In Association With: ISSUE 4 – April 2011 – CLICK HERE to ENTER >> Turnkey Module Production Lines

In Association with: ISSUE 4 – April 2011 – www.futurepv.com CLICK HERE TO ENTER >> Turnkey Module Production Lines

Our technological innovations mean that our customers are able to Flexible and modular system manufacturemodulesofhighestqualityatlowestTotalCostofOwnership. - fastest line expansion at minimum cost We cover all the key technologies which are important for the production Turnkey lines from a single source of solar modules. - saving time and costs due to simplification of work Excellence in know-how transfer and certification - save up to 6 months in time to market

3S Swiss Solar Systems AG / Schachenweg 24 / CH-3250 Lyss / Switzerland Phone +41 (0) 32 391 11 11 / www.3-s.ch

A member of Meyer Burger Group Future Photovoltaics CONTENTS |

“The quest to replace fossil-fueled electrical grids…” T. Patrick Walsh - p21

Welcome… FUTURE VISIONS SILICON WAFER & CURRENT CONCERNS PHOTOVOLTAICS …to Future Photovoltaics’ fourth issue, mark- 13 Introduction 32 Introduction ing our first full year of issues (FPV is tri- annual). As we’re getting increased input and Wim C. Sinke – ECN Solar Energy Jörg Müller – Q-Cells SE feedback from a readership that’s diversifying 14 Driving Into the Future: 33 N-type Bifacial Solar Cells at a surprising rate (38.6 percent of readers Solar Roadways for Industrial Application are from China now; welcome!), we’re trying Scott Brusaw – Solar Roadways V.D. Mihailetchi,1 A. Edler,1 R. Harney,1 to bring new ideas and concepts to you in R. Kopecek,1 T.S. Böscke,2 D. order to break out of what’s been described to 21 Kerosene Parity Stichtenoth,2 T. Aichele,2 J. Lossen2 – 1 us as a “solar farm mentality.” T. Patrick Walsh – International Solar Energy Research Greenlight Planet Inc. Center - ISC Konstanz, Germany 2Bosch Solar Energy AG, Germany With this in mind, we’d like to highlight the first in a series of articles that bring ground- breaking PV ideas to your attention in the NEW TECHNOLOGIES works of Greenlight Planet and Solar & MATERIALS Roadways. Both ideas use PV technologies to solve energy issues in very different ways, 25 Introduction and serve to highlight the prospect that PV Jef Poortmans – imec can not only change energy generation, but 26 Thin Is In, But Not Too Thin! change it at a societal level. These articles will be followed by submissions from other K.V. Ravi – Crystal Solar, Inc. visionaries; some more well known, others that we hope will inspire your own ideas.

The Future Photovoltaics team

“The interest in harnessing the sun for energy…” Louay Eldada - p40

THIN FILM MANUFACTURING EQUIPMENT PHOTOVOLTAICS & MATERIALS 39 Introduction 53 Introduction Danielle Merfeld – Craig Hunter – Intermolecular GE’s Global Research Center 40 Monolithic CIGS Photovoltaic Modules Manufactured by 54 Thought Reactive Transfer Leadership Profile Louay Eldada, Dingyuan Lu, Scott Hinson, Billy J. Stanbery – HelioVolt Corporation 56 Xjet’s FabGrade Inkjet 48 Brave Steps Needed in Thin Film For Solar Cell Metalization Silicon R&D to Harvest Optimal Robert Even – Xjet Ltd. Technology Torsten Brammer – Photovoltaics research consultant

“The goal for production is to maximize equipment uptime...” Jean Randhahn - p61 For the full versions of the following biographies, please click here

METROLOGY, TEST MODULE & Robert Birkmire & FAILURE ANALYSIS PANEL CONSTRUCTION Director, Institute of Energy Conversion; Professor of Materials Science, Professor of Physics 60 Introduction 67 Introduction Robert Birkmire is director of the Institute of Energy Conversion, a U.S. Department of Energy Alain C. Diebold – CNSE William Richardson – Center of Excellence for Photovoltaic Research and Education, as well as professor of mate- SOLON Corporation rials science and engineering with a secondary appointment as professor of physics. He is 61 Optimizing Approach for author of over 150 technical publications and is inventor on eight U.S. patents. Thin Film PV Production 68 Encapsulating Thin Cells: The i-module Approach Jean Randhahn,1 Stefan Gritsch2 – Richard Bozicevich 1Dr. Randhahn solar engineering K. Baert, J. Govaerts, J. Poortmans – imec 2Dr. Schenk GmbH Vice President of Business Development; TUV Rheinland Photovoltaic Testing Laboratory - PTL

Mr. Bozicevich is the VP of Business Development for TUV Rheinland Photovoltaic Testing Laboratory, a division of TUV Rheinland Group. He is responsible for the market development PHOTOVOLTAIC of photovoltaic testing services in the North and South American Markets. Richard received a B.S. in electrical engineering from Michigan Technological University, and currently sits on a SYSTEMS number of commercial and technology advisory boards for solar technology implementation. 76 Introduction Lior Handelsman – Christoph J. Brabec SolarEdge Technologies i-MEET Chair, Friedrich Alexander University Erlangen-Nürnberg 68 A Simpler BOS = A Faster ROI Christoph holds the chair "materials for electronics and energy technology (i-MEET)" at the Fannon Lim – Solar Research Pte Ltd., materials science department of the Friedrich Alexander University Erlangen-Nürnberg. He Luna Road LLC is also the scientific director of the Erlangen division of the Bavarian research institute for renewable energy (ZAE Bayern, Erlangen). He received his Ph.D. (1995) in physical chemistry from Linz University. Christoph has authored and co-authored more than 150 papers and holds 200 patents and patent applications.

Torsten Brammer Torsten Brammer, photovoltaics research consultant

Torsten Brammer has focused on photovoltaics since 1993. He graduated from the University of Karlsruhe (Germany) and the Australian National University, Canberra (Australia). After his degree, he joined a solar company in Ghana (West Africa). At the Research Center Jülich (Germany), he did his dissertation on thin film PV. In 2004, he joined Q-Cells SE, where he was responsible for the optimization of the crystalline silicon solar cell production process. In 2006, he became technical managing director of the Q-Cells spin-off Sontor GmbH, which was merged with Sunfilm AG in 2009. At Sontor and Sunfilm (now Schüco TF GmbH), he was responsible for R&D. Since 2011, he has been offering consulting services for technology, product design and business planning. Also since 2011, he has been offering consulting services in the area of metrology systems together with his associate Jörn Suthues.

Alain C. Diebold Hansjörg Lerchenmüller Empire Innovation Professor of Nanoscale Science, College of Nanoscale Science Co-Founder of Concentrix Solar, now a division of Soitec and Engineering, University at Albany; AVS Fellow; Senior Member of IEEE

Alain’s research focuses on the impact of nanoscale dimensions on the physical properties of Hansjörg Lerchenmüller is the co-founder of Concentrix Solar, now a division of Soitec. He materials. He also works in the area of nanoelectronics metrology. Alain is a member of the graduated with a degree in physics from the University of Karlsruhe, Germany. Before. International Metrology Technical Working Group, founder and co-chair of the U.S. Metrology Lerchenmüller became CEO of the Fraunhofer spin-off Concentrix Solar, he worked for 10 Technical Working Group for the 2007 International Technology Roadmap for Semiconductors, years at the Fraunhofer Institute for Solar Energy Systems ISE in Freiburg, Germany. Mr. and chair of the manufacturing Science and Technology Group for the American Vacuum Society. Lerchenmüller was among the initiators of the CPV Consortium, a globally acting industry organization for which he is a member of the board of directors.

Bryan Ekus Dean Levi Managing Director, International PV Equipment Association Principal Scientist National Center for Photovoltaics; National Renewable Energy Laboratory

Bryan Ekus is the managing director of the International PV Equipment Association (IPVEA). Dean Levi is currently a line manager and senior technical staff member in the National In this role, he is responsible for overseeing the associations operations and strategic posi- Center for Photovoltaics at the National Renewable Energy Laboratory. He received his Ph.D. tion in the PV industry. Bryan is also the managing director of the MEDIA-TECH Association in physics from the University of Illinois at Urbana-Champaign in 1990. that serves the packaged media industry. He has over 25 years experience working interna- tionally and currently resides in Orlando, Florida. Philippe Malbranche Research Programme Manager, CEA-INES

Robert E. Geer Philippe Malbranche is the research programme manager at CEA-INE. He received his engineering degree at the Ecole Centrale de Paris. He is also a member of the Steering Committee Professor, Nanoscale Science; Vice President, Academic Affairs of the European Photovoltaic Technology Platform. College of Nanoscale Science and Engineering; University at Albany

Robert E. Geer is a professor of Nanoscale Science and vice president for Academic Affairs in Brad Hines the CNSE at the University at Albany. He has published more than 100 technical articles, Vice President of Engineering, Idealab books and book chapters. Professor Geer’s research in emerging energy applications, nanoelectronics and nanomaterials has been supported by the National Science Foundation, the Office of Naval Research, the Department of Energy, the Air Force Office of Scientific Brad Hines is VP of engineering at Idealab, a creator of a number of cleantech, Internet and Research, the Semiconductor Research Corporation, DARPA, International SEMATECH and technology startups. He graduated from the Massachusetts Institute of Technology with a the New York office of Science, Technology and Academic Research. B.S. and M.S. in electrical engineering.

Danielle Merfeld Oliver Mayer Director, Solar Technology Platform; GE’s Global Research Center Principal Scientist for Solar Systems; Head of Quality at GE Global Research, Munich

Danielle Merfeld is the director of the Solar Technology Platform at GE’s Global Research Oliver Mayer is a principal scientist at GE Global Research in Munich, responsible for solar Center. She is responsible for managing the PV-related projects across the Center with top- system technologies and energy concepts for complex systems such as cities or hospitals, ics ranging from material development to grid controls. Danielle received her B.S. in electri- etc. For more than 20 years he has worked in the field of solar systems, having gained field cal engineering from the University of Notre Dame, and her Ph.D. in electrical engineering experience by installing PV systems in such countries as Jordan, Eritrea, Uganda, Chile and from Northwestern University. She has authored or co-authored over 60 papers in refereed Germany. Oliver received his Ph.D. from UniBwM, and is an honorary professor for solar sys- technical journals and has given scientific presentations at conferences and symposiums tems at the Munich University of Applied Science. around the world.

Craig Hunter Steve Roberts Vice President and General Manager Automation Engineer, Heliovolt Corporation Clean Energy Technologies, Intermolecular

Craig Hunter runs Intermolecular’s Clean Energy Technologies Group. Some of Mr. Hunter’s Currently an automation engineer with Heliovolt Corporation, Steve recommends and previous roles include: senior manager for the E-beam Test (EBT) and PVD Products of AKT, implements SCADA, MES, SPC, automated material handling and equipment control systems Inc.; CFO of Evercare Corp.; and director of M&A at The Beacon Group. He received a B.A. in for thin films photovoltaic panel manufacturing. For more than 20 years, he has been a soft- East Asian Studies from Harvard College and graduated with high distinction from Harvard ware/control systems product manager and engineer in the semiconductor and electronics Business School. manufacturing industry. Christopher Case Mike Moore Independent Consultant, Solid State Solutions Vice President, SVTC Solar

Christopher Case is currently an independent consultant with Solid State Solutions (UK), Mike Moore is vice president of SVTC Solar, where he is responsible for guiding the compa- serving the photovoltaics, electronics, chemicals and gases industries. He chairs the ny’s Solar strategy and overseeing day-to-day operations of the Solar facility. He holds a BSEL Interconnect Working Group of the ITRS. At Brown University, Christopher pursued photo- from Cal Poly State University, San Luis Obispo, and an MSEE from San Jose State University. voltaics, where he was an assistant professor of engineering and director of the Thin Film Institute. Christopher was a Fulbright-Hays scholar at the Université de Bordeaux and earned a Ph.D. in materials science from Brown University. Brent Nelson Group Manager, Process Development & Integration Lab; Jörg Müller National Renewable Energy Laboratory Director R&D Cells, Q-Cells SE Brent Nelson is PDIL group manager for the NREL, where he oversees a laboratory designed to integrate as many of the PV technologies and material characterization techniques as possi- Jörg Müller is director of R&D Cells at Q-Cells SE, one of the leading worldwide PV compa- ble. He has authored 18 publications as well as co-authored an additional 77 other scientific nies with a strong technology focus. In this role, he is responsible for the development of publications. Brent holds a B.S. in engineering physics from the Colorado School of Mines. the crystalline silicon solar cell technology. Jörg holds a master’s degree in physics from the University of Munich and a Ph.D. from the University of Hannover. Kristian Peter Rommel Noufi CEO, ISC Konstanz Principal Scientist, National Renewable Energy Laboratory

Dr. Peter is a co-founder of ISC Konstanz and a member of the board of directors, as its CEO. Rommel Noufi is a principal scientist with the National Renewable Energy Laboratory in Since January 2007, he has been working full time as a researcher, leading the department Golden, Colorado. He is also a visiting professor at Stanford University. Rommel received his of advanced solar cell processes. Kristian obtained his M.S. in physics in 1993 at the Ph.D. in analytical/physical chemistry from the University of Texas. He has published more University of Konstanz, and his Ph.D. in applied solid state physics in 1997. than 170 papers and has been issued eight NREL patents.

Jef Poortmans William Richardson Department Director Solar and Organic Technologies, imec Head of Research & Development, SOLON Corporation

Jozef Poortmans is program director of the Strategic Program SOLAR+ at imec. He has William Richardson heads Research & Development for SOLON Corporation, where he over- authored or co-authored nearly 350 papers that have been published in conference proceed- sees module testing and product development as well as strategic management and evalua- ings and technical journals. Jef is a board member of the EUREC agency and was general tion of new technologies for the North American market. Bill holds dual degrees in Renewable chairman of the 21st European Photovoltaic Solar Energy Conference & Exhibition. He Natural Resources and Electrical Engineering from the University of Arizona, where he cur- received his Ph.D. in 1993 on strained SiGe-layers. rently sits on the advisory board for the Material Sciences and Engineering Department.

<< PREVIOUS PAGE | NEXT PAGE >> 10 | Future Photovoltaics | April 2011 www.futurepv.com | 11 Future Photovoltaics EDITORIAL PANEL | FUTURE VISIONS & CURRENT CONCERNS PRINT E-MAIL For the full versions of the following biographies, please click here Click here to return to Table of Contents this article this article

Wim C. Sinke Wim C. Sinke Staff Member Solar Energy Energy Research Centre of the Netherlands Staff Member Solar Energy Energy Research Centre of the Netherlands Prof. dr. Wim C. Sinke is staff member Program & Strategy in the Solar Energy unit of the Energy research Centre of the Netherlands (ECN) and part-time professor at Utrecht University on “Science and Applications of Sustainable Energy Systems.” He is an active member of the European Photovoltaic Technology Platform. Sinke has 30 years’ experience Where technology and passion meet ... in photovoltaics. In 1999, he received the Royal Dutch/Shell Prize for Sustainability and Energy for his work in the field of photovoltaic solar energy. Photovoltaics: from milliwatts to coin. Photovoltaics is like a big box of Lego (or Robert David Vinje gigawatts with only one building block – the K’nex or Meccano or Lincoln Logs or whatever solar cell. From small stand-alone systems you grew up with).You can do almost anything Managing Director, SunPower Malaysia for rural use via medium-sized grid-connect- with it and you will never get bored. ed, building-integrated systems to large-scale It only takes a creative kid to build an Robert Vinje is managing director for SunPower’s 3rd Cell Manufacturing Facility in Melaka, power plants with just one component: the endless number of “products”; products that Malaysia. He resides on the Board of Trustees of Investors for First Philippines Industrial solar module. Examples illustrating two of are fun to create, but also very inspiring to Park, and chairs the Malaysian Alternative and Renewable Industry committee in Malaysia. Vinje has an electronics degree, an electrical engineering degree and an MS degree in man- the major strengths of photovoltaics: modu- others, and useful, if not indispensible: aiding agement of technology from the University of Minnesota. larity and versatility. development and improving health in rural This issue of Future Photovoltaics brings areas of our globe, combating climate change, Lior Handelsman us two exciting and, above all, personal stories increasing energy security, creating green VP Product Strategy & Business Development, Founder; SolarEdge Technologies about very different types of applications. At jobs, and more. first sight they have little in common: One is T. Patrick Walsh of Greenlight Planet Inc., about rural lighting powered by photovoltaics, and Scott Brusaw of Solar Roadways share Lior Handelsman is responsible for the company's product management and business the other about driving into the future on solar their passion and entrepreneurial drive with development activities. Mr. Handelsman holds a B.Sc. in electrical engineering (cum laude) and an MBA from the Technion, Israel Institute of Technology in Haifa. roadways. When reading them, however, I the readers of Future Photovoltaics. Read and found that they are just two sides of the same enjoy.

Publisher and CEO | Nikki Wood Editorial and Digital Projects Director | Matt Grimshaw Future Photovoltaics – © 2011 Mazik Media, Inc. Art Director | David Witcomb The entire contents of this publication are protected by copyright, full details of which are available from the pub- Editor | Lisa Barr lisher. All rights reserved. No part of this publication may be Web | Leslie Sumrall reproduced, stored in a retrieval system or transmitted in any form or by any means - electronic, mechanical, photo- Production Manager | Jane Ellis copying, recording or otherwise - without the prior permis- Europe & Asia sion of the copyright owner. Account Manager | Stuart Blacklock Production/Admin Intern | Nicki Rosenberg While every effort is made to ensure the accuracy of the contents of this book, the publisher will accept no responsibility for any errors or omissions, or for any loss or damage, consequential or otherwise, suffered as a result of any Published by: Mazik Media Inc. material here published. The publisher assumes no respon- 38 Miller Ave | Suite 9 | Mill Valley sibility for statements made by their advertisers in business CA 94941 | USA competition, nor do they assume responsibility for state- O: +1 415 381 6800 | F: +1 415 381 6803 ments/opinions expressed or implied in the articles of this Email: [email protected] publication.

