$1 per w photo voltaic systems

8/7/2019 $1 Per W Photo Voltaic Systems

1/28

U.S.DepartmentofEnergy

AdvancedResearchProjectsAgencyEnergy EnergyEfficiencyandRenewableEnergy

$1/WattWhitePaper 1|P a g e

$1/WPhotovoltaicSystemsWhitePapertoExplore

AGrandChallengeforElectricityfromSolar

DISCLAIMER: Thepurpose of thispaper is tofacilitate discussion amongparticipants in the

$1/WSystems:AGrandChallengeforElectricityfromSolarWorkshop,tobeheldonAugust

1011,2010inWashington,DC. Thispaperdoesnotrepresent,reflect,orendorseanexisting,

planned, orproposedpolicy of the U.S. Government, including but not limited to the U.S.

Department of Energy. The U.S. Department of Energy does not guarantee the accuracy,

relevance,timeliness,orcompletenessofinformationherein,anddoesnotendorseanysources

usedtoobtainthisinformation. Assuch,thispaperisnotsubjecttotheInformationQualityAct

andimplementingregulationsandguidelines.


2/28




I. IntroductionAkeyplankoftheObamaAdministrationsEnergyPolicyistoputthecountryonapathtoreduceGreen

HouseGas(GHG)Emissionsby80%by2050. Solarenergytechnologyhasthepotentialtoplayamajor

roleinachievingthisgoalbuttodatehasbeenlimitedbyhighcosts.

TheU.S.DepartmentofEnergy(DOE)estimatesthata$1/wattinstalledphotovoltaicsolarenergy

systemequivalentto56cents/kWhwouldmakesolarwithoutadditionalsubsidiescompetitivewith

thewholesalerateofelectricity,nearlyeverywhereintheUS. Asolarenergysystempricedat$1/Watt

wouldunlockthepotentialofthesuntoprovidelowcost,cleanlimitlesselectricitytotheU.S.andthe

restoftheworld,atthesamecostofcoalbasedgeneration. Atthisprice,solargeneratedelectricity

combinedwithaffordablestoragetechnologiescouldthenmeetallconventionalelectricityenergy

needs,providingsolarenergypotentially24hoursaday. Meetingthischallengewouldresultina

revolutionintheworldsgenerationanduseofenergy.

Withthecurrentrateofprogress,thecostofautilitysizedphotovoltaic(PV)systemislikelytoreach

$2.20/wattby2016,and$2.50/wattand$3.50/watt,forcommercialscaleandresidentialscalesystems

respectively. Reductionssignificantlybeyondthatinthenextfourtoeightyearsareunlikelyabsent

dramaticallynewideasandsignificantinvestment.

PreliminaryDOEanalysisonrequiredcomponentcoststoreacha$1/wattinstalledPVsystemimplies

thefollowingbreakdown: 50cents/wattforthemodule,40cents/wattforthebalanceofsystemand

installation,and10cents/wattforthepowerelectronics. Privateinvestmentisunlikelytobringabout

thetypesofambitiousadvancesrequiredtomeetthesegoalsasthecapitalmarkethasfocusedfunding

onshortertermcommercializationgoalsandinternationalmarketswithattractivegovernmentsupport

andhighelectricityrates.

ThiswhitepaperwasdevelopedbyDOEstafftostimulateadialogueabouttechnologypathwaysto

achieve$1/watt.Thepapercontainsinitialideasonhowtoachievesignificantreductioninthecostof

modules,powerelectronicsandbalanceofsystem/installation. Forexample,analysissuggeststhat

moduleswith25%efficiency,powerelectronicswithdoubletheratedlifetime,andsimplerandquicker

PVinstallationmethodsareallrequired.

II. TheTechnicalChallengeInstalledPVarraypricesforutilityscalesystemswere$8/wattin2004andbidsbelow$3.50/wattare

expectedbytheendof2010ifnotsooner. Residentialandcommercialpricesareover$6/wattsince

theyaremuchsmallerinscaleandincurmuchlargerinstallationpricesandretailmarkups. With

currentmarkettrendsandcostreductionopportunities,utilityscalesystemcostsareexpectedtoreach

$2.20/wattby2016ifnonewprogramislaunched. The$1/wattgoalwillrequireamajorchangeinthe

rateofinnovation. (SeeTable1.)

Reachingthegoalwillrequiredramaticimprovementsinatleastthreeareas(eachisdiscussedin

greaterdetailinAppendicesEandF):


3/28




InstalledSystemPrice($/W)

2010 2016 $1/Watt

Module 1.70$ 1.05$ 0.50$

BOS/Installation 1.48$ 0.97$ 0.40$

PowerElectronics 0.22$ 0.18$ 0.10$

3.40$ 2.20$ 1.00$

CostofEnergy($/kwh)

2010 2016 $1/Watt

Module 0.063$ 0.037$ 0.018$


PowerElectronics 0.008$ 0.006$ 0.004$

O&M 0.013$ 0.009$ 0.003$

0.139$ 0.086$ 0.038$

Table1:Potentialutilityscalesystemcostbreakdowntoreach$1/watt

(notecapacityfactorsassumedare26%in2010and28%in2016)

Arrays: Higharrayefficiencyisessentialbothforreducingarraycostsandcuttingthearea(andthuscost)ofinstalledarrays. Fortunately,therearetechnologypathwaysthatmightachieve

thisgoal. Itisexpected,however,thatthemostpromisingislikelytobeonethatdoesnot

requireglassandcanbedepositedinexpensivelyonathinmetalorpolymersubstrate.

Reachingthegoalwillrequirefindingwaysto(a)designmanufacturablecellscapableof

achievingefficiencies

demonstratedinlaboratories

and(b)makinguseofthekinds

ofrolltorollproduction

devicesorotherapproaches

thatgreatlysimplify

manufacturingprocesses.

PowerElectronics:ConvertingtheDCoutputof

arraysintohighqualityACat

usefulvoltagesrequires

equipmentthataddsbothto

initialinstalledcostandto

maintenancecostssince

currentdesignsoftenfailafter

10years. Buildingonongoing

powerelectronicsworkatDOE,

atleasttwopromising

approachescouldbepursued:

(a)radicalredesignofcurrentlargeinverterunitsanduseofinnovativecomponents,and(b)

designingmodularinvertersthatcouldbecheaplymassproducedandattachedtoeachmodule.

Installation:Thecostofmountingandwiringarraysandtheassociatedequipmentisabouthalfthecostoftodayssystems. Twoapproacheswillbepursuedtoachievethedramaticcost

reductionsrequired:(a)installingarraysinfieldsonlightweightframeswithequipmentthathas

thesophisticationofagriculturalcombinescapableofcoveringhundredsofacresaday,and(b)

findingwaysofbuildingPVarraysintobuildingcomponentssuchasroofingsothatthe

incrementalinstallationcostcouldbeverylow. Arraysthatfollowthesunaresomewhatmore

expensivethaninstallationsthatdontmovebutcanproducemoreelectricityperyearperwatt

ofinstalledPVandcanproducemoreenergylateinthedaywhenmanyutilitiesneedmost

power. Trackingisusuallyalsoneededforunitsthatconcentratesunlightonhighefficiency

cells. Concentratingsystemsaddtocostsbutcanreducetheareaandcostofthephotovoltaic

devices.

