Solar Photovoltaics Overview 2005- J. El Spec Rel Ph 2006- p 105- Kazmerski

Journal of Electron Spectroscopy and Related Phenomena 150 (2006) 105–135

Solar photovoltaics R&D at the tipping point:A 2005 technology overview

Lawrence L. Kazmerski∗National Center for Photovoltaics, National Renewable Energy Laboratory, Golden, CO 80401, USA

Abstract

The status of current and coming solar photovoltaic technologies and their future development are presented. The emphasis is on R&D advancesand cell and module performances, with indications of the limitations and strengths of crystalline (Si and GaAs) and thin film (a-Si:H, Si,Cu(In,Ga)(Se,S)2, CdTe). The contributions and technological pathways fornow and near-term technologies (silicon, III–Vs, and thin films) andstatus and forecasts for next-next generation photovoltaics (organics, nanotechnologies, multi-multiple junctions) are evaluated. Recent advancesin concentrators, new directions for thin films, and materials/device technology issues are discussed in terms of technology evolution and progress.Insights to technical and other investments needed totip photovoltaics to its next level of contribution as a significant clean-energy partner in theworld energy portfolio.©

K

1

aCrrtegrdnoasptiDbi

livesullerallysts ofastnted inriti-to

our

annual

ually35%

ndes foresultany

iciesmar-

sidies,

0d

2005 Published by Elsevier B.V.

eywords: Photovoltaics; Solar cells; Thin films

. Introduction and progress

Just over 50 years ago, this solar–electric technology markedtipping point [1] at Bell Telephone Laboratories when Darylhapin, Gerald Pearson, and Calvin Fuller suddenly turned a

esearch curiosity into a viable electricity producer[2]. Theiresearch innovation brought the performances of these crys-alline silicon devices from “laboratory interest” (conversionfficiencies hovering at 1%) to efficiencies five to eight timesreater, earning consideration of these solar-powered devices aseal electrical power sources. Although this threesome worked toevelop a practical power supply for Bell’s remote telephone sig-al transmissions here on earth, they actually created the technol-gy thatfirst blossomed to power our early satellites—leading torevolution in wireless communications that was not yet envi-

ioned within their own forward-looking communications com-any. They were creative scientists and engineersahead of their

ime. The revolution in the terrestrial markets they were address-ng was delayed until at the end of their 20th century—much asaryl Chapin himself had contemplated[3]. First, there woulde a demonstration of technology readiness with remote power

to help electrify our homes and many aspects of our daily[4,5]. As this 21st century is starting, Chapin, Pearson, and Fwould be pleased to finally see this PV market growth finreaching respectable levels that are capturing the intereinvestors and users[4–6]. The research progress over the p25–30 years has been substantial and steady, as represeFig. 1. Photovoltaics is poised at what may be its most ccal tipping point[1]; the one that will cause this technology“spread like wildfire” as it finally becomes a major part ofworld’s energy portfolio.

PV as a technology and a business has just surpassedsales of 1 GW and U.S.$ 10B. As represented inFig. 2, theseworldwide shipments have been growing above 30% annfor the past decade; in fact, they have averaged abovefor the past 5 years[7–9]. PV is a real business now—ashould continue to exhibit such substantial annual increassome time to come. Much of this growth has been the rof government incentives, mainly in Japan and in Germ[5,6,10,11]. Both these governments have shown that polmake a difference—using quite different approaches. Theket stimulation in Japan has been based on cash sub

n the 1980s[4]. Then, in the late 1990s, this “PV” would emerge

∗ Tel.: +1 866 270 2962.E-mail address: larry [email protected].

initially buying down the price to the consumer. Starting in 1994with a 50% rebate, this program has followed its design to grad-ually phase down the government portion as the price for thePV system decreased. This coming year, the programs successis indicated by the more than 140,000 installations and reaching

368-2048/$ – see front matter © 2005 Published by Elsevier B.V.oi:10.1016/j.elspec.2005.09.004

106 L.L. Kazmerski / Journal of Electron Spectroscopy and Related Phenomena 150 (2006) 105–135

Fig. 1. Efficiency evolution of best research cells by technology type. This table identifies those cells that have been measured under standard conditions andconfirmed at one of the world’s accepted centers for standard solar-cell measurements.

the point that no subsidies are needed in their electricity pricemarket. On the other hand, Germany introduced a “feed-in tar-iff” in 2000 that offers solar PV users a guaranteed “Euro perkWh” over a 20-year period, with each year the guaranteed priceis reduced by 5%. With the availability of low-interest loans,the German markets have heated from less than 20 MW/yearto above 150 MW/year currently. The cost is spread over theentire electricity user rate base so that utilities are not negatively

impacted and the government does not have to appropriate thefunds annually.

The successes of these policies sometimes overshadowanother important component—technology advancement. The18–22 U.S. cents/kWh that has been reached also required aprogression of substantial and creative R&D improvement inmaterials, devices, fabrication, characterization, and processing,leading better device performance and reliability, and lowered

F
ig. 2. Photovoltaic annual module shipments as reported by the industry. Sh ipments for U.S., Japan, Europe, and rest-of-world sectors have been identified [7].

L.L. Kazmerski / Journal of Electron Spectroscopy and Related Phenomena 150 (2006) 105–135 107

systems costs that the “policies” have leveraged. This electric-ity price breaches into some electricity markets[11]. But it isstill too high for the next wave of grid-tied applications (con-sumer side of the meter prices) and almost an order or magnitudetoo high for wholesale (central utility) generation. To tip thistechnology to its next level—building first multi-GW marketstoward the terawatt levels, and manufacturing plants to hun-dreds of megawatts then perhaps the GW annual capacities, PVtechnology requires evenmore creativity, science, and engineer-ing to meet the growing and diversified technical and consumerdemands.

This paper looks at the current PV technology sta-tus, the reasons underlying recent improvements, compar-isons of approaches, with an emphasis on R&D needs anddirections—and a bit of a look forward what future generationPV might encompass. There is a focus on the fundamental build-ing blocks of the PV system—the cells and modules. The crys-talline silicon, crystalline III–V, and polycrystalline and amor-phous thin films that dominantnow and near-term expectationswill receive the primary attention. However, thenext generationsof PV are already on the laboratory research benches—organics,nanotechnologies, multi-multiple junctions, bandgap engineer-ing, thermo-tuned concepts—all aimed either the ultra-high effi-ciency or the ultra-low cost regimes that the technologies of ourquarter- to mid-century will likely require. Because even nowour R&D is both undergoing and causing rapid and continuousc y ot meni cea nget

2. The technologies

Photovoltaics technology includes a number of significantcomponent performance “gaps” for various crystalline, poly-crystalline, and amorphous; bulk, as well as thin-film technolo-gies. Thefirst is the difference between the theoretical limits, theattainable levels, and what has been demonstrated under the bestconditions in the laboratory (the headline or record cells). Theseare shown inFig. 1for various crystalline, polycrystalline, bulk,and thin-film technologies[12]. Underlying these differences arelosses that are inherent to the conversion process (theoretical toattainable), and the ability to fabricate the cell with the ensem-ble of optimal, interrelated properties, and parameters. The “gapbetween what can been attained and what has been reached” is amajor focus for researchers (represented inFig. 3) – a process ofidentifying, understanding, and minimizing losses – collectingevery incident photon, allowing these to create the maximumnumber of electron–hole pairs, and then making these carrierslive long enough to contribute to the current generation process.The second is the difference between the laboratory efficiencyof cells and those produced in commercial lines. This has to dowith scale up of the processing to larger areas, variations of mate-rials (starting wafers, substrates, coatings, etc.), less controlledconditions, and higher required throughputs. Thethird gap isthat between the cell efficiencies and those of the modules. Thisdepends upon the ability to minimize the losses when wiringt dulet ans-m nedb ” areo nd are

n the

hange, this review can only present “snapshots” of todaomorrow’s technologies. However, the attention and investn this R&D is critical to realizing the new levels of performannd reduced-costs needs, both for the now and for the lo

erm success of solar photovoltaics.

Fig. 3. Performance gaps between best device efficiencies i

rt

r-

he cells into circuits, bringing the active area of the moo be closer to the cell area, and maximizing the optical trission of the protective or support layers that are positioetween the cells and the incident sunlight. These “gapsnes that can and have to be addressed and minimized—a

laboratory and attainable efficiencies for several solar cell technologies.


active areas of research and development for all PV technologies[3–5].

2.1. Crystalline silicon

Of the more than 1 GW of the PV commercial shipmentsin 2004, a dominant 94% were single-crystal, multicrystalline,ribbon, and sheet silicon[7]. In fact, the markets would haveabsorbed even more except that the supplies were limited dueto insufficient manufacturing capacities and supplies[13,14].Although these capacities are expected to more than doublein the next 3 years (beyond a 25% per year growth), this overdemand for the “semiconductor foundation” of the photovoltaicsindustry will likely lag behind the marketplace—as long as theincentives in Europe grow, and even more so, if those in the U.S.are implemented. The spectacle of “zero-inventory warehouses”is one that the industry relishes, but under which consumers andthe environment suffer.

How does this Si cell differ in 2005 from that “invented” atBell Labs a half century ago? Indeed, there has been incred-ible progress in the understanding of the semiconductor andthe device. (This “understanding” has been a major advan-tage for Si cells since it has leveraged the R&D accomplishedfor other more prevalent electro-optical technologies that havebecome parts of our everyday lives.) It is also interesting tosee how the fundamental understanding of that 1950s teame ll# PVmc allowd ep-a t arem att . Byc un-P nf pingp

ft inc atoryp pho-t rted“ linec ple,c lvedt imeda holeg o bec -t -type(s SBC)[a me 10y iesc hasn parts

of the world, particularly in Europe and Japan, where it has beenrecognized that there are significant improvements for both cur-rent commercial approaches—and especially for a potential nextgeneration of Si solar technologies[23]. Moreover, the “26% cal-culated limit” for crystalline Si technologies has recently beenre-evaluated—with the result that the “bar” has now been raised.

Fig. 4. Historical Bell Telephone Laboratoriescrystalline Si cell reported in1954. (a) Bell cells and first solar module photograph and (b) device represen-tation from Bell Laboratory (from page 68 of laboratory notebook of C. Fuller,witnessed by D. Chapin 1953).

ndures today.Fig. 4a shows a photo of the April, 1954 “Be1”, measured at about 6% efficiency along with the firstodule, deployed for use in late 1954[2]; Fig. 4b shows the

ross-section of that device. Its design has a relatively shiffused junction (certainly the research tipping point to srate it from previous grown junctions), and contacts thaade from theback of the cell—minimizing any obscuration

he front surface in order to accept every incident photonomparison,Fig. 5a show the high-efficiency design of the Sower cell[15]—strikingly similar in these two critical desig

eatures that brought about an early PV technology “tipoint”.

Some have had a misconception that there is little lerystalline Si research. After all, cells have reached laborerformances converting nearly one-fourth of the incident

ons into electrical power—reaching about 92% of its repotheoretical limit”. (Table 1presents a summary of the headells in various Si device technologies.) The relatively simonventional p–n junctions of the early 1980s have evooward more complicated designs and structures—all at capturing every incident phonon, maximizing electron–eneration, and prolonging the lifetime of those carriers tollected for maximizing current generation[16]. The evoluion of these designs has included metal/insulator/n-type/pMINP) [17], passivated-emitter solar cells (PESC)[18], single-ided and doubled-sided buried contacts (SSBC and D19,20], point contact[21], and bifacial cells[22]. It wouldppear that the single-crystal research phase plateaued soears ago (Fig. 1), leaving an impression that other technologan only gain on this frontrunner. This limited viewpointot clouded the research thinking and strategies of some


Fig. 4. (Continued ).

Swanson and others have refined the calculations based on thegreater understand and materials and device knowledge that hasevolved over the last decade. This limit is now proposed to benear “29%” under standard conditions[24–26]. This not only hasreinvigorated a community of researchers to find how the crys-talline Si device can enter into this expanded efficiency regimeof 26–29%. It is also a partial basis of a new “Si TechnologyInitiative” proposed in the U.S.[23]—and such cell improve-ments may recast the thinking whether silicon can compete at thesubstantially sub-$1/W system price that other emerging tech-nologies have taken as their exclusive real estate value in thelonger term. Silicon is real business—and it is also a focus forreal research excitement again.

It must also be recognized that there are other importantprogress metrics. For example, there is a real and current eco-nomic benchmark for maximum “W/m2” for many rooftopapplications, adding value to higher efficiency, manufacturing-friendly approaches. This has partially driven the developmentof the three major commercial cells having efficiencies in excessof 20%, as represented inFig. 5. Additionally, “performance” is

a marriage between the conversion parameters and the reliabilitycriteria. These devices must maintain their high power outputsfor decades—meeting warranties that are now 20–25 years, with30 years and beyond expected to be the industry standard in thenot too distant future. Stability, durability, and operating life-time are equal partners in the economic viability (energy andcost recovery) and consumer acceptance (electrical power ondemand) of our PV technology, Silicon leads right now in thesecombined categories—and has potential for additional improve-ment. It also establishes valuable targets and proofs of conceptfor the other PV approaches as systems gain from installationexperience and use.

The terrestrial commercial “20% club” represented by thecells in Fig. 5 show three different approaches that are lead-ing toward the 20% crystalline Si module. The SunPower cell(Fig. 5a) has been discussed briefly above[27]. It is engi-neered to contact from the rear side, with no light obscurationon the front surface—a design developed by R. Swanson inthe 1990s (first termed the “point contact” cell). This utilizedinterdigitated n+ and p+ diffusions and grid lines to collect the


Fig. 5. Commercial crystalline Si solar cell “20% club” showing cross-sections of: (a) SunPower A-300, (b) BP Solar Saturn, and (c) Sanyo “HIT” solar cells.

photogenerated currents and low series resistance loss. Cells(125 mm× 125 mm) have been verified with efficiencies to21.5%. The design incorporates a great deal of engineering, withsome special attention to light trapping. The cell production anddevelopment is based upon sophisticated statistical process con-trol and design of experiments—quality control leveraged fromtheir parent Cypress Semiconductor to greatly tighten processspecification limits. SunPower has begun to market modulesbased on these cells that have high packing density for the cells,and give the illusion of being uniformly black—designed foraesthetic enhancements for the built environment. Variationsof this cell design are also being used in concentrating PV(100×–300×) with peak efficiencies of 25%.Fig. 5b shows theBP Solar buried contact cell[28], of the design by Martin Greencited above. It uses advanced light trapping and minimizing thefront contact “area” by “burying” the contacts in laser-producedgrooves that lower the resistance through enhanced sidewallareas. This design has yielded production line cells with efficien-cies to 20.5%. The third cell is the Sanyo heterojunction withintrinsic thin layer (HIT) device (Fig. 5c) [29]. This employs ann-type single-crystal wafer surrounded by ultra-thin a-Si:H filmsin a bifacial structure. Headline cells at 20.1% have been mea-sured. Module efficiencies exceeding 16% conversion efficiencyare now commercially available using this innovative technol-ogy.

