Relative to silicon, thin film is even in decline in terms of market share. And take away the might of First Solar, and there is no disguising a malaise of sorts.
So what do experts put this down to?
“The efficiencies of thin film still lag those of crystalline silicon,” said Dr. Roland Schindler of the Fraunhofer Center for Sustainable Energy Systems. “Availability of materials is another big concern.”
He rolled out the hearse with his analysis of the long term prospects for thin film staples like Tellurium (Te) and Indium (In), as well as the toxicity of Cadmium (Cd).
“There just aren't enough of these elements to sustain a high volume of thin film usage and GW capacity installation,” he said. “If you end up having to do direct mining, this means CIS and CdTe are no better in kg per kWh compared to polysilicon.”
His solution is simple. Instead of squandering a limited supply of Te and In in power generation while trying to compete toe to toe with silicon and other technologies, he recommends the supply be preserved for value-add areas such as in building integrated PV (BIPV) where aesthetics might play a greater role in decision making.
Avoiding rare earths
Not everyone is so bleak about thin film's prospects, however. Gregory Ashley, Vice President and COO of Solar Frontier Americas, countered Schindler's arguments by saying that the thin film market actually doubled in 2010 – it is just not growing as quickly as silicon. He said that both Copper Indium Selenium (CIS) and Copper Indium Galium Selenium (CIGS) held steady at 2% of the market last year according to GTM Research. He added that aggressive recycling of In and greater care during manufacturing were going a long way towards reducing waste.
That said, he agrees that rare earths aren't the only way forward in thin film. In fact, his company is working with IBM on a copper zinc tin solar initiative. But he sought to distance his company's thin film efforts from those based on CdTe. A small amount of In is harnessed in CIS/CIGS, and almost all of it is recovered as part of the production process, he added.
According to Ashley, CIS and CIGS have over 13% efficiency, making it the leader in the thin film category. Solar Frontier Americas has measured 17.2% on a smaller sub-module. Further benefits include the fact that it includes no toxic components such as lead and Cd, and has a broader spectral response than mono-crystalline, poly-crystalline and CdTe. He added that CIS and CIGS are not polarity sensitive, don't have durability issues and are not bothered much by light induced degradation (LID).
“There is no initial LID in CIS although it gains 5% to 10% in the field due to the light soaking effect,” said Ashley. “But this is still better than poly-crystalline and mono-silicon in this area.”
While poly-crystalline silicon may claim higher efficiency, field testing by Solar Frontier Americas indicates that CIS outperforms it in raw kW output. Similarly, it outperforms CdTe by a few percent, though the gap narrows in high temperatures during three months of testing.
“4.76 MW of CIS delivers the same kWh as a 5 MW poly-crystalline system,” said Ashley.
Dr. Subhendu Guha, president of United Solar Ovonic, is another who sees a bright future for thin film. He outlined its recent expansion in usage – annual production has gone from 198 MW in 2007 to 1.3 GW in 2010. Manufacturing costs, he said, are being reported at less than US$1 per watt by at least one Chinese firm.
“The key challenge for thin film is the improvement of conversion efficiency,” he said. “Researchers worldwide are working on this issue.”
First-generation cells, he said, were single junction, had low efficiency and suffered from a lot of LID. That led to the introduction of more advanced designs such as double junction, dual gap and triple junction. While single junction is 9.3% efficient, double junction ups that to 10.1%, dual gap to 12.4% and triple junction to 13%. United Solar Ovonic uses triple junction thin film cells made from amorphous silicon and germanium. The cells absorb red, green and blue light through three different layers each with varying amounts of germanium. Nano-crystalline silicon has replaced silicon germanium as it is more efficient, gives better absorption of infra-red and there is no LID when used in the bottom cell of a multi-junction structure.
“This design set a 16.3% world record for initial cell efficiency,” said Guha.
That may only be the beginning as nano-crystalline technology remains in its infancy. Challenges include how to improve the quality when it is deposited on a textured substrate. To address this, United Solar Ovonic has devised a method of hydrogen profiling as well as a cathode design to raise deposition rates. Questions remain to be answered such as whether it might be best to use nano-crystalline silicon germanium as the bottom cell, and how to best engineer the back reflector.
“With a nano-crystalline back reflector we have demonstrated 16.3% efficiency so far,” said Guha. “With the active participation of international researchers, we will move towards 25%.”
Better glass, material alternatives
Upgrades to thin film are moving forward is several areas. Professor Tonio Buonassisi, principal investigator MIT Lab for Photovoltaic Research, for example, was critical of the wastage in silicon. With wire sawing losing about 50% of the raw material to sawdust, he noted, it was no wonder that it costs around US$1.70 to manufacture silicon today.
“Thin wafers without kerf and more efficient cells can mean that we might be able to achieve a rate of 44 UScents for silicon manufacturing,” he said.
His emphasis was on the need for scalable and abundantly-available materials. While silicon has got plenty of the R&D attention, he sees a lot of potential in such areas as FeS2 and SnS. Early success with silicon had led researchers to focus on that one element. He laid out several other possible materials and urged others to invest the time and energy in these areas to see which will bear fruit.
While they may appear unattractive at the moment, they have not gone through the same learning curve as silicon, to see just how realistic an alternative they might be. Only time will tell. He hoped that one or two could play a role in reducing manufacturing costs while boosting efficiency.
Cuprous oxide is another possibility for thin film, he added. Buonassisi said MIT is working on improving its current level of 4% efficiency in various ways. But there are other options.
“Tin Sulfide is another with high promise which is undergoing a lot of testing,” he said.
He cautioned researchers to avoid obscure or extremely rare/expensive materials in their search for answers. To make his point, he displayed the periodic table and colour coded many abundant elements that could be employed.
“We have to be focused on earth-abundant materials, explore all the options and in the meantime, continue to use silicon.”
As well as better materials, gains can also be achieved through improved glass design and manufacture. “Innovation in glass can impact all fields of PV,” said John Duke, the Business Development Director for PV Glass Technologies at Corning. “If you take the thickness below 2 mm and reduce the weight, you can increase performance.”
Standard soda lime glass (as used in windows everywhere) is only 10% efficient. Corning engineered glass, on the other hand, ups that to the 12% to 13% range. The company makes thin film glass for CdTe, silicon-tandem and CIGS. Over the past 10 years, the holders of record efficiency for CdTe have utilised Corning glass which has fewer impurities, he said.
In silicon tandem thin film, for example, Corning provides textured glass that enhances light trapping to boost efficiency. According to Duke, this means five times more light scattering compared to soda lime.
In Part 5, is crystalline silicon expected to be the mainstay PV technology for some time to come?
About the author:
Drew Robb is a graduate of the University of Strathclyde in Glasgow. Currently living in Los Angeles, he is a freelance writer focusing on engineering and technology.