Current Status of Solar Photovoltaic Technology Platforms, - - PowerPoint PPT Presentation

Photonic Spectra Webinar “PV manufacturing and research” Hegedus 09/27/12 #1

Current Status of Solar Photovoltaic Technology Platforms, Manufacturing Issues and Research

Steve Hegedus Institute of Energy Conversion University of Delaware

With assistance from IEC staff: Brian McCandless (CdTe), Bill Shafarman (CIGS)

Photonic Spectra Webinar “PV manufacturing and research” Hegedus 09/27/12 #2

Outline

 Introduction to IEC at U of Delaware  PV trends, growth in scale, contribution to energy production  Crystalline Si PV status, baseline technology, near term focus:

• New methods emitter formation, passivation and device

architecture

 Thin Film PV status, baseline technology and near term focus:

• Cu(InGa)Se2 : wide gap alloys, improved 2-step selenization
• CdTe: higher deposition T , substrate
• a-Si/nc-Si (briefly) : multijunction, collapse of the a-Si industry

 Two common PV myths

Photonic Spectra Webinar “PV manufacturing and research” Hegedus 09/27/12 #3

Institute of Energy Conversion at U of Delaware

 Founded in 1972 to perform thin-film PV research  World’s oldest continuously operating solar research facility  First 10% efficient thin film solar cell (1980)  Dept of Energy University Center of Excellence for Photovoltaic Research and Education (1992)  Soft funded - government and industry contracts  2012 staff: 11 professional, 3 tech, 2 admin, 5 post doc, >20 grad students (4 depts)  Recently rec’d $8.4M from DOE (3 year grants)

First flexible 10% cell 4x4 inch minimodule

Photonic Spectra Webinar “PV manufacturing and research” Hegedus 09/27/12 #4

IEC Research Program Goals

 Expand the fundamental science and engineering base for thin film and c-Si photovoltaics to improve performance  Transfer these technologies to large-scale manufacturing

• IEC has been responsible for growth of several PV start-ups

through technology transfer and validation

 Provide workforce with PV scientists and engineers

• >40 graduates since 1992 (PV Center of Excellence)

Photonic Spectra Webinar “PV manufacturing and research” Hegedus 09/27/12 #5

IEC Technology Thrust Areas

 Thin film polycrystalline CuInGaSe2-based (CIGS) solar cells  Thin film polycrystalline CdTe solar cells  Silicon-based solar cells

• Front and back contact heterojunction (a-Si/c-Si)
• Thin film tandem a-Si and nc-Si

Photonic Spectra Webinar “PV manufacturing and research” Hegedus 09/27/12 #6

IEC Facilities: complete capability for fabrication and characterization of thin film and c-Si solar cells

 Over 20 thin-film deposition systems: PECVD (vhf/rf/dc), HWCVD, PVD, Vapor Transport, sputtering, H2S/H2Se reaction, chemical bath  Materials characterization: XRD, GIXRD, VASE, EDS, SEM, AFM, AAS, XPS, FTIR, Raman, optical trans+refl, Hall effect  Device fabrication: complete capability for high efficiency solar cells: c-Si (front heterojunction and IBC), CdTe, Cu(InGa)Se2 , a-Si  Laser and mechanical scribing for monolithic module fabrication  Device characterization: J-V, J-V-T, QE, C-V, OBIC, accelerated life stress (damp-heat, ambient, light)

Photonic Spectra Webinar “PV manufacturing and research” Hegedus 09/27/12 #7

The Big Picture: PV applications and achievements

Photonic Spectra Webinar “PV manufacturing and research” Hegedus 09/27/12 #8

PV can be installed anywhere, 10’s Watts to 100’s Megawatts

Photonic Spectra Webinar “PV manufacturing and research” Hegedus 09/27/12 #9

Recent worldwide achievements

 Installation: 17 GW in 2010 (100% growth), 30 GW in 2011 (70% growth)

• Average annual growth >50% p/y for decade

 EU: PV providing 2-4% of annual electricity in Spain, Germany, Italy

• May 2012 Germany received >10% from PV
• On one day >40% (22 Gigawatts peak supply out of 27 GW installed)

 US: 5.7 GW installed, 2 GW in CA

• Over 70% of 2011 installations are ‘utility scale’ or > 100 kW
• Worlds largest PV power plant 250 MW Aqua Caliente Project (CA, thin film)

 Creating hundreds of thousands of jobs

• 400,000 in Germany; 100,000 in US
• R&D, manufacturing, supply chain (materials), system design, installation

Photonic Spectra Webinar “PV manufacturing and research” Hegedus 09/27/12 #10

Trend in PV applications: 1990-2009

 1995: PV demand driven by off-grid applications  After 1995: Innovative policy in Japan, Germany stimulated market for grid-connected residential and commercial  >2008: Asian Si modules drove down prices, increased installations  >2010: Significant growth in utility scale > 1 MW projects  2012: First year of flat or  negative growth in decades, projected to recover 2013

