Future PV Roundtable at Solar Power International 2017

Pushing the boundaries: the latest steps to decrease LCOE

Initiative partner Gold sponsors Bronze sponsor

September 12, 2017 – Future PV Roundtable at SPI 12:00 Welcome speeches

Session 1: Cell level and module level

Agenda 12:10 Introductory talk from NREL.

12:25 Pitch 1: Heraeus

Part I 12:39 Pitch 2: SunPower

12:48 Pitch 3: GCL

12:57 Pitch 4: Hanwha Q CELLS

September 12, 2017 – Future PV Roundtable at SPI Session 2: Applications

13:18 Pitch 1: Recurrent Energy

Agenda 13:27 Pitch 2: Draker

13:36 Pitch 3: Nel Hydrogen Part II 13:45 Pitch 4: First Solar

14:00 END

September 12, 2017 – Future PV Roundtable at SPI Session 1 Cell and module level

September 12, 2017 – Future PV Roundtable at SPI I Introduction

September 12, 2017 – Future PV Roundtable at SPI Future trends and challenges to PV

Dr. Pauls Stradins Principal Scientist

National Renewable Energy Laboratory Golden, Colorado - USA

12 September, 2017 Silicon is the dominant PV technology today

Si PV

Thin –film PV

Record lab cell efficiencies: Silicon PV: 26.7% for mono- Si; 21.9% (n, ISE) 21.3% (p, Trina) for multi-crystalline Si Thin films: CIGS 21.7%; CdTe 21.0%. Source: Fraunhofer ISE Photovoltaics Report 2017 From Al BSF to p-PERC to passivated contact and IBC cells

Current mainstream: Al-BSF η ~ 19% After p-PERC  Next generation Si PV: > 25% Limited to ~ 20%; likely phase out in 5 years Passivated contacts, IBC, n-type wafers: Record 26.7%. Production: SunPower, Panasonic.

HIT® cell 24.7%

De Wolf et al., green 2012

n+ n+ + + p p+ i n+ i p

Diffused IBC cell Trina 24.1% Next mainstream: p-PERC η > 20%; R&D 22.6%; est. limit ~ 24%

Schmidt et al., JPV 2013; Altermatt PVSC 2017 8 p-PERC: An Upcoming Dominant PV Cell Technology

Ye et al., PVSC 2016 Future (in ~ 3 yr) Mainstream: P-PERC cell ~ 22%; potential for 24%

n+ n+

Schmidt et al., JPV 2013; Altermatt PVSC 2017 Adapt conventional Al BSF lines 2 new steps: back passivation and laser grooving; 40% upgrade cost. Stabilization against light-induced degradation (“regeneration”) Allows for bifacial modules. • IHS predicts that the total global capacity of PERC cells in 2016 (11.9 GW) will double 2015 capacity (5.6 GW). Hanwha Q cells, JA Solar, Trina leading the transition to p-PERC. • Al-BSF technology, by contrast, seems to be close to reaching its technology limit for efficiency improvement. IHS predicts PERC will become in the next mainstream technology and that there will be very little new investment for Al-BSF lines from 2019.

IHS PV Cell Technology Report 2016 Long-term stability of the regenerated (stable) state C that eliminates LID and enables >21% efficient p-PERC cells is still questionable.

State C destabilizes in dark at elevated T:

“Extrapolating the Arrhenius correlation to lower temperatures gives Wilking et al., SolMat 2014 destabilization time constants of … ≥ 53 (+110/-7) years at 60oC.”

“However, long-term understanding of the dynamic equilibrium of the H-passivated state requires an understanding of the evolution of H in the bulk. Over the course of time, H may be gettered at sites in the bulk crystal or otherwise become unavailable to the regeneration process”. H. Hieslmair, IEEE PVSC 2016 Summary of future trends and challenges

➢ P-PERC is rapidly advancing to become dominant PV technology

➢ Advantages of p-PERC: o High efficiency - up to 24% cells projected o Allows for bifacial modules o Inexpensive.