A typical breakdown of costs associated with the makingo[ Si

cell is associated with producing the high-perfection wafer. Oneapproach has been to employ less energy intensive processesto make the wafer; sacrificing efficiency points for the bene-fits of lower energy production and perhaps facilitating the useof lower-purity feedstock material. Casting (and several sib-ling technologies) has become the manufacturing convention[31–33]. The bulk Si methods have greatly benefited from theevolution of sawing techniques that utilize multiple thin wiresthat waste less of the precious semiconductor. The eliminationof this cutting process affords potential materials and produc-tion savings—the logic behind the development sheet and ribbonapproaches over the past 20 years. Two major ones that have cur-rent commercial significance are edge film-fed growth (EFG)[34] and the string ribbon processes[35]. The former methodinvolves shaping of the Si through a special die, forming a long(tens of meters) hollow octagon (although other shapes, evencylinders have been successfully produced) structures (meter-size in diameter). A laser is used to separate the flat faces intoindividual cells. In thelatter technique, two high-temperature“strings” are pulled vertically through a shallow molten Si melt,which crystallizes between the stings. This process has the pos-itive feature of being continuous—the melt can be replenished,the strings are fed in from spools, and the emerging Si can beto desired lengths without interrupting growth. Athird sheetmethod involves the growth of the Si layer by a high-speed pro-cess on a variety of substrates, including ceramics)[36,37]. Thec thosem r of

f a commercial single-crystal Si cell is represented inFig. 630]. A major consideration with the cost of producing the

ell efficiencies are similar to these sheet processes toade from the ingot-cut multicrystalline wafers—of the orde


Table 1Summary of confirmed, selected Si solar-cell efficiencies and related parameters, under standard measurement and reporting conditions[194]

Voc (mV) Jsc (mA/cm2) FF (%) Area (cm2) Efficiency (%) Organization Comments

Monocrystalline silicon706 42.2 82.8 4.00 24.7 UNSW PERL (3/99)702 41.6 80.3 22.10 23.4 UNSW PERL (5/96)704 41.5 81.0 22.1 23.7 UNSW PERL (8/96)678 39.5 80.3 148.9 21.5 SunPower Rear contact, FZ Si, large area (9/03)714 37.6 78.1 100 21.0 Sanyo HIT large area (3/01)625 36.3 80.6 147.5 18.3 BP Solar Laser Grooved, large area (9/02)634 36.3 81.6 4.02 20.5 Spire Conv. p–n (10/85)600 31.6 80.7 1.02 15.4 Westinghouse Dendritic Web (5/85) [28◦C]

Multicrystalline silicon664 37.7 80.9 1.002 20.3 FhG-ISE (5/04)632 35.9 77.7 144 17.6 Univ. of Konstanz, Photowatt,

and BP SolarMechanically textured, large area (2/03)

654 38.1 79.5 1.09 19.8 UNSW/Euro-solare (2/98)636 36.5 80.4 1.00 18.6 Georgia Tech HEM Si (12/95)610 36.4 77.7 100.00 17.2 Sharp Textured mech. (3/93)

Thin silicon (on thick low-cost Si substrate)608 33.5 81.5 0.98 16.6 AstroPower Silicon-film (700�m substrate) (2/97)602 10 15.3 Sharp Silicon-film (400�m substrate (5/05)

Thin-film poly-, micro-, and nano-crystalline silicon (<100�m)699 37.9 81.1 4.04 21.5 UNSW 47�m Si (8/95)668 37.1 77.5 4.00 19.2 FhG-ISE 46�m on SiOx (4/97)651 32.6 81.4 4.00 17.0 ANU 20�m (thinned Si) (9/94)539 24.4 76.8 1.12 10.1 Kaneka 2�m on glass (12/97)645 32.8 78.2 4.017 16.6 University of Stuttgart 45�m thin film transfer (7/01)589 35.6 76.3 95.8 16.0 Mitsubishi 77�m VEST (2/97)661 32.8 81.4 4.00 17.6 UNSW CVD Si on Cz Si (5/95)25000 3.26 68.0 661 8.2 Pacific Solar PECVD on glass-submodule; 1–2�m Si (5/95)

Concentrator cells0.15 26.5 Stanford U. Point contact: 140 suns (5/87)

20.00 21.6 UNSW Laser groove: 11 suns (9/90)1.60 26.8 SunPower Rear contact: 96 suns (10/95)1.21 15.7 SunPower Rear contact: 74 suns (7/93)

11–13% (although smaller grained sized sheets have somewhatlower performance). The best bulk multicrystalline Si cell wasverified in this past year by the Fraunhofer Institute at 20.1%[38], with several other cells in this materials category above18% efficiency[39,40].

Closing the differences between these less perfect materi-als and the single-crystal cell has involved considerable R&D.This has included understanding the nature of the defects result-ing from the growth and processing process—and then devisingsurface and materials treatments to minimize the carrier lossesat unwanted active sites. These have included processes involv-ing hydrogen, lithium, aluminum arsenic, and phosphorus forfront and back surfaces[41–43]. The development of plasmaand other nitriding processes by the semiconductor electronicsindustry has led to improvements in single- and polycrystallineSi cells. Additionally, the economics of using less material hasled the industry to investigate the production and use of thinnerwafers—with objectives of trimming the industry standard ofabout 300�m thick wafers by a third or more. This has fueledresearch on maintaining efficiencies for thinner layers. The flex-ibility of these wafers and devices, such as shown inFig. 7, havesome very desirable application features, but present consider-

able difficulties in handling and processing because of mechan-ical fragility for wafers of 100�m thickness and extremes intemperature cycling and thermal stresses they must endure dur-ing processing[44].

A number of thin or thinned wafer technologies have emergedin this new century, including an interesting concept termed theSliver® cell [45], developed by the Australian National Uni-versity. Silicon strips are produced by micro-machining narrow∼100�m deep grooves into a Si wafer. The cell processing(P and B diffusions, texturing the faces, passivation, metal-lization) is completed while the Si strips arestill supportedby the Si substrate at the bottom edge. The cells are thenremoved from the wafer support—providing narrow (1 mm),long (50–100 mm), thin (50–70�m) (thus, “slivers”) bifacialcells. These have had efficiencies up to about 18%, with proto-type minimodules (560 cm2 area) above 12% efficiency. Thecells are being considered for transparent and concentratingmodules, in addition conventional module use. The inven-tors argue for not only the materials savings these cells offer,but also that the manufacturing process is greatly simplifiedand that the required manufacturing plant requires less capitalinvestment.


Fig. 6. Cost and energy breakdowns structures for crystalline Si technology: (a) PV system, (b) Si module, and (c) energy required to make Si module components.Module finishing includes encapsulation materials, glass, frame, contacting/wiring, junction boxes, etc.

Ongoing research in crystalline Si aims to improve thethroughput and yield of the growth processes. This is particularlytrue of the developing ribbon and sheet processes). The roles ofdefects and impurities – and their interactions and impacts onsolar cell performance – remain important. Low-cost processingand manufacturing for high-efficiency technologies is a majorportion of efforts aimed at improving and simplifying all aspectsof cell and module production. Incremental improvements willcontinue to help reduce these manufacturing costs to keep sil-icon as the major part of the PV business for the near future.However, some revolutionary developments in manufacturingor cell innovation could accelerate industry expectations signifi-cantly. Finally, there remain questions about the resources for Sifeedstock[4,13]. The surge in markets combined with a seriesof falls in the supplies needed for the semiconductor electron-ics manufacturers have elevated PV to be the major customer.However, this has caused the price for the feedstock to fall to theU.S.$ 30–35/kg level that the solar industry can afford. However,as the competing microelectronics and optoelectronics businessagain build – an industry that can pay far more for these sup-plies – the question of whether Si PV can continue to follow90% learning curves1 that it has followed since the late 1990s

1 Aside: There were always termed “cadmium sulfide solar cells”. But in fact,both the major PV activityand especially the problems were with the otherh ss frob

for its product. There is some discussion over these feedstockissues. These range from evaluations that shortages are tem-porary to some dire arguments that additional investments areneedednow to build plants to specifically supply this growingPV need.

2.2. Thin films

Thin-film photovoltaics is always looked at as the “youngercousin” of the silicon technology—poised to take over theenergy production responsibilities of its older relative, but neverquite fulfilling its expectations or potential. In fact, the thin-film “efficient” counterpart of the Bell Labs Si cell surfacedwithin a year of the Bell announcement. These first “5%” cellswere cuprous sulfide/cadmium sulfide (Cu2S/CdS)1, growingout of work by Reynolds et al. in 1955[46]. The argumentsfavoring thin-film PV have been based on materials utiliza-tion, large-scale manufacturing advantages, and better energyeconomy for production and product recovery. The first 25years of thin films were dominated by that “cadmium sulfide”device—until its stability problems were judged to be insolvable.Now the major approaches include amorphous silicon:hydrogen(a-Si:H), copper–indium–gallium–selenide (CIGS) and relatedcompounds, cadmium telluride (CdTe), and silicon. Most of thework on these materials surged with the interests in terrestrialp waso

eterojunction partner. So, the CdS perhaps received some early bad preeing in the right place at the wrong time!.

mhotovoltaics in the early- to mid-1970s—although CdTef interest to earlier space PV work along with the Cu2S/CdS


Fig. 7. Flexible cells: (a) thin-layer Si sliver cell and (b) thin-film polymer cell.

Technology concerns and issues

• Manufacturing (costs, yields, complexities;quality assurance; diagnostics; robotics).

• Manufacturing capacity (meeting mar-ket needs/growth; plant scales toward500 MW/year, investments).

• Silicon feedstock (critical issue for the near-term: availability, cost, competition from semi-conductor electronics industry; need for solargrade?).

• Current research (thin wafer production, pro-cessing, and handling; manufacturing, process-ing, process integration; materials optimization;diagnostics).

• R&D initiative (new efficiency regimes, novelapproaches, new devices—next generations ofSi technology).

device owing to requirements for high power-to-weight ratiosthat still exist today in those important markets. Thin films haveadvanced significantly over these past 30 years (Figs. 1 and 3),and single-junction polycrystalline research cells are approach-ing 20%. More importantly, these technologies are now enteringthe world commercial markets, in demand partially from thecapacity problems with crystalline Si. These thin-film tech-nologies have more than doubled their sales over the previousyear—and have a window of opportunity to further prove theirmerit, worth, and competitiveness.

2.2.1. Thin-film silicon: amorphousThe introduction of this new class of semiconductors in the

mid-1970s seemed to have posed the ideal photovoltaic candi-date absorber[47]. Having no long-range and perhaps only lim-ited short-range order, its physics was completely different thanthe crystalline Si model. Because of the defects associated withthe “dangling Si bonds”, the amorphous Si was hydrogenatedto reduce the bandgap states and to allow the development ofopen-circuit voltages. Its bandgap could be varied over tenth ofeVs by changing the hydrogen content. Its light optical charac-teristics make it 100 times more effective in absorbing the sun’sirradiance than crystalline Si. It also benefited technologicallybecause it leveraged the R&D interests from other electronictechnologies (transistors, flat-panel displays).Figs. 1 and 8showthe evolution of the a-Si:H cell’s performance, with a summaryo

PVt tanti hisd onkiel un-d rablec nifi-c lema ngt s, thei sses( ices ndm s aren o thisp ding,d duc-t linea

timeso at isbt oper-a erciala rys-t g., ato sb reas.( H/Si

f champion devices inTable 2.The development and larger-scale adoption of this

echnology has been impaired primarily by single, imporssue—stability. Early in the research investigation of tevice, an inherent light instability, (named the Staebler-Wrffect after its discoverers) was observed[49]. Devices could

ose 50% or more of their power output over the first hreds of hours of its operation exposed to light—not a desiharacteristic for a solar cell! Though there has been sigant investment in both identifying the origin(s) of this probnd solving it, this technical limitation still exists. Followi

he guidance provided by several research investigationnstability has been minimized by engineering layer thickneprimarily the intrinsic layer) and by the use of multiple devtructures[50–52]. The “cure” has not been found, but cells aodules with less than 10% change in output characteristicow attainable. Research groups continue to give attention troblem—with several recent new paths toward understanepositing the material, and further stabilizing the semicon

or [52–55]. This includes combining or using nanocrystalnd/or microcrystalline Si in the device structure[56–58].

There is a benefit to the a-Si:H technologies that is someverlooked—a temperature performance characteristic thetter than its other semiconductor competitors[59]. Because

he performance tends to recover and improve at higherting temperatures and the bandgap is higher, the comm-Si:H module might actually have efficiencies better than c

alline Si counterparts under some specific conditions (e.perating temperatures above 70◦C). This characteristic haeen exploited for applications for tropical and desert aThis “temperature benefit” has also been cited for the a-Si:


Fig. 8. Thin-film amorphous-Si:H solar cells, showing progression of performances for small-area and large-area cells.

heterojunction in the last section.) To date, the major applicationof this technology has been in successful integration into roof-ing products[60]—with these products in the 5–7% efficiencyrange typically. These have 20-year warranties—as well as aes-thetic benefits of direct integration into the built environment. Itis the protective architectural roofing material—that also gener-ates useful electricity.

2.2.2. Nanocrystalline and microcrystalline silicon-films;R&D at the edge

In the mid-1990s, many research groups started to look at thefirst stages of crystallization of their a-Si:H films into nanocrys-talline (nc) and microcrystalline (�c) regimes. The use of theselonger-range order films “at the edge” of the ordering processwere deemed to provide the path toward more stable and higherperformance devices. Some tagged this as the evolution of theamorphous technology toward thin-film crystalline silicon.

The first progress was the introduction of the “micromorph”solar cell (Fig. 9, showing “generic” and cells grown on glassand stainless steel), a combined a-Si:H/�c-Si:H into a stacked

structure[61]. Initial cells with 7.7% efficiency were reportedfor this arrangement—and a-SiH/�c-SiH tandem cell with 10%efficiency (both stabilized). The micromorph cell was furtherimproved with the introduction of a ZnO layer as an intermedi-ate reflector[62]. A large research effort on the microcrystallineand nanocrystalline films exists today[63–68]. The micromorphconcept has progressed to commercial reality. Kaneka offersa number of products, ranging from rooftop to semitranspar-ent designs for building integration[69]. Recently, they havereported cells in the 13% range and a module with an 11.8%stabilized (aperture-area) efficiency (seeTable 2) [70].


• Stability, reliability, durability.• Perceptions of device lifetime issues; motiva-

tion for new “players” in PV industry.

Table 2Summary of confirmed, selected a-SiH based solar-cell efficiencies and related parameters, under standard measurement and reporting conditions[194]


Single-junction cells

D

T6)8)96)

887 19.4 74.1 1.00 12.7897 18.8 70.1 1.08 11.5886 17.46 70.4 0.99 10.9

ual-junction cells1621 11.72 65.8 0.28 12.51685 9.03 68.1 0.76 10.3

riple-junction cells2375 7.72 74.4 0.27 13.52541 6.96 70 0.27 12.42289 7.9 68.5 1.00 12.4

Sanyo a-Si:H (not stabilized) (4/92)Solarex a-Si:H (not stabilized) (4/87)Glasstech a-Si:H (not stabilized) (9/89)

USSC/Cannon a-Si:H/a-SiGe:H/ss (not stabilized) (1/92)Solarex a-Si:H/a-SiGe:H (not stabilized) (10/87)

USSC a-Si:H/a-Si:H/a-SiGe:H (not stabilized) (10/9EDC a-Si:H/a-Si:H/a-SiGe:H (not stabilized) (12/8Sharp a-Si:H/a-Si:H/a-SiGe:H (not stabilized) (10/

SiGe:H (not stabilized) 12/92)


Fig. 9. Cross-section of micromorph thin-film solar cell: (a) glass substrate configuration, with ZnO intermediate reflector and (b) stainless-steelsubstrate configu-ration.

• Manufacturing costs and capacities (vacuum-based production, yields, volume; substratesand packaging).

• Amorphous Si:H deposition rates (increaserates needed for manufacturing).

• Research (modeling, characterization, analysis,deposition R&D).

2.2.3. Thin-film silicon: polycrystallineFrom consideration of improving materials utilization, thin-

film Si was always the logical progression toward the idealsolar cell. Early work in this area was limited to cells hav-ing efficiencies in the 5% regime—much below expectations[71]. These were mainly grown on foreign substrates, suchglass and graphite, using vacuum deposition and chemical vapordeposition—but always producing films with small grain sizesand high defect densities that limited carrier lifetimes. Addi-tionally, the relatively poor optical absorption characteristics ofthis indirect bandgap semiconductor required relatively thicklayers to produce adequate electron–hole populations from theincident photons. For the purposes of learning more about theprocesses in thin-film Si, there has been some progress in boththinned and in epitaxial layers of Si on Si[72–75]. In addition,there has been some progress in polycrystalline thin-film Si onforeign substrates[76–78]—including some recent commercialventures[79,80].

Several demonstrations of thin Si cell performances haveappeared in the literature over the past 10 years, most withthe purpose of demonstrating viability of “thin-Si” technol-ogy from the performance and device engineering perspectivesThese have ranged from thin expitaxial Si films on Si sub-strates to Si on glass or ceramics, the goal being able to utilizinexpensive support structures in the latter case. A brief sum

mary is shown inTables 2 and 3for these various technologydemonstrations.