Photonic Spectra Webinar “PV manufacturing and research” Hegedus 09/27/12 #11

Industry in turmoil: ‘roller coaster ride’

 Significant consolidation, bankruptcy, closures in past 2 years  Top companies for years suddenly quit PV or bankrupt  Worldwide capacity ~ twice demand yet demand still growing  Huge excess inventory  Shrinking profits - many companies selling at loss to compete  c-Si done much better in price and efficiency than many expected, squeezing thin film start ups  Renewed emphasis on improving performance since costs so low

Photonic Spectra Webinar “PV manufacturing and research” Hegedus 09/27/12 #12

Brief Overview of PV Basics

Photonic Spectra Webinar “PV manufacturing and research” Hegedus 09/27/12 #13

What is a PV device?

Direct converter of light into electricity: photons in, electrical current (DC) out Three critical processes:

Light Absorption + Carrier Generation + Carrier Collection (current flow) (deliver P to load)

e - h+ e - h+ load

charge separation V+ V-

Photonic Spectra Webinar “PV manufacturing and research” Hegedus 09/27/12 #14

Cell efficiencies vs. bandgap EG

 Record performance single junction cells vs. theoretical limit

• Expect maximum

performance with EG ≈ 1.5 eV

• But theor eff >25%

possible EG ≈ 1 – 1.8 eV

• Many thin film, III-V options

a-Si/nc-Si 2J

Photonic Spectra Webinar “PV manufacturing and research” Hegedus 09/27/12 #15

Commercial Scale PV Devices

 Single crystal or multicrystalline Si wafers

• Dominate market: 85-90% of sales
• Solar grade Si, lower qual than IC
• Module efficiency: 14-20%
• Low cost Asian Si driving prices down

 Thin films (1-3 µm polycrystalline or amor)

• Ultimately lower cost than Silicon wafers (??)
• On glass, metal or plastic foils
• Diverse materials, techniques
• Lower quality, imperfect crystallization, more defects
• Module efficiency: 8-14%
• Unique advantages in building integrated products
• 10-15% of market, #2 PV company is TF CdTe (First Solar)

Photonic Spectra Webinar “PV manufacturing and research” Hegedus 09/27/12 #16

One common PV challenge: reducing gap between champion cell and module efficiency

Wolden et al, JVST-A 29 (2011) 030801

Multi c-Si and TF CIGS both ~20% cell efficiency. But mc-Si has more mature module technology.

Photonic Spectra Webinar “PV manufacturing and research” Hegedus 09/27/12 #17

Why efficiency matters – fixed BOS costs

Wang et al Renewable and Sustainable Energy Rev (2011)

Levelized cost of Energy:

• Lifecycle costs/energy
• LCOE costs include

Balance of System which scale with # modules, area

• Lower eff = higher BOS$
• More rack, wiring, install $
• y-intercept is system price

without module

• Si modules ‘selling’ at $0.9/W

Or $120/m2

Photonic Spectra Webinar “PV manufacturing and research” Hegedus 09/27/12 #18

Crystalline Si (c-Si) Technology and Advanced Concepts

Photonic Spectra Webinar “PV manufacturing and research” Hegedus 09/27/12 #19

Standard commercial Si PV cell process

start finish

900°C 400°C 900°C

Photonic Spectra Webinar “PV manufacturing and research” Hegedus 09/27/12 #20

Commercial Si Solar Cell, Eff ~ 15-17%

Front contact (Screen printed Ag fired through SiN) (~ 0.3 µm)

Random textured F and R

Photonic Spectra Webinar “PV manufacturing and research” Hegedus 09/27/12 #21

World record Si solar cell: PERL

Cell Voc (V) Isc (mA/cm2) FF(%) Eff(%) PERL 0.70 42 81 24.7

• PERL cell: Passivated Emitter, Rear contact Locally diffused
• 2-step emitter (thin n between contact and thick n+ under contact)
• UNSW, AU, 1998; very complex design, not manufacturable

Photonic Spectra Webinar “PV manufacturing and research” Hegedus 09/27/12 #22

Conflicting emitter properties: pn junction vs resistance vs absorbing ‘dead layer’

property Advantage Disadvantage Increase thickness

• Reduce lateral R for

current flow to Ag contact

• Prevent melting Ag SP

metal penetrate to base

• Increase absorption in

highly defect layer (photons not converted to e-h pair), lower blue QE and Jsc Increase doping

• Reduce lateral R
• Reduce contact R with

Ag or other metal grid

• Increase defects and

recombination (Io) so decrease Voc Increase bandgap

• Decrease absorp loss
• Increase band bending,

reduce recomb (Io)

• Increase lateral

resistance significantly, req second conductive layer ($$)

Photonic Spectra Webinar “PV manufacturing and research” Hegedus 09/27/12 #23

Why not ‘tune’ the emitter to have properties it needs only where it needs them?

