Rubin Sidhu, David E. CarlsonBP Solar

April 13, 2010

Crystalline Silicon Solar Cell Technology

Agenda

• Global PV Market• Nomenclature• PV Technologies• Crystalline Silicon Solar Cell Value Chain• Solar Cell Physics (no equations!)

• Why is solar cell efficiency only ~15%?• Reaching Grid Parity using BP Solar’s Crystalline

Silicon PV Technology• The Solar America Initiative• Silicon Solar Technology Roadmap

• Conclusion: Future of PV

PV Experience Curve

Cost of Materials Limit (20% Modules)Cost of Materials Limit (20% Modules)Cost of Materials Limit (20% Modules)Cost of Materials Limit (20% Modules)Cost of Materials Limit (20% Modules)Cost of Materials Limit (20% Modules)

PV module prices have followed an experience curve with a slope PV module prices have followed an experience curve with a slope of of ~ 80% ~ 80%

20% decrease in price with every doubling of cumulative producti20% decrease in price with every doubling of cumulative productionon

Cost of Solar Electricity

Shipments of Photovoltaic Modules

Projection (Lux Research)

PV shipments actually increased about 20% 2009.

PV Shipments & Countries of Origin

(Navigant Consulting)

Europe has been the largest consumer of PV in recent years (> 50% of all installations). China and Taiwan have become the largest producers.

PV Shipments in 2009

• World solar PV market installations reached a record high of 7.3 GW in 2009

growth of 20% over 2008

• $43 billion in global revenues in 2009, up 8% on the prior year

• World solar cell production 9.34 GW in 2009, up from 6.85 GW in 2008

• Crystalline Si module price average for 2009 crashed 38% from previous year

European countries accounted for 5.60 GW, or 77% of world demand

Germany, Italy and Czech Republic accounted for 4.07 GW

United States grew 36% to 485 MW

Grid Parity

GTM Research

LCOE and kWh/kWp

Levelized Cost Of Electricity (¢/kWh):LCOE is the minimum price at which energy must be sold for

an energy project to break even

Including: initial investment operations and maint. cost of fuel cost of capital

Typical Sunny Day

0

200

400

600

800

1000

1200

0 6 12 18 24

(W/m

²)

7.5 Sun-Hour Day24-Hour Equivalent

7.5 kWh Spread over 24 Hours

6 AM Noon 6 PM

Insolation

kWh/kWp

PV Facts

Using today’s PV Technology, an array field that is 300 miles on each side could produce the entire electrical energy used by the United States in a year.

PV Facts

Over its lifetime, a typical PV module, in a sunny climate, will produce over twenty times the electricity initially used to manufacture it.

Technology Overview: Nomenclature

A solar PV cell is the smallest semiconductor device that can convert sunlight into electrical energy

A module is an assembly of cells in series or parallel to augment voltage and/or current

A panel is an assembly of modules on a structure

An Array is an assembly of panels at a site.

Multi (Poly) Cell Mono Cell

Large Roof-Top PV Array

Conversion Efficiencies vs. Time (NREL)

There has been steady progress in the improvement of conversion efficiencies for a number of PV technologies over the last few decades.

The Major Players

Sharp SolarPower Kyocera BP Solar Q-Cells Mitsubishi SolarWorld Panasonic (Sanyo) Schott Solar Isofoton Motech Suntech Evergreen Solar JA Solar

United Solar Kaneka Fuji Electric Sharp Mitsubisihi Schott Solar SunTech EPV PowerFilm AMAT licensees Orelikon licenses

Nanosolar Avancis Solar Frontier Wurth Solar Global Solar Honda Soltec

First Solar Antec Solar Abound Solar PrimeStar Solar Calyxo

Crystalline Si a-Si/µc-Si CIGS CdTe

There are currently more than 300 companies developing or producing solar cells.

With prices continuing to decrease, and more companies entering the market, many small companies and start-ups are likely to fail.

Source: pvsociety

Manufacturing Process for mc-Si PV Modules

% of % of % of % ofTotal Total Total Total0.00% 0.00% 0.00% 0.00%0.27% 2.03% 0.34% 0.85%2.18% 13.23% 2.23% 2.53%

SIZING WAFERING

FINISHTEST

INTER-CONNECTLAMINATION

ANTI-REFLECTIVECOATING

CLEANINGCASTING DIFFUSION

METALLIZATION

% of % of % of % ofTotal Total Total Total0.00% 0.00% 0.00% 0.00%0.27% 2.03% 0.34% 0.85%2.18% 13.23% 2.23% 2.53%

SIZING WAFERING

FINISHTEST

INTER-CONNECTLAMINATION

ANTI-REFLECTIVECOATING

CLEANINGCASTING DIFFUSION

METALLIZATION

Most companies are manufacturing PV modules based on screen-printing contacts on multicrystalline silicon wafers

CZ - Silicon for Solar

Silicon Feedstock

Mono solar cell

Mono-crystalline wafer Drawbacks of CZ Solar

– Cylindrical ingot (module packing)

