thin-film pv - a system designer's guide - solarpro

7/30/2019 Thin-Film PV - A System Designer's Guide - SolarPro

1/9

By Rick Holz, PE

A SystemDesignersGuide

Thin-Film PV:


2/9

spfessin.cm|SolarPro35

The thin-flm PV indus-

try has seen extraordi-

nary growth in the past 8

years. According to GTM

Research, about 17 MW

o thin-flm modules were

produced globally in 2002,

accounting or a 3% mar-ket share. anks in large

part to the global polysilicon shortage o the mid-

2000s, not to mention the $4-per-watt crystalline

silicon (c-Si) PV module prices that ollowed, thin-

flm commercialization activities increased. e

search or alternative technologies led to a tidal wave

o investment and entrepreneurial activity in thin

flm, writes Shyam Mehta, senior analyst at Green-

tech Media, with 46 companies entering the market

between 2004 and 2008, as well as $1.8 billion in ven-

ture capital investment in the space. Displaybank, a

market research and consulting authority, reported

that thin-flm cell production reached 1.9 GW in 2009,

or a 19.8% market share, and the company orecasts

2.8 GW o thin-flm cell production or 2010.

Notably, 2002 was the year that First Solar

launched production o its commercial cadmium

telluride (CdTe) products. Fast-orward to today

and the company expects to achieve a produc-tion capacity o 1.4 GW in 2010. While production

capacity and module production are not the same,

First Solar reportedly produced 1.1 GW o mod-

ules in 2009 out o a total production capacity o

1.2 GW, proving that demand or its CdTe products

was strong. Mark Osborne, the senior news editor

at PV-tech.org, notes that the companys 2009 pro-

duction is a record or the industry. is is not only

a record or the thin-flm market, but also a record

or the PV industry as a whole.

The largest

supplier ofPV modules

in 2009 was a

US-based thin-film

manufacturer.

With thin film

poised to deliveron its much-

heralded promise,

what do system

designers and

integrators

need to know?Cortesyjwisolar


3/9

36 SolarPro|Decembe/Jnuy2011

Thin-Fi lm PV

Despite these very encouraging signs or the thin-flmindustry, competition with traditional c-Si PV manuacturers

has never been fercer. As the polysilicon shortage passed, theglobal economy altered. is led to a material and productoversupply situation, one characterized by downward pricepressures. Crystalline silicon module prices have more thanhalved since 2007. Hennig Wicht, PhD, senior director andprincipal PV analyst or iSupply, expects to see an averagesales price o $2 per watt or c-Si PV modules in 2010, versusan average sales price o $1.40 or thin-flm PV. In some cases,c-Si PV modules (especially Asian multicrystalline products)destined or large projects are reportedly selling at pricesapproaching those or thin-flm modules.

As a result o this narrowing price gap and the dicultcredit market, many analysts believe that the thin-flm indus-try is acing signifcant consolidation. Signs o this are evidentin recent headlines, such as Applied Materials exit rom thethin-flm market. Price pressures have translated into greaterdemands on the perormance and durability o thin-flmproducts. Tight credit markets have created an emphasis onthe bankability o thin-flm PV products and manuacturers.

Some thin-flm sectors, particularly amorphous sili-con, ace increasing skepticism rom investors and integra-tors. is skepticism involves manuacturer credentials, aswell as product perormance and durability. In this article, Iaddress the latter concerns, especially in terms o thin-flmmodule deployment. I seek to answer the all-important ques-

tion: What can you as system designers do to improve theperormance and durability o systems that use thin-flm PV

modules? One valuable tool or designers in this regard isthe comprehensive thin-flm product specifcations table on

pages 4849. Granted, thin-flm technologies are quite dier-ent rom the c-Si PV products that you are more amiliar with.But as long as you are armed with meaningul and completedata about these modules, then you can deploy systems thatmeet or exceed perormance expectations.

What is thin-film PV?

