Yesterday I went to an interesting talk here by Eric Toone, deputy director of ARPA-E, what is supposed to be the blue-sky high-risk/high-reward development portion of the US Department of Energy. He summarized some basic messages about energy globally and in the US, gave quite a number of examples of projects funded by ARPA-E, and had a series of take-home messages. He also gave the most concise (single-graph) explanation for the failure of Solyndra: they bet on a technology based on CIGS solar cells, and then the price of silicon (an essential component of the main competing technology) fell by 80% over a few months. It was made very clear that ARPA-E aims at a particular stage in the tech transfer process, when the basic science is known, and a technology is right at the edge of development.
The general energy picture was its usual fairly depressing self. There are plenty of fossil fuels (particularly natural gas and coal), but if you think that CO2 is a concern, then using those blindly is risky. Capital costs make nuclear comparatively uncompetitive (to say nothing of political difficulties following Fukushima). Solar is too expensive to compete w/ fossil fuels. Other renewables are also too expensive and/or not scalable. Biomass is too expensive. Batteries don't come remotely close to competing with, e.g., gasoline in terms of energy density and effective refueling times.
The one thing that really struck me was the similarity of the replacing-fossil-fuels challenge and the replacing-silicon-electronics challenge. Fossil fuels have problems, but they're sooooooo cheap. Likewise, there is a great desire to prolong Moore's law by eventually replacing Si, but Si devices are sooooooo cheap that there's an incredible economic barrier to surmount. When you're competing against a transistor that costs less than a millionth of a cent and has a one-per-billion failure rate over ten years, your non-Si gizmo better be really darn special if you want anyone to take it seriously....
Interesting comparison. But when I worked on silicon, it seemed to me that a big part of the barrier was the trillions of dollars that had already been invested making silicon cheap, by making ever-huger batches of ever-smaller devices. Any competing technology has to find a way to make money to support its learning curve, even if it could eventually be cheaper.
ReplyDeleteThe question for energy is whether there is the same issue, of cumulative investment making late entry prohibitively expensive. Certainly the technology for deep-water drilling and horizontal fracking allows the petrochemists to stay competitive even when oil doesn't just ooze out of the ground. There's also a lot of infrastructure (gas stations) and legal and financial support for fossil fueles. But maybe there's also just an inherent cost advantage to using pre-existing energy-packed molecules, compared to other sources. Even a factor of two in intrinsic cost is a big deal.