Moore’s Law and Photovoltaics – Why Solar Panels Are Different From Semiconductors

Moore’s Law states that the number of transistors on integrated circuits doubles approximately every two years. For ‘number of transistors’ read ‘processing power’ and you have a road map for a silicon based device that becomes increasingly more cost effective over time. It was tempting to believe the same would hold true with photovoltaic devices and that over time these too would provide more watts per square metre and hence more watts per Dollar or Euro. And initially efficiency did increase quite rapidly from around 6% to over 15% - but that took longer than two years and progress has now slowed. We may not see photovoltaic panels with an efficiency of 30% on the market in two years time - or of 60% four years from now. The current maximum for a cell is around 24% and there is something called the Shockley-Queisser efficiency limit that provides a ceiling of 33.5%.

So if we want to replicate Moore’s Law we have to increase the area of a panel - and use more silicon. The cost of panels has fallen rapidly: especially over the last few months as the slowing of Europe’s economy and scaling back of feed in tariff subsidies kicks in. There is an overcapacity at foundry level and accusations of dumping. So does this mean that solar panels, at least while the market digests the glut, will stand on their own two feet economically.

Perhaps not. There is still that issue of efficiency. True if you have sixteen square metres of roof going spare you can cover it with PV panels that produce around 4Kw peak at the current industry efficiency. But even if the panels were free it would be better to cover that sixteen square metres with technology that was more efficient and produced more power. There is an opportunity cost here as you can only use your roof once. As the major cost is now installation and ancillary equipment why choose a technology rated at less than 20% efficiency? And there are solar energy systems whose efficiency already exceeds the theoretic limit of PV panels. Perhaps it is also worth noting that little word that sneaked into this paragraph ‘peak’ - more of that in another posting.

Not wanting to appear totally negative (pardon the pun) about PV, compared to turning sunlight into electricity via oil PV is incredibly efficient. Mother nature took millions of years to turn Gigawatts of solar energy into a mere barrel of oil – and has now pretty much stopped doing this. We can use silicon to churn out Kilowatts in seconds.

However we may have been seduced into thinking the PV market will behave in much the same way as the high tech market. Here is a device that operates in the opposite way to other electrical devices. You plug it in and it generates electricity. Magic, don’t we just love devices? And perhaps over time, like PCs they will just get cheaper and more powerful. Just as we have gone from a clunky IBM PC to a supercomputer in our pocket so we will eventually see a Megawatt generator on every roof, won’t we?

Well obviously not and this delusion is going to cause a lot of pain down the line – most of it when subsidised feed in tariffs come to and end. And there are a few other issues we need to consider, ones we looked at long and hard when designing our solar energy capture storage system.

Peter Kruger

And follow me on Twitter