12 | Future Photovoltaics | April 2011 << PREVIOUS PAGE | NEXT PAGE >> www.futurepv.com | 13 FUTURE VISIONS & CURRENT CONCERNS

Driving Into the Future: cells inside from the static and dynamic were looking for reasons that it couldn’t be forces of a rogue overloaded 18-wheeler done, but none of these experts doubted locking up its brakes at 80 mph? After all, that a solar panel could be built to with- Solar Roadways what is the black box on an airplane other stand the abuse of traffic or that a solar than a structurally engineered case to pro- road could become the new power grid. tect sensitive electronics through the Armed with this knowledge, we now Scott Brusaw PRINT E-MAIL this article this article worst of airplane disasters? So I rolled up believed that we could make glass-cov- Solar Roadways my sleeves and got to work. ered solar panels that could be driven This case would have to be transpar- upon. We went back to work brainstorm- ent on the top to allow the sunlight to ing new features of our Solar Roadways reach the solar cells. I looked up the No. 1 system. Living in north Idaho, we get a lot Abstract labeled myself an environmentalist. My and 2 materials labs in the nation. They of snow. To keep the solar cells function- A Solar Roadway is built out of Solar wife had always considered herself one were Penn State University’s Materials ing, we’d have to prevent snow from accu- Road Panels: structurally engineered though. At her request, we saw Al Gore’s Research Institute and the University of mulating on the road surface. That meant cases that can be driven upon and con- movie, “An Inconvenient Truth,” and Dayton’s (UD) Research Institute respec- adding a heating element to the surface: tain electronics including solar cells and began educating ourselves on the climate tively (I received my bachelor’s and mas- something similar to the rear window of a LEDs. The generation of electricity allows crisis debate. We eventually bought the ter’s degrees in electrical engineering car. Melting the snow wasn’t enough, the Solar Roadway (or parking lot, airstrip, movie and book. I don’t know how many from UD and had no idea that they even however: We couldn’t just let the water playground, bike path, etc.) to pay for barrels of oil were required to make the had a materials research lab!). I visited run off the side of the road and refreeze, itself over its life span. DVD or how many trees had to give their both universities and learned quite a bit. lives for the hard copy of the book, but Plastic “yellows” under the sun, which we’ve since bought a Kindle, which will affects its transmissivity (the ability of the Somewhere around six years of age, I save future trees, but caused Borders material to pass light through to the solar received my first slot car track. Remember bookstore to file for Chapter 11. No one cells beneath). I learned that glass falls on the little electric cars that rode on metal said that becoming an environmentalist the hardness scale between steel and rails? The more you squeezed the trigger would be easy. stainless steel, and you can do a whole lot of your hand-held controller, the faster During our studies, we learned that more with it than just make windows that they’d go. I was amazed and started think- renewable energy seemed to be the solu- can be knocked out with a well-placed ing about real roads. If we could only tion to the majority of greenhouse gases baseball.They actually have glass that can make them electric, then kids could drive (50 percent from coal-fired electricity bend like a piece of paper. They also make and relieve our parents of the burden. plants and another 25 percent from our glass that is bulletproof and even bomb My wife and I live in north Idaho, exhaust pipes) that were believed to be resistant. It’s a lot tougher than I had ever where the air is clean, the birds are causing the climate crisis. One day while imagined. I learned that glass can be tex- always singing and life seemed perfect. we were adding rabbit manure to our all- tured to provide traction, so we wouldn’t That is, until Al Gore came along and organic garden, my wife asked if I could have to worry about all the vehicles slid- ruined it for us by introducing phrases make my electric roads out of solar pan- ing off the road the first time it rains. like “global warming” and “climate crisis.” els. I suppressed a chuckle and explained We started meeting with civil engi- At first I didn’t pay much attention to the that solar panels were so fragile that you neers, structural engineers, other electrical problem. I figured somebody was bound couldn’t even stand on them, let along engineers, power transfer engineers, to fix it, right? drive on them. hydrologists, forestry experts, utility com- I’ve always hated litterbugs as much as But I started thinking: What if we could panies and other experts in every field we Figure 1 – Author’s Point of Enlightenment the next guy, but I would have never make a case that would protect the solar could imagine would affect our project. We

<< PREVIOUS PAGE | NEXT PAGE >> 14 | Future Photovoltaics | April 2011 www.futurepv.com | 15 Driving Into the Future: Solar Roadways FUTURE VISIONS & CURRENT CONCERNS

or it would cause heaving and destroy our their destination in one piece. Instead of walks, playgrounds, bike paths, race probably half a dozen other agencies road. A storm water redistribution system painting road lines, what if we placed yel- tracks, amusement parks, airports, cruise would have to not only approve it, but was the answer: Remove the particulates low and white LEDs in the case to light up ship decks, oil tanker decks, etc. – basical- learn to work together to make it happen. through something similar to a French the road lines from beneath? Instead of ly anything with a hard surface that is It began to sink in that this could be the drain and store the storm water in tanks straining to see where the road lines were exposed to the sun can be covered with hardest part of the whole project – the below the frost line. This cool water could at night, it would be like driving on a run- structurally engineered solar panels. actual engineering paled in comparison. be used to warm the surface during the way or in a video game. We learned that So what does one do when one has a A trio who called themselves YERT winter, or cool the surface during the similar experiments in the U.K. reduced great idea and no business or marketing (Your Environmental Road Trip – summer. Or, the water can be pumped to nighttime accidents by 70 percent. experience? Naturally, you start a website yert.com) was in the process of touring all a filtration facility, an agricultural center, Adding LEDs meant we’d have to add a and share your plans with the world. 50 states in 52 weeks to find solutions to an aquifer, a sewer system or just the microprocessor to drive them. Adding a Slowly, we started getting noticed and a global warming. They had read an online nearest body of water. microprocessor makes each Solar Road few websites started writing articles article about us and asked if they could Painting road lines over solar cells Panel a communications device, which about us. People began to email us to stop by while in Idaho for a quick inter- would also present a problem. We live on allows a central control station to recon- either tell us we were brilliant or insane. view. It worked out well for both of us: a long, winding mountainous road. Now figure the road lines instantaneously. We weren’t really surprised: Some of our They got a story and I got our first video to that we were in our late 40s, we were hav- They can reroute traffic, spell out warn- children were beginning to question our post on our website. ing a harder time seeing those road lines ings in the road, create detours, etc. This sanity. Although we were familiar with In 2008, a company called Booz Allen at night. Many people tell us they have not only solved a problem, but inadver- blogs and bloggers, we’d never had their Hamilton sent us an email requesting we the same problem and pray they get to tently increased the “coolness factor” ten- full attention directed at us. It was quite a do a presentation about Solar Roadways fold according to resounding responses learning experience: While the truly edu- at their headquarters in Virginia. We had we received. cated seemed to recognize the potential never heard of them and they sounded My wife requested a system for pre- of such a system, the failed science like an alcohol company, so we didn’t venting vehicle/animal collisions, so back majors were having a field day. The possi- respond. Several days later, my wife to the drawing board I went. Placing load bilities kept growing and features contin- looked up the company. To our horror, we cells in the panels would allow the road to ued to be added, with the bloggers on had been ignoring the oldest consulting know when something was on its surface: dozens of new articles both singing our firm in the nation, which had contracts If a deer walked onto the road, the road praises and telling us we couldn’t possibly with the DOT, DOE, EPA, Homeland would know it. The panels could warn the be serious. I often wondered what the Security and probably half a dozen other oncoming driver (also detected by the Wright Brothers would have done if there agencies that we would need. They had road) of a problem on the road ahead via had been an Internet around 1900 and merged the NFL and the AFL! Before you an LED message in the road itself. I real- they had announced they’d fly a heavier- could say, “Coal can’t be made clean,” we ized this would also allow suspected ter- than-air machine. The bloggers would were on a train headed East. rorists to be tracked by the road in real have eaten them alive. Hopefully they It was our first big presentation, and it time via RFID tags placed on their vehi- would have done what we learned to do: went well. In the crowd was a representa- cles. Hazardous materials shipments develop a thick skin and a good sense of tive from the Federal Highway could be tracked in real time the same humor and press forward. Administration (FHWA). After the presen- way. Autonomous vehicles could finally A bigger problem was looming on the tation, he asked if we’d be in town long become a reality. Homeland Security was horizon: Would the powers that “rule the enough to do our presentation for his going to love this. road” allow us to just go out and start group. We figured if anyone was going to Our wheels were really turning now: slapping our panels onto public roads and throw us out on our ear, it’d be these guys. Figure 2 – Exploded View of the Initial Solar Road Panel Courtesy of Dan Walden Why limit our product to roads? There are highways? We began to realize that the Two days later, we presented the Solar plenty of driveways, parking lots, side- DOT, DOE, EPA, Homeland Security and Roadways to the FHWA. Miracle of mira-

<< PREVIOUS PAGE | NEXT PAGE >> 16 | Future Photovoltaics | April 2011 www.futurepv.com | 17 Driving Into the Future: Solar Roadways FUTURE VISIONS & CURRENT CONCERNS

cles, they were friendly and helpful. What then met with one of our own Idaho con- now shot over 10 hours of Solar Roadways This holds true of every kind of renewable they particularly liked was the thought of gressman, who told us he could earmark video and plans to produce a full-length energy. Solar Roadways could become the a road that could pay for itself through us on a bill in January. He was soundly documentary about the origins of the backbone for all forms of renewable – the generation of electricity. It turns out defeated in November, probably because Solar Roadways project. past, present and future. that road construction materials tend to of excessive earmarking. Since the Solar Roadways double as The FHWA has told us they want us to follow the price of a barrel of oil: Asphalt In 2009, the DOT came out with an the nation’s power grid, we’ll be killing target parking lots first, which makes a lot is petroleum-based, after all. We didn’t SBIR solicitation for a new paving system two birds with one stone (only metaphor- of sense: Parking lots typically have slow- realize it at the time, but our project could that could pay for itself over its life span. ically, fellow environmentalists). Our gov- moving lightweight vehicles. They want not have come at a more perfect time in We learned that the Federal Highway ernment is looking for an intelligent high- us to learn our lessons and perfect our history: Our infrastructure was falling Trust Fund was broke, as were the state way system, a self-healing, decentralized, technology before moving out onto public apart, the price of asphalt and other tradi- DOT budgets. We applied for the contract clean renewable power generation sys- roads. Smart folks. tional materials for building roads had with our Solar Roadways project and were tem, and a smart grid. That’s precisely The Achilles’ heel of electric vehicles skyrocketed, and our country was ready awarded a six-month $100,000 Phase I what Solar Roadways offers. (EVs) is that they have a limited range: to do something about our antiquated contract to develop a crude prototype With a decentralized power generation usually around 100-150 miles. That makes highway system and start thinking about Solar Road Panel. We turned in our final and distribution system, a DC grid makes them great for around-town driving, but a true smart grid. report and a video demonstration of the a lot more sense. Currently, each Solar you’re not currently able to take them for Between the presentations, we met system. YERT visited us once again and Road Panel (or any typical solar panel) a long-distance family vacation, unless with several congressmen and senators at filmed a demonstration of the prototype would need its own micro-inverter to per- your destination is the side of a road 150 their offices across the Potomac. While and placed it on YouTube. This technical form independently and feed into an AC miles from home. Suppose we could per- they were all encouraging and friendly, video is now approaching 1 million views: grid. By generating power near the end suade one national fast-food chain to none of them threw large sums of cash at http://www.youtube.com/watch?v=Ep4L1 user, AC power may no longer be needed. green their image by retrofitting their us as we had hoped. “It’s an election year,” 8zOEYI. I’m still amazed – we now have a Most of the devices we plug into our wall parking lots nationwide with Solar Road they said, which was apparently the worldwide fan base and can’t begin to outlets immediately convert the AC to DC Panels. Let’s use McDonald’s for our wrong time to be looking for funding. We keep up with our email inbox. YERT has to perform the work. If we created a DC example. If all I have to do is find the standard for household use, then manu- golden arches to recharge my EV, then I facturers of appliances could remove the can drive it from California to New York. I AC-to-DC conversion circuitry, which challenge anyone to find a 100-mile would cut down on costs and conversion stretch of American highway without a losses. We’d also eliminate the need for set of golden arches. DC-to-AC inverters. Thomas Edison would With this capability, EVs could finally be ecstatic. become practical. Drivers could start trad- Since Solar Roadways are, by defini- ing in their gas guzzlers for EVs. Our tion, already in the right-of-way, no spe- dependency on oil would begin to sub- cial land need be set aside for solar farms. side. Taco Bell, KFC and Burger King would The roads are already there – why not put see the writing on the wall: McDonald’s them to use? One of the problems facing would be getting all of the business from wind farms is the challenge of getting the the increasing number of EV owners. wind power to the grid. If the nearest road More and more businesses would now serves as the grid, then all you’d need do be using solar rather than coal. is “pave” the road leading up to each Environmentalists would be dancing in Figure 3 – Artist’s Rendition of a Solar Roadway Courtesy of Dan Walden windmill with Solar Road Panels. Can you the streets, along with all of the solar cell imagine solar solving the wind problem? manufacturers.

<< PREVIOUS PAGE | NEXT PAGE >> 18 | Future Photovoltaics | April 2011 www.futurepv.com | 19 Driving Into the Future: Solar Roadways

There are over 28,000 square miles of percent? This same Solar Road Panel Kerosene Parity asphalt and concrete surfaces exposed to would then be capable of producing four the sun in the U.S. alone. Covering these times more electricity than it did previous- surfaces with 12’ by 12’ Solar Road Panels ly. This allows the system to keep up with T. Patrick Walsh PRINT E-MAIL this article this article would require roughly 5 billion panels. the increased demand in electricity over Greenlight Planet Inc. Even at today’s standard of 15 percent the years and it also provides job security efficient solar cells, that area could pro- for many of you reading this article! duce three times more electricity than It is said that necessity is the mother of we’ve ever used as a nation (for a break- invention. Not just a 6-year-old’s desire to down of the numbers, see our numbers drive a car, but rather the need to change The quest to replace fossil-fueled elec- our prior life experiences are often a world page: http://www.SolarRoadways.com/ the way we currently think and do things. trical grids with solar energy has been the apart from villagers’. But the effort to numbers.shtml). That’s almost enough to Our world is waiting for us to make some underlying motivator for most progress in understand and address this market is power the entire world. If the Solar big and drastic changes. Lobbyists and blog- photovoltaics – the focus of researchers, worth it: Designing products, businesses Roadways provided power during daylight gers will continue to try to tell us otherwise, entrepreneurs and policymakers. Human- and policies to get photovoltaics into the hours, and (assuming that no clever indi- but we all know that solar is the only real ity’s desperate need for a renewable alter- hands of underserved rural consumers is vidual is ever able to create a practical solution to our future energy needs. native makes grid parity both a noble and good, it’s profitable and it’s downright fun. energy storage system) wind, hydro and Note: The FHWA invited us back to lucrative goal. But the most desperate As a kid growing up in Chicago, I nuclear can take up the slack at night, apply for Phase II funding (two-year quarter of humanity, namely the 1.6 billion thought I was afraid of the dark. But in then we could finally be done with coal. $750,000 contract to build the first Solar rural people living without electricity in truth, the ever-present glow of a city Manufacturing 5 billion Solar Road Road Panel parking lot). We submitted our developing countries, pay far higher energy meant that true darkness was something Panels would do to unemployment what a application in January 2011. prices than do grid-connected people. For I had barely experienced. It wasn’t until good solar array can do to a utility meter: In 2010, Solar Roadways won the prize developing-world villagers, existing photo- driving up a dirt road into an Indian vil- We could spin it backward and produce for the most votes in GE’s Ecomagination voltaic technology is already the most eco- lage in the summer of 2005, my sopho- more jobs than we could fill. It’s not just Challenge. As of this writing, Solar nomical energy source for myriad needs. more year of college, that I experienced the final assembly, installation and main- Roadways is the most supported entrant Understanding, let alone designing for, life beyond the reach of power lines. I had tenance, but every component that goes in round 2 of the Ecomagination these needs is difficult for most of us in the set out with a handful of friends from col- into the panels: the manufacturers of Challenge, which is focused on products photovoltaics community, if only because lege as a newly formed chapter of solar cells, LEDs, microprocessors, glass, for the homeowner. etc., will have to ramp up. We could manufacture our way out of this economic About the Author slump and become an exporter of energy for the first time in decades. Scott Brusaw is president & CEO of Solar At the end of their anticipated 20-year Roadways Inc. He is an electrical engineer life span, the Solar Road Panels would be (MSEE) with over 20 years of industry expe- refurbished rather than discarded. Most of rience. Brusaw has multiple patents, and his the parts should be reusable. Certain ele- hardware and software have been sold inter- ments, including the solar cells, would be nationally. updated to the latest and greatest. For instance, if we go into production tomor- Click here to return to Table of Contents row using 15 percent efficient solar cells, then 20 years from now we’d replace them PRINT E-MAIL Woman Cooking With Woman Cooking With Dim Dark Kerosene Lamps the Bright Light of the Sun King with the off-the-shelf norm of the day – 60 this article this article

<< PREVIOUS PAGE | NEXT PAGE >> 20 | Future Photovoltaics | April 2011 www.futurepv.com | 21 Kerosene Parity FUTURE VISIONS & CURRENT CONCERNS

