3 Canadian Solar-Guoqiang XING

CSIQ NASDAQ Listed Mass Production Technology Advancement and Challenges in PV Guoqiang Xing, Ph.D Sr. VP and CTO, Canadian Solar SNEC, Shanghai, China June 4, 2019 2Canadian Solar Inc. Canadian Solar overview What makes CSI different Global/Local Module + PV Project Development Profitability ahead of market share Prudent, yet aggressive Innovation Module Shipments - MWFounded in Ontario, 2001 Listed on NASDAQ (CSIQ) in 2006 Presence in 18 countries / territories No.1 Silicon module solar plant developer by GTM 2017 No.1 Module supplier for Quality and Performance/Price by IHS Module Customer Insight Survey BNEF Tier 1 solar company PV Project Development 9.5 GWp Total pipeline 6.6 GWp Early to mid-stage development pipeline 2.9 GWp Late-stage project pipeline in 13 countries 1.15 GWp Solar power plants in operation in 6 countries 4.6 GWp solar power plants built and connected in 15 countries 33 GW modules shipped 3Canadian Solar Inc. Moore s Law in PV – increasing learning rate ➢ Learning rate (LR) of PV modules is now at 39.8% (2006-2018) ➢ Technology advancements are the engines of such high LR, in addition to economies of scale ➢ Milestones of global cumulative module shipment： 2012: 100GW 2015: 200GW 2017: 400GW 2018: 500GW 2022?:1000GW ➢ LR back to the historical mean of 23.2% for the next doubling to 1TW before 2022? Source: International Technology Roadmap for PV (ITRPV), 2019 10th Edition, https://itrpv.vdma.org 4Canadian Solar Inc. Solar cell technology evolutions Source: Bloomberg NEF, 2018 PV Manufacturing Overview: from Polysilicon to Module ➢ PERC becomes the mainstream, first mono PERC, then mc-Si PERC in 2020/21 ➢ p-type c-Si technology keeps 95% market share, to reach 50:50 splits between mono and mc- Si by 2021 ➢ n-types do not see significant market share climb? 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% 2016 2017 2018e 2019e 2020e 2021e N -t y pe m ono I B C N -t y pe m ono H I T N -t y pe m ono PER T P-t y pe m ono PE R C / PER L P-t y pe m ono Al- BSF P-t y pe m ult i PER C P-t y pe m ult i Al-BSF Low H igh Ef fic ienc y mc-Si Al-BSF Mono PERC mc-Si PERC Mono Al-BSF 5Canadian Solar Inc. mc-Si technology advancement in the last decade G5→G6 HPM P3: DWS+MCCE P4: Multi-PERC P5: Cast-mono mono efficiency at multi cost Larger sized wafer 157mm → 166mm 6Canadian Solar Inc. Care-about for PERC and bifacial Full Al-BSF PERC Bifacial PERC • Mono and Multi • multi Black-Si is the mainstream • mc-Si module are the majority of the 500GW installations • Mono and Multi • PERC cell efficiency +1% • PERC makes LID/LeTID more important • Higher wattage makes hotspot an issue. Half-cut becomes essential • Mono and Multi • PERC cell make bifacial possible • PERC bifacial + double glass make high volume bifacial module possible • Module + system integration, LCOE driven. 7Canadian Solar Inc. LID “Light-Induced Degradation” LID & LeTID – What’s behind the barbaric acronyms? LeTID “Light and Elevated Temperature Induced Degradation” PID, LID solved. LeTID remain an issue for the ones without technology understanding. ➢ Fast degradation within 1st month in field + recovery. ➢ Mechanism well-known related to Boron- Oxygen defects. ➢ PERC is more sensitive than non-PERC ➢ Mono is more sensitive related to bulk material. Can be solved by Ga-doping, low- oxygen, etc. ➢ Slow degradation + recovery. ➢ Mechanism not fully known, theories of Hydrogenation induced degradation, passivation degradation, etc. ➢ Higher temperature accelerate degradation significantly. ➢ Mono, multi and n-type all sensitive. Innovative technology needed to mitigate. PL co un t PL co un t EL of LeTID sensitive module Time (Hour) Time (Hour) 1000 1000 8Canadian Solar Inc. LeTID comparison – multi vs. mono ➢ Both mono and multi-PERC showed LeTiD ➢ All multi-PERC modules LeTID 2%. And on average mono PERC LeTID is higher by 1.7% (absolute) R. Gottschalg et al., Fraunhofer CSP, 35th EUPVSEC, 2018 9Canadian Solar Inc. LeTID mitigations by Canadian Solar ➢ Ingot/Wafer: impurity and dopant control ➢ Cell: optimized process to suppress defect formation and proper hydrogen passivation ➢ “CSAR”: Canadian Solar Advanced Regeneration Material Quality Optimized Cell Process Advanced Regeneration Int ensi fie d Inl ine Contr ol 10Canadian Solar Inc. 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% 17-Q1 17-Q2 17-Q3 17-Q4 18-Q1 18-Q2 18-Q3 18-Q4 19-Q1 19-Q2 19-Q3 19-Q4 20-Q1 20-Q2 20-Q3 20-Q4 P5 P4 P3 P2 P4 phased in Q3/2017 P5 phased in Q3/2018 Fast cell technology evolutions at Canadian Solar • P3: All DWS+MCCE since 2017/9 • P4: 1GW by end of 2017, and 4GW by end of 2018 • P5: Start to phase in from Q3,2018 • All-PERC by Q2,2019 – many years ahead of ITRPV P2 out in Q3/2017 P3 to be out by Q2/2019 mc-Si Cell Capacity by Technology 11Canadian Solar Inc. PERC efficiency roadmap Fine line SP Oxidation Baseline Emitter Resistivity Plating Bulk lifetime 20202018 Simulated With Quokka3. 