➢ Challenges to industry: o Can p-PERC fully realize the efficiency potential? o Are p-PERC panels stable > 25 years against LID? o Will n-Cz based, high performance Si PV become mainstream after p-PERC as next generation PV? II Thesis

“Cell technology improvements can reduce LCOE by 20 % in 5 years"

September 12, 2017 – Future PV Roundtable at SPI METALLIZATION TECHNOLOGY EVOLUTION & REVOLUTION - REDUCE LCOE FURTHER BY 20 % IN 5 YEARS

Dr. Weiming Zhang, EVP & CTO Heraeus Photovoltaics A LONG LEGACY IN THE PHOTOVOLTAICS INDUSTRY

› More than 45 years of experience in thick film pastes › Over 30 years history in solar cell paste technology (from integration of Ferro Corporation’s solar paste business) Heraeus Share 40% 35% 30% 25% Formation of 20% Heraeus Photovoltaics 15% 10% 5% Expansion of business Purchase & 0% model towards solutions integration of the 2013 2014 2015 2016 2017 2018 beyond paste Ferro Corporation‘s solar paste business Opening of Opening of Opening of Opening of regional Opening of the technical service lab full service site production line Product APAC Technology in Taiwan in Singapore in Taiwan Modification Labs in and Development China and Taiwan Center in China + Product › 1st Chinese modification lab customer Japan › 1st Taiwanese customer

2009 2010 2011 2012 2013 2015 2016 2017

14 HERAEUS PHOTOVOLTAICS - LEADING THE FUTURE OF PV HERAEUS PHOTOVOLTAICS IS CLOSE TO YOU – ALL AROUND THE WORLD

Hanau, Germany Seoul, Korea Tokyo, Japan › Faster customization › One week lead time TS, Sales, Conshohocken R&D, TS, Sales Shanghai, China TS, Sales › Availability around all time R&D zones USA › Largest and deepest product portfolio Production,TS Taipei, Taiwan R&D, Sales, R&D, › Leading IP position Production, APAC TDC TS, Sales

› Over 45 years experience Singapore in silver paste Production, TS, Sales, R&D

Production, TS, Sales

TS Technical Service APAC TDC Asia Pacific and Development Center

15 HERAEUS PHOTOVOLTAICS - LEADING THE FUTURE OF PV TOTAL SOLUTION COMMITMENT TO PV INDUSTRY

Infrared Lamps for various drying/curing processes High Mesh Screen 红外灯管高性能网版可适应多种烘干、烧结工艺 Diffusion Consulting 扩散工艺咨询服务 Cell Module Si/Crucible Metallization Wafer Processing Assembly 铸锭金属化硅片电池工艺组件封装 Silver Pastes for PERC, Black-Silicon, © HeraGlaze Double-Print, N-type and ECA Impurity diffusion barrier Knotless™ Screen for Shingled modules coating for crucibles SOL系列导电银浆导电胶 HeraGlaze©坩埚涂层为PERC、黑硅、二次印为叠瓦组件定制用于坩埚隔绝杂质的涂层刷、N型以及无网结等工艺 Efficiency Gain, Lower LCOE 提高效率！降低度电成本

16 HERAEUS PHOTOVOLTAICS – LEADING THE FUTURE OF PV HERAEUS METALLIZATION PASTE CONTRIBUTION “Fine “Mono” “Double line” Print” ) 120

2 “Super

“Low Adhesion” 0.1% □ 110 ∕cm laydown” 0.1% 100 0.1% SOL9651 “Great 90 wall” 0.1% SOL9641B Q3 2017 0.1% SOL9631C 80 Q3 2016 0.2% SOL9621H/M Q3 2015 70 SOL9610M/Y 0.15% Q3 2014 60 0.3% SOL9411 Q3 2013 +1.2% SOL9273 50 0.15% Q4 2011 = Sheet Resistance Resistance ( Sheet SOL9235H/L 40 Q2 2011 -6.4% SOL9117 Q3 2009 SOL950 LCOE Q1 2009 Q3 2008

Goal: 1.0% efficiency gain in next 5 years = - 5.4% LCOE

17 9/17/2017 Presentation title | Author | Department SILVER (PASTE) USAGE PER CELL HAS BEEN CUT >60%!

