Technion, 25/2/07

Where are we today in y Research in Photovoltaics ?

(efforts in Israel are a separate story)

Weizmann Institute of Science

David Cahen

THANKS to many colleagues, esp. Robert Birkmire R b t Bi k i

Institute of Energy Conversion, U o Delaware DoE University Center of Excellence y for Photovoltaic Research & Education
Tom Surek Bolko von Roedern

NREL NREL

The Photovoltaic (PV) effect:

Light Absorption + Carrier Generation + Carrier Collection e charge e separation load + h h+

Solar (photovoltaic) cells work like an electron slide

A PV device:

Types of PV Devices
Primarily based on solid state electronic material systems solid-state
Elemental Semiconductors
Single or multi-crystal multi crystal Polycrystalline Amorphous thin film

Si, Ge
GaAs InP CdTe CuInSe2

Inorganic Compound Semiconductors

Single crystal Polycrystalline thin film

Organic Compounds

Polycrystalline film interpenetrating network i t t ti t k Nanocrystalline film; dye-sensitized

PhenylenePhenylene vinylidene Ru dye-TiO2

Quan um n r n Quantum conversion a Threshold process

Solar Energy Spectrum

ultra ultra- visible -violet (UV)

Infra Infra-Red (IR)

Effect of bandgap on efficiency

Theoretical limit of single junction solar cells as function of bandgap for AM 1.5 spectra along with best efficiency for some materials

PV best cell efficiency time lines

25 20
crystalline Si amorphous Si nano TiO2 (dye) CIS/CIGS CdTe

E Efficienc (%) cy

15 10 5 Cu2S/CdS
nanoporous cell

1950

1960

1970

1980

1990

2000

Year

Japan Residential Systems: Cost History

Inverter Module
Mounting, Electrical Installation, Installation Engineering

For Si, focus on module R&D...

Inverter Module
Mounting, Electrical Installation, Installation Engineering

World Solar Module Production and Cost

World PV prices and production 40 35 vo olume sold (MW) 30 cost (U US$/Wp) 25 20 15 10 5 0 0 1975 1980 1985 1990 1995 2000 2005 year 400 $/Wp MW 800 1200

Increase in production accompanied by decrease in cost Is it really good news?

Dominates commercial terrestrial PV market Use purified Si..

grid coverage, reflection losses, spurious absorption p p

Lets look at crystalline Silicon (c-Si)

Single crystal or large grain multi-crystalline

h< Eg

Si cell power loss distribution @ AM1 d b AM1 M1 M

h> Eg
quantum efficiency

Shockley-Queisser limit Engineering achievement of 25 years f

Martin Wolf

Eg> qVOC

fill factor

Energy Convers. 11 (1971) 63

final cell

Crystalline Silicon: PERL Cell

Passivated emitter and rear locally diffused lower diffused, recombination at back surface: higher Voc and FF

What else can we do: III-V Based PV

III-Vs III V can be alloyed for: n b ll d f :
wide range bandgap engineering 0.5 eV - over 2 eV lattice matching to form complex device structures

High cost of fabrication for terrestrial applications. Maybe for concentrators High performance multijunction cells > 35% efficiency, used in space.

To better use solar spectrum -> mult junct on device multi-junction dev ce structures
Thermodynamic Efficiency Limits non-concentrated Sunlight (AM 1.5)
# of Junctions 1 2 3 4 infinite Efficiency 30% 42% 49% 53% 68% Optimum EG (eV) 1.3 1.9 - 1.0 2.3 - 1.4 - 0.8 2.6 1.8 1.2 0.8 26-18-12-08

Improve performance using concentrated sun light

farewell to most .. diffuse radiation?