– Smaller throughput rate• CZ: 5.3 kg/h at 2mm/min• Cast: 12.9 kg/h at 0.3mm/min

– High oxygen content -> LID– Higher energy use– Feedstock limitations– Importance of skilled operators

Mono module

Multicrystalline Silicon Casting

Silicon Feedstock

Silicon Ingot

Silicon Bricks

Multi-crystalline

Wafer

Multicrystalline Solar Cell

Multicrystalline Module

Casting Station cross-section

Heater

Crucible Liquid Si

Solid Si

Insulation

Drawbacks:• Crystal defects

• Iron from crucible

• Inclusions from coating, furnace

• Not compatible with pyramid texturing

The Typical Silicon Solar Cell

This device structure is used by most manufacturers today.• The front contact is usually formed by POCl3 diffusion

• The rear contact is formed by firing screen-printed Al to form a back-surface field

The cell efficiencies for screen-printed multicrystalline silicon cells are typically in the range of 14 – 17%.

Operation of a Solar Cell

The theoretical limit for a crystalline silicon solar cell is ~ The theoretical limit for a crystalline silicon solar cell is ~ 29%.29%.

Backsheet

EVA encapsulant

Solar cells

EVA encapsulantGlass

J-boxCable

Frame w/adhesives

Typical module construction

BP Solar now uses a polyester backsheet to protect the rear of the module

Also a new encapsulant 5% lower temperature 2 – 3% gain in energy production

PV Module Costs

Silicon 23%

Other Materials 27%Labor 25%

Equipment 15%

Electricity 10%

Si Wafer PV Module Production Cost $2.10/W

Major Energy Consuming Steps

Feedstock 3%

Ingot 3%

Wafering 2%

Cell 2%

Silicon 23%

Other Materials 27%Labor 25%

Equipment 15%

Electricity 10%

Si Wafer PV Module Production Cost $2.10/W

Major Energy Consuming Steps

Feedstock 3%

Ingot 3%

Wafering 2%

Cell 2%

About half of the total module cost is associated with the cost of materials

Silicon prices have been dropping rapidly over the last 18 months

Source: George Nemet, U Wisc.

Cost of Materials Limit (20% Modules)Cost of Materials Limit (20% Modules)Cost of Materials Limit (20% Modules)Cost of Materials Limit (20% Modules)Cost of Materials Limit (20% Modules)Cost of Materials Limit (20% Modules)

Cost Reduction

Solar America Initiative

This program is a BP Solar cost shared Technology Pathways Partnership under the DOE Solar Energy Program.

The program addresses all aspects of crystalline silicon PV manufacturing and systems deployment.

Overall Program Objective

Accelerated development of crystalline silicon technology using thin Mono2 TM wafers as the platform.

Module designed for use in residential and commercial markets with products designed specifically for these applications.

System components designed to add value to electricity produced.

PV Product Value Chain

Metallurgical Silicon

Semi-conductor Silicon

Crystal Growth

Wafering Cell Fabrication

Module Assembly

Mono2 TM Silicon Casting

Silicon Feedstock

Silicon Ingot

Silicon BricksCasting Station cross-section

Heater

Crucible Liquid Si

Solid Si

Insulation

Mono2

WaferMono2

Solar CellMono2

Module

Mono2 TM Lifetime Uniformity

61934.7116.4175 s40%

Mono2 TM, pyramid etched

61833.5515.8775 s40%

Mono2 TM

61033.1515.4675 s80%

Multi

VocmV

JscmA/cm2

%

Lifetime Median

Std. Dev.

Material

Simulation results

Multi

Mono2 TM

PV Product Value Chain

Metallurgical Silicon

Semi-conductor Silicon

Crystal Growth

Wafering Cell Fabrication

Module Assembly

c-Si Solar Cells: Loss Mechanisms

0.7%

2.4%

2.2%

2.5%

6.3%

ReflectionShadowingOther (Optical)OhmicRecombination

Thermodynamic limit forcrystalline Si solar cells ~29%

Sources of loss relative tomaximum achievable efficiency– Front surface reflection– Emitter contact shadowing– Poor rear reflectance for light

trapping– Surface recombination (i.e.,

surface cleanliness)– Bulk recombination (i.e.,

material quality)– Resistive losses at the cell

and module level

Cell Process

Texture Etch– ICT

o Preferred option for Mono2 and multi o Cell efficiency improvements demonstratedo Improvements lost when cells are encapsulated

– Alkaline texture etcho Good for Mono2 but not multio Increased cell and module efficiencies by ~ 0.5% absolute.

Cell Process

Back Print Optimization– Maximize area covered with Al paste for improved

performance– Minimize area of Ag pads to reduce cost– Implemented new pattern in production with 0.6% gain in cell

efficiency and 1% increase in module efficiency.

Advanced Metallization– Looking at new approaches to increase metallization thickness,

to reduce series resistance while also reducing finger width andtherefore shadowing.

– Looking at aerosol jet printing, ink jet printing and offset printing.

– Successfully printed 80 to 90 µm wide and ~ 35 µm thick

c-Si Solar Cells: Advances

Source: pvsociety

The Selective Emitter

A number of organizations are developing solar cells with selective emitters in order to use thinner emitters and improve the short-wavelength spectral response.

