Nature - Photovoltaics

8/6/2019 Nature - Photovoltaics

http://slidepdf.com/reader/full/nature-photovoltaics 1/20

MaY 2008

Photovoltics





nature photonics | VOL 2 | MAY 2008 277

CONTENTS | TECHNOLOGY FOCUS

Considering how abundant the Sun’senergy is, it is surprising that only 0.1%o the world’s electricity is made using

solar power. Even in Germany, the mostenthusiastic supporter o solar power, thisgure is only 0.2% (page 281). Te reason orthis huge disparity is that at present the costo producing 1 kWh o electricity using solar

power is much higher than using oil, coal ornuclear power. But times are changing.

Te solar-power industry is experiencinggrowth rates that other industries can only dream about. Compound annual growthrates o around 30% until 2013 have beenpredicted by one market research company (page 290), and these predictions match theresults o several key players in the market,with one even reporting an annual salesgrowth o more than 50%. Let’s hope thisgrowth continues because the industry hassome serious catching up to do i it wants tobecome competitive. o become economical

and thereore take a signicant slice o theelectricity market, the price per kWh mustbecome comparable to that o a typical cleancoal power station. Winried Homann,president o the European PhotovoltaicsIndustry Association, believes that, despitehealthy growth, this will still take around20 years (page 292).

So what can the solar-power industry doto speed up this process? New technologiesare helping to make solar cells cheaper,thinner and more ecient, thus bringingdown the cost per kWh. Silicon-basedsolar cells dominate the market now, but

will in the uture lose market share to othertechnologies. Compound semiconductorsystems are reporting record eciencies o more than 40% (page 284), and developerso fexible polymer-based cells claim theirtechnology has what it takes to bring costsdown to a competitive level (page 287).

Only time will tell which technology will win in the end, but with such a lucrativemarket at stake, there will probably be roomor everyone to have a slice o the cake.

Coer imageHighly ecientmultijunction cells canbe used in concentratorphotovoltaic systems tobring down the cost o solar power.

Indtry Perpectie p284

EDITORs: NADYA ANsCOMBE,

OlIvER GRAYDON

PRODuCTION EDITOR: ChRIs GIllOCh

COPY EDITOR: ANNA DEMMING

ART EDITOR: TOM WIlsON

[email protected]

REsEARCh hIGhlIGhTs279 Improvingefciencies usingdipoles,nanotubes andparallelheterojunctions

INDusTRY PERsPECTIvE281 Manufacturing technology:

Fabricationinnovations NigelMason

284 Multijunction cells: Recordbreakers RichardR.King

287 Organic materials: Fantasticplastic RussellGaudianaand ChristophBrabec

BusINEss NEWs290 Newplants,collaborative

research,jointventuresand recordgrowth

PRODuCT hIGhlIGhTs291 Innovativemanuacturing systemsorphotovoltaiccells

INTERvIEW292 Acceleratingadoption Interviewwith WinriedHofmann

Paying catc-p

mailto:[email protected]






RESEARCH HIGHLIGHTS | TECHNOLOGY FOCUS

Payback time

Energy 33, 224–232 (2008)

Using solar panels to generate electricity is ultimately an environmentally friendly process, despite the huge amounts of energy used to produce the panels, according to

an Italian researcher who has carried out adetailed life-cycle assessment of electricity generation from photovoltaic cells.

Anna Stoppato, from the University of Padova, looked at mass and energy flows overthe whole production process, starting fromsilica extraction to the final panel assembly,and considered the most advanced andconsolidated technologies for polycrystallinesilicon panel production. She found that themost critical phases are the transformation of metallic silicon into pure polysilicon and thepanel assembly, which uses energy-intensivematerials such as aluminium and glass.

“The most important results of theanalysis are the calculation of a gross energy requirement of 1,494 MJ per panel (0.65 m2 surface area) and of a global-warmingpotential of 80 kg of equivalent CO2 perpanel,” says Stoppato. “I also evaluated theenergy pay-back time and I estimate it to beshorter than the panel operation life even inthe worst geographic conditions.”

Stoppato also looked at the energy returnfactor, which is defined as the ratio betweenexpected panel life (28 years) and the energy pay-back time. It represents how many timesa typical photovoltaic plant with 36 cells pays

back the energy needed for its production.She found that a photovoltaic plant can pay back this energy more than eight times.

Dipole alternative Appl. Phys. Lett. 92, 053507 (2008)

Researchers at the University of Toledo,USA, have developed an alternative way of creating the electric field required inphotovoltaic cells to effectively separatethe photo-generated electrons andholes. Instead of a field formed by the

electric contact between p- and n-typesof semiconductors, Diana Shvydka andVictor Karpov suggest that the electric fieldcan be generated by aligned electric dipolesembedded in a photoconducting host.

These electric dipoles are simply semiconductor nanoparticles, such aswurtzite CdS and CdSe or similar materialsthat are pyro- and piezoelectric. Someferroelectric nanoparticles may alsobe suitable.

“These dipoles do not even have to bein electric contact with the host material,which makes the system extremely flexibleto implement,” says Karpov. “In contrast,

p–n and similar junctions require a high-quality electric contact between the twomaterials, which sets major restrictions on thephotovoltaic cell design.”

The researchers claim that this conceptcan be used with a variety of host materials

that can be polymers, liquids, amorphousor polycrystalline. “We have shown thatthe generated field from these dipoles canbe uniform and strong enough, around3 × 104 V cm–1, to separate electron–holepairs and run significant drift currents,” saysShvydka. “Our suggested structure does notrely on p–n or Schottky junctions and can betunable in a broad range of parameters.”

Shvydka and Karpov do not projectextremely high absolute efficiencies fornanodipole photovoltaics and expect themto be at the level of the best existing thin-film photovoltaics (12–15%). “However the

relative cost of their generated power (indollars per watt) can be ten times lower thanthat of any of the existing photovoltaics,”says Karpov.

Nanotube hybridsNanotechnol.19,115601 (2008)

Although research on the use of single-walled carbon nanotubes (SWNTs) as theacceptor in polymer photovoltaic cells ismaking great progress at present, theirpoor dispersion in a polymer matrix hasgreatly hindered the overall performance

of the devices. Researchers in China think they have come up with a solution. They havedeveloped a bulk heterojunction structurebased on a poly(phenyleneethynylene)/SWNT composite. They found that itachieved better dispersion and higher

performance when compared with acommon control device based on a poly(3-octylthiophene)/SWNT composite layer.

Qian Liu and colleagues from theTianjin University of Technology andNankai University claim that theirs is a newphysical approach to the improvement of thedispersion of SWNTs in a polymer matrixbased on the design of the polymer molecule.