 Why not a 2 step emitter – spatially specific?

• different thickness and doping where needed?
• acknowledge that current flow is 2D not 1D

 Why not replace with wider bandgap material?  Why not get rid of the emitter all together* * On the front of the device

Photonic Spectra Webinar “PV manufacturing and research” Hegedus 09/27/12 #24

Industrially proven high efficiency Si solar cell concepts

 Three commercial proven enhancements (full size wafer results)

• PERL/SE: passivated+selective emitter, rear localized contact
• HIT: ‘HJ intrinsic thin’ a-Si/c-Si heterojunction
• IBC: ‘interdigitated back contact’ rear emitter and base contact

Photonic Spectra Webinar “PV manufacturing and research” Hegedus 09/27/12 #25

Hi Eff concept #1: the 2-step ‘selective emitter’

Conventional 1 step thick emitter 2 step emitter: thinner n+ everywhere except under metal n++ Applied Materials website Increased blue response with thinner n+

Photonic Spectra Webinar “PV manufacturing and research” Hegedus 09/27/12 #26

SE option 1: laser doping + plating metal

http://www.photonics.com/Article.aspx?AID=40098

Photonic Spectra Webinar “PV manufacturing and research” Hegedus 09/27/12 #27

SE option 2: deposit thicker n++ then etch back

• requires alignment of front metal

to thicker n++ mesa

Photonic Spectra Webinar “PV manufacturing and research” Hegedus 09/27/12 #28

Thinner vs selective emitter: Voc, Isc, FF, Eff

Gauthier “Industrial Approaches of Selective Emitter on Multicrystalline Silicon Solar Cells” 24th Eu-PVSEC (2009) 2-CV.5.46

Photonic Spectra Webinar “PV manufacturing and research” Hegedus 09/27/12 #29

Hi Eff #2 Heterojunction Solar Cell: deposited a-Si passivation layers reduce surface recombination

 Device: n-type c-Si wafer and 5-10 nm PECVD a-Si layers (EG=1.7-1.8 eV)

(i) a-Si:H surface passivation layers (both sides) (p)a-Si:H emitter (front) (n)a-Si:H back contact (rear) All a-Si and contacts deposited <200°C (low $, less defects, no warping thin wafers)

 High efficiency and VOC (Sanyo/Panasonic): η = 23% champion cells, 19% modules VOC = 740mV

10nm (i)a-Si:H 30nm (n)a-Si:H 10nm (p)a-Si:H

60nm ITO

300μm n-c-Si wafer

Rear Contact (Al) Front Contacts (Ag)

Hole Current Electron Current EV EF EC

hυ

Photonic Spectra Webinar “PV manufacturing and research” Hegedus 09/27/12 #30

Two modes of c-Si Surface Passivation by a-Si:H

c-Si Substrate a-Si:H film EF EG,c-Si EG,a-Si:H EC EV PECVD a-Si:H provides best passivation and processing < 300 °C, high VOC Si surface cleaning critical to good passivation Well-established, slow deposition rate, easily control thickness ~ 5-10 nm

Defect neutralization by H atoms:

Reduce c-Si surface dangling bonds Reduce recombination (IO), increase VOC

(n)c-Si Substrate (i)a-Si:H film EF EG,c-Si EG,a-Si:H EC EV

Field effect passivation:

Increased band bending at junction Repel/separate majority or minority carrier Reduce Io, increase VOC

Photonic Spectra Webinar “PV manufacturing and research” Hegedus 09/27/12 #31

Hi Eff Si #3: Interdigited back contact (IBC) cell

Photonic Spectra Webinar “PV manufacturing and research” Hegedus 09/27/12 #32

Standard front junction vs all back contact

Photonic Spectra Webinar “PV manufacturing and research” Hegedus 09/27/12 #33

IBC spectral response higher in short (blue) and long (IR) wavelengths

Photonic Spectra Webinar “PV manufacturing and research” Hegedus 09/27/12 #34

Sunpower IBC: highest efficiency module

Photonic Spectra Webinar “PV manufacturing and research” Hegedus 09/27/12 #35

Integration of both device concepts: SHJ-IBC Cell

Interdigitated back contact (IBC) solar cell. n-type c-Si

intrinsic a-Si p-type a-Si TCO intrinsic a-Si n-type a-Si TCO

Silicon heterojunction (SHJ) solar cell.