In contrast to industry-standard monocrystalline or multi-crystalline silicon cells, which are currently produced com-mercially with a thickness o 150 to 200 microns or more,thin-flm PV technology deals with very thin layers o semi-conducting materials on the order o only 2 microns thick.e semiconducting materials used in thin-flm PV cellsrange rom the ubiquitous, like silicon, to the exotic, such astellurium or gallium. ese thin material layers are typicallydeposited, oten in multiple layers, on rigid or exible sub-strates made o metal, glass or polymers.

e promise o thin flms, which are sometime reerredto as second-generation solar cells, is their potential or costsavings. Signifcant raw material reductions, coupled with theuse o manuacturing processes well suited to mass produc-tion, should translate into material supply, energy use andenvironmental benefts, and ultimately lower prices. e lat-ter is illustrated by First Solars average cost o $0.76 per wattin Q2 2010 or modules having an average eciency o 11.2%.

In addition, improved energy payback times or thin flms arewell documented.

Four thin-flm technologies have reached the point ocommercialization on a large scale: amorphous silicon(a-Si); micromorphous silicon (a-Si/c-Si); cadmium tellu-ride (CdTe); and copper indium (di)selenide (CIS) or copperindium gallium (di)selenide (CIGS). Modules rom these ourgeneral technology groups are included in the companionproduct specifcations table.

Amorphous silicon. e frst a-Si cell was developed at RCALaboratories by David Carlson and Christopher Wronski inthe mid-1970s. e term amorphous reers to the random,

noncrystalline structure o the atoms making up these mate-rials. Compared to other thin-flm technologies, a-Si is rela-tively easy and low-cost to manuacture, but it is also the leastecient o the commercialized thin flms. Module ecienciestypically range between 5% and 7%.

O the many a-Si PV module manuacturers worldwide,United Solar, a subsidiary o Energy Conversion Devices,stands out or two main reasons. First, the companys roll-to-roll manuacturing process is unique. ree a-Si cell layers,each with a complementary spectral response, are depositedon a exible stainless steel substrate and encapsulated withina rugged polymer. ese triple-junction a-Si rolls are then cut

up or use in ramed UNI-SOLAR modules, C o n t i n u e d o n p a g e 3 8

Back contact

Top glass

Encapsulation

Top cell (a-Si:H)

Bottom cell (c-Si:H)

Transparent conductive

oxide (TCO) front contact

Bottom glass

Micromorphous silicon A typical tandem-junction a-Si/c-Si

thin-lm cell is shown here in cross section.


4/9


as well as or a variety o building-integrated and electronics-integrated PV products. Second, as Greentech Medias Mehtapoints out in his article e Future o in Film: Beyond theHype, going into 2010, United Solar was one o only two thin-flm manuacturers to produce in excess o 100 MW annually.

e other was market leader First Solar.Micromorphous silicon. An increasingly common way or

manuacturers to improve on a-Si eciencies is to depositan additional microcrystalline silicon (c-Si) absorber layer,resulting in a tandem-junction micromorphousa-Si/c-Si cell.is is still considered thin flm because once both layers aredeposited, the tandem-junction cell is still on the order o 2microns thick. e a-Si/c-Si cell is more ecient than ana-Si cell, owing to the complementary spectral response o themicrocystalline layer. Micromorphous PV products listed in thecompanion table are generally between 8% and 9% ecient.

Most industry analysts group amorphous and micromor-

phous silicon module manuacturers together, and manycompanies in this feld produce both types o modules. Repre-sentative manuacturers include Kaneka, Oerlikon and Sharp.Oerlikon oers complete production lines that the companyclaims will be capable o producing modules at a cost o $0.70per watt or less by year end.

Cadmium telluride. It was a CdTe cell, produced byresearchers at the University o South Florida, that in 1992frst broke the 15% eciency barrier or thin-flm PV. Com-mercial eciencies are currently in the 9% to 11% range orthese glass-on-glass modules. Only two CdTe manuacturerscurrently oer UL- and CEC-listed products: First Solar and

Abound Solar.

