Solar PV industry in India: Strategy for success
To avoid repeating the wafer fab fiasco for solar PV too, government policies to promote it must be realistic
by Dev Gupta, CTO, APSTL
Scottsdale, USA & Bangalore, India, September 23, 2008
Source: CIOL.com
http://www.ciol.com/Semicon/Solar/PV/News-Reports/
Solar-PV-industry-in-India-Strategy-for-success/23908110718/0/
Nearly four years since promoters first descended on India to hawk multi-billion dollar wafer fabs for semiconductor chips and for nearly as long since local booster associations (composed mostly of software/design types with rather thin credentials in expertise e,g. physics or materials science, critical to semiconductors) jumped on the bandwagon to advocate wafer fabs, not a single new fab has come up anywhere in India!
Even the Government Ministries responsible for facilitating the belated start of semiconductor manufacturing in India seem to have focused solely on the financial aspects (e.g., subsidies), and at the same time, underestimated the overwhelming importance of securing scarce technical knowhow that still dominates the viability of this most high tech of industries. The fact that nanoscale devices are already in production at the latest wafer fabs seems to have escaped them as well.
Instead of a parallel underwriting of technology research at various Indian laboratories and institutes to accelerate the implementation of basic silicon semiconductor technology into fabs, they have decided to dabble in fashionable longer term technologies like compound semiconductors and nanotechnology with no compulsion to produce any tangible results in the near term.
Just like computer chips, solar photovoltaics too use semiconductors as the active material to convert the sun’s rays into electricity. The complexity of processing these materials and the need to squeeze the maximum performance out of the photovoltaic devices thus made should not be underestimated. In order to stay viable, the private industries embarking on manufacture of solar PV in India will have to export their products for some time to come. Relying on technology and tools bought off-the-shelf will not be a guarantee to remain competitive in the international market.
Compared to India, China has higher labor productivity and they are already far ahead in solar PV production. And then, there is the fact that as yet electricity generated by solar PV has NOT attained grid parity (in fact, it is thrice as expensive) and to make them self sustaining (i.e., without having to depend on government subsidies ) the cost of solar panels must be reduced by a factor of 3.
There is an intense effort worldwide to develop technologies to do just that. Thus, the technology for solar panels is in a state of flux and many competing technologies are being developed in parallel.
Need for realistic policies
To keep up with this dynamic scenario continuous improvement fueled by domestic R&D capability will be critical. This can be ensured only if a parallel national research and training program on solar and alternative energies is launched without further delay at competent scientific laboratories and technical institutes of India. To avoid repeating the wafer fab fiasco for solar PV too, government policies to promote it must be realistic and should be formulated not only by the technical bureaucrats entrenched in Delhi ministries or subsidized government laboratories but also include experts from the semiconductor manufacturing industry.
The predominant technology for building solar photovoltaic cells and modules today uses crystalline silicon wafers made from polysilicon. The performance of the cells depend on the quality of the silicon used e,g. the degree of chemical purity, even how perfectly the atoms are arranged in a wafer. The most perfect silicon, called single crystal semiconductor grade silicon can convert over 22 percent of the sunlight falling on it into electricity. But more common are cells made from polycrystalline silicon wafers that have some imperfections due to the cheaper manufacturing process used and yield conversion efficiencies between 15-18 percent.
To make polysilicon, first metallurgical grade silicon is gasified, then purified and next deposited as chunks of solid polysilicon. The polysilicon chunks are next melted in an ultra pure environment and cast into ingots ( either as a perfect but expensive single crystal or a cheaper and imperfect poly crystal ), then sawn into wafers and polished. These wafers are next processed using steps common in wafer fabs for chips to make photo voltaic cells. The processed cells are tested and assembled into modules.
Approximately 10 gms of silicon is needed to build modules capable of generating a watt of electricity under the strongest sun. Polysilicon is expensive and the modules thus built cost up to $ 4.50 per watt of power generated. At present, there is a polysilicon shortage in the world which has raised their price from $ 40 per kg to $ 100 per kg even for long term contracts. A new polysilicon plant of annual capacity of 2,000 T py (good for modules that will generate 200 MW of power) would cost $200-300 million and take two to three years to build and stabilize production.