IBC-SHJ solar cell (IEC structure)

First published results on SHJ-IBC By IEC (APL 2007)

• intrinsic a-Si buffer
• separate a-Si p

and n regions

Photonic Spectra Webinar “PV manufacturing and research” Hegedus 09/27/12 #36

IEC Multichamber PECVD for HIT and IBC-SHJ

4 chambers plus 2 load lock, DC/RF/VHF plasma Multiple substrate sizes (1x1 up to 12x12 inch)

Photonic Spectra Webinar “PV manufacturing and research” Hegedus 09/27/12 #37

Thin Film PV

 Common Features

• Monolithic Integration via laser patterning: enabling

technology

• Lower efficiency: TF PV best suited for BIPV, large power

plants

 Status and Critical Issues

• Cu(InGa)Se2
• CdTe
• A-Si/nc-Si (briefly)

Photonic Spectra Webinar “PV manufacturing and research” Hegedus 09/27/12 #38

Monolithic Series Interconnection : laser scribe

 Allows structuring of large area uniform thin film layers into series connected junction diodes; critical technology for TF PV  Three laser scribing steps (patterning steps P1, P2, P3)

• P1) bottom conductor; P2) semiconductor junction; P3) top conductor

 Width of cells determines module current (ISC), # in series determines VOC

Chapter 12, Handbook of Photovoltaic Science and Eng (Luque, Hegedus), Wiley 2011

Photonic Spectra Webinar “PV manufacturing and research” Hegedus 09/27/12 #39

TFPV applications: BIPV

 Appearance preferred for building-integrated PV

85kW Shell Solar Cu(InGa)Se2 in Wales Semitransparent a-Si Architectural skylight

Photonic Spectra Webinar “PV manufacturing and research” Hegedus 09/27/12 #40

 Flexible PV for roll-out rooftop installation (USSC triple junction/SS)

Flexible a-Si on SS: BIPV (flex laminate)

Photonic Spectra Webinar “PV manufacturing and research” Hegedus 09/27/12 #41

4 MW of CdTe installed by Tucson Electric

Photonic Spectra Webinar “PV manufacturing and research” Hegedus 09/27/12 #42

15 MW of 3Sun Tandem Thin Si in Altomonte, Calabria Italy

Photonic Spectra Webinar “PV manufacturing and research” Hegedus 09/27/12 #43

Help from Bill Shafarman

Institute of Energy Conversion University of Delaware

Cu(InGa)Se2

Thin Film Solar Cells

Photonic Spectra Webinar “PV manufacturing and research” Hegedus 09/27/12 #44

Why thin film CuInSe2 alloys for PV?

 Direct bandgap chalcopyrite materials with high absorption coefficient  Extraordinary compositional tolerance  Alloy with Ga, Al, Ag, S to engineer bandgap  improve performance, TF tandem  Can be deposited on glass or light flexible substrates: polymer, foils  Highest device and module efficiency of any TF PV technology  Multiple deposition technologies with promise of scalability  Attracted considerable private investor funding  Outdoor module stability demonstrated

Photonic Spectra Webinar “PV manufacturing and research” Hegedus 09/27/12 #45

Cu(InGa)Se2 Thin Film PV

 Performance

• Highest cell efficiency = 20.3% (ZSW 2010)

 Efficiency ≥ 18% from several laboratories  Sub-module eff. = 17.8% with area > 800 cm2 (Solar Frontier 2012)  12–14% module efficiency from companies worldwide

 Manufacturing

• Many companies with various approaches
• 1. Reaction of metal precursors (2-step)

 Low cost deposition of metals  Batch process: “selenization”

• 2. Multi-source evaporation (1-step)

 In-line process, high temp

Substrate Mo Cu(InGa)Se2 Grid CdS ZnO:Al

Photonic Spectra Webinar “PV manufacturing and research” Hegedus 09/27/12 #46

Cu(InGa)Se2 Optical Absorption

 High optical absorption of sunlight

• Direct bandgap
• Complete absorption in ~ 1 µm thickness (CdTe very similar)
• Reduces requirements for minority carrier transport

0.2 0.4 0.6 0.8 1 0.001 0.01 0.1 1 10 100 1000

R e la tive A b so rp tio n thickness (µm) Si CIGS

Photonic Spectra Webinar “PV manufacturing and research” Hegedus 09/27/12 #47

Cu(InGa)Se2 Grain Structure

 Films are polycrystalline with rough surface

• Grain size ~ 0.1 – 1 µm depends on deposition conditions

e.g. substrate temperature during evaporation

• But device performance is remarkably insensitive to grain size,

morphology

1µm

Cu(InGa)Se2 Mo

deposited at 400°C deposited at 550°C

Wilson, Birkmire, Shafarman Proc. 33rd IEEE PVSC (2008)

Photonic Spectra Webinar “PV manufacturing and research” Hegedus 09/27/12 #48

CuInSe2 Alloys with wider EG: Al, Ga, S

 CuInSe2 has EG=1.0 eV, limits eff., major focus 20 yrs is to raise EG  Wide range of ternary, quaternary alloy options