DARPA: Very High Efficiency Solar Cell

Goal: 50% Efficient Solar Module
Prototype-to-be: 0.5 W, 10 cm2 Reduce weight of b tt i carried b soldier .. R d i ht f batteries i d by ldi

Back to earth PV Growth Drives Si Feedstock Production

PV Growth Drives Si Feedstock Production

Losing 2-3% in (absolute) efficiency g ( ) y negates any cost advantages for mg-Si Anticipate continuing progress down to 150 m with current processes Thinner wafers (100 150 m) will require (100-150 significant innovation and new processes Based on module spec sheets, efficiencies average 12.5- 13.5% The average increased ~1% / 5 yr Can we assume this rate to accelerate?

What aboutThin Film PV ?

Challenges/Opportunities
c Si c-Si technology approaches lower limit for production costs ?? Si shortage gave 2-4 year window of opportunity .. Low cost potential of Thin-film PV modules requires >100 MW factory size Critical Issues Equipment largely unique and custom designed. q p g y q g Lab -> large-scale manufacturing much more difficult than expected. Limited Science and Technology base impedes scale up (WIS) Lower performance than c-Si is probably acceptable f BIPV etc., L f th Si i b bl t bl for t IF cost gain offsets BOS costs. Lower performance disadvantage for fixed-area applications like rooftops (dominates market!)

Thin Films vs. Crystalline Si

Thin Films
Highest efficiency: 19.5%

Crystalline Si
Highest efficiency: 24.7% in l 24 7% single crystal t l 19.8% multi-crystalline g High materials cost wafer thickness > 300 m poor material use due to sawing d

Low material costs film thickness < 5 m high utilization

High throughput continuous g manufacturing Monolithic integration

Large # of processing steps Wafer binding Module assembly Established market

Thin Film PV: Basics

1.0 3.0 0.8 b 0.6 Q QE 0.4 0.2 0.0 400 600 800 1000 1200 1400 Wavelength (nm) a c d e Energy (eV) 2.0 1.5 1.0

QE of high efficiency TFSCs: (a) a-Si thin top cell of a triple junction with no back reflector; (b) a-SiGe, bottom cell with graded bandgap and back reflector; (c) CdTe; (d) Cu(InGa)Se2 with Ga/(In+Ga)= 0.3 and Eg 1.2 eV;4 (e) CuInSe2

Why a-Si:H? Strong optical absorption (-> thin film); alloys cover much of solar spectrum Can be doped both p- and n-type Performance (stabilized) Highest lab./ module efficiency : 13.3% / 11% (??) Typical module efficiency: 6-7% Less temperature-sensitive than c-Si (10-20% more kWhr / kW) Manufacturing Plasma CVD -> Low temp. processing and minimal processing steps m mp p g m m p g p On many substrates, e.g., polymer, stainless steel -> OK for BIPV Stability degrades during first 100 hours of light exposure by 10-20% (rel.) (rel ) Multi-junction structures improve efficiency and stability Well-characterized, reversible, self-limiting; partial summer annealing Improve further: nano-crystalline-Si embedded in a-Si tissue (HUJI)

a-Si and Related Alloys

Why CdTe? y By alloying with Hg, Zn, S, can cover much of solar spectrum Strong optical absorption (-> thin film) Can be doped both p- and n-type p p yp Performance Highest lab. /module efficiency: 16 5% / 10 7% lab 16.5% 10.7% Typical module efficiency: 9.0% Several Manufacturing Methods (can be up scaled low-cost potential) up-scaled, low cost Vapor transport (at First Solar) Very high growth rate @ high temperature In-line processing n l ne process ng Electrochemical deposition Batch: many plates at once, very slow, but 100% utilization, Post-deposition processes activate junction (critical and complex) p p j p Environmental problems associated with Cd manageable

CdTe PV

CuInSe2 Based PV
Why CuInSe2? Wh C Very strong optical absorption (-> thin film) (Tolerant in terms of manufacturing Alloys can cover part of solar spectrum Intrinsically stable, self-healing -> good module outdoors stability selfPerformance Highest lab. efficiency:19.5% ( multi-crystalline Si) lab efficiency:19 19. (~ multiHighest module efficiency:13.4% (Showa Shell Cu(InGa)(SeS2)) efficiency:13. Manufacturing f Glass or flexible foils or plastic substrates MultiMulti-source evaporation -> continuous in-line deposition inReaction of metal film or nano-particle precursor in H2Se or H2S nano-