2017

40µm

140 FS (mg/cell) Laydown Paste Ag 250 Screen Finger Opening (um) Opening Finger Screen 120 200 100

150 80

60 100

40 50 20

0 0 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 finger opening (um) laydown (mg)

18 9/17/2017 Presentation title | Author | Department FINER AND FASTER METALLIZATION LINES

19 9/17/2017 HERAEUS PHOTOVOLTAICS - LEADING THE FUTURE OF PV CELL OPTIMIZATION FURTHER CONTRIBUTION ENABLE BY PASTE

21,5

Recombination 20,5 Optical 1% = - 5.4% LCOE

20 Resistance EFFICIENCY 19,5

19 Ref. Cell Emitter BSF Bulk Contact Grid ARC Light Lateral Contact Metal recomb. recomb. recomb. recomb. shading trapping resist. resist. resist. IMPROVEMENTS

20 9/17/2017 HERAEUS PHOTOVOLTAICS - LEADING THE FUTURE OF PV 360° APPROACH TO REDUCING LCOE

o LID Friendly Metallization Paste Increased o Strong adhesion

Reliability: 1.8% 1.8% o Improved TC200 Performance

Low Low

Cost Ongoing Ongoing

o New technologies (PERC/PERL, Increased Device Super Fine Line Print) Efficiency: 12.7% 12.7% o Superior metallization contact on

High LDE Energy Energy Harvest o Selectively coated ribbon o Cell Optimization Consultancy Increased 20% Operational o Higher yield (Heraglaze +4% yield) Excellence 3.5% o Higher throughput (Metallization Paste screen printing speed 2x)

Lower Material o Electrically Conductive Adhesives Cost o Lower metallization paste Laydown Low InitialLow Cost 2.0% o Screen Bundle o Lamp Bundle

21 9/17/2017 HERAEUS PHOTOVOLTAICS - LEADING THE FUTURE OF PV REDUCING LCOE* OF RENEWABLE SYSTEM

Renewable system LCOE-reduction within next 5 years: 1/3 possible with Heraeus technologies

Reliability: 1.8% Efficiency: 12.7% Operational: 3.5% Materials: 2.0%

*LCOE = Levelized Cost of Electricity

22 9/17/201703/01/2017 Dr. Weiming Zhang |3rd PV paste and metallization workshop, 2017 THANK YOU 133+ GW MADE WITH HERAEUS

23 9/17/2017 HERAEUS PHOTOVOLTAICS - LEADING THE FUTURE OF PV II Voting

Thesis: “Cell technology improvements can reduce LCOE by 20 % in 5 years"

Agree? Disagree? Don’t know?

September 12, 2017 – Future PV Roundtable at SPI III Thesis

„Using front contact cells in shingled modules it is possible to increase power plant project net present value by more than 3 c/W compared to standard p-type modules “

September 12, 2017 – Future PV Roundtable at SPI SunPower® Performance Series Solar Panels

Josh Moore, SunPower Director

© 2017 SunPower Corporation All Rights Reserved. SUNPOWER and the SUNPOWR logo are registered trademarks of SunPower Corporation in the U.S. and other countries as well. SunPower >30 Years of Experience

• Founded in 1985, Silicon Valley

• Publicly traded NASDAQ (SPWR)

• Total majority investor

• Innovation platform: >750 patents

• $2.7B revenue in 2016 (non-GAAP)

• >7,000 employees

• Approximately 1.6GW/year solar cell capacity

• >7 GW of solar PV and >3 GW of trackers deployed

Conventional Panel 7% higher efficiency1 P-Series Tiled-cell Panel

+6% 350W +4% +3% Larger 310W More light: format Reduced no reflective Electrical copper lines Resistance and less white space

16% efficient 17% efficient

13% more power: 7% higher efficiency than Conventional panels plus 6% larger size1 1 Compared to a Conventional Commercial Panel (310W, 16% efficient, approx.1.94m2).