A commercial entry that has attracted some attention andinterest recently has been the “crystalline silicon on glass”(CSG) technology[79–83]. This approach had been under devel-opment by Pacific Solar in Australia since the mid-1990s, grow-ing out of their analysis of the highest payoff paths to thin-filmsolar cell market penetration. Success with the prototype mod-ule technology led to the reformation the initial startup into CSGSolar, which has now been taken under Q-cells, with productionfacilities deployed in Germany. The banner for this technologyis that is combines the proven strengths of wafer crystallineSi technology, while posing potential improvements over mostcurrent thin-film manufacturing approaches and offering somereliability, materials availability, and environmental advantagesover some of the other PV thin films.Fig. 10presents a cross-section of the CSG cell structure. The substrate is textured glasshaving a Si nitride deposited surface layer. Three layers of a-Si films with appropriate doping and a silicon oxide cappingfilm are deposited in a single deposition chamber. These lay-ers are crystallized and hydrogenated using plasma techniques,providing the active cell for the patterning process. Inkjet orplasma methods form grooves (Fig. 10). A low-cost “opticallynon-absorbing resin, electrically insulating” (micron thickness)is applied and patterned in two stages: (1) the first producing thenegative crater contact regions (where the Si is etched throught p-l tires r intot ell.T sings d fort ion-a ntials ces.P eenr ging

.

e-

o then + layer) and (2) the positive dimple contacts to theayer. Aluminum (for example) is depositing over the enurface and patterned appropriately by either injet or lasehin strips joining the n+ to the p+ region of the adjacent che cells offer integration with a small number of procesteps. The Si has sufficient conductivity to avoid the neehe usual transparent conducting oxide (TCO) film. Additlly, some creative “strip interconnection” circumvents potehunts that can be issues with thin-film module performanrototype modules into the 8% efficiency regime have b

eached, and the expectation is to reach 10% levels in brin


Table 3Summary of confirmed, selected thin-film solar-cell efficiencies and related parameters, under standard measurement and reporting conditions[194], except (*),which are reported but not confirmed


Cu-ternary and multinaries678 32.5 85.3 0.449 18.8 NREL ZnO/CdS/CIGS (12/98); also, 18.2%, 1.1 cm2 cell

(1/99)693 35.7 79.4 0.410 19.5 NREL ZnO/CdS/CIGS (9/04)669 35.73 77.1 1.039 18.4 NREL Large area (3/01)605 36.19 68.6 0.462 15.0 NREL ZnO/CIGS(1/99) Cd-free cell666 30.51 75.6 0.418 15.4 NREL ZnO/CdS/CIGS (electrodeposited) (2/99)636 34.64 71.5 0.442 15.7 NREL ZnO/[Cd-doped CIGS] (2/01)

671 34.0 77.6 0.15 (Active area) 17.7 Ritsumeikan University Active area efficiency; ZnS buffer, small area (11/00)17.4 NREL CIGS on stainless steel (flexible) 2/00

539 33.7 73.6 0.192 13.4 Siemens Solar ZnO/CdS/CIS (11/92)736 510.1 80.5 0.102 21.1 NREL Concentrator: 14.3× (21.5% direct, 14.1× (3/01)

CdTe843 25.09 74.5 1.047 15.8 Univ. South Florida MgF2/7059 glass/SnO2/CdS/CdTe/C/Ag (6/92)

15.8 NREL MgF2/7059 glass/SnO2/CdS/CdTe/glass (4/99)

848 25.86 75.5 1.131 16.4 NREL CdSnO/CdS/CdTe/ glass (2/01)845 25.90 75.5 1.132 16.5 NREL CdSnO/CdS/CdTe/ glass (9/01)840 26.1 73.1 1.0 16.0 Matsushita 3–5�m CSS CdTe; question QE-current (3/97)

Other advanced types795 19.4 71.0 0.25 11.0 EPFL Nanocrystalline dye (Gratzel (12/96)795 11.3 59.2 141.4 4.7 INAP Nanocrystalline dye (Gratzel) submodule (2/98)726 15.8 71.2 2.36 8.2 ECN Nanocrystalline dye (Gratzel) (7/01)522 22.7 70 4.00 7.8* Toshiba GLE (polymer gel electrolyte) Photoe-electrochemical

cell (5/00)

835 6.3 63 3.3* Bell Labs/Lucent “Plastic Cell” (ITA/Pentacene) (5/00)4.9 NREL “Plastic Cell” (8/05)

Advance tandems4.0 25.8 Kopin/Boeing GaAs/CIS thin film (11/89)2.4 14.6 ARCO a-Si:H/CIGS (6/88)

0.768 25.5 68.90 2.4 13.8 NREL Transparent CdTe cell0.357 6.06 68.01 1.47 CIS cell

1.14 15.3 Glass/Cd2SnO4/ZnSnOx/CdS:O/CdTe/CuxTe—Glass/Mo/CIGS/CdS/ZnO CdTe/CIS 4-terminal mechanicalstack (12/04)

this into a viable commercial product. This concept has pro-gressed rapidly from demonstration to prototyping and certainlyhas advantages if the performance levels, manufacturing cost,energy payback, and reliability parameters are realized in itsfirst-time manufacturing phase. It is an ambitious program, butone that currently has significant technical and funding resourcesbehind it.


• Manufacturing (cell and module performance,costs, yields, complexities; quality assurance;diagnostics).

• Manufacturing capacity (requires demonstra-tion); industry involvement.

• Energy payback.

• Cell and module efficiencies approaching the15% levels (materials quality, crystallinity,reproducibility; device fabrication).

• R&D initiative (part of next generation Si tech-nology).

2.2.4. Thin-film copper indium selenide, its alloys, andrelated chalcopyrites

Interest in the Cu-ternary semiconductors began in the early1970s for non-linear optics[84]. The bandgaps of several mem-bers (including CuInX2, with X = S, Se, and Te) of this chalcopy-rite family exhibited properties well suited for PV consideration.These were typically direct bandgap semiconductors, capable ofeither p- or n-type conduction, having high optical absorption,stable electro-optical properties, and bandgaps matched to the


Fig. 10. Cross-section of thin-film Si solar cell (crystalline silicon on glass or CSG technology).

solar spectrum. Initially, the emphasis was on the heterostructureCdS/CuInSe2 (CIS), which was first demonstrated with efficien-cies at 12% for single crystals[85] and at 6% for thin films inthe 1974–1976 timeframe[86]. The device evolved into a alloycousin, Cu(In,Ga)Se2 (or “CIGS”) and Cu(In,Ga)(Se,S)2, whichhave slightly higher bandgaps (to about 1.2 eV for usual cellcompositions compared to 1.04 eV for CIS) for better voltageoutput for this “heterojunction” solar cell[87].

The cross-section of the device is represented inFig. 11,with an SEM micrograph of the various layers. Each of thelayers, thicknesses, interfaces, and compositions are ascribed

F lenidet ited-l

to the engineering of the cell for optimal performance andreliability. The most widely used back contact is a sputter-deposited molybdenum thin film. The usual substrate is sodalime glass, with the Na having some benefit to the CIGSproperties and the electro-optical characteristics of the device[87,88]. (In fact, technology not using soda lime glass incor-porated the Na during the deposition process.) The issue ofNa remains a research topic, especially optimizing concentra-tions and modeling its effects. Two recent investigations areof interest to the device optimization from the materials pointof view. The first relates to the compositional grading withinthe CIGS (the relative ratios of the Cu:In, in particular). Stud-ies by Sites et al. have identified the reasons underlying theoptimal compositional profiles (Fig. 12) [89]. Additionally, theeffects of the grain boundaries have always been of concern inthese polycrystalline devices. Modeling and analysis by Zunger[90] and by Noufi[91] have shown that the grain boundariesin the p-type CIGS do not affect the carrier conduction—andthe polycrystalline semiconductor behaves almost a single-crystal device for the minority-carrier transport. Thus, passi-vation and other such schemes have been unnecessary, andthis phenomena resulting from normal deposition underliesthe high light-generated current densities—a remarkable prop-erty for a polycrystalline semiconductor having micron-sizedgrains.

Several CIGS and CIGSS deposition methods have been suc-c hesei hav-i etalsaS apord , ande edu out5 hasb

ig. 11. Cross-sectional representation of copper–indium–gallium–sehin-film solar cell, including electron micrograph showing actual deposayer configurations.

essfully used to realize greater than 10% solar cells. Tnclude deposition from the elemental constituents (thisng produced the highest efficiencies), sputtering of the mnd post-treatments in selenium-containing gas (e.g., H2Se) ande-containing solids (including elemental Se), physical veposition and post-Se treatments, chemical depositionlectro-chemical techniques[92]. The CdS is usually depositsing immersion in liquid baths to provide a very thin (ab0A) layer [93]. Sputtering and chemical vapor depositioneen used for the ZnO[94].


Fig. 12. Analytical study of effects of Cu and In gradients in the CIGS activeexplaining the resulting performance and optimization of these thin-film solarcells[89].

The best research cells have been validated at a remarkable19.5% efficiency[95], with several headline cells in this familysummarized inTable 3. This device technology has also pro-vided the first better than 20% efficiency for a polycrystallinecell—at 21.1% under 14.3× concentration[96]. Certainly, thepositive and perhaps unique factors that favor this thin-film tech-nology are stability and large-area production potential—withperformance characteristics for smaller area cells similar tothe module performances. The best commercial modules havereached 13% with 4 ft2 areas (Fig. 13), and manufacturers in theU.S. and Europe report 10–11.5% average efficiencies from theirmanufacturing lines[97,98]. Research centers on the effects ofalloying (with materials like Ga and S), replacing the CdS win-dow layers with Cd-free layers (including ZnS and ZnO, withthe best such cell ZnO/CIGS at 16.5%), and the use of non-glasssubstrates. The most successful of the non-glass approaches hasbeen the use of flexible stainless steel, and commercial productsfor battery charging for military and for recreational applicationshas efficiencies in the 8–10% range[99]. Recently, a commercialmodule of this technology was verified with 10.1% efficiency,providing both lightweight and flexibility for the “power roof-ing” applications[100].

Looking toward the materials demands for multi-megawattsof this technology, current research has focused on reducing the

Fig. 13. Current–voltage characteristics of headline commercial CIGSeS mod-ules: (a) showa shell (13.4% conversion efficiency) and (b) shell solar (12.9%conversion efficiency).

thickness of the absorber layer from the 2.5�m regime to lessthan 1�m [101]. One objective is to reduce the amount of Inused—a material that has issues with availability (abundancy),competition from the large and growing flat-panel display indus-try, and recent rising prices. The requirement is obviously to


Fig. 14. Scanning electron micrographs of CIGS absorber layers for: (a) 1�mthick absorber deposited by three-stage process (16.2% efficiency); (b) 0.75�mthick absorber, deposited by coevaporation (12.5% efficiency); (c) 0.5�m thickabsorber deposited by coevaporation (8.7% efficiency). For completeness,1�m absorber deposited by coevaporation yielded a 16.1% efficiency in thisstudy.

reducethickness without adversely affecting the cellperfor-mance. There have been some initial successes (see deviccross-sections inFig. 14a–c), but device efficiencies decreasewith absorber thicknesses <1�m (seeTable 4). These cells have

a lower Voc, attributed to recombination in the back contactregion. The current collection is expected to improve with theuse of back reflectors, and with the texturing of the back surfaceto enhance the path length of the longer wavelength light. Inorder to improve energy payback and to enable the use of flexiblepolymer substrates, there is also work at producing these thinnerlayers at lower temperatures (nearing 400◦C). Initial results at450◦C with a 1.4�m absorber thickness, cell efficiencies near13% have been reached (12.9%,Voc: 0.585V;Jsc: 31.1mA/cm2;FF: .706 with no antireflection coating)[101]. Finally, theseresearch and current commercial devices use high-purity mate-rials (99.999% or better)—maximizing the performances byminimizing the introduction of possibly detrimental impuri-ties. Recent studies for use of lower-purity sources for possiblecost reduction have been initiated because very little is knowabout the affects of various impurities on ternary semiconductorand device properties. Preliminary cells using 99.9% elementalsources of Cu, In, and Ga have led to efficiencies about 2 per-centage points less than the high-purity material control samples[101]. Finally, the throughputs of devices in the manufacturinglines can be enhanced greatly by the ability to use higher depo-sition rates for the absorber. The results of doubling the ratesfrom the current∼0.06�m/min show that efficiencies∼20%less than control samples are produced. The results of higher-rate deposition indicate problems with loss inVoc—likely dueto the outdiffusion of Gas being limited by the reaction rate oft

ortso[ newp lls( vicer h-n fore Sec-o keda t hasr issuet h uti-l xide[ cedt thea ori of am

TA r cells[

T

11000

a

e

he binary selenides during the second processing stage[101].Work on other Cu-ternaries continues, with periodic rep

f research progress on CuGaSe2, CuGaS2, and CuInS2102–106]. These have some additional importance forolycrystalline device directions—multi-junction solar cediscussed later in this paper). Several single-junction deesults are cited inTable 4. Central to discussions of this tecology relate to materials. First is the availability of the In,xample, if hundreds of megawatts are to be deployed.nd is a current one—that of the cost of the In itself (tracs a bi-product of the zinc refining process)—a price thaisen over an order of magnitude in the past year, and anhat has also caused concern in the flat-panel industry whicizes this element in its transparent conducting indium–tin-o107]. Most of those working in the technology are convinhat indium supply is sufficient, especially if the designs ofctive layer reach the less than 1�m that they are targeting (

f this much thinner cell eventually is successful as partultiple approach)[108].

able 4bsorber thickness-dependent parameters for best CIGS thin-film sola

101]

hickness (�m) Voc (V) Jsc (mA/cm2) Fill factor (%) Efficiency (%)

.0 (3-stage) 0.654 31.6 78.3 16.2

.0 (codep) 0.699 30.6 75.4 16.1

.75 (codep) 0.652 26.0 74.0 12.5

.5 (codep) 0.607 23.9 60.0 8.7

.4 (3-stage) 0.565 21.3 75.7 9.1



• Research (chemical paths to materials realiza-tion, window, process development and inte-gration, contacts, role of sodium, alloy compo-sition optimization).

• Complexity (manufacturing costs, control).• Stability of modules (encapsulations, operation

in hot and humid climates); cells (efficiencyincreased with light exposure); transient effects.

• Devices (lower Voc, higher Isc than preferred).• Elimination of CdS window (environmental

acceptance, less complex processing).• Scale-up (deposition techniques, throughputs,

yields, suitable large-scale deposition tech-niques/conditions for “20%” quality devices).

• Materials (availability, especially of the In, cost,environmental concerns).

• Manufacturing base (currently embryonic com-mercial and product base).

2.2.5. Cadmium tellurideSince the 1960s, CdTe has been a candidate PV

material—first for space, then as the “next-in-line” among thepolycrystalline thin films[109], and now the leading terres-trial product (with sales expected to reach 20 MW this year)[110]. Having a nearly ideal bandgap for a single-junction solarcell, efficient CdTe cells have been fabricated by a variety ofpotentially scalable and low-cost processes, including physicaldeposition, spraying, screen printing/sintering, and electrode-position [111]. Inherent to most cell processing is a CdCl2chemical treatment[113], eitherliquid (in CdCl2:methanol solu-tions [113]) or a currently preferreddry vapor process [114](because of better process control and ability to be incor-porated in the manufacturing line)—both treatments done attemperatures of∼400◦C for periods of 8–20 min. The bene-ficial effects (increases in efficiency 10–25%) of this processhave been attributed to enhanced grain size, evolution of ap–I–n or heterojunction, surface alteration/passivation, alter-ation of shallow, and/or deep electronic levels, improvement inmorphology, and the formation of an interfacial CdSTe layer,a solid solution with 2–15% sulfur[115–116]. Some alter-natives to this process have been investigated, including anoxygen-alloyed CdS, with a resulting bandgap between thatof the CdS and up to∼3.2 eV for higher oxygen contents[117].

A representative cross-section of the device is shown inF thev encyi or-t 11%e -u elate

Fig. 15. Cross-sectional representation of cadmium telluride thin-film solar cell,including electron micrograph of this region.

to contacting, contact stability, ability to control the CdTeconductivity with oxygen and other extrinsic dopants, chem-ical and heat treatments, the transparent conducting oxidesat the top surface of the cell, and the packaging critical forlong-term life of the module[120]. In fact, this has initiallycaused some concern for the product operating in outdoorconditions, but attention to new packaging techniques andprocesses have been successful in overcoming most of theproblems.