In spite o its large market share at present,the rise o CdTe in the thin-flm ranks has not

been without challenges. Early testing o CdTecells at the National Renewable Energy Labo-ratory (NREL) indicated lower than advertisedpower output, high degradation rates and vul-nerability to moisture. Furthermore, cadmiumis a heavy metal waste product rom zincand copper mining that is known to be toxic,resulting in justifable concerns about harm tothe environment and to human health. FirstSolar has proactively addressed all o theseissues and created a worldwide recycling pro-gram or its modules.

Copper indium (di)selenide and copper indium

gallium (di)selenide. CIS cells include thin lay-ers o copper, indium and selenium; CIGS cellsadd gallium to that mix. In the laboratory,CIGS cells produced by NREL have broken the20% eciency barrier. Commercial CIS/CIGSproducts generally have module eciencies inthe 9% to 11% range. Glass substrates are most

common; however, exible substrates are also being used orbuilding-integrated PV (BIPV) products.

ough ARCO Solar felded the frst CIS modules in themid-1980s, many researchers and analysts consider this tech-nology to be still in its inancy. Early CIS/CIGS cells tested by

NREL experienced higher degradation rates than other thinflms. (See Dirk Jordan and Sarah Kuntzs article on pages2428.) is may explain in part why the commercializationo CIS/CIGS has generally lagged behind that o other thin-flm technologies. is situation is rapidly changing, however,as a great deal o CIS/CIGS manuacturing capacity is nowbeing brought online. Notable manuacturers include DowSolar, Miasol, Nanosolar, Solar Frontier and Solyndra.

thin-film ProPerties and CharaCteristiCs

e accompanying specifcations table lists the propertiesand characteristics o thin-flm modules that are most rel-

evant or system designers and integrators. When workingwith these data, however, it is important to recognize andaccount or the ways that specifc thin flms dier in theirbehavior rom traditional c-Si PV modules. While in somecases these dierences cut across cell types, in others theyare unique to a technology or product.

Voltage characteristics. One o the frst things a designermay notice when working with thin-flm modules is thatthey may have undamentally dierent voltage character-istics than the designer is used to seeing in c-Si modules.Relatively speaking, low voltage and high current are char-acteristic o c-Si PV modules, whereas many thin-flm mod-

ules are just the opposite, exhibiting high voltage and low

Cadmium telluride The layers that make up a CdTe thin-lm cell from Abound

Solar are shown here. Front and back glass is typical.

Back electrode

Front glass

Encapsulant

Cadmium Sulfide (CdS)

Cadmium Teluride (CdTe)

Transparent conductive oxide (TCO)

Back glass

Thin-Fi lm PV


5/9


current. is dierence is signifcant when considering elec-trical string confgurations.

While c-Si PV source circuits operating at less than 600Vdc commonly consist o eight to 16 modules in series, sourcecircuits designed using representative thin-flm modulestypically have two to fve modules in series. In most cases,thin-flm modules also have relatively lower power ratings,which results in a proportional increase in the number omodules. e balance o the dc electrical equipment needsto support confgurations with shorter and more numerousstrings, as I discuss in more detail in the Balance o SystemRequirements section (pages 5054).

O particular interest to designers is the act that the ratiobetween Vmp and Voc or thin-flm modules is less than that

or c-Si PV modules. is is illustrated in Figure 1. For the thin-flm PV modules in the companion table, Vmp as a percentageo Voc generally ranges around 75%, whereas in the c-Si PVspecifcations table that SolarPro published in August/Sep-tember 2010, the average ratio is closer to 80%. is ratio issignifcant because as it decreases the voltage spread betweenVmp and Voc increases, suggesting that array-to-invertermatching becomes more dicult. Other actors, o course,are at worklower-voltage modules inherently provide moredesign exibility than higher-voltage modules, or example

so string-sizing options or thin-flm arraysare generally more limited than or c-Si PV.