Whileeachoftheseareascouldbepursuedasaseparatetask,thesuccessoftheprojectdependson

ensuringthateachoftheprogramsunderstandsthechallengesfacedbyotherareas. Arraysshould,for

example,bemassproducedforeasyinstallation.


4/28




III. OtherRequirementsAnumberofotherobjectiveswillneedtobemetinordertomakea$1/wattsystemcommercially

viableandscalabletomeetlongtermenergyneeds. Inordertorepresentasignificantadvancement

fromexistingindustrytrendsandmakeasignificantimpactonAdministrationGreenhouseemission

goals,the$1/wattgoalshouldbedemonstratedby2017. Thetargetcouldbedemonstratedbyfull

systemsorincrementaltoexistingsystemcosts,suchasforPVsystemsintegratedwithinnewroofing

systems. Toensurelargevolumescalability,the$1/wattsystemshouldbebasedonearthabundant

materialsandcapableoffullrecycling. Finally,the$1/wattsystemshouldmeetallapplicablesafetyand

environmentalstandards.

$1/wattinstalledby2017:DefiningtheObjective

By2017: Demonstrationofallkeycomponentsandinstallationmethodsinsystemsatleast5MWinsizeandinitialproductionordersforequipmentcapableofdelivering$1/watt

installedsystemsin2017

Includesallcomponents,equipmentandinstallationprocessestoproducegridcompatibleelectricity

Targetcouldbemetwithsystemsinstalledonthegroundoronbuildings Earthabundantmaterials Recyclablecomponents Meetsallapplicablesafetyandenvironmentalstandards


5/28




Appendices

AppendixA: PotentialImpactsof$1/watttoU.S.ElectricitySystem

AppendixB: BusinessasusualandthechallengeoftheUSelectricitymarket

AppendixC: CostGoalsoftheCurrentDOEprogramand$1/WGoalforutilityscalesystems

AppendixD: ManagementAlternatives

AppendixE: PotentialPathwaystoCostReduction

AppendixF: ChallengesforCostReductioninArrayProduction

AppendixG: PreliminaryanalysisofEfficiency,Cost,andReliabilityBarriers

AppendixH: TheneedforGovernmentfunding


6/28




AppendixA:PotentialImpactsof$1/watttoU.S.ElectricitySystem

PreliminaryNRELanalysisconductedwithReEDS1andSolarDS2suggeststhatifthe$1/wattgoalis

reachedby20203,morethan100GWofPVcouldbeinstalledcumulatively,representingabout5%of

thenationselectricgenerationcapacity.4 By2030,installedcapacitycouldgrowto389GW

representing14%ofU.S. generationcapacity.

Ifthesecapacityadditionsarerealized,utilitysystemswillneedtoadjusttheirgenerationmix

andoperationstoadapttolargeramountsofsolarenergygeneration. Theanalysisindicates

thatif14%ofautilitysenergycomesfromsolar,theseadjustmentscouldberelativelysmall,

onlyminimalamountsofnewtransmissionandstoragewouldberequired. Thetotalamount

ofnaturalgaspoweredintermittentandpeakingcapacitywouldfall,butahigherfractionof

thisequipmentwouldneedtobemaintainedas"spinningreserve. Inexpensivestoragewould

cutoverallcosts.AttheselevelsofPVadoption,consumerpricesofelectricitycouldbe2%lowerin2030ifthe

$1/wattgoalismet. Further,by2030CO2emissionsfromthepowersectorcouldbereduced

by

approximately

213

MMT

CO2

annually

and

the

growth

of

CO2

emissions

from

the

sector

couldbecutinhalfby2030.

1NRELsRenewableEnergyDeploymentSystem(ReEDS)isalinearcapacityexpansionmodelthatoptimizestheregional

expansionofelectricgenerationandtransmissioncapacitywithin356regionsofrenewableresourcedatawhilespecifically

addressingthevariablenatureofsomerenewableresources.http://www.nrel.gov/analysis/reeds/2NRELsSolarDeploymentSystem(SolarDS)isamarketpenetrationmodelforcommercialandresidentialrooftopPV,which

takesasinputregionalelectricityprices,financialincentives,regionalsolarresourcequality,androoftopavailability. Denholm,

P.,Drury,E.,andMargolis,R.,2009, TheSolarDeploymentSystem(SolarDS)Model:DocumentationandSampleResults,

NREL/TP61245832.3TheReferenceCaseusespreliminarytechnologycostandperformanceassumptionsbeingdevelopedforanother EEREstudy

andassumesa30%ITCforsolarthrough2016duetoARRAandthentheITCexpiresafter2016thathasnotyetbeenpeer

reviewed. The$1/Watt2020CaseusesthesametechnologyandpolicyassumptionsastheReferenceCase,butthecostofPV

rampsdownto$1/Wattby2020forutilityscaleandcommercialapplications.Forresidentialapplications,PVisavailablein

2020at$1/Wattfornewconstructionwithroofresurfacingmodeledbya5%houserebuildrateandretrofitsata40%cost

penalty(or$1.40/Watt).Forbothcases,ReEDSandSolarDSwereruntogetherinaniterativefashion. Thefossilandnuclear

numbersreflectpreliminaryestimatesbyanengineeringconsultingfirmandhavenotbeenreviewedbyDOE. Theyarebeing

usedhereasplaceholdersonly,toallowtheexplorationofthepotentialimpactof$1/WinstalledPVscenariosforthisstudy.4Theseresultsarepreliminary,notpeerreviewedorvettedforcitationorquotation.Properstudyofthisissuewouldrequire

additionalReEDS/SolarDSanalysistoprovideamorerobustunderstandingofthepotentialimpactofreachingthiscostgoalon

theU.S.electricitysystem.

2030Results ReferenceCase $1/WattnoITCpost2016

%GenerationofPV 1% 14%

CumulativeInstalled

Capacity(GW)

40GW 389GW,

232GWUtilityscale,157GWDistributed

NationalAverage

ElectricityPrice

10.45cents/kWh 10.23cents/kWh

AnnualCO2emissions

fromthePowerSector

2,408MMTCO2 2,194MMTCO2

Table2:Potentialim actontheU.S.electricit s stemin2030fromthede lo mentofa 1 WPVs stemb 2020


7/28




Figure1:GeographicDiversityofPVDeployment

Figure1showsthatwiththe$1/Wattsystemssolarelectricitycouldbecostcompetitivewithother

formsofgenerationinalmosteverystateinthecountryby2030.

SummaryResultsfromReEDSandSolarDSCasesInstalled

Capacity

2010 2030 2050

(GW) Reference $1/WCase Reference $1/WCase Reference $1/WCase

SolarPV .3 .3 40 389 83 592

Wind 35 35 49 38 118 70

Storage 21 21 23 24 24 30

NG 389 389 516 412 635 567

Coal 308 308 302 303 355 301

Nuclear 100 100 96 96 57 57Table3:InstalledCapacityinReferenceand$1/WattCases

TheresultsfromtheReferencecaseandthe$1/WattcasearepresentedinTables3and4andarenot

anofficialDOEperspectiveonthefuture,butpossibleoutcomesbasedononesetoftechnologycost

andperformanceprojectionsandotheraspectsoftheReEDSandSolarDSmodels,withparametersas

suppliedbyanexternalengineeringconsultingfirm,whichhavenotbeenpeerreviewednorreviewed

byDOE. Thesevaluesfornuclear,fossil,andrenewabletechnologiesarebeingusedhereas

placeholdersonly,toallowtheexplorationofthepotentialimpactof$1/WinstalledPVscenariosfor

thisstudy.Marketpenetrationlevelsofdifferenttechnologiescanbeattributedtoanumberoffactors

includingprojectionsoncostandperformance.

