Dreaming of a Solar Future

When I was researching my article last week, I noticed that there seems to be a plethora of new photovoltaic technologies in the lab and just beginning to come to market. The Wikipedia article is a good overview of how solar cells generally work and some of the interesting stuff currently in research.

The important point is this: right now, nearly all commercial solar cells are based on silicon semiconductors, and the majority of those are fabricated using fairly conventional silicon wafers and techniques similar to integrated circuit fabrication.

This is a problem because, while chip fabrication is relatively cheap for things like Intel processors--which use only a few square centimeters of silicon wafer to make something quite valuable--it is expensive for solar collectors which can only collect so much energy per square meter.

In other words, you don't get the benefit of Moore's Law with solar cells because the solar cell can't get physically smaller without reducing the amount of energy available to collect.

It also means that solar cell manufacturing competes for raw silicon against all sorts of other semiconductor products, and a silicon shortage (like what we're seeing right now) will push the price of solar cells way up.

There are other solar cell technologies in development, though, some of which have the chance to dramatically cut the cost of solar cells. Some of these are based on organic dyes, photochemical cells, and other radically different approaches. None of these can beat silicon for conversion efficiency (yet), but some look extremely promising for reduced cost per watt. Some have other promising characteristics, such as better-than-silicon performance in low light, which could allow them to outperform silicon on a total energy delivered per nominal watt basis.

It's always risky predicting the future of new technologies currently in the lab, but with the diversity of different approaches, there's a reasonable chance that at least one will deliver on its promise. If energy prices remain high (an important factor), I would guess that there's about a 25% chance we'll see an order of magnitude cost-per-watt breakthrough in photovoltaics within the next ten years.

Difference In Degree Becomes Difference in Kind

I have a personal rule of thumb that any time the cost of something changes by an order of magnitude, it leads to not just radically less expensive technology, but entirely new usage patterns. So what would happen if the price per watt of a solar cell dropped from $7.50 installed (today) to $0.75 installed (in ten years)?

First, let me point out that this magnitude of decline implies that the cost of installation and other equipment will also have to drop, since the solar panels are only about 75% of the cost of a complete system today. On the other hand, an order of magnitude drop in the price of solar cells could lead to a substantial (though perhaps not a full order of magnitude) drop in the price of installation and other components. That's because when the price of the cells is so much cheaper, you can afford to be sloppier: you can use installation techniques which are less expensive but lead to the occasional dead cell or panel, you can afford less-than-optimal site selection, and the price of other system components is likely to drop as sales volume increases.

The most obvious effect of an order-of-magnitude drop in the price of solar cells is that photovoltaic power would be (over the long term) half the price of the stuff from the power company. The average Minnesota home uses about 8,000 kWh of electricity per year, and today you need to spend $60,000 upfront to generate all that electricity using solar cells. Cutting that price to $6,000 will induce a lot of homeowners (and nearly all commercial properties) to go solar. In fairly short order, only people who couldn't afford the upfront cost would lack solar electric homes.

The first ripple effect of this would be on the power companies. They would be in the odd position of being fed lots of power during peak times, and having to generate power only at night and on cloudy days. The role of an electric utility would have to shift from power generation and delivery to power storage--and the current net metering schemes would have to change to keep the utilities from going bankrupt. Barring some breakthrough in electric storage technology, the cost of electricity at certain times might actually go to zero, as the utility could be forced to find ways to safely dissipate excess power being delivered from all those solar systems. Offsetting that would be much more expensive electricity at times when solar is unavailable.

(Ironically, in this scenario the people hurt the most would be the very early adopters of solar power: the people who put in solar systems while they were still expensive, counting on net metering to defray some of the cost. Those people would see their net metering benefit cut dramatically, yet be stuck with very expensive and outdated systems.)

If electricity storage rather than generation is the biggest problem, one logical place to store lots of cheap and excess electricity is in our cars. Right now, on a per-watt-delivered-to-the-wheels basis, electricity is about half the price of gasoline for powering cars (that's because the cost-per-BTU of electricity and gasoline is about the same, but electricity can be turned into usable work at about twice the efficiency of gas). The drawback of electric cars has always been recharging (it takes too long, and the car doesn't go far enough on a charge). For a commuter vehicle--which is most cars most of the time--this isn't so much of an issue. With solar electricity (hypothetically) costing one-quarter as much as gasoline, it makes a lot of sense to keep a spare battery pack at home and charging during the day. Then when you get home, you swap battery packs (the car would have to be specifically designed to allow this, since battery packs are big and heavy).

Or, alternatively, your employer might provide a recharging station at the office. A small solar panel on the roof of the car would give a little extra range during the daytime, but wouldn't be enough for continuous driving such as on an extended road trip.

We could also see a drive towards the construction of much larger electric grids. Northern latitudes could draw power from more southerly locations in the winter (when there's not much daylight up north), and conversely, the south could get electricity from the north in the summer, when us Minnesotans get tons of daylight and the folks in Mississippi are running their air conditioners around the clock. East-west power transmission would also make sense.

If someone could figure out how to turn electricity into liquid fuel at over 50% efficiency (as far as I know such a process doesn't yet exist), then excess solar power would also be turned into fuel for airplanes, trucks, and other vehicles where solar electric power is impractical (at 50% conversion efficiency under this scenario, such a liquid fuel would cost about the same to produce as gasoline sells at retail). This would remove one of the last remaining uses for fossil fuels.

As an aside: I don't consider hydrogen a practical fuel, even though it can be easily produced from electricity. Hydrogen is simply too hard to store and transport to be useful as a vehicle fuel.

Pipe Dreams?

The scenario I've spun sounds like a fantasy, and in many ways it is. Consider, though, that the premise (an order of magnitude drop in solar cell prices) may not be likely, but it is certainly possible. There's no law of physics or economics which says that converting sunlight to electricity must be expensive.

I started thinking about this topic when Scooter (now 8 years old) mentioned that he's really worried about global warming. When I was 8 years old, the Big Issue I worried about was nuclear war with the Soviet Union and the resulting destruction of all life on the planet.

Amazingly, though, by the time I graduated from high school, the Soviet Union had completely disappeared. Back in the late 1970's, hardly anybody would have thought this was even possible, much less that it would happen within a decade. In retrospect, the seeds of this dramatic change were already sprouting, but nobody gave the scenario enough credibility to take it seriously.

Today, our Big Issue is the increasing cost of fossil fuels and the increasing levels of carbon dioxide in the planet's atmosphere. But if you look, you can see the seeds of a dramatic change beginning to sprout, nurtured by the very high energy prices and environmental problems which have people so worried.

I would not want to place bets on any particular technology which might give us cheap solar power--they're all too immature--but as long as the price of energy stays high, the search for a cheaper alternative will continue. If such technology exists, I'm confident we'll find it sooner or later.