Possibilities for Technological Progress

Eff.(%)
rated / minimum

Manufacturer

Technology

module/cell

BEST

17.7 / 16.3 17.0 15.3 17 0 / 15 3 15.5 / 13.5 14.4 / ??? 13.3 / ??? 13.2 / 11.9 11.0 / 10.3 9.4 8.3 94/83 6.4 /??? 6.4 / 5.7

SunPower Sanyo BP Kyocera Shell Schott

Single-crystal Si non-standard jnctn Single-crystal Single crystal Si non standard jnctn non-standard Single-crystal Si non-standard jnctn Multi crystal Si standard jnctn Single-crystal Si standard jnctn g y j EFG(ribbon) Si standard jnctn

72% 69% 63% 68% 63% 62% 56% 57% 64% 53%

WuerthSolar CIGS First Solar CdTe Mitsubishi Uni-Solar a-Si, single jnctn a-Si, triple jncnt

Energy Pay-back Time for PV Cells

PV M d l P d ti E Module Production Experience ( L i (or Learning) Curve i ) C

1976 80% Learning Curve: Module price decreased by 20% for every doubling of cumulative production

2005

80%

PV M d l P d ti E Module Production Experience ( L i (or Learning) Curve i ) C

1976 80% Learning Curve: Module price decreased by 20% for every doubling of cumulative production

2005

Crystalline Silicon 35% per year growth

80%

2010

2015

PV M d l P d ti E Module Production Experience ( L i (or Learning) Curve i ) C

1976 80% Learning Curve: Module price decreased by 20% for every doubling of cumulative production

2005

80% 2010 2015

PV M d l P d ti E Module Production Experience ( L i (or Learning) Curve i ) C

1976 80% Learning Curve: Module price decreased by 20% for every doubling of cumulative production x-Si module prices are expected to increase to ~2008, then decrease 2005

90% 80% 2010 2015 80%

PV M d l P d ti E Module Production Experience ( L i (or Learning) Curve i ) C

1976 80% Learning Curve: Module price decreased by 20% for every doubling of cumulative production

2005

90%

Evolutionary x-Si technology advances: x Si advances Thinner wafers (to 150 m) Improved processing Improved performance Lower feedstock prices Needed for ~80% learning curve

80% 2010 2015

80%

PV M d l P d ti E Module Production Experience ( L i (or Learning) Curve i ) C

1976 80% Learning Curve: Module price decreased by 20% for every doubling of cumulative production

2005

90%

Disruptive silicon technologies: Much thinner wafers (<100 m) Much improved performance (>25%) Innovative processing and processes Needed for ~70% learning curve

80% 2010 2015

80% 70%

PV M d l P d ti E Module Production Experience ( L i (or Learning) Curve i ) C

1976 80% Learning Curve: Module price decreased by 20% for every doubling of cumulative production

Thin Film Technologies

2005

2005

80%

PV M d l P d ti E Module Production Experience ( L i (or Learning) Curve i ) C

1976 80% Learning Curve: Module price decreased by 20% for every doubling of cumulative production

Thin Film F m Technologies

2005

2005

80%
50% per year growth

2010

2015

PV Module Production Experience (or Learning) Curve M d l P d ti E i ( L i ) C

1976 80% Learning Curve: Module price decreased by 20% for every doubling of cumulative production

Thin Film Technologies

2005

2005

90% 80% 80%

2010

2015

PV Module Production Experience (or Learning) Curve M d l P d ti E i ( L i ) C

1976 80% Learning Curve: Module price decreased by 20% for every doubling of cumulative production

Thin Film Technologies

2005

2005

90% 80% 80%

Significant opportunities for evolutionary and disruptive advancements: Improved performance and reliability Improved processes and processing Reduced manufacturing costs

