Progress in Thin-Film Solar Cells from Inorganic to ...mx.nthu.edu.tw/~hytuan/courses-files/2012energy&environment/12.14.pdf · 新能光電 Sunshine PV Progress in Thin-Film Solar

新能光電Sunshine PV

Progress in Thin-Film Solar Cells from Inorganic to Organic Materials

William W. Hou

12.14.2012新能光電

Sunshine PV

��


Presentation Outline

1. Background– Energy need for next century– Market trend & strategy– Cost vs Efficiency

2. Current Status - CIGS– Material property– Processing technology

3. Emerging Technology– CZTS– Dye-sensitize solar cell

4. Feed in Tariff

2


Global Energy Consumption

3

0

10

20

30

1990 1995 2000 2007 2015 2020 2025 2030 2035

terawatt-year

Projections History

11.89 12.50 13.58

16.56

Projections History

18.18 19.75

21.36 22.96

24.71

Source: www.eia.gov/emeu/international

• averaged growth of 1.6TW every 5 years• need to install additional 8.2 TW by 2030


How to meet the increase in energy demand

• Coal – CO2 emission

• Oil – CO2 emission, limited resource

• Nuclear – Waste management issues– Public confidence– Shortage of supply: 18 new reactors every month.

4


Cost Comparison between Nuclear and Solar Energy

• In the region of North Carolina, USA, electricity cost from solar is cheaper than nuclear.

• Solar energy still required new grid structure to solve the intermittency in power generation.

5

SUMMARYSolar photovoltaic system costs have fallen steadily for decades. They are projected to fall even farther over the next 10 years. Meanwhile, projected costs for construction of new nuclear plants have risen steadily over the last decade, and they continue to rise.

In the past year, the lines have crossed in North Carolina. Electricity from new solar installations is now cheaper than electricity from proposed new nuclear plants.

This new development has profound implications for North Carolina’s energy and economic future. Each and every stakeholder in North Carolina’s energy sector — citizens, elected officials, solar pow-er installers and manufacturers, and electric utilities — should recognize this watershed moment.

SOLAR AND NUCLEAR COSTS – THE HISTORIC CROSSOVER 3

Solar-Nuclear Kilowatt-Hour Cost Comparison

0

5

10

15

20

25

30

35

1995 2000 2005 2010 2015 2020 2025Year

2010

Cen

ts p

er k

Wh

Solar PV Nuclear Solar Trendline Nuclear Trendline

Solar-Nuclear Kilowatt-Hour Cost Comparison

0

5

10

15

20

25

30

35

1995 2000 2005 2010 2015 2020 2025Year

2010

Cen

ts p

er k

Wh

Solar PV Nuclear Solar Trendline Nuclear Trendline

Figure 1: The Historic Crossover — Solar photovoltaic costs are falling as new nuclear costs are rising.1

The Solar PV least-squares trendline is fit to data points representing the actual cost of producing a kilowatt-hour in the year shown through 2010 and for cost projections from 2010 to 2020. The nuclear trendline is fit to cost projections made in the year shown on the x-axis of eventual kilowatt-hour cost if projects reach completion. See complete methodology in Appendix A.

NUCLEAR

SOLAR PV

History Projection

Source: www.ncwarg.org - “Solar and nuclear cost - the historic cross over” (2010)

http://www.ncwarg.org

http://www.ncwarg.org

新能光電Sunshine PV 6

1.22 x 105 TWper year

16.5 TWper year*

* Based on EIA analysis for 2007, (www.eia.gov/emeu/international)

http://www.eia.gov/emeu/international



14.9 TWper HOUR

* Based on EIA analysis for 2007, (www.eia.gov/emeu/international)

16.5 TWper year*




Market Trend

8


Growth of solar industry in different regions of the world

9Source: EIPIA energy outlook 2010

8 MARKET EVOLUTION

3MARKET EVOLUTION3.1. Historical PV market development

From the first PV applications in space to the GW systems planned today, more than 40years have passed. Over the last decade, PV technology has acquired the potential tobecome a major source of power generation for the world. That robust and continuousgrowth is expected to continue in the years ahead. At the end of 2008, the world’scumulative installed PV capacity was approaching 16 GW. One year later it was 23 GW. In 2010, almost 40 GW are installed globally and produce some 50 TWh ofelectricity every year.

In terms of global cumulative installed capacity, the EU leads the way with almost30 GW installed as of 2010. This represents about 75% of the world’s total cumulativePV capacity (up from 70% in 2009). Japan (3.6 GW) and the USA (2.5 GW) are someway behind, while China has already entered the Top 10 of the world’s PV markets and should reach its first GW in 2011. China is expected to become a major player inthe coming years, and the size of its domestic market so far is only a small indication ofits potential.

45,000

40,000

35,000

30,000

25,000

20,000

15,000

10,000

5,000

MW 0

China 893

1,191

1,844

2,727

3,622

29,252

39,529

APEC

Rest of the world

North America

Japan

EU

Total

Figure 1 - Evolution of global cumulative installed capacity - 2000-201020

103

9,5

29

373

718

1,427

1,744

2,632

16,006

22,90020

092

2,9

00

145

466

1,303

1,205

2,149

10,387

15,655

2008

15

,65

5

100

170

1,190

856

1,919

5,257

9,492

2007

9,4

92

80

112

1,128

645

1,708

3,307

6,980

2006

6,9

80

68

80

1,010

496

1,422

2,324

5,399

2005

5,3

99

64

66

1,000

379

1,132

1,319

3,961

2004

3,9

61

55

57

971

287

860

613

2,842

2003

2,8

42

45

49

894

222

637

414

2,261

2002

2,2

61

30

43

814

177

452

275

1,790

2001

1,7

90

19

38

758

146

318

181

1,459

2000

1,4

59

8 MARKET EVOLUTION

3MARKET EVOLUTION3.1. Historical PV market development

From the first PV applications in space to the GW systems planned today, more than 40years have passed. Over the last decade, PV technology has acquired the potential tobecome a major source of power generation for the world. That robust and continuousgrowth is expected to continue in the years ahead. At the end of 2008, the world’scumulative installed PV capacity was approaching 16 GW. One year later it was 23 GW. In 2010, almost 40 GW are installed globally and produce some 50 TWh ofelectricity every year.