Engineers Without Borders, planning to dized. (African villagers often spend twice light than the kerosene lamp, all at an off- major source of indoor air pollution. They spend the summer installing an experi- or three times as much on the open mar- the-shelf cost of around $15. By purchasing significantly reduce global greenhouse gas mental bio-diesel generator in Badaka- ket.) For people who earn less than $2 per such a solar lantern, most villagers could emissions from lighting. They save villagers mandara, population 800. day, this is a huge cost for a few hours of earn back their investment through money. They allow men and women to pur- That first evening in town, swarmed by dim, smoky lighting: The brightness is reduced kerosene costs within half a year, sue cottage-industry work at night, and chil- children who were as fascinated by my equivalent to roughly 5 percent of a 60- and reduce their overall lighting expendi- dren to study, all without worrying about the digital camera as I was in taking their pho- watt incandescent bulb. And since light ture by 80 percent over a three-year prod- cost and fire hazard of kerosene. And per- tos, the setting sun prompted me to walk from a flame is hard to direct downward, uct lifetime. haps most importantly to villagers, the new inside the one-room building where I most goes straight up, to be wasted on the From my view in Badakamandara, it lamps are cool: They are a means to, and would sleep for the next two months, and hut’s dark ceiling. seemed like a foregone conclusion that this also a symbol of, a higher standard of living. reach for the light switch. But as I groped Starting around 2005, photovoltaic and convergence would kill the kerosene lamp As photovoltaics becomes more efficient along the concrete wall, my fingers found light-emitting diode (LED) technology con- within a decade. I decided to set aside the and the power requirements of other elec- nothing. Being in town specifically to verged to the point where an amazing new centralized plans and philanthropy of our tronic technologies fall, new opportunities install an electrical generator, it should combination became possible: With a tiny, NGO and start a sustainable, for-profit busi- come into play every year. In 2005, to find a have been no surprise that there was no half-watt solar module, a small battery ness to commercialize the technology. So pay phone, I had to drive half an hour from light switch. But the sudden realization and an efficient LED, one can now assem- far, Greenlight Planet Inc. has lit up half a Badakamandara. By last year, in a startling that it was getting dark, and I was com- ble a solar LED lantern that charges during million lives with our clean, safe, affordable shift seen around the developing world, I pletely without power for the foreseeable the day, and at night yields more usable Sun King™ solar lanterns, designed with vil- watched villagers tend to cattle, eyes all the months, inspired an unexpectedly visceral lage needs in mind. The lamps eliminate a while glued to text messages on inexpensive shock: It was a mild but real sense of des- peration that I will not forget. As night fell, villagers lit kerosene lanterns; first a handful, and soon dozens. I had only seen these flickering, smoky relics before in movies. But it immediately became apparent how essential the kerosene lamp was, however dim and inefficient. As a month passed and our complicat- ed, expensive generator installation approached completion, the simplicity and ubiquity of the kerosene lamps stuck in my mind. Despite being woefully outdated technology, it remained the de facto standard for one reason: In the absence of gov- ernment-organized utilities, it was the light source that villagers could purchase individually, to solve their personal, immediate need for light, without relying on any network. A typical Indian village family spends Brothers Study in Kenya Family Living With Sun King $2 per month on kerosene, which is subsi-

<< PREVIOUS PAGE | NEXT PAGE >> 22 | Future Photovoltaics | April 2011 www.futurepv.com | 23 Future Photovoltaics Kerosene Parity NEW TECHNOLOGIES & MATERIALS PRINT E-MAIL Click here to return to Table of Contents this article this article mobile phones. But with mobile-phone pen- For photovoltaics researchers who may etration soaring, power lines are still a dis- be impatient for grid parity, the knowledge tant dream, leaving villagers to pay 15 to 25 that millions of kids are now studying by cents at a local shop every time they need the light of solar lanterns (while saving Jef Poortmans to charge their phone: This charging often their lungs, and their family’s precious Department Director, Solar and Organic Technologies, imec costs more than the phone service. So after income to spend on something other than countless requests, Greenlight Planet’s fossil fuel) should be an enormous source newest product includes adapters for charg- of pride: Decades of research have already ing a wide variety of these low-cost phones. pushed photovoltaics well past kerosene Designing appropriate new photovoltaic parity, which was the key metric for the applications for these rural markets is now people in most desperate need of an ener- Reducing PV production costs ductor industry. Dr. Ravi especially goes into a principal challenge. But in terms of pho- gy alternative. Although this massive con- with thinner cells the efforts of the industry to reduce the tovoltaics themselves, much important sumer market is not as obvious or as The PV industry is rapidly maturing. We thickness of absorber layers, lowering the innovation remains to optimize the under- accessible to Western technologists, are on a fairly long and stable path to lower production costs by using less material, the lying technology for the village. At the investors and policymakers as markets the costs of energy generated with solar material cost being an important part of the semiconductor level, thin films will take closer to our experience, the corporate modules. It is a path that is dictated by the total costs. This can be done by moving to over this market in the future if they can strategist C.K. Prahalad had it right when economic realities. On the one hand, we have materials with higher absorption coefficient achieve even modest efficiency gains: For he predicted a fortune waiting at the “bot- to produce solar cells cost-effectively. That (a-Si:H, CIGS, CdTe, etc.). But also for crys- small modules, the benefit of doing away tom of the pyramid.” And getting there is inevitably leads to larger-scale production talline Si solar cells, thinner active layers can with manual soldering of individual crys- as fun as it is profitable. and industry consolidation. On the other yield higher efficiencies, through the use of talline cells is enormous, but above about 1 hand, there is still some way to go before PV proper optical designs. Dr. Ravi makes the watt, the low efficiency of thin films means About the Author technologies reach grid parity. So we have to point that the current cell thickness is large- the modules are too bulky to efficiently keep improving the technology, and look for ly imposed by the difficulty of handling much package and ship with a low-cost con- T. Patrick Walsh is founder and chief tech- advanced and novel concepts that can per- thinner wafers. He then discusses the tech- sumer product. At the module level, due to nology officer of Greenlight Planet Inc. He manently close the gap with grid parity. nology that is being developed at Crystal the small size of the modules, the cost and founded the company while in college at the In the article in this section, K.V. Ravi from Solar – fabricating 50 µm thin wafers through design of encapsulation technology is criti- University of Illinois, where he studied eng- Crystal Solar, Inc. discusses two related top- epitaxial deposition. Of course, the handling cally important: Solar cells themselves ineering physics and economics, graduating ics. The more general topic is scaling in PV issues still apply, so Crystal Solar is also account for a large fraction of the cost of a in 2007. The company distributes a direct technologies, which is quite different from working on new approaches for handling, large solar module, but the semiconductor replacement for kerosene lamps: affordable how scaling is understood in the semicon- processing and packaging. accounts for less than half the cost of the solar-powered LED lanterns that villagers finished module in the case of a 1-watt buy to light up their own homes, with panel, using traditional assembly and no philanthropy or public infrastructure encapsulation processes. In the realm of required. Patrick has become a recognized solar concentrators, a very small-scale con- mass-affordability product designer, gar- centrator could be a major boon for the vil- nering notable awards from Lemelson-MIT, lage but would require redesigning with an UNESCO and the Lighting Africa program. entirely different cost/benefit analysis in mind relative to traditional large-scale con- Click here to return to Table of Contents centrators, considering, for example, how to minimize setup and maintenance com- PRINT E-MAIL plications for millions of end consumers. this article this article

24 | Future Photovoltaics | April 2011 << PREVIOUS PAGE | NEXT PAGE >> www.futurepv.com | 25 NEW TECHNOLOGIES & MATERIALS

Thin Is In, tor. Silicon, an indirect band gap semicon- films on rigid substrates such as glass or ductor, has a low absorption coefficient metal sheets. for light, necessitating the use of fairly With the current state of photovoltaic But Not Too Thin! thick wafers (~180 microns today) to technology, there is a trade-off between absorb a large portion of the sun’s spec- reduced materials usage with thin film trum that silicon is sensitive to and, more compound semiconductor (and amor- K.V. Ravi PRINT E-MAIL importantly, the difficulty of handling and phous silicon) photovoltaics and the rel- this article this article Crystal Solar, Inc. processing silicon wafers of thicknesses atively low energy conversion efficiency much below current ~180 micron thick- of these products. In contrast silicon ness. A response to the need to use thick wafer based devices, although consum- silicon wafers has been the development ing too much material are the undisput- of photovoltaic devices using direct band ed leaders in achieved high energy con- gap semiconductors such as amorphous version efficiencies. Abstract improve the economics of manufacturing. silicon and various compound semicon- Recently there has been a great deal of The trade-off between thick (~170 Typically there have been wafer size (diam- ductors with cadmium telluride (CdTe) interest in CdTe and CIGS based photo- microns) silicon-based PV and thin (a few eter) increases roughly every 10 years for and copper indium gallium selenide voltaics with CdTe being the current leader microns) film non-silicon and amorphous about the last ~20 to 30 years, with the cur- (CIGS) being the current favorites. These in volume manufacture. However several silicon PV is addressed by the develop- rent leading-edge factories being based on materials have superior absorption coeffi- issues confront these materials if truly ment of single crystal silicon wafers of 300 mm diameter wafers. cients, enabling the use of very thin films large scale deployment is contemplated. thicknesses of ~50 microns produced by Scaling in photovoltaics has taken a dif- of these materials and manufacturing This includes toxicity of these materials epitaxy. This approach has the cost ferent path. Clearly the analogy with semi- technologies based on depositing thin although many companies claim to have advantages of thin film technologies and conductor critical dimension scaling does the efficiency, reliability and non-toxicity not apply to photovoltaics where the 30 of earth-abundant silicon PV. device dimensions are equal to the wafer dimensions. In crystalline silicon photo- 29.5 voltaics, wafer size scaling is also not a 29 Semiconductor technologies, of which very important factor, since very large photovoltaics is an increasingly large part, wafers would lead to extremely unwieldy 28.5 have had an obsession with dimensions devices with large currents and low volt- 28 and scaling over their history for increasing ages limited by the band gap of silicon. product functionality and reducing manu- Consequently, manufacturing cost reduc- 27.5 facturing costs. The most familiar and the tions in photovoltaics are not dominated 27 most impactful is the ubiquitous scaling of by wafer size changes. Efficiency - % Efficiency 26.5 gate dimensions in integrated circuits as The primary scaling factor in photo- described and predicted by Moore’s Law. voltaics has been the thickness of the 26 This remarkable scaling of the critical semiconductor or the absorber driven by 25.5 dimensions in transistors in integrated cir- manufacturing cost reasons, since mate- 1 2 3 5 13 20 50 100 200 500 1000 cuits has been the driving force for the rials constitute the major portion of man- information, communication and enter- ufacturing costs. There are, generally, two Wafer thickness - microns tainment industries. An additional impor- approaches to thickness scaling. The tant scaling in semiconductor manufacture more fundamental one is based on the Figure 1 – Efficiency as a Function of Wafer Thickness has been silicon wafer size changes to absorption coefficient of the semiconduc-

<< PREVIOUS PAGE | NEXT PAGE >> 26 | Future Photovoltaics | April 2011 www.futurepv.com | 27 Thin Is In, But Not Too Thin! NEW TECHNOLOGIES & MATERIALS

this under control. Another much more thicker than needed for achieving high thin silicon wafers are basically impossi- absorber (wafer) thickness that is, as difficult issue with these materials is their energy conversion with the thickness being ble tasks. The technology limitation of Goldilocks would say it, it is neither too availability.As Martin Green states, “In suc- driven by the difficulty of handling and current wafer-based technology from the thick nor too thin, but just right! This cess CdTe and CIGS technologies will ulti- processing wafers of thicknesses much perspective of wafer thickness can be technology is based on fabricating single mately guarantee their eventual failure. below about 160 µm. Figure 1[2] shows that seen in Figure 2. crystal wafers by depositing silicon from This will be by pushing Te and In prices the minimum wafer thickness required to In loose analogy with the continuing the vapor phase on to appropriately pre- beyond the threshold for profitability, as achieve the theoretical maximum efficien- reduction in the critical dimensions for pared substrates using an epitaxial depo- recently with polysilicon prices.”[1] cy from silicon solar cells is ~40 to 50 integrated circuits, the continuing reduc- sition process. The positioning of this In view of the above, although “thin is microns, not the current ~180 microns. tion in wafer thickness has been called the technology is depicted in Figure 3. in” a case can be made for not being too However, in addition to the basic prob- Moore’s Law for PV (Figure 2). In further This approach enables the fabrication, thin! This case is based on the fact that sil- lem of slicing silicon ingots into wafers of analogy with Moore’s Law for ICs, where handling, processing and packaging of icon wafer thicknesses in today’s high-vol- these thicknesses and the inevitability of continuing reduction in the CD is slowing very thin (< 50 microns thick) single crys- ume manufacturing technology are much kerf loss, handling and processing such down due to fundamental limitations, tal silicon wafers and solar cells. This will wafer thickness reductions for PV are also substantially reduce the amount of sili- slowing down, due to fundamental limita- con utilized with a major impact on the 450 tions. Both technologies need radical overall materials usage in PV manufac- IC - CD - nm departures from conventional scaling. turing. When this technology is transi- 400 Crystal Solar is developing such a rad- tioned into manufacturing, we project Solar wafer thickness ical departure from conventional practice direct manufacturing costs well under - microns 350 for manufacturing solar cells with an $1/Wp to be achievable with high-effi-

300 Interdigitated back-contact cells Current wafer-based Limit of sawing 250 PV technology α-Si_c-Si heterojunction devices

200 20 Conventional Crystal Solar mono c-Si 150 Crystal 15 Multi-c-Si solar technology CIGS/CdTe 100 Critical Dimension for ICs - nm Solar wafer thickness - µm ITRS (%) Efficiency 10 50 5 α-Si/micromorph a-Si thin thin film 0 Films 10 20 40 80 160

1980 1990 1996 1997 1999 2001 2003 2007 2008 2009 2010 2012 2014 Cell Thickness (µm)

Figure 2 – Moore’s Law for PV? Figure 3 – Cell Efficiency for Various Thickness Semiconductors

<< PREVIOUS PAGE | NEXT PAGE >> 28 | Future Photovoltaics | April 2011 www.futurepv.com | 29 Thin Is In, But Not Too Thin! NEW TECHNOLOGIES & MATERIALS

ciency PV modules with a direct impact Although high-quality, very thin sili- At Crystal Solar, such processes have on lowering systems costs. This technol- con wafers can be produced by this been developed for fabricating very thin ogy will have the disruptive potential to approach, novel and innovative approach- solar cells and packaging them for the dramatically reduce manufacturing costs es have to be developed for handling, completed PV module. Figure 5 shows an as follows: processing and packaging. example of mini modules ~1 ft. X 1 ft. • Silicon utilization is reduced to about using ~50 micron thick solar cells. 20 percent of that utilized in current When transitioned into volume manu- wafer-based technology – ~300 µm facturing, this technology is expected to (slice + kerf) for current technology enable the lowest manufacturing costs of versus < 50 µm using epitaxial tech- all PV technologies with the cost advan- nology. tages of thin film technologies and the Don’t Miss Out • The traditional supply chain – polysili- efficiency, reliability and non-toxicity of con production in Siemens reactors, earth-abundant silicon PV. Sign up NOW for our crystal growth or casting, ingot crop- FREE biweekly newsletter ping, squaring and wafering – are References eliminated by the direct gas (tri- 1. “Price and Supply Constraints on Te chlorosilane)-to-wafer process involv- and In Photovoltaics,” Martin A. Green, ing epitaxy (Figure 4). This reduces IEEE PV SEC, 2011 process complexity and is substantial- 2. “Limiting efficiency of Silicon Solar Figure 5 – ~1 ft. X 1 ft. Mini Modules With Square, ly more capital-efficient as compared ~50 Micron Thick Solar Cells Produced by Epitaxy Cells,” Tiedle et al. IEEE Trans. Electron to traditional technology. Devices, vol. ED-31, No. 5, May 1984.

About the Author A roundup of the industry’s top R&D stories in one place K.V. Ravi is the chief technology officer of Crystal Solar, where he is responsible for the Be notified when there is a new development of technology for the manufac- Cell Production/Manufacture Future PV blog post Poly- Ingots Wafering Cell Module turing of low-cost silicon wafers, solar cells silicon Saw Emitter Passivation/ Metalization Edge Inspect/ interconnect/ damage/ formation AR coating isolation Test Laminate and modules. His other affiliations have Be notified about our partner Texture etch been Applied Materials, Intel, Motorola, developments and partner events Texas Instruments and Mobil Solar Energy Crystal solar Modified Corporation. He has a Ph.D. in materials module technology Modified cell processing for thin wafers science from Case Western University. Thin wafer (< 50 µm) processing for thin wafers Sign up NOW Click here to return to Table of Contents Materials Production Materials Processing Packaging click here PRINT E-MAIL this article this article Figure 4 – The front end of the current PV supply chain involving the production of polysilicon from trichlorosilane (TCS), the growth of ingots and the machining and wafering of ingots are eliminated with the Crystal Solar direct gas-(TCS)-to-wafer process.