Poly(phenyleneethynylene) is a class of rigid-rod conjugated polymer composedof aromatic rings and alkyne functionalgroups, which can form a hybrid withSWNTs. The good dispersion of the SWNTs

in the poly(phenyleneethynylene) matrixmakes this composite a potential candidatein photovoltaic applications. At present,a limitation of the donor material, thepoly(phenyleneethynylene), is its narrowabsorption band, which hinders itsphotovoltaic performance, especially interms of its current density. The researchersbelieve this problem can be overcome by introducing other functional groups into thebackbone or side chain to improve the ability of harvesting solar energy. By broadening itsabsorption band, a higher current density and energy-conversion efficiency can thenbe expected.

RESEARCH HIGHLIGHTS | TECHNOLOGY FOCUS

acceptor to form simultaneously active,parallel heterojunctions with both donors.Moreover, the spectral photoresponseof these multiple heterojunction cells issignificantly broadened compared with

the conventional single donor–acceptorheterojunction cell. The researchersanticipate that optimization of the choiceof material and crystal size will furtherimprove the efficiency of this typeof solar cell.

Increased efciency Appl. Phys. Lett. 92, 053310 (2008)

Researchers in the USA have broadenedthe spectral response and increased thepower-conversion efficiency in a singleorganic thin-film photovoltaic cell by

creating simultaneously active, parallelheterojunctions with two donor materials.

The researchers used copperphthalocyanine (CuPc) and tin(II)-phthalocyanine (SnPc) as donor materialsand a single C60 acceptor. In previouswork using this combination of materials,an ultrathin intermediate donor layer wasgrown in discontinuous islands. However,when thicker than 5 nm, this layer formeda continuous layer that blocks interactionsbetween the acceptor and the otherdonor material.

In this new work, Fan Yang and

colleagues from Princeton University andthe University of Michigan, demonstrate atwo-donor, one-acceptor cell. Controlledcrystallization of the intermediatedonor layer ensures the formation of nanocrystalline islands that allow the

© 2

0 0 8 A I P

nature photonics | VOL 2 | MAY 2008 | www.nature.com/naturephotonics 279

http://www.nature.com/naturephotonics




From daily science updates to investigative

journalism, from community commentary

to editorial opinions, Nature brings you themost in-depth science news coverage online.

Visit our new blog, The Great Beyond, for a

comprehensive round-up and expert

discussion of today’s science news from

around the world.

Subscribe to Nature and receive full access

to Nature News as well as weekly issues of

the most influential scientific journal in the

world, in print and online.

Science news from

a different angle

www.nature.com/news





282 nature photonics | VOL 2 | MAY 2008 | www.nature.com/naturephotonics

INDUSTRY PERSPECTIVE | TECHNOLOGY FOCUS

and microelectronics applications. Tisprocess is expensive and increases the costo silicon rom around $2 per kilogramor MG-Si to around $40 per kilogramor polysilicon. Tere are a number o

alternative processes or the puricationo silicon. One that oers good prospectsor signicant cost reduction at presentcomprises the selective removal o key electronic impurities in molten MG-Si,ollowed by zone rening o solid siliconblocks. Successul pilot production trialshave been reported, and output rom therst production-scale units is expected overthe next two years.

cryStal growth

Once the silicon has been puried, two

orms o bulk silicon crystals or solar-cellmanuacture can be made. Te rst ormis monocrystalline silicon typically grownby the Czochralski process, where a seedcrystal o silicon located on the end o arod is dipped into molten polysilicon. Teseed crystal’s rod is pulled upwards andsimultaneously rotated. By controlling thetemperature gradients, the rate o pullingand the speed o rotation, it is possible toabricate a large, single-crystal ingot (alsoknown as a boule) o silicon rom themelt. Te second orm is multicrystallinesilicon, which is cast in a quartz crucibleunder controlled cooling. Te presence

o grain boundaries and associatedelectronic deects in multicrystallinesilicon means that the eciency o theresulting solar cells is normally lowerthan those made with monocrystallinematerial. However, or multicrystallinesilicon the unit investment cost or themanuacture is lower and the throughputis higher than or monocrystallinematerial. Current developments areocused around the casting process andincreasing the throughput and ingot size.Te standard production ingot masstoday is around 260 kg, but experimentalingots o up to 600 kg have been reported

by the manuacturers Solar Worldand Photowatt. A signicant innovationin this technology is the announcementby BP Solar o a process to growmonocrystalline-like material by a cast

process, which oers the prospects o ahigher-eciency low-cost product.

wafer SliciNg

Te bulk mono- or multicrystallinesilicon is rst sawn into bricks using aband-saw. A cutting fuid comprisingsilicon carbide particles in ethylene glycolmedium and a 160-μm-diameter steelwire travelling at a velocity o 12 m s –1 slices the bricks into waers. Te processproduces square waers with a side lengtho 125 mm, 200 μm thick, but with an

associated ker loss o 50%. Over 4,000waers are produced in a typical ve-hoursaw cycle. Over the coming three to veyears it is expected that waers will be cutusing wire with a diameter o 120 μm,to a waer thickness o 150 μm, givinga 30% increase in the waer yield perkilogram o brick. Sawn waers that are just 80 μm thick have been demonstratedunder laboratory conditions. Tis processoers the potential or urther savings,provided that the yields in subsequentstages o the cell manuacture can bemaintained. Te growth o silicon ribbonsdirectly rom the melt is a process that

In the waer-based crystalline siliconsolar-cell process described, thepolysilicon eedstock today accountsor some 20% o the cost o the inishedmodule. hin-ilm solar cells oer

a potentially lower-cost solution by reducing the amount o semiconductorneeded. he approach deposits thinsemiconductor ilms directly on to aglass support and optical window or onto a lexible metallic or metal-coatedpolymer substrate. In most thin-ilmphotovoltaic products, the top electricalcontact is a transparent conducting ilmo indium tin oxide (IO) applied to theglass surace. Various semiconductingmaterials have been used in these deviceswith those oering the most commercialprospects being: amorphous silicon

(a-Si), Cde and copper indium galliumdiselenide (CIGS).

In most thin-lm processesa sequence o conductor andsemiconductor layers is depositedusing vacuum-coating processes. Cellsare ormed by isolating areas o the

multilayer lm through laser ablation o the conductor or semiconductor layers,and monolithically integrating the seriesconnection with a nal conductinglayer. Tis nal layer connects the

rear o one cell with the ront o theadjacent cell through the interveninglaser-ablated trench. Te eciency o today’s commercial thin-lm productsapproaches 10%, but laboratory results onsmall area Cde and CIGS devices oerthe potential or higher perormance inthe uture. Nevertheless, the long-termprospects o thin-lm products are notwithout signicant challenges. Althoughthe material consumption is lower thanwith crystalline silicon, the materialsthemselves are not as abundant or as low-cost as silicon. Indium is the most notable

example. It is produced as a bi-producto metallic zinc and tin production. Itscost has increased more than 10 timesin recent years due largely to its use infat-panel displays. Its use in very high volumes or photovoltaic applicationsraises the prospects o having to establish

a mining and rening operation orphotovoltaic indium.