In terms of global cumulative installed capacity, the EU leads the way with almost30 GW installed as of 2010. This represents about 75% of the world’s total cumulativePV capacity (up from 70% in 2009). Japan (3.6 GW) and the USA (2.5 GW) are someway behind, while China has already entered the Top 10 of the world’s PV markets and should reach its first GW in 2011. China is expected to become a major player inthe coming years, and the size of its domestic market so far is only a small indication ofits potential.

45,000

40,000

35,000

30,000

25,000

20,000

15,000

10,000

5,000

MW 0

China 893

1,191

1,844

2,727

3,622

29,252

39,529

APEC

Rest of the world

North America

Japan

EU

Total

Figure 1 - Evolution of global cumulative installed capacity - 2000-2010

2010

39

,52

9

373

718

1,427

1,744

2,632

16,006

22,900

2009

22

,90

0

145

466

1,303

1,205

2,149

10,387

15,655

2008

15

,65

5

100

170

1,190

856

1,919

5,257

9,492

2007

9,4

92

80

112

1,128

645

1,708

3,307

6,980

2006

6,9

80

68

80

1,010

496

1,422

2,324

5,399

2005

5,3

99

64

66

1,000

379

1,132

1,319

3,961

2004

3,9

61

55

57

971

287

860

613

2,842

2003

2,8

42

45

49

894

222

637

414

2,261

2002

2,2

61

30

43

814

177

452

275

1,790

2001

1,7

90

19

38

758

146

318

181

1,459

2000

1,4

59

• Since 2000, there is a total of ~40GW of solar energy installed worldwide.


Global Solar Cell ProductionGlobal Solar Cell Production

Estimated Actual Production

IEK/Photon (2010/4)

Production capacity of different solar technology

IEK/Photon (2010/04)


Race to Grid Parity

11

DELOIT TERE VIE W.COM Deloitte Review

59SOL AR’S PUSH TO REACH THE MAINSTREAM

Solar Incentives: a Pareto Improvement

In most areas of the world, solar energy is too expensive to compete directly with

traditional fossil fuels — primarily because the industry is relatively immature and

lacks economies of scale. Grid parity refers to the point at which the cost of solar

electricity (or other alternative energy source) rivals that of our more traditional

sources, such as coal, oil, natural gas or nuclear. While many areas of the United

States are expected to reach this point for solar by 2015, grid parity actually varies

geographically as a function of local climate, utility rates and government sup-

port, to name a few. In the United States, residential electricity rates vary between

about $0.07 to nearly $0.25 per kWh;4 consequently, different areas will reach grid

parity at different times. Globally, the variance is even more pronounced, with a

significant portion of the world’s population not served by a grid at all.

The solar industry faces a chicken and egg dilemma: lower costs will stimulate

demand and increase economies of scale, but achieving lower costs now requires

scale — or market subsidies. To achieve grid parity, industry players need to sur-

vive the current economic recession, reduce operating costs, and pursue technolog-

ical advancements. Simultaneously pursuing all three initiatives will be difficult;

Projected PV Solar Market Penetration

Falling PV system prices will spur a sharp increase in demand from 2010 to 2020

Source: U.S. Department of Energy. Solar Energy Industry Forecast (2008)

Source: US DOE Solar Energy Forecast (2008)

• Grid parity = cross-over when the price of solar energy equals that of grid price.• At grid price of $0.1/kWh, the module cost would required to fall below $1/Wp.


Cost is the most critical factor of success in solar business

• Grid parity < $1/Wp

• Efficiency is a part of the competitive scheme.

• Success for a solar manufacturer relies on a competitive cost structure instead of demonstrating highest efficiency.

• Unlike the IC industry, customers are unwilling to pay more for most advanced technology.

12

Cost

Material

Equipment Efficiency


Advantage of Thin-Film

13

Thin-Film

Crystalline Si

Why Thin Film for The Future?Why Thin Film for The Future?yyTF has much shorter value chance, much less

materials but greater challengeFeedstockCu/In/Se/Ga

Module System

materials, but greater challenge.

TF:

Si:Si:

13

Feedstock

Why Thin Film for The Future?Why Thin Film for The Future?yyTF has much shorter value chance, much less

materials but greater challengeFeedstockCu/In/Se/Ga

Module System

materials, but greater challenge.

TF:

Si:Si:

13

• Thin-film solar cells required a shorter value chain• Long value chain presents difficulties with reaching grid parity.



Market size of different solar technologies in 2009

15

• Both poly-Si and mono-Si combined a total 82%

• First Solar’s CdTe was the first company to exceed 1GW/yr production, captured 13% market share.

• CIS/CIGS only has 1% market share.

issue of material availability for existing technologies.Lastly, recommendations are made for research investment,with an emphasis on those areas that are expected to havecross-cutting impact.

