ContentslistsavailableatScienceDirect
ThinSolidFilms
journalhomepage:www.elsevier.com/locate/tsf
AnalysisofperformanceanddeviceparametersofCIGSPVmodulesdeployedoutdoors
C.Radue⁎,E.E.vanDyk,E.Q.Macabebe
PhysicsDepartment,P.O.Box77000,NelsonMandelaMetropolitanUniversity,PortElizabeth,6031,SouthAfrica
articleinfoabstract
Two20Wcopperindiumgalliumdiselenidephotovoltaicmodulesweresubjectedtoathoroughindoorassessmentprocedure,followedbyoutdoordeploymentattheNelsonMandelaMetropolitanUniversityaspartofanongoingstudy.Theinitialindoormeasurementofmaximumpoweroutput(PMAX)ofoneofthemoduleswasconsiderablyhigherthanthemanufacturer'srating(E.E.vanDyk,C.RadueandA.R.Gxasheka,ThinSolidFilms515(2007)6196).Themodulesweredeployedonadual-axissolartrackerandcurrent–voltagecharacteristicswereobtainedweekly.InadditiontothenormalPVparametersofshort-circuitcurrent,open-circuitvoltage,PMAX,fillfactorandefficiency,shuntandseriesresistanceswerealsomonitored.Theperformancesofthetwomodulesarecomparedandanalyzedandtheresultspresentedinthispaper.
©2008ElsevierB.V.Allrightsreserved.
Availableonline8November2008Keywords:CIGS
PhotovoltaicmodulesI–V
Parasiticresistances
1.Introduction
Copperindiumgalliumdiselenide(CIGS)isapromisingchalcogenidematerialusedforthinfilmphotovoltaic(PV)applications.Efficienciesofmassproducedmodulesaretypically11–12%witharecordof13%for85WmodulesatWürthSolar[2].Improvementsinconfigurationandmanufacturingtechnologieswillresultinhigherefficiencies.SomestudieshaveshownthatthesePVcellsandmodulesarestable[3,4]whileothershaveshownthattheydegrade[5–7].InorderforCIGSPVmodulestocompetewithcrystallinesiliconmodules,thestabilityissueswillneedtobedealtwith.Servicelifetimesforcommercialmodulesof20yearsandmorearedesirable,buthavenotyetbeendemonstrated,asthefirstCIGScommercialmoduleswereonlyavailablein1998([8]p.369).TheperformanceoftwoCIGSPVmodulesdeployedoutdoorsisbeinginvestigated,andtheresultsofthepreliminaryassessmentsarepresented.ThispaperispartofanongoingstudyintotheperformanceandservicelifetimesofCIGSPVmodules.2.Experimentaldetails
TwonominallyidenticalCIGSPVmoduleswereboughtfromthesamemanufacturerandhadthesameoutputpowerratingsof20W.TheywerelabelledCIGS-1andCIGS-2,andweresubjectedtoadetailedindoorassessmentcarriedoutatSTC(StandardTestConditions:25°C,1000W/m2,AM1.5).Efficienciesof11.2%and9.4%andmaximumoutputpowersof24.13Wand20.29Wrespectively,weremeasured[1].The
reasonsforthedifferencesintheperformancesofthesemoduleshavebeendiscussedpreviously[1].Themodules,eachconsistingof42monolithicallyintegratedseries-connectedcells,weredeployedout-doorsattheOutdoorResearchFacilityoftheNelsonMandelaMetropolitanUniversity,situatedinPortElizabeth,SouthAfrica,atalatitudeof34°S.Themodulesweremountedonadual-axissolartrackersoastoalwaysbeperpendiculartothesuninordertomaximizeexposuretoirradianceandminimizetheeffectsofdifferencesinincidentangle.Duringtheperiodunderconsideration,themodulesreceived,onaverage,27%moreirradiancethantheywouldhaveonafixedrackatanangleof34°.
