A recent article notes that 2 U.S. solar energy startup firms, Semprius and Alta, have broken solar-panel efficiency records by using GaAs (Gallium Arsenide) cells as opposed to the conventional Silicon. Although this material is more expensive, the fact that it’s much more efficient at absorbing sunlight means overall it could be more cost effective than conventional solar cells per kwh produced . The cells made by Semprius are basically a completely new design, using mirrors that concentrate light onto extremely tiny cells, and these 3 layers of cells (as opposed to 1 normally) can also absorb a wider spectrum of the sun’s energy. For this reason they achieve 33.9 % efficiency compared to 15% currently for the average solar panel. Furthermore, these cells aren’t prototypes unlikely to actually be sold on the market, but rather “designed for commercial use” .
Of course, the issue in deciding whether an alternative energy is viable is almost always the cost per kw-h of the electricity produced by it. To this end, Semprius’ panels only use 1 micron thickness of GaAs and a novel production technique using chemical etching on an “inexpensive substrate.”  In MIT’s “technology review” article, Joe Carr, CEO of Semprius, believes the panels will eventually produce electricity at a cost of 10 cents per kw-h. According to the eia (U.S. Energy Information Administration), the national average retail cost of electricity (the price consumers pay to companies per kw-h) is 9.83 cents . That means that if Semprius’ prediction becomes reality, there’s a decent likelihood that solar energy will make a significant part of the U.S. electricity portfolio. However, I doubt that it will match coal’s use or completely displace it until at least 30 years, depending on a number of political, technology, and climatic factors. Furthermore, the DOE (Department of Energy) predicts that if the price of solar energy reaches 6 cents per kw-h by 2020, “solar panels would account for 15-18% of US electricity generation by 2030” as opposed to less than 1% currently .
A huge issue with alternative energy sources is also government subsidies. In the engineering finance class I took last semester, we studied the economic feasibility of alternative energy in items such as hybrids cars, electric cars, and solar panels in depth. The study accounted for a huge variety of parameters, including: inflation, value of the panels, psychological effects of the users, uncertainty in solar flux, deterioration in panel efficiency over time, aesthetic and practical effects of the panels on home value, the real increase in electricity cost (without inflation), change in usage between months, and the social cost of CO2. This was done from society’s and an individual’s perspective. The study concluded that current panels are highly economically unfeasible, from both perspectives, for 2 different setups, even with large government subsidies. This was based on Sharp panels with a nominal efficiency of 13.7 %. Society lost an average of $18,000 over the 25 year panel life and the user lost over $7,000, even with subsidies, and that’s for 1 house! However, interestingly enough, our analysis concluded that the panels would have to be 29-31 % efficient for the user to break even, but 39-57 % efficient for society to break even since society pays for subsidies through taxes. According to the “technology review” article, CEO Carr believes he can reach the 10 cents per kw-h target “without the help of government subsidies” .
For the reasons outlined above, I’m a bit skeptical but still optimistic for the future. As noted in the end of the last paragraph, the new panels have reached an efficiency that could be feasible for the user. I believe that it’s only through new technological advances such as Semprius’ that solar energy has a chance to be successful in the marketplace, and this new technology seems very promising if, and only if, it can be implemented in the manner desired. Right now, coal and natural gas account for 45% and 24% of electricity generation in the U.S., respectively . If this technology continues to advance at the rate computer technology has, for example, I believe the prediction is surprisingly reasonable. My skepticism is about the cost estimates and feasibility of mass production, because no matter how promising a technology and how “green” any given person is, it doesn’t matter if the product is too expensive to actually use. I hope Semprius’ partnership with Siemens, who have “experience in manufacturing and in building complete solar farms,” will actually “make it easier to secure financing from banks,” and thus make the product cheap enough to make a real difference on a national scale.