II. PRESENT STATUS OF PV MANUFACTURING

This article is not intended to provide a comprehensivereview of the history of photovoltaics but rather to update thesubstantial changes that have occurred since the turn of thecentury. This date was chosen for a couple of reasons. First,there are a number of very good reviews detailing the growthand evolution of the PV industry up to around 2003.6,8,9

Moreover, there have been two significant changes that haveaccompanied the growth experienced during the past 7 years.In 2003 various forms of silicon accounted for !99% of thePV market, and manufacturing was lead by Japan andEurope.9 Today crystalline silicon !c-Si" still accounts for85% of the market, but China has become the world’s lead-ing manufacturer.10 Table I lists the top 10 crystalline Simanufacturers and their output in 2009.11 In a very shortperiod of time China has moved from being a negligibleplayer to supplying 50% of the worlds PV modules. Thisgrowth was spurred by strategic government investment, ac-

cesses to cheap capital, and has benefited from low Chineselabor costs that enable module production at lower costs thanits competitors in Japan, Europe, and the U.S.

The second major change in the past decade has been theemergence of cadmium telluride !CdTe" thin-film technol-ogy, which was a negligible contributor in 2003. Figure 3displays the global market share held by the various PV tech-nologies in 2008 and 2009.12 A single company, First Solar,has brought CdTe to mass production. In 2009, it became theworld’s first PV manufacturer to exceed 1 GW/yr productionrate and captured 13% of the global market !Fig. 3". Whilesome of this market share has come at the expense of c-Si, ithas also eroded the foothold of competing thin-film tech-nologies such as amorphous silicon !a-Si:H". However, thedistribution of market share continues to be very fluid, withc-Si expected to regain market share in 2010 due to contin-ued expansion of PV manufacturing operations in China. Themost important point is that the emergence of CdTe attests tothe fast change the PV industry can undergo within a rela-tively short period of time. Such shifts are driven by innova-tions in manufacturing and fundamental science coupledwith creative and aggressive entrepreneurship.

Figure 4 displays the champion overall power conversion

FIG. 2. !Color online" Workshop demographics.

TABLE I. Top 10 crystalline silicon manufacturers in 2009 !source: Green-tech Media, May 2010".

Rank Company2009 module production

!MW dc"

1 Suntech Power 7042 Sharp 5953 Yingli Green Energy 5254 Kyocera 4005 Trina Solar 3996 Sunpower 3987 Canadian Solar 3268 Solarfun 3139 SolarWorld 28810 Sanyo 260

FIG. 3. !Color online" Market size and the share held by relevant technolo-gies in 2008 and 2009.

FIG. 4. !Color online" Champion efficiencies reported for cells and commer-cial modules for the established PV manufacturing technologies.

Wolden et al.: Photovoltaic manufacturing: Present status030801-3 030801-3

JVST A - Vacuum, Surfaces, and Films

Total of 7.86 GW


Market Strategy

• CdTe:– New comers would face high barrier set up by First Solar claiming a cost of

$0.76/Wp.– Lack of long term solution for Cadmium (toxic) – The availability of Te becomes problematic beyond 100GW

• Amorphous-Si:– Stable efficiency is not cost competitive– tandem structure offers ~10%, but capex is high– investor confidence is low based on pessimistic results of AMAT SunFab.

• CIGS:– Great potential for high efficiency(>20%) and low cost (<$0.5/Wp). Currently

suffer from lack of mature technology for mass production

16


Challenges in CIGS technology


Cost Reduction

• Cost of CIGS module is higher than that based on crystalline Si technology.

2

$1.75

$1.50

$1.25

$1.00

$0.75

$0.50

COGS ($/W)

2009 2010 2011 2012 2013 2014 2015

mc-Si TF-Si CdTe CIGS (sputtering on glass)

Source: Lux Research Inc - “Module Cost Structure Breakdown” Oct. 2010


CIGS - Material Properties and Device Physics

3


Development of CuInSe2-based solar cell

4

MATERIAL PROPERTIES 571

Even though the behavior of CuInSe2 provides a good basis for the understanding ofdevice-quality material, there are pronounced differences when Ga and Na are presentin the films. More recently, Cu(InGa)Se2 has been reviewed in the context of solar cellswith an emphasis on electronic properties [27].

In this section the structural, optical, and electrical properties of CuInSe2 arereviewed along with information about the surface and grain boundaries and the effect ofthe substrate. In each case, as appropriate, the effect of the alloying with CuGaSe2 to formCu(InGa)Se2 and the impact of Na and O on the material properties will be discussed.

13.2.1 Structure and Composition

CuInSe2 and CuGaSe2 have the chalcopyrite lattice structure. This is a diamond-likestructure similar to the sphalerite structure but with an ordered substitution of the group I(Cu) and group III (In or Ga) elements on the group II (Zn) sites of sphalerite. Thisgives a tetragonal unit cell depicted in Figure 13.3 with a ratio of the tetragonal latticeparameters c/a close to 2 (see Table 13.1). The deviation from c/a = 2 is called thetetragonal distortion and stems from different strengths of the Cu–Se and the In–Se orGa–Se bonds.