• Recent focus on alloys with Ga/(Ga+In), S/(S+Se), Ag/(Ag+Cu)

 Alloying changes EG also lattice constants (no epi!), band alignment

x = alloy fraction: Al/(In+Al) Ga/(In+Ga) S/(Se+S)

Ga/(In+Ga) a x=0 b x=0.24 c x=0.61

Photonic Spectra Webinar “PV manufacturing and research” Hegedus 09/27/12 #49

Increasing efficiency and VOC at higher EG

Increasing EG with alloy (Ga, Ag, S) to push efficiency at EG >1.3 eV Failure to capture benefit of larger EG due to VOC/ EG<1 is critical issue

Contreras et al, 37th IEEEE PVSC, Seattle 2011

Photonic Spectra Webinar “PV manufacturing and research” Hegedus 09/27/12 #50

Process 1: Elemental Co-evaporation

• Simultaneous delivery of elemental vapors

to hot substrate

• Makes higher efficiency devices than 2-step
• Independent control of each element

Ga/(In+Ga) gradient  bandgap gradient

To vacuum pump Pbase ≈ 1x10-6 Torr Prun ≈ 2x10-5 Torr

Substrate Heater Substrate at 400-600°C Film Growth Monitor Thermal Evaporation Sources for Cu, In, Ga, Se, (S)

Photonic Spectra Webinar “PV manufacturing and research” Hegedus 09/27/12 #51

IEC Cu(InGa)(SeS)2 Co-evaporation

 Five source system for Cu-In-Ga-Se-S

• Boron nitride Knudsen cells
• Typical temperatures

T(Cu) = 1350°C T(In) = 1000°C T(Ga) = 1100°C T(Se) = 300°C

• Source design and

T control are critical features

Photonic Spectra Webinar “PV manufacturing and research” Hegedus 09/27/12 #52

Process 2: 2-step Precursor Reaction

• Advantages: lower cost, higher uniformity + materials utilization
• Many precursor deposition options with Cu, In, Ga –

 sputtering – commercially available  electrodeposition – high utilization, non-vacuum, batch  ink printing – high utilization, non-vacuum, continuous

• Reaction in hydride gases (H2Se, H2S) or elemental vapors (Se,S)
• Multi-step reaction pathway to form Cu(InGa)Se2

Mo/Cu/Ga/In Mo/Cu(InGa)Se2 Reaction H2Se/H2S

• r Se/S

400 - 600°C

Photonic Spectra Webinar “PV manufacturing and research” Hegedus 09/27/12 #53

Cu(InGa)Se2 Precursor Reaction

 IEC H2Se / H2S reactor

• reaction in quartz tube of glass/Mo/Cu-In-Ga precursor layers
• atmospheric pressure with flowing H2Se / H2S / Ar / O2
• Temperature-time cycle critical to uniformity, manufacturibility

Photonic Spectra Webinar “PV manufacturing and research” Hegedus 09/27/12 #54

CIGS Module Process Flow

Clean Substrate Mo dc sputtering P1 laser scribe CIGS deposition P3 mech. Scribe EVA/glass lamination CdS bath deposition HR ZnO deposition P2 mech. scribe ZnO:Al deposition Module completion Attach buss bars

Photonic Spectra Webinar “PV manufacturing and research” Hegedus 09/27/12 #55

Manu- facturer Deposition Technology Champion product % apa Current nameplate capacty MW/a Deposition Process Pro´s Deposition Process Con´s comments Manz (Würth) Solibro 1-stage co- evaporation 15.1 14.4 30 120 Simpler, more advanced process Sacrifices efficiency Glass-glass Cd- buffer Global Solar Energy 3-stage co- evaporation 15 (cell) 13 (mod) Highest known TF module efficiency Complex process SS substrate. glass/polymer MiaSolé Reactive sputter 15.7 >40? good efficiency potential, rel. small capex exp. Complex process SS substrate, cut + stitch, glass-glass CdS-dry Solar Frontier, Stion/ TSMC 2-step: Sputter+ H2Se/S- Selenization SF: 17.8 SF: 14.1 (manu) Stion: 14.5 TSMC 15.1 980 5 + 135 +300 advanced process, potentially higher CapEx Sacrifices efficiency Higher OpEx than evap. glass-glass CdS glass-glass Cd-free Avancis, Hyundai 2-step: Sputter+Se- evap.+ RTP- cryst./H2S 30 + 100 +100 Glass-glass CdS (Cd-free) Solo Power 2-step electroplate- Selenization 15.1% cell 13.5% mod 30 + 400 Good metal utilization, rel. low CapEx Sacrifices efficiency SS substrate Polymer CdS

