A word of caution for next-gen solar startups


It’s been painfully apparent that many of the well-funded startups building next-gen solar tech have yet to ship solar panels in any large volumes and have missed their targets. Even the folks in the solar industry are starting to take each other to task publicly on this issue.

At the Photon Thin Film conference in San Francisco last week, well-known researcher Rommel Noufi, who heads the thin film research at the National Renewable Energy Laboratory, challenged startup AQT Solar over promising and under delivering.

AQT Solar’s CEO, Michael Bartholomeusz, gave a pretty standard talk about the company’s shift from using the material combo CIGS (stands for copper-indium-gallium-selenide) to using CZTS (copper-zinc-tin-sulfide) for its solar cells. The Silicon Valley startup hasn’t met the goals it set a few years back with the manufacturing of CIGS solar cells, but it recently raised $18.7 million and is now charging ahead with its new CZTS technology instead. He promised to commercialize the CZTS technology in 2013.

And after Bartholomeusz spoke, Noufi stepped up to the microphone that was set up for a Q&A, and promptly put Bartholomeusz on the spot and cautioned about making empty promises: “The only advantage of CZTS is when indium and gallium disappear. We have to be careful about not giving another technology a bad name.”

In his own defense, Bartholomeusz, pointed out that the amount of money that has been pumped into developing CIGS technology has yet to make a significant dent in the market. He said: “I don’t think any commodity since the dawn of time hasn’t responded to supply and demand. So with indium and gallium there will have an issue.”

Failure of CIGS

The exchange ended there, but the feeling that the CIGS technology, which has attracted more than $1 billion in venture capital, hasn’t become a key player in the solar market lingered. Markus Beck, who until last December was heading the secretive CIGS development lab at First Solar (s FSLR), gave a talk at the conference that outlined why the CIGS community hasn’t achieved success. 

Some of the missteps involved experimenting with too many different ways to making CIGS panels and falling short on figuring out how to run production lines efficiently, Beck said. He noted the reliance on venture capital is partly to blame – VCs want to bet on cutting-edge technology because it could deliver a huge return if done right. But that in fact encouraged the development of fancy but expensive products that can’t compete, and Solyndra is a prime example, said Beck, who was the lead scientist at Solyndra before joining First Solar in 2008.

“If I say I have a new sexy process, then people give you money. We have diluted the money that is available to us,” he said.

Aside from AQT, other CIGS companies that still hope to become major players in the market include MiaSole, Nanosolar, HelioVolt, Stion, SoloPower, Soltecture, Global Solar, Solibro and Avancis. Several of them have brought in new CEOs in recent past. The biggest CIGS solar manufacturer is Solar Frontier, which brought online a 900 MW factory in its native Japan last year. The company last month announced a deal to supply up to 150MW of solar panels for a project in California. That project is now touted as a sign that demand for CIGS technology is there. Now if only other CIGS solar manufactures could work fast enough to meet demand.

Photo courtesy of Micky Aldridge via Flickr and Nanosolar



No worries, the Chinese will buy our CIGS companies for pennies on the dollar then get money from their state owned investment banks to make their own silicon based pv industry obsolete. It’s not even a secret, it’s in the latest five year plan for you to read. If you read Mandarin.


I don’t think anyone saw the price of 7N silicon falling as low as it has. At this point, balance of system costs are 2x or more module costs. The real name of the game right now is to increase efficiency to reduce BOS costs.

IMHO, nanosolar has the edge on all of their CIGS piers (assuming they can actually ship product). Their larger panel size combined with higher voltage and simplified assembly helps to reduce BOS costs.

With that said, it’s going to be very difficult for CIGS to compete so long as silicon stays cheap. The fact that monocrystalline panels with ~26% efficiency are on the near horizon makes things just that much worse.


CIGS is done. Just going to take awhile for everyone to realize it. Cost of crystalline solar continues to drop, efficiency is better, technology is more reliable and robust, with a multi-decade track record.

Ucilia Wang

Sigh. I’m trying not to think that way. Certainly the number of successful CIGS companies will not be nearly as high as some investors believed. First Solar’s pulling out of CIGS was a big hit for the CIGS sector — the company has the high-volume manufacturing experience to push CIGS’ adoption, and that experience is sorely missing in many startups.

Vincent Kapur

Consider the following:
1) c-Si has taken in way more $$$ in development over the past 4 decades than CIGS, not including the ability to leverage tech investment inSi-use in semicon
2)Tier 1 c-Si still requires >$0.60/Wp in non-Si costs, despite economies of scale. Thus if CIGS can beat this cost, c-Si will not be competitive with poly-Si priced at $0/kg
3)Yield is critical to achieving low costs for CIGS. c-Si production was expensive but achieved low cost by pumping $$$ into capacity expansion. This strategy hasn’t worked for CIGS players because expansion with poor yield is counter-productive. First Solar has shown that low double digit efficiency can compete on cost by effective EoS. Therefore, it seems apparent that companies which have yet to scale up are not ready to demonstrate tight yields close to their reported champion modules.
4) Nevertheless, CIGS is not close to being ‘done’ because c-Si cost structure is so weak it can’t compete with itself. Who has a strongly positive Operating Profit Margin in pure-play c-Si module manufacturing (not counting in-house EPC) today?
5) First Solar has a strong background in Vacuum processing, which is much easier to control for binary CdTe than with the four elements of CIGS. Non-vacuum faces a strong negative bias among the industry, but is really the only road forward for CIGS at this stage to achieve cost-targets.
6) Drop-in CIGS adds materials & CAPEX costs along with process complexity which negatively impact competitiveness. Solar Frontier, Avancis, Stion, Q-Cells and others have shown that high quality can be achieved on full-size in contrast to the oft-repeated claims that drop-in cells improve yield and performance. A lot of VC money has gone to the drop-in model.
7) Too much $$$ has gone to replicating complicated lab processes to achieve high efficiencies at a larger scale, rather than developing processes from the outset which are conducive to cost/yield in volume production. A good example is buffer layer deposition, which has no strongly cost-competitive standard solution for mass production. Waste processing can be more expensive than the deposition process itself. These under-optimized operations become cost-drains at scale. Careful engineering before committing $$$ to capacity ramps can scrub out substantial amounts of cost and provide clear paths to sustainable cost advantage.
8) Finally, 2012 will demonstrate that more c-Si will be shuttered than CIGS efforts by a far greater number, technology track record aside. There is plenty of reason to be bullish about CIGS mass production if the right combination of process development experience & robust tool design can be focused on achieving optimal cost&yield.
(Full disclosure: I work for ISET, Inc.)

Comments are closed.