The possible phases in the Cu–In–Se system are indicated in the ternary phasediagram in Figure 13.4. Thin films of Cu–In–Se prepared under an excess supply of Se,that is, normal conditions for thin-film growth of Cu(InGa)Se2, have compositions that fallon, or close to, the tie-line between Cu2Se and In2Se3. Chalcopyrite CuInSe2 is located onthis line as well as a number of phases called ordered defect compounds (ODC), becausethey have a lattice structure described by the chalcopyrite structure with an ordered inser-tion of intrinsic defects. A comprehensive study of the Cu–In–Se phase diagram has beencompleted by Godecke et al. [32]. A detail of the Cu2Se–In2Se3 tie-line near CuInSe2 isdescribed by the pseudobinary phase diagram reproduced in Figure 13.5 [32] Here ! isthe chalcopyrite CuInSe2, " is a high-temperature (HT) phase with the sphalerite structure,and # is an ODC phase. It is interesting to note that the single phase field for CuInSe2 atlow temperatures is relatively narrow as compared to earlier beliefs, and does not contain

Cu

In

Se

Figure 13.3 The unit cell of the chalcopyrite lattice structure

Chalcopyrite structureCu2S

CuInSe2

Cu(In,Ga)Se2

?CuGaSe2

CuAlSe2

CuAlS2

(AgCu)(InGa)Se2


In Theory - a simple system

5

p-type Semiconductor

n-type Semiconductor

Electrode

Electrode

pn-junction


p-type

ECB

EVB

EF

n-typeECB

EVB

EF

pn-junction under illumination


p-type

ECB

EVB

EF

n-typeECB

EVB

EF



p-type

ECB

EVB

EF

n-typeECB

EVB

EF



Depletion Region

p-type

ECB

EVB

EF

n-typeECB

EVB

EF

E

eVbi



Depletion Region

p-type

ECB

EVB

EF

n-typeECB

EVB

EF



Depletion Region

p-type

ECB

EVB

EF

n-typeECB

EVB

EF



p-type

ECB

EVB

EF

n-typeECB

EVB

EF



Absorption

7

300 400 500 600 700 800 900 1000 1100 12000.0

0.2

0.4

0.6

0.8

1.0

1.2

1.4

1.6

1.8

Wavelength (nm)

0.0

0.2

0.4

0.6

0.8

1.0

CIGS

CdTe c-Si

Rel

ativ

e S

pect

ral R

espo

nse

Sol

ar s

pect

ral i

rrad

ianc

e (W

m-1 n

m-1)


CuInSe21.04 eV

CuInS21.53 eV

CuGaSe21.68 eV

Bandgap Engineering of Absorber Layer


CIGS - processing technology

9


Anti-reflective coating

Metal grid

Window - ZnO(150-250 nm)

Buffer - CdS (50 nm)

Absorber - CuInGaSe2(1-2 μm)

Back-contact -Mo(1 μm)

Substrate (glass, metal foil)

CIGS Solar Cell Structure


Coevaporation- [vacuum] evaporation from elemental source- demonstrated champion efficiency 20.3% by NREL’s 3-stage process (Wurth Solar, Solibro (Q-cell), Global Solar)

Selenization of metallic precursors- [vacuum] sputtering + selenization Rapid Thermal Processing: (Sunshine PV, Avancis, Johanna Solar) H2Se: (Axuntech, TSMC, Solar Frontier, Honda Soltec)- [solution] electrodeposition + selenization (Solopower)- [solution] nanoparticle deposition + selenization (Nanosolar)

Solution Process- [solution] hydrazine precursors (IBM, UCLA)- in research scale development.

Processing of CIGS Solar Cell


Processing of CIGS Solar Module

12

CIGS Processing TechnologyCIGS Processing Technologyg gyg gySequential deposition/reaction

From B. Dimmler’s presentation, PV Taiwan, 2009


60 cm! 120 cm size is currently common in the commer-cilization, although more robust, highly reproducible, low-cost and longer lifetime reaction furnace for this techniqueis needed. Another important request to the furnace designis to be applicable to larger substrate size.

5.2 Absorber formation process of ShowaShell Sekiyu K.K.

Showa Shell Sekiyu K.K. started the commercialization byestablishing its 100% subsidiary of Showa Shell Solar K.K.in 2006. The developed absorber formation process byShowa Shell Sekiyu K.K. was the SAS method as shown inFigure 9 [33], which was transferred to Showa Shell Solar.In this method, selenium from the H2Se gas is incorporatedinto the metal precursor from the surface at the selenizationstep, and then sulfur from the H2S gas is incorporated intothe selenide at the sulfurization step. The final chalco-genide absorber is a graded band gap Cu(InGa)Se2absorber with a thin Cu(InGa)(SSe)2 surface layer.

In the SAS method, both selenium and sulfur, as a resultof decomposition of the reaction gases of H2Se andH2S respectively, are incorporated from the surface. Thesurface is the reaction front, so that the fill factor (FF) isstrongly influenced by its quality. Over (or excess)sulfurization causes shunt at the pn heterointerface,because low-resistivity semi-metallic compounds likeCu–VI(Se, S) ternary alloys are formed by sulfur etchingof indium.While Mo–Se binary or Mo–Se–S ternary alloysare not considered to disturb the carrier flow at the Mo andabsorber interface, because favorable orientation of Mo–Se

binary or Mo–Se–S ternary alloys fabricated by the SASmethod is likely to be prepared at a lower absorberformation temperature than the case of co-evaporation.This would be a proof of higher reactivity of selenium andsulfur from the reaction gases [39]. These Mo chalcogen-ides are removed to some degree at the patterning 2 step,which is effective to remove the CIS-based absorbermechanically and form a direct contact of Mo and n-typeZnO:B (BZO) window layers. Therefore, FF is a goodindicator of the reaction severity (or intensity), especiallysulfurization intensity.

Possible reaction pathways on the absorber formation inthis method are proposed in Figure 10 based on the results

Figure 9. Temperature profile of SAS method as an absorber

formation process.

Figure 10. Possible reaction pathways of absorber formation in the SAS method [33].

460 Prog. Photovolt: Res. Appl. 2010; 18:453–466 ! 2010 John Wiley & Sons, Ltd.DOI: 10.1002/pip.969

CIGS absorbers and processes S. Niki et al.