Source: Markus Beck / Photon San Francisco‚ Feb. 2012 plus adds from the author

CIGS thin film manufacturing status: > 30MW real capacity

Photonic Spectra Webinar “PV manufacturing and research” Hegedus 09/27/12 #56

Two Current research issues: Cu(InGa)Se2

 Fundamental understanding of relation between wide EG alloys and device performance with multisource evaporation  Control uniformity and reduce process time with Se/S reaction

Photonic Spectra Webinar “PV manufacturing and research” Hegedus 09/27/12 #57

• 1. New wide gap alloy (AgCu)(InGa)Se2
• Ag addition to Cu(InGa)Se2 lowers melting temperature, better surface

mobility, potential for improved structural hence electronic quality

• Ag increases bandgap by up to 0.25 eV,
• Single phase over entire range of Ag–Cu and Ga–In alloying

Photonic Spectra Webinar “PV manufacturing and research” Hegedus 09/27/12 #58

Notable wide EG cell results with (AgCu)

• Cells: SLG/Mo/(AgCu)(InGa)Se2/CdS/ZnO/ITO/grid/MgF2
• High Eff = 17.6% with Eg ≈ 1.3 eV

High VOC = 890 mV with Eg ≈ 1.6 eV

Hanket, et al., Proc. 34th IEEE PVSC

Photonic Spectra Webinar “PV manufacturing and research” Hegedus 09/27/12 #59

• 2. H2Se/H2S reaction of Cu-Ga-In Precursors:

Ga segregation at rear limits EG and Voc

10 20 30 40 50 60 20 40 60 80

composition (%) sputter tim e (m in) Se Cu I n Ga Mo S

10 20 30 40 50 60 20 40 60 80

composition (%) sputter tim e (m in) Se Cu I n Ga Mo S

• 30’ H2Se@450°C, 15’ H2S@450°C
• Complete H2Se reaction prior to H2S
• Ga segregated at back, none at front
• Low EG at front junction, low Voc
• 15’ H2Se@450°C, 15’ H2S@450°C
• Partial H2Se reaction prior to H2S
• Ga uniform, higher at front
• Higher EG at front junction, higher Voc

Hankett et al, Proc 4th World PVSEC, 2006

Photonic Spectra Webinar “PV manufacturing and research” Hegedus 09/27/12 #60

Single step H2Se process vs. Three-step H2Se/H2S process

550 Time, min Temp, °C

20 50 10 10

1st step : H2Se

20 10

400 Time, min Temp, °C

20

Single step: H2Se 450

2nd step : Ar 3rd step : H2S

60 ~ 90

Mo Mo

Ga accumulation Ga homogenization

Process optimization  Ga distributed uniformly with 3 step H2Se/H2S  Major improvement in VOC and Eff.

Single step process Three-step process

• 0.2

0.0 0.2 0.4 0.6 0.8

• 40
• 20

20 J (mA/cm

2)

V (V)

Eff Voc Jsc FF Single-step : 8.3% 0.383 V 37.0 mA/cm

2 58.7%

3-step : 14.2% 0.599 V 32.2 mA/cm

2 73.5%

Photonic Spectra Webinar “PV manufacturing and research” Hegedus 09/27/12 #61

CdTe Thin Film Solar Cells

Help from Brian McCandless

Institute of Energy Conversion University of Delaware

Photonic Spectra Webinar “PV manufacturing and research” Hegedus 09/27/12 #62

Why CdTe TF solar cells?

 Chemically stable, simple phase diagram, easy surface passivation  Film deposition by a variety of scalable techniques  Optimized cells require post-dep halide (Cl) exposure at ~ 400C  Easily adapted to monolithic integration  Low cost: thin absorber, low cap ex equipment costs, high dep rate  Best laboratory cell efficiency >17%, best module 14% (First Solar)  Presently lowest price PV available (First Solar)

• First Solar <$0.75/W manufacturing cost
• 2-3 hours from glass to module with 13% efficiency
• Responsible for US lead in module manufacturing
• Thin film success story!

Photonic Spectra Webinar “PV manufacturing and research” Hegedus 09/27/12 #63

Superstrate CdTe/ CdS cell configuration

Glass superstrate (SLG or advanced) SnO2 (comm or more complex TCO) HR layer (intrinsic TCO, ~ 20 nm) CdS (n-type, 30-50 nm) CdTe (p-type, 2-4µm) + CdCl2 step Contact (contains Cu as dopant)

Photonic Spectra Webinar “PV manufacturing and research” Hegedus 09/27/12 #64

CdTe/CdS heterojunction cell structure: grain boundaries, S-Te interdiffusion, QE

Contact ( Cu+ other) p-CdTe n-CdS Zn 2SnO4 Cd2SnO4 glass 400 500 600 700 800 900 1.0 0.8 0.6 0.4 0.2