Astandardizedmodulemeasurementsystemwasusedtocharacter-izetheperformanceparametersofmodulesdeployedoutdoors.Directirradiance,planeofarrayirradiance,backofmoduletemperatureandambienttemperaturewererecordedinadditiontothecurrent–voltage(I–V)curves,inordertofacilitatefurtheranalysisandcomparisonwithindoormeasurements.TheI–Vcurveswereusedtodetermineshort-circuitcurrent(ISC),open-circuitvoltage(VOC),efficiency(η),fillfactor(FF)andmaximumpoweroutput(PMAX).Shunt(RSH)andseries(RS)resistance,idealityfactor(n)andsaturationcurrent(I0)werealsodeterminedbyfittingthetwo-diodemodeltotheI–Vcurves.ThefittingroutineemployedaPSO-based(particleswarmoptimization)algorithm,thedetailsofwhichareoutsidethescopeofthispaper,andwillbereportedelsewhere.Thetwo-diodemodelisgivenby[9,10]:V+IR
SqðV+IRSÞ=n1kTqðV+IRSÞ=n2kT
−1−I02e−1−I=IL−I01e
RSH
ð1Þ
⁎Correspondingauthor.Tel.:+27415042269;fax:+27415042573.E-mailaddress:chantelle.radue@nmmu.ac.za(C.Radue).0040-6090/$–seefrontmatter©2008ElsevierB.V.Allrightsreserved.doi:10.1016/j.tsf.2008.11.011
whereIListhelightgeneratedcurrent,I01isthereverseordarksaturationcurrentcorrespondingtodiffusionandrecombinationof
2384C.Radueetal./ThinSolidFilms517(2009)2383–2385
electronsandholesinthequasi-neutral(bulk)regions,I02isthereverse(dark)saturationcurrentcorrespondingtorecombinationinspacecharge(depletion)region,n1andn2arethediodeidealityfactors,qistheelectroncharge,kisBoltzmann'sconstantandTisthetemperatureinKelvin.Theidealityfactorstypicallyhavevaluesof1foradiodedominatedbyrecombinationinthequasi-neutralregions,andvaluesapproaching2fordiodesdominatedbyrecombinationinthedepletionregion.
Moduleperformancewasmonitoredweeklyatsolarnoon.Themeasurementswerenormalizedwithrespecttotemperature(25°C)andirradiance(1000W/m2),andtheresultsarepresentedbelow.3.Resultsanddiscussion
Fig.1showstheperformancesofthetwomodulesindoorsatSTC(symbollines),atdeploymentoutdoors(solidlines),andafter4monthsofexposureoutdoors(brokenlines).ItiswellknownthatoutdoorperformanceseldommatchestheperformanceobtainedindoorsatSTC.ThisisclearlyseeninFig.1andthePMAXvaluesofCIGS-1andCIGS-2measuredoutdoorsatdeploymentare11%and4%lowerrespectivelythantheinitialSTCmeasurements.
Fig.1.I–VcharacteristicsobtainedindoorsatSTC(symbollines),outdoorsatdeployment(solidlines),andafter4monthsofexposureoutdoors(brokenlines)for(a)CIGS-1,and(b)CIGS-2.
Fig.2.PMAXasafunctionoftimeoutdoorsforthetwoPVmodules.PMAXfortheinitial14daysisshowninregionA,indicatedbythesolidlines,andmonthlyaveragesofPMAXforthe4monthsareshowninregionB,indicatedbythebrokenlines.