Some o the materials used inthin-lm products contain toxic heavy metals. For example, cadmium is

gradually being removed rom many manuactured components owing toenvironmental legislation. Manuacturerso Cde thin-lm products now oerend-o-lie recycling o their productto accommodate uture disposalrestrictions, and this will add signicantly to the initial product cost. All thin-lmproducts are at an inherently greater risk o degradation than waer crystallinesilicon. By denition, the active layersare thin and at risk o electromigrationo ions in the presence o an electriceld over extended periods. Whereas

crystalline silicon products have operatedin the eld or more than 25 years andcontinue to deliver electricity, the currentthin-lm products will need extensiveeld experience beore customers canbe condent o a high-perormance lieexpectancy o 20–40 years.

Box 1 Thin-film alternatives

B P

S O L A R

gmn s n pv nsns.

h s v s s.







avoids waer sawing completely. However,current ribbon growth rates are slowand their suraces are uneven, requiringspecialized techniques or cell production.

cell fabricatioN

Te most common photovoltaic cells

comprise a p–n junction photodiodenear the top surace o a silicon waer,where light is incident. Metal contacts areapplied to the top and bottom suracesto extract the current, with the topsurace requiring a ne-line metal grid tomaximize the surace area that is exposedto the Sun. Screen-printing is used toapply aluminium and silver conductinglayers to the top or bottom o the cells,which are then red at 700–800 °C toorm suraces on which external copperconductor strips can be soldered.

he eiciency o typical commercial

screen-printed photovoltaic cells rangesrom 14% to 16%, but a number o notable high-eiciency cells have cometo the market. One o the irst was BPSolar’s laser-grooved buried contactcell, which or almost a decade rom1992 was the world’s highest-eiciency commercial product. One o thedistinctive attributes o this cell was thetop metal contact. his was produced by orming narrow but deep trenches in thetop silicon surace (which incorporateda dielectric coating) using a laser, andelectroless plating the silicon grooves

with copper. hese cells are produced onmonocrystalline waers and are typically 18% eicient. Sanyo has pioneeredthe development o a cell in which the junction is ormed by doped amorphoussilicon layers applied to both suraceso an n-doped monocrystalline siliconwaer. Commercial cells o 20% eiciency are being produced. Most recently,SunPower has introduced a cell whereboth n and p contacts are integratedinto the rear surace o the waer,leaving the top surace completely reeo shading by conductors. Commercial

products containing 22% eicient cellsare expected on the market in 2008.Eiciency improvements oer the

greatest potential or reducing the unitcost per watt o solar power. However,

cost reductions are also expected romeconomies o scale in high-throughputautomated production lines and atransition rom printed silver to platedbase metal conductors.

Module aSSeMbly

Te assembly o the module, or panel,is the nal stage in the productionprocess. Under standard illuminationtest conditions, each 125 mm × 125 mmcrystalline silicon cell produces a typicalcurrent o 5 A at a potential o 0.5 V. In

a module, cells are connected in series,with a typical 72-cell module deliveringa power o 180 W at a potential o 36 V.Te module is a laminate comprising asheet o high-transmission glass, layerso ethylene vinyl acetate copolymer eachside o the solar cells, a breglass mattinglayer and a back sheet o polyester oraluminium-coated efon as a moisturebarrier. Te assembly is placed in a vacuum laminator at a temperatureo 130 °C, where the air and moistureare removed and the copolymer ormsstable and durable crosslinked chains.

A well-manuactured product shouldhave a lie expectancy o up to 40 years,and most quality manuacturers oer a

25-year warranty on their products. Longlie expectancy is vital to low-cost solarelectricity production. Te uel is ree andmaintenance costs are low, so the cost o energy ($ per kWh) is governed by themanuacturing cost, conversion eciency and operating period. Recent innovationsin module assembly have included the

application o a durable antirefectionlayer on the external glass surace, andbypass chip diodes on printed circuitboards incorporated into the laminate,thus avoiding external junction boxcircuits. In the uture, low-cost miniatureinverters may be attached at the individualmodule level to provide a.c. current outputand enable lower connection costs to theelectrical grid network.

Te energy used in the productiono a photovoltaic module, at present,takes between 1.5 years and 3.5 yearsto generate through sunlight energy

conversion (depending on irradiancelevels at the site o installation). Teexpected uture reduction on material(largely silicon) consumption per moduleand higher-eciency devices will reducethe current energy pay-back timeover the coming ve years. In terms o greenhouse-gas reduction, each installedkilowatt o solar power can avoid up to40 tonnes o carbon dioxide emissionduring its operating lie cycle.

outlook

Te prospects or photovoltaic solarenergy are exciting. Government targetsor 20% o electricity production romrenewable sources by 2020 will continueto provide market pull or the technology to establish the industry, and will drivecosts down to the level o retail electricity prices over the near term. Tere isan opportunity or new photovoltaictechnologies to demonstrate their merits,and existing technologies will continueto drive down costs in line with historictrends. In the longer term, new emergingtechnologies, based on materials such

as organic and polymer chemicals ornanostructured or nanocompositematerials, may become competitive.

B P

S O L A R

Pv s n nn s svn n

m ss. ts s-v n

s 140 μm . w ss 80 μm v

n mns.





Rhard R. KgSpectrolab, 12500 Gladstone Avenue, Sylmar,

California 91342, USA

e-mail: [email protected]

Flat-plate solar panels orhome-owner electricity generationhave been popping up on rooops

at an ever-increasing rate in recentyears. Te solar photovoltaics industry is enjoying growth at present that ewother industries have seen: globalphotovoltaic cell production has had asustained increase o more than 30%each year or the past 10 years, withincreases o more than 50% in someyears. Solar-cell electricity generatingcapacity, measured in megawatts peryear, exceeded 1 GW per year or the rsttime in 2004, about the equivalent o aconventional coal or nuclear power plant,and has been growing rapidly since. As

the process o energy conversion romsunlight to electricity in a photovoltaiccell produces no CO2 or other greenhousegases, supplying our electricity andtransportation needs with photovoltaicsis one o our best hopes or developinga carbon-neutral orm o powerproduction, and staving o the perils o global climate change.

Within the eld o photovoltaics,there is a new technology just comingout o the labs and onto the marketplacethat is arguably growing even aster thanthe rest o the industry. Te technology

is multijunction concentrator solar cellsbased on III–V semiconductors. Testriking eature o such cells, and thephotovoltaic systems made rom them,is their ability to convert sunlight toelectricity with very high eciency. Tisreduces the area o all components o thesystem and drives down cost. Te cellsachieve their high eciency by combiningseveral solar cells, or p–n junctions, intoa multijunction cell. Each solar cell inthe multijunction cell, called a subcell, iscomposed o a dierent semiconductormaterial, each with a dierent energy bandgap. Te subcells are electrically

connected in series as shown in Fig. 1.Te subcells are also positioned in optical

series such that the highest bandgap is inthe subcell on top, which uses the highestenergy photons in the solar spectrum.Photons with energy lower than thebandgap are transmitted to the subcellbeneath, which has the second highestbandgap, and so on. In this way themultijunction solar cell divides the broadsolar spectrum into wavelength bands,each o which can be used more eciently by the individual subcells than in thesingle-junction case.