<< PREVIOUS PAGE | NEXT PAGE >> 30 | Future Photovoltaics | April 2011 Future Photovoltaics SILICON WAFER PHOTOVOLTAICS PRINT E-MAIL Click here to return to Table of Contents this article this article N-type Bifacial Solar Cells

Jörg W. Müller for Industrial Application Director, R&D Cells; Q-Cells SE

V.D. Mihailetchi,1 A. Edler,1 R. Harney,1 R. Kopecek,1 PRINT E-MAIL T.S. Böscke,2 D. Stichtenoth,2 T.Aichele,2 J. Lossen2 this article this article 1International Solar Energy Research Center - ISC Konstanz 2Bosch Solar Energy AG

Photovoltaics is one of the fastest-grow- industrial process for n-type silicon similar to ing electricity generation technologies in the today’s p-type screen-printed standard world. Average annual growth rates of global process without sacrificing the efficiency PV installations have been around 45 percent potential of n-type wafers. Increasing the conversion efficiency of nique of screen printing is only now start- for the last 15 years, which, in combination Researchers from the International Solar solar cells is a measure to reduce the cost ing,[2] because processing of n-type with typical learning rates of 20 percent, lead Energy Research Center in Konstanz are of energy generation of a photovoltaic wafers comprised mainly three unre- to fast and ongoing cost reductions in the working in close cooperation with Bosch system. Currently, worldwide production solved issues: cost-effective emitter for- industry. Solar Energy on such a simple industrial of solar cells is more than 80 percent mation, passivation of p+ emitters and Today the photovoltaics marketplace is process for n-type solar cells. They demon- based on crystalline silicon, from which contacting of p+ emitters. experiencing rapid decline in average selling strate very high efficiencies exceeding 19 more than 90 percent are produced on p- The potential for higher cell efficien- prices of solar cells and modules. This price percent on large-area wafers using a boron type doped wafers. However, the highest cies on n-type base material led many decline is a mandatory prerequisite to mak- front emitter, a phosphorus back-surface conversion efficiencies of commercially groups, including ours, to develop a cost- ing PV competitive with fossil and nuclear field and conventional screen printing. Apart available crystalline silicon solar cells and effective process to fabricate n-type cells power in terms of retail or even wholesale from the achieved cell efficiencies and the modules are achieved by the company for a subsequent industrial implementa- electricity prices. Manufacturers are thus simple process sequence, the exciting fea- sunpower on n-type doped wafers. tion. We review here our progress and being challenged to dramatically reduce ture of this cell structure is its outstanding Sunpower uses an advanced cell concept achievements of an n-type cell concept operational costs while at the same time bifacial capability. The measured quantum called interdigitated back contact, achiev- developed by ISC Konstanz e.V. and Bosch increase cell efficiency and quality in order to efficiencies under front and rear illumina- ing cell efficiencies of up to 24.2 percent. Solar Energy AG. Our cell process features remain competitive. tion are almost equal. This allows fabrication But additionally for screen-printed a homogenously diffused boron emitter, One way to increase cell efficiencies is the of bifacial modules with significantly solar cells, the n-type approach holds sig- which is topped by a stack of a passivat- use of n-type silicon, which is known to yield increased energy yield compared to standard nificant benefits over advanced p-type ing layer and microwave plasma- higher-quality wafers compared to p-doped modules of equal conversion efficiency. approaches and might therefore become a enhanced chemical vapor deposition wafers commonly used in the PV industry. Depending on the mounting conditions of cost-effective alternative to the predomi- (PECVD) of hydrogenated silicon nitride SunPower and Sanyo, for example, manufac- bifacial modules, the additional energy yield nant p-type solar cell process with alloyed (SiNx) anti-reflection coating layer. The ture solar cells with efficiencies exceeding 20 generated by illumination from the rear side aluminium back surface field (BSF).Typical rear side of the cell comprises a phospho- percent using n-type silicon. Both companies could be almost as high as the contribution advantages of n-type material have rous-diffused BSF subsequently passivat- are employing a cell process different from of the front side. These new n-type bifacial already been comprehensively reviewed, ed by a SiNx layer. The cell is contacted by the industrial standard, e.g., interdigitated solar cells have the potential to reduce the such as the higher tolerance to common standard screen printing using an h-grid back contact cells and heterojunction with levelized cost of electricity drastically, paving metal impurities[1] and the lack of a pattern on both the front and the back intrinsic thin layer, respectively. the road to PV grid parity and fuel parity, boron-oxygen-complex-related degrada- side. A schematic cross section of the cell A more favorable path toward high cell consequently making PV a major energy tion. However, the mass production of n- is shown in Figure 1. In contrast to a cell efficiencies would be a simple and low-cost resource on a global scale. type solar cells using the low-cost tech- with an Al-alloyed BSF, this cell structure

<< PREVIOUS PAGE | NEXT PAGE >> 32 | Future Photovoltaics | April 2011 www.futurepv.com | 33 N-Type Bifacial Solar Cells For Industrial Application SILICON WAFER PHOTOVOLTAICS

allows a bifacial usage with a high bifa- BBr3 is introduced into the tube at elevat- process (drive-in). At this stage of the dif- mized boron diffusion not only to obtain ciality factor. In appropriate applications, ed temperature and mixed with oxygen, fusion process, a very high concentration good sheet resistance, homogeneity and the bifacial character of this cell concept which causes it to oxidize and form boron of boron can build up at the glass-silicon high throughput, but to maintain high yields higher energy generation as com- oxide and bromine following the reaction: interface, resulting in a boron-rich layer bulk lifetime after the diffusion process. pared to a monofacial cell concept of (BRL) formation, identified as silicon To test the bulk stability, we diffused equal front-side conversion efficiency. → (1) boride[3]: boron at various temperatures on n-type 4BBr3 + 3O2 2B2O3 + 6Br2. Cz wafers and subsequently etched back Boron Diffusion and Since the vapor pressure of B O is → (3) the diffused layers. A passivation layer 2 3 Si + 6B SiB6. Bulk Lifetime Stability rather low at typical diffusion tempera- consisting of silicon nitride (SiNx)was Boron is the commonly used dopant to tures (only 10-8 atmospheres at 920°C), The BRL is insoluble in hydrofluoric then applied on both surfaces, followed form the emitter in n-type p-n junction liquid B2O3 condenses on the silicon acid and it has been suspected to be by a firing step to activate the surface solar cells, especially in high-efficiency wafer surface and on the furnace walls. responsible for bulk lifetime degradation passivation. The bulk lifetime was then cell concepts. Many diffusion sources for The chemical reaction occurring at the if its thickness grows more than 5-10 measured by mapping the surface using diffusion of boron into silicon exist. There silicon surface reduces the B2O3 to ele- nm.[4] It is known that the growth of the the microwave-photoconductance decay are gas phase diffusion sources, i.e., boron mental boron as follows: BRL can be suppressed if the oxygen con- (µW-PCD) method. Figure 2 shows the tribromide (BBr3), boron trichloride (BCl3) centration in the atmosphere increases, bulk lifetime and the sheet resistance of and boron diborane (B H ), but also → (2) according to the reaction: boron emitters using our optimized 2 6 2B2O3 + 3Si 4B + 3SiO2, boron-doped oxide as a solid or liquid recipe. The initial lifetime measurement phase diffusion source. The most fre- Where SiO partially dissolves in liquid SiB + 11 O → SiO + 3B O . (4) was performed on etched and passivat- 2 6 2 2 2 2 3 quently used method to bring boron into B2O3 to form the mixed phase B2O3-SiO2 ed neighboring wafers before the boron silicon is diffusion from a BBr3 source in system (borosilicate glass, BSG). The ele- For an industrial process, however, diffusion process, and serves as a refer- an open-tube furnace. In the first stage of mental boron diffuses into silicon mostly where wafers are closely packed in a rela- ence. No degradation in bulk lifetime is the diffusion process (deposition), the during the second stage of the diffusion tively long boat, homogeneous diffusion observed using our optimized boron dif- over the entire batch in a relatively short fusion that covers a sheet resistance process time is desirable. This implies, at range of 45 to 260 Ω/. This is an indica- the first stage of the diffusion, a very good tion that boron diffusion, if the chemical front contact control of BBr oxidation in the tube (reac- reactions (1-4) that govern the diffusion passivating layer/SiNx 3 tion 1) in order to slowly deposit B2O3 uni- process are properly controlled, is not a formly on all wafers. The formation of liq- detrimental step in the solar cell uid B2O3 during BBr3 step and the buildup process. of the BRL at the surface are the key + n-type Si (p ) Emitter parameters to control in boron diffusion. Solar Cells Results and Outlook Therefore, homogeneous boron diffusion The emitter formation and its passiva- (n+) BSF (from a gas source) at PV industrial tion are the most crucial steps in fabricat- requirements is more difficult to realize ing high-efficiency solar cells. We applied than the more commonly used phospho- our knowledge and development on rous diffusion in the fabrication of p-type boron diffusion to fabricate n-type cells

SiNx solar cells. having the same structure that is back contact In our lab, we put considerable effort schematically depicted in Figure 1. The into optimizing the boron diffusion for front surface has a random pyramid tex- Figure 1 – Schematic Cross Section Of Our N-Type Solar Cells n-type solar cells in a quartz tube fur- ture formed in KOH/IPA solution, while nace of an industrial scale. We opti- the rear side is chemically polished to

<< PREVIOUS PAGE | NEXT PAGE >> 34 | Future Photovoltaics | April 2011 www.futurepv.com | 35 N-Type Bifacial Solar Cells For Industrial Application SILICON WAFER PHOTOVOLTAICS

enhance the internal rear reflection. The dard processes used in the industry for the potential to surpass the standard p- standing passivation of the diffused sur- n+ BSF is formed on the rear side by dif- the fabrication of p-type Al-BSF solar cells. type-based concepts. faces and on the other hand the very high fusing phosphorous from a phosphorus No customized equipment is required to Moreover, this cell concept is bifacial, as lifetime of the material, which obviously oxychloride (POCl3) source in a quartz apply these steps in the industrial fabrica- it has an open-back side grid and passiva- does not suffer substantially during the tube furnace. The BSF is passivated then tion of n-type cells. The only exception is tion layer, allowing it, in appropriate appli- process. by a standard SiNx layer. On the front for the boron diffusion, which requires a cations, to generate more power than a con- A symmetrical IQE from both sides is an boron emitter, we have applied our own dedicated diffusion furnace system. ventional monofacial cell under the same excellent characteristic because it theoreti- developed passivation method consisting Therefore, our work emphasizes the conditions. Therefore, the performance of cally allows fabricating bifacial cells with of a thin passivating layer and a SiNx stack. understanding and development of boron these n-type solar cells, when illuminated the same power conversion efficiencies for The metalization was applied on both diffusion for industrial applicability. from the rear side, is very important. front- and back-side illumination, provid- sides of the cells by screen printing metal Typical cell results we obtain with our All our solar cells processed so far have ing that the back surface is also textured. pastes. The cells went through a firing n-type process are summarized in Table 1. a flat back side, since they were optimized The major difference between front- and step to complete the contact formation. We show average values over more than to have maximum efficiency only from the back-side performance of the cells is in the It is important to note that in our fabri- 100 cells, along with the results for the front side. This results in different short-cir- Jsc. Therefore, work is under way to adapt cation process of n-type solar cells, the best cell. The efficiency averaged at 19.1 cuit current densities (Jsc) under front- and the cell process to allow for both sides to be + formation of the n BSF layer, its passiva- percent, and the best cell was found to back-side illumination. However, this effect textured in order to match the Jsc. tion and metalization, chemical cleaning, have 19.4 percent. This result underlines can be eliminated by comparing the inter- etching or texturization steps are all stan- that the demonstrated cell concept has nal quantum efficiencies (IQE) measured under front- and back-side illumination conditions. Figure 3 shows the IQE measurement of our best cell (black circles) illuminated from the front side. The response observed at short wavelengths (below 0.5 µm) confirms the excellent passivation of the boron emitter (60 Ω/). Moreover, Figure 3 also shows the IQE data of an optimized bifacial cell, which has different BSF sheet resistance and metalization grid (but still a flat back side) as compared with a cell optimized only for the front side. The IQE results of the bifacial cell show that, independent from which side the cell is illumi- Figure 3 – Internal Quantum Efficiency (IQE) of nated, the response of the cell is the same. our Best Fabricated Cell, and of a Bifacial Cell Illuminated From the Front and Back Side This demonstrates on one hand the out-

2 2 η Area [cm ] Jsc [mA/cm ] Voc [mV] FF [%] [%] Best cell 241 38.8 643 77.6 19.4 Average, 100 cells 241 38.3 642 77.5 19.1 Figure 2 – Mean bulk lifetime measured on 156 mm wafers using µW-PCD on 156 mm n-type Cz wafers before and after boron diffusion at various temperatures. The corresponding sheet esis- tance and standard deviation of the diffused layer are also shown. Table 1 – Best Solar Cell and the Average of 100 Cells

<< PREVIOUS PAGE | NEXT PAGE >> 36 | Future Photovoltaics | April 2011 www.futurepv.com | 37 Future Photovoltaics N-Type Bifacial Solar Cells For Industrial Application THIN FILM PHOTOVOLTAICS PRINT E-MAIL Click here to return to Table of Contents this article this article It could be demonstrated that the About the Authors combination of proven mass production technologies and the electrical superiori- ty of n-type material holds great potential Valentin D. Mihailetchi, Ph.D. – Researcher for cost-effective manufacturing of high- and Coordinator of Advanced Solar Cell Danielle Merfeld efficiency cells in high volume. With the Concepts and Devices; ISC Konstanz Director, Solar Technology Platform; GE’s Global Research Center challenge of emitter formation, its passivation and metalization presently being Alexander Edler – Ph.D. student in the solved, there is no reason for n-type sili- Advanced Solar Cell Concepts and Devices con solar cells to remain a niche product. group; ISC Konstanz Thin film PV continues to capture the atten- and novel ideas about what could enable Conclusions Rudolf Harney – Industrial Solar Cells tion of the industry with the promise of a them to compete effectively in the race We have presented a simple industrial Department Head; ISC Konstanz departure from the “s-curve” described by the toward our solar future. A team from process to fabricate n-type solar cells with current c-Si PV technology era of solar. Today HelioVolt Corporation describes an interest- boron front emitter and phosphorous BSF Radovan Kopecek, Ph.D. – Researcher, the range of thin film PV materials is wide and ing approach to depositing CIGS, which that leads to highest efficiency of 19.4 Member of the Executive Committee, the approaches varied; however, the mantra for reduces the thermal budget – and associated percent on large-area (241 cm2) mono- Managing Director of Advanced Cell success is the same: Use technology to drive costs – typically required with this material crystalline Cz-Si substrates and an aver- Concepts and Co-Founder; ISC Konstanz high efficiency at low cost. Some approaches system. Torsten Brammer, a photovoltaics age efficiency of 19.1 percent. We demon- are of particular interest based on the opportu- research consultant, offers an enlightening strated that this cell concept has an out- Tim S. Böscke, Ph.D. – Leads the nity to deliver a high-efficiency product, as evi- commentary on the many reasons why a-Si standing bifacial capability. This feature Development of Rear Side Passivated Solar denced by recent lab-scale records. Others ben- is an attractive choice as the technology allows for the fabrication of bifacial mod- Cells; Bosch Solar Energy AG efit from manufacturing scale and a communi- path. He argues that a-Si delivers not just the ules with higher energy yield as com- ty of researchers creating advancements on a cost and performance metrics, but also a pared to the standard module of equal Daniel Stichtenoth, Ph.D. – Scientist, Project well-known material system. path to sustainability; claims supported by conversion efficiency. Manager in the Engineering Department of In this issue, we will hear about two dif- the plethora of well-respected companies Crystalline Cells; Bosch Solar Energy AG ferent thin film approaches (CIGS and a-Si) pursuing this path today. Acknowledgments This work is financially supported by Tilo Aichele, Ph.D. – Crystal Growth the German goverment (BMU) within the Research and Engineering; Bosch Solar EnSol project, contract number 0325120A. Energy AG

Endnotes Jan Lossen – R&D Manager; Bosch Solar 1. D. Macdonald, L.J. Geerligs, Appl. Phys. Energy AG Lett. 85, 4061 (2004) 2. A.R. Burgers et al. Proc. of the 25th European Photovoltaic Solar Energy Click here to return to Table of Contents Conf. (2010) PRINT E-MAIL 3. T.L. Aselage, Journal of Materials this article this article Research, 13, 1786 (1998) 4. M.A. Kessler, et al. Semicond. Sci. Technol. 25, 055001 (2010)

38 | Future Photovoltaics | April 2011 << PREVIOUS PAGE | NEXT PAGE >> www.futurepv.com | 39 THIN FILM PHOTOVOLTAICS

Monolithic CIGS achieved. However, it has been challeng- benefits of high speed, high uniformity ing for CIGS manufacturers to deliver a and precise compositional control.[3] reliable and rapid manufacturing process Photovoltaic Modules that can be scaled effectively. In this CIGS Manufacturing Technology paper, we describe such a process devel- Figure 1 shows the HelioVolt CIGS mod- oped by HelioVolt and scaled up on its ule manufacturing flow.The reactive trans- Manufactured by first 20 MW nameplate capacity produc- fer process can form CIGS on a variety of tion line in Austin, Texas. The process is materials, including glass, metals and plas- Reactive Transfer based on reactive transfer and has the tics. In the first product release, soda lime

PRINT E-MAIL Louay Eldada, Dingyuan Lu, Scott Hinson, Billy J. Stanbery this article this article HelioVolt Corporation

Abstract quality CIGS films with large grains were In recent years, thin film photovoltaic fabricated on the production line, and (PV) companies started realizing their low high-performance monolithic modules manufacturing cost potential, and have with a form factor of 120 cm x 60 cm were been grabbing an increasingly larger mar- produced. With conversion efficiency lev- ket share. Copper indium gallium els around 14 percent for cells and 12 per- selenide (CIGS) is the most promising thin cent for modules, HelioVolt started com- film PV material, having demonstrated mercializing the process on its first pro- the highest energy conversion efficiency duction line. in both cells and modules. However, most CIGS manufacturers still face the chal- Introduction lenge of delivering a reliable and rapid The interest in harnessing the sun for manufacturing process that can scale energy through PV technologies has effectively and deliver on the promise of increased tremendously in recent years, this material system. HelioVolt has devel- as the importance of using renewable oped a reactive transfer process for CIGS energy has moved to the forefront of absorber formation that has the benefits social consciousness. With their cost of good compositional control and a fast, advantage, thin film PV technologies have high-quality CIGS reaction. The reactive been attracting significant attention. transfer process is a two-stage CIGS fabri- Copper indium gallium selenide (CIGS) is cation method. Precursor films are the most promising thin film material, deposited onto substrates and reusable having exhibited the highest thin film cover plates in the first stage, while in the energy conversion efficiency. Record effi- second stage, the CIGS layer is formed by ciencies of 20.3 percent for cells[1] and Figure 1 – Schematic of the Reactive Transfer Process rapid heating with Se confinement. High- 15.1 percent for modules[2] have been