60 cm! 120 cm size is currently common in the commer-cilization, although more robust, highly reproducible, low-cost and longer lifetime reaction furnace for this techniqueis needed. Another important request to the furnace designis to be applicable to larger substrate size.

5.2 Absorber formation process of ShowaShell Sekiyu K.K.

Showa Shell Sekiyu K.K. started the commercialization byestablishing its 100% subsidiary of Showa Shell Solar K.K.in 2006. The developed absorber formation process byShowa Shell Sekiyu K.K. was the SAS method as shown inFigure 9 [33], which was transferred to Showa Shell Solar.In this method, selenium from the H2Se gas is incorporatedinto the metal precursor from the surface at the selenizationstep, and then sulfur from the H2S gas is incorporated intothe selenide at the sulfurization step. The final chalco-genide absorber is a graded band gap Cu(InGa)Se2absorber with a thin Cu(InGa)(SSe)2 surface layer.

In the SAS method, both selenium and sulfur, as a resultof decomposition of the reaction gases of H2Se andH2S respectively, are incorporated from the surface. Thesurface is the reaction front, so that the fill factor (FF) isstrongly influenced by its quality. Over (or excess)sulfurization causes shunt at the pn heterointerface,because low-resistivity semi-metallic compounds likeCu–VI(Se, S) ternary alloys are formed by sulfur etchingof indium.While Mo–Se binary or Mo–Se–S ternary alloysare not considered to disturb the carrier flow at the Mo andabsorber interface, because favorable orientation of Mo–Se

binary or Mo–Se–S ternary alloys fabricated by the SASmethod is likely to be prepared at a lower absorberformation temperature than the case of co-evaporation.This would be a proof of higher reactivity of selenium andsulfur from the reaction gases [39]. These Mo chalcogen-ides are removed to some degree at the patterning 2 step,which is effective to remove the CIS-based absorbermechanically and form a direct contact of Mo and n-typeZnO:B (BZO) window layers. Therefore, FF is a goodindicator of the reaction severity (or intensity), especiallysulfurization intensity.

Possible reaction pathways on the absorber formation inthis method are proposed in Figure 10 based on the results

Figure 9. Temperature profile of SAS method as an absorber

formation process.

Figure 10. Possible reaction pathways of absorber formation in the SAS method [33].



As-deposited

Source: Niki et al., Prog. Photovolt. Res. Appl. 18 453 2010.

Complexity of CIGS Processing


CIGS Technology by various companies

14

solar module manufacturers. The first column indicates thegeneral purpose and thickness of the layers. There are twomajor deposition pathways for the ‘‘heart’’ of the solar cell,the CIGS absorber layer, co-evaporation, and selenization.Wurth Solar and Global Solar use the in-line evaporationprocess for the CIGS absorber layer deposition which willbe described more in detail in Section 4.

Figure 1 also includes manufacturers which apply theselenization/sulfurization approach—Avancis, JohannaSolar, Showa Shell, Sulfurcell, and Honda Soltec. Thisapproach involves the deposition of precursors containingCu, In, and Ga as metals, alloys, or chalcogenidecompounds followed by a selenization and/or sulfurizationprocessing step using either a thermal treatment in thechalcogen atmosphere or reactive chalcogen compoundslike hydrogen selenide or hydrogen sulfide. Alternative,low-cost, non-vacuum methods may also be applied forprecursor deposition and concepts for large-scale pro-duction are in progress.

Most manufacturers use rigid glass substrates. GlobalSolar and Solarion use flexible stainless steel andpolyimide (PI) foils, respectively. Solyndra is pursuingan interesting concept by coating glass rods. Although allmanufacturers use a molybdenum back contact, someinclude a barrier for elements diffusing from the substrateand/or electrical insulation, as indicated.

The buffer and window layers are necessary forcompleting the heterojunction. The most common bufferlayer is CdS deposited in a chemical bath coupled with asputtered intrinsic ZnO layer. Cadmium-free alternativeslike suphides and hydroxides of zinc or indium are used bysomemanufacturers. The ZnOwindow is doped with eitheraluminium or boron and is deposited either by sputtering orwith an MOCVD process.

These layers are deposited sequentially to form the solarcell and patterning steps integrate the cells into modules.Figure 2 indicates the basic process flows in CIGS modulemanufacturing as developed by the ZSW/Wurth Solar as an

Figure 1. Layer stacks as applied by a selection of manufacturers of CIGS-based thin-film solar modules.

Figure 2. Process flows for the manufacturing of CIGS solar modules as used by the ZSW and Wurth Solar.





Fabrication of a solar module

15


Interconnect Interconnect

Cell Cell Cell

Cell vs Module

• Series Connection (increase voltage)Vtotal = V1 + V2 + V3 .....Itotal = I1 = I2 = I3

• Parellel Connection (increase current)Vtotal = V1 = V2 = V3 .....Itotal = I1 + I2 + I3 .....

=

SS1114120F / SS1114120N SS1114120F / SS1114120N is a new developed in the large-sized

CIGS Thin Film solar module. It offers an excellent power output

with high product quality and low cost-effective solution.

*SS1114120F framed / SS1114120N frameless

Product Specification Electrical Characteristics

Power Output (Pmpp) 120Wp Max. Power Voltage (Vmpp) 57.2V Max. Power Current (Impp) 2.08A Open Circuit Voltage (Voc) 82.1V Short Circuit Current (Isc) 2.5A

Max. System Voltage 1000V Mechanical Characteristics

Size 1100mm x 1400mm Depth 7.7mm

Weight (Frame/Frameless) 33kg / 28.5kg Type of Frame Al Frame / Frameless Front Cover Float glass Back Cover Tempered glass

Encapsulation Glass / EVA / Glass

Limited Warranty

Product Guarantee Warranty of the Materials and the Workmanship for 2 years.