Generation 2010 2030 2050

(TWh) Reference $1/WCase Reference $1/WCase Reference $1/WCaseSolarPV 0.4 0.4 60.5 654 128 986

Wind 115 115 170 129 427 250

NG 767 767 1025 675 1313 1188

Coal 1886 1886 2235 2138 2629 2129

Nuclear 790 790 757 757 448 448Table4:ElectricityGenerationinReferenceand$1/WattCases

>30

20 30

10 20

5 101 5

0.1 1


8/28




AppendixB: BusinessasusualandthechallengeoftheUSelectricitymarket

Figure2belowhighlightsthecurrentstateoftheartforthetwoleadingphotovoltaictechnologies,

waferbasedsiliconandthinfilmCdTe. ThecostsarebasedonsystemsinstalledinPhoenix,AZanddoes

notincludefederal,state,orutilityincentives.

Figure2:ComparisonofcurrentandprojectedSolarPVcosts(Phoenix,AZ)toUSwholesaleelectricityrates(wholesalerates

ofleadingsolarenergyadoptedcountriesincludedforreferencetoUSratesonly)

TheUSandothercountriesnationalaveragewholesaleelectricityratesarerepresentedbybandson

Figure2(Source:EIA). Theindustryiscurrentlyfocusedonthosemarketswithhighelectricitypricesand

highsubsidies. SincePVisalreadyattractiveinthoseregions,manufacturersarefocusedonservicing

thoseopportunitiesthroughincrementalimprovementsthatincreasetheirprofitabilityandmarket

share. Reaching$1/WisachallenginggoalthatismoreimportantfortheUSthansomeofthoseother

countries. Projectionsforcostreductionsfortheleadingsiliconandcadmiumtelluridetechnologies

suggeststhatutilityscalesystems,whilealreadyattractiveinotherregionsoftheworld,willnotbe

broadlycompetitivewiththeUSaveragewholesalerateofelectricitywithoutsubsidiesby2016.6,7,8,9

Thecostreductionsneededareunlikelytobeachievedwithtechnologiesnowinwidespread

production. AnalysisshowningreaterdetailinAppendixFsuggeststhatcurrentcrystallinesiliconand

cadmiumtelluridetechnologiesarereachingthelimitsofwhatcanbeachievedthroughincremental

improvementofcurrentproductionmethods. Dramaticallynewideasarerequired.

6AppendixC

7AppendixE

8AppendixF

9FirstSolarAnalystInvestorMeeting,LasVegas,June24,2009

0.0

5.0

10.0

15.0

20.0

25.0

30.0

$0.25 $0.75 $1.25 $1.75 $2.25 $2.75

LCOE(centsperkWh)

NonModuleCosts($perWp)

UnsubsidizedUtility

Scale

Solar

PV

Energy

CostsMinimumSustainableModulePrice,MedianTechnologyEfficiency

Phoeniz,AZ;FixedPower(20MW)GroundMount

CdTeToday(2010),

$0.98/Wp,10.8%

CdTeProjected(2014),

$0.68/Wp,14.4%

cSiToday(2010),

$1.70/Wp,

14.4%

cSiProjected(2016),

$1.05/Wp,17.4%

2015ProjectedUSUtility WholesaleElectricityPrice

Ja an Wholesale Electrict Price

GermanWholesaleElectricityPrice

ItalyWholesaleElectricityPrice

DollarperWattGoal


9/28




AppendixC: CostGoalsoftheCurrentDOEprogramand$1/WGoalforutilityscalesystems(Figure3

&5)

Figure3:CostGoalsoftheCurrentDOEProgram(2016)vs.$1/WGoal(2017)forcSiutilityscalesystems

$8

$1.70

$1.05$0.50

$0.22

$0.18

$0.10

$1.48

$0.97

$0.40

$

$1

$2

$3

$4

$5

$6

$7

$8

$9

2004 2010(Est.) 2016(CurrentGoal) 2017($1/WGoal)

InstalledSystem

Cost($/W)

BOS/Installation

Inverter

PVModule


10/28




Table5:$1/WGoalofUtilityScaleSystemsbasedoncSi

ComponentCost($/W)

2010

(Est.)

2016

(Current

Goal)

2017

($1/W

Goal)

PVModule 1.70$ 1.05$ 0.50$Semiconductor 0.54$

RawMaterials(Sifeedstock,sawslurry,sawwire) 0.36$

Utilities,Maintainence,Labor 0.04$

Equipment,Tooling,Building,CostofCapital 0.06$

Manufacturer'sMargin 0.08$

Cell 0.45$

RawMaterials(eg.metallization, SiNx,dopants,chemicals) 0.18$




Module 0.70$

RawMaterials(eg.Glass,EVA,metalframe,jbox) 0.26$



Shipping 0.08$


RetailMargin $

Inverter 0.22$ 0.18$ 0.10$

Magnetics 0.03$

Manufacture 0.05$

BoardandElectronics(Capacitors) 0.07$

Enclosure 0.04$

Power

Electronics 0.03$


MountingandRackingHardware 0.25$

Wiring 0.14$

Other 0.17$

Permits 0.01$

SystemDesign,Management,Marketing 0.15$

InstallerOverheadandOther 0.19$

InstallationLabor 0.38$

Total 3.40$ 2.20$ 1.00$


11/28




AppendixD: ManagementAlternatives

Severaldifferentresearchmodelsshouldbeconsidered. Theseinclude:

1. SEMATECH: TheSEMATECHresearchconsortiumwasconceivedin1986outofaconcernthattheUnitedStateswasabouttolosetheentiresemiconductormanufacturingindustry

totheJapanese. JointlyfundedbytheDefenseAdvancedResearchProjectsAgency

(DARPA)andindustry,itreportedtoaboardwithonegovernmentmember. The

consortiumwonacompetitivesolicitation. IthadaClass1cleanroomoperatingin32

weeksandattractedthebestpeopleintheindustry. Itdevelopeddetailedroadmaps

showingwhatneededtobedonetoproducenextgenerationCMOSsemiconductorsand

focusedresearchonchallengesthecorporatemembersshared. Governmentfunding,and

governmentboardmembership,endedin1995,andtheorganizationcontinuesasan

independentorganization.

a. Advantages: Theorganizationensuresthatthetopicschosenarecloselyalignedtotherealneedsofindustry. Itcouldoperateflexiblyandquicklywithoutfederalred

tape.

b. Disadvantages: Theorganizationwasabletoidentifyanumberofcriticalareas(suchasmanufacturingequipment)whichwereneededbyallmembersbutwere

notpartofthebusinesseslinesofthememberswhichrevolvedaroundthedetails

ofchipdesignnotmaterialsormanufacturingtechnology. Thismeantthatthey

couldsharetheintellectualproperty. ConversationswithPV manufacturershave

raisedconcernsthatthismodelmaybedifficulttoapplytothearrayindustrysince

materialsandmanufacturingdetailsareatthecoreoftheirintellectualproperty.