Temperature coefcients. System design-ers are well aware o the inverse relation-ship between temperature and voltage asit relates to the perormance o PV materi-als. Lower voltages also produce less power.ese eects are quantifed using tempera-ture coecients, which are essentially mea-sures o the voltage drop in a diode as itstemperature increases. As a rule, tempera-ture coecients o voltage and power orthin flms are quite a bit less than those orconventional c-Si PV modules. Representa-tive temperature coecients by cell type are

illustrated in Figure 2. e lowest thin-flmtemperature coecients are seen in a-Simodules, whereas CIS/CIGS products havethe highest. e good news is that lowertemperature coecients generally beneftthe designer and may improve system per-ormance as well.

According to Charly Bray, vice presi-dent o project engineering and operationsat Sky Solar Group: Lower temperaturecoecients o voltage generally work tothe beneft o the designer because this

limits the operating voltage range o thesystem. While the spread between Vmpand Voc is larger or thin flms at STC, thelower temperature coecients o voltagetend to moderate the spread between themaximum system voltage and the mini-mum operating voltage in the feld. Whiletemperature coecients are assumed tobe constants, some manuacturers havedemonstrated that their products exhibita nonlinear dependence on tempera-ture. is may have the eect o lowering

the maximum system voltage rom that

-0.15%

-0.20%

-0.25%

-0.30%

-0.35%

-0.50%

-0.45%

-0.40%

a-Si a-Si/c-Si CdTe CIS/CIGS pc-Si mc-Si

%/C

Figure 1 This graph shows Vmp as a percentage of Voc averaged by cell tech-

nology. Comparative data for polycrystalline silicon (pc-Si) and monocrystalline

silicon (mc-Si) is included.

Figure 2 The average temperature coefcient of maximum power (Pmp) is

shown here according to cell technology. Lower values are benecial when cell

temperatures are greater than 25C.

60%

65%

70%

75%

80%

85%



6/9


calculated using the published coecients. is emphasizesthe importance o working closely with product applications

engineers, especially when working with a product or thefrst time.

In terms o system perormance, lower temperature coe-fcients suggest that thin-flm products tend to outperormc-Si modules at higher temperatures. Ratios o PTC to STC,or example, are higher or thin flms (93%94%) than orc-Si (87%90%). is is important because PV modules donot spend very much time at a single set o operating con-ditions, especially not at STC. In general, more energy is

harvested rom a PV system when cell temperatures exceed25C. is is one o the reasons that the annual energy yieldper rated watt (kWh/kWp) or a thin-flm array may exceedthat or a c-Si array.

Perormance transients. All PV technologies have some elec-tronic hysteresis, writes Ken Zweibel in a 1999 conerencepaper or the Electrochemical Society. e ormer programleader or NRELs in-Film PV Partnership goes on to explain:is means that their sunlight exposure and electronic his-tory inuence perormance. As an example, c-Si PV modulesundergo an initial light-induced degradation (LID) in the frstew hours o outdoor exposure. Because this LID is rapid and

specifcally aects short-circuit current, system designers need

not pay it any mind. e same cannot be said or thin-flm per-ormance transients.

A small amount o hydrogen, or example, is incorporatedinto the a-Si cell structure. e hydrogen atoms enhance theelectrical properties o the a-Si. But whenthese hydrogen atoms are pushed away romthe silicon, a process energized by the pres-ence o photons, deects occur in the atomicstructure and the perormance o the a-Sicell is gradually reduced. is is reerred toas the Staebler-Wronski eect, ater the RCALaboratories researchers who discovered it.How this is expressed in the feld is impor-tant or system designers and integrators tounderstand, as it may impact inverter selec-tion and source-circuit design.