Power Guarantee 90% of the minimal rated Power Output for 10 years after shipment

80% of the minimal rated Power Output for 20 years after shipment.


Monolithic Integration

17

solar module manufacturers. The first column indicates thegeneral purpose and thickness of the layers. There are twomajor deposition pathways for the ‘‘heart’’ of the solar cell,the CIGS absorber layer, co-evaporation, and selenization.Wurth Solar and Global Solar use the in-line evaporationprocess for the CIGS absorber layer deposition which willbe described more in detail in Section 4.

Figure 1 also includes manufacturers which apply theselenization/sulfurization approach—Avancis, JohannaSolar, Showa Shell, Sulfurcell, and Honda Soltec. Thisapproach involves the deposition of precursors containingCu, In, and Ga as metals, alloys, or chalcogenidecompounds followed by a selenization and/or sulfurizationprocessing step using either a thermal treatment in thechalcogen atmosphere or reactive chalcogen compoundslike hydrogen selenide or hydrogen sulfide. Alternative,low-cost, non-vacuum methods may also be applied forprecursor deposition and concepts for large-scale pro-duction are in progress.

Most manufacturers use rigid glass substrates. GlobalSolar and Solarion use flexible stainless steel andpolyimide (PI) foils, respectively. Solyndra is pursuingan interesting concept by coating glass rods. Although allmanufacturers use a molybdenum back contact, someinclude a barrier for elements diffusing from the substrateand/or electrical insulation, as indicated.

The buffer and window layers are necessary forcompleting the heterojunction. The most common bufferlayer is CdS deposited in a chemical bath coupled with asputtered intrinsic ZnO layer. Cadmium-free alternativeslike suphides and hydroxides of zinc or indium are used bysomemanufacturers. The ZnOwindow is doped with eitheraluminium or boron and is deposited either by sputtering orwith an MOCVD process.

These layers are deposited sequentially to form the solarcell and patterning steps integrate the cells into modules.Figure 2 indicates the basic process flows in CIGS modulemanufacturing as developed by the ZSW/Wurth Solar as an

Figure 1. Layer stacks as applied by a selection of manufacturers of CIGS-based thin-film solar modules.

Figure 2. Process flows for the manufacturing of CIGS solar modules as used by the ZSW and Wurth Solar.




• The advantage of thin-film technology over crystalline Si is the ability to create electrical connections in situ, which offers a shorter supply chain


Matching Champion Efficiencies in Cell and Module.

• CIGS presents the largest gap in the efficiency of cell and module.

• Combined effort from industrial R&D and academic research on the absorber material is required.

• Multidisciplinary efforts are needed from Materials, Chemistry, and Physics.

18

issue of material availability for existing technologies.Lastly, recommendations are made for research investment,with an emphasis on those areas that are expected to havecross-cutting impact.

II. PRESENT STATUS OF PV MANUFACTURING

This article is not intended to provide a comprehensivereview of the history of photovoltaics but rather to update thesubstantial changes that have occurred since the turn of thecentury. This date was chosen for a couple of reasons. First,there are a number of very good reviews detailing the growthand evolution of the PV industry up to around 2003.6,8,9

Moreover, there have been two significant changes that haveaccompanied the growth experienced during the past 7 years.In 2003 various forms of silicon accounted for !99% of thePV market, and manufacturing was lead by Japan andEurope.9 Today crystalline silicon !c-Si" still accounts for85% of the market, but China has become the world’s lead-ing manufacturer.10 Table I lists the top 10 crystalline Simanufacturers and their output in 2009.11 In a very shortperiod of time China has moved from being a negligibleplayer to supplying 50% of the worlds PV modules. Thisgrowth was spurred by strategic government investment, ac-

cesses to cheap capital, and has benefited from low Chineselabor costs that enable module production at lower costs thanits competitors in Japan, Europe, and the U.S.

The second major change in the past decade has been theemergence of cadmium telluride !CdTe" thin-film technol-ogy, which was a negligible contributor in 2003. Figure 3displays the global market share held by the various PV tech-nologies in 2008 and 2009.12 A single company, First Solar,has brought CdTe to mass production. In 2009, it became theworld’s first PV manufacturer to exceed 1 GW/yr productionrate and captured 13% of the global market !Fig. 3". Whilesome of this market share has come at the expense of c-Si, ithas also eroded the foothold of competing thin-film tech-nologies such as amorphous silicon !a-Si:H". However, thedistribution of market share continues to be very fluid, withc-Si expected to regain market share in 2010 due to contin-ued expansion of PV manufacturing operations in China. Themost important point is that the emergence of CdTe attests tothe fast change the PV industry can undergo within a rela-tively short period of time. Such shifts are driven by innova-tions in manufacturing and fundamental science coupledwith creative and aggressive entrepreneurship.

Figure 4 displays the champion overall power conversion

FIG. 2. !Color online" Workshop demographics.

TABLE I. Top 10 crystalline silicon manufacturers in 2009 !source: Green-tech Media, May 2010".

Rank Company2009 module production

!MW dc"

1 Suntech Power 7042 Sharp 5953 Yingli Green Energy 5254 Kyocera 4005 Trina Solar 3996 Sunpower 3987 Canadian Solar 3268 Solarfun 3139 SolarWorld 28810 Sanyo 260

FIG. 3. !Color online" Market size and the share held by relevant technolo-gies in 2008 and 2009.