Although there are concerns, perceptions, and misconcep-tions about the environmental, safety, and health effects ofthe Cd in this device, extensive studies indicate that all safetyissues can be handled with modest investments in cost, recy-cling of the materials and modules, and tracking of deployedproduct[121]. A major attribute of this technology (contrast-ing other thin films) is that its manufacturing is not limitedby the deposition of the active semiconductor layer. In fact,this “front end” deposition step in the module fabrication isaccomplished in several minutes. Thus, when considering hun-dreds of megawatts of production, CdTe thin films lead all othertechnologies. The commercial segment is growing, more thandoubling capacities in the next couple of years both in the U.S.and Europe. This is spurred partially by market opportunities(world demand coupled with the Si supply slowing), but alsowith the resolve and confidence this industry has for and in itsproduct.

ig. 15, with a corresponding SEM micrograph showingarious layers. The best confirmed research-cell efficis 16.5%[118]; the CdTe film produced using close vapransport. The champion commercial module has reachedfficiency (Table 5) [119]; with off-the-shelf commercial prodcts in the 7–9% range. Areas of concern for devices r


Table 5Summary of confirmed, selected PV module efficiencies, and related parameters


Crystalline silicon5.6 3.93 80.3 778 22.7 UNSW/Gochermann PERL (9/96)14.6 1.36 78.6 1017 15.3 Sandia HEM multicrystalline (10/94)20.1 2.72 73.6 3931 10.3 Texas Instr. Spheral Si (9/92)20.1 1875 20.3 Sandia/UNSW/ENTECH Crystalline Siconcentrator module (80 suns) 12 cells

Cu-ternary based23.42 2.83 67.9 3651 12.1 Siemens Solar CIGSS, 44.3 W (3/99)9.33 0.181 73.7 90 13.9 Univ.Stuttgart Submodule, 15 cells (4/96)2.643 1.334 75.1 16.0 16.6 Uppsala, Sweden CIGS, 4 cells (3/00) submodule6.949 0.057 71.7 18.9 14.7 Siemens Solar Submodule 12 cells series (7/99)7.46 0.173 68.0 69.1 12.7 Siemens Solar CIGSS (4/94)31.2 0.625 68.9 3459 13.4 Shell Solar CIGSS (8/02)

CdTe8760 10.5 BP-Solarex 91.5 W: highest power thin-film module (5/00)

46.45 3.07 64.3 8670 10.5 BP-Solarex Large-area, hi-power

26.21 3.205 63.3 4874 10.7 BP-Solarex 53.9 W (4/00)6728 9.1 First Solar 61.3 W (6/96)5432 9.7 Matsushita 52.7 W (12/99)

34.91 0.653 63.2 1376 10.5 Matsushita (2/99)

Amorphous Si:H4.353 3.285 66.0 905.1 10.4 USSC a-Si:H/a-Si:H/a-SiGe:H (stabilized) (10/98)

9276 7.6 USSC a-Si:H/a-Si:H/a-SiGe:H (stabilized), 70.8 W (9/97)1200 8.9 Fuji a-Si:H/a-Si:H (stabilized) (9/93)

12.5 1.3 73.5 100 12.0 Sanyo a-Si:H/a-Si:H (unstabilized) (12/92)

Polycrystalline Si25000 3.26 68.0 661 8.2 Pacific Solar PECVD on glass-submodule; 1–2�m Si (5/95)

III–V34 27.0 ENTECH GaInP/GaInAsGe(10× conc.) (5/00)41.4 25.1 Sandia GaAs/GaSb (57× conc.) (3/93)

Aperture area is used unless indicated[194].


• Research (process development, modeling,interface optimization, contacts, chemical treat-ments, role of oxygen).

• Substrates (use of sodium-containing glasses,cost of other substrates).

• Environmental concerns and perceptions (Cd,potential of environmental regulation limitingdeployment in some countries, existing controlof product currently in some major markets).

• Materials availability.• Stability (Cu diffusion and oxidation of contacts,

contact degradation, packaging especially forhot and humid climates).

• Modules (encapsulation).• Ability to process other device layers (win-

dows, contacts, etc.) at rates compatible withthe active layer rates.

• Manufacturing base; recycling of modules.

2.3. Very-high efficiency and concentrator devices

Higher-cost semiconductors, such as GaAs, GaAlAs,GaInAsP, InSb, and InP (Table 6) have been receiving attentionas PV converters because they have exceptional performancedemonstrations that have the potential to convert more than athird of the sun’s terrestrial power into electricity[122]. Costis the overriding consideration for terrestrial applications inconventional flat-plate technologies. One means for improv-ing both the PV efficiency, reducing the high-value converterarea, and significantly reducing the systems cost is the use ofconcentrators—lenses, reflectors, or other optics that focus thesunlight onto the collection area of the solar cell. Concentratorshave been used successfully with crystalline silicon technol-ogy[123], with concentrations up to 400×, efficiencies to 27%,and larger-scale modules at 20% using 25% commercial cells.Single-junction GaAs cells have been measured at 28% at 1000×concentration[124]. The economics of these approaches havebeen argued for decades—but it has been the leveraging of themultiple-junction III–V cell technologies for space applicationsthat have brought renewed interest and investment into the ter-restrial concentrator system.


Table 6Summary of confirmed, selected III–V solar-cell efficiencies and related parameters, under standard measurement and reporting conditions[194]


Single-junction, one-sun cells1022 28.2 87.1 3.91 25.1 Kopin GaAs,AlGaAs window (3/90)1011 27.6 83.8 4.00 23.3 Kopin GaAs Cleft (4/90)4034 6.55 79.6 16.0 21.0 Kopin GaAs Cleft 5mm submodule (4/90)1018 27.56 84.7 0.25 23.8 Spire/Purdue MBE GaAs (3/89)878 29.3 85.4 4.02 21.9 Spire InP (4/90)813 27.97 82.9 0.108 18.9 NREL ITO/InP (8/88)994 23.0 85.4 4.011 18.2 RIT Poly-GaAs on Ge (11/95)

Multi-junction, one-sun cells2488 85.6 81.5 4.0 30.3 Japan Energy GaInP/GaAs (4/96) 2-terminal, monolithic876 28.2 82.9 0.310 20.9 NREL InP/GaInAs (8/90)

337 21.94 72.1 0.312 5.3 3-terminal tandem26.2

4.00 25.8 Kopin/Boeing GaAs/CISx (11/98)

1402 13.92 86.8 4.00 17.6 Varian AlGaAs: 1.93 eV

1000 13.78 83.0 16.0 GaAs (3/89)27.3 3-terminal tandem

2622 14.37 85.0 3.989 32.0 Spectrolab GaInP/GaAs/Ge 2-terminal, monolithic (1/03)2910 12.22 87.54 0.237 31.1 NREL GaInP/GaAs/GaInAs 2-terminal, monolithic, lattice-mismatch,

inverted structure (3/05)

Concentrator cells0.126 27.6 Spire GaAs [255 suns]0.075 27.5 NREL GaInAsP [171 suns] Entech cover0.075 24.3 NREL InP [pp suns] Entech cover0.53 32.6 NREL GaAs/GaSb 4-terminal, mechanial stack [100 suns]0.63 31.8 NREL InP/GaInAs 3-terminal, monolithic [50 suns]0.103 30.2 NREL GaInP/GaAs 2-terminal, monolithic [180 suns]

32.3 Spectrolab/NREL GaInP/GaAs/Ge 2-terminal, monolithic [100 suns] (9/99); also,same efficiency (5/00) at 560 suns

3039 2435 88.4 0.2639 35.7 Boeing/Spectrolab GaInP/GaAs/Ge 2-terminal, monolithic [175.2 suns] (3.04)3120 1392 88.2 0.2428 37.9 NREL GaInP/GaAs/GaInAs 2-terminal, monolithic, lattice-mismatch,

inverted structure [10.1 suns] (4/05); also, AM0 efficiency of29.7%

3089 3377 88.24 0.2691 39.0 Boeing/Spectrolab GaInP/GaAs/Ge 2-terminal, monolithic [236 suns] (5/05)

The first laboratory demonstrations of greater than 30%efficiency terrestrial multiple-junction devices were actuallyreported at the beginning of the 1990s[125–128]. Various 2-,3-, and 4-terminal configurations, shown inFig. 16, have beeninvestigated over this last 20 years (see alsoFig. 17). The param-eters for the record devices are summarized inTable 6. Thesecells attain “beyond the conventional” efficiencies because theyuse multiple devices in the same area, each tuned to a differentportion of the solar spectrum (e.g., the top cell tuned to the redand the bottom cell to the blue). Thus, the total device is moreeffective in utilizing the total frequencies of the sunlight thatintercepts its surface.

The space markets of the late 1980s and 1990s demanded highpower to weight ratio—and the multiple-junction approacheswere ideal[129]. Both 2- and 3-junction cell technologies weredeveloped. The latter configuration realized AM0 cell efficien-cies approaching 30% for the non-concentrator space arrays.These notable results rekindled interest in the terrestrial use ofthese devices – and led to forecasts of cells with efficienciesof 40% and systems in the 30–35% range – the realization of

“a third of the sun’s energy” converted to consumer electricity.The use of these in concentrators provided some benefits: low-ering the system’s cost by minimizing the area of the expensivecell (by 400–1000 times for concentrations of 400×–1000×),and making use of the increase in efficiency of these devicesat higher concentration with increased current densities withoutsignificant lowering of the other cell parameters[130]. The bestterrestrial triple-junction monolithic cell has been confirmed at39% under 360× concentration (Spectrolab/Boeing)—the high-est efficiency attained to date for any solar-cell technology[131].

Two major device configurations have evolved for the high-efficiency terrestrial concentrators, both 3-junction, 2-terminalmonolithic structures. The first is the lattice-matched device,shown in Fig. 17a, Ge/GaAs/GaInP (the same configurationdeveloped for space). The champion cell is 39.0% under 400×concentration (see IV characteristics inFig. 18a). This cellcaptures the blue and green portions of the spectrum effec-tively, but Ge is not the ideal third junction. Research centerson developing a “1 eV” cell, such as GaInAsN that can bethe bottom cell—or a fourth junction in this configuration.


Fig. 16. Cross-sectional representations of multiple-junctions solar cells, show-ing (a) 2-terminal monolithic, (b) 3-terminal monolithic, and (c) 4-terminalmechanically stacked configurations.

Problems, however, with incorporating the nitrogen into theGaInAs without accompanying intraband defect levels has lim-ited the progress of this approach[130–133]. Certainly, the4-junction Ge/GaInAsN/GaAs/GaInP structure has been modeled and is projected to exceed the 40% goal for the concentratotechnology.

The lattice mismatch approach,Fig. 17b, presents some inter-esting possibilities[134,135]. It has already demonstrated one ofthe highest efficiencies confirmed for a solar cell (37.9% shownin the IV characteristic ofFig. 18b). The benefit to this approachis that the bottom cell as a higher bandgap than for the latticematched configuration, and the modeling predicts greater tha40% can be attainedwithout a fourth junction. In addition, theprevious lattice-match device presents some challenges for thdeveloping an anti-reflection coating that can effectively coverthe large spectral range. It has not yet been realized and potetially requires new materials development and engineering omultiple, complex layers that may add to the cost and processing complexity of the device. This 3-junction lattice-mismatcheddevice avoids this problem. Of course, other problems may provide difficulties—such as the stress that can cause bowing anmaking the final processing and handling of the cell difficult.

These concentrator technologies are primarily aimed at largeutility scale applications (for high solar insolation regions suchas the southwestern U.S.)[136]. However, a number of organi-zations have been pursuing “rooptop” potential systems—thac per-h

case, the “concentrator” has developed a new life—thanks tothe investment in space technology and to the persistence ofthis R&D community for terrestrial solar power service. Thistechnology that has always been dubbed the “application of thefuture” may have made its first viable footprints in the nearerterm markets with high efficiency and high electricity value.“Roadmaps” predict significant markets for such utility-scalePV in the 2020–2025 timeframe, and “CPV” is positioning toserve those.


• Cost (cells, materials, manufacturing, and pro-cessing; hardware; materials availability).

• Industry (primarily directed to space applica-tions; capacities).

• Utility market (must compete with electricitycosts 1/3–1/4 those of rooftop PV ¢/kWh prices;markets harder to penetrate).

• Research (development of 1 eV “fourth junc-tion” for lattice-matched cell; question ofantireflection coatings; processing of lattice-mismatched structure, cell engineering, systemdevelopment).

2

era-t reen[ gen-e lms,a s, andw rfor-m awattsb om-p n then trug-g t form ltagea ers’fin itingt fort havee theo-rst

an certainly be projected for commercial buildings andaps for some residences as well in the future[137,138]. In any

-r

-n

e

n-f-

-d

,

t

• Systems (concentrator hardware suitable forPV; cost-effective structures, tracking and elec-tronics, long term reliability, cooling of the cell,maintenance, daily energy output or two-accesstracker compared to other approaches).

.4. PV technologies: the race toward the next generations

Some of the possible contenders for the next PV genions have started their journeys in the laboratory. Martin G139] brought attention to this future when he classified 1stration PV as crystalline Si, 2nd generation as the thin find 3rd generations as a host of evolving devices, upstartild ideas that have lined up in the race to meet the peance and cost goals needed to deliver those 15–30 tery mid-century. Whereas the “2nd generations” might be ceting in the analogue of the 100 m dash to surpass Si iow to near-term, the 3rd generations are in marathon sle that must not only bring them to commercialization, buost—even to demonstrate their abilities to generate vond current for the very first time. This is the PV researcheld of dreams. It is also theparking lot of nightmares for theear-term real business of photovoltaics—delaying or inhib

he adoption of real and working technologies that will servehe next 20–30 years in order to wait for one that might notven been demonstrated to generate electricity yet but inetically promise performance beyond Olympic levels. (This isomething many of us have experienced awaiting the next, thenhe next, speed bump up in computer microprocessors—and we


Fig. 17. Cross-sections of triple-junction, high-efficiency solar cells: (a) lattice-matched design, (b) lattice-mismatched (metamorphoric), and (c) lattice semi-mismatched, thin, inverted structure. ARC is the antireflection coating.

may never purchase a computer!) There must be an understand-ing and patience—knowing that the investment in these researchareas is important for both future technology ownership and forreadying the next generation(s) of solar electricity for many gen-erations of consumers to come.

2.4.1. Fooling mother natureBiomimetics (or mimicking nature’s use of chlorophyll)

was underlying the first nanotechnology approaches to evolveultralow-cost technologies (seeFig. 19) [140]. Using dyemolecules, Michael Gratzel reported new success in PV conver-sion[141–145]—in which the sunlight creates bound electronsand holes – excitons – that travel as a unit and separate only afterreaching some material boundary. In the case of Gratzel’s dyecell shown inFig. 20, the interface is that between the titaniumdioxide and the dye. After separation, the electron is conductedin the TiO2 (the same material that is used in paint and tooth-paste) and the hole returns to the counter electrode throughthe electrolyte. The organic dye region is very thin—and theexceiton is quenched very rapidly by injecting the electron intothe TiO2 particle. This is an advantage of this cell in that the

diffusion of the hole through the thin dye does not limit the per-formance as much as it does for organic cells reported in the nextsection. This device has reached 11% conversion efficiency in itssimple form[145]. Recently, a dual-junction dye cell (buildingon the efficiency of spectrum capture pioneered by several othertechnologies) has been reported with a 15% efficiency[146].However, not much has been released on this cell and its struc-ture. There is some question on whether it is a tandem usingtwo dye-sensitized cells (mechanical stack equivalent) or onethat uses a dye cell in conjunction with a conventional solar cell(perhaps Si). Apparently, intellectual property issue has heldoff the release of this important information at the time of thiswriting and the report[146].

The dye cell has been in development for a relatively longperiod compared to a number of the other next-generationapproaches. Most in the PV community associate the prob-lems and issues of the dye cell primarily with the dye. Thematerials for the dye are expensive today because it is a spe-cialty molecule, but these materials should not be showstoppersfrom either cost or availability in the future. The manufacturingis extremely low cost because there is no vacuum equipment


Fig. 18. Current–voltage characteristics of the champion cells for the: (a) lattice-matched triple-junction Ge/GaAs/GaInP monolithic concentrator cell [39%] and(b) inverted, thin lattice semi-mismatched triple-junction monolithic concentra-tor cell [37.9%].

involved. However, the liquid dye and a solid–liquid interfaceraise some questions about 30-year lifetimes. Gels and solid dyeare being investigated, but have not yet provided the same peformance levels as the original designs[147–149]. In general,many worry about the long-term viability of this technology.However, there have been some encouraging tests completethat indicate that this can achieve reliability level competitivewith other PV approaches. It remains a serious member of th2nd generation approaches—with some prototype manufactuing underway[150].

Fig. 19. Pathways toward next-generation photovoltaics technologies for ultra-low cost and ultra-high efficiency, indicating electricity cost targets and expec-tations of current U.S. PV program and those for longer-term (“breakthrough”)approaches.