2. SkunkWorks: TheLockheedAdvancedDevelopmentProjects(akaSkunkWorks)puttogether

a

team

that

crafted

some

of

the

most

amazing

aircraft

ever

built

including

the

U

2andtheSR71anddiditinrecordtime. TheSkunkWorksteamdesignedandbuilta

prototypeoftheXP80,oneofthefirstturbojets,inonly143daysandplayedamajorrolein

theKoreanWar.

a. Advantages: Allowsahighlycreativeteamtoaddresscomplexapplieddesignandfabricationproblemswithfewconstraints.

b. Disadvantages: Thismodelmaynotworkwithoutanexclusiverelationshipwithanindividualcorporation. Italsoreliesonextraordinaryleadership. Thecreativityfor

whichtheorganizationwasfamousbegantodissipatewhenleadershipchanged.

TheinfamousF22aircraftwasdevelopedbythesameorganization.

3. HUBs: HUBshavetheabilitytopulltogetheradiverseresearchteamandformtightallianceswithindustrypartnersensuringactivemovementofideasoutandproblemsin.

a. Advantages: Modeliswellestablishedandcouldbesetupquicklyb. Disadvantages: TheHUBsarewelldesignedtoaddressarangeofindividual

researchproblemsbutarenotmanagedaroundtightroadmapsaimedatmeeting

hardprice/performancetargetsbyadatecertain. Itmightbepossibletochange

this.


12/28




4. GeneralGroves: TheManhattanprojectsucceededbecauseofcreative,butveryhierarchical,management,ledbyLt.Gen.LeslieGroves. Someofthenationsmostcreative

scientistswereallowedroomforcreativity,buttightlyfocusedonthepracticalproblemsat

hand.

a. Advantages: Withtherightmanagementitcanbuildcreativealliancesbetweenscientificresearchandthedesignofpracticaldevices.

b. Disadvantages: Theprojectdidnotneedtodevelopcommercialproductsorconcernitselfwithintellectualpropertyorsimilarissues. Itwasabletoavoidmost

ofthelimitationsthatcreatedelaysinfederalprocurementandhiring. Anditwas

extraordinarilyexpensivesincethefederalgovernmentpickeduptheentirebill.


13/28




AppendixE: PotentialPathwaystoCostReduction

Atoplevelanalysisoffirstestimatepotentialsystemandcomponentperformancerequirementsand

characteristicsofa$1/WsystemissummarizedinTable6. Oneofthegreatestopportunitiesforcost

reductionliesinmoduleandsystemefficiencyenhancements. Efficiencyadvancesreducenotonlythe

PVmodulecostbutalsomuchoftheBalanceofSystems(BOS)andfixedsystemcosts. Forexample,if

modulesweretwiceasefficient,withthesamelaborandmountinghardware,twicethepowercanbe

obtained,

thereby

reducing

the

BOS

cost.

AnalysisofthemanufacturingandinstallationcostsofPVsystemsrevealsseveralpathwaystoachieve

the50cents/wattmodulecosttargetthatisconsistentwitha$1/watttotalinstalledsystemcost. The

primaryrequirementsarethatmoduleefficiencyneedstobe25%orgreaterandproductlifetimeneeds

toexceed30years.10 MeaningfulBOScostreductionisnotpossiblewithoutraisingmoduleefficiency.

Productlifetimeneedstobelongenoughtoamortizetheinstallationanddisposalcosts.11

A. LowCostArraysThegoalistobuildanarrayatacostwhichmakesapriceof50centsperwattfeasible,whileachieving

efficiency

greater

than

25%

and

a

lifetime

of

at

least

30

years

with

minimal

maintenance.

The

modules

mustbebuiltattremendousvolumewhichrequiresthattheybebasedonearthabundantmaterial.

Andtheymustberecyclable. Thelowcostsuggeststhattheyalmostcertainlywillnotusetraditional

thickglassorsignificantamountsofpurifiedsiliconthatcontributesignificantlytotodaysarraycosts.

10AppendixF

11NonModuleCostSensitivitytoEfficiency,DOE$1/WWorkshopPhotovoltaic(PV)IndustryPrimer,July2010.

Characteristic ValueorQualifier

Module

Efficiency >25%

Substrate Lowercostandweightthanglass

Reliability 30yearsorcanbereplacedwithminimumlabor

Materials Earthabundant, nontoxicorestablished

recyclingplan.

BOS/Installation

Labor Canbedonewithnonspecializedlabor

Process Lightweight(easeofhandling,nospecial

equipment)

Assembly Snaptogethermechanicalandelectrical

PowerElectronics

Efficiency >95%,improvedmodulepeak

powermanagement

Reliability 30years

Assembly Integrationofwiring,componentstominimize

electricalconnectionsTable6:SuccessCharacteristics


14/28




Devicescapableofmeetingthegoal:

AsshowninFigure4,commercialcellsoperatefarbelowtheoreticalpotentialefficienciesandwell

belowefficienciesdemonstratedinthelaboratory.12,13 Duetothecompoundedsystemcostbenefits

associatedwithefficiencygains,closingthesegapsiscritical.

Wafersiliconisamaturetechnologyandefficienciesofcommercialcellsareapproachingtheoretical

limits. Thinfilmapproachescanreach25%orhighermoduleefficiencyoninexpensivesubstrateswith

potentiallyscalableandcosteffectivedepositionmethodologies,butsincetheyarecomparativelynew,

thegapseparatingtheoreticalfromproductionarraysisquitelarge. Cadmiumtelluridearraysarein

largescaleproductionbuthaveonlyachieved11%moduleefficienciesinproduction;and17%in

laboratorydevices. Thereisalsoconcernthatthecostofreclaiming,andrecyclingcadmiumcontaining

PVmoduleswillescalateinthefuture.14 ThinfilmCIGS,whichalsocurrentlycontainsverysmall

12http://www.tfp.ethz.ch/Lectures/pv/thinfilm.pdf

13http://scitation.aip.org/getpdf/servlet/GetPDFServlet?filetype=pdf&id=APPLAB00009500001616330

2000001&idtype=cvips&prog=normal&doi=10.1063/1.3243986&bypassSSO=1 14

http://seekingalpha.com/tag/transcripts?source=headtabs

Figure4:Gapsinefficiencybetweenbestlaboratoryresultsandtheoreticallimitsandbetweenproductionandbest

laboratoryresultsprovideopportunitiesforimprovement. (TheoreticalbasedonShockleyQueisserlimitandbandgapof

semiconductor. NRELverifiedbestcellefficiency)


15/28




amountsofcadmiumsulfide,hasevengreateropportunityforefficiencyimprovementwith10%to11%

inproductionandmorethan20%inthelab. ThefuturecostofCIGSmodulesandthistechnologys

abilitytocontributetoverylargescalePVpowermarketsdependsupontheavailabilityofindiumand

gallium. CZTS,amaterialsystemverysimilartothinfilmCIGSbutwithindiumandgalliumreplacedby

moreearthabundantzincandtin,providesarelativelynewopportunityforatransformationalchange

ofthePVmarketinsupportofreaching50cents/wattmodulecosts.