Wavelength (nm) Quantum efficiency

CdCl2 HT as-deposited

CdS depletion CdTe1‐xSx alloy Bandgap shift

Photonic Spectra Webinar “PV manufacturing and research” Hegedus 09/27/12 #65

CdTe/ CdS m odule processing

Clean glass/ SnO2 Substrate Deposit CdS, CdTe First Scribe Post Dep CdCl2 Treat Back Contact Second Scribe Third Scribe Tabbing, Junct box Encapsulate Test CdTe Surface Etch

Photonic Spectra Webinar “PV manufacturing and research” Hegedus 09/27/12 #66

CdTe deposition technology

Atmospheric High Cl O2 Vacuum No Cl Low O2 T > 500°C T < 400°C Spray Screen Print CSS VT PVD ED rf Sp

Golden Photon Matsushita Abound, USF First Solar, Calyxo PrimeStar/GE, NREL, IEC Canrom BP Solar XunLight 26

Photonic Spectra Webinar “PV manufacturing and research” Hegedus 09/27/12 #67

IEC CdTe Research using Vapor Transport Dep

CdTe Source

Vapor Transport System CdCl2 Reactor

Develop process for Eff > 15% on moving glass substrate ( 3 cm/min) to provide basis for in-line manufacturing Optimize: CdTe deposition, substrate, contacts, CdCl2 annealing  Surface chemistry, interdiffusion, impurities, grain growth Currently IEC CdTe work is proprietary; evaluating alternative glass, TCO, buffer layers

Photonic Spectra Webinar “PV manufacturing and research” Hegedus 09/27/12 #68

CdTe film growth: TSS, thickness, rate

AFM 20 x 20 m

Substrate temp Film thickness Growth rate 500ºC 5 m, 8 m/min 570ºC 5 m, 8 m/min 550ºC 1 m, 8 m/min 550ºC 6 m, 8 m/min 550ºC 23 m, 8 m/min 550ºC 6 m, 81 m/min

Faceted morphology, grain size increases with substrate temperature Grain size increases proportional with CdTe thickness Grain size inversely proportional to CdTe growth rate

Photonic Spectra Webinar “PV manufacturing and research” Hegedus 09/27/12 #69

Fundamental CdTe issues

 Compensating defects → doping and defect control difficult

• Difficult to achieve NA>5E14 cm3

 High hole affinity → non-ohmic back contact (needs p+ Cu2Te)  Unable to increase Voc>0.85V which is only 60% of EG  Highest efficiencies with non-commercial substrate

• Replace SLG with BSG glass ($) : more transparent, higher T deposition
• Replace SnO2:F (FTO) with Cd2SnO4/Zn2SnO4 (CTO/i-ZnO)
• Higher CdTe dep T allows improved grain structure (higher Voc, FF)

Photonic Spectra Webinar “PV manufacturing and research” Hegedus 09/27/12 #70

New substrates enable higher CdTe dep T  higher Eff

[McCandless et al, to be submitted (2012)]

Commercial Tec10 SLG/SnO2:F/i-ZTO vs R&D GL/Cd2SnO4/i-ZTO

Record CdTe FF>81% With new substrates

13.5% 16.5%

Photonic Spectra Webinar “PV manufacturing and research” Hegedus 09/27/12 #71

Manu- facturer Deposit Tech. Champion

product % apa

Current nameplate capacity MW/a Depo Process Pro´s Depo Process Con´s module

First Solar Low Press Vapor Transport 14.4 mod

Champ lab cell: 17.3%

2,700 closing several lines

Simple and matured process, high thruput Global player, Quality ??

Glass- glass Primestar/ GE “Thermal evap.“ 12.8 mod 30, started construct 400 MW, on hold

?

• Techn. Status

?

Glass- glass Abound Solar (CSU) Low Press Thermal evap. 15.7 cell Started construct, Closed 2012

Glass in, module out. Lower efficiency?

Glass- glass Calyxo (Q-cell)

• Atmos. press.

thermal evap. 13.4%

Champion lab cell: 16.2%

80

Potential low cost, high rate Quality?

• Techn. Status

?

Glass- glass Source: Schock / SNEC Shanghai‚ May 2012 plus adds from Dimmler 38th IEEE PVSC

CdTe thin film manufacturing status

Photonic Spectra Webinar “PV manufacturing and research” Hegedus 09/27/12 #72

Brief discussion of a-Si based multijunction PV

Steven Hegedus

Photonic Spectra Webinar “PV manufacturing and research” Hegedus 09/27/12 #73

Why amorphous Si (a-Si:H) for thin film PV?