<< PREVIOUS PAGE | NEXT PAGE >> 40 | Future Photovoltaics | April 2011 www.futurepv.com | 41 Monolithic CIGS Photovoltaic Modules Manufactured by Reactive Transfer THIN FILM PHOTOVOLTAICS

glass (SLG) with a form factor of 120 cm x high-performance thin film CIGS absorber ature, enabling lower cost and higher after the CIGS reaction, driving the manu- 60 cm is used as the substrate material. As layer.[5] In the first stage, Cu-In-Ga-Se- throughput. In the second stage, these pre- facturing cost even lower. SLG has a low strain point around 518°C[4], based precursor films are deposited onto a cursors are rapidly reacted under pressure, After the reactive transfer p-type CIGS good thermal uniformity is required in pro- substrate and a cover plate, forming the with the plates being in proximity (not in formation, an optimized n-type thin cessing. The monolithic interconnection chemical basis of CIGS. This film stack contact) and the precursor film stacks fac- buffer layer is deposited with a wet scheme involves three main pattern steps, allows precise control of the composition ing each other. A controlled amount of chemical process. After mechanically including laser and mechanical scribing. and the crystalline structure, and affords material transfers in the vapor phase from scribing the active materials, a layer of After the initial deposition step of sputter- different choices of processing conditions. the cover plate (a reusable “source” plate) to transparent conducting oxide (TCO) ing the molybdenum (Mo) back contact In addition to physical vapor deposition the substrate, thereby conditioning both material is sputtered onto the structure, film and patterning it, a two-stage field- (PVD), other rapid deposition methods, the bulk and the surface of the resulting followed by the final isolation scribe. assisted simultaneous synthesis and trans- such as liquid precursor spray printing, can CIGS film. By pulse-heating the films or by Module packaging is finished by edge fer (FASST®) reactive transfer process also be utilized.[6] Furthermore, precursors rapid thermal processing, the overall ther- sealant application, lamination and J-box (Figure 1) is utilized for the formation of a can be deposited at a low substrate temper- mal budget is significantly reduced. The attachment after bus bar and tab connec- lower heat requirements enable the poten- tion. HelioVolt’s unique edge sealant tial use of low-cost, less thermally stable solution is applied for higher moisture 106 substrate materials, such as polyimide. resistance. This helps the guarantee of a (a) Sufficient pressure can substantially pre- 25-year lifetime, including in high- Sample 5-8 5-8-a:CsCu+(25.00%)(<) vent the loss of selenium (Se) from the humidity environments. Both the encap- 5-8-a:CsGa+(25.00%)(<) reaction zone, thereby achieving highly sulant sheet and low-iron tempered 5-8-a:CsSe+(25.00%)(<) 5 10 5-8-a:Cs113In+(25.00%)(<) efficient incorporation of Se into the com- superstrate glass used in lamination have position layer. The cover plates are reused high optical transparency to improve CIGS

104 Mo (a) (b)

103 Count rate (counts/s) 60 (b) 50

102 30

0 20 25 30 35 40 45 50 55 60

101 01020 30 40 Time (minutes)

Figure 2 – (a) SIMS depth profile of a CIGS film formed by the reactive transfer process; Figure 3 – SEM micrographs showing (a) cross-sectional and (b) top views of CIGS films synthesized (b) XRD pattern of a CIGS film fabricated by the reactive transfer process. in HelioVolt’s production line in Austin, Texas.

<< PREVIOUS PAGE | NEXT PAGE >> 42 | Future Photovoltaics | April 2011 www.futurepv.com | 43 Monolithic CIGS Photovoltaic Modules Manufactured by Reactive Transfer THIN FILM PHOTOVOLTAICS

photon transmission. Various tests, such shows a secondary ion mass spectroscopy age (Voc) was 630.5 mV, the short-circuit predicted equivalent efficiency for mod- 2 as efficiency and wet dielectric measure- (SIMS) depth profile of a CIGS thin film current density (Jsc) was 30.9 mA/cm and ules is around 12 percent, if the material ments, are conducted on every module to processed by reactive transfer. The pre- the fill factor was 71.8 percent. quality achieved in a cell is achieved uni- ensure good quality control and product cursors for the film are deposited by PVD. HelioVolt started scaling up its PV tech- formly across the area of the module, with safety. Tests such as thermal cycling, The uniform elemental distribution indi- nology from the cell level to the module the power loss from 14 percent to 12 per- damp heat and outdoor module array cates that a complete reaction of the pre- level by first creating a generation 1 (G1) cent being mainly due to the non-active testing are done on selected samples to cursors took place, and the X-ray diffrac- platform for modules. The G1 form factor interconnect scribes area and to lateral ensure exceptional product reliability tion (XRD) analysis (see Figure 2(b)) con- is 30 cm x 30 cm. With a 14 percent effi- current conduction loss in the finite- and performance in the field. firms the absence of deleterious phases ciency level achieved in HelioVolt cells, the resistance TCO layer. Current crowding other than CIGS. All the XRD peaks are CIGS Formation based on Mo and chalcopyrite-type CIGS The reactive transfer process can syn- structure, with a preferred 220/204 CIGS Device ID: DAAPANEL-4560 Device Temperature = 24.3°C thesize CIGS from the precursors in a orientation. Evidence indicates that Jan 06, 2011 13:37:57 MT Device Area = 6420.3 cm2 matter of several minutes. As we continue 220/204-oriented films may help junction Spectrum: ASTM G173 global Irradiance = 1000.0 W/m2 to reduce the cycle time, the reaction time formation and improve solar cell per- can be shortened to seconds. Figure 2(a) formance.[7] This unique processing LACSS IV System approach results in a significantly PV Performance Characterization Team decreased thermal budget compared with 2.0 common CIGS manufacturing methods, such as co-evaporation and two-step sel- enization processes. This advantage leads 1.5 to higher throughput, and therefore significantly lower cost. 1.0 Figure 3(a) shows a tilted cross-sectional SEM micrograph of a CIGS layer (on

Mo) formed with PVD precursors on our Current (A) 0.5 production line. High-quality CIGS films with large grains are yielded by the reac- 0.0 tive transfer process. Figure 3(b) shows a SEM top view of a CIGS layer in another sample on the production line. The -0.5 -10 0 10 20 30 40 50 60 70 80 faceted surface enables high efficiency under various lighting conditions. Voltage (V) V = 57.09 V V = 75.90 V max Device Results and Module Scale-up oc I = 1.534 A I = 1.325 A Solar cells with a conventional device sc max Fill Factor = 65.0% structure of glass/Mo/CIGS/buffer/TCO Pmax = 75.64 W were fabricated. The CIGS absorber was Efficiency =11.8% Figure 4 – Photograph of a G2 (120 cm x 60 cm) fabricated by the reactive transfer process module produced at HelioVolt's first factory in with PVD-based precursors. Cell efficien- Austin, Texas (inset: view of HelioVolt’s automat- Figure 5 – Light current-voltage (I-V) curve measured by NREL on a HelioVolt G2 (120 cm x 60 cm) ed manufacturing line). cies of 14 percent were achieved.[5] With a module. The efficiency is 11.8±0.6 percent and the output power is 75.64 W. cell area of 0.66 cm2, the open-circuit volt-

<< PREVIOUS PAGE | NEXT PAGE >> 44 | Future Photovoltaics | April 2011 www.futurepv.com | 45 Monolithic CIGS Photovoltaic Modules Manufactured by Reactive Transfer THIN FILM PHOTOVOLTAICS

within the active segments can also con- high uniformity over large areas. CIGS About the Authors tribute to the power loss, which can be films with large grains on the order of improved by choosing the right segment microns are routinely produced, and Louay Eldada is the chief technology officer Billy J. Stanbery is chairman and chief strat- width based on CIGS and TCO films char- exhibit optimal crystallographic orienta- for HelioVolt Corporation, heading its R&D egy officer of HelioVolt Corporation, having acteristics. We measured in our G1 mod- tion. Cells with 14 percent efficiency and and engineering departments, and leading founded the company to develop and com- ules conversion efficiencies as high as modules up to 120 cm x 60 cm with 12 the design and development of photovoltaic mercialize the FASST® process for manufac- 12.03 percent, and this performance level percent efficiency are produced with this technologies, processes, equipment, modules turing CIGS thin film PV that is now covered was independently verified by the process. HelioVolt started to commer- and systems. Dr. Eldada has published 200 by 11 patents issued and nine additional National Renewable Energy Laboratory cialize its reactive-transfer-based CIGS papers and 40 books, organized and chaired patent applications. Prior to founding (NREL) at 11.7±0.6 percent efficiency.[5] technology at its first factory in Austin, 160 conferences, presented 180 invited and HelioVolt, he managed Boeing’s terrestrial The 20 MW manufacturing line at Texas. keynote talks, garnered 50 technical awards PV program. During that 17-year tenure, his HelioVolt’s first factory in Austin, Texas is and holds 47 patents. He received his Ph.D. team manufactured and deployed PV for producing G2 modules, with a form factor Acknowledgments in optoelectronic devices and systems from spacecraft, and achieved the world record in of 120 cm x 60 cm. Figure 4 shows one The authors would like to acknowledge Columbia University. multi-junction thin film PV conversion effi- such G2 module, with the inset showing the entire HelioVolt team for their contribu- ciency. Dr. Stanbery completed his Ph.D. part of the automated manufacturing line tion in the successful scale-up of the CIGS in chemical engineering in 2001 at the at HelioVolt. reactive transfer technology. We would also Dingyuan Lu is a process design scientist at University of Florida, having previously During the G2 size module production, like to thank Keith Emery and his team at HelioVolt Corporation, where he is responsi- obtained an M.S. in physics from the it is critical to have good uniformity con- NREL for the verification testing. ble for the development of processes for the University of Washington and a B.S. in trol in each of the processes and in-line formation of next-generation, higher-efficien- physics and mathematics from the metrology for monitoring. For example, References cy and lower-cost CIGS absorber thin films. University of Texas. extra attention has been paid to the uni- 1. http://www.renewableenergy He also works on photovoltaic device analy- formity of the thickness and composition focus.com/view/11941/solar-thinfilm- sis and modeling for the fabrication of CIGS of the precursor films to avoid deleterious reaches-203-efficiency solar modules. Dr. Lu received his Ph.D. in Click here to return to Table of Contents effects on module performance. Heating 2. http://www.avancis.de/en/presse/ semiconductor optoelectronic devices from uniformity, especially at the high temper- pressemeldungen/anzeigen/meldung/ the University of Texas at Austin. PRINT E-MAIL this article this article atures during the reactive transfer, can neuer-weltrekord-avancis-erzielt-151- also be very important to ensure uniform wirkungsgrad CIGS reaction and avoid glass substrate 3. http://www.heliovolt.com/technology/ Scott Hinson is a senior electrical product warpage. presentations.php engineer at HelioVolt Corporation, where he The scale-up to the G2 form factor was 4. O. S. Narayanaswamy, Journal of the leads module packaging, performance test- successful, with HelioVolt achieving American Ceramic Society, 64 (1981) 109. ing, reliability and certification efforts. Prior approximately 12 percent module effi- 5. L. Eldada, B. Sang, M. Taylor, P. Hersh to working at HelioVolt, he worked in the ciency. NREL confirmed the G2 efficiency and B. J. Stanbery, Proc. SPIE, 7409 (2009) military, medical, consumer and oil indus- at 11.8±0.6 percent, as shown in Figure 5. 74090N. tries developing power supplies, precision For this large format module, this high- 6. C.J. Curtis, M. van Hest, A. Miedaner, J. measurement equipment and inductive heat- efficiency level corresponds to an output Nekuda, P. Hersh, J. Leisch and D. ing technologies. Scott received his B.S.E.E. power of 75.64 W. Ginley, Proc. IEEE Photovolt. Special. Conf., from The University of Texas at Austin with 33 (2008) 1065. undergraduate specialization in communica- Conclusion 7. S. Chaisitsak, A. Yamada and M. tion systems and power distribution. HelioVolt’s reactive transfer process Konagai, Jpn. J. Appl. Phys., 41 (2000) 507. produces high-quality CIGS films with

<< PREVIOUS PAGE | NEXT PAGE >> 46 | Future Photovoltaics | April 2011 www.futurepv.com | 47 THIN FILM PHOTOVOLTAICS

Brave Steps Needed in Thin Applied Materials was not happy with With that, the situation at the two the drop in sales, so it was just a matter of most prominent tool and manufacturing time when the management had to cut line suppliers – Applied Materials and Film Silicon R&D to Harvest back on efforts and with that, on expens- Oerlikon – is described. When it comes to es. But single-tool sales continue; most panel manufacturing, the companies Optimal Technology importantly, the sales of the PECVD tool with longest track records and the lead- that forms the heart of a thin film silicon ers in volume and technology are Uni- production line. In fact, most of the other Solar, Sharp and Kaneka. All three panel TM Torsten Brammer PRINT E-MAIL tools in the Sunfab manufacturing line manufacturers had been around well this article this article Photovoltaic research consultant were supplied by OEMs anyway. The stat- before Applied Materials and Oerlikon ed continuation on R&D efforts, even if on entered the business; hence, they do not a smaller scale, also underscores that rely on Applied Materials and Oerlikon. Applied Materials sees a future potential This fact clearly states that the ups and When Applied Materials announced this marks the apparent failure of thin in thin film silicon. downs at Applied Materials and Oerlikon the discontinuation of their product film silicon. But is this based on a thor- This potential was confirmed impres- do not fully represent the situation of SunfabTM, a fully integrated line for manu- ough analysis? sively by Applied Materials and many thin film silicon. facturing thin film silicon solar panels, Let us look at the press release from other companies and research institutes The picture of thin film silicon becomes many took this as a clear signal that thin Applied Materials: http://www.appliedma- at the most recent European PV trade quite interesting when we look at the situa- film silicon is not a viable technology. terials.com/news/articles/applied-materi- fair and technical conference, held in tion at Sharp. Sharp started R&D on solar According to analysts, the material to als-announces-restructuring-energy-and- Valencia (EU PVSEC 25) in September cells in 1959. Work on thin film silicon start- work with is CIGS (new material that environmental-solutions-segment). The 2010. Basically, all major thin film silicon ed in 1983. Sharp is one of the top players promises high efficiencies) and crys- only words that are usually looked at are: PV labs and companies presented signifi- in crystalline silicon and has reached effi- talline silicon (mature and dominant “As part of the restructuring, Applied will cant progress on efficiency and commer- ciencies of more than 35 percent with market share). CdTe is doing fine anyhow. discontinue sales to new customers of its cialization. Oerlikon made the strongest a compound solar cell. Sharp began opera- While there are surely some facts that SunFab™ fully-integrated lines for manufac- claim, with the launch of their newest tion of its Japanese thin film silicon fab support this analysis, it is time to see if all turing thin film solar panels … .” manufacturing line, called ThinFab, which with an initial capacity of 160 MW in facts are considered and to look at thin apparently allows production costs of March 2010 (http://www.sharp-world.com/ film silicon more closely. Most readers never made it to the fol- 50 € cents/Wp at 10 percent total area corporate/news/100329.html). Sharp will Now is the post-hype time for thin film lowing words: panel efficiency (http://www.oerlikon.com/ also start the production of thin film silicon silicon. The hype was created by the “Applied … will offer individual tools for solar/thinfab/). Operation of a ThinFab panels in Italy in the second half of 2011, prominent battle between Swiss Oer- sale to thin film solar manufacturers, includ- could start in mid-2012. This clearly chal- with an initial capacity of 160 MW. Plans likon and U.S. Applied Materials on mar- ing chemical vapor deposition (CVD) and lenges First Solar’s cost target of 52-63 $ include an expansion to 480 MW ket shares and intellectual property. Both physical vapor deposition (PVD) equipment. cents/Wp in 2014, as presented by First (http://sharp-world.com/corporate/ companies achieved considerable sales … The company will support existing SunFab Solar at the same conference. Oerlikon news/100802_2.html). When questioned if volume of half- and fully integrated man- customers with services, upgrades and capac- representatives claim that their cost cal- Sharp sees a certain pressure on sales after ufacturing lines between 2006 and 2008. ity increases through its Applied Global culation method is equivalent to the one all the bad press on thin film silicon, a Both companies have sold none to little Services segment.” done by First Solar. So when the cost Sharp representative during the Solar volumes in the recent past, which was potential of thin film silicon is chal- Power International trade show in October considered a first signal of potential fail- Just as important is this: lenged, there are answers. According to a 2010 in Los Angeles pointed to a massive ure of thin film silicon. In July 2010, the “R&D efforts to improve thin film panel press release from Oerlikon dated poster showing a utility-scale PV power discontinuation of Applied Materials’ efficiency and high-productivity deposition February 22, 2011, Dong Xu Ltd. is con- plant using thin film silicon panels. The SunfabTM was announced. For many in PV, will continue.” vinced of the competitiveness. Sharp representative firmly stated that the