Organicphotovoltaic(OPV)anddyesensitizedcell(DSC)deviceshavereceivedsignificantinterestas

promisingpathstoveryinexpensivesolarmodules. Tobecommerciallyviable,thesetechnologiesmust

solvesignificantissueswithefficiencyaswellaslifetime. Significantbreakthroughsintheseareasare

stillneeded.

ConcentratedPV(CPV)takesadifferentapproachtoconventionalflatmodulesbyleveragingadvanced

opticalsystemstoshiftthecostbalanceofcellmanufacturingtomodulemanufacturing. Higher

efficiencyCPVdevicesaswellasinnovativesystemsapproachestoCPVmoduleassemblycouldincrease

thetypical25%orlowerefficientmodulescurrentlyproducedtogreaterthan35%moduleefficiencyif

thesemiconductordevicecanexceed50%efficiency. Advancesinopticaldesign,manufacturingand

assemblywouldalsoberequired. Analternativeapproachistoapplyliftoffprocesstothehigh

efficiencyIII/VcellsusedinCPVmodules. Theseprocesseshavethepotentialtosignificantlyreducethe

costsofIII/Vcellsandmakethemavailableforunconcentratedflatplateorflexiblemodules.

Whiletheuncertaintiesaremuchhigher,itispossiblethatinexpensivearrayswithefficienciesmuch

higherthanthoseshowninFigure4canbebuiltusingadvancedmultijunctionthinfilmsformanyIII/V

andII/VIsystemsthroughbandgapengineeringsuchastandemjunctionstructuresonthinfilm

materials(Ex:InSbonthinfilmCdTe,orGaPonthinfilmSi). Suchinnovationswouldrequire

breakthroughsintunneljunctionformationoverlargeareasacrosspotentiallyhighlylatticemismatched

polycrystallineinterfaces.Perhapsrecentadvancesinthermoelectricdevicescouldmakeitviablefor

photovoltaicdevicesinaconcentratingsystem.

ModuleProduction

Technology:

Lowproductioncostsare

clearlyessentialtomeetthe

pricegoals. Thiswill

requiredramatic

innovationsthatcancut

bothcapitalandlaborcosts

inarraymanufacture.

BasedonDOEanalysis,cellsproducedbelow50centsperwattwouldlikelymean10centsperwattfor

capitaldepreciation,23centsperwattformaterials,and6centsperwattforlabor.

2010 2016Proj

Cost Cost Cost($ /W) Cost($/m2)

Capital $0.24 $0.20 $0.10 $28

Materials $1.11 $0.49 $0.23 $68

Labor $0.27 $0.12 $0.06 $17

Margin $0.79 $0.24 $0.11

TotalModule $1.70 $1.05 $0.50

$1/WTarget

Table7:DetailedcostbreakdownforcSiPVmodule($/m2valuesassume29%

efficiency


16/28




Capitalcosts:Assumingasevenyeardepreciationperiod,initialcapitalcostswillneedtobeonthe

orderof70cents/watt. Thisisapproximately50%lowerthanthecostsoftodaysstateoftheartthin

filmPVfacilities.

Materialcosts: ifweassumeanextremelyambitiousmoduleefficiencyof29% thelimitofwafered

siliconreachingamaterialscostof23cents/wattwouldbeequivalenttoanareacostof$68/m2

. Even

atthishighefficiency,currentcommoditymaterialssuchcoverglass,substratesandbacksheets,silicon,

silver,aluminumframingcouldconsumetheentirebudgetwithlittlecurrentpotentialforcost

reduction(Figure6showsagraphicaldepictionoftheseandothermaterials). Atcurrentcosts,glass

withTransparentConductingOxides(TCO)andantireflectioncoatingmighttakeupasmuchas16%

to25%ofthistotal,indicatingtheneedforcheapersubstratesandbettercoatingtechnology. Frames

andotherstructuralpartsmighttake

upasmuchas25%ofthetotal

indicatingtheneedforframeless,

possiblyflexiblemoduletechnology.15

Productioncostscanbereduced

sharplywithprocessinnovations

includinglowcosthighlyscalable

approachessuchasdepositiononlow

costnonwoventextilesorrolltoroll

filmprocessing(Figure6).

Innovationsinanumberofareas

couldsharplycutproductioncostsfor

advancedmoduledesignsincluding:

Lowcostorvirtualsinglecrystalenablingsubstratesreducingwasteassociatedwithsawingwafers,therebyreducingmaterialsusage(gramsperwatt)byafactorof2ormore

Ultrathinwafersthatnotonlyenablegreatermaterialutilization,butalsoinherentlyimprovecertainlossmechanisms(ie.bulkrecombinationlosses);optimizewaferthickness

Lowstresstabbingtechniques,suchasprepatternedbacksheetsthatnotonlyenablethestringingofultrathincellsintomodulesbutmayalsoenablesemiflexibleroofinglaminates,

greatlyreducingdeploymentcosts.

Defectengineeringandadeeperscientificunderstandingoftheroleofimpuritiesandstructuraldefectsaswellaswaystomakethemelectricallyinactivecanimprovetheefficiencyofsolar

cells

FlexiblePVsolutionsthatenableinnovativeBIPVandfielddeploymentapproachesthatcouldsignificantlyreduceBOScosts.PVmodulesthatserveadualpurposeasabuildingfacade,or

roofingmembranefurtherreduceinstallationcosts.

15NRELAnalysismemo,SolarManufacturingCostModels,AlanGoodrich,April29,2010.

Figure5:SiliconPVmodulecontainsmanycommoditizedraw

materialsincludingsilicon,glass,andaluminum. Thechallengewill

betoreducethecostsorutilizationofthesematerialswhile

improvingmoduleefficiency.(Graphicsource:Hisco)


17/28




Inexpensiveandhighperformance(conductivity,transmissivity;absorption)TCOsordesignsthateliminatetheneedforvacuumprocessingofTCOsentirelyreducestheseriesresistanceof

PVmodules.

AtomicbarriersthatcanprovidebettermoistureimpermeabilitytothePVdevicethanglass,therebyincreasingreliabilityandfunctionalmodulelifetime.

HighgrowthrateepitaxyofGroupIV,III/V,andII/VIsemiconductorsonlowcostsubstrates.VeryhighfrequencyplasmadepositionorotherprocesseshasthepotentialtoincreasePV

depositionrateandreduceequipmentcostby10Xwhilemaintaininglargeareauniformityand

materialqualitythroughnew approachestosourceandprocessdesign.

Breakthroughapproachestodepositionfromatmosphericpressureliquidprocessingcouldsignificantlyincreasethroughputandlowerdepositioncosts

Plasmonicsandnanowirethinfilmscanenhancelighttrapping

Combinatorialapproachestomaterialsdiscoveryandcharacterizationaswellas

deviceoptimizationcangreatlyspeed

upthedevelopmentofnew

technologiesandallowefficiencyto

approachtheoreticallimits. APV

MaterialsGenomeProjectcouldenable

researcherstorapidlyscreencandidate

PVabsorbersystemsandlinktheoretical

understandingPVmaterialswith

experimentalconfirmationofmaterialproperties.