Out o the box, amorphous and micro-morphous PV modules eectively exceedtheir nameplate power ratings. Upon expo-sure to light, their eective power gradu-ally decreases until it settles out within thepower tolerance o the STC rating. is pro-cess is reerred to as light soaking. e num-ber o junctions generally determines thelevel o decay due to light-soaking eects.Single-junction a-Si products, or example,can experience up to 30% decay; tandem-

junction products, including a-Si/c-Sicells, might experience as much as 20%decay; and decay or triple-junction prod-ucts is usually around 10%. Typical stabili-zation times or the Staebler-Wronski eect

are 6 to 16 weeks. is is a ully reversible eect. I you coverthe a-Si module or a period o time and then uncover it, youwill see the same gradual degradation all over again.

Subsequent long-term degradation in a-Si is not lin-ear, due to the seasonal annealingeect. At cooler times othe year, the Staebler-Wronski eect is relatively strongerand module eciency is relatively lower as a result; warmer

weather results in relatively improved eciencies. (is rela-tionship to temperature is counterintuitive since it is oppositeo the diode power-temperature relationship.) See Figure 3 oran idealized a-Si power degradation curve.

Dark soak is a perormance transient that aects CdTe,and CIS/CIGS thin-flm modules in particular. Ater mod-ules are removed rom their boxes and exposed to sunlight,their output has been observed to increase by as much as 6%.Stabilization can take up to a ew weeks on large systems.According to Rommel Nouf, PhD, principal scientist andthin-flm group lead at NREL: is phenomenon is not wellunderstood. It is due either to junctions being in a nonequi-

librium state when the cells are made and C o n t i n u e d o n p a g e 4 2

Powe

r(Pmp

)

1 2 3 54

Time (years)

0.8

0.9

1.0

1.1

~February

~August

Figure 3 Long-term degradation from nameplate-rated power, representedhere by the dashed red line, underlies two a-Si performance transients: an initial

power degradation due to the Staebler-Wronski effect and seasonal variation due

to thermal annealing.

Thin-Fi lm PV


7/9


packaged, or to diering electrical biases in darkness versusthe presence o light. Whatever the cause, dark soak shouldnot complicate system design or operation.

Fill actor. In Jim Dunlops textbook or the National JointApprenticeship and Training Committee, Photovoltaic Sys-tems, the termll factor(FF) is defned as the ratio o maxi-mum power to the product o the open-circuit voltage and the

short-circuit current.

FF = pm / (Vc x isc)

Fill actor not only describes a cells I-Vcurve but also its quality compared toan idealized cell.

System designers are amiliar withthe classic shape o the I-V curve ora c-Si PV module. While I-V curvesor thin-flm modules are similar, theyhave noticeably more rounded cor-

ners, meaning the maximum power isless pronounced. Figure 4 shows thepublished I-V curve or a UNI-SOLARtriple-junction a-Si module. e round-ing o the I-V curve is more pronouncedin a-Si than in CIS/CIGS or CdTe, butit is nevertheless present or all thin-flm technologies.

e maximum power or an ideal-ized PV cell is equal to the product oits Isc and Voc, resulting in a rectan-gular I-V characteristic and a fll actor

o unity (FF = 1.0). According to Roger

Messenger and Jerry Ventre, authors o Photovoltaic Sys-tems Engineering:e secret to maximizing the fll actor is

to maximize the ratio o photo-current to reverse currentwhile minimizing series resistance and maximizing shuntresistance within the cell. Series resistance, which manu-acturers seek to minimize, describes the cells internalresistance to current, resulting rom the semiconductormaterial itsel as well as rom metallic contacts and inter-connections. Shunt or parallel resistance, which manuac-turers seek to maximize, describes the devices resistance toleakage current.

In general, thin-flm modules have higher series resistancethan c-Si cells do, resulting in lower fll actors and eciencies.Shunt resistance can be aected more by module construc-tion than PV technology. ese resistances have improved(increased) in thin-flm modules as module construction hasimproved over the years. Representative fll actors by tech-nology are shown in Figure 5.

