnal.Chem.,2003,22,456–469.

23R.O.KadaraandI.E.Tothill,Anal.Bioanal.Chem.,2004,378,770–775.

24Weareawarethatthisdeviceissensitivetotemperatureduetotemperature-dependentrateofwickingofflow.Thisproblem,however,canbefixedbyimprovedengineering.

25J.Wang,J.M.Lu,S.B.Hocevar,P.A.M.FariasandB.Ogorevc,Anal.Chem.,2000,72,3218–3222.

26G.H.Hwang,W.K.Han,J.S.ParkandS.G.Kang,Sens.Actuators,B,2008,135,309–316.

27G.Cui,S.J.Kim,S.H.Choi,H.Nam,G.S.ChaandK.J.Paeng,Anal.Chem.,2000,72,1925–1929.

28G.Cui,J.H.Yoo,J.Yoo,S.W.Lee,H.NamandG.S.Cha,Electroanalysis,2001,13,224–228.

29K.C.Honeychurch,J.P.HartandD.C.Cowell,Electroanalysis,2000,12,171–177.

ThisjournalisªTheRoyalSocietyofChemistry2010LabChip,2010,10,477–483|483