Tiled-cell panel Solar Cells Conventional Cell (Front)

1. Thin screen-printed metal lines on the front of the cells are protected from 1. High-stress solder joints between the long copper ribbons and crystal solar cell corrosion by SunPower’s specially engineered encapsulant • As the panels get hot in the day and cold at night the copper expands but the silicon cell does not. 2. No soldered ribbons along the length of the cell – one of the major failure modes of using traditional cells has been designed out of the panel. • Over time, this repeated stress causes cells to crack and solder bonds to break.

3. Cells are connected across their length, creating many redundant paths for 2. Single points of failure on copper ribbons between cells. electricity, and no single point of failure. 3. Very thin screen-printed metal lines on the front of the cell are susceptible to corrosion over time

Thermal Cycling: IBC & P-Series Vs. Conventional Corrosion Test: IBC & P-Series Vs. Conventional

Industry standard Tougher SunPower standards Industry standard Tougher SunPower standards

100% 1

0,8 95%

0,6 90%

0,4

Panel Power Panel Panel Power Panel 85% 0,2

20 years in a high humidity climate5 80% 0 0 200 400 600 800 1000 1200 0 1000 2000 3000 4000 5000 Number of Thermal Cycles Hours in Damp Heat

Conventional Panels1,2 SunPower® P-Series3 SunPower® IBC Number of Cycles: -40 to 85°C (-40 to 185°F), 5 cycles per day Hours in Damp Heat: 85°C (185°F), 85% relative humidity 1 Meakin, “PV Durability Initiative for Solar Modules.” 2013. 2 Ferrara, “PV Durability Initiative for Solar Modules: Part 2.” 2014. 3 Renewable Energy Test Center results, 2015 5 Kempe, et. al. Modeling the Ranges of Stresses for Different Climates/ Applications. PVMQA Forum. 2011

Conventional Panel P-Series Panel

POWER IN POWER OUT POWER IN POWER OUT

• SunPower operates a power plant with several brands of solar panels in Arizona, a hot and dry desert climate.

• Relative energy production is measured with panels on adjacent trackers in identical conditions.

• SunPower® P-Series panels demonstrate 5-6% more yield due to their superior performance in high temperatures1.

Energy Solar Panel Yield SunPower P-Series +5.5% Competitor A +2.4% Competitor B +1.4%

Competitor C +0% % Yield Relative to Conventional to Relative Yield %

1 Zweibaum. “Performance of P-Series installation at TEP – Technical summary.” 2017.

$0,05

$0,04 $0,034 $0,005 $0,013 $0,03

$0,02 $0,008 NPV ($/W) NPV

$0,01 $0,002 $0,007

$0,00 CT1 - 320W BOS savings O&M Savings Module Degradation Bankability / SunPower Performance Brand P17-350 Gain $3.4 million value increase per 100MW

Let’s change the way our world is powered.

Thesis: „Using front contact cells in shingled modules it is possible to increase power plant project net present value by more than 3 c/W compared to standard p-type modules “

Agree? Disagree? Don’t know?

September 12, 2017 – Future PV Roundtable at SPI IV Thesis

„Combining black silicon and perc technologies it is possible to increase IRR by 2,5 % and the next technology steps will increase it further"

September 12, 2017 – Future PV Roundtable at SPI GCL presentation

September 12, 2017 – Future PV Roundtable at SPI IV Voting

Thesis: „Combining black silicon and perc technologies it is possible to increase IRR by 2,5 % and the next technology steps will increase it further"

Agree? Disagree? Don’t know?

September 12, 2017 – Future PV Roundtable at SPI V Thesis

Steel frame PV modules with advanced cell technologies enable the reduction of ~20% in the installation and material costs of PV plants, as well as > 15% in the manufacturing carbon footprint of PV modules. It also has significant impacts on logistics, staff training and O&M cost structures.