 Unique among thin film PV technologies

• Easily vary doping (p or n), bandgap, crystallinity (a-nc), thickness,
• Well-established commercially viable multijunction process (>20 yrs)

 Minimal deposition steps: PECVD plus sputter back contact (<200°C)  Highest cell efficiency (3-junction a-Si/a-SiGe/nc-Si) : 14% (stabilized)  Commercial module stabilized efficiency (2J a-Si/nc-Si) : 9-10%  Least difference between best cell and typical module efficiency: Oerlikon, Sharp, AMAT have tandem 12% cell and 10% module  Challenge: native and light induced defects low mobility+lifetime  limit efficiency; even as lowest cost PV in market, cannot compete

Photonic Spectra Webinar “PV manufacturing and research” Hegedus 09/27/12 #74

Multijunction multibandgap a-Si solar cells

Chapter 12, Handbook of Photovoltaic Science and Eng (Luque, Hegedus), Wiley 2011

‘micromorph’ a-Si/nc-Si tandem

• ptimum trade-off

between cost and efficiency

Eff=6-7% Eff=9-11% Eff=12-13-% Eff=12-13%

Photonic Spectra Webinar “PV manufacturing and research” Hegedus 09/27/12 #75

Current a-Si PV commercial status: deathwatch

 ~ 10-15 companies bought Oerlikon or AMAT turn-key fab lines 2007- 2009 during Si feedstock shortage and PV price increase

• Low efficiency, high cap ex, new a-Si PV at disadvantage
• Most now closed (bankrupt), few operating in Asia
• Subsequent c-Si overcapacity and PV module price collapse squeezed a-

Si from the cost side (its strength) now multi-Si comparable cost

• Appl Matl closed their a-Si fab line 2010, Oerlikon Solar closed 2012

 United Solar (USSC/ECD) flex roofing product: closed 2012 after 30 yr  Sharp (?), Panasonic, Kaneka, NexPower, Astraenergy-Chint; 3Sun (Italian JV Sharp-Enel) are leading players, ~9-10% efficient 2J

• 58 MW Sharp micromorph tandem installed in CA in 2012

Photonic Spectra Webinar “PV manufacturing and research” Hegedus 09/27/12 #76

Dispell Two PV Myths

Photonic Spectra Webinar “PV manufacturing and research” Hegedus 09/27/12 #77

I heard it takes more energy to make a solar module than they can ever produce?

• NO!!!!
• How many years of operation before reach energy break-even point?
• Energy payback is <1.5 years for today’s crystalline Si wafer modules,

even less for next generation thin film modules

• With 25 year warrantees, today’s PV modules will be net producers of

clean, CO2-free electricity for at least 23 years!

Photonic Spectra Webinar “PV manufacturing and research” Hegedus 09/27/12 #78

I heard you would have to cover the country with solar modules to make any ‘real’ energy?

Total energy: 200x200 miles (50% coverage) Electricity only: 100x100 miles (50% coverage)

Photonic Spectra Webinar “PV manufacturing and research” Hegedus 09/27/12 #79

Questions? Then buy this book!

2nd Edition (2011) of the most comprehensive PV book

• 1100 pages
• 6 new chapters
• 8 different cell technologies
• TCO’s
• performance characterization
• batteries, inverters, system
• policy, economics

Photonic Spectra Webinar “PV manufacturing and research” Hegedus 09/27/12 #80

Back-up slides

Photonic Spectra Webinar “PV manufacturing and research” Hegedus 09/27/12 #81

2012 TF PV industry expected production

200 400 600 800 1000 1200 1400 1600 First Solar - CdTe Solar Frontier - CIGS Sharp - thin Si Astroenergy - thin Si NexPower - thin Si Trony - thin Si Miasole - CIGS T-Solar - thin Si 3Sun - thin Si Solibro - CIGS 2102 Top Thin Film PV manufacturers 2012 MW produced 20 30 35 40 80 90 120 180 620 1530

Very fluid, obtained from Renewable Energy World on-line 6/4/2012

Photonic Spectra Webinar “PV manufacturing and research” Hegedus 09/27/12 #82

Double vs Triple Junction: United Solar/ECD

Guha, SPIE Photovoltaics 2009

Photonic Spectra Webinar “PV manufacturing and research” Hegedus 09/27/12 #83

Cadmium toxicity: perception problem

 Widely studied by Brookhaven Natl Lab and NREL  Cd: lung, kidney, bone carcinogen  CdTe: less soluble, more stable, less toxic  Cd is by-product of Zn mining  Choices: react Cd with Te and encapsulate behind glass in controlled environment and generate clean energy; or leave it in exposed ore tailings  Not released to environment during roof-top residential fire  EU: granted CdTe PV an exemption; politically vulnerable  US: CdTe PV not classified as toxic waste (EPA)  Japan, Korea: banned CdTe PV and closed research  First Solar recycling/insurance program, “behind the fence”  www.nrel.gov/cdte