2011 was hardly a vintage year for thin-film solar in the US. Doubts about revenue-ready technologies based on copper, indium, gallium and selenide coalesced around the Solyndra bankruptcy in August. In December, First Solar decided it could not replicate its success in cadmium telluride, and shuttered its CIGS division.
So far this year, companies such as Stion and SoloPower, both in San Jose, California, are forging ahead with new manufacturing plants in Korea, Mississippi and Oregon. MiaSolé's future remains secure for now as the CIGS startup searches for a company to partner with and two of Silicon Valley's largest venture capital companies, Kleiner Perkins Caulfield Byers and VantagePoint, are unlikely to watch their thin-film darling get shot down in flames, taking around $350m with it.
But have venture capitalists spread their bets too widely when it comes to this particular emerging technology and inadvertently created barriers to high volume manufacturing? At least one industry insider thinks so.
Markus Beck was Solyndra's chief scientist until 2007. First Solar snapped up the former National Renewable Energy Laboratory scientist to lead the $37.7m R&D facility in Santa Clara, California.
“There are a few things we haven't historically done right,” Beck told a PHOTON conference in San Francisco last week. “I want us to be frank about the mistakes that we've made in the past and preconceptions that we have that are often times wrong and not data driven. If we addressed those reasons, we would have been a lot smarter and could actually move CIGS where it belongs.”
VC-backed startups had been experimenting too widely with sputtering, evaporation, ink printing and substrates such as glass, steel and polyimide, Beck said.
“These opportunities are actually a curse,” he said. “The fact that you can make CIGS in many different ways has misguided us and we have defocused our efforts on too many R&D projects, too many reinventions of the wheel.”
Beck claimed to have patented an optimal covaporation technique on glass substrate that could scale more cheaply and easily to high-volume manufacturing. He said that at the time of its collapse, Solyndra still had the best technology in the industry with a 20.5% lab efficiency using a 3-stage covaporation technique.
“The data says that the highest efficiency devices are made by covaporation or the two-step process – so why did we defocus our efforts and had this notion that we had to come up with new processes to CIGS?”
“To some extent it's the way this whole thing is being financed. People wouldn't give you money if say, hey I'm going to covaporate, it's a proven process. But if you say I have this sexy new process, I don't know whether it's going to work, they would give you money. No one was in a position to evaluate whether can you actually manufacture it or how long is the research phase of the R&D is going to take. We've diluted the money that was available to us.”
Solar Frontier, Global Solar, Solibro and Avancis have opted for an evaporation process while Stion and MiaSolé use sputtering. Having won a Department of Energy loan for $197m last August, SoloPower is now under pressure to prove out its pioneering electrodeposition process on flexible steel at its new 300MW facility in Portland.
Pallavi Madakasira, a CIGS engineer who left MiaSolé in 2006, and is now a Lux Research analyst said that technological diversity is good for the industry.
“As a technologist I don't think there are too many processes that are being used to develop CIGS,” said Madakasira. “Evaporation and sputtering are probably going to be the two more mainstream methods that will get companies to where they want to be. It's a question of time, and not everybody is patient.”
But the investor profile for CIGS has already begun to alter from VCs to conglomerate acquisitions or strategic partnerships. Global Solar now sells its shingle PowerFLEX Technology to Dow and Avancis last year announced a joint venture to produce up to 400MW by 2015 with Hyundai last year.
Madakasira said: “Strategic partnerships will more likely shape the future of the CIGS industry because they are not looking to have an IPO. They have the patience because they've developed technology before and want to transfer their knowledge. They are more focused on technology improvements as opposed to an IPO or acquisition.”
Contrary to conventional wisdom of VC-backed startups, technical teams at large companies may also be more ready to adapt to the rapidly changing dynamics that are particular to the CIGS market, said Beck.
“CIGS has been plagued. The biggest problem in CIGS is with discipline and funding. It's not a technology problem it's a financing problem – it's a curse. Once you go to a VC, they want to hear that this is the holy grail, this is a gamechanger. You can't change course. If you do, you're doomed. You have to stick with the idea.”
Beck also claimed that before leaving Solyndra, he implored the chief executive, Chris Gronet, to ditch the cattle grid design. But Gronet apparently told him that VCs had bought into the design that ultimately proved an expensive novelty that added no efficiency.
“I told him that the design was flawed he freaked out because there was too much money invested in the company,” said Beck. “I told him that we should switch to flat glass – but that would have killed the idea.”
Madakasira authored last month's report from Lux Research, Sorting through the Maze of CIGS Technologies: Who Will Cash in on the Breakout Year, which identified winners who have already peeled away from the pack, while the losers would be those who had not gone far enough beyond the science project stage.
Madakasira said they had tracked around 18 CIGS companies and five came out on top. “Solar Frontier is clearly the outright winner now in terms of scale,” she said.
Solar Frontier is owned by Showa Shell Sekiyu, a Shell subsidiary, which last year opened its 900MW CIS (copper, indium and selenium) plant in Miyazaki, Japan.
“It's a huge advantage to be a subsidiary of a large company or in a strategic partnership with a large conglomerate, especially given the dynamics in the solar industry. In CIGS there is a huge amount of risk and liability because the technology is new and is largely unproven on a commercial or industrial scale.”
Madakasira is convinced that there is still room for venture-backed innovation in CIGS, such as AQT's announcement last month of a sputter deposited copper-zinc-tin-sulfide (CZTS) process. But even though AQT recently secured $18.7m in venture funding, investments in CIGS may become an increasingly attractive technology proposition for large companies.
But changing tack and retooling a CIGS process once an idea is sold to VCs is very difficult, she said. “I agree that you can't go back to a VC. When you bring something innovative to the table and you want to switch all that out go with something that's proven and already been done – there is clearly an issue with that.”
Eighty-five percent of the solar market will continue to be dominated by silicon until at least 2025, according to IHS Emerging Energy Research. Dominance by incumbents makes it all but impossible for newcomers to enter the polycrystalline market, forcing them to seek investments in emerging technologies such as CIGS, which Lux Research estimates will reach 2.3GW with $2.35bn in revenue by 2015.
“Our outlook is bullish in terms of the technology and its promise,” Madakasira said. “Forget about the cost and anything else associated with market trends – the technology in itself has the potential to compete very effectively with multicrystalline silicon.
“There's a huge learning curve to get there and only a handful of companies from around the world are going to be able to do it. It took First Solar 10 years to get there with cadmium telluride and CIGS is only really in its 5th or 6th year of development.”
Felicity Carus will be a regular blogger on PV-Tech.org.