<< PREVIOUS PAGE | NEXT PAGE >> 48 | Future Photovoltaics | April 2011 www.futurepv.com | 49 Brave Steps Needed in Thin Film Silicon R&D to Harvest Optimal Technology THIN FILM PHOTOVOLTAICS

discussions about thin film silicon have not item 7 – the present efficiency level in most the leader in this technology. The speed at needs, and it’s doubtless some of those will altered Sharp’s business, that Sharp’s productions worldwide is 8-10 percent. which Sharp will take these steps will be be very price- or volume-sensitive, which choice for utility-scale plants is thin film sil- Instead, the high availability and minimum regarded as the equivalent for the confi- could give thin film Si an edge. The bottom icon and that the sales volume is well above usage of silicon, the non-toxicity and the dence level of Sharp in thin film silicon from line is that PV technologies change very production volume. In summary, for Sharp, inherent long-term stability of silicon might a technological as well as from a marketing fast. One minute polycrystalline silicon is a global technology company with a long have played an important role. The often- point of view. All players will need to take in short supply and so is expensive; the history in PV, thin film silicon plays an criticized usage of NF3 for PECVD chamber brave steps to upgrade their existing lines to next it’s the opposite. Raw silicon costs important role for its future expansion cleaning can be avoided by using fluorine. eliminate unexpected bottlenecks and were the reason for the interest in thin film strategy. However, the recent announce- The high-performance ratio of thin film sil- upgrades to close the gap in efficiency Si but were also responsible for the shine ment that Sharp will use part of its Sakai icon is studied by many and will be promot- and/or costs to CdTe, CIGS and, of course, wearing off; however, whenever there is a operations for back-contact c-Si cells clear- ed even more strongly in the future (see, crystalline silicon. The steps will take some specific cost of materials issue, such as ly shows that Sharp also sees the challenge e.g., the agenda of the 3rd Thin Film Solar courage due to the instable political frame- could be the case with c-Si and the ele- with thin film silicon. Summit Europe, March 3-4, 2011, Berlin). work of this industry and large oversupply. ments of CIGS/CdTe, thin film Si will come What could be the reasons for Sharp or Even if the “softer” items 1-6 listed above Future expansion, essential to cut down on to prominence. So don’t write this thin- Inventux to continue with thin film sili- might have played the decisive role for the overhead costs, will take additional R&D film Si off just yet, as you could regret it. con? There are several key aspects that technology choice for the present players in budget up front to develop the next product Times have changed radically, no doubt. need to be fulfilled by any PV technology thin film silicon, the efficiency and cost-per- generation. This would be facilitated by The hype that was initially triggered by the to reach grid parity in a sustainable way: watt challenge needs to be tackled for thin high-risk R&D work at the public level. For entry of Applied Materials and Oerlikon, in 1. Environmentally friendly along the film silicon. In the present situation, fulfill- that, policymakers will need to be con- addition to PV panel undersupply and high whole supply chain (from mining of ing announced targets is key to regain trust vinced that R&D projects for thin film sili- silicon feedstock prices, is finally over. the resources to the end-of-life of the outside the thin film silicon community. con are worth being funded. In order to Stakeholders in the thin film silicon indus- panel) Oerlikon has made big claims with its new improve competitiveness, the thin film sili- try will now need to show more results 2. High availability of resources and/or product ThinFab. It will be interesting to see con community might have to move closer with less funding, but that’s no different minimized material consumption to whether Oerlikon can compete with other together to share funds and resources and than the teething troubles of many tech- allow scaling and low material cost thin film turnkey suppliers such as Manz align on a common road map. nologies. At the EU PVSEC 25th conference volatility and Centrotherm for CIGS or Roth & Rau for The other large element that is left out in Valencia, many promising results were 3. High inherent long-term reliability CdTe. Similarly interesting to see will be the of many people’s minds is the fact that the presented and many presenters promised 4. High performance ratio (kWh/kW at a outcome of Applied Material’s efforts to PV industry is still relatively young, espe- even better results in the future. We as an specific site) improve efficiency and productivity. The cially when viewed as a consumer-based industry need to give them more time to 5. Good aesthetics more-focused-since-downsized activity, in business. Most people in the industry are allow for a new phase in the evolution of 6. Fit between PV technology and target- combination with no new turnkey business still struggling to think beyond solar farms thin film silicon to progress. ed application in parallel and less-global attention, might and cannot see how solar technologies can 7. High efficiency now and high future even help the Applied Material team to gain be used in diverse and unique environ- efficiency potential momentum in specific areas. Sharp’s goal ments such as mobile applications (back- 8. Low production costs now and high announced in 2008 to gradually move packs, portable power generation), novel About the Author future cost-reduction potential toward 1 GW is still in the mind of many ideas such as the solar roadway and the industry experts (http://www.sharp- various forms of building-integrated PV Torsten Brammer – See page 7 In order to make a sound technology world.com/corporate/news/080327.html). technologies (roof tiles, 6 square-meter choice, all items need to be analyzed and Together with Kaneka (Sharp’s long-term building facade elements as offered by Click here to return to Table of Contents weighted both for the present and for the Japanese rivals in this technology), and Schüco, etc.). All can have an impact on the future. It could be that Sharp and other thin Schüco (the new German heavyweight after development of differing PV technologies PRINT E-MAIL film manufacturers have not only looked at buying Sunfilm’s two sunfabs), Sharp will be like thin film Si, as each has its own unique this article this article

<< PREVIOUS PAGE | NEXT PAGE >> 50 | Future Photovoltaics | April 2011 www.futurepv.com | 51 Future Photovoltaics MANUFACTURING EQUIPMENT & MATERIALS PRINT E-MAIL Click here to return to Table of Contents this article this article

Craig Hunter Vice President and General Manager Don’t Miss Out Clean Energy Technologies, Intermolecular Sign up NOW for our FREE biweekly newsletter The PV industry has for some years been waged regarding specific process steps in hoping for better solutions to the vexing and standard cells. For the “evolutionary” crys- expensive problem of making contact to crystalline silicon roadmap, what specific tools talline silicon solar cells. The materials, equip- and processes will prevail in texturing, diffu- ment and process involved in screen-printing sion, passivation, metalization – continuous- contacts impact virtually every element of ly improved standard tools and processes, or solar cell cost and performance – conversion something completely different? efficiency (in several ways, including fill factor The claims being made by newcomer and effective collection area/shadowing), Xjet, which offers inkjet tools and inks, are mechanical and electrical yields, wafer thick- aggressive indeed – lower capital costs, half ness, capital and material cost. Needless to the materials usage, higher conversion effi- A roundup of the industry’s top say, there are enormous efforts afoot to ciencies (narrower lines, ability to tailor improve every element of screen-printing con- materials specifically for seed layers vs. bulk), R&D stories in one place tacts – lower-cost materials, materials that thin-wafer enabling and portable to back- make contact to lightly doped emitters, tools contact cell designs. Be notified when there is a new that print narrower or higher-aspect ratio With pilot tools reported to be in the field Future PV blog post lines (i.e., double printing), to name just a few. already, we will be waiting with baited breath While many industry participants debate for updates in the coming months and years Be notified about our partner “revolution vs. evolution” in solar cell materi- on this intriguing potential breakthrough in developments and partner events als and architectures, the same debate can be crystalline silicon metalization.

<< PREVIOUS PAGE | NEXT PAGE >> www.futurepv.com | 53 Thought Leadership Profile MANUFACTURING EQUIPMENT & MATERIALS

Click here to return to Table of Contents PRINT E-MAIL this article this article Swagelok Company supports the global semiconductor marketplace with Generating Energy for skilled associates, cutting-edge fluid system technology, and high-purity PV-Specific Standards and ultrahigh-purity manufacturing operations. The company’s manufacturing, research, technical support, and For the global photovoltaic (PV) power new Swagelok Photovoltaic Process distribution facilities support a global industry to prosper, it must become more Specification (SC-06) is the photovoltaic network of more than 200 authorized competitive with the cost of traditional industry’s first specification for process- sales and service centers in 57 coun- electric power. That means reducing the ing stainless steel fluid system compo- tries. A wholly owned subsidiary of cost per kilowatt-hour (kWh) associated nents. It’s based on our work in the Swagelok Company, Swagelok with solar-generated power to a level that industry, and continues Swagelok’s lead- Semiconductor Services Company yields grid parity between solar power ership in process-specific cleaning, (SSSC) facilitates a focus on industry and traditional electricity. which began in the 1980s with ultrahigh- needs. Among the factors that weigh on the purity components. PV industry’s ability to achieve grid parity With Swagelok SC-06, you can now Swagelok Semiconductor Services Company are standards that add to the total cost specify fluid system components with per kWh. Following standards is a pre- cleanliness and purity levels that closely 3200 Scott Blvd. Santa Clara, CA 95054 ferred practice as specifications ensure match the true process requirements for www.swagelok.com quality in components, processes, manu- solar cell production. Products processed facturing and end-product performance. with SC-06 bridge the gap between ultra- Business Contact high-purity standards and the more limit- Al Bousetta ed requirements of general industry prod- Market Manager However, in some cases, PV standards ucts. Applied to gas handling components Phone 408.764.0240 may be stricter than what is necessary to in your system, this specification is [email protected] ensure the safe, efficient generation of designed to help lower overall cost of sys- power. By their nature, tighter specs are tem ownership. All with the peace of more costly. As PV energy producers and mind that comes with the quality and market suppliers work toward reducing reliability of Swagelok® products. Using costs, they need to examine existing stan- SC-06, our customers take one step closer dards and determine if less stringent to grid parity. specifications can yield acceptable results while saving money. Swagelok is addressing one such area of PV manufacturing that defines specifi- To learn more about Swagelok SC-06 cations for stainless steel components and applicable products, visit: used in the production of solar cells. The www.swagelok.com/solar

<< PREVIOUS PAGE | NEXT PAGE >> 54 | Future Photovoltaics | April 2011 www.futurepv.com | 55 MANUFACTURING EQUIPMENT & MATERIALS

Xjet’s FabGrade Inkjet inkjet solution includes multiple metals in From a mass production perspective, one printing step, enabling a transition to inkjet is a much safer investment than lower-cost metals. Xjet’s front-side metal- other exotic deposition strategies being for Solar Cell Metalization ization printing consists of three types of proposed and promoted. Inkjet is a com- inks in one path: fritted silver for seed; monplace, mature industry and technolo- pure silver on top; and a low-cost bus bar gy. Each year, hundreds of millions of dol- Robert Even PRINT E-MAIL metal providing superior adhesion at a lars are invested by world-leading compa- Xjet Ltd. this article this article lower cost. In the future, additional metals nies such as Epson, HP, Ricoh, Konica may easily be added onto this single print- Minolta, Fuji and many others in the er to further reduce costs by layering mul- development and production of more reli- tiple metals on top of each other. able, precise and lower-cost inkjet heads. The company is now developing back- The graphic arts market drives these Abstract Boosting efficiency through double contact pastes. Xjet’s printers are upgrad- demands and provides the PV industry a Xjet develops and manufactures inkjet print adds 30 percent more metal per cell, able to back-side printing at minimal cost. “free,” fully subsidized and guaranteed production printers and inks for solar cell but fails to provide a return, as it adds Producing front and back metalization on roadmap for better, finer inkjet heads. metalization. The company’s design over- metal when it should be reducing metal. an inkjet printer in a single step will Xjet intends to transform inkjet tech- comes inherent inkjet drawbacks and is New-generation pastes demonstrate an reduce capital costs for the cell producer nology from industrial grade for short- covered by over 20 patents and pending improved contact on high-sheet rho by close to 50 percent and become a runs to fabrication-grade technology for patents. wafers, but they also require about 30 strong driver for thinner wafers and fur- 24/7 production. Starting some 20 years percent more metal, with fingers as tall ther cost reduction. ago with HP Scitex (formerly Idanit), a as 35 microns. In addition, despite Inkjet technology has been long envi- world-leader in inkjet printing of bill- Metalization is becoming too expen- reduced silicon prices, wafer costs still sioned as the industry’s “silver bullet.” On boards, the Idanit printer completed a 3.5 sive. The rapid decline in $/watt price ver- represent about 65 percent of the cell paper, inkjet has all that is required to x 1.6 meter poster in less than 45 seconds. sus the spiraling price of metal has blown price, forcing the industry to renew its become the industry’s leading deposition A few years later, Objet, a 3D inkjet print- metalization out of proportion. Its cost as pursuit of wafer thinning. Double print tool – much lower cost than photolithog- er for rapid prototyping, was launched. a percentage of c-Si cell price has rocketed adds substantial mechanical pressure to raphy, an additive process eliminating the Using UV curable polymers, this innova- from 3 percent in 2008 to 12 percent in the wafers and does not support this material waste of spin coating and con- tive inkjet printer converts CAD images to 2010 (normalized for efficiency). Photon trend for thinner wafers. tactless printing, unlike the wafer-crack- colorful plastic objects within minutes. Consulting recently published a forecast Consuming less than half the metal ing squeegee used in screen printing. In for a continued decline in cell pricing from and printing 35 micron features without addition, inkjet “counts” and deposits the present $1.2/watt to $0.7/watt by the making contact, Xjet’s inkjet technology nano-droplets, ensuring very high unifor- end of 2012. Even under the unlikely provides a solution to cut metalization mity across the substrate as well as sub- assumption of a freeze in raw metals costs by about 40 percent. Over the past 12 strate-to-substrate. price, the decline in cell price alone would months, Xjet has demonstrated consistent drive the cost of metalization to an unac- output of 18.8 percent average efficiency ceptable 20 percent of the cell price. and 19.2 percent best cells on production Metalization must advance quickly and floors. Recent development has shown an its cost must be cut in half. New solutions 80 percent fill factor with Xjet’s full inkjet must consider the use of lower-cost metals front metalization solution. Xjet is demon- and a decrease in the quantity of metal. strating printing speeds above 20 cm per Xjet Printed Seed Layer, 35 Microns Wide, Lowering the capital cost of metalization second, which translates to a throughput A Bicycle Chain, Printed on Objet’s 3D Inkjet Printer With 25 Microns Core equipment would also be beneficial. of 5,000 wafers per hour. Moreover, this

<< PREVIOUS PAGE | NEXT PAGE >> 56 | Future Photovoltaics | April 2011 www.futurepv.com | 57 Xjet’s FabGrade Inkjet For Solar Cell Metalization MANUFACTURING EQUIPMENT & MATERIALS

Objet, like Idanit, soon became a market material to inkjet, direct metalization is cement through your showerhead – how tion into two stages – first, a very shallow leader in its field. much more difficult to implement. Inkjet long would it take the showerhead to clog seed layer achieves a “perfect” contact fol- In these two companies, with thou- experts can quickly compute that given and fire astray? In the PV world, the equiv- lowed by pure silver on top, for lower sands of installed printers and multiple the pattern, drop volume and deposition alent of cement is metal and the equiva- resistivity, minimizing the amount of sil- innovations, inkjet was driven further rate required for metalization, hundreds lent of the showerhead is the 100,000 30- ver consumed. Fraunhofer ISE, Dr. Ebong than what was deemed possible. The of inkjet heads, operating simultaneous- micron diameter nozzle array. With hun- from Georgia Tech and Xjet have proven next major challenge was to develop the ly, are required to meet industry stan- dreds of inkjet heads and over 100,000 in numerous experiments that this met- fastest, most reliable inkjet possible. dards. An additional challenge is to jet nozzles firing 10,000 droplets per second alization architecture ensures good ohmic Newcomers to inkjet may assume that droplets with abrasive and heavy materi- per nozzle, all registered to 5 micron accu- contact on shallow emitters > 100 investment in ink-to-head compatibility als (glass frits and silver) with a precision racy, one begins to understand the chal- ohm/sq., eliminating the need for selec- will eventually yield a solution to known of 5 microns. lenges Xjet had to overcome. tive emitters. Efficiency has been proven inkjet deficiencies such as clogged and Even a home showerhead, jetting Companies have tried to integrate above 19 percent at customer sites, and misfiring nozzles. This is incorrect, as water from relatively big nozzles, gets inkjets in LCD or RFID for almost two research institutes show promise for even heads will always clog and inks will never clogged and nozzles stray fairly quickly. decades. They were rarely successful in a higher efficiency to be harvested through be good enough. Unlike inkjetting wax Water is the equivalent of wax or aqueous demanding, production environment. further process optimization. (used for masking), which is the easiest solutions in inkjet. Now think of jetting Without the FabGrade qualities, which The company highlights that front- address the problematic clogging and grid metalization is just the first stepping- poor repeatability of inkjets, it is hard to stone, and back-contact patterning with imagine how inkjet could be adapted to advanced materials is on the company’s the demanding fab (cell production) immediate roadmap. Xjet has already environment. shipped pilot printers to the world’s lead- Xjet’s FabGrade inkjet prints wafer after ing cell manufacturers. The company wafer at high speed with sub-50 micron expects to perfect the product through lines and a high degree of uniformity. pilot runs in 2011 and be ready for a pro- Every droplet is measured and every noz- duction ramp-up toward the end of the zle is tracked by the system in real time to year. shut off automatically if clogged or firing stray. Inkjet heads are automatically cleaned and maintained. A metrology system measures finger widths in real time About the Author and automatically selects the best nozzles to jet as narrow fingers as possible. Robert Even is the marketing manager for Each connector and plug on the print- Xjet. He has over 12 years of high-tech mar- er is monitored – if one is disconnected, keting experience with inkjet and automatic the printer shows when and where this inspection capital equipment. Robert holds a takes place. Every moving part is meas- B.Eng from McGill University, and an MBA ured for lifetime expectancy and preven- from INSEAD, France. tive maintenance. If the system is knocked out of alignment, the printer Click here to return to Table of Contents quickly recalibrates itself. Xjet-Printed Full Conductor Layer, 43 microns Wide, 25 microns High Similar to screen printing double print, PRINT E-MAIL Source: Solar Energy Research Institute of Singapore, 2010 this article this article the Xjet printer separates front metaliza-