Figure6:RolltoRollcontinuousprocessingcansignificantlyreducemanufacturingcostshasithasdoneinanumberof

otherindustries

Figure7:Highlyautomatedagricultureequipment

revolutionizedharvestingofcrops.http://www.businessweek.com/magazine/content/08_22/b4086

072681496.htm?chan=search


18/28




B. BalanceofSystemandSystemInstallationTheBOSandinstallationcostsincludeforexample:mountingandrackinghardware,installeroverhead,

permitsfees,landpreparation,andinstallationlabor. This

currentlyaddsabout$1.48/watttothecostofaninstalled

system. The$1/wattgoalrequiresareductiontoabout

$0.44/watt(Table1). Itwillprobablybenecessarytocut

installationlaborcoststoapproximately5cents/wattand

mountingandrackinghardwaremustreducetoabout20

cents/watt. Thiswillrequirefundamentallynew

approaches.

Asabenchmark,amodern20MWutilitysolarplantcovers

approximately150acres. Theinstallationtakesacrewof

100peoplesixmonthstoinstallcompletely.16 Laboralone

adds10centsto15cents/watt. Twoprincipalstrategiesfor

reducinginstallationcostswillbein

competition: (1)reducingthecostofinstalling

hugearraysinopenfields,and(2)

incorporatingthearraysintobuilding

componentsthatcouldsubstituteforstandard

roofingmaterials.

FieldInstallations:

Modernagricultureuseshighlyproductive

machinescalledcombinesthatcanprocess

(reaping,binding,andthreshing)200acresaday,as

showninFigure7. Thesemachinesofferamodelof

whatcouldbepossibleemployinginnovative

roboticapproachestoPVfieldinstallation.

LimitedautomationforPVinstallationisalreadyin

existencewherecurrentlysomePVinstallations

machinesdrivepostsintotheground,asshownin

Figure8,butthisdeviceisdesignedonlytoputtens

ofthousandsofpostsintheground. Muchgreater

16Quote:DataprovidedbyaninstallerforcSiarraydeployedwith1axistracking.

Figure8:Lowlevelsofautomationforsolar

fieldinstallationsarejustbeginningtobe

developed. http://www.philadelphiasolar.eu/philadelphia_solar_gallery.html

Figure9:PlasticcoveringastrawberryfielddemonstratingthepotentialofrollingoutPV.

http://photosbygarth.com/sampleslg/050115_202p_9584lg.jpg

Figure10:Solarrooftilesdemonstratingthepotential

forlargeautomatedsnaptogetherPV.

http://www.reuk.co.uk/OtherImages/solartiles.jpg


19/28




costreductionscouldbeachievedusingadevicethatcombined1)postdigging;2)rackhardware

installation;3)modulemountingand4)electricalconnection.

Inadditiontoautomatedinstallation,innovativearraydesignscanleadtomajorcostreductions. These

include:

RollingOutPV(i.e.,PVonaflexiblesubstrate,similartohowplasticisrolledacrossfarmfieldsforfumigation,Figure9).

ContinuousPV(i.e.,modulesthatcannearlycontinuouslybedeployedanalogoustopavingcrewsthatdrivedowntheroadremovingpavementandinstallingfreshpavementinacontinual

process)

BuildingInstallations:

Installationcostscanalsobecutsharplyifthephotovoltaicarrayscanbeinstalledasabuildingmaterial.

Arraysintegratedintoroofingmaterialscouldcombineweatherproofingwithelectricgeneration. The

incrementalcostofinstallingthearrayscouldbequitesmallandlittleornoadditionalstructuralsupport

wouldberequired. Thesystemswouldbeeasiesttodesignfornewconstruction,butitispossiblethat

thedevicescouldbeintegratedinmembranesorotherequipmentusedtoreplaceworncommercial

roofing. OnesuchconceptisshowninFigure10usingmodulesthatrapidlyinterconnectanddisplace

thecostsofcurrentroofingmaterialsandlabor.

Centraltoeachoftheseproposedinstallationcostreductionstrategiesistheneedtoreducethe

amountofspecializedlaborthatisrequired. Thefollowingsectionsproposemethodsforintegrating

intothemodulethepowerelectronicsandothersystemcomponents,therebysignificantlydecreasing

PVinstallationlabor.17

C.

Power

electronics

Thepowerelectronicsspecificallytheinverteristheinterface

betweenthemoduleandthegridconvertingthedirectcurrent

outputofthearraysintothehighvoltagealternatingcurrent

neededformostpowerapplications. Itcurrentlyaddsabout22

cents/watttosolarinstallationsandthismustbecuttoabout8

cents/watttomeetthegoal.18 Thecostandperformanceofthe

invertermaynotbeadominantportionofthetotalinstalled

systemscostbutitisasignificantportionandneedstobe

addressedinordertoachieve$1/watt. Today,thepower

electronics(1)arethedominantpointoffailurefortheinstalled

systemandareamajorcomponentofmaintenance,(2)are

17Electricalcomponentinstallationlaborcanaccountforasmuch78%ofthemanhoursrequiredforautilityscale

system. Thenationalaverage(source:RSMeans,2010)burdenedelectricianrateis$72.85/hour. PVhardware

installedbygeneralorroofingcontractorswouldprovideasavingsofupto19%.18

Dataprovidedbyananonymoussystemsinstaller.

Capacitor

Type

FailureRate

(%/1000h)

Electrolytic 0.2

Tantalum 0.1

Paper 0.05

Ceramic 0.025

Table8:ApproximateReliabilityof

capacitortypes


20/28




responsibleforalossofapproximately4%ofalloftheelectricitygenerated,and(3)addcomplexityand

costtowiringandinstallation. Megawattscaleinvertersweighinexcessof10,000pounds,andoccupy

morethan500cubicfeetofspace.

Currentgenerationinvertersareonlyexpectedtolastabout10to15years,

requiringatleastonereplacementduringananticipatedlifetimeof30years

orlongerforaPVarray. Thislimitedlifemeansthattheinverters

contributiontoalevelizedcostofelectricityishigherthantheoriginal

installedcost. Onereasonforlimitedlifetimeistheuseofelectrolytic

capacitorswhichhavefailureratesthatare10timesgreaterthanthatof

lowerenergydensitythinfilmcapacitors(seeTable8,Figure11). Itis

projectedthatcurrentcostreductionsintheinverterwillbeachieved

throughadvancedcircuitarchitecturethatavoidstheneedforelectrolytic

capacitors;however,thisaloneisnotsufficienttopermitlargescale

deployment. Besidescost,inverterperformanceandfunctionalitywill

becomeanincreasinglyimportantfactorwithhigherlevelsofpenetrationofrenewableelectricityon

thegrid. BettercommunicationsandfunctionalitywillrequireradicallydifferentarchitectureforPV

powerelectronics.

Approachestomajorimprovementsinpowerelectronicsfallintotwocategories:(1)majorredesignsof

existinglargescaleinverters,and(2)developingsmallunitsthatcanbemassproducedandattachedto

individualmodules.

Majorredesignsofexistinglargescaleinverters:

ExistingMWscaleinvertersemployhighvoltagesiliconswitchesandlargemagnetictransformers. The

highvoltagesonthetransmissionsideoftheinverteraremanagedthroughacombinationoflarge60Hz

transformersandstackedsiliconswitches. Advancedhighvoltage,highfrequencyswitchcomponents,

lowlossmagneticmaterials,andnovelcircuitarchitectureshavethepotentialtosignificantlyreducethe

sizeandcostofMWscaleinverterswhilesimultaneouslyincreasingtheoverallefficiency.