2.4.2. Just one word—plasticsOrganic photovoltaics also operate through excitonic pro-

cesses. Pioneering work by Heeger, MacDiarmid, and Shirawayin the late 1970s for organic electronics led to some early inves-tigations for PV in that same period—notably using polyacety-lene, perylene, and phthocyanine[151,152]. R&D was largelyabandoned during the 1980s because of drops in available fund-ing for PV, but organic semiconductors remained of interest tothe electronics industry because of radiation resistance (the coldwar paranoia) and lightweight. Now with the quest for new PVtechnologies and the growing interest in organic light-emittingdiodes, organic solar cells have again come to the limelight (actu-ally, the full solar spectrum!). This research comes with whatseems like an infinite number of possibilities—ranging fromsmall molecules (<104 molecular weight) to polymers (>106

molecular weight)[154,155].Despite all the leveraging R&D in the organic opto-

electronics areas, these organic cells are certainly still in theirinfancy. Only a few technologies have validated performances(see Fig. 1), and those still in the 3–5% range. But the

sr-

d

er-

Fig. 20. Device representation of dye-sensitized solar cell.


scientific research community is growing in this technologyarea—bolstered by the work with organic LEDs and the promiseof ultralow-cost technologies for materials and manufacturing.Of course, there are concerns and issues. Organic materials arenotoriously sensitive to ultraviolet radiation that can changebonding and lead to instabilities. Higher temperature operationmay also provide limitations for some organics. Although initialevidence indicates cells have better performances under con-centrated light[156]—the combined effects of UV and thermalenergy may limit this approach. More positively, the organicmaterials can be chemically controlled to respond to differentwavelengths—making multiple-junction approaches a poten-tial pathway to increased efficiencies[156]. Organic PV hasa great deal to overcome before becoming part of the PV energyportfolio – but cost, manufacturability and performance poten-tials – and the leveraging from other electronic technologies –help maintain a serious interest in these approaches. The dreamof unraveling our electricity from “Saran-Wrap-like” rolls ofplastic may someday become reality. But many questions needanswers, many technical barriers need to be surmounted, andmajor R&D efforts need to be successful before these innova-tions make it to their major commercial phase.

2.4.3. Using more of the sun, minimizing expensive realestate

Multiple-junction solar cells are among the few very-highe rciai prv g tom As-b 9%a ).

2 c-t , haa ata dc in thr numb nces

TC giesC ser ai

Iw

CLMISHTTISS

lattice-induced stresses, current matching, defect generation,inter-cell contacting and losses, light absorption/transmission,etc. Demonstrating a 50% 4 or 5 junction cell in the laboratorymay well be come within a decade or less—but translating sucha complex device from the laboratory bench to the commercialmarket will entail a massive effort covering a spectrum of com-plexities from the scale-up of the growth through the engineeringof the concentrator systems. Not easy, but also not beyond theingenuity and persistence of the PV technical community. Withthe time compressions experienced for technological advance-ment in the last five decades and the growing leveraging oftechnical creativity within the vast opto-electronics arena, the50% multiple-junction cell may not be that far off if sufficientresources are sustained for these approaches. Of the 3rd genera-tions approaches, this one may really be a closer 2nd generationcontender.

2.4.3.2. Polycrystalline tandems. Multiple-junctions are notthe sole property of the “III–V” semiconductors. Besides thetandem and triple-junction a-Si:H cells and modules, thin-filmpolycrystalline technologies are currently being developed, withthe prediction of exceeding 25% efficiency for two cells stacksand modules in the 20–25% regime[159]. Several polycrys-talline thin-film tandem structures are being developed, withsome initial demonstrations of device feasibility. A summaryo en-t hb chan-ica tiono ver,m arentc lithica and as lose

F cell( ersione

fficiency approaches that have been realized and commezed. Successes with 2- and 3-junction stacks have beeniously cited—namely, (a) a-Si:H-based thin films (leadinoderate 13% laboratory cells for 3 junctions); and (b) Gaased monolithic designs (with very-high efficiencies to 3lready confirmed for two different, 3-junction approaches

.4.3.1. Multi-multiple junctions. The use of a number of junions, each tuned to a specific energy of the solar spectrumlimiting efficiency for an infinite number of such junctionsbout 87% (seeTable 7) [158]. An infinite number of stackeells may not be practical, but concepts of 4–6 junctions areesearch stage. With each cell added, the complexity ander of problems increases with thermal expansion differe

able 7alculations of maximum efficiencies for various next-generation technoloarnot efficiency and calculations by Landsberg and Schockly and Queis

ncluded[132]

deal Converter (atT = 300 K andTs = 6000 K)ith isotropic illumination)

Efficiency

arnot 0.950andsberg 0.933ultiple-junction 0.868

mpact ionization (highest Q factor) 0.868olar thermal 0.854ot electrons 0.854hermophotovoltaic 0.854hermophotonic 0.854

ntermediate band (multi-quantum dot, alloy) 0.632econd photon pumped multiple-quantum well 0.632hockley and Queisser 0.403

l-e-

s

e-,

,re

f progress with these technologies, still in their developmal stages, is provided in Table 8[160]. The choice of the higandgap top cell semiconductor dictates the use of a me

cal or monolithic design. The mechanical stack (Fig. 21, thehampion cell to date with an efficiency of 16.5%[161]) hasdvantages of independent cell fabrication with minimizaf cross contamination, interdiffusion, and testing. Howeore materials are required (antireflection coatings, transp

onducting oxides, substrates) in the structure. The monopproach has one grid or contact on each side, one TCO,ingle ARC. However, current matching becomes critical, c

ig. 21. High-performance polycrystalline mechanically stacked tandemCdTe–CIGS) with IV data for this concept that has reached 15.3% convfficiency.


tolerances for the connecting junction (e.g., a tunnel junction),and the less-than-flat topography complicates the growth ofsequential layers.

Much of the R&D has focused on the top cell—becausetransmission through this structure and the performance ofthis device is critical[160]. Typically, these cells are in thebandgap range of 1.5–1.9 eV, and the semiconductor requiresminimum sub-bandgap absorption. Examples for this widerbandgap device include: (1) CuGaSe2 (CGSe), with a bandgapof 1.68 eV [162]. For this application, the surface of theCGS is modified, and devices with efficiencies up to 10.2%(Voc: 0.823 V;Jsc: 18.61 mA/cm2; FF: 0.668) have been real-ized. (2) Cu(In,Ga)(Se,S)2 with a bandap of 1.5 eV[163].The best bell with element ratios of Ga/(In + Ga) = 0.51 andS/(Se + S) = 0.33 has been 10.1% (Voc: 0.80;Jsc: 17.9 mA/cm2;FF: 0.707). (3) Cd1−xZnxTe with 0≤ x ≤ 0.8 having corre-sponding bandaps 1.5–2.1 eV[165]. For compositions withx ∼ 0.05 (Eg = 1.53 eV), cells with efficiencies to 11.9% havebeen fabricated. For higher 1.8 eV bandgaps (x ∼ 0.5), effi-ciencies to 8.4% have been validated. These cells have lowerthan expectedVocs, leaving considerable room for improve-ment. (4) BaCuTeF, deposited via pulsed laser deposition hasa bandgap of 2.3 eV[164], and is considered a possible cou-ple with CdTe. (5) MCuSnQ4 (with M = Ba, Sr; Q = S, Se) havebeen shown to have appropriate bandgaps[166], and the com-pound BaCuSnSeS has been measured with a 1.77 eV directb ctorsr andt theirs

butp or-m ically,t cer nciesi ch-n o bea

2r risk

a en-t eringt aps,o ns,t thersw archc ereu y top aturea teri-a dings ity ofs r andw wew on-s ely

Fig. 22. Conceptual representations for: (a) thermophotovoltaic and (b) ther-mophotonic devices.

include concepts or relatives of those that many consider to be atleast a bit risky if not approaching the radical fringe right now!

2.4.4.1. Thermophotovoltaics. The first for “solar” is a tech-nology that works, is used, but has applications utilizing theinfrared (primarily) portions of the spectrum and not currentlya major consideration for terrestrial PV. Thermophotovoltaics(TPV—the conversion of “thermal” wavelengths into electric-ity, Fig. 22a) has been under consideration for some time withorigins back to the 1960s[167]. TPV has made it to the nicheconsumer commercial markets[168,169]—but has some widerinterest in military ones. The technology awaited the develop-ment of suitable low-bandgap semiconductors—and the abilityto process them, with the swell in R&D coming in the 1990s.Because the active cells are tuned to a very specific wavelength(responding, for example, from the emission from a selectiveemitter or reflector), system efficiencies are usually consideredthe benchmark of this technology. These lie in the range of15–25% currently, but could exceed 40% for applications suchas the generation of electricity from “power sources” aboardnuclear submarines or converting the waste heat from automo-bile exhausts or from the float-glass process that brings us mega

2 2andgap. Of course, the complexity of these semiconduaises concerns about compositional control and stability—hey require considerable more investigation to determineuitability.

This area is of immense technical interest—high-risk,otentially high payoff with the dual promise of high perfance and low cost. These structures are complex—phys

opographically, and electro-optically. Initial performanesults are encouraging, but feasible structures with efficien excess of 20% are a challenge. Like other thin-film teologies, stability and lifetime remain issues that have tddressed as well.

.4.4. The far sidePV science and technology have always included highe

pproaches in its R&D portfolio; alternatives to the convional nearing or at the outer fringes of science and enginehat might provide breakthroughs, significant progress ler even new technologies. High-efficiency multiple-junctio

hin-films, nanotechnologies, organic solar cells, and oere all part of the intellectual domain of the solar-cell reseommunity in the 1950s though the 1970s. “Inventors” wsually far ahead of the ability of hardware and technologrove the concepts—and remained part of the valued literwaiting to be realized with the evolution of processing, mals, and characterization. Impatience (especially among funources!) sometimes shut the door too early on the ingenucientists and inventors—but fortunately, good ideas lingeait their turn as that next possible “breakthrough”. Whatill be using in solar PV at the end of this century will be ciderably different that what we have now—and will also lik


square miles of windows each year[169]. R&D on the cellshas probably kept in front of the critical needs for the emittersand reflectors in the thermophotovoltaic system. It is the systemperformance (efficiency) that is critical, including the lifetimeof the components outside the solid-state TPV converter itself.These issues have been discussed in recent TPV conferencesand in reviews of the technology[168,169]. There have alsobeen research and prototypes combining burning of bio-products(e.g., wood powder) to activate thermal emitters in the 1500◦Cregion that would match the responses of several semiconductortechnologies near 0.5 eV[170]. The problem of heat recoveryfrom low temperature gradients (such as waste heat) is a limitingfactor in system performance and economics.

2.4.4.2. Thermophotonics. Related is thermophotonics(Fig. 22b), which uses a diode that emits a photon with energyabove the bandgap (the radiator is an electroluminescent devicesuch as a light-emitting diode)[172]. The heat is supplied to onediode to maintain it at a higher temperature than the other. Thelower temperature one is maintained at ambient temperature.The two diodes are thermally isolated but optically coupled.When both diodes operate at the radiative limit, heat suppliedto the warmer diode is converted to power in an electricalload that is connected between the two devices. The efficiencycan approach the Carnot limit for conversion between thetemperatures of the warmer and cooler device (Table 7). Theset lizey

2 stt theya hnoo[w encyu uceb tionh g thq elect ze ot st rs it giono ogliew con-d ntlyr comp toat (i.e.,t prod rierit ch aa rriert ctrono to

moderate energy levels. The high-energy carrier may be pro-duced using an electric field, as is done in many electronicdevices, or from a high-energy photon, as happens in solarcells. Photonic impact ionization using the photons availablefrom sunlight is very inefficient in bulk semiconductor mate-rials, so much so as to be virtually unnoticeable. As a result,for all intents and purposes a bulk semiconductor solar cell pro-duces only one electron–hole pair per photon, no matter howmuch energy the photon has. Any excess energy carried bythe photon is wasted as heat. This multiple-exciton produc-tion has now been demonstrated by several research groups(using quantum dots of Se, PbS)[176]. These are significantresearch results—the first stages possibly leading to this newregime of ultra-high efficiency devices. Green et al.[183] haveproposed some interesting approaches using Si and high temper-ature processing integrated with low-temperature quantum dotcells. These are artificial semiconductors using one-dimensionalquantum confinement to extend the bandgap. The control ofthe bandgap has been demonstrated, and it is anticipated that adevice-proof of concept will be realized by the end of the decade.It is speculated that a 25%-efficient device could be ready by2020. However, it has to be emphasized that the photovoltaiceffect in any of these quantum-dot structures has still not beendemonstrated, quantum-dot solar cells are still in theirtheoret-ical, concept, and demonstration stages, and it may be sometime before the research community produces a true proof ofc

s semi-c e har-v rum.T -dots n-t PVc eb gies tob dgap.T diateb roupsh su

hermophotonic devices have been modeled, but not reaet in the laboratory.

.4.4.3. Nanotechnology approaches. Other concepts are juhat, with no full operation of cells demonstrated thoughre theoretically possible. This includes several nanotecgy and hot-electron approaches such as thequantum dot cell

171]and relatives using quantum rods and quantum pods[172],hich has the a potential thermodynamic conversion efficip to 66% by utilizing hot photogenerated carries to prodoth higher photocurrents and voltages. Several configuraave been proposed for these devices—including imbeddinuantum dots in polymers and other semiconductors. The

ronic properties are controlled to a large extent on the sihe nano-particles themselves (Fig. 23) [173]. Quantum dot cellake advantage of quantum size effects. When the carriehe semiconductor are confined by potential barriers to ref space that are comparable or smaller than their deBravelength (or to the Bohr radius of excitons in the semiuctor bulk), the hot-carrier collection rates may be significaeduced. Then the rate of impact ionization could becomeetitive with the rate of carrier cooling. The critical point isvoid the thermalization of the carriers (excitons)[174]. A keyo the efficiency increase is multiple-exciton generationwo or three excitons generated per incident photon). Theuction of multiple carriers from a single high-energy car

s a well-known process called impact ionization[175]. Mul-iple carriers are produced when an energetic carrier, sun electron, transfers enough of its energy to bound ca

o free them. Thus, by creating a single high-energy elene may ultimately produce two or more electrons with low

d

l-

se-f

ns

-

-

ss

oncept.Related areintermediate bandgap solar cells [179]. These

tructures can also used quantum dots of binary or ternaryonductors to create and intermediate band that allows thesting of a much larger portion of the available solar spectheoretically, efficiencies are about 3/4 those for quantumtructures (Table 7). The structure inFig. 24a uses the quaum dots sandwiched in an intrinsic portion between twoells[180]. These quantum dots (Fig. 24b) form the intermediatand of discrete states that allow sub-bandgap photon enere absorbed. The extracted carriers are limited by the banhe energy states of the quantum dots forming the intermeandgap can be controlled by the size of the dot. Several gave been investigating structures[183–186], but no devicesing this approach have yet been validated.


• Research concepts and technology (requireresearch creativity, patience and funding toprove or develop; government funding sourcesare typically shortsighted).

• Industry (needs firm proof that these are signif-icantly “better”—cost and performance; capac-ity or experience curve penetration needs to beaccelerated).

• Still small research efforts that need growth.• Very high risk (though very high potential pay-

off).


Fig. 23. Micrographs of quantum-dot structures used in “quantum dot”, “intermediate band” and other nanotechnology-based solar cells. The dimensions of thesenanostructure provide the ability to tune the device to the incoming radiation.

• “Over-promising” inhibits adoption of currenttechnologies (“waiting for solar godoh”).

3. Module technologies

In the business of photovoltaics, the module usually repre-sents the first-level selling component of the PV system[187]. (Itis currently also the single, most expensive part of the PV system,representing between 50 and 60% of the cost. That is also the rea-

son that worldwide R&D is centered on it because it representsthe biggest and nearest potential in lowering the total systemcost for the consumer[4–6].) It is the configuration of cells,electrically connected into series and parallel strings (either sep-arately, as for crystalline Si, or process-integrated, as for thinfilms) to deliver a desired voltage and current—and encapsu-lated into a supporting structure for environmental protectionand strength. A summary of the headline modules reported todate are presented inTable 5. The module construction deter-mines not only its cost, but also its reliability. Therefore, moduleissues are important because them directly impact the perfor-mance, lifetime, and cost of the PV technology. This packageis the first line of defense in protecting the cells and the circuit


Fig. 24. Intermediate band cell: (a) band diagram and (b) device cross-sectio

structures for the 25–30 year lifetimes that are now expecteof this PV technology[11,188]. It is important to understandthat “packaging” is a major (perhaps monumental!) step to takefollowing the delivery of an efficient solar device to the produc-tion of this first-line commercial product. One that provides theintegrity for ensuring performance for 30 years or more! R&Dis as important here—especially for the emerging technologiewhich have more complexities and sensitivities than imposed bcurrent crystalline Si approaches.