While it is not uncommon or inverter manuacturers totout their products exceptional ability to fnd the MPP ora thin-flm array, there is no evidence that system designersneed to make special accommodations or lower fll actors.e MPPT algorithms in modern inverters are both ast andhighly accurate. In terms o system perormance, low fll ac-tor may actually be advantageous because it translates intolower mismatch losses.

Most designers and project developers accept that c-Si

modules o diering power ratings should not share thesame inverter. e situation with thin-flm modules is moreambiguous, however. Lower fll actors and rounded I-Vcurves mean that thin-flm modules are relatively insensitive

50%

55%

60%

65%

70%

80%

75%


0 24 28 32 444036 480

40

20

50

60

30

10

4 8 12 16 20

Voltage (V)

Current(A)

STC (1000 W/m2)

Figure 4 This representative a-Si thin-lm I-V curve is

based on published data for UNI-SOLAR PV laminates.

One potential benet of a rounded I-V curve is increased

tolerance to mismatch.

Figure 5 The average I-V curve ll factor for different PV cell technologies is

shown here.

Thin-Fi lm PV


8/9


to nonoptimal currents and voltages. Some thin-flm man-uacturers seek to use this thin-flm characteristic to theiradvantage. For example, a Solyndra manual states, Panelso dierent power ratings do very well when wired in seriesand also do well when wired in parallel, but slightly less well.is undoubtedly makes some integrators and procurementmanagers very happy.

Shade tolerance. in-flm cells are usually very long and

narrow. While shading is always a problem or PV modules,it becomes especially problematic when one or more cellsare completely shaded and unable to deliver any current atall. With long, narrow thin-flm cells, the likelihood o total

cell shading is diminished, provided that the designer has cor-rectly oriented the module. e rule o thumb is to keep thelong dimension o the cells perpendicular to the ground. Inthis manner, narrow interrow shade bands that occur earlyor late in the day do not shade ull cells and thereby disableportions o the array. Similarly, i the module is ramed, thisorientation better resists the deleterious eects o dirt build-up at the rame. I present, bypass diodes help mitigate shad-

ing eects, but the best strategy is to avoid shading altogether.Module construction. in-flm cells are much more likely to

be sandwiched between two sheets o glass than are c-Si PVcells. In the case o Solyndra products, the CIGS cells are sand-

wiched between two glass cylinders. Typ-ically, these glass-on-glass modules donot have a metal rame. Eliminating therame, usually made o anodized alumi-num, not only reduces cost and embed-ded energy, but it also helps with theelectrical isolation o the module. Havingmetal in contact with glass can provide

an unwanted conduction path or leak-age currents. Frameless modules, alsocalled laminates, do lack the structuraledge support and edge impact protec-tion that rames provide. For design pur-poses, this may have consequences orthe type o racking required. It defnitelyhas implications regarding the modulehandling and installation techniquesthat are employed.

Module efciency. Currently, c-Simodule eciencies run between 13%

and 19% and will likely edge up 2% to

Recommended

Shade

Module orientation Consult

the manufacturers design

and installation manual forrecommended module orien-

tation. According to Sharps

thin-lm PV system design

book, for example, permanent

damage to the product can

result when cells are oriented

horizontally and shaded by

dust, dirt or snow.

Figure 6 Average module efciency for UL- and CEC-listed thin-lm products are

shown here, based on data from SolarPro product specication tables, grouped

according to cell type.

0%

2%

4%

6%

8%

16%

14%

12%

10%



9/9


3% over the next 5 to 10 years. in-flm PV module ecien-cies are currently running between 6% and 12%, as shown in

Figure 6 (p. 43) and and could close the eciency gap withpolycrystalline silicon over the next several years. Accord-ing to NRELs Nouf: CIGS has a technology path to achievecomparable eciencies to polycrystalline silicon in the next 5to 10 years, given the current levels o R&D unding, and per-haps sooner i unding increases. With urther R&D in CdTetechnology, specifcally relating to substrate confguration,CdTe eciencies could also approach those o polycrystallinesilicon. In act, the theoretical eciencies o CdTe are slightlyhigher than that or CIGS.