In the past eight years, the outputpower o terrestrial III–V multijunctioncells used in a solar concentrator design

(an approach where optics are used toocus and concentrate sunlight to achieve

higher light intensities) has increased by 25%. Te workhorse multijunction cellor terrestrial concentrator photovoltaics,which itsel is the result o a great dealo innovation and vigorous renementin the past decade, is the lattice-matchedthree-junction GaInP/GaInAs/Ge cell,shown in Fig. 1a.

But the eciency o this type o cell can be improved still urther,with an approach called metamorphicmultijunction solar-cell design. Te solarspectrum turns out to be best suitedto dierent semiconductor bandgapsthan those that are readily accessible

Multijunction solar cells used in concentrator photovoltaics have enabled record-breaking

efciencies in electricity generation rom the Sun’s energy, and have the potential to make solar

electricity cost-eective at the utility scale.

Multijunction cells

Record breakers

GalnP

top cell

n+-GalnAsn-AllnP window

n-GalnP emitter

n-GalnAs emitter

n-GalnAs buffer

n+-Ge emitter

p-Ge baseand substrate

Nucleation

Contact

p-GalnP base

p-GalnAs base

n-GalnP window

p-AlGalnP BSF

p-GalnP BSF

p++-TJ

n++-TJ

p++-TJ

n++-TJ

GaInAs

middle cell

Gebottom cell

Wide-bandgap TJ

TJ

Buffer region

Contact

AR

Contact

AR

n+-GalnAsn-AllnP window

n-GalnP emitter

n-GalnAs emitter

p-GalnAsStep-graded

buffer

n+-Ge emitter

p-Ge baseand substrate

Nucleation

Contact

p-GalnP base

p-GalnAs base

n-GalnP window

p-AlGalnP BSF

p-GalnP BSF

p++-TJ

n++-TJ

p++-TJ

n++-TJ

Lattice matched Lattice mismatched or metamorphic

Figure 1 shma r-a dagram hr- gra. a, la-mahd ad,

b, mamrph GaiP/GaiA/G, rrpdg 40.1% (a mahd) ad 40.7%

(mamrph) r rar . (AR: arf ag; BsF: bak-ra d; tj: .)









in the lattice-matched three-junctioncell. Specically, the solar wavelengthdistribution avours lower bandgaps orthe upper two subcells than nature hasprovided or GaInP and GaInAs grownat the lattice constant o a germaniumsubstrate. Te lower bandgaps canbe reached by increasing the indiumcontent, but this increases the latticeconstant, typically causing the ormationo dislocations in the crystal lattice whengrown on a germanium substrate. Tisin turn has a detrimental eect on the voltage in the subcells.

In the metamorphic multijunctionsolar cell (Fig. 1b) the crystal dislocationsare allowed to orm in a metamorphicbuer — a region with a gradedsemiconductor composition, grownrst on the substrate. Te crystalstructure relaxes, so that by the endo the metamorphic buer growth, anew, larger lattice constant has beenreached, which can be used as a virtualsubstrate or growth o semiconductorswith high crystal quality at the largerlattice constant. Te central challengewith metamorphic solar cells is keeping

the dislocations that propagate upwardinto the active cell layers to a minimum,through careul buer design.

Recent advances in III–Vmetamorphic multijunction solar-celldesign or terrestrial use have resultedin ever greater photovoltaic eciency.Metamorphic three-junction cellshave been demonstrated in the labwith high-indium-content Ga0.44In0.56Ptop subcells, and Ga0.92In0.08As middlesubcells, at a lattice mismatch o 0.5%with respect to the germanium substrate,which also serves as the third and bottomsubcell. Such metamorphic three-junction

concentrator cells developed at Spectrolabhave reached a record eciency o 40.7%under the terrestrial solar spectrum,when the sunlight was concentrated by aactor o 240 (Fig. 2). Tis is the highestsolar-energy conversion eciency yetachieved or any type o photovoltaic cellunder standard spectral conditions, and

is the rst solar cell to reach over the 40%milestone. Te lower bandgaps o thetop two subcells result in a more optimaldivision o the solar spectrum among thesubcells, pushing the eciency higheror this type o cell. Lattice-matchedthree-junction GaInP/GaInAs/Geconcentrator cells developed at Spectrolabare not ar behind, with independently conrmed eciencies o 40.1%.

Te shape o the solar spectrum alsoavours a higher bandgap than that o the germanium bottom subcell in thelattice-matched three-junction design. In

another type o metamorphic cell, underinvestigation in several labs around theworld, the germanium third subcell isreplaced with a metamorphic GaInAssubcell that has a bandgap o around 1 eV,grown in an inverted conguration witha transparent graded buer region. Tistype o metamorphic cell design has giventhe best perormance (33.8% eciency)under unconcentrated sunlight in cellsdeveloped by the National RenewableEnergy Laboratory.

In photovoltaics, eciency has adirect impact on the cost per watt o

electricity delivered. A high cost perwatt in the past is the basic reasonwhy solar electricity has not displacedconventional electricity productionso ar. Cost per watt has been steadily declining, and is nearing the tippingpoint at which photovoltaic technology may be deployed on a very broad scale,

replacing a signicant raction o theelectricity-generation inrastructure,but it is still not low enough. A largepart o the cost o photovoltaic systemsis in the semiconductor material o thesilicon or other types o solar cells used,but the dominant cost is that o moreconventional materials, such as glass,metal and plastic. Solar energy is a airly dilute energy resource, and requirescollectors with a large surace area togenerate useul amounts o power. I the eciency o the system is doubled,then only hal the collector area is

needed or the same power production,halving not only the amount o expensivesemiconductor material required or thesolar cells, but also those conventionalmaterials that contribute so much to thesystem cost.

Trough a combination o higher lightconcentration and better use o the solarspectrum due to its multijunction design,III–V three-junction concentrator cells with37–40% eciency can easily enable overallconcentrator photovoltaic system ecienciesgreater than 30%, roughly double that o thetypical silicon fat-plate solar module.

1mm

S P E C T R O L A B

Figure 2 Phgraph a sprab iii– V m

rar ar , h yp ha ha rahd a

rrd 40.7% y.

A M O N I X

Figure 3 examp yp rar ar phva ym rm Amx, g a parq Fr

rm a array may ma gh pavy d ar .





