What thin-film PV has in common with Mark Twain
September 17, 2010 - When asked by a reporter about his illness, poverty, and possible upcoming death, Mark Twain replied: "James Ross Clemens, of St. Louis, a cousin of mine, was seriously ill two or three weeks ago in London, but is well now. The report of my illness grew out of his illness. The report of my death was an exaggeration." Lately there have been whispers about the death of amorphous silicon (a-Si) and micromorph, along with hints about other thin-film PV technologies -- but as with the illness and death of Mark Twain, these are exaggerated.
Just a couple of years ago, with demand soaring and polysilicon in tight supply, thin films were considered the future of photovoltaics. It was assumed back then (the recent olden days) that the potential of lower cost would outweigh higher efficiency. Companies looking to enter PV from industries with flatter growth profiles saw thin films as the future. Investors (and government agencies) wanted to invest in new and exciting technologies, viewing higher-efficiency crystalline silicon as having had (pardon the pun) its day in the sun. (Note that thin-film technologies have been around for ~20 years.) Forecasts of thin-film market shares at 30%, 40%, and even 60% of the market were commonplace, and woe to anyone who raised a cautionary point or two.
The primary cautionary points are these:
- All PV technologies chase the twin goals of higher efficiency and lower manufacturing costs. Crystalline leads in efficiency and are getting cheaper to manufacture. In the efficiency race, crystalline wins. In the low-cost manufacturing race, First Solar -- oops, CdTe wins.
- When higher-efficiency crystalline modules are cheap, thin films must be (on average) 12% cheaper -- meaning constrained margins. And, when c-Si is priced aggressively cheaper the majority of thin films are uncompetitive.
Many investors leapt enthusiastically into turnkey manufacturing expecting to be ramped up to full production in six months and sold out into the foreseeable future. All investors with these expectations were (and likely still are) disappointed. During the 2004-2008 period customers lined up to buy product and everyone reported sold-out conditions -- whether or not they had produced even one kilowatt of PV. What was purchased was essentially vaporware -- an old software term for a technology or product which is announced before (or during) development and may never be released. Investors lost faith, companies lost funding -- and with the rise of utility-scale, multi-megawatt investor installations came a new term: bankability. The moral of this story: don't make promises you can't keep to the people with the checkbooks, even if they are made in the spirit of true-believer-ism.
To turnkey or not to turnkey: Machines rule
With Applied stepping back from its thin-film turnkey unit and Oerlikon continuing to lose money on its offering, it would be easy to pronounce thin-film turnkey manufacturing dead. This may not be the case, however (remember the report of Mark Twain's death). At the start of the turnkey craze, many long time PV participants warned that historically solar equipment has been sold one machine at a time and that at some point it is part proprietary black-arts, part science, part engineering, and part manufacturing.
A feed-in-tariff (FiT) boom in demand for solar technologies in Europe encouraged manufacturers of large LCD panels for televisions to assume that the transfer of technology would be simple, and that fully commercialized amorphous and micromorph manufacturing capacity would soon be efficiently and profitably functioning worldwide. Well, not quite. Photovoltaic technologies take time (and a lot of it) to progress from R&D -- lab cell records are just a starting point, and are themselves not trivial -- through pilot scale production to commercialization. From pilot-scale to commercial production is a long slog through repeatability, and there are no shortcuts. In the relatively short (yet at the same time, long) history of the PV industry, many have entered from other industries believing that they had a shortcut -- and all these shortcuts turned into long journeys.
Currently the most successful thin-film turnkey manufacturer is First Solar, and it took ~20 years (and some name changes) to perfect its proprietary manufacturing process -- which it has successfully replicated in several facilities. The successful replication and rapid ramp-up of equipment is the definition and the key to a successful turnkey business, and it takes a lot of time to make it look this easy. Newcomers to the PV industry, who entered without really understanding what incentive-driven meant, found themselves adrift in a volatile, start-up industry with exacting technology requirements - and with the realization that it could be years before market entry. In the rapidly changing incentive environment, the market you delay entering today will surely be a completely different one tomorrow.
With two of Applied's customers in bankruptcy protection, the future of turnkey thin-film manufacturing appears foggy. Yet, with the entry of Manz into CIGS turnkey manufacturing, and other companies continuing to market thin-film lines and equipment, it is too early to call thin-film turnkey dead. As a caution, in an industry that traditionally buys one machine at a time, machines (not lines) still rule. Though turnkey promises to offer newcomers to PV an easy entry point, in the end if you cannot commercialize rapidly and profitably it is an awfully expensive Lego set.
PV manufacturing -- particularly thin-film manufacturing -- can be a brutal arena, particularly when crystalline prices are at today's bargain basement levels. There is, however, no substitute for learning the hard way. Back to Twain: "A man who carries a cat by the tail learns something he can learn in no other way."
Will thin films return to single-digit market share?
There is no sugarcoating the lesson currently being experienced by all thin-film manufacturers (except low-cost leader First Solar): when the price of crystalline technologies falls, the price of thin films must fall even lower. The reason that thin films must be priced cheaper than crystalline products comes down to efficiency. As said in politics, particularly around elections: it is the economy, stupid. In solar, it is efficiency, efficiency, efficiency. Thin films make up for the area penalty (more land/roof, more hardware) suffered because of lower efficiency with lower price points for modules. Over time, the average difference between the price of a thin-film module and a crystalline module is around -12%. New entrants chose to believe that the promise of lower-cost manufacturing was the only requirement for success, and they were by and large disappointed.
|Technology shipments through 2014.|
Because of downward price pressure brought about by decreasing incentives, and growth of the utility-scale market (which values cheap kilowatt hours over other attributes of solar), thin films are in for a difficult ride. Manufacturers must deal with bankability issues (brought about partly by the failure to commercialize as rapidly as promised), and with strong competition by inexpensive and high-quality crystalline silicon technology. Manufacturing costs must be brought down in order for thin films to have comfortable (or even slightly uncomfortable) margins. In the meantime, and likely for at least a couple of years, the dominant thin-film manufacturer will be First Solar, continuing to enjoy lower costs and stronger margins. Over the next few years, thin-film market share (of total shipments) will likely hover around 13%-15%, with crystalline continuing to dominate for the foreseeable future (see chart above).
A few years ago, it was assumed that thin films belonged in large fields, while crystalline would dominate on the rooftop. This assumption does a disservice to both technologies -- but particularly to thin films, by broadcasting that they are not competitive on rooftops. This is categorically untrue, and thin-film manufacturers should work to overcome this unhelpful marketing. In the future, BIPV will help buoy up thin-film market share, but BIPV has struggles of its own. Figure 1 presents a forecast for thin film technologies and crystalline technologies. The solar industry is moving into multi-gigawatt deployment, and thin-film technologies will play a part in the future of the industry.