TheinitialoutdoorI–VcharacteristicsinFig.1alsorevealthatCIGS-1hasagreateroperatingcurrentthanCIGS-2,resultinginahigherPMAX.ThefinaloutdoorI–Vcharacteristicsshowthatinbothcases,theperformanceofthemodulesseemstohaveimproved.TheI–VcharacteristicsshowthatCIGS-1isstilloperatingbelowtheinitialindoorSTCperformance,whileCIGS-2isperformingbetterthaninitiallyatSTC.TheFFforCIGS-2hasimprovedfrom0.57[1]to0.60,theVOCisunchanged,butthephotogeneratedcurrenthasnotyetreachedtheSTCvalue.Theperformanceofeachmoduleisdiscussedinthesucceedingparagraphs.3.1.Cigs-1
Fig.2illustratesthemaximumpowerforthefirst14daysofexposure(regionA—solidlines)andthenthemonthlyaveragesforthe4monthsofoutdoorexposure(regionB—brokenlines).BothVOCandphotogeneratedcurrentshowedaninitialincreaseinthefirst2weeksofoutdoorexposure,leadingtoanincreaseinPMAX,asillustratedinFig.2regionA.Thisinitialimprovementinperformancehasbeenshownbyotherauthors,whoascribeittolight-inducedpowerenhancement[2].Aftertheinitialimprovement,VOCcontinuestoincrease,butatamuchslowerrate.However,thephotogeneratedcurrentdecreases,resultinginadecreasingPMAXwithtime(Fig.2regionB).ThisincreaseinVOCiscommonlyobserved([8]p.393)andisthoughttobeduetopersistentphotoconductivityintheCIGSabsorbermaterial,whichaffectsVOCexclusively([8]p.394).Someauthorsreportthataftertheinitialimprovement,theperformanceofCIGSisstable[2,4],whileothershaveobserveddegradation,althoughthisistypicallyforcopperindiumdiselenide(CIS)duringdampheattesting(e.g.[5,6]).
Thephotogeneratedcurrentismainlyaffectedbytheavailabilityofphotonsandtheabsorptioncoefficientofthematerial[11].Asthephotogeneratedcurrentdecreases,thefillfactor(FF)decreases.AdecreaseinFFthussuggeststhatthematerialhasdegraded.Increasedparasiticresistancescausingpoorercollectionofphotons,orchangesintheabsorptioncoefficient,couldresultinsuchdegradation.Noobviousvisualdegradationofthematerialhasbeenobserved,butthecorrosionofcontactsmayhaveincreasedRS,causingthephoto-generatedcurrenttodecrease.Table1liststhedeviceparametersatthebeginningofthestudyandafter4monthsoutdoors.ItcanbeseenthatRShasincreasedovertimefrom0.035Ωto0.052Ω.Thiscouldbeduetocorrosionofcontactsanddailythermalcycling,resultinginadecreaseinFF.FFisalsonegativelyaffectedbycellmismatch,andlaserbeaminducedcurrent(LBIC)studiesonthesemoduleshaveshownthatthecurrentmismatchbetweenthedifferentcellsoftheentire
C.Radueetal./ThinSolidFilms517(2009)2383–2385
2385
Table1
Deviceparametersasamonthlyaverageforthefirstmonthofoutdoorexposure,andafter4monthsoutdoors,forCIGS-1andCIGS-2
CIGS-1CIGS-2Month1Month4Month1Month4Average
AverageAverageAverageRS(Ω)0.0350.0520.0320.041RSH(Ω)199.9
29.0
20.74
20.38
I01(A)1.92×10−44.96×10−53.59×10−41.22×10−5n12.44
4.08
2.56
2.01
I02(A)5.01×10−51.41×10−51.49×10−42.56×10−5n22.50
1.95
2.45
2.74
moduleishigh[1].Celldegradationisthusthemostlikelydegra-dationmodetoexplainthedecreasingFFinCIGS-1.ThedecreaseinRSHbynearlyafactorof10overtheperiodofthestudycanalsobeseeninTable1.Thishasledtoadecreaseintheamountofcurrentreachingtheintendedload,possiblyduetoanincreaseinthenumberofmetastabledefectsinthematerial.Theseoccurduetothepersistentcaptureofphotogeneratedelectronsintraps([8]p.394)thatleadcurrentawayfromtheintendedload.ThislargedecreaseinRSHalsohastheeffectofdecreasingFF.