<< PREVIOUS PAGE | NEXT PAGE >> 58 | Future Photovoltaics | April 2011 www.futurepv.com | 59 Future Photovoltaics METROLOGY, TEST & FAILURE ANALYSIS PRINT E-MAIL Click here to return to Table of Contents this article this article Optimizing Approach for Alain C. Diebold Empire Innovation Professor of Nanoscale Science; Executive Director, Thin Film PV Production Center for Nanoscale Metrology, College of Nanoscale Science and Engineering, University at Albany; AVS Fellow; Senior Member of IEEE Jean Randhahn,1 Stefan Gritsch2 PRINT E-MAIL this article this article 1Dr. Randhahn Solar Engineering The opportunity for manufacturing control-based cost reduction 2Dr. Schenk GmbH Industriemesstechnik

Amid the effort to reduce the costs asso- time monitoring of materials properties and ciated with solar energy, there is a golden equipment status. This information can be opportunity to import some of the estab- sent forward to the next process step, and Abstract all production lines feature a sun simula- lished principles of manufacturing control software can be utilized to alter the process In current thin film solar manufactur- tor to determine the final module quality from the silicon semiconductor community. for optimum results. This information can ing, module quality and process conditions in terms of their electrical rating. Factory control is highly automated using also be used to correct equipment faults. are primarily determined by physical met- The goal for production is to maximize an integrated computer management system The solar industry is introducing similar rics and the final electrical metrics of the equipment uptime and yield and to stabi- that regulates the flow of wafers through the approaches to manufacturing control. product, the solar module. For established lize all processes. For all of these steps, equipment and tracks each process step. Especially important are correlations well-known manufacturing processes the above-mentioned metrology can be Metrology information is captured with between physical and electrical characteris- such as optical disc, this is feasible, as the sufficient. The next step is raising product enough detail to determine if the measure- tics and yield. As the solar industry works effects and causes of defects have been quality, which is equal to an increase in ment equipment itself is properly calibrated toward improved efficiency and yield, it will thoroughly studied and documented. For power output and the securing of guaran- and operating at required specifications. Film be keeping watch over production costs. It TF PV production, however, this is not the teed product longevity. The decision of thickness, defect location and a variety of will be very interesting to observe the evolu- case, which leads to several shortcomings how to best achieve this is based on the additional information are all available for tion of manufacturing control and the imple- in process control and production opti- solar module output, the result of which process control, which allows for higher mentation of new metrology. mization. This article shows one optimiza- is obtained in the final electrical test. This yields and optimum circuit performance. In this issue, the papers cover topics asso- tion approach and some of its pitfalls and final test, which is performed on a (often An integral part of this strategy is the use ciated with improved process control. proposes a quality indicator. pulsed light) sun simulator, yields the of advanced equipment control (AEC) and Randhahn and Gritsch discuss the relation- current-voltage characteristic (I-V curve) advanced process control (APC). AEC and APC ship between physical and electrical proper- Minimizing Loss Mechanisms of the solar module. Since a PV module use sensor-based collection of information ties and a new measurement capability In current thin film solar cell manufac- comprises many individual cells, this inside the process equipment to provide real- known as the Solar Measure I-V Curve Tracer. turing, module quality is primarily deter- characteristic is the result of the interac- mined by physical and basic electrical tion of these individual cells. The interac- metrics such as layer thickness, resistivi- tion between the cells, however, remains ty and optical haze. Some sites perform invisible to those measurement devices; offline spectral response and I-V meas- only their result is detectable.The interac- urements on special samples. The sam- tion involves effects such as: pling rate of such offline measurements • current limitation due to single defec- and their special design often render tive cells, e.g., deposition irregularities results unfit for process control though, • current and efficiency gradients over since they do not necessarily represent the module due to layer and interface the true production process. Additionally, properties

<< PREVIOUS PAGE | NEXT PAGE >> 60 | Future Photovoltaics | april 2011 www.futurepv.com | 61 Optimizing Approach for Thin Film PV Production METROLOGY,TEST & FAILURE ANALYSIS

• shunted cells causing a current drain metrology are misinterpreted and lead size of 1 cm2. In this way, the results for Resistance Issue • suboptimal working points of cells to either unnoticed quality loss or different manufacturers or labs can be The most obvious way to contact a sin- cost-intensive corrective actions that compared and side effects are kept to a gle cell of interest on a full module is to use Despite the fact that these effects are yield only limited improvement. minimum. Since cell dimensions on PV its exposed contact and the exposed con- well-known on the individual cell level modules are significantly larger, side tact of the adjacent cell. The latter is con- from laboratory tests, they are often not One possible solution to this dilemma effects must be accounted for. This espe- nected to the buried contact of the cell of observed in production environments due is the measurement of the I-V curve of cially includes resistance issues due to interest (Figure 1). The cell’s contact layer to lack of suitable equipment. Still, it has every single cell. A few years ago, this limited contact layer resistivity, spectral then forms the highest resistance in the been readily acknowledged that only the might have seemed like a tremendous response (SR) for multi-junction cells, and measurement path. Due to this resistance, understanding of loss mechanisms in the task; however, there have been many cell contacting and illumination depend- the current from a single-point contact module will lead to significant perform- major improvements in sun-simulator- ence on module configurations. over the cell is not uniform, which causes ance improvements.[1] Not knowing the related technologies that simplify this interaction on the cell level reduces the job. Today cell I-V recording systems are potential of power output optimization. able to work inline in a production flow. Two cases present themselves: (1) Effects on cell level pass unnoticed Cell I-V Curve Measurement and corresponding defects in absorber A common procedure to measure the production are never corrected. I-V curve of a cell of any thin film technol- (2) Existing defects seen with other ogy is to cut the samples to a standard

Sunlight

Single cell contacting

substrate superstrate

_ Cell side view Sunlight _ Cell top view

Multiple contacts per cell

Single contact per cell

Figure 1 – Substrate and Superstrate Configuration, Single Cell and Cell String Contacting and Single Contact vs. Multiple Contacts Per Cell Figure 2 – Two IV-& Power Curves Showing Same ISC and Power Loss Due to Working Point Mismatch

<< PREVIOUS PAGE | NEXT PAGE >> 62 | Future Photovoltaics | April 2011 www.futurepv.com | 63 Optimizing Approach for Thin Film PV Production METROLOGY,TEST & FAILURE ANALYSIS

errors in shunt and serial resistance ular multi-junction cells, multiple Module Configurations point value is obtained by using the mod- (RP,RS) and therefore in fill factor (FF). weighed semi-monochromatic illumi- In present thin film cell technologies, ule’s PMPP value and breaking it down to One possible solution is to cut out one nations are required to obtain single cell both substrate and superstrate configura- the cells. For example, a module contain- or multiple small cells on the module and information. tions are used (Figure 1). While the super- ing a series connection of cells will fea- perform an I-V measurement on them. Unfortunately, the common rating for strate configuration allows for easy ture a constant current through all cells Although this method delivers repro- continuous spectra illumination (e.g., access to individual cells, the substrate on a first approach. This leads to a specif- ducible results, it is destructive and only assuming AM 1.5 G spectrum as in IEC configuration presents an obstacle in ic working point for each cell in the mod- partially representative for the module. 60904) does not make sense for semi- obtaining cell I-V curves (Figure 1). This is ule, which does not have to correspond to A second solution is to use a (nearly) monochromatic light sources. An alterna- because the sunny side is the same side current at maximum power point (IMPP) full cell contact, such as multiple contact tive solution is the use of a sun equiva- on which the cells are accessible during (Figure 2). More refined approaches may pins, which allows current to distribute in lent. This presumes a linear correlation their production. Substrate configuration even take into account the slight devia- the contact pins rather than in the cell’s between the cell’s short circuit current is used, for instance, by CIGS technolo- tion from this behavior due to other contact layer (Figure 1). (ISC ) and irradiance (G). As ISC can be cal- gies. Since regular contacting schemes effects in the cell. A third solution is to connect multiple culated from the spectral response of a (e.g., pins) cause partial shading, I-V cells on the module in series (Figure 1). cell (SR(λ)) and the spectral irradiance measurements will always be impaired to Solar Measure I-V Curve Tracer This allows for a uniform current distri- (G(λ)), a factor k between the theoretical a certain degree. Such shading especially Recently, Dr. Schenk GmbH Industrie- bution on all contacted cells except for ISC under AM 1.5 G and a semi-monochro- influences the parameters ISC, FF and messtechnik introduced a non-destruc- the first and last ones. This method yields matic light source can be computed. maximum power. tive, high-speed metrology system. It is good spatial resolution in combination capable of inline I-V curve measurement with multiple contacts. Proposed Quality Indicator of each cell in a module.[2] The system A fourth way is to establish a model Two factors are taken into account integrates solutions for contacting sub- for the occurring error and compensate regarding the quality rating of a module strate and superstrate configurations ref cell AM1.5G the deviation in the I-V curve (mainly at λ . λ λ and the individual cells: 1) the power out- with various contact schemes. Its LED I=SC SR ( ) G()d RS and FF). The constant geometry of the put capacity of the cell; and 2) the power light source supports several cell tech- cells should yield a consistent behavior. output losses due to interaction of the nologies including multi-junction. Due to However, RS and FF are calculated from I=monoSR cell (λ) . G( mono λ)dλ individual cells. The power output capaci- its LED light source, it is easily tunable in the non-linear section of the I-V curve SC ty is often simplified to the use of the ISC irradiance to conform to the proposed and are sensitive to even the smallest as it is strongly related to many light con- approach using a sun-equivalent factor. changes in illumination and cell per- I mono version properties, such as absorber spec- The available irradiance adjustment adds formance. k = SC tral response, light trapping and front- the benefit of testing low light behavior. I ref contact spectral window. Power loss due Since the system records full I-V curves Spectral Response Issue SC to cell interaction can be seen directly in for each cell, both homogeneity of cell The goal of the measurement is to the power output value. Yet it is not the properties and cell interaction are made check single-cell properties in order to maximum power point (PMPP) of each cell, visible. obtain information on cell interaction. This factor k can be interpreted in the but rather the power in the working point A simple semi-monochromatic light following way: the multiple of normed (PWP) that is in effect. The reason is that Summary source (narrow bandwidth source) is suns that the cell in question sees as power loss can be caused by driving a Production stabilization and quality sufficient for this purpose. This can be being emitted by a semi-monochromatic good cell into an unbeneficial working improvement are ever-ongoing tasks in a realized with LED technology that today light source. In this way, multiple light point simply because of its series connec- production line. As product quality in is far more stable and more readily sources become comparable and can be tion with other cells. Therefore, the differ- terms of power output and longevity available than conventional sun simula- adjusted to enable meaningful data in the ence between PMPP and PWP is suggested directly affects product prices, it is of tor light sources. Considering the pop- I-V measurement of multi-junction cells. as a reasonable indicator. The working great interest for maximization. One

<< PREVIOUS PAGE | NEXT PAGE >> 64 | Future Photovoltaics | April 2011 www.futurepv.com | 65 Future Photovoltaics Optimizing Approach for Thin Film PV Production MODULE & PANEL CONSTRUCTION PRINT E-MAIL Click here to return to Table of Contents this article this article approach is to minimize output power About the Authors losses by understanding cell interaction in the module and tracking power losses Jean Randhahn is a freelance consultant. to their root causes. Such an approach His focus lies on process control design and Bill Richardson requires more information than a full application on factory level, process qualifi- Head of Research & Development, SOLON Corporation module-based sun simulator can yield. cation as well as design and development of A cell I-V measurement is required. metrology systems. Dr. Randhahn previously Predominant issues in this kind of meas- worked at Oerlikon Solar in the process con- urement were shown, and power output trol and metrology team. He received his in the cell working point was proposed as diploma in electrical engineering in 2001 a quality indicator. and his Ph.D. in engineering in 2008 from In the PV industry’s never-ending quest to tion. These windows allow for deposition of The main benefit of such an approach the University of Rostock, Germany. compete, the game of wringing cost out of the metalization. Contacts can then be weld- lies in the reduction of optimization every produced electron is played on many ed in whatever scheme desired before the efforts by identifying the most severe Stefan Gritsch joined Dr. Schenk GmbH as levels: module level, system level, cell level, final layer of encapsulant and backsheet is causes of power loss and channeling project engineer and product manager in the etc. Sometimes (often) a cost saver on one added. The theory goes that reduced cell han- resources into meaningful improvement Optical Disc and Solar division in April 2005. level leads to a headache on one or more of dling results in less breakage, but also that projects. He graduated with a diploma in physics the others. For example, using thinner cells this process allows for easy integration of from the Technical University Munich, where means lower cost. It also likely means more miniaturized components and potentially References he also received his Ph.D. in 1997. In Dr. fragility and consequently, higher breakage increased throughput. 1. L. Feitknecht, “Manufacturing Progress Gritsch’s professional career, he has worked rates, a manufacturing headache to be sure. Why stop there though? Taking things for Amorphous Silicon Based Solar as the senior manager of the metrology Or is that an opportunity? In the following one step further results in the – what else – Cells,” Future Photovoltaics, vol. 3, engineering team for a leading manufacturer article, Drs. Kris Baert, Jonathan Govaerts and “i2-module” (integrated interconnect mod- December 2010 of semiconductor materials. Jef Poortmans of imec present their proposed ule). In this case, the actual processing of the 2. S. Gritsch, “In-line Metrology for Thin- elixir to the headache of incorporating ever- cell and junction creation is carried out while Film PV Production,” 25th EUPVSEC, Click here to return to Table of Contents thinner silicon cells into modules. the wafer is supported on the substrate, Valencia, Spain, 2010 Their “i-module” (interconnect module) which the authors argue should allow for the PRINT E-MAIL concept involves replacing standard tabbing scaling benefits associated with parallel pro- this article this article and stringing of back-contact cells with a cessing as well as utilization of even thinner more integrated metalization technique. In cells. They rightly point out that cell sorting essence, cells are adhered front-side down to would then no longer be possible, but then the substrate and then windows are opened perhaps we can leave the cure for that partic- in the rear encapsulant through laser abla- ular headache to a future issue.

66 | Future Photovoltaics | April 2011 << PREVIOUS PAGE | NEXT PAGE >> www.futurepv.com | 67 MODULE & PANEL CONSTRUCTION

Encapsulating Thin Cells: processes would be viable that will be applied for ultra-thin solar cells particularly challenging for metalization whichever way they have been fabricat- in particular. ed. Rear-contacted heterojunction solar The i-module Approach Imec is therefore developing a new cells could especially provide here a module technology for back-contact unique advantage in the sense that the solar cells, referred to as the i-module cell process, the inter-cell metalization K. Baert, J. Govaerts, J. Poortmans PRINT E-MAIL (interconnect-module). The concept orig- and embedding of additional compo- imec this article this article inates as a novel method for fabricating nents inside the module can be com- solar panels starting from < 120 µm thin bined in one integrated cell-module- back-contact crystalline silicon solar integrated component flow. cells. This concept is aiming at an Figure 1 gives an overview of the basic improved reliability due to embedding in process flow of the i-module approach. silicone and interconnection of back- On a clean glass module substrate, an contact solar cells through alternative adhesive layer is applied, and the back- module-level metalization techniques, contact solar cells are placed with the Abstract route toward ever-thinner cells. This as compared to tabbing and stringing of front side toward the glass, so that the A novel approach is proposed for the serves the purpose of cutting cost by the cells and subsequent lamination in contacts remain available for processing module encapsulation of thin silicon reducing the amount of silicon needed, state-of-the-art module manufacturing. at the rear. After this, the full area is cov- back-contact solar cells. The concept dif- especially beneficial in view of a predict- Additionally, it can by extension be ered with an encapsulant, so that the fers from existing concepts by the fact ed shortage of solar-grade silicon in the that cells are mechanically fixed to the (near) future. module glass prior to interconnection. The current industrial approach for The approach is compatible with the crystalline Si PV-modules consists in pro- long-term trend to ultra-thin silicon solar cessing the individual crystalline Si cells, cells and can lead to improved perform- and then stringing them together into ance and reliability. As a next step in the modules. However, according to our own module roadmap, we envisage an evolu- roadmap [1,2] and supported by other tion toward module-level fabrication of roadmaps (see e.g., the CTM-roadmap[3]), wafer-based thin silicon solar cells. back-contacted solar cells made on significantly thinner Si-foils will prevail in Si- Introduction PV on the longer term. As a result, the With crystalline silicon solar cells cur- current cell manufacturing and module rently being the most prevalent type of integration approach will undergo drastic photovoltaic technologies (and for the modifications. The first problem to be predictable future), the production cost tackled is wafer breakage – this could be should still be substantially lowered to get improved by non-contact wafer handling to grid parity and beyond. There are procedures, but today the approach is numerous ways to reduce production speculative below 80 µm. More funda- costs, ranging from lower-temperature mentally, however, warping considera- processing and cheaper materials to high- tions would severely restrict the process er-throughput systems and improved effi- window of most of the process steps – in Figure 1 – i-module Process Flow for Back-Contacted Solar Cells ciencies. Here the idea is to embark on the practice, only virtually stress-free