In concentrator systems like thoseshown in Figs 3 and 4, with concentrationratios o around a actor o 500, thecell area required can be approximately 500 times smaller than the aperture areao the primary concentrating optics. Tedramatic reduction in cell area allows

III–V multijunction cells, which arecostly per unit area, to be used eectively in a concentrator system. Tese III–Vmultijunction solar cells have becomethe cell architecture o choice or mostconcentrator photovoltaic systems beingdesigned today because o their higher

eciency. Although III–V multijunctioncells are more complex and expensivethan conventional silicon cells, thereduced system cost that their higheciency brings more than makes upor their cell cost. Tis lower system costand corresponding drop in cost per kWho solar electricity is expected to enable

photovoltaics to compete not only in theresidential market, but also in the morechallenging commercial rooop andutility markets, where prices paid perkWh are very low.

Te production levels o III–Vmultijunction concentrator photovoltaiccells are projected to reach severalgigawatts per year in the next decade,catching up with and augmentingthe burgeoning market or its sistertechnology, fat-plate silicon solarcells. Continuing research in advancedmultijunction cell architectures around

the globe, many o which incorporatemetamorphic semiconductor materials,can be expected to increase practicalconcentrator-cell eciencies to 45%or even 50%. By using eciency as apowerul lever to bring down the cost,it will become possible to deploy solar-celltechnology on a larger scale than everbeore. Tis will undamentally change theway we generate electricity, and orm a key part o the solution to the energy-security and global climate-change perils that weace today.

S O L A R

S Y S T E M S P T Y

Figure 4 examp ahr yp rar ar phva ym rm sar sym Py, g a arg

rfv dh a d array phva .







Ru Gud* d

Chroph Brbc†

Konarka Technologies, 100 Foot of John Street,

Boott Mill South, 3rd Floor Lowell, Massachusetts

01852, USA

e-mail: *[email protected];†[email protected]

Much o the early work onphotoactive materials orphotovoltaics ocused on

crystalline silicon, which dominatesthe commercial solar-energy eldtoday. Several other materials, such asamorphous silicon (a-Si), cadmiumtelluride (Cde) and copper indiumgallium selenide (CIGS), are also now in various stages o commercialization, andare known as thin-lm technologies. Telower manuacturing costs and higherproduction throughput o these materialspotentially translate into lower electricity

costs. Current thin-lm technologies areexpected to bring costs reasonably close to$1 per watt o electricity produced at peak solar power.

here is, however, anothertechnology that has the potential tobring this cost down even urther. Bulk heterojunction technology, using organicsemiconductors and roll-to-roll coatingand printing techniques, could becomethe technology that makes solar energy aordable to the general public. hetechnology uses abundantly availablenon-toxic materials, is based on a

scalable production process with highproductivity, and requires low investmentrom the manuacturer.

A bulk heterojunction is a blend o p- and n-type semiconductors, whichorms molecular p–n diodes all overthe bulk layer. On light absorption,photo-induced charges are producedby ultraast charge transer (within aew emtoseconds) between the twosemiconductor types. Various materialsystems have been suggested or bulk heterojunction solar cells, including:organic semiconductors, inorganic andorganic semiconductor nanoparticles

and nanorods, metal oxides stained withdye molecules, as well as combinationso these.

O all these technology platorms,organic photovoltaics is generatingconsiderable interest (Box 1). Asthe name implies, this technology comprises carbon-based materials asdonor and acceptor molecules. he

most popular class o organic donormolecules are conjugated polymers,such as polythiophenes, polyluorenesor polycarbazoles. he material choiceor acceptors is much narrower — ormore than ten years substitutedullerenes have given by ar the bestphotovoltaic perormance.

he eature that dierentiates thistechnology rom all o the others isits compatibility with high-speed andlow-temperature roll-to-roll processing.he processes are typical o those usedin the printing and coating industry inthat solutions o the active materials,

dissolved in organic solvents or water,are applied to a plastic sheet by means o a coating applicator.

Various printing and coatingtechnologies have proven theircompatibility with organicsemiconductor processing, among themgravure printing, lexo printing, screenprinting, slot die coating and, most

recently, ink-jet printing. he printingsolvent is evaporated when heated tomoderate temperatures, producing adried layer o the photoactive polymer.he modules are encapsulated betweenthin, lexible over-laminates, whichprotect the active layers rom mechanicalabrasion and the environment. Capitalcosts are very low, and the printing andcoating processes can be done at highspeed with no obvious limitation in thesubstrate width. he processing steps aresummarized in Fig. 1.

he combination o low-temperatureprocessing paired with high production

Polymer materials could bring down the cost o electricity production using photovoltaic

technology to below $1 per watt or the frst time, and enable mass-market, portable

applications or photovoltaic technology.

ORGaniC mateRials

Fantastic plastic

Transparent

packagingTransparent

electrode

Substrate

Angle viewEnd view

Printed active

material

Primary

electrode

Transparent electrode

External load

ElectronsTransparent packaging

Light

Primary electrode

Substrate

Active material

(polymer blend)

K O N A R K A

Figure 1 Poyr-bd phoovoc c c b ucurd ug drd prg proc.









throughput suggests that attractiveenergy payback times — the time it takes

to generate energy equivalent to thatoutlaid during abrication — should bepossible. With large-volume manuactureand reasonable eiciencies o 5% to10%, these printed solar cells shouldhave the potential to go signiicantly below $1 per watt o electricity. So ar,

no complete lie-cycle analysis has beencompleted, but expectations are that the

energy payback time can be as low as aew weeks.

he eiciency o organic photovoltaictechnology is low when compared withsilicon or compound semiconductortechnologies. However, many signiicantmodiications and improvements have

been made to the polymer structuresover the past ew years (Fig. 2). hesemodiications include the incorporationo comonomers that withdraw electronsrom the sea o electrons on the polymerbackbone, causing a large shit in theabsorption band towards the inrared.Some o these modiications result in

absorption at wavelengths as long as1,000 nm and more, enabling the cellto absorb more than 50% o the totalradiation rom the Sun. In addition, thestructures are versatile enough, roma molecular architecture standpoint,that the entire visible spectrum can becovered as well. Modelling indicatesthat some o these polymers, whencombined with ullerene, will exhibit celleiciencies between 7% and 10%.

Even higher eiciency values areexpected or tandem or multiple junctiongeometries, where solar cells o dierent

bandgaps are stacked on top o eachother and interconnected in series. Eachcell absorbs at a dierent wavelength,reducing the amount o uncapturedradiation that is lost as heat and enablinghigher eiciencies. he materials ortandem cells can come rom a variety o

All photovoltaic cells have severalcommon eatures: they must have two

electrodes and a layer o photoactivematerial that absorbs light (a photon)and generates current (an electron).he key to the success o organicphotovoltaics is its two-componentactive layer, which on coating anddrying orms a very unique morphology (shown schematically in Fig. B1)reerred to as a bulk heterojunction.he main eature o this heterojunctionmorphology is the intertwining o phases o each o the components, whichspontaneously occurs when the solventis evaporated. In current designs o

polymer photovoltaic cells, one o thecomponents is a polymer that has threeunctions: to absorb light; to inject anelectron into the second component;and to carry the resultant hole to oneo the electrodes. he other componentis a ullerene derivative. Its unction isto accept an electron and carry it to theother electrode. he primary advantageo the bulk heterojunction is the very high surace area that is ormed betweenthe two phases, which directly aectsthe eiciency o charge transer betweenthe polymer and the ullerene phases.