Theidealityfactors(n1andn2)arealsoshowninTable1.Idealityisakeyindicatorofexcessforwardcurrentattributedtoe–hrecombinationthroughtrapsinthejunctionregion[12].Inaddition,VOCiscontrolledbythesingledominantmechanismwiththehighestcurrent[13].Theidealityfactorcandistinguishbetweenrecombinationinthequasi-neutralregionorattheCIGS/CdSinterface(n=1),thedepletionregion(1bnb2),orviatunnelling[14]attheCIGS/CdSinterface(nN2).IthasbeenshownthatinCIGSPVcellsthediodecurrentiscontrolledbyShockley–Read–Hallrecombination[13],wherenisbetween1and2.Therecombinationislargestthroughdeeptrapstatesinthedepletionregion.Thevariationofnbetween1and2dependsonthetrapenergiesofthedeepleveldefects.Asnapproaches1,bandtobandbulkrecombinationstartstodominate.
SinceI01hasinitiallythehighestcurrentof1.92×10−4A,andn1=2.44,itmaybeconcludedthatthedominantcurrent-reducingmechanismcontrollingVOCisrecombinationattheCIGS/CdSinterface.BothI01andI02havedecreasedovertheperiodunderconsideration,i.e.thediodequalityhasimproved.Afterthe4months,I01isslightlylargerthanI02andrecombinationatthisdiodestilltendstodominatetheI–Vcharacteristics.Sincen1hasincreasedton1=4.08overthefourmonthperiod,therehasbeenincreasedtunnellingenhancedrecombinationattheinterface.Theidealityfactorfortheseconddiode,n2,hasdecreasedfromn2N2ton2b2,andwethusconcludethattunnellingenhancedrecombinationattheinterfacedominatedtheseconddiodeinitially,andthenShockley–Read–Hallrecombinationintheabsorberlayerdominated.3.2.Cigs-2
ThebehaviourofVOCandISCarethesameasthoseforCIGS-1:bothincrease,accompaniedbyaninitialincreaseinPMAX,shownbythesolidlineinFig.2(regionA).However,theaveragemonthlyPMAXoverthe4monthsofoutdoorexposureisconstantwithinexperimentalerror,indicatedbythelinearfitinregionB(brokenline)ofFig.2.IthasbeenshownthatwhilePMAXmaybestable,changesmayoccurinthedevicecharacteristics[4],asobservedforCIGS-2.TheFFforCIGS-2is,unlikethatofCIGS-1,increasing.Thisparameter(FF)dependsontemperatureandidealityfactor([8]p.393).
I01andI02initiallyhaveverysimilarvalues.However,inthelowvoltageregionoftheI–Vcurve,I02isnegligiblecomparedtoISCandthethirdterminEq.(1)canthusbeignored.I01thusdominatesinthelowvoltageregions.AscanbeseeninTable1,n1hasdecreasedovertimetonearly2,meaninginterfacerecombinationisdecreasingandShockley–Read–Hallrecombinationintheabsorberlayerisbeginningtodominate
theI–Vcharacteristicsinthelowvoltageregion.Thedecreaseinn1alsoresultsinanincreasedFF,despitetheincreaseinRSwhichshouldreduceFF.AnotherfactoraffectingFFisthetypeIIImetastabilityfoundinCIGSmaterial([8]p.394).ThisisanincreaseinFFduetoilluminationwithlightthatisabsorbedinthebufferlayerorextremesurfaceregionoftheCIGSabsorber,i.e.thebluepartofthespectrum.