To reach grid parity, modules should cost below $2.00 per watt, a range of new technologies are under development to meet this goal. Some minimize the use of silicon (e.g., by making cells out of thin film amorphous silicon using only about a hundredth as much as a wafer), some use no silicon at all (e.g., make cells from thin films of metallic alloys of Cadmium and Tellurium or CdTe, Copper-Indium-Gallium Selenide or CIGS, organic dyes, etc.).
Attaining grid parity
In addition to consuming a lot of expensive silicon the traditional wafer route is also labor intensive as it requires quite a bit of wiring and mechanical assembly (which is why most modules by this route are now assembled in China and some even in Bangalore). So, another thrust in reducing the cost of solar modules is to design them to require minimal assembly, the same idea that led to the spectacular success of integrated circuits. To this end continuous roll to roll processes have been developed and put into production, the leading example of which today is the technology developed and used by United Solar (Michigan, USA) to deposit thin films of silicon and silicon-germanium alloys on sheets of stainless steel.
What would come as a surprise to most Indians today is that this technology had its genesis in Calcutta and United Solar of USA is led by Dr. Subhendu Guha with a PhD from Calcutta University! Compared to the traditional thick silicon wafer solar cells, thin film solar cells consume much less precious semiconductor material and on a per watt basis cost only 60-70 percent as much.
At present, First Solar Corp. of Phoenix AZ, USA, the world’s largest thin film solar cell company, uses its proprietary cadmium telluride thin film technology to deposit solar cells on glass sheets and claims the lowest cost in the world ( $ 1.25 per watt ). They will soon ramp their production from 300 MW py to 1 GW py. However, be it silicon or non-silicon ( like Cd Te or CIGS ), in the absolute efficiency scale thin film cells in production are somewhat deficient as due to various unavoidable imperfections in crystal structure, their solar conversion efficiency, seldom exceeds 10 percent, less than half that of the best silicon wafer solar modules.
The production processes are also more sophisticated and thus cost more (turnkey plants made by Applied Materials for thin film silicon cells on glass sheets may cost over 3 million USD per MW py capacity). APSTL, the authors company in Scottsdale, AZ, is developing silicon PV cells that will consume only a fraction as much polysilicon as the traditional wafers yet have conversion efficiencies well above that of thin film cells.
The highest solar conversion efficiency today is produced by cells made of multi-junction compound semiconductors. Spectrolab, located north of LA, makes cells that show efficiency of up to 40 percent (nearly double that of the best silicon wafer cells). However, due to the slow and exact process used to grow these semiconductors they are extremely expensive ($2/sq. mm of die ) and most frequently can be used only for defense and satellite applications. Mars lander robots, including the one that recently confirmed water in Mars, use these multi-junction solar cells to generate enough electricity to operate a power shovel to dig into the hard Martian soil or run a whole chemistry lab on board!
To make these very expensive multijunction solar cells affordable for commercial applications they need to be integrated with concentrator optics (mirrors with cassegrain optics) with attendant cooling systems as well as sun-tracking devices.
Maintaining competitiveness in Indian solar/PV
The above thumbnail sketch should be enough to convince both government policymakers and entrepreneurs that solar PV is an industry deeply rooted in physics and hardcore sciences and it is no place for those with software/design background to dabble in and mess up. The long term goal should be to identify strategies to maintain competitiveness in the nascent solar PV industry in India even if to get itself off the ground it first uses standard off the shelf tools and technologies that are available to competitors, e.g., China etc. too.
This requires a policy of continuous improvement in cost and/or performance via technology, to sustain, which requires a capable domestic R&D and hardware base. To this end the government ministries must formulate policies that are well grounded in the technical and business realities of the semiconductor industry worldwide as well as adopt a systems approach. That is, it is not just offering financial incentives to set up semiconductor plants, but simultaneously, finance the development of the missing technical infrastructure for semiconductors, viz training of manpower in physics, materials sciences and construction of semiconductor production and testing equipment.
Research based training/education programs on solar photovoltaics and alternative energy in general should be launched at select Indian universities. Lastly, a new Semiconductor Hardware Association of India, composed primarily of physicists and hardware engineers (rather than the unqualified software/design type pretenders who have contributed to the wafer fab fiasco), should be created at the earliest so as to co-ordinate the development of the semiconductor hardware manufacturing industry in India with minimum avoidable delays.
The author is the Chief Technical Officer of APSTL llc, of Scottsdale, AZ, USA, a company that specializes in developing key semiconductor technologies and licensing them worldwide.