<< PREVIOUS PAGE | NEXT PAGE >> 68 | Future Photovoltaics | April 2011 www.futurepv.com | 69 Encapsulating Thin Cells:The i-module Approach MODULE & PANEL CONSTRUCTION

cells are fully embedded. The contacts are The potential advantages, as compared absorption. It is also expected to improve Technology then accessed by drilling vias through the to the conventional approach as used in, reliability,[8] because of a lower glass tran- Coating of the different silicone lay- encapsulant material, and subsequently e.g., the SunPower modules,[4] and the sition temperature and Young’s modulus, ers can be done by blade coating, dis- depositing the metalization, patterned ECN approach,[5] are summarized in Table a better UV stability and reduced moisture pensing or screen printing silicones according to the preferred interconnec- 1. Wet coating as opposed to dry lamina- takeup, and a higher heat and flame such as Dow Corning’s PV6010 Cell tion scheme for the cells. Outside con- tion could result in an increased speed resistance, which is interesting from both Encapsulant. The flow has been devel- tacts are provided by welding strings to and throughput, and allow for thinner a processing (e.g., the silicone can with- oped, at first with dummy cells and sev- some designated areas of the metaliza- cells, as any uneven pressure distribution stand soldering temperatures) and safety eral trials, and subsequently demon- tion, and finally, another layer of encap- during lamination could result in break- point of view. strated as proof-of-concept with 2x2 cm2 sulant, possibly in combination with a age. The use of silicones as adhesive and Another distinctive enabler as com- functional 120 µm thick interdigitated back sheet, is applied to protect the met- encapsulant instead of EVA is considered pared with regular c-Si or thin film mod- back-contact cells as well as standard alization and reinforce the relatively weak to be beneficial in terms of optical per- ules is that miniaturized chips could be thickness MWT cells.[10] CO2 laser interconnection points. formance [6,7], due to a reduced UV integrated at the cell level and seamlessly drilling was used to ablate vias to the interconnected with the PV cells. This contacts of the embedded cell. For inter- State-of-the-Art imec Potential allows a seamless integration of minia- connection, feasibility (as proof-of-con- SunPower ECN i-module Advantages turized components that may be a critical cept) was proven through either indus- feature enabling smart PV modules.[9] trially compatible ribbon soldering for Dry lamination Dry lamination Wet coating • Speed, throughput • Thin cell capability EVA foils EVA foils Silicones • Optical performance • Reliability Stand-alone cell-to-cell Module-level Module-level • Reduced cell-level soldering interconnection interconnection handling/processing (tabbing/stringing) (conductive foil (soldering, plating) (breakage) and adhesives) • Thin cell capability

Table 1 – Summary of Comparison of the i-module Approach Against Other Module Manufacturing Approaches for Back-Contact Solar Cells

Figure 2 – View of experimental i-modules. Four small-area back-contact solar cells[11] with a thickness of 120 µm are embedded in a mini-module (left); industrial MWT cells[12] are embedded in 1- Figure 3 – Example of Stress Simulation in a 40 µm Resp.; 120 µm Crystalline Si Solar Cell cell-modules for reliability testing.[10] as a Function of Glass Thickness and the Embedding Glue Thickness[13]

<< PREVIOUS PAGE | NEXT PAGE >> 70 | Future Photovoltaics | April 2011 www.futurepv.com | 71 Encapsulating Thin Cells:The i-module Approach MODULE & PANEL CONSTRUCTION

industrial MWT (metalization wrap- Longer term, we are looking toward Reliability possible on module level. In this con- through) cells, or low-temperature Ag electroplating of Cu. The reasoning Concerning reliability, initial steps cept, the cells are fixed to the glass paste for next-generation IBC (interdigi- behind this is threefold. First, such a type have been taken toward predicting fail- superstrate as soon as possible; so, after tated back-contact) cells. For bringing of interconnection could be integrated ures in modules based on i-module front-side processing at cell level (which out the module power, outside contacts with the cell metalization, as discussed technology. Starting from a so-called must be done either stand-alone or were incorporated by contact welding further in the i2-module concept. The sec- failure mode and effect analysis (FMEA), using carrier substrates). Then the junc- standard Sn-coated Cu tabbing strings. ond consideration would be the improved the issues most likely to occur in the tion is created at the (still-accessible) Initial experiments have resulted in conductivity of bulk Cu (typically also module are related to cracking of the back of the wafer, and base and emitter functional modules with small-area 120 used for the ribbons in standard module cells, cracking of the interconnections regions are contacted, at the same time µm thin back-contact cells, and prelimi- technology) versus screen-printed Ag. and delamination of the silicone layers. also providing an electrical interconnec- nary outdoor tests are promising. Some Thirdly (though not less important), Cu is To predict these possible issues, simulation between the neighboring cells. The impressions of the final result are given considered to be a more sustainable tions of a module model are very help- conceptual flow is shown in Figure 4. in Figure 2, showing the interconnection option compared to Ag, which will proba- ful: By using finite element modeling, Two main advantages can be identi- of four cells resulting in a module effi- bly also translate into a cost advantage in stresses and strains in the different lay- fied for such a flow compared to con- ciency of 16.1 percent with relatively low the longer run.[14] Cu plating has also ers may be predicted. If these values are ventional crystalline Si module manu- loss as compared to the efficiency of the been demonstrated already for PERC-style compared to the material parameters facturing, where cells are first fabricat- original rear-contacted cells. solar cells.[15] (silicon yield strength, metalization ed and only then integrated into the strength and silicone adhesion peel module. The first one encompasses all strength), this gives a good indication of the scaling benefits that may be associ- possible problematic areas in the mod- ated with parallel instead of serial pro- ule. As an example, Figure 3 shows the cessing: throughput, cost, etc. The sec- stress in the silicon cell as a function of ond one is that it offers the possibility the thickness of the glass, adhesive and of processing much thinner cells: As of the encapsulant layer, and as they are supported during processing by depending on Young’s modulus of the the module substrate, the breakage can silicone. In this example, it seems that be expected to be significantly lower, stress in the silicon is minimized by with anticipated higher yields (we pre- using thin silicone layers, silicone with viously already demonstrated embed- a low Young’s modulus, and a thick ding of 40-micron-thin foils[17]). The glass substrate. Another outcome of the critical point for such an approach is modeling applicability is the analysis of the fact that post-process cell sorting is strain in the silicone (which might lead no longer possible, such that the (uni- to delamination at the sides during formity) requirements for the (larger- thermal cycling) or the strain in the Cu area) equipment will need to be even interconnect bridges between adjacent more stringent compared to stand- cells (which might lead to failure by alone cell processing equipment. fatigue).[16] This i2-module approach is targeting an a-Si heterojunction IBC (interdigitat- i2-module ed back-contact) structure, as the a-Si The i2-module approach (for inte- heterojunction has a proven potential Figure 4 – i2-module Process Flow grated interconnect-module), aims at for high efficiency [18] and is compati- carrying out cell processing as much as ble with processing at low temperatures

<< PREVIOUS PAGE | NEXT PAGE >> 72 | Future Photovoltaics | April 2011 www.futurepv.com | 73 Encapsulating Thin Cells:The i-module Approach MODULE & PANEL CONSTRUCTION

that is required for the glass and (sili- itrpv_03_2010_web.pdf 13.M. Gonzalez et al. “Thermo-mechani- About the Authors cone) adhesive used for fixing the cells 4. D. Rose et al. “Development and cal Challenges of Advanced Solar to the glass.[10] Development in this manufacture of reliable PV modules Cell Modules,” to be presented at Kris Baert obtained his Ph.D. from Leuven concept has up to now mainly been with >17% efficiency,” Proceedings of EuroSimE 2011, April 18-20, 2011, University, Belgium, in 1990 on PECVD focused on depositions of a-Si by the 20th EUPVSEC, 2005. Linz, Austria. of thin film c-Si. From 1990 till 1992, he PECVD, and values in the range of 50 5. P.C. De Jong et al. “Single-step lami- 14.J. Bartsch et al. “Copper as Conduc- worked on TFT-LCDs with Mitsubishi Elec- fA/cm2 for the emitter saturation cur- nated full-size PV modules made ting Layer in the Front Side tric (Japan). In 1992, Dr. Baert joined imec rent have already been obtained. with back-contacted mc-Si cells and Metallization of Crystalline Silicon (Belgium), where he managed research and conductive adhesives,” Proceedings Solar Cells – Challenges, Processes development in various areas of MEMS and Conclusions and Outlook of the 19th EUPVSEC, 2004. and Characterization,” Proceedings Integrated Microsystems. Since 2008, he has We have described our vision on the 6. K. McIntosh et al. “An optical com- of the 2nd Workshop on Metalliza- been program manager of silicon solar cells evolution of back-contacted crystalline parison of silicone and EVA encapsu- tion, 2010, pp. 32-37. in the SOLO department. Si solar cell modules. The proposed lants for conventional silicon PV 15.J.L. Hernandez et al. “Application of “i-module” concept is based in essence modules: a ray-tracing study,” Pro- CMOS Metal Barriers to Copper Jonathan Govaerts obtained an engineering on an embedding technology using ceedings of the 34th IEEE PVSC, 2009. Plated Silicon Solar Cells,” Proceed- degree and his Ph.D. in electrical engineering back-contacted solar cells that are 7. B. Ketola et al. “Silicones for photo- ings of the 25th EUPVSEC, 2010, from Ghent University, Belgium, in 2004 and interconnected on module level. A voltaic encapsulation,” Proceedings pp.1479-1483. 2009 respectively. He has worked for imec, straightforward extension of the proof the 23rd EUPVSEC, 2008. 16.J. Govaerts et al. “Progress in i-mod- Belgium, since 2004. Dr. Govaerts first did posed embedding technology allows the 8. M. Kempe et al. “Acetic acid produc- ule: towards improved performance, research on the assembly and interconnection integration of additional components tion and glass transition concerns reliability and module-level fabrica- of microelectronics on and in flexible sub- and interconnections. A next-genera- with ethylene-vinyl acetate used in tion technologies for BC PV modules strates at the associated lab CMST (Centre tion, even more disruptive concept is photovoltaic devices,” Solar Energy based on ultra-thin Silicon solar for Microsystems Technology) at Ghent Univ- the i2-module, which proposes transfer- Materials and Solar Cells, vol. 91(4), cells,” submitted for 26th EC-PV Conf. ersity. In 2009, he shifted toward the SCT ring the majority of the solar cell pro- pp. 315-329, 2007. 17.J. Govaerts et al. “Performance of a (Solar Cell Technology) group within imec cessing steps from stand-alone wafer to 9. J. Poortmans et al. “Linking Nano- new type of module based on back- and is currently working on module integra- module-level processing. technology to GigaWatts: Creating contact solar cells,” Proceedings of tion of solar cells. Building Blocks for Smart PV- the SPIE, vol. 7773, 2010. Modules” Proceedings of the 25th 18.H. Sakata et al. “R&D Progress of Jef Poortmans received his Ph.D. in electronic Endnotes EUPVSEC, 2010. Next-Generation Very Thin HIT Solar engineering from the Katholieke Universiteit 1. K. Baert et al. “Crystalline Si Solar 10.J. Govaerts et al. “A novel concept for Cells,” Proceedings of the 25th of Leuven, Belgium, on strained SiGe-layers in Cells: Applying the Experience from advanced modules with back-contact EUPVSEC, 2010, pp. 1102-1105. June 1993. He then joined the photovoltaics Microelectronics to Improve Effi- solar cells,” Proceedings of the 25th group of imec. Currently Dr. Poortmans is ciency and Reduce Cost,” 24th EUPVSEC, 2010, pp. 3850-3853. program director of the Strategic Programme European Photovoltaic Solar Energy 11.J. Robbelein et al. “Industrial type SOLAR+ at imec. He has authored or co- Conference September 21-25, 2009, passivation on interdigitated back authored nearly 500 papers. Since 2008, Hamburg, Germany. junction solar cells,” Proceedings of he has also been guest professor at the 2. K. Baert and J. Poortmans, “Improving the 24th EUPVSEC, 2009. Katholieke Universiteit Leuven. cell efficiency and reducing costs: 12.A. van der Heide et al. “Industrial applying experiences in microelec- Fabrication of multi-crystalline MWT Click here to return to Table of Contents tronics,” Photovoltaics International, cells with interconnection flexibility Ed. 7, 76, 2010. of 16.5%,” Proceedings of the 24th PRINT E-MAIL this article this article 3. http://www.itrpv.net/doc/roadmap_ EUPVSEC, 2009.

<< PREVIOUS PAGE | NEXT PAGE >> 74 | Future Photovoltaics | April 2011 www.futurepv.com | 75 Future Photovoltaics PHOTOVOLTAIC SYSTEMS PRINT E-MAIL Click here to return to Table of Contents this article this article A Simpler BOS =

Lior Handelsman A Faster ROI VP Product Strategy & Business Development Founder, SolarEdge Technologies Fannon Lim PRINT E-MAIL this article this article Solar Research Pte Ltd., Luna Road LLC

The term “photovoltaic system” spans a This evolution is also gaining from the wide variation. From off-grid to on-grid and latest developments in power semiconduc- from 1 W to 100 MW, the variety of systems tors and components such as new control and usage is incredible. The implementation schemes based on digital processing and new The words “solar energy” and “solar They have a much simpler “balance of of such a system has always included – other switching elements and materials, among power” convey the thoughts of relatively system” (BOS). than the photovoltaic source – cabling and which we can count vertical FETs, SiC and medium- to large-scale projects and A complete PV energy system typically power electronics to harvest the sun’s ener- GaN. To these we can add the leapfrog installations; however, through the use of comprises three subsystems: gy and redirect it to the where it is needed: advances in energy storage devices, initially creativity and innovation, we are now 1) On the power-generation side, a sub- the load. driven by commercial mobile devices, and able to integrate the same concept of system of PV devices (cells, modules In the last several years, we have seen a lately by the automobile industry. solar energy into smaller products that and arrays) that converts sunlight to rapid evolution and revolution in the way PV The following article describes “inde- may have even bigger applications and direct-current (DC) electricity. systems are implemented. The balance-of- pendently integrated solar products (I2SP),” impacts around the globe. 2) On the power-use side, the subsystem system (BoS) developers – driven by the con- which utilizes the advancement made in Quite often, when people think of consists mainly of the load, which is stant need to increase performance and technology, as described above, to offer using solar to generate power for lighting the application of the PV electricity. improve financial gain – have introduced smaller, more integrated and more sophisti- applications or other usages, the thoughts 3) Between these two, we need a third new concepts, methods and components to cated photovoltaic off-grid products to the of huge rooftop spaces or lands are subsystem that enables the PV-gener- the PV systems. modern environment. required to house the photovoltaic (PV) ated electricity to be properly applied energy system. This is particularly true if to the load. This third subsystem is the user is thinking of harnessing the often called the BOS. sun's energy to power electronic devices inside the building that all possess high The BOS typically consists of structures power consumption and demands. for mounting the PV arrays or modules and However, if such a system is set up to the power-conditioning equipment that power outdoor devices, it is not as eco- adjusts and converts the DC electricity to nomically attractive as using “independ- the proper form and magnitude required ently integrated solar products.” by an alternating-current (AC) load. The Independently integrated solar prod- BOS can also include storage devices, such ucts (I2SP) are products that are innova- as batteries, so that the PV-generated elec- tively designed to be stand-alone solar tricity can be used during cloudy days and product solutions fulfilling a specific at night to charge controllers that regulate need. These products can work inde- the power being transferred from the solar pendently and are completely modular. panels to the batteries.

<< PREVIOUS PAGE | NEXT PAGE >> 76 | Future Photovoltaics | April 2011 www.futurepv.com | 77 A Simpler BOS = A Faster ROI PHOTOVOLTAIC SYSTEMS

Compared to a PV system, the BOS for within it. Examples of such products that Another key advantage of integrating I2SP is simpler. An I2SP also consists of a we have integrated are road markers, street solar panels into I2SP is the return on solar panel and a load, but the BOS does lamps and illuminated signages, which are investment (ROI). The ROI for solar not require a power inverter or structures all installed outdoors as stand-alone sys- becomes more attractive as compared to to mount the PV arrays. All the compo- tems serving a specific purpose and need. using PV systems. Currently, a PV system nents are integrated within the product Using a PV system to power outdoor takes typically 10-12 years to break even; itself, and the product will be able to self- devices such as lighting applications is however, by using I2SP, the ROI can be as generate electricity and power the load inefficient and ineffective as compared low as two years. This is largely due to a to using I2SP, for the following reasons: simpler and integrated BOS, where the a) I2SP uses DC electricity throughout the user will save costs from underground BOS, which minimizes losses as com- cabling, long cables and lower efficiencies. pared to using inverters in PV systems. With deployments of solar made easi- er and cheaper from these innovatively i.e., The I2SP generates DC electricity that is Figure 3 – Luna Eyes Illuminating designed I2SPs, Luna Road believes in sav- a Floating Bridge With no AC Supply stored in batteries to power DC loads, such ing lives, protecting and preserving our as LED lightings, without inverting the environment, in addition to beautifying electricity to AC, which will have some nighttime roads all around the globe. The losses. Luna Road products will deliver safety, beauty and eco-friendliness wherever b) Having a PV system on the rooftop or they are installed. The future of nighttime on open land means long cables are illumination is here! required to be routed to power the out- Figure 1 – Luna Road Markers Illuminating a Private Condominium Driveway door devices at different locations. This will usually cause cable losses. To minimize such losses due to voltage About the Author drop, the simplest and most obvious way is to use larger cables, which in Fannon Lim received his Ph.D. and B. Eng. Figure 4 – Luna Sign – Unlighted (left) turn yields higher costs. vs. Lighted (right) in Daylight degrees from the University of Glasgow in 2003 and 1999, respectively. His interest in c) Being modular means that I2SP can be entrepreneurship and green energy led him added at a later stage as compared to a d) Being stand-alone means that I2SP can to join Solar Research as chief operating fixed PV system. easily be added conveniently at any officer, and head the Asia headquarters of location. Often underground cables are Luna Road in Singapore. i.e., When a PV system is designed to laid for the predetermined lamps; to power 50 lamps, and within a year it is add new lamps at a new location Click here to return to Table of Contents decided that another 20 lamps are would incur high costs of routing new required, the PV system may not have cables. I2SP can operate individually PRINT E-MAIL this article this article enough power generated to power all 70 without any cables at any location. lamps. However, with I2SP, you can fairly simply add another unit to obtain the extra Figure 2 – Luna Eyes Illuminating lighting applications. a Public Walkway

<< PREVIOUS PAGE 78 | Future Photovoltaics | April 2011 www.futurepv.com | 79