FIG. 4. !Color online" Champion efficiencies reported for cells and commer-cial modules for the established PV manufacturing technologies.

Wolden et al.: Photovoltaic manufacturing: Present status030801-3 030801-3

JVST A - Vacuum, Surfaces, and Films


Future of CIGS

19

TF Has Opportunities in AsiaTF Has Opportunities in AsiappppMorgan Stanly Research

1. The race for dominance has not even started in Asia2 Japan Korea & Taiwan have relevant experience from2. Japan, Korea & Taiwan have relevant experience from

semiconductor, TFT LCD & Optical Media manufacturing3. China has the end markets and ability to reduce costy Source: Morgan Stanley Research

• Taiwan has the BEST overall manufacturing advantage!


Sunshine’s 3KW CIGS Demo System

20


Material Abundance for terawatts capacity

1


Best Research Cell Efficiencies

2


Inorganic semiconductor from earth abundant elements - CZTS

3


Copper-Zinc-Tin-Sulfur (Kesterite)

• Current production of Indium is 600 ton/yr, which can support ~70GW/yr of CIGS production.

• Copper, Zinc, Tin, Sulfur are able to provide on a TW level, assuming similar efficiency as CIGS has been achieved.

4

1.0E%02'

1.0E%01'

1.0E+00'

1.0E+01'

1.0E+02'

1.0E+03'

1.0E+04'

Copper' Indium' Gallium' Zinc' Tin' Sulfur' Selenium'

ppm#

Cu2ZnSnS4


Cu2ZnSn(S,Se)4 Solar Cell

• Highest cell efficiency demonstrated is 9.7%, compare to 20.1% for CIGS.

• Industrial effort: IBM Corp. DelSolar Co, Solar Frontier

• Potentially fast maturation due to similarities with CIGS processing

• First time higher cell efficiency was demonstrated using solution-based processing than vacuum deposition.

5

the compounds in the above series. Two counter-examples areCu2FeSn(S,Se)4, which has been determined to be stannite [15], andCu2ZnSnS4 (CZTS), which was originally described as stannite, butis now recognized as kesterite [15,20,21]. Cu2ZnSnSe4 (CZTSe) hasalso been described as stannite in the literature [16,22–24].However, the data presented for this compound are largely powderX-ray diffraction (XRD) data, which cannot reliably be used todistinguish between the two possibilities. Theoretical calculationspredict that the lowest energy configuration for CZTSe (as for CZTS)is kesterite [22,25], although the difference in energy betweenkesterite and stannite is much smaller for the selenide vs. thesulfide, suggesting a stronger tendency of the selenide to incorpo-rate stacking faults involving alternative layering sequences. Wewill describe all compounds in the Cu2ZnSn(S,Se)4 (CZTSSe) familyas kesterite, with the recognition that experimental verification ofthis assignment for the high Se-content samples must still beprovided. Lattice constants and band gaps of several more thor-oughly studied kesterites and stannites are given in Table 1.

2.2. Kesterite-related photovoltaic films

Table 1 demonstrates that many of the kesterite-related phasesexhibit a direct band gap in the optimal range for photovoltaicenergy conversion (all are reported to be p-type semiconductors).Despite this fact, relatively few of these compounds have beenreported in actual solar cell devices. In 1988, Ito et al. [29] examinedthe electrical and optical properties of CZTS, a material which hasall earth-abundant and relatively non-toxic elements, as well asCu2CdSnS4 (CCTS). Both compounds were reported to be directband gap p-type semiconductors with band gaps of 1.45 and1.06 eV for the zinc and cadmium compounds, respectively. Thefirst CZTS solar cell was constructed from a heterostructure withcadmium tin oxide, yielding an open circuit voltage of 165 mVunder AM1.5 illumination (no efficiency reported). Earlier work on

the CCTS system [35], employing a heterodiode of a CCTS singlecrystal with a CdS film, yielded a power conversion efficiency of1.6% (reported to be limited by a large series resistance), an opencircuit voltage of 500 mV and a short circuit current density of7.9 mA/cm2.

By far the most extensive subsequent work on kesterite-relatedPV devices has been performed on CZTSSe materials, which are theprimary focus of this review, leading to continual device perfor-mance improvement (Fig. 2) [37]. In 1996, Katagiri et al. [28,37]reported on sequentially evaporated CTZS (pure sulfide) films,yielding a glass/Mo/CZTS/CdS/ZnO:Al device efficiencyof 0.66%andopen circuit voltage of 400 mV. Within a year, the StuttgartUniversity group had pushed the power conversion efficiency ofa similarly prepared device to 2.3% [38]. Currently, the recordpower conversion efficiency for vacuum-deposited CZTS hasimproved to 6.8% [37]. For analogous selenide CZTSe devices, in1997 Friedlmeier et al. [38] reported on vacuum-fabricated films,obtaining a device efficiency of 0.6%. By 2009 the efficiency forCZTSe devices had increased to 3.2% [31]. The current CZTSSekesterite record of 9.7% (Fig. 2) was achieved using a hydrazine-based solution deposition approach with mixed sulfur/seleniumanions [26]. More detailed discussion of the various depositionmethods for CZTSSe films, as well as approaches for improvingefficiency, will be given in the next section, while analogousdiscussion on in-depth device characterization will be given inSection 4.