12Canadian Solar Inc. Canadian Solar’s P5 - another big step for multi mono efficiency @ multi cost A 5cm*5cm mono-like brick was grown based on the mono seed by Ted Ciszekt in 1979. Mono area Mono areaMono area Multi Multi Multi Multi 13Canadian Solar Inc. Black Si-invert pyramid Alkaline-random pyramid Extending MCCE to P5 100.0% 100% 99.1% 100.13% 95.0% 96.0% 97.0% 98.0% 99.0% 100.0% 101.0% 电池 组件 Cell eff and module power +0.13% Alkaline Black Si Alkaline Black Si Cell Module 14Canadian Solar Inc. CSI P5 cell and module performance • Cell efficiency increased 1.5% vs P4 • The cell efficiency difference with mono is within 0.3% • Module power output is 16Watts higher than P4 • The power difference with mono is within 5Watts P5 module P5 module EL P4 P5 Mono Cell Eff 20.5% 22% 22.3% Module Power 370 386 390 15Canadian Solar Inc. Canadian Solar sets world record – 22.28% based on P5 ✓ Ultra-low dislocation p-type multi crystalline wafers ✓ MCCE black silicon ✓ Selective emitter ✓ SiO2 passivation ✓ Multi-layer ARC ✓ Al2O3 back passivation ✓ Advanced metallization • P-type 157mmx157mm large area multi crystalline wafer • Tested and certified by Fraunhofer ISE of Germany 16Canadian Solar Inc. 2016 2017 2018 Module technology evolutions 5 Bus bar (5BB) More wattage, less crack risk MBB dual cell Lowest crack risk More energy per watt Dual cell w/166mm wafer 400W Poly module Lower LCOE Bifacial dual cell 5-20% more energy 30 years warranty 2019 Double glass 30 years warranty Bifacial dual cell with 166mm wafer 400W Poly module 5-20% more energy 30 years warranty Lowest LCOE All black shingling ~20% module eff. Perfect all black More powerful More reliable More energy per watt More tailored Dual cell More wattage, cooler temperature, lower hotspot risk, better shading tolerance Shingling High power density, lowest hotspot risk Better shading tolerance 17Canadian Solar Inc. Canadian Solar module product roadmap Standard Full Cell 2017 Ku Module Half Cell P3 (Poly BSF/Black Silicon) P4 (Poly PERC) P5 (PERC) 400 W+!2017 HDM Shingled Cell 2019 HiDM Shingled Cell 2018 HiKu High Power + Ku 2018 BiKu Bifacial + Ku 400W+! 2019 HiKu+P5 High Power + Ku All Black 2019 BiHiKu Bifacial + HiKu Tracker Optimized 430W+! 5BB MBB 1000V 1500V Mono BSF/PERC Mono PERC Backsheet Bifacial Double Glass 18Canadian Solar Inc. Canadian Solar module capacity by technologies 19Canadian Solar Inc. 144 cells 360 ~ 400 W multi or mono 992x2000 mm 120 cells 300 ~ 330 W multi or mono 992x1675 mm Dual-cell module advantages Lower NMOT: 1~2 °C lower Lower temperature coefficient (Pmax): -0.37 %/°C Better shading performance Lower hotspot temperature: 20 °C lower than full cell ➢ Resistive power loss reduced by: 75% ➢ Module power output increased by: 4% i 1/2i Full-cell Dual-cell HiKu 3W 400W in 2018 144 cells 400 ~ 410 W multi 1048x2108 mm Ku 3U Ku 3K 20Canadian Solar Inc. -1.6% -1.7% -1.7% -1.0% -2.8% US Brazil South Africa -2.2% China Australia Italy Saudi Arabia -1.7% Canadian Solar’s 400+W HiKu module–LCOE vs mono PERC ➢ Expected LCOE reduction in the 1.0%~2.8% range for different markets and geographies. 21Canadian Solar Inc. SAM Bifacial – Hot or Hype? How to best optimize bifacial systems? Is bifacial system yield modeling accurate? Are bifacial modules and systems reliable? B ➢ Key challenges hinder bifacial system higher market penetration A C 22Canadian Solar Inc. Changshu 24kW BiKu, SAT 1P Albedo ~0.25 Diffuse ratio ~58% 8.9% 1 year SANDIA 11kW BiKu, SAT 1P Albedo ~0.2 5.2% 3 months Wuhai 150kW BiKu, SAT 1P Albedo ~0.3 Diffuse ratio ~50% 7.3% 5 months U.S. China Mojave Desert 22kW BiKu, SAT 1P Albedo ~0.3 Diffuse ratio ~30% 5.0% 4 months BITEC 18kW BiKu, SAT 2P Albedo ~0.2 / 0.6 6.4 - 15.6% 7 months 4.5% 6 months Wuhai 150kW BiKu, Fixed Changshu 12kW Albedo ~0.62 17.5% 3 months Suzhou 80kW BiKu/BiHiKu, Fixed Albedo ~0.6-0.8 Diffuse ratio ~58% Waiting grid connection Boulder 80kW BiKu, SAT 1P Albedo ~0.3 12% 2 weeks Canadian Solar’s bifacial testing sites Darlington ~2MW BiHiKu, SAT 1P Albedo ~0.3 Australia 23Canadian Solar Inc. • Technology and product evolutions have been at a much faster pace in the 2-3 years. • While PERC technology is becoming the mainstream cell technology, bifacial, half-cut, shingling, and bigger wafer are the module technologies driving down LCOE. • Much closer interaction between module and system technologies. • Ability to develop and adopt new technologies are essential to PV companies old and new. Summary 24Canadian Solar Inc. THANK YOU