Several mis-steps have impeded module development in thpast. First, it was only considered as a mechanical supportfor the expensive and high-tech converters that do the “reawork” of photovoltaics. These components have numerous inti

mate interfaces among metals, insulators, and semiconductors,energy transfer locations, varying electrical fields, and large sur-face areas that require micro- as well as macro-area design tomeet the demands of operation in varying ambient conditions.The module is more: it is a composite optical and electricaland mechanical structure that has required the collaborating andoverlapping knowledge of physics, chemistry, materials science,and engineering to ensure its viability.Second, it is an evolvingarea of photovoltaic technology and industry development; onethat has to adapt, redesign, and reconfigure with the progress ofthe various cell approaches already discussed[189]. The mostdeveloped module concepts are for crystalline Si, and the mis-take was made at first just to use these for the developing thinfilms. Different technologies present different challenges. Thinfilms have extremely high surface-to-volume ratios than the bulktechnologies. They also use materials that have different sensi-tivities to the ambient (such as water, water vapor, other gases).Thus, ingress issues are considerably different—more criticalto component lifetime and durability. Much was learned fromthe a-Si:H experience[190–192], which was a radical changeto existing module approaches—presenting areas for shunt-ing (bridges, contact openings, electromigration, interdiffusion,delamination, and microdefects that affect macroscale electri-cal behavior[193]. Technologies that incorporate liquids (dyecells) and those that operate at higher temperatures (concentra-tors) have added orders of difficulty in packaging design and

rtingur-iteria

oper-kW,o areource

powerof

nstal-eardslogy.

ss isatelyastric-

ar PVenesss thatd,any.

c andedssneedline

n.

d

sy

e

l-

integrity.International standards for testing modules and repo

performance are in place[194]. However, many of these are crently being re-evaluated to determine if these reporting crmeet the needs of the installers and consumers to ensureation in the real world. Energy ratings (annual kWh, kWh/etc.) are important for installation designs and use. More sthe ambient parameters—such as the available solar resand temperature. The current standard for measuring theof a module at a 25◦C temperature provides a standardcomparison—but does not meet some requirements for ilations which commonly operate at 40, 50◦C, or even abov70◦C. This remains a concern for those working in the standarea—as well as for the installers and users of this techno

4. Summary and future

Currently, photovoltaics as a technology and a businecomposed a complex network of co-dependent and intimrelatedtipping points [1]. First, it is a real business that hreached $10B levels: it is clearly the fastest growing elecity source over the past year and past five years. But, solneeds attention to government policy and consumer awarand acceptance to take it to its next levels—those pushewill make it “spread like wildfire” in markets around the worlgrowing the bonfires that have been lit in Japan and GermThese have shown technology worth, as well as economiemployment value. Policy is important, but the wildfire neadditional and new fuels to make it endure.Second, solar PV hato tip to its next stages of technology development—this thefor R&D to improve now and near technologies in crystal


Si and thin films—and to develop the next generations that willfuel the wildfire of business and deployment. This investmentin R&D is essential to bringing down costs, as well as to ensur-ing our next generations of consumers have technologies readyto meet the mounting demands for energy in this 21st century.Photovoltaics has advanced incredibly from its Bell Laborato-ries beginnings in 1954—the next decade will likely produce50 times more technically than that first half century. It has thepotential to grow as an energy resource 50 times more. However,this is one of true “intelligent design”. We have to provide thetechnical expertise, the resources, the creativity and innovation,and the belief—and solar photovoltaics will be significant in ourclean-energy future.

Acknowledgements

The author expresses sincere gratitude and appreciation tocolleagues with the National Center for Photovoltaics at theNational Renewable Energy Laboratory, who help in review-ing this material. Special thanks goes to Keith Emery of NRELfor his counsel and sharing his warehouse of knowledge on PVperformance and characterization. This paper represents primar-ily the thoughts, insights, and observations of the author, basedupon his some 40 years in PV R&D. This paper was preparedpartially through the support of the U.S. Department of Energy

,

6.nain

d:ts

ferIP

l98

n-

lar

ort

ies

st

ap

1st

[13] M. Schmela, Photon International: The Photovoltaic Magazine, May2005, Solar Verlag GmbH, Aachen, Germany, 2005, pp. 24–35.

[14] M. Rogol, “Investing in Solar Power: Hot Spots and Shadows”, Pre-sented at the U.S. DOE Solar Energy Technologies Review Meeting,Denver, Denver, Colorado, October 2004 (this is unpublished and notavailable in the proceedings from this meeting, but it is availabledirectly from the presenter:[email protected]).

[15] W.P. Mulligan, D.H. Rose, M.J. Cudzinovic, D.M. De Decuester, K.R.McIntosh, R. Swanson, Proceedings of the 19th European PhotovoltaicsSolar Energy Conference Paris, WIP, Munich, 2004, pp. 387–390.

[16] M. Green, Solar Cells: Operating Principles, Technology and SystemApplications, Prentice Hall, Englewood Cliffs, NJ, 1982.

[17] M.A. Green, A.W. Blakers, S.R. Wenham, S. Narayanan, M.R. Willi-son, M. Taouk, T. Szpitalak, Proceedings of the 18th IEEE PhotovoltaicSpecialists Conference, IEEE, New York, 1986, pp. 39–42.

[18] M.A. Green, A.W. Blakers, S. Jiqun, E.M. Keller, S.R. Wenham,R.B. Godfrey, T. Szpitalak, M.R. Willison, Proceedings of the 17thIEEE Photovoltaic Specialists Conference, IEEE, New York, 1984, pp.386–389;Also I. Zhao, A. Wang, P. Altmatt, M.A. Green, Appl. Phys. Lett. 62(1987) 243.

[19] C.M. Chong, F. Zhang, A. Sproul, S.R. Wenham, M.A. Green, Pro-ceedings of the 20th IEEE Photovoltaic Specialists Conference, IEEE,New York, 1988, pp. 529–531.

[20] A.U. Ebong, S.H. Lee, C.B. Honsberg, S.R. Wenham, Proceedings ofthe 25th IEEE Photovoltaic Specialists Conference, IEEE, New York,1996, pp. 513–517.

[21] R. Swanson, Solar Cells 17 (1986) 85.[22] H. Eguren, J. del Alamo, A. Cuevas, A. Luque, Proceedings of the 15th

IEEE Photovoltaic Specialists Conference, IEEE, New York, 1981, pp.1343–1348;

hoto-251.

nergy

New

enver,

nergy.tentld

ata,ltaic866–

hoto-pp.

D.oltaic, pp.

24th, pp.

983)

r, A.theWIP,

-,

4

Also T. Warabissako, A. Luque, Proceedings of the 23rd IEEE Pvoltaic Specialists Conference, IEEE, New York, 1993, pp. 248–

[23] Crystalline Silicon Initiative, SEIA, Washington, DC (www.seia.org).[24] R. Swanson, Proceedings of the 20th European PV Solar E

Conference-Barcelona, James and James, UK, in press.[25] R. Swanson, Proceedings of the 31st IEEE PVSC-Orlando, IEEE,

York, in press.[26] R. Swanson, Proceedings of the DOE Solar Program Review, D

NREL, Golden, CO, 2005.[27] R. Swanson, Proceedings of the European Photovoltaics Solar E

Conference-Paris, James and James, UK, 2004, pp. 1078–1082[28] See, www.bp.com/genericarticle.do?categoryld=3050557&con

=3060005.[29] M. Taguchi, M. Tanhaka, E. Maruyama, S. Kiyama, H. Sak

Y. Yoshime, A. Terakawa, Proceedings of the IEEE PhotovoSpecialists Conference-Orlando, IEEE, New York, 2005, pp.871.

[30] See, for example T. Jester, Proceedings of the 31st IEEE Pvoltaic Specialists Conference-Orlando, IEEE, New York, 2005,1667–1770, Also see Ref. 188.

[31] T. Ciszek, J. Cryst. Growth 66 (1984) 655.[32] M. Acciarri, S. Binetti, C. Savigni, S. Pizzini, F. Feerrazza,

Margadonna, Proceedings of the 13th European Solar PhotovEnergy Conference, H.S. Stephens & Assoc., Bedford, UK, 19951336–1340.

[33] C.P. Khattak, F. Schmid, W.K. Schubert, Proceedings of theIEEE Photovoltaic Specialists Conference, IEEE, New York, 19951351–1355.

[34] J.P. Kalejs, J. Cryst. Growth 104 (1990) 14.[35] J.J. Hanoka, Solar Energy Mater. Solar Cells 65 (2001) 231.[36] C. Dube, H.I. Hanoka, B. Sandstrom, Appl. Phys. Lett. 44 (1

425.[37] R.G. Sidendsticker, J. Cryst. Growth 39 (1977) 17.[38] O. Schultz, S.W. Gluntz, J.C. Goldschmidt, H. Lautenschlage

Leimenstolt, E. Schneiderlochner, G.P. Willeke, Proceedings of19th European Photovoltaic Solar Energy Conference-Paris,Munich, 2004, pp. 604–607.

under Contract No. DE-AC36-98-GO10337.

References

[1] M. Gladwell, The Tipping Point, Little, Brown and Company, Inc.New York, 2000.

[2] D.M. Chapin, C.S. Fuller, G.L. Pearson, J. Appl. Phys. 25 (1954) 67[3] Strategies Unlimited, Global Analysis of PV Markets and Applicatio

Forecast, Report PM-54, June 2004, Strategies Unlimited, MountView, California, 2004–2005.

[4] The U.S. PV Industry Roadmap Through 2030 and Beyonwww.seia.org; See, Proceedings of the IEEE Photovoltaic SpecialisConferences, IEEE, New York, 1960 (present).

[5] EPIA Roadmap:www.epia.org/05Publications/EPIAPublications.htm;See, Proceedings of the European Photovoltaic Solar Energy Conences, D. Reidel Publ., Kluwer Publ., James and James Ltd., and W1977 (present).

[6] PV Roadmap 2030:www.nedo.go.jp/englishy/archives/161027.htm;See, Proceedings of the Photovoltaic Solar Energy Conferences, 1(present).

[7] P.D. Maycock, PV News, vol. 24, no. 3, PV Energy Systems, Warreton, VA, March 2005.

[8] M. Schmela, Photon International: The Photovoltaic Magazine, SoVerlag GmbH, Aachen, Germany, March 2005, pp. 66–82.

[9] Strategies Unlimited, Manufacturer Shipments 2004/2005, RepPM-57, April 2005, Strategies Unlimited, Mountain View, CA,2005. See also, M. Rogol, Sun Screen: Investment Opportunitin Solar Power, (available athttp://photon-magazine.com). Also, see“Solar Power Through 2015” posted onhttp://esd.mit.edu/esdreports/summer2005/solarpower.html.

[10] Strategies Unlimited, Photovoltaic Five-Year Market Foreca2005–2010, Strategies Unlimited, Moutain View, CA, 2005.

[11] Our Solar Power Future; The U.S. Photovoltaics Industry RoadmThrough 2030 and Beyond, SEIA, Washington, DC, October 2004.

[12] R.W. Birkmire, L.L. Kazmerski, in: V.K. Kapur, R.D. McConnell, D.Carlson, G.P. Caesar, A. Rohatgi (Eds.), Photovoltaics for the 2Century, Electrochem. Soc., NJ, 2000, pp. 274–399.

http://www.seia.org/

http://www.epia.org/05publications/epiapublications.htm

http://www.nedo.go.jp/englishy/archives/161027.html

http://photon-magazine.com/

http://esd.mit.edu/esd_reports/summer2005/solar_power.html

http://esd.mit.edu/esd_reports/summer2005/solar_power.html

mailto:[email protected]

http://www.seia.org/


[39] N. Garayanan, J. Zolper, F. Yun, S.R. Wenham, A.B. Sproul, C.M.Chong, M.A. Green, Proceedings of the 21st IEEE Photovoltaic Spe-cialists Conference, IEEE, New York, 1990, pp. 678–680.

[40] A. Rohatgi, S. Narasimha, S. Kamra, P. Doshi, C.P. Khattak, K. Emery,H. Field, Proceedings of the 25th IEEE Photovoltaic Specialists Con-ference, IEEE, New York, 1996, pp. 741–744.

[41] B. Sopori, J. Electron. Mater. 32 (2003) 1034.[42] C.E. Dube, J. Hanoka, Proceedings of the IEEE Photovoltaic Special-

ists Conference, IEEE, New York, 2005, pp. 883–888.[43] B. Sopori, Y. Zhang, R. Reedy, K. Jones, Y. Yan, M. Al-Jassim, B.

Bathey, J. Kaleijs, Proceedings of the 31st IEEE Photovoltaic Special-ists Conference-Orlando, IEEE, New York, 2005, pp. 1039–1042.

[44] See, www.sunpowercorp.com/html/Technical%20Papers/pdf/PVSC00-MCCellPaper.pdf; www.earthlsci.org/energy/solar/page3html;www.oksolar.com/solarpanels/evergreen.htm;www.aist.go.jp/aiste/aisttoday/200413/feature2/feature10.html;See J.H. Sohlegemuth, M. Narayanan, R. Clark, T. Koval, Ju. CFrea-ger, S. Rowcin, M. Benett, D. Cunningham, D. Amin, Proceedingsof the 31st IEEE Photovoltaic Specialists Conference-Orlando, IEEE,New York, 2005, pp. 1023–1027;A. Mueller, M. Reinecke, A. Bachmann, C. Seifert, T. Trommer,A. Schindler, Proceedings of the 31st IEEE Photovoltaic SpecialistsConference-Orlando, IEEE, New York, 2005, pp. 1019–1022.

[45] K.J. Weber, A.W. Blakers, M.J. Stocks, J.H. Babaei, Va.A. Evertt, A.J.Neuendorf, P.J. Verlinden, Electron. Device Lett. 25 (2004) 27;Also A.W. Blakers, K. Weber, S. Deenapanray, V. Everett, E. Franklin,Proceedings of the 31st IEEE Photovoltaic Specialists Conference-Orlando, IEEE, New York, 2005, pp. 991–994.

[46] D.C. Reynolds, G. Leies, L.L. Antes, R.E. Marburger, Phys. Rev. 96(1954) 533.

[47] D.E. Carlson, C.R. Wronski, Appl. Phys. Lett. 28 (1976) 671;

g, J.

EE, pp.

1.. Li,etti,ro-

, pp.

lski,Sola, pp.

ed-EEE,

of, WIP

ner,.Sola

ingsParis

oto-, pp.

[60] S. Guha, Proceedings of the 31st IEEE Photovoltaic SpecialistsConference-Orlando, IEEE, New York, 2005, pp. 12–16.

[61] D. Fischer, S. Dubail, J.A. Anna Selvan, N. Pilaton Vaucher, R. Platz,Ch. Hof, U. Kroll, J. Meier, P. Torres, H. Keppner, N. Wyrsch, M.Goetz, A. Shah, K.-D. Ufert, Proceedings of the IEEE PhotovoltaicsSpecialists Conference, IEEE, NY, 1996, pp. 1053–1056.

[62] D. Domine, J. Steinhauser, L. Feitknecht, A. Shah, C. Ballif, Pro-ceedings of the 20th European Photovoltaic Solar Energy Conference-Barcelona, James and James, UK, in press.

[63] A.V. Shah, E. Vallat-Sauvain, J. Meier, M. Vanecek, S. Fay, V. Terraz-zoni, Ju. Guillet, L. Feitknecht, J. Bailat, C. Ballif, I. Pola, Proceedingsof the IEEE Photovoltaic Specialists Conference-Orlando, IEEE, NewYork, 2005, pp. 1353–1358.

[64] J. Yang, S. Guha, B. Yah, G. Yue, Proceedings of the IEEE Photo-voltaic Specialists Conference-Orlando, IEEE, New York, 2005, pp.1359–1364.

[65] M. Konagai, S. Miyamjima, Y. Yashiki, T. Watahiki, K.L. Narayanan,A. Yamada, Proceedings of the IEEE Photovoltaic SpecialistsConference-Orlando, IEEE, New York, 2005, pp. 1424–1427.

[66] J. Bailat, Proceedings of the 19th Euorpean Photovoltaic Solar EnergyConference-Paris, WIP, Munich, 2005, pp. 1541–1545.

[67] A. Shah, J. Non-Cryst. Solids 338–340 (2002) 639.[68] M.K. van Veen, R.E.I. Schropp, J. Appl. Phys. 93 (2003) 121.[69] www.kaneka.com/siliconpv.html.[70] K. Yamamoto, M. Yoshimi, T. Suzuki, T. Nakata, T. Sawada, A.

Nakajima, K. Hayashi, Proceedings of the 28th IEEE PhotovoltaicSpecialists Conference-Anchorage, IEEE, New York, 2000, pp. 1428–1431.

[71] T.L. Chu, S.S. Chu, K.Y. Duh, H.I. Yoo, IEEE Trans. Electron Dev.ED-24 (1977) 438;A.K. Ghosh, C. Fisherman, T. Fang, J. Appl. Phys. 50 (1979) 3454;

. in–125.orp,

onfer-

gs ofork,

s ofsion-

nergy

lsh,O.oux,ence-

04)

edingsYork,

gs ofelona,

peanh, in

oltaic938.

ialists

S.R. Ovshinsky, J. Non-Cryst. Solids 2 (1970) 99;W.E. Spear, P.G. LeComber, Philos. Mag. 33 (1976) 935;Also, Appl. Phys. Lett. 28 (1976) 671.