Forexample,widebandgapsemiconductorssuchasSiChavethepotentialtoswitchat13kVwith

frequenciesashighas50kHz. Thehighervoltagereducesthepackagingcostandcomplexityofthe

system. Thehigherswitchingfrequenciesdramaticallyreducethesizeandcostofthetransformer

becauseforfixedimpedance,theinductancescalesinverselywiththeswitchingfrequency. Advanced

magneticmaterials,suchasnanocrystallinecompositeswithlowelectricalconductivityandlow

hysteresis,canenableswitchingfrequenciesthatare1001000timeshigherthanemployedtoday.

Scalingtheswitchingfrequencyfrom60Hzto50kHzcanallowthecoretransformertoscalefrom8,000

lbs.tolessthan100lbs.

Powerelectronicsattachedtomodules:

Electroniccomponentsandcircuitarchitecturesthatcanbeembeddedinthemoduleframewouldallow

maximumpowerpointtrackingtooccuratthemoduleorsubmodulelevel. Theresultingmodules

wouldbemoretoleranttopartialshading(allowingforpotentiallydenserinstallations)andcould1)

Figure11:Exampleofa

failedelectrolytic


21/28




routepoweraroundsuboptimalmodulesorcells(improvingsystemavailabilityandmitigating

reliabilityofindividualcells),and2)directlyproduceACvoltages(simplifyingresidentialinstallation

whileincreasingthesafetyandreliability).

Onepossibleapproachwouldbetocreatemoduleintegrated

powerelectronicswhereadvancedsemiconductorswouldneed

tobeintegratedwithbeyondstateoftheartmagneticand

dielectricmaterialstorealizelowcost,smallformfactor,batch

manufacturedpowerconvertersthatcouldpotentiallybe

installedoneachmodule(Figure12). Theseelectronicswould

requireadvancedsemiconductormaterialsthatcouldwithstand

hightemperatures(100Cbackskintemperatures)andmaintain

highfrequencyswitching.

Thesameapproachofhigherinternalswitchingfrequencythat

wasdiscussedforutilityscaleinverterscouldalsobeappliedto

moduleintegratedpowerelectronics. Suchanarchitecture

wouldpermitmodulestobeeasilyconnectedandimprovethe

overallenergyefficiencyofthesystembyallowingeachmicroinvertertofrequencymatcheach

module.

Figure12:Microconvertersoneach

modulepromotebettercommunication

andfunctionalityamongstmodules.


22/28




AppendixF: ChallengesforCostReductioninArrayProduction

SolarPVtechnologieshave,todatemadesignificantprogressreducingthecostofmodulesbuta

detailedanalysisofthefutureofcrystallineSiliconandCadmiumTelluridecellssuggestthattheyare

unlikelytomeetthetargetpricesgiventechnologiesdrivingcurrentlearningcurves. Thefigurebelow

showsthatcrystallinesiliconmodulepricescouldreacharound73centsperwattpricesiftheyareable

toachieve80%ofthetheoreticallimitofsinglecells(29%)andifwaferthicknesscanbereducedto80

microns(theyareabout180micronsthicktoday).

CadmiumTelluridemoduleshaveachievedsharppricereductionsandareonapathofcontinuouscost

reductiondriveninpartbycontinuousimprovementsinefficiency. Currentcellshaveanefficiencyof

approximately11%butefficienciesof17%havebeenachievedinthelaboratoryandtheoretical

efficiencies

are

approximately

29%.

The

experience

of

the

past

few

years

suggests,

however,

that

efficienciesabove11%areextremelydifficulttoachieveinpracticeandtherateofefficiencyincreases

hasslowedconsiderablyinrecentyears. Makingtheoptimisticassumptionthatefficienciescanreach

14%whilethecostofproducingaunitofarrayareadeclinesby26%,arraypriceswouldstillbearound

63centsperwatt.

Neartheoreticallimits:~24%efficientcells;low

($32/kg)siliconprice,80micronkerflesswafers


23/28




Figure13QuarterlyreportedmoduleefficienciesfortheleadingCdTemanufacturerFirstSolarsuggeststhatsignificant

innovationisrequiredtocontinuetoadvancethetechnology.

Figure14:GlobalsolarPVmodulepricetrend:siliconwaferbasedandCdTe,historicandshorttermforecast19

19Sources:Mints,Navigant;BloombergNEF;FirstSolarEarningsReports;NRELinternalsiliconPVcostmodel

Costreductionnecessarytoachieve

$0.52/Wcost($0.63/Wprice)at

14%efficiencyrequiresa26%

reductioninmanufacturingcosts

whileincreasingefficiencyby32%


24/28




Nevertheless,pastperformanceisnotnecessarilyanindicatoroffutureprogress. Thetrendofmodule

pricesdepictsthecompositeimpactofseveralcomplicatedfactors,includingforexample:

manufacturingefficiencies,economiesofscale,andinnovation.20 Asthesetechnologiesmature,the

contributionofmanufacturingefficienciesandeconomiesofscaletofurthercostreductionswill

diminish. Theimportanceofcontinuedandimpactfulinnovationswillbecomecriticaltomaintaining

thehistorictrend.

ForcSi,withoutnewmoduleencapsulationmaterials,costswillreachanaturalasymptotedefinedby

thecostofcommoditymaterialslikeglass(asymptote,orlowerlimitofcurrentcSimodulecosts

depictedinFigure14bybluedashedline).

ForCdTe,theleadingthinfilmtechnology,furthercostreductionwillrequirehigheraveragemodule

efficiency,expectedtoreachover14%by2014basedonaroadmapprovidedbyFirstSolar. Whilethe

pathwaytothisgoalisknownandinvolvestechnicalinnovations,implementingtheseinnovationsina

highvolumeproductionscenariowillbechallenging. Greaterchallengesexisttoincreasethis

technologysefficiencybeyond14%,suchasimprovingthelifetimeoftheabsorberlayer. Itisgenerally

believedthatinordertoimprovethequalityoftheabsorber,thegrainsizewillneedtobeincreased.

Onewaytoaccomplishthisistoslowdownthedepositionprocess,butthishastheundesirableeffect

ofincreasingthecost. Innovationinmaterialstechnologyisrequiredtoeliminatethistradeoff.

Ifweassumeabreakdownbetweencapital,materials,andlaborfollowingtheproportionsprojectedfor

atypicalcSimodulein2016,giveninTable7,thenthis48centswillbecomposedof9cents/wattfor

capitaldepreciation,22cents/wattformaterials,and5cents/wattforlabor. Thisbreakdownprovides

furtherillustrationofthechallengestoreach$1/watttotalsystems. Forexample,assumingaseven

yeardepreciationperiod,initialcapitalcostswillneedtobeontheorderof63cents/watt,

approximately50%betterthanFirstSolarsand25%to33%ofcurrentwaferedsilicon.