September 12, 2017 – Future PV Roundtable at SPI Innovative Steel Frame PV Module with High Performance Q.ANTUM CELLS for Reducing System Costs and Carbon Footprints

Wai Lo Senior Director (Products) Hanwha Q CELLS America Hanwha Q CELLS Steel Frame PV Module: Q.PEAK RSF L-G4.2

An innovative PV module designed for: • Reducing PV plant installation and materials costs by ~ 20% – easy installation, simple parts • Simplifying mounting parts logistics – no multiple mounting parts • Streamlined O&M process – now with installation process on the backside only • Minimizing installer training efforts – simply parts, easy installation • Reducing PV module carbon footprints by >15% – steel vs aluminum frame and parts • Increasing PV plant energy production – high performance mono-Q.ANTUM cells

How do we achieve them:

• Hanwha Q CELLS reinforced steel frame technology • Hanwha Q CELLS easy mounting system • Hanwha Q CELLS Q.BOX junction box • Hanwha Q CELLS Q.ANTUM technology

Confidential 42 Hanwha Q CELLS Hanwha Q CELLS Reinforced Steel Frame Technology

• Frame formed with folded steel sheet process

• Better water drainage and minimized soiling impact with

front surface opening

• Better backside cooling with frame openings

• Zine-aluminum-magnesium coating provides long term

corrosion protection

• Increases conductivity because of clinching

• Reduces energy consumption and carbon footprint

Confidential 43 Hanwha Q CELLS Hanwha Q CELLS Easy Mounting System

• Easy installation – simple mounting key insertion, 90o twisting and self- tapping screws fixing • Simple part – only mounting keys needed • Easy O&M – installation, un-installation and re-installation can all be done from the backside • Minimal staff training required – thanks to the simple parts and installation process • Installation, logistics and material costs saving – simple installation process, only 78g mounting keys per panel and minimum training needed • Low carbon footprint – with only 78g mounting key per panel, which is to be compared with ~ 400 grams with aluminum clamps

Confidential 44 Hanwha Q CELLS Hanwha Q CELLS Q.BOX Junction Box

• Better heat dissipation – due to the de-centralized positioning of the

diodes

• Reduces cable length – due to the de-centralized junction boxes

• Reduces module resistive loss – shorter current paths

• Reduces handling and installation related cable and junction box

damages

• Reduces backsheet damages during installation

• Reduces BOS costs and carbon footprint – due to the substantial

cable length reduction

Confidential 45 Hanwha Q CELLS Hanwha Q CELLS Q.ANTUM Technology

• Q.ANTUM products: cell technologies that lead to high module power and efficiency, anti-LID, anti-PID, Hot-spot Protect, TRA.Q benefits • Exceptional low light behavior and temperature coefficients • Extensive Q.ANTUM manufacturing experience: > 2.5 GW Q.ANTUM shipped since 2012 • Strict quality assurance system • Industry top ranking warranty • All designed, developed and deployed to generate the best values for the PV plants

Initial certification tests Q CELLS Yield Security VDE Quality Tested Q CELLS Quality Program

Confidential 46 Hanwha Q CELLS In case of questions, please contact:

Wai Lo, Ph.D. Senior Director, Products Hanwha Q CELLS America Email: [email protected]

47 Hanwha Q CELLS V Voting

Thesis: Steel frame PV modules with advanced cell technologies enable the reduction of ~20% in the installation and material costs of PV plants, as well as > 15% in the manufacturing carbon footprint of PV modules. It also has significant impacts on logistics, staff training and O&M cost structures.

Agree? Disagree? Don’t know?

September 12, 2017 – Future PV Roundtable at SPI Session 2 Applications

September 12, 2017 – Future PV Roundtable at SPI I Thesis

"As developer it is possible to significantly reduce LCOE in the coming years by using flexible EPC contracting and pushing for promising technologies“

September 12, 2017 – Future PV Roundtable at SPI Reducing LCOE by Driving New Technology Adoption

September 2017 Recurrent Energy Business Overview

Recurrent Energy Has The Resources, Experience, Technical Expertise, and Access to Capital to Deliver Utility Solar at Any Scale