3. Approaches for CZTSSe film deposition

The growing perception of CZTSSe as a potentially ideal photo-voltaic absorber material is reflected in the abundance and varietyof routes that have been undertaken for thin-filmdeposition. Theseinclude a wide spectrum of vacuum and atmospheric-pressureapproaches. For each of these deposition approaches one keybarrier toward a reliable and low-cost process is the complexand incompletely understood nature of the multinary Cu–Zn–Sn–S–Se phase diagram and control over the phase progression duringfilm formation [39–41], which presents a challenge for preparingsingle phase films. A second common theme generally encountered

Table 1Lattice constants and band gaps for selected families of kesterite/stannite materialsfor which both structural and band gap information have been reported.

Compound a (A) c (A) V (A3) Eg (eV) References

Cu2ZnSnS4 5.427 10.848 319.5 [19]5.432 10.840 319.9 [26]5.435 10.822 319.7 [27]5.426 10.81 318.3 1.45 [28]

1.45 [29]1.51 [30]

Cu2ZnSn(Se2.4S1.6) 5.567 11.168 346.1 1.21 [26]Cu2ZnSn(Se3.9S0.1) 5.668 11.349 364.6 1.03 [26]Cu2ZnSnSe4 5.681 11.34 366.0 [14]

5.684 11.353 366.8 0.94 [31]5.688 11.338 366.8 [23]

1.017 [32]Cu2ZnGeS4 5.341a 10.509a 299.8 [33]

5.344 10.513 300.2 [27]Cu2ZnGeSe4 5.622 11.06 349.6 [14]

5.606 11.042 347.0 1.63 [16]5.618 11.04 348.4 1.52 [34]

Cu2CdSnS4 5.586 10.834 338.1 [19]5.593 10.840 339.1 [27]

1.06 [29]1.16 [35]

Cu2CdSnSe4 5.832 11.389 387.4 0.96 [16]0.89 [36]

5.829 11.418 388.0 [23]5.832 11.392 387.5 [27]5.814 11.47 387.7 [14]

Cu2CdGeSe4 5.657 10.988 351.6 [19]1.29 [36]

a Also reported [19] as orthorhombic with lattice constants a!7.504 A,b!6.474 A, c!6.185 A.

Fig. 2. Record CZTSSe thin-film device performance vs. year, showing consistentprogress in power conversion efficiency. Circles represent data for pure-sulfur-containing materials (CZTS), whereas the triangle represents data for kesterites inwhich Se has been introduced. The laboratories responsible for the record are noted.References for the data points are given in the text.

D.B. Mitzi et al. / Solar Energy Materials & Solar Cells 95 (2011) 1421–1436 1423

Source: Mitzi et al. Sol Energ Mat Solar C 95 1421 (2011)


Dye-Sensitized Solar Cell

6


Best Research Cell Efficiencies

7


Dye-Sensitized Solar Cells

8

• Unlike a conventional solar cell that operates based on a pn-junction, dye-sensitized solar cell is based on a photo-electrochemical effect

• TiO2: electron acceptor and matrix for the dye.

• Dye: Light absorption.

• Electrolyte: electron donor and prevent recombination

• Industrial efforts: G24i, Solarprint, Dyesol, Sony

Image was adopted from M. Gratzel, Inorg. Chem. 44 6841 (2005).


Dye-sensitized solar cell

• Unlike conventional inorganic-based solar cell, DSSC showed higher performance under diffuse and low light conditions

• This seems to be a characteristic for organic-based solar cell.

9Source: M. Gratzel, Inorg. Chem. 44 6841 (2005).


Ru-based dye material provides a variety of colors

10Source: M. Gratzel, Inorg. Chem. 44 6841 (2005).

新能光電Sunshine PV 11Source: www.sony.net

http://www.sony.net

http://www.sony.net


Challenges: - Perfomance

•Progress:– Since the demonstration of 10% efficiency in 1993, progress has improved

1% (absolute) over the past 18 years.

•Dye (Ru):– limited absorption range (< 750 nm) – Potentially limited material supply to reach TW range due to Ruthenium.

• Iodine solution:– I3-/I- limits Voc at 550mV– Replacing with solid state electrolyte only resulted in lower efficiency.

12


Challenges - Stability in Commercialized product

•Demonstration:– Dyesol showed over 20 years under accelerated conditions.– Outdoor testing by Toyota showed a 6% (relative) power loss for 2.5 years.

•Ecapsulation process:– Encapsulation process has a strong influence on stability of module.– Champion cell (11%) demonstrated so far has been limited to using a liquid

electrolyte

•Requirement a hermetic seal:– The material much be mechanically and thermally stable– prevent mass-transport between cells.

13


Feed in Tariffs

14


A Few Thoughts about Feed In Tariff

• A tool to jump-start the solar industry • With the help of FIT, European solar market grew 80%, whereas the US market,

without FIT, grew 30%.

• A fund design problem: policy makers sets price that are inadequate for the market demand. If the subsidy is too deficient, nothing happens. If it's too generous, it creates a surge in solar production followed by overstocking.

• Case of Germany: generous incentive resulted in growth of solar industry by 200% in 2005 and 2006.

• Case of Spain: planned for a total of 400 MW from 2007 through 2010. By Sept. 2007, installation already reached 344 MW.

• FIT is a temporary solution.• Ultimately market should set the price for a technology. A technology is

sustainable only if it’s cost structure is able to match with the market price.

15


“Energy is the biggest business in the world, the mother of all markets. The winners of the race to reinvent energy will not only save the planet, but also make megafortunes.”

16

John DoerrPartner at KPCB


Thank you

Documents

Progress in Thin-Film Solar Cells from Inorganic to ...mx.nthu.edu.tw/~hytuan/courses-files/2012energy&environment/12.14.pdf · 新能光電 Sunshine PV Progress in Thin-Film Solar