[49] D.L. Staebler, C.R. Wronski, Appl. Phys. Lett. 31 (1977) 292.[50] H. Branz, Phys. Rev. B59 (1999) 5498.[51] A.H. Mahan, J. Carapella, B. Nelson, R.S. Crandall, I. Balber

Appl. Phys. 69 (1991) 6728.[52] M.S. Bennett, K. Rajan, K. Kritikson, Proceedings of the 23rd IE

Photovoltaic Specialists Conference PC IEEE, New York, 1993850–856;Y. Hamakawa, T. McMahon (Eds.), Solar Cells, vol. 9, 1993, p.

[53] D.E. Carlson, R.R. Arya, M. Bennett, L.-F. Chen, K. Jansen, Y.-MN. Malley, J. Morris, J. Newton, R.S. Oswald, K. Rajan, D. Vezzr. Willing, L. Yang, Proceedings of the 13 NREL. Photovolaics Pgram Review, AIP Conference Series 353, AIP, New York, 1995304–311;Also, Q. Wang, Y. Xu, R.S. Crandall, pp. 473–480.

[54] A.E. Delahoy, T. Tonon, J.A. Cambridge, M. Johnson, L. MichaF.J. Kampas, Proceedings of the Eighth European PhotovoltaicEnergy Conference, Kluwer Acad Publ, The Netherlands, 1988646–652.

[55] V.L. Dalal, J. Leib, K. Muthukrishanan, D. Stieler, M. Noack, Proceings of the IEEE Photovoltaic Specialists Conference-Orlando, INew York, 2005, p. 1448.

[56] J.A. Anna Selvan, Y.-M. Li, S. Guo, A. Delahoy, Proceedingsthe 19th European Photovoltaics Solar Energy Conference-ParisMunich, 2004, pp. 1346–1351.

[57] J. Meier, U. Kroll, J. Spitznagel, S. Benagli, T. Roschek, G. PfanCh. Ellert, G. Androutsopoulos, A. Hughi, G. Buchel, D. Plesa, ABuchel, B. Rech, Proceedings of the 19th European PhotovoltaicsEnergy Conference-Paris, WIP, Munich, 2004, pp. 1328–1333.

[58] Y. Mai, S. Klein, J. Solff, A. Lambertz, X. Geng, F. Finger, Proceedof the 19th European Photovoltaics Solar Energy Conference-WIP, Munich, 2004, pp. 1399–1402.

[59] D.E. Carlson, G. Lin, G. Ganguly, Proceedings of the IEEE Phvoltaics Specialist Conference-Anchorage, IEEE, New York, 2000707–712.

r

,

r

,

Also, Appl. Phys. Lett. 25 (1974) 283.[72] J.H. Werner, R. Bergmann, R. Brendel, in: R. Helbig (Ed.), Adv

Solid State Phys., vol. 34, Vieweg, Braunsschweig, 1994, pp. 115[73] M.A. Green, A. Wang, J. Zhao, S.R. Wenham, P. Campbell, D. Th

Proceedings of the 13th European Photovoltaics Solar Energy Cence, H.S. Stephens & Assoc., Bedford, UK, 1995, pp. 13–16.

[74] G.F. Zheng, A.B. Sproul, S.R. Wenham, M.A. Green, Proceedinthe 25th IEEE Photovoltaic Specialists Conference, IEEE, New Y1996, pp. 465–468.

[75] A.B. Aberle, P.I. Widenborg, A. Straub, N.-P. Harder, Proceedingthe 3rd World Conference on Photovoltaic Solar Energy ConverOsaka, IEEE, New York, 2003, pp. 1194–1197.

[76] B. Rech, Proceedings of the 20th European Photovoltaic Solar EConference-Barcelona, James and James, UK, in press.

[77] A.G. Aberle, P.I. Widenborg, D. Song, A. Straub, M.L. Terry, T. WaA.B. Sproul, P. Campbell, D. Inns, B. Beilby, M. Griffin, Y. Huang,Kunz, F. Martin-Brune, S.R. Wenham, J. Weber, R. Begs, V. BarrProceedings of the 31st IEEE Photovoltaic Specialists ConferOrlando, IEEE, New York, 2005, pp. 877–882;S. Reber, A. Hurrle, A. Eyer, G. Willeke, Solar Energy 77 (20865.

[78] A. Slaoui, E. Pihan, A. Focsa, G. Beaucarne, J. Poortmans, Proceof the 31st IEEE Photovoltaic Specialists Conference, IEEE, New2005, pp. 1127–1130;R. Reber, A. Eyer, F. Haas, N. Schillinger, S. Janz, Proceedinthe 20th European Photovoltaics Solar Energy Conference-BarcWIP, Munich, in press;R. Auer, B. Gazuz, M. Schulz, Proceedings of the 20th EuroPhotovoltaics Solar Energy Conference-Barcelona, WIP, Municpress.

[79] P.A. Basore, Proceedings of the 3rd World Conference on PhotovSolar Energy Conversion-Osaka, IEEE, New York, 2003, pp. 935–

[80] www.q-cells.com.[81] P.A. Basore, Proceedings of the 31st IEEE Photovoltaic Spec

Conference-Orlando, IEEE, New York, 2005, pp. 9967–9970.[82] M.A. Green, Solar Energy 77 (2004) 857.[83] www.csgsolar.com/downloads.html.


[84] J.L. Shay, J.H. Wernick, Ternary Chalcopyrite Semiconductors:Growth, Electronic Properties and Applications, Pergamon Press, NewYork, 1975.

[85] S. Wagner, J.L. Shay, P. Migliorato, H.M. Kasper, Appl. Phys Lett. 25(1974) 434.

[86] L.L. Kazmerski, M.S. Ayyagari, G.A. Sanborn, F.R. White, J. Vac. Sci.Technol. 13 (1976) 139;Also L.L. Kazmerski, in: G. Holah (Ed.), Ternary Compounds 1977,Institute of Physics, London, 1977, pp. 217–228.

[87] Some recent review of CIGS.[88] M.A. Contreras, B. Egaas, P. Dippo, J. Webb, J. Granata, K.

Ramanathan, S. Asher, A. Swartzlander, R. Noufi, Proceedings ofthe 26th IEEE Photovoltaic Specialists Conference, IEEE, New York,1997, pp. 259–262.

[89] J. Sites, Proceedings of the International Conference on Ternary andMultinary Compounds (ICTMC-2004), Denver, Colorado, J. Chem.Phys. Solids, in press.

[90] A. Zunger, Proceedings of the International Conference on Ternaryand Multinary Compounds (ICTMC-2004), Denver, Colorado, J. Chem.Phys. Solids, in press.

[91] R. Noufi, Proceedings of the International Conference on Ternatry andMultinary Compounds (ICTMC-2004), Denver, Colorado, J. Chem.Phys. Solids, in press.

[92] See, for example L.L. Kazmerski, Renewable Sustainable Energy Rev.1 (1997) 71.

[93] T.M. Friedimeier, D. Braunger, D. Hariskos, M. kKaiser, H.N. Wanka,H.W-. Wchock, Proceedings of the 25th IEEE Photovoltaic SpecialistsConference, IEEE, New York, 1996, pp. 845–848;also J. Kessler, Sixth International Photovolaic Science and Engineer-ing Conference, 1992, pp. 1010–2005.

[94] M. Ruckh, D. Hariskos, U. Ruhle, H.W. Schock, Proceedings ofork,

oto-884;, R.Ngs ofork,

oung

og.

pan,Spe-2.H.-pean, pp.

, R.ecial-

orld

31stork,

ceedence-

ilms

er,oltaic219–

[104] T. Enzenhofer, T. Unold, R. Scheer, H.-W. Schock, Proceedings ofthe 20th European Photovoltaics Solar Energy Conference-Barcelona,WIP, Munich, in press.

[105] M. Kharchenko, B. Book, O. Chernykh, Proceedings of the 20th Euro-pean Photovoltaic Solar Energy Conference-Barcelona, WIP, Munich,in press.

[106] V. Nadenau, U. Rau, A. Jasenek, H.W. Schock, J. Appl. Phys. 87(2000) 584;D.I. Young, J. Keane, A. Duda, J. Abusarma, C.L. Perkins, M. Romero,R. Noufi, Prog. Photovoltaics 11 (2003) 535;J.H. Yun, R.B.V. Chalapathy, S.K. Kim, J.C. Lee, J. Song, B.T. Ahn,K.H. Yoon, Proceedings of the 31st IEEE Photovoltaic SpecialistsConference-Orlando, IEEE, New York, 2005, pp. 445–448;J.A. AbuShama, R. Noufi, S. Johnston, S. Ward, X. Wu, Proceedings ofthe 31st IEEE Photovoltaic Specialists Conference, IEEE, New York,2005, pp. 299–301.

[107] www.compoundsemiconductor.net/articles/magazine/11/5/5/1.[108] www.radiochemistry.org/periodictable/elements/49.html.[109] K. Zanio, Cadmium Telluride, Academic Press, New York, 1975.[110] See,www.firstsolar.comand www.antecsolar.de.[111] D. Bonnet, Proceedings of the 19th European Photovoltaic Solar

Energy Conference-Paris, James and James, UK, 2004, pp. 1657–1662.[113] B.E. McCandless, S.S. Hegedus, Proceedings of the 22nd IEEE Pho-

tovoltaic Specialists Conference, IEEE, New York, 1991, pp. 967–972.[114] X. Wu, P. Sheldon, Proceedings of the 16th European Photovoltaic

Solar Energy Conference, James and James, UK, 2000, pp. 341–344;Also X. Wu, P. Sheldon, Y. Mahathongdy, R. Ribelin, A. Mason, H.Moutinho, T.J. Coutts, Proceedings of the 15th NREL/SNL PV Pro-gram Review Meeting, Denver, NREL, Golden, CO, 1998, pp. 37–41.

[115] S.H. Demtsu, J.R. Sites, Proceedings of the 31st IEEE PhotovoltaicSpecialists Conference-Orlando, IEEE, New York, 2005, pp. 347–350,

ou,29th, pp.

da,17th

James,

EEE005,

gs ofYork,

31st, pp.

s ofYork,

31st, pp.

of the005,

gies

nce-

kins,ergy5.

the 25th IEEE Photovoltaic Specialists Conference, IEEE, New Y1996, pp. 835–838;Ellmer, R. Wendt, R. Cebulla, Proceedings of the 25th IEEE Phvoltaic Specialists Conference, IEEE, New York, 1996, pp. 881–J.R. Tuttle, T.A. Berens, J. Keane, K.R. Ramanathan, J. BranataBattachararya, H. Wiesner, M.A. Contreras, R. Noufi, Proceedinthe 25th IEEE Photovoltaic Specialists Conference, IEEE, New Y1996, pp. 797–800.

[95] M.A. Contreras, K. Ramanathan, J. AbuShama, F. Hasoon, D.L. YB. Egaas, R. Noufi, Prog. Photovoltaics 13 (2005) 209.

[96] See M.A. Green, K. Emery, D.L. King, S. Igari, W. Warta, PrPhotovoltaics 13 (2005) 49.

[97] R. Wieting, R. Gay, H. Nguyen, J. Palm, R. Rischmiller, A. SeaD. Tarrant, D. Willett, Proceedings of the 31st IEEE Photovoltaiccialists Conference-Orlando, IEEE, New York, 2005, pp. 177–18

[98] M. Powalla, B. Dimmler, R. Schaeffler, G. Voorwinden, U. Stein,D. Mohring, F. Kessler, D. Hariskos, Proceedings of the 19th EuroPhotovoltaics Solar Energy Conference-Paris, WIP, Munich, 20041663–1667.

[99] M.E. Beck, S. Wiedeman, R. Huntington, J. VanAlsburg, E. KantoButche, J.S. Britt, Proceedings of the 31st IEEE Photovoltaic Spists Conference-Orlando, IEEE, New York, 2005, pp. 211–214;Also, 2004 R&D 100 Award (R&D Magazine).

[100] D. Tarrant, R. Gay, V. Probst, F. Karg, Proceedings of the 3rd WConference on PV Energy Conversion, IEEE, New York, 2003;Also B. Dimmler, M. Powalla, R. Schaetter, Proceedings of theIEEE Photovoltaic Specialists Conference-Orlando, IEEE, New Y2005, pp. 189–194.

[101] K. Ramanathan, J.C. Keane, B. To, R.G. Dhere, R. Noufi, Proings of the 20th European Photovoltaics Solar Energy ConferBarcelona, WIP, Munich, in press.

[102] K. Ramanathan, G. Teeter, J.C. Keane, R. Noufi, Thin Solid F480–481 (2005) 499.

[103] N. Meyer, I. Luck, U. Ruehle, J. Klaer, R. Klenk, M. Ch. Lux-SteinR. Scheer, H.-W. Schock, Proceedings of the 31st IEEE PhotovSpecialists Conference-Orlando, IEEE, New York, 2005, pp.222.

.

,

-

See also, Ref. 116.[116] X. Wu, Solar Energy 77 (2004) 803.[117] X. Wu, R.G. Dhere, Y. Yan, M.J. Romero, Y. Zhang, J. Zh

C. DeHart, A. Duyda, C. Perkins, B. To, Proceedings of theIEEE Photovoltaic Specialists Conference, IEEE, New York, 2002531–535;Also, Phys. Stat. Sol. 1062 (2003).

[118] X. Wu, J.C. Keane, R.G. Dhere, C. DeHart, D.S. Albin, A. DuT.A. Gessert, S. Asher, D.H. Levi, P. Sheldon, Proceedings of theEuropean Photovoltaics Solar Energy Conference, James andUK, 2001, pp. 995–1000.

[119] B. von Roedern, H.S. Ullal, K. Zweibel, Proceedings of the 31st IPhotovoltaic Specialists Conference-Orlando, IEEE, New York, 2pp. 183–188.

[120] S.S. Hegedus, D. Desa, D. Ryan, B.E. McCandless, Proceedinthe 31st IEEE Photovoltaic Specialists Conference, IEEE, New2005, pp. 319–322;K.L. Barth, R.A. Enzenrothy, W.S. Sampath, Proceedings of theIEEE Photovoltaic Specialists Conference, IEEE, New York, 2005323–326;T.J. McMahon, T. Berniard, D. Albin, S.H. Demtsu, Proceedingthe 31st IEEE Photovoltaic Specialists Conference, IEEE, New2005, pp. 363–366;V.G. Karpov, D. Shvydka, Y. Roussilon, Proceedings of theIEEE Photovoltaic Specialists Conference, IEEE, New York, 2005437–440;J.H. Scofield, P.S. Cohen, S. Gould, C. Tronson, Proceedings31st IEEE Photovoltaic Specialists Conference, IEEE, New York, 2pp. 1623–1626.

[121] www.nrel.gov.[122] L.L. Kazmerski, Proceedings of the DOE Solar Energy Technolo

Program Review, Denver, NREL, Golden, CO, 2004;Also, Proceedings of the IEEE Photovoltaic Specialists ConfereNew Orleans, IEEE, New York, 2003.

[123] P.J. Verlinden, R.M. Swanson, R.A. Crane, K. Wickham, J. PerProceedings of the 13th European Photovoltaic Solar EnConference-Nice, Kluwer, The Netherlands, 1995, pp. 1582–158

http://www.firstsolar.com/

http://www.antecsolar.de/


[124] M. Yamaguchi, Semiconductors 33 (1999) 961;T.V. Torchynska, G. Polupan, Superfieies y Vacio 17 (2004) 22;K. Takahashi, S. Yamada, T. Unno, Hitachi Reviews, 523.9-7 (2004)1 (www.hitachi-cable.co.jp/en/review/17/review07.pdf).

[125] M.W. Wanlass, J.S. Ward, K.A. Emery, T.A. Gessert, C.R. Osterwa.d,T.J. Coutts, Solar Cells 30 (1990) 363.

[126] L.M. Fraas, J.E. Avery, V.S. Sundaram, V.T. Kinh, T.M. Davenport,J.W. Yerkes, J.M. Gee, K.A. Emery, Proceedings of the 21st IEEE Pho-tovoltaic Specialists Conference-Kissimmee, IEEE, New York, 1990,pp. 190–195.

[127] M.W. Wanlass, T.J. Coutts, J.S. Ward, K.A. Emery, T.A. Gessert, C.R.Osterwald, Proceedings of the 22nd IEEE Photovoltaic Specialists Con-ference, IEEE, New York, 1992, pp. 38–45.

[128] Other.[129] R. Surapudi, T. Hamilton, D. Rapp, P. Stella, N. Mardesich, J. Mondt,

B. Nesmith, R.L. Bunker, J. Cutts, S.G. Bailey, H.B. Curtis, M. Piszc-zor, E. Gaddy, D. Marvin, L.L. Kazmerski, Solar Cell and ArrayTechnology for Future Space Science Missions, JPL D-24454, Spe-cial Report, December 2003.