External quantum eiciencies o up to80% have already been demonstrated,though today’s power-conversioneiciency is only in the regime o 5–6%.

he current generated by the cellis related to the absorption spectrumo the polymer, which is determinedby its molecular structure. hestructural design o these moleculesis adjusted to absorb broadly acrossthe solar spectrum rom the blue to

the near inrared. As the active-layercoating is a raction o a micrometre

thick (100–200 nm), the polymers musthave a very high absorptivity as well.When choosing which polymers to useas the donor and acceptor molecules,material scientists must look at thedierent energy states o the molecularelectrons in the donor and the acceptormolecule. All charge carriers needto be transported across the bulk tothe electrodes beore recombinationtakes place. At typical carrier lietimeso a ew microseconds, charge-carrier mobilities o 10–3 cm2 V–1 s–1 or higher are required or loss-ree

carrier collection.Tese selection criteria signicantly narrow down the polymeric structuresthat have potential or ecientphotovoltaic energy conversion. Findingsuitable polymers or this applicationis challenging because a compromisemust always be reached between thechoice o bandgap (which dominatesthe short-circuit current) and the positiono the electronic levels (which dominatesthe open-circuit voltage). Te product o the short-circuit current and the open-circuit voltage dominates the eciency.

Box 1 The structure of an organic photovoltaic cell

Figure B1 th ch o chrg rr d

rpor buk hrojuco rucur.

Figure 2 th po how rcord cc chvd wh orgc phoovoc choogy crd by h

no Rwb ergy lb (nRel). a, Ru or dvc wh cv r o 1.024 c 2 d ccy

o 5.21%. b, Ru or dvc wh cv r o 0.685 c 2 d ccy o 5.24%.

10

8

6

4

2

0

–2

–0.2 0.0 0.2 0.4 0.6 0.8 1.0

12

10

8

6

4

2

0

–2

–0.2 –0.1 0.10.0 0.2 0.3 0.4 0.5 0.6

C u r r e n t ( m A )

Voltage (V) Voltage (V)

C u r r e n t ( m A )

N R E L

n-type carriers:

electron acceptors

+

++

e– e–

e–

p-type carrier:

electron donators

Positive charge

Negative charge

Primary electrode







conjugated polymers, always combininga narrow-bandgap semiconductor witha wide-bandgap one. Eiciencies over6% have already been reported or atandem architecture based on polymerand ullerene composites (Kim, J. Y. et al. Science 317, 222–225; 2007).

Although still low compared with

silicon and compound semiconductortechnologies, these ecienciesrepresent a big step orward in polymerphotovoltaic technology. Polymersystems also have many advantages overconventional photovoltaic technology.For example, organic photovoltaicmodules are lightweight, have a highpower-to-weight ratio (more than100 mW g–1 at 5% eciency) and they are mechanically fexible (Fig. 3). Tismakes them particularly useul orportable applications, which represent therst target markets or the technology.

Potential uses include battery chargers ormobile phones, laptops, radios, fashlights,toys, and almost any handheld devicethat uses a battery. Te modules can beadhered to the outside ace o a briecaseor a piece o clothing, or they can beincorporated into the housing o a device.Tey can also be rolled up or olded orstorage in a pocket when not in use. Terst products serving this market shouldbe available later in 2008.

Organic photovoltaic modules willprobably also be used in architecturalor building-integrated applications.

Potential applications are inelectricity-generating awnings that canbe rolled up or storage or windows inoice buildings and greenhouses, wheretwo transparent electrodes are used sothat the module is semitransparent.

he biggest market or organic solarpanels is large-area rootop applications

in residential and commercialbuildings. his will require eicienciesbetween 7% and 10% and lietimes o 7–10 years. Konarka is now runningextensive lietime investigations, andthe company’s cells regularly pass morethan 1,000 hours under accelerateddegradation. o veriy the lietimeo 7–10 years, the degradation overaround 3–5 years must be measuredand then extrapolated. So ar, Konarkahas been measuring the lietimeo its cells or more than one year(extrapolated out to three years) with no

degradation observed.his shows the enormous progress

that has been made over the past ewyears in designing environmentally stableorganic materials and interaces. Untilrecently it was thought that organicsolar-cell lietimes were restricted to1–2 years, but today it is clear that thetechnology could reach lietimes o 5–10 years, making it a strong contenderor consumer solar-cell applications.

Figure 3 Poyr phoovoc choogy h hgh

powr-o-wgh ro, whch ow h odu h

hv h ddd dvg o bg ghwgh d

fxb, kg d or porb ppco.

K O N A R K A



290 naturephotonics|VOL2|MAY2008| www.nature.com/naturephotonics

BUSINESS NEWS|TECHNOLOGY FOCUS

Ersol and Schott to develop

thin-flm technology

German companies Ersol Thin Film andSchott Solar have agreed to jointly developso-called micromorphous technology forthin-film solar cells. The two companieswill be combining their resources in the

area of research and development, andwill exchange staff between the two sites.In contrast to the amorphous version, amicromorphous thin-film module has adouble-layer structure consisting of anamorphous and a microcrystalline siliconfilm. The arrangement results in improvedexploitation of sunlight, because the twosilicon layers convert the whole lightspectrum into power. “We believe that thisso-called micromorphous tandem cell canachieve up to a 50% increase in the moduleefficiency, and therefore also in the moduleyield in comparison with amorphous

technology,” explains Christian Koitzsch,managing director of technology for ErsolThin Film.

Q-Cells becomes ‘world’s largest

manuacturer o solar cells’

German company Q-Cells claims it is nowthe world’s largest manufacturer of solar cells,after reporting a 54% increase in productionin 2007 to 389.2 MW. In 2007, sales roseby 59% to €858.9 million ($1,300 million)compared with €539.5 million ($830 million)for the previous year. Earnings before interest

and taxes grew by 52% to €197.0 million($300 million), compared with€129.4 million in 2006 ($200 million). Forfiscal year 2008 as a whole, Q-Cells continuesto expect total sales of approximately €1.2 billion ($1.85 billion).

To allow rapid growth to continue,Q-Cells has decided on furtherexpansion in its production capacity.

In its core business, the production of monocrystalline and polycrystalline solarcells based on silicon wafers, Q-Cellshas completed the ramp-up of the firsttwo phases of expansion of productionline ‘V’ and is now starting work on theconstruction of production line ‘VI’ inBitterfeld-Wolfen, Germany. The firstphase of expansion with a productioncapacity of 130 MW is to start productionin the fourth quarter of 2008. The company has also decided to construct its seventhproduction line in Malaysia.