• Leading solar project developer helping our world • 4 GWp project pipeline, more than 2.3 GWp of sustainably meet its energy needs power contracts executed, more than 1.9 GWp • Over $8B in capital secured to date from leading of operating projects lenders and investors active in the energy sector • Seasoned leadership team with experience in • Proven access to capital through a network of conventional and renewable power businesses financial partners that enables us to deliver utility • Strong technology and supply chain expertise solar at any scale enable delivery of solar projects at market- leading cost 52 www.recurrentenergy.com Leading Offtakers, Counterparties & Partners

Offtakers Debt Finance

Equity Finance Partners

53 www.recurrentenergy.com Technology and LCOE Improvements Are Not Smooth or Orderly

Utility Scale PV BOS Price Decline

- Step changes in BOS and LCOE - Smooth curve over time

54 www.recurrentenergy.com Quick Case Studies Adopting new technology takes effort

55 Proven Partner to Technology Providers Recurrent Energy has a track record of using new technology Geoff Baxter SVP, Engineering, Procurement, Construction & Operations P: 415.501.9407 [email protected] www.recurrentenergy.com I Voting

Thesis: "As developer it is possible to significantly reduce LCOE in the coming years by using flexible EPC contracting and pushing for promising technologies“

Agree? Disagree? Don’t know?

September 12, 2017 – Future PV Roundtable at SPI II Thesis

"By integrating monitoring and analytics with SCADA and power plant controllers, it is possible to reduce downtime and LCOE, and additionally it is possible to increase the earnings with the solar energy produced"

September 12, 2017 – Future PV Roundtable at SPI High Performance Monitoring

The Future of Monitoring and Management Trends for PV Assets

Christos Georgopoulos: CEO Inaccess Steve Ritacco: CTO Draker/BlueNRGY WHO WE ARE

Monitoring & Control Systems

Utility-scale, C&I

State-of-the-art SCADA Technology Platform & end-to- end solutions

7GW of Solar installations under monitoring worldwide

bluenrgy.com FUTURE PV MANAGEMENT TRENDS

Open Architecture More accurate and higher systems and Portfolio resolution Business Consolidation Analytics

Integration of Incorporation of Energy Monitoring and SCADA Storage Solutions systems

Integrated Power Plant Management Platforms

bluenrgy.com BENEFIT 1: More efficient O&M

Example: Integrated Monitoring + SCADA Fast reaction times => less downtime remotely located 2 MW PV plant at 8.30am Recloser goes off due to a demand surge

Traditional O&M and Monitoring Diagnosis Alerts, Charts, RCA Truck Roll On-site Recloser Reset } 24h O&M with Integrated Monitoring and Control

Diagnosis Alerts, Charts, RCA Remote Recloser Reset } 1h

$1,000 (avoided truck roll) + Savings: $2,000 $1,000 (avoided loss, assuming a 10MWh/day production)

bluenrgy.com BENEFIT 2: Maximize Asset Performance

Example: Minimize unexpected downtime: Predict problems before they happen with smart analytics

Minimize Loss during downtime: 20 MW PV plant Data-driven corrective O&M using Performance Modelling

20 MW*1400KWh/yr => 28,000MWh/yr 1% availability difference => 280MWh or $28,000

Availability +5% +$140,000

Production

99% 94%

bluenrgy.com BENEFITS OF INTEGRATED POWER PLANT PLATFORMS Summarizing

Reducing O&M part of LCOE • Optimize Corrective Maintenance • Optimize costs and uptime via predictive maintenance • Efficiently manage equipment outages

Maximize Technical Asset Performance • Predict problems before they happen with smart analytics • Data-driven corrective O&M using Performance Modelling • Performance Optimization analytics

bluenrgy.com PLUS: EVOLVING GRID REQUIREMENTS Spinning Reserve Control Capabilities for all size plants Example • Manual and Automated Controls Typical MWh price: Grid Interactive $100/MWh • Prepared for emerging requirements • Avoid unnecessary curtailments • Forecasting for market participation Spinning reserve price: (from 2x up to >10x) Storage Integration • Convert Solar plants to Dispatchable $1,000/MWh assets • Participate in Frequency Response or Spinning Reserve market Opportunities Market Additional bluenrgy.com OPTIMIZE O&M /

MAXIMIZEIt’s not far – it’s happening PERFORMANCE right now Global Owner/IPPs use Integrated Platforms on several GW of installed power

bluenrgy.com II Voting

Thesis: "By integrating monitoring and analytics with SCADA and power plant controllers, it is possible to reduce downtime and LCOE, and additionaly it is possible to increase the earnings with the solar energy produced"

Agree? Disagree? Don’t know?