InthehighervoltageregionoftheI–Vcurve,I02cannotbeneglected.I01andI02areofthesameorderofmagnitudeandbothhaveidealityfactorsgreaterthan2.WeconcludethattheinitialdominantrecombinationmechanismforCIGS-2isthustunnellingenhancedrecombinationattheinterface.Afterthe4months,n2hasincreased,resultinginincreasedinterfacerecombinationinthehighcurrentregionsoftheI–VcurvewhereI02cannotbeneglected.I02isnowgreaterthanI01anditisthusinterfacerecombinationwhichdominatestheI–Vcharacteristics.BothI01andI02havedecreasedovertime,resultinginimproveddevicequality.4.Summaryandconclusion
BothCIGS-1and-2initiallyshowedanincreaseinPMAXinthefirst2weeksofoutdoorexposure.AverylargedecreasewasobservedfortheRSHofCIGS-1.ThislargedecreaseinRSHdominatedthedevicecharacteristicsofthemodule,leadingtoanoveralldecreaseinPMAX.RSHwasfairlyconstantforCIGS-2,andtheimprovingdevicecharacteristicsandtypeIIImetastabilityresultedinthestableperformanceofCIGS-2.ThedominantrecombinationprocessinCIGS-1wastunnellingenhancedrecombinationattheCIGS/CdSinterface,andinCIGS-2thisrecombinationalsodominates,althoughShockley–Read–Hallrecombinationintheabsorberlayerisbeginningtoemerge.TheinitialindoorassessmentofthesemodulesindicatedthatCIGS-1wouldbethebettermodule.However,outdoortestinghasshownthatCIGS-2isstable,whileCIGS-1isdegrading.CIGS-2hasadifferentsurfaceappearancetoCIGS-1andhasmanylighterdiscolouredareas[1].Asaresult,CIGS-2doesnotheatupasmuchasCIGS-1duringoutdoordeployment,resultinginabetteroutdoorperformance,especiallyinsummer.
Acknowledgements
TheauthorswouldliketothankTelkomundertheCentreofExcellenceProgramme,TotalFacilitiesManagementCompany(TFMC),theTHRIPprogrammeoftheSouthAfricanDepartmentofTradeandIndustry,theErnstandEthelEriksenTrust,theSchlumbergerFounda-tionFacultyfortheFutureProgram,theNelsonMandelaMetropolitanUniversityandtheSouthAfricanNationalEnergyResearchInstitute(SANERI)fortheirfinancialsupport.
References
[1]E.E.vanDyk,C.Radue,A.R.Gxasheka,ThinSolidFilms515(2007)6196.[2]B.Dimmler,R.Wächter,ThinSolidFilms515(2007)5973.[3]F.H.Karg,Sol.EnergyMater.Sol.Cells66(2001)645.
[4]T.Kojima,T.Koyanagi,K.Nakamura,T.Yanagisawa,K.Takahisa,M.Nishitani,T.Wada,Sol.EnergyMater.Sol.Cells50(1998)87.
[5]W.Durisch,K.H.Lam,J.Close,Appl.Energy83(2006)1339.[6]E.L.Meyer,E.E.vanDyk,Renew.Energy28(2003)1455.
[7]J.Wennerberg,J.Kessler,L.Stolt,Sol.EnergyMater.Sol.Cells75(2003)47.
[8]U.Rau,H.W.Schock,in:T.Markvart,L.Castañer(Eds.),PracticalHandbookofPhotovoltaics:FundamentalsandApplications,Elsevier,Oxford,2006,p.369.[9]J.L.Gray,in:S.Luque,S.Hegedus(Eds.),HandbookofPhotovoltaicScienceandEngineering,JohnWileyandSons,Chichester,2003,pp.91.[10]E.E.vanDyk,E.L.Meyer,Renew.Energy29(2004)333.
[11]E.L.Meyer,E.E.vanDyk,17thEuropeanPhotovoltaicSolarEnergyConference,Munich,Germany,22–26October,2001,ConferenceProceedings,2001,p.524.[12]J.R.Sites,Sol.EnergyMater.Sol.Cells75(2003)243.
[13]W.N.Shafarman,L.Stolt,in:S.Luque,S.Hegedus(Eds.),HandbookofPhotovoltaicScienceandEngineering,JohnWileyandSons,Chichester,2003,pp.597.[14]
U.Rau,Appl.Phys.Lett.74(1)(1999)111.
因篇幅问题不能全部显示，请点此查看更多更全内容