Sustainable cell manuacture

A team of scientists from eight universitiesacross the UK is embarking on one of theUK’s largest ever research projects intophotovoltaic solar energy.

Led by Durham University, the£6.3 million ($12.5 million) PV-21programme will focus on makingthin-film light-absorbing cells for solarpanels from sustainable and affordablematerials. The eight universities will work together with nine industrial partnerstowards a “medium- to long-term goal”

of making solar energy more competitiveand sustainable. Principal investigatorKen Durose highlights the price of indium($660 per kilogram) as a particular costbarrier. “At present you would need tens of tonnes of very rare and expensive materialsfor large-scale production of solar cells toproduce sizeable amounts of power,” he

says. “Some of the materials currently usedmay not be sustainable in 20 years time,which is why we have to conduct researchinto alternative materials that are cheaperto buy and more sustainable.”

Thin-flms to eat into

silicon’s dominance

The global market for photovoltaics isexpected to be worth more than $16 billionby the end of 2008, according to a marketreport by BCC Research. This is expectedto increase to over $32.2 bil lion by 2013, acompound annual growth rate of 14.9%.

According to the report, the energy produced by photovoltaic cells and

modules from global shipments reached2,875.1 MW in 2007. This is projected togrow by 28.6% to reach 3,697.3 MW by the end of 2008, and with a compoundannual growth rate of 30% it will reach13,724.4 MW by 2013.

Silicon technology, which accountedfor about 89% of the market in 2007, willcontinue to dominate in the near future,with shipments of multicrystalline siliconcells growing by 285% by 2013. Recentimprovements in this traditional technology and its reliability will keep it in the forefront,and BCC predicts that silicon will still

represent 79% of the market in 2013.That said, thin films, which only have10% of the market at present, will grow by 45% to account for 19% of the market in2013. Improvements in efficiencies and theuse of these materials on flexible substrateswill account for their rapid growth.

New technologies, such asnanostructured thin films and silicon anddye-sensitized solar cells, accounted for just under 0.5% of the market in 2007,but will grow by 34% to reach an energy production of 19.2 MW in 2008, and thenexhibit 50% annual growth to achieve145.7 MW by 2013.

Abengoa Solar has signed a contractwith Arizona Public Service Company (APS) to build, own and operate whatwould be the largest solar power plant inthe world if operating today.

The plant, called the SolanaGenerating Station and scheduled to go

into operation by 2011, is located 70 milessouthwest of Phoenix, near Gila Bend,Arizona, USA. It will sell the electricity

produced to APS over the next 30 years fora total revenue of around $4 billion.

The plant will have a total capacity of 280 MW, enough to power 70,000 homes.It will use proprietary concentrating

solar power (CSP) trough technology developed by Abengoa Solar, and willcover a surface of around 1,900 acres.

The solar trough technology usestrackers with high-precision parabolicmirrors that follow the Sun’s path andconcentrate its energy, heating a fluid toover 371 °C and using that heat to turnsteam turbines. The solar plant will alsoinclude a thermal-energy storage systemthat enables electricity to be produced asrequired, even after the Sun has set.

Abengoa Solar is now operating theworld’s first commercial CSP solar tower

plant in Spain, and is building threemore in the country with a total capacity of 120 MW.

A B E N G O A

S O L A R

Giant solar plant or Arizona

B C C

R E S E A R C H





PRODUCT HIGHLIGHTS | TECHNOLOGY FOCUS


A exible system or making

exible photovoltaic cells

www.solarcoating.de

German company Solar CoatingMachinery has developed a modular pilotline for the production of flexible solarcells. Called Click&Coat, the system ismade up of 12 units, which click together

to form a pilot line. These are a winderunit, coating unit, infrared dryer unit,flotation unit, sintering unit, vacuumunit, bath unit, another winder unit,edge-guiding unit, slitting unit, laminatingunit and an edge-sealing unit. Thecompany claims its pilot lines can be usedto scale up from research to production volumes and suit a variety of photovoltaictechnologies, including amorphous silicon,dye-sensitized cells, organic, copperindium selenide (CIS) and copper indiumgallium selenide (CIGS). The company says that its products are particularly

suitable for manufacturing flexible solarcells, performing encapsulation, packagingand substrate bonding.

Non-contact solutions or handling

crystalline waers and solar glass

www.coreflow.com

Israeli company CoreFlow has developeda number of non-contact solutions forthe transport, handling, and processingof crystalline wafers and solar glasssubstrates. CoreFlow’s air-cushionsolutions enable high-speed handlingand transport (up to 100 m per minute)for large thin-film solar substrates. Thisnon-contact procedure enables very high

process speeds and reduced vibrationduring the manufacturing process,resulting in improved throughputand higher manufacturing latitude.CoreFlow’s Pressure Atmospheretechnology floats substrates with an airgap of up to 300 μm in size to ensurenon-contact, and is ideal for most inline

and offline transport applications. ThePressure Vacuum system is ideal forcritical process zone applications, suchas laser-scribing and surface-inspectionstations. The company claims its systemscan enable higher material throughputand minimize glass stress and damageduring manufacturing.

Metallization urnace combines

solar-cell drying and fring in one unit

www.btu.com

The company BTU International hasintroduced a metallization furnace thatcombines solar-cell drying and firingin one unit for optimized throughputand reliability. The most recent additionto the PVD-600 Series furnaces, thePVD624, combines the rapid thermalprocessing of a 600 Series Solar CellFiring Furnace and a D-900 Series Dryeras an integrated system. By combiningdrying and firing in one unit, the solarcells transfer directly from the dryer intothe furnace without additional handling.This improves efficiency by saving factory

floor space and reducing automationcosts. In addition, the PVD-624 offersgreater throughput with the use of awider, 24-inch belt. The PVD-624 alsofeatures highly effective, closed-loopconvection cooling, which enables precisecontrol of cooling rates. This new designalso provides a significantly advancedwater cooling system. Quartz tubes set ina herringbone pattern maximize thermaluniformity throughout the PVD624.Individually removable, sliding side portsin each zone allow easy cleaning belowthe belt, and an enhanced exhaust design

helps to provide superior uptime.

Flexible laser system processes

thousands o silicon waers per hour

www.innolas.de

The ILS 700 laser system from Germancompany Innolas is a dedicated toolfor processing monocrystalline andpolycrystalline silicon solar cells.Examples of the laser-processingtechniques available with the system are:laser edge isolation, laser-fired contactsprocesses, micro through-hole drilling,

SiN and SiO2 ablation, downsizing,surface modification, laser scribing andsurface structuring.

This gantry-type machine enables theprocessing of wafer sizes ranging from100 mm × 100 mm to 210 mm × 210 mm,selected using a software program.Process performance does not degrade as

the wafers increase in size, and automaticlaser treatment of pseudo square formatsis also possible. The machine can besupplied in various configurations,including a version that suits integrationinto fully automatic manufacturing lines.