[130] Summary of concent McConnell.[131] R.R. King, C.M. Fetzer, R.A. Sherif, D.C. Law, D.D. Krut, J.H. Ermer,

N.H. Karam, Proceedings of the European Photovoltaics Solar EnergyConference-Barcelona, WIP, Munich, in press.

[132] O.K. Jani, C.B. Honsberg, A. Asghar, D. Nicol, I. Ferguson, A. Doolit-tle, S. Kurtz, Proceedings of the 31st IEEE Photovoltaic SpecialistsConference-Orlando, IEEE, New York, in press.

[133] W. Aalukiewicz, J. Wu, K.M. Yu, J.W. Ager III, E.E. Haller, J. Jasin-ski, Z. Liliental-Weber, H. Lu, W.J. Schaff, Proceedings of the 19thEuropean Photovoltaics Solar Energy Conference-Paris, WIP, Munich,2004, pp. 30–33;Also, Phys. Rev. B 66 (2002) R201403;

man,A.

ltaic530–

, K.Pro-rs foCD-

lt-#05-

nerg

lPho-ther-

R.ialists

[146] M. Graetzel, Proceedings of the 20th European PhotovoltaicsSolar Energy Conference-Barcelona, James and James, UK, inpress.

[147] M. Freemantle, Chem. Eng. News 80 (December (51)) (2002) 6.[148] S.M. Zakeeruddin, M. Graetzel, Chem. Commun. 24 (2002) 2972.[149] M. Durr, A. Schmid, M. Obermaier, S. Rosselli, A. Yasuda, G. Nelles,

Nat. Mater. 4 (2005) 607.[150] www.dyesol.com (www.sta.com.au).[151] J.R. Waldrop, M.J. Cohen, A.J. Heeger, A.G. MacDiarmid, Appl. Phys.

Lett. 38 (1981) 53.[152] M.J. Cohen, J.S. Harris Jr., Tech. Dig. Intl. Electron Dev. Mtg. (1978)

278;K.P.J. Reucroft, K. Takahashi, H. Ullal, Appl. Phys. Lett. 25 (1974)664;C.W. Tang, A.C. Albrecht, Nature 254 (1975) 507.

[154] P. Peumans, A. Yakimov, S.R. Forrest, J. Appl. Phys 93 (2003) 3693.[155] See, J. Halme, Masters Thesis, Dy-Sensitized Nanostructured

and Organic Photoovoltaic Cells: Technical Review and Pre-liminary Tests, www.tkk.fi/Units/AES/studies/dis/halmeabs.htm;http://solar-club.web.cern.ch/solar-club/SolPV/autres/organics.html;www.sciam.com/article.cfm?chanID=sa003&articleID=0004C094-02CC-1CD0-B4A8809EC588EEDF.

[156] S. Forest, Proceedings of the SPIE Meeting, Denver, SPIE, Bellingham,WA, 2003.

[158] See, for example C. Honsberg, Proceedings of the DOE Solar EnergyTechnologies Program Review, Denver, NREL, Golden, CO, 2004;M.A. Green, Prog. Photovoltaics 9 (2002) 257.

[159] M. Symko-Davies, Proceedings of the 19th European PhotovoltaicSolar Energy Conference-Paris, James and James, UK, 2004;M. Symko-Davies, R. Noufi, T.J. Coutts, Proceedings of the 31st IEEEPhotovoltaic Specialists Conference-Orlando, IEEE, New York, 2005,

oltaicpress.

, M.

gs ofrence-

no,and

05).

.ltaicork

f theElec-180.

EE005,

ia/

hoto-ess.ation,

Pro-rence,

Also, Appl. Phys Lett. 80 (2002) 4741.[134] M. Wanlass, J.F. Geisz, S. Kurtz, R.J. Wehrer, B. Werns

S.P. Ahrenkiel, R.K. Ahrenkiel, D.S. Albin, J.J. Carapella,Duda, T. Moriarty, Proceedings of the 31st IEEE PhotovoSpecialists Conference-Orlando, IEEE, New York, 2005, pp.535.

[135] M. Wanlass, S.P. Ahrenkiel, D.S. Albin, J.J. Carapella, A. DudaEmery, J.F. Geisz, K. Jones, S. Kurtz, T. Moriary, M.J. Romero,ceedings of the International Conference on Solar Concentratothe Generation of Electricity or Hydrogen, Scottsdale, AZ, NREL/520-38172, August 2005, NREL, Golden, CO, in press;M. Wanlass, Approaches for Ultra-High Efficiency, Monolithic, MuBandgap, Tandem solar Photovoltaic Energy Converters, NREL IR05 (patent pending).

[136] www.aps.com/mycommunity/Solar/Solar20.html.[137] www.sunpowercorp.com/html/Resources/concbasic.html, andwww.

energyinnovations.com/and http://solar.sharpusa.com/files/soldow2005products.pdf.

[138] www.wired.com/wired/archive/13.07/solar.html, andwww.futurepundit.com/mt/mt-comments.cgi?entryid=2900andhttp://lawnorder.blogspot.com/1990/01/rooftop-solar-panels-could-finally-see.html.

[139] M. Green, Proceedings of the 17th European Photovoltaic Solar EConference-Glasgow, James and James, UK, 2000.

[140] www.rdg.ac.uk/Biomim/about.htm,;www.bath.ac.uk/∼en2jsr/examples1.htmlwww.treehugger.com/files/2005/04/theelectronna.php;www.californiasolarcenter.org/solareclips/2002.07/20020709-6.htm.

[141] M. Graetzel, A.J. McEvoy, Proceedings of the 12th Europeantovoltaic Solar Energy Conference, Kluwer Acad. Publ., The Nelands, 1994, pp. 579–582.

[142] M. Graetzel, Prog. Photovoltaics 8 (2000) 171.[143] M. Graetzel, Nature 414 (2001) 338.[144] B. O’Reagan, M. Graetzel, Nature 353 (1991) 737.[145] N. Koide, Y. Chiba, A. Islam, R. Komiya, N. Fuke, A. Fukui,

Yamanaka, Proceedings of the 31st IEEE Photovoltaic SpecConference-Orlando, IEEE, New York, 2005, pp. 175–176.

r

y

pp. 410–413.[160] M. Symko-Davies, R. Noufi, Proceedings of the European Photov

Solar Energy Conference-Barcelona, James and James, UK, in[161] MECH STACK BEST.[162] D. Young, J. Keane, A. Duda, J.S.M. Abushama, C. Perkins

Romero, R. Noufi, Prog. Photovoltaics 11 (2003) 535.[163] M. Gossla, W. Shafarman, Thin Solid Films (2004).[164] T.P. Niesen, S. Visbeck, J. Palm, V. Probst, F. KaRG, Proceedin

the 20th European Photovoltaic Solar Energy ConferenceConfeBarcelona, James and James, UK, in press; Also, ref. 163.

[165] H. Yanagi, J. Tate, S. Park, C.-H. Park, D.A. Keszler, M. HiraH. Hosono, Solid State Commun. (2005), See also, Ref. 159160.

[166] M. Price, J. Tate, C.-H. Park, D.A. Keszler, Thin Solid Films (20[167] T.J. Coutts, Renewable Sustainable Energy Rev. 3 (1999) 77.[168] www.jxcrystals.com/ThermoPV.htm.[169] T.J. Coutts, M.C. Fitzgerald, Sci. Am. (September) (1998) 90–95[170] L. Broman, J. Marks, 1st NREL Conference on Thermophotovo

Generation of Electricity, AIP Conference Series 321, AIP, New Y1994, pp. 221–228;L. Broman, K. Jarefors, J. Marks, M. Wanlass, Proceedings oSecond NREL Conference on Thermophotovoltaic Generation oftricity, AIP Conference Series 358, AIP, New York, 1996, pp. 177–

[171] www.grc.nasa.gov/WWW/RT2001/5000/5410bailey1.html.[172] A. Freundlich, A. Alemu, S. Bailey, Proceedings of the 31st IE

Photovoltaic Specialists Conference-Orlando, IEEE, New York, 2pp. 137–140;www.cchem.berkeley.edu/pagrp/, andwww.berkeley.edu/news/medreleases/2000/03/03-03-2000.html;G.J. Conibeer, M.A. Green, Proceedings of the 20th European Pvoltaics Solar Energy Conference-Barcelona, WIP, Munich, in pr

[173] See, for example, Basic Research Needs for Solar Energy UtilizU.S. Department of Energy, Washington, DC, 2005.

[174] M. Neges, K. Schwarzburg, L. Fundlach, L. Toeben, F. Willig,ceedings of the 20th European Photovoltaic Solar Energy ConfeWIP, Munich, in press.

[175] www.sciencenews.org/articles/20040306/note13.asp.

http://www.hitachi-cable.co.jp/en/review/17/review07.pdf

http://www.energyinnovations.com/

http://www.energyinnovations.com/

http://solar.sharpusa.com/files/sol_dow_2005_products.pdf

http://solar.sharpusa.com/files/sol_dow_2005_products.pdf

http://www.futurepundit.com/mt/mt-comments.cgi?entry_id=2900

http://lawnorder.blogspot.com/1990/01/rooftop-solar-panels-could-finally-see.html

http://www.bath.ac.uk/~en2jsr/examples1.html

http://www.treehugger.com/files/2005/04/the_electron_na.php

http://www.californiasolarcenter.org/solareclips/2002.07/20020709-6.html

http://www.sta.com.au/

http://www.tkk.fi/units/aes/studies/dis/halme_abs.htm



[176] A.J. Nozik, NCPV Program Review Meeting, NREL/CP-590-31011,NREL, Golden, CO, 2001;www.nrel.gov/news/press/2005/1805quantumdot.html?print;S.G. Bailey, D. Wilt, S.L. Castro, B.J. Landi, C. Evans, R.P. Raf-faelle, S. Sinharoy, C.W. King, J. Cowen, F. Ernst, Proceedings ofthe 20th European Photovoltaics Solar Energy Conference-Barcelona,WIP, Munich, in press.

[179] A. Luque, A. Marti, Phys. Rev. Lett. 78 (1997) 5014;Also A. Marti, L. Cuadra, A. Luque, Proceedings of the 28th IEEEPhotovoltaic Specialists Conference-Anchorage, IEEE, New York,2000, pp. 940–943.

[180] A. Luque, A. Marti, P. Sahnon, L. Cuadra, C. Tablero, C. Stanley, A.McKee, D. Zhou, R. Konenkamp, R. Bayon, A. Belaidi, J. Alonso,J. Ruiz, J. Fernandez, P. Palacios, N. Lopez, Proceedings of the 29thIEEE Photovoltaics Specialists Conference-New Orleans, IEEE, NewYork, 2002, pp. 1190–1193.

[183] M.A. Green, E.C. Cho, Y. Cho, Y Huang, E. Pink, T. Trupke, A Lin, T.Fangsuwannarak, T. Puzzer, G. Conibeer, R. Corkish, Proceedings ofthe 20th European Photovoltaics Solar Energy Conference-Barcelona,WIP, Munich, in press.

[184] P. Wahnon, P. Palacios, J.J. Fernadez, K Sanchez, Proceedings of the20th European Photovoltaics Solar Energy Conference-Barcelona, WIP,Munich, in press.

[185] N.P. Harder, Proceedings of the 20th European Photovoltaics SolarEnergy Conference-Barcelona, WIP, Munich, in press.

[186] Y. Okada, N. Shiotsuka, H. Komiyama, K. Akhane, N. Ohtani, Pro-ceedings of the 20th European Photovoltaics Solar Energy Conference-Barcelona, WIP, Munich, in press;P. Palacios, P. Wahnon, S. Pizzinato, J.C. Onesa, Proceedings of the20th European Photovoltaics Solar Energy Conference-Barcelona, WIP,Munich, in press;

sser,peanh, in

air-

Fun-

and

/

nical

ThinO,

Also N. Dhere, Proceedings of the IEEE Photovoltaics SpecialistsConference-Orlando, IEEE, New York, 2005.

[190] K. Fukae, C.C. Lim, M. Tamechika, N. Takehara, K. Saito, I. Kajita,E. Kondo, Proceedings of the 25th IEEE Photovoltaic Specialists Con-ference, IEEE, New York, 1996, pp. 1227–1230.

[191] S. Guha, Proceedings of the 25th IEEE Photovoltaic Specialists Con-ference, IEEE, New York, 1996, pp. 1017–1022.

[192] C. Hof, M. Ludi, M. Goetz, D. Fishcher, A. Shah, Proceedings ofthe 25th IEEE Photovoltaic Specialists Conference, IEEE, New York,1996, pp. 1057–1061.

[193] W. Calahan, R. McDonald (Eds.), Flat Plat Solar Array Project, FinalReport, vol. I–VIII, JPL, Pasadena, CA, 1986 (JPL Publ. 86-31).

[194] C.R. Osterwald, R. Hammond, G. Zerlaut, R. D’Aiello, Proceedingsof the 24th IEEE Photovoltaic Specialists Conference, IEEE, NewYork, 1995, pp. 885–888;Also J. Bishop, H. Ossenbrink, Proceedings of the 25th IEEE Photo-voltaic Specialists Conference, IEEE, New York, 1996, pp. 1191–1196;http://www.toolbase.org/tertiaryT.asp?DocumentID=3209&CategoryID=1215, andwww.oja-services.nl/iea-pvps/pv/glossary.htm.

Further reading

[48] Y. Hamakawa, in: T.J. Coutts, J.D. Meakin (Eds.), Current Topics inPhotovoltaics, Academic Press, New York, 1985, pp. 111–169;Also, S. Guha, Proceedings of the 14th NREL Photovoltaics Pro-gram Review, in: C.E. Witt, M. Al-Jassim (Eds.), AIP, New York,1997.

[112] B.E. McCandless, R.W. Birkmire, Solar Cells 31 (1991) 527.[153] D. Wohrle, D. Meissner, Adv. Mater. 3 (1991) 129;

P. Fairley, Technol. Rev. (July/August) (2004) 35–40;. 11

Pho-

, M.

y 6,

Pro-

3rdNew

A. Bessiere, K.W.J. Barnham, I.M. Bailard, D.C. Johnson, M. MaJ.P. Connolly, G. Hill, J.S. Roberts, Proceedings of the 20th EuroPhotovoltaics Solar Energy Conference-Barcelona, WIP, Municpress.

[187] A.F. Williams (Ed.), The Handbook of Photovoltaic Applications, Fmaont Press, Atlanta, 1986;T. Markvart, L. Castaner, Practical Handbook of Photovoltaics:damentals and Applications, Elsevier, Oxford, UK, 2002;A. Luque, S. Hegedus (Eds.), Handbook of Photovoltaic ScienceEngineering, J. Wiley and Sons, New York, 2004;Also, see PVWATTS performance calculator:http://rredc.nrel.govsolar/calculators/PVWATTS/.

[188] U.S. DOE Solar Energy Technologies Program Multiyear TechPlan 2005-2010, U.S. DOE, Washington, DC, 2005.

[189] See, for example, T. McMahon, N. Dhere, Proceedings of theFilm Module Reliability Project, 2003–2005 (NREL, Golden, C2003, 2004, 2005);

C.J. Brabec, N.S. Sariciftci, J.C. Hummelen, Adv. Funct. Mater(2001) 15;P. Ball, Nat. Sci. Update (February) (2001) 9;H. Hoppe, N.S. Sariciftci, J. Mater. Res. 19 (2004) 1924;C. Brabec, V. Dyakonov, J. Parisi, N.S. Sariciftci (Eds.), Organictovoltaics: Concepts and Realization, Springer, UK, 2003.

[157] J. Drechesel, B. Mannig, F. Kozlowski, D. Gebeyehu, A. WernerKoch, K. Leo, M. Pfeiffer, Thin Solid Films 451–452 (2004) 515;D. Meissmer, J. Rostalski, Synth. Met 121 (2001) 1551.

[177] R. Shaller, V. Klimov, Phys. Rev. Lett. 92 (2004) 186601.[178] P. Weiss, Sci. News 165 (2005) 259, also, Science Daily, Ma

2004.[181] M. Romero, Proceedings of the DOE Solar Energy Technologies

gram Review Meeting, Denver, NREL, Golden, CO, 2004.[182] L. Cuadra, A. Marti, N. Lopez, A. Luque, Proceedings of the

World Conference on Solar Energy Conversion-Osaka, IEEE,York, 2003, pp. 3–8.




Solar Photovoltaics Overview 2005- J. El Spec Rel Ph 2006- p 105- Kazmerski

Documents