September 12, 2017 – Future PV Roundtable at SPI III Thesis

„Large scale production of hydrogen from PV power plants is already feasible and profitable today and will help financing PV plants already in the near future"

September 12, 2017 – Future PV Roundtable at SPI Large scale hydrogen production from PV power plants - feasible and profitable today

Bjørn Simonsen Vice President Market Development and Public Relations, Nel ASA Large scale renewable hydrogen production since 1927

One of several renewable hydrogen production plants at 135 MW / 30,000 Nm3/h

A Standardized 50MW electrolyser plant from Nel Hydrogen can support more than 40.000 FCEVs on annual basis. This corresponds to 100% renewable hydrogen supply to all FCEVs in California by 2022 as projected by the ARB. Additional electrolyser capacity can easily be added and more plants build as hydrogen demand increases.

40.000 FCEVs per year

24 x NEL A-485 electrolysers – 50MW

Hydrogen pump price of $7.1/kg matches gasoline on a fuel cost per driven mile basis. Renewable electricity at <$50/MWh and hydrogen infrastructure at scale can achieve this.

<$50/MWh <$6.7/kg Renewables Electrolysis Distribution Dispensing

Pump price Fuel economy Fuel cost / mile Toyota Camry GASOLINE $3/gallon → 28 MPG → $0.11/mile ↓ 66.85 miles/kg HYDROGEN $7.1/kg ← ← $0.11/mile Toyota Mirai (65 MPG equivalent)

One 50MW electrolyser can provide renewable hydrogen to 40,000 FCEVs in California by 2022, at a cost competitive level with gasoline

Factors which will delay going to scale: ● Uncertain if renewable share will be 100% (33% required today) ● Market demand will be addressed by multiple electrolyser plants (competition) ● Sufficient demand for one plant will require more than 40,000 FCEVs

Public contributions could address gasoline price gap until sufficient scale is achieved: ● CAPEX support (e.g. competitive grants) ● OPEX support (e.g. LCFS or waived grid fees)

Hydrogen production connected to renewables can reduce risk of curtailment and enable a higher renewable share in the grid ● Day time: Electrolyser utilizes solar power or grid power and reduces risk of curtailment stemming from higher RE shares ● Night time: Electrolyser draws power from the grid or wind and reduces production during high demand periods ● Drawing power from both grid and renewables enables a more steady and less costly hydrogen production (more operation hours)

Despite that hydrogen production would help the grid, various grid fees still apply when hydrogen production draws from the grid. Waiving of grid fees for hydrogen production e.g. under the Net Energy Metering (NEM 2.0) will help reduce hydrogen costs, incentivize energy companies to install hydrogen production assets, and reduce the “Duck Curve” effect.

MW Solar production MW Wind production

Hydrogen Reduced curtailment Grid fees Reduced curtailment Grid fees production Grid fees

Time of day Time of day

1. Hydrogen production from renewable energy can be done at a large and cost competitive scale 2. At below 0.05 $/kWh hydrogen can be competitive with gasoline at the pump in California, all costs included 3. Market demand for hydrogen fuel is needed to make the business case fly 4. Electrolysers connected to renewable energy generating assets and the grid, will enable an increased share of renewables, while reducing the need for curtailment, alleviating “Duck Curve” issues

Nel has the necessary technology & know-how to design an optimal solution for solar to hydrogen

Hydrogen fueling stations Alkaline and PEM electrolysers Fast fueling of all types of vehicles, enabling Converting water and electricity to same long range as today hydrogen and oxygen

www.nelhydrogen.com III Voting

Thesis: „Large scale production of hydrogen from PV power plants is already feasible and profitable today and will help financing PV plants already in the near future“

Yes? No? Maybe?