Passivated Rear Point Contacts

Fraunhofer ISE has used laser firing of aluminum to fabricate high efficiency (21.7%) solar cells with passivated rear point contacts.

The PERL Solar Cell

p p-silicon

n

p

p

oxide

+

+p ++

n +

rear contact oxide

PERL Cell Structure

p p-silicon

n

p

p

oxide

+

+p ++

n +

rear contact oxide

PERL Cell Structure

The PERL solar cell has a passivated emitter with a rear locally diffused base contact, and efficiencies as high as 25% have beenobtained with this structure.

SunPower Back Contact Solar Cell

The SunPower cell has all its electrical contacts on the rear surface of the cell.

Production cells ~ 22.4% efficiency; new prototypes at 23.4%.

Diffusion lengths > 3 x cell thickness (using 145 m thick CZ-Si at end of 2008).

Sanyo HIT Solar Cell

The HIT cell utilizes amorphous Si intrinsic layers (~ 5 nm) as passivation layers. The cell is symmetric except for the a-Si p+ emitter layer (~ 10 nm) on the front and the a-Si n+ contact layer (~ 15 nm) on the rear.

Best lab efficiency = 22.3% (open-circuit voltages as high as 739 mV).

PV Product Value Chain

Module Assembly defines product reliability and longevity

Material choice is key to meet life time goal Good choice of materials allows life time to exceed

warranty period Balance of performance and cost

Metallurgical Silicon

Semi-conductor Silicon

Crystal Growth

Wafering Cell Fabrication

Module Assembly

Module Packaging Materials

Encapsulant: Glue for the whole package: Key component

Back sheet: Protect cells against the environment and provide electrical isolation

Superstrate glass: Cell protection and light capture Frames and framing adhesives: Mounting and load

bearing J-box/Cables: Electric isolation and longevity Tabbing ribbons: For cell interconnection

Module Development

AR Coated Glass– Qualified and implemented Gen 1 AGC Flat Glass NA AR

coated glass with 1.5 to 2% increase in STC power and 4% increase in overall energy collection

– Gen 2 now under development with improved cosmetic uniformity and ~ 2.5% STC power gain.

EVA cure level determination– Standard approach is to use chemical extraction of un-cross

linked EVA (gel content)– Developed differential scanning calorimetry (DSC) as a fast

and reliable way of measuring cross-link density.

ARC vs. standard module

Normalized energy generation of a day– 6% to 7% more energy at early and late periods of the day

-1%

1%

3%

5%

7%

9%

6:00 7:00 8:00 9:00 10:00 11:00 12:00 13:00 14:00 15:00 16:00 17:00

Time of Day

kWh/

kWp

diffe

renc

e be

twee

n A

R c

oate

d &

Unc

oate

d gl

ass

Perc

enta

ge k

Wh/

kWp

diff

eren

ce

Time of Day-1%

1%

3%

5%

7%

9%

6:00 7:00 8:00 9:00 10:00 11:00 12:00 13:00 14:00 15:00 16:00 17:00

Time of Day

kWh/

kWp

diffe

renc

e be

twee

n A

R c

oate

d &

Unc

oate

d gl

ass

Perc

enta

ge k

Wh/

kWp

diff

eren

ce

Time of Day

Standard modules

AR coated modules

BP Solar testing beyond industry standards

BP Solar ensures the long-term dependability and reliability of our products by testing beyond industry standards to mimic 25 years of

exposure and real world conditions.

BP Testing Implications

Thermal cycling at 2.5 times IEC Standards

Damp heat exposure at 25% more than IEC Standards

Addition of wind simulation to IEC humidity freeze cycles

Load testing as much as twice the IEC Standard

Thermal cycles stress and crack module interconnects; failures of interconnects lead to significant loss of power

Moisture and heat can lead to corrosion leading to significant power loss or module failure

Modules experience significant flexing due to wind pressure thatcan cause cell cracks and lead to reduced module output or failure

Modules experience significant flexing due to snow loads that can cause cell cracks and lead to reduced module output or failure

Residential Building-Integrated PV

Building-integrated PV may become ubiquitous in the next few decades.

Large Grid-Connected PV Arrays

The levelized cost of electricity should fall to ~ 6 ¢/kWh by 2015 for large grid-connected arrays

Solar Energy – the Long-Term Solution?

Source: German Advisory Council on Global Change

Some forecasts predict that solar will provide most of our energy needs in the latter half of this century.

Projections for the Future of PV

The levelized cost of PV electricity could fall to ~ 6 ¢/kWh by 2015

Disruptive technologies with theoretical limits of > 60% may emerge in the next few decades

Assuming a CAGR of 35% (average over the last few decades), the cumulative PV production would be ~ 3.5 TWp by 2026.

3 TWp of solar electricity will reduce carbon emissions by about 1 Gton per year (7 Gtons of carbon were emitted as CO2in 2000)

Thus, by about 2030 PV could be producing about 10% of the world’s electricity and start to play a major role in reducing CO2 emissions

Questions?