While STC eciency is an important cell technologyparameterit provides a relative measure o array ootprintit is not all-important. Eciency at non-STC irradiances, lightspectra, temperatures and energy yield per rated watt are alsovery important parameters to consider.

In contrast to most c-Si modules, thin-flm eciencies

may be higher at lower irradiance levels than at the 1,000W/m2 standard test condition. A c-Si module with low seriesresistance does not see as much I2R loss between low and highoutput and consequently has less eciency change. A thin-flm module with high series resistance, however, has a largereciency dierence between low and high output. Becauseall modules are rated at the high-output end o the eciencycurve, the relative change in eciency or a thin-flm moduleis greater than or a c-Si module when moving rom high tolow output on the curve. is bonus yield over c-Si could beas high as 10% or some thin-flm technologies, most nota-bly a-Si/c-Si and CdTe. Defnitive studies demonstrating

this phenomenon are dicult to fnd, however. In addition,

because low irradiance levels oten involve changes in lightspectrum, it is dicult to separate the two.

Spectral response.All photovoltaic materials absorb andconvert light more eciently at some wavelengths and less

eciently at others. One advantage o thin-flm PV materialsis that they can be stacked so that the aggregate o these di-erentially responding layers is a cell that perorms well over abroader light spectrum. A good example o this is UNI-SOLARstriple-junction a-Si product that has three complementaryabsorber layers to capture blue, green and red portions o thevisible light spectrum. Amorphous silicon and microcrystallinesilicon can also be stacked. With its heterojunction with intrin-

sic thin-layer (HIT) cells, Sanyo evenlayers amorphous silicon with a mono-crystalline silicon cell.

e spectral content o sunlightchanges with time o day and varyingatmospheric conditions, such as cloudsand pollutant levels. is is why test-ing standards use identifers like AM1.5, which reers to an air mass that is1.5 times thicker than when the sun isdirectly overhead. in-flm cells havesomewhat better perormance in diuselight than do c-Si cells due to somewhatimproved spectral response in the yellowand red wavelengths o light (600800

nm) in combination with improved e-ciency at low irradiance. When optimiz-ing PV plant perormance, the best wayor system designers to account or thisphenomenon, and other complex thin-flm perormance variables, is through

the use o perormance modeling tools (see sidebar p. 46).

inVerter ComPatibility Considerations

While the design process is largely the same, there are impor-tant dierences to consider when matching a thin-flm arrayto an inverter: lower temperature coecients, transient out-

put eects, diering voltage characteristics and the poten-tial or harmul ion migration internal to the module. esedierences impact source-circuit design, inverter sizing andinverter selection.

Source-circuit sizing.Voltage issues at the high (Voc) andlow (Vmp) ends o the inverter input voltage window need tobe looked at careully when dealing with thin-flm technolo-gies. is process is not unique to thin flms. According toSky Solar Groups Bray, a proessional engineer and industryveteran: It is advisable in designing any PV system, whetherthin flm or crystalline, to look at the ull operating range othe module in terms o temperature to make sure that the sys-

tem will be staying within the maximum C o n t i n u e d o n p a g e 4 6

0 600 800700 900

0

40

20

50

30

10

300 400 500

Wavelength in mn

Relativeintensity Triple junction

technology

blue-absorbing

cell

green-absorbing

cell

red-absorbing

cell

Spectrum AM 1.5

Based

on

united

Solardata

Figure 7 The UNI-SOLAR triple-junction a-Si cell is an aggregate of three distinct

absorber layers, each with a unique and complementary spectral response.

Thin-Fi lm PV

thin-film pv - a system designer's guide - solarpro

Documents