2010

2015

PV Module Production Experience (or Learning) Curve M d l P d ti E i ( L i ) C

1976 80% Learning Curve: Module price decreases by 20% for every doubling of cumulative production

Thin Film Technologies

2005

2005

90% 80% 80% 70% 2010 2015

Significant opportunities for evolutionary and disruptive advancements: Improved performance and reliability Improved processes and processing Reduced manufacturing costs

PV M d l P d ti E Module Production Experience ( L i (or Learning) Curve i ) C

1976 80% Learning Curve: Module price decreases by 20% for every doubling of cumulative production

Concentrators

2005

2005

80%

PV M d l P d ti E Module Production Experience ( L i (or Learning) Curve i ) C

1976 80% Learning Curve: Module price decreases by 20% for every doubling of cumulative production

Concentrators

2005

2005

80%

2010

WGA Solar Task Force Report

2015

PV M d l P d ti E Module Production Experience ( L i (or Learning) Curve i ) C

1976 80% Learning Curve: Module price decreases by 20% for every doubling of cumulative production

Concentrators

2005 80% 0

2005

90% 80%

2010

2015

PV M d l P d ti E Module Production Experience ( L i (or Learning) Curve i ) C

1976 80% Learning Curve: Module price decreases by 20% for every doubling of cumulative production

Concentrators

2005

2005

90% 80%

Significant potential for evolutionary g p y 80% and disruptive advancements: Low manufacturing costs with scale-up Very high cell efficiencies Improved reliability p y 2010 Innovative concentrating concepts

2015

Critical Engineering R&D Issues

CdTe-based solar cell technology Develop ways to increase module performance from ~9% ->15% Reduce CdT thickness R d CdTe thi k CuInSe2-based solar cell technology Pathway to 15% module known but need to develop engineering base to transfer laboratory results to manufacturing ( yield ! ) Develop wide bandgap CuInSe2 based materials for improved module performance and potential for multi-junction structures Si-based solar cell technology Develop high rate a-Si deposition technology for multi-junction structures using nano-c-Si bottom cell Develop heterojunction c-Si cells using low temperature c Si processing to facilitate use of thin, < 150 m, Si wafers

Critical Engineering R&D Issues

Basic research for PV technologies
Needed to underpin manufacturing capability and progress: c-Si has largest science base, but significant advances over past 5 - 10 years came from R&D a-Si has reasonable science base, but significant advance in , g high rate manufacturing & stability from R&D CdTe, CuInSe2, Cu2S, Dye cell-based thin film materials: ALL have limited scientific b h limit d i ntifi base; p im ; primary use for PV; f --> no complimentary R&D to augment development --> R&D critical to their long term success g

PV M d l P d ti E Module Production Experience ( L i (or Learning) Curve i ) C

1976 80% Learning Curve: Module price decreases by 20% for every doubling of cumulative production

Efficiencies beyond the Shockley-Queisser Limit

(40% 50% b beyond?) d?)
2005

Revolutionary Generation) Technologies: Quantum dots Nanotechnology Multi-multi-junctions Thermophotonics, Thermophotonics voltaics Intermediate band . . .

(3rd

80%

World PV module production, is growing ~~ exponentially (1.65 GW; 2005) (1

But . PV module manufacturing cost decreases asymptotically in past 12 yrs $ 1/Wp ( 10 /kWh @ 10 yr lif ) (~10 /kWhr life)

NEED BREAK THROUGH(S) LATERAL BREAK-THROUGH(S), THINKING

Time-line of Photovoltaic Devices

What is the hope for Photovoltaics?
Efficiencies for different types of solar cells
III-V, incl. tandem cells

NEW materials and approaches! h !

Dye cells

Make better use of sunlight bad materials good cell fabrication easy HOW? Smarter Optics Self-healing materials Self-assembly Solar paint

NEEDED New Materials, Better Optics Materials

Roll to roll

Rethink established approaches