The system offers the possibility of integrating up to two laser sourcesin one machine for processing twowafers simultaneously or for subsequentprocessing with different laser sourcesand wavelengths. This guaranteesflexibility and a very high throughput of up to 2,000 wafers per hour, depending

on the application.

System combines water and

laser cutting in one operation

www.synova.ch

Swiss company Synova claims its LaserMicroJet can cut photovoltaic cellswithout affecting electrical efficiency orinducing chips, burrs, mechanical stressor damage from heat. This technology enables omnidirectional cutting, drilling,scribing, grooving, edge grinding,thinning, marking and other specificapplications for thin photovoltaic cells.

The Laser MicroJet combines theadvantages of both water and laser cuttingin one operation. Using the difference in therefractive indices of air and water, a laserbeam is guided to the workpiece within athin jet of water, which acts as a light guide.

Laser MicroJet technology enables thebeam to be guided over a distance of upto 10 cm. The water jet cools the substratewhile removing the molten material fromthe cut and avoiding contamination.The result is accurate cutting of porousor layered materials, with minimalthermal and structural distortion and afinely cut edge.

S Y N O V A

S O L A R

C O A T I N G

M A C H I N E R Y

C O R E F L O W

http://www.solarcoating.de/

http://www.coreflow.com/

http://www.btu.com/

http://www.synova.ch/




http://www.btu.com/

http://www.coreflow.com/

http://www.solarcoating.de/



INTERVIEW | TECHNOLOGY FOCUS

Why is so little o the world’s electricity generated by solar power?Cost is the one and only limitationthis industry has had to its growth.Until recently, it simply has not beeneconomical or the average householderto buy a solar-power system or electricity generation. But this will change

dramatically in the next 20 years. Te costo generating power using solar cells hasbeen decreasing by 5% every year, and Ibelieve the industry is able to maintainthis trend. In 1991, the electricity production cost rom a 1-kW residentialrooop system under German sunlightconditions was €1.1 per kWh. oday, thisis more like €0.40 per kWh. In 20 yearstime, electricity production costs will becomparable to those o a 500-MW cleancoal power station, and this is when solarpower will nally take a considerably larger market share o the energy market.

Can the industry survive without theheavy subsidies it receives at present?Tere are already certain places in theworld, such as Caliornia and Japan,where a liberalized utility market exists,and in these places an investment insolar power would pay o, even withoutsubsidies. But at the moment, most o the industry is reliant on subsidies. Tishas been necessary to get volume up andto get us through the price–experiencecurve as quickly as possible so we can getto the next stage. All these subsidies are

doing is speeding up the process. Withoutthem, the industry would still haveprogressed, but more slowly. Germany is a particular success story. en yearsago, there were about 5,000 peopleworking in the photovoltaics industry inGermany. oday this gure is more than40,000. It is predicted that in 20 yearstime, the photovoltaics industry willhave a turnover similar to that o thesemiconductor industry, so the Germangovernment has been wise to invest soheavily in this area.

But the subsidies cannot go on orever.Te cost o electricity production rom

solar cells depends a lot on the amount o sunlight available in a particular location.In 5–8 years, sunny parts o the world willbe able to survive without subsidies, but inother areas, such as the UK and Germany,this will take closer to 20 years. Oncethe cost o making electricity rom solarcells is comparable to that rom a largepower plant, the industry will not needany subsidies.

How has the silicon shortage afected thephotovoltaics industry?When an industry is growing as astas the solar-cell industry is, there willinevitably be shortages in some materials.Te shortage in solar-grade silicon hasbeen a huge challenge or the industry to overcome. Making polysilicon takeschemical expertise that cannot be learnedovernight. Rectiying the shortage hastaken 2–3 years o investment and over thecourse o 2008 much greater quantities o silicon eedstocks have become available.By 2010 we will have three times more

polysilicon than we had two years ago.Although the shortage has been painul, ithas also had some good points. Necessity isthe mother o invention and the industry’sengineers have been very inventive duringthe shortage. A ew years ago no-one wouldhave believed that you can make solar cellswith waers just 200 μm thick, but today

this is a reality.

Will silicon technology always dominatethe market?Other technologies, such as thin-lm,polymer and III–V semiconductorsystems, will take an ever-increasingmarket share in the uture, but I believecrystalline silicon will continue todominate the market over the next20 years. Tin-lm technologies aregrowing ast: every 5 years, they increasetheir market share by 5%. I have alwaysbelieved it is the applications that choose

the technologies and not the other way around. So i you have limited spacebut plenty o sunshine, silicon is themost appropriate technology to use. I the application is in partial shade or,or example, a photovoltaic system isbeing integrated into a semitransparentwindow, then thin-lm technologies aremore appropriate. Organic photovoltaicswill be particularly suited to applicationswhere their shorter lietimes are not anissue. For example they can be used inmobile devices, where a lietime o lessthan ve years is not a problem. I believe

that, while the efciencies o silicon solarcells are continuously increasing, they will never match the efciencies o III–Vsemiconductor multijunction systems. So,although these systems are expensive, orsome applications they are the systems o choice because o their higher efciency.It all comes down to the intelligent use o technology or each specic application,but or the oreseeable uture I believesilicon-based systems will still dominatethe market.

Nadya Anscombe is a freelance science and technology journalist based in the UK.

Winfried Hoffmann is president of the European

Photovoltaics Industry Association.

The photovoltaics industry is growing ast, but it still needs to bring down costs beore it can reach

its true potential. Nadya Anscombe talks to Winried Homann, president o the EuropeanPhotovoltaics Industry Association, to fnd out more.

Accelerating adoption






Taking LightFurther

BreaultResearch

USA | Canada 1.800.882.5085

Worldwide 1.520.721.0500

Breault.com

Nowshipping.

Request your ree ASAP DVD-ROM DEMOat www.Breault.com/DVD and see

optical engineers as they discusseatures and applications o the

world’s most sophisticatedoptical sotware, ASAP.

Over 50 ASAP movies(6+ hours) cover:

Display DesignImaging Systems

Stray Light AnalysisIllumination Engineering

Wave Optics…and more!

See rsthand why opticalengineers and designers in

35 countries choose ASAP ortheir virtual prototyping needs,

and witness the unmatched capability,fexibility, speed, and accuracy o ASAP.

•

•

•

•

•

Breault.com/DVD



©2008 Newport Corporation

True power is knowingwhen not to use it.

The Mai Tai®

DeepSee™

laser—more fluorescence with less power.

Now you can preserve more living tissue while looking deeper into life’s

mysteries. Mai Tai DeepSee is the first multiphoton laser source to penetrate

so deeply into the sample and create brighter fluorescence while minimizing

tissue damage. Make new discoveries at greater depths with less power and

better signal to noise than with other lasers. Take your research further with the

Mai Tai DeepSee. And discover the very real power of doing more with less.

1.800.775.5273 newport.com/deepsee08

Documents

Nature - Photovoltaics