The effort to increase the yield of solar cells is relentless. Today the conversion rate of solar energy into electricity is below 25% (it was around 20% at the end of last century and it has basically levelled out) although there are some solar cells sitting on research bench labs that reach 40%. Commercial solar panels are more likely to have an efficiency between 15 and 20%.
The improvement in efficiency has been made possible by using sophisticated materials that cost a lot of money, which would translate into too pricey solar panels with no economical sustainability.
Hence the interest for this news from Semprius of a new approach to design solar panels yielding an efficiency of 43% (and they claim it shouldn't be impossible to push it up to over 50%) whilst keeping the price at the level of today's commercial panels.
One of the problem in today's solar cells is that the semiconductor used to convert the sun light into electrical power is actually using only a few wavelengths of the incoming light. What Semprius did was to stack different kind of semiconductors (silicon with different doping materials) to be able to capture much more of the Sun light spectrum. This has tripled the output of the solar cell.
Once this translate into a commercial product, which according to Semprius should not take a long time since it is compatible with current manufacturing processes, it would significantly decrease the price per W making it competitive with fossil fuel based electricity. Of course, the problem than to get electrical power you need sunlight and you don't get it when the Sun is down or the sky is clouded still remains...