September 12, 2017 – Future PV Roundtable at SPI IV Thesis

„It is possible to achieve solar penetrations higher than 50%“

September 12, 2017 – Future PV Roundtable at SPI The Future of Solar PV

Raffi Garbedian CTO, First Solar

Enabling a world powered by reliable, affordable solar electricity. Images from: LucentLabs /Bell PV STARTED AS A VISION… A AS STARTED PV …of abundant, clean,…of abundant, affordable electricity. 81 17 September 2017 © Copyright First Solar, Inc. • • • Image of THE VISION IS REALITY. WE NOW LIVE IN A WORLD WHERE… WORLD A IN LIVE NOW WE REALITY. IS VISION THE to highpenetration Is the nextGrid Flexibilitychallenge resolved been has The challenge cost affordablePV sources. by safe, scalecarbon clean, at large be suppliedElectricitycan Topaz Topaz Solar Farm from FirstSolar - free, free, and 82 17 September 2017 © Copyright First Solar, Inc. THE CAISO “DUCK CURVE” IS A CALL FOR GREATER SYSTEM FLEXIBILITY SYSTEM GREATER FOR CALL A IS CURVE” “DUCK CAISO THE Data source: CAISO. Annotations: Solar First Limits on downward flexibility: — — — Abnormal imports Other Inflexibleresources ( generation The need for intra for need The Load self Biomass - high hydrohigh scheduled scheduled generation conditions conditions 2017in - hour & multi & hour nuclear and must , CHP - take Solar ) - hour flexible capacity flexible hour Wind Nuclear Net Net Interchange Thermal Hydro 83 17 September 2017 © Copyright First Solar, Inc. Data Data Annotations:source:First Solar CAISO. ANOTHER WAY TO LOOK AT THE GENERATION STACK GENERATION THE AT LOOK TO WAY ANOTHER High Load High DayLoad Low Load Low Load Day Must Minimum Minimum - Run 84 17 September 2017 © Copyright First Solar, Inc. HYPOTHETICAL FUTURE GENERATION STACK GENERATION FUTURE HYPOTHETICAL High Load High DayLoad Low Load Day “Must Minimum Minimum - Run” 85 17 September 2017 © Copyright First Solar, Inc. HIGH SOLAR PENETRATION | | PENETRATION SOLAR HIGH • • outside of solar hours solarofoutside to Netpeakshifts load peakloads day of a part Solaris - time load offsetting near time offsetting load NEW SOLAR Grid Friendly and Load MARKET Solar Net Load Dispatchable NEW CONSTRAINTS AND OPPORTUNITIES EMERGE OPPORTUNITIES AND CONSTRAINTS NEW • • solar more Responsesupports FlexibilityDemand + & Operational Market penetration forvital valueits increasing & Flexible& controllable solar Net Load MATURE SOLAR Load MARKET Solar … Next Phase of SolarExpansion Solar System System Ramp by by Targeted Reduced Reduced Control Control • • Net Load SOLARMARKET 2.0 penetrationofsolar evenenables higher flexibility grid Increased penetrationpotential valueits increasing and Storage time Load Solar - shifts solar, shifts Storage 86 17 September 2017 © Copyright First Solar, Inc. Sources:Climate PolicyInitiative 2016Analysis; fromData WE’RE AT THE START, NOT THE END OF THIS JOURNEY THIS OF END THE NOT START, THE AT WE’RE If energy could be shifted by these many could days If energy be shifted 10%Wind 9% Solar 2016 California Energy Commission Commission EnergyCalifornia With added flexibilityandWith added market adjustments market >40% of demand>40% – Tracking Progress could meet could California with PV with 87 17 September 2017 © Copyright First Solar, Inc. IV Voting

Thesis: „It is possible to achieve solar penetrations higher than 50%“

Agree? Disagree? Don’t know?

September 12, 2017 – Future PV Roundtable at SPI Future PV Roundtable at Solar Power International 2017

Pushing the boundaries: the latest steps to decrease LCOE

Initiative partner Gold sponsors Bronze sponsor

September 12, 2017 – Future PV Roundtable at SPI