Formaterials,ifweassumeanextremelyambitiousmoduleefficiencyof29% theShockleyQueisser

limitofwaferedsiliconthe22cents/wattwouldbeequivalenttoanareacostof$65/m2. Evenatthis

highefficiency,currentcommoditymaterialssuchcoverglass,substratesandbacksheets,silicon,silver,

aluminumframingcouldexceedthesystemsbudgetformodulecost. Atcurrentcosts,glasswithaTCO

andantireflectioncoating

mightspecificallytakeupas

muchas16%to25%ofthis

total,indicatingtheneedfor

cheapersubstratesand

coatingtechnology. Frames

andotherstructuralparts

mighttakeupasmuchas

25%ofthetotalindicating

20GregoryF.Nemet,Beyondthelearningcurve:factorsinfluencingcostreductionsinphotovoltaics,Energyand

Policy34(2006)3218323,August2005

2010Es t. 2016Proj

Cost Cost Cost($/W) Cost($/m2)

Capital $0.20 $0.20 $0.09 $27

Materials $0.79 $0.49 $0.22 $65

Labor $0.09 $0.12 $0.05 $16

Margin $0.62 $0.24 $0.11$1.70 $1.05 $0.48

$1/WTarget

Table9:DetailedcostbreakdownforcSiPVmodule($1/WTargetassumes29%

efficiency)


25/28




theneedforframeless,possiblyflexiblemoduletechnology. Goingtowardflexibleorotherlightweight

structuralmoduleapproachescouldalsosignificantlyreduceshippingcosts,currently8centsto10

cents/watt.21

Table9assumessignificant

potentialcostreduction

throughproductionmodels

thatcondensethesupply

chain. Waferedsilicon

modulescurrentlyhaveas

manyasfourdifferentsteps

inthesupplychain

(polysiliconprocessing,wafering,cellproduction,andmoduleassembly),eachextractingaseparate

marginwhichaddscosttothefinalproduct. Whiletherehasbeensomeattemptsatintegration(mostly

throughcombiningpolysiliconprocessingand/orcombiningwaferingandcellandmoduleproduction)

therehasalsobeencountertrendstowarddisaggregation. Movingtowardthinfilmtechnologiesthat

areinherentlymoreintegratedwouldsignificantlyreducemargins. FirstSolarforinstancebringsin

basicmaterialssuchasglass,depositiongases,andothermaterialsinoneoftheirfactoriesandships

outfinishedmodulesfromtheotherend. Thereiscurrentlynoanalogousproductionfacilityfor

waferedsilicon. Movingtheindustrytowardthinfilmtechnologieswouldalsocountercurrent

competitiveadvantagesofChinesewaferedsiliconmanufacturersandpotentiallycreatemore

opportunitiesforU.S.exports.22

SimilaranalysiscouldbedonetolookatthechallengeswiththeBalanceofSystems(BoS)and

Installationcostcomponents. ReferringtoTable10,installationlaborandoverheadtoinstallthe

$1/watt

system

would

be

approximately

19

cents/watt.

For

large

utility

systems,

only

one

fifth

of

the

laborisformechanicalinstallationwiththemajoritybeingforelectricalconnections. Thisindicatesthat

thereexistsanopportunitytoreduceinstalledcoststhroughefficientcomponentdesign. Forexample,

microinvertorsintegratedintoeachmodulemightbedevelopedtoexploitnewinnovativeconnection

schemesthatreduceelectricalinstallationlaborcosts. Otherpathsforcostreductionmightinclude

havinglargerpanels,potentiallyinstalledasrollsofflexiblematerial,withmanyoftheelectrical

connectionsintegrated.

21NRELAnalysisinternalmemo,SolarManufacturingCostModels,AlanGoodrich,April29,2010.

22Asia:65%siliconwaferglobalmarketshare;CompanyproductioncapacitiesSiliconforSolarCell,RTS

Corporation,September2009

2010 2016Proj

Cost Cost Cost($/W) Cost($/m2)

Mounting,Wiring,Other $0.87 $0.38 $0.17 $50

InstallationLabor,OH,othe $1.10 $0.41 $0.19 $54

Permitting,Design,Mgt $0.23 $0.18 $0.08 $24

$2.20 $0.97 $0.44

$1/WTarget

Table10:DetailedcostbreakdownforBOS/Installation($1/WTargetvalues

assumes29%efficiency)


26/28




AppendixG:PreliminaryanalysisofEfficiency,Cost,andReliabilityBarriers(Table11)

FollowingtheexampleoftheSematechroadmap,theabovetable,dividesproblemsintofour

categories:

Manufacturablesolutionsexistandarebeingoptimized

Manufacturablesolutionsareknown

Interimsolutionsareknown

ManufacturablesolutionsareNOTknown

ForCPV,targetis 35%Module,currently


27/28




AppendixH: TheneedforGovernmentfunding

Venturecapitalandotherprivatecapitalsourceshavefueledsignificantgrowthinthedevelopmentand

manufacturingofPVandothersolartechnologiesoverthelast5years,asshowninFigure15. These

investmentsfocusedlargelyontechnologiesthatcouldbecommercializedin2to4years,inorderto

respondtogovernmentincentiveprograms. Thistrendhaspeaked,however,and,asshownfor2009,

overallVCandprivateequitycapitalinvestmentsinsolararedecreasing. Thisisduetothecurrent

economicconditionswhichhavelimitedopportunitiesforprivatecapitalsourcestoexittheir

investmentsthroughpublicmarkets,alsoshowninFigure15. Thedevelopmentofsolartechnologiesis

capitalintensiveandmanyprivatecapitalsourcesarefocusedonseeingtheirexistinginvestments

throughtomaturityratherthanconsideringnewinvestments. Theseandotherfactorshavetherefore

limitedtheavailablepoolofprivatecapitaltoinvestinthenextgenerationofsolartechnologies.

ThecurrentenvironmentthereforeprovidesaneedandanopportunityforFederalR&Dfundingtofilla

gapinthesolartechnologyinvestmentpipeline. TheFederalgovernmenthasplayedthisrolebeforein

launchingthecurrentsolartechnologyindustryaswellasothernewindustriesinelectronicsand

biotechnology.

Inadditiontoaddressingtheadministrationsenergyandenvironmentalgoals,aninvestmentbythe

Federalgovernmentthattargetstheutilizationofthecountrysextensiveresearchandmanufacturing

infrastructurewillcontributesignificantlytowardsexportandeconomicgrowthinitiatives. Itis

Figure15: USCapitalInvestmentsinSolarEnergy

$0

$500

$1,000

$1,500

$2,000

$2,500

$3,000

2000 2001 2002 2003 2004 2005 2006 2007 2008 2009TotalU.S

.Investment(MillionsNominal$)

Year

Solar Public Equity Activity

VC & PE Investments

TotalInvestmentIncrease

33%CAGR 20002009

36%CAGR 20002005

30%CAGR 20052009

U.S. Capital Investments in Solar Energy


28/28



estimatedthatby2012,andassumingstandardsolarPVcosttargetsandglobaldemandprojections

exportopportunitiesthroughoutthesolarPVsupplychainwillexceed$18.0billion.23

Othercountries,includingChinaandMalaysiahavemadearenewedcommitmenttotheirnationssolar

PVindustry,attractingglobalmanufacturers,includingcompanieswhosetechnologieswere,insome

cases,atleastpartiallydevelopedatUSinstitutions.

23NRELanalysismemoinsupportofthePresidentsNationalExportInitiative,SOLARENERGYTECHNOLOGIES:

USExportOpportunities,May2010

$1 per w photo voltaic systems

Documents