Solar Power Wins Enthusiasts but Not Money

Interesting New York Times article (and video) on solar power.

But for all the enthusiasm about harvesting sunlight, some of the most ardent experts and investors say that moving this energy source from niche to mainstream — last year it provided less than 0.01 percent of the country’s electricity supply — is unlikely without significant technological breakthroughs. And given the current scale of research in private and government laboratories, that is not expected to happen anytime soon.

Even a quarter century from now, says the Energy Department official in charge of renewable energy, solar power might account for, at best, 2 or 3 percent of the grid electricity in the United States.

That number of 7 billion kWh in 2030 and accounting for just .1% of electricity in the graph seemed low to me. So I ran my own projections.

According to Solar Buzz, Solar Electric Energy demand has grown consistently by 20-25% per annum over the past 20 years. Worldwatch Institute shows growth of 30-35% over the last 5 years.

The following table shows how many billion kWh and the percentage of total US electricity generation (based on 6,000 billion kWh) at various years with various growth rates.

growth rate	2030		2035		2050		2100
8.1%	7.0	0.12%	10	0.17%	33	0.55%	1,635	27%
15%	32.9	0.55%	66	1.10%	539	9%	>10,000	100%
20%	95.4	1.59%	237	3.96%	3,657	61%	>10,000	100%
25%	264.7	4.41%	808	13.46%	>10,000	100%	>10,000	100%
30%	705.6	11.76%	2,620	43.67%	>10,000	100%	>10,000	100%
35%	1,812.8	30.21%	8,129	100%	>10,000	100%	>10,000	100%

The EIA's projection of 7 bil kWh of solar in 2030, assumes a growth of just 8.1% a year. But if you use a 20% growth rate estimate, you get to 95.4 bil kWh a year, over 13 times as much! 20$ is a more reasonable estimate in my opinion. But this would still only account for 1.59% of total electricity generation.

If you use a 30% growth rate projection, as has been the growth rate over the last 5 years, in 2030 you get to 705 billion kWh and 12% of total generation. I think this kind of growth is unlikely, but possible.

You can really see the power of exponential growth as you go out to 2035, 2050 and 2100. If these growth rates are maintained (and looking at Moore's law for semiconductors, it is possible to keep growth rates like this up for a long time), by the end of the century we can be 100% powered by the sun. Even by 2050, under growth rates of anything greater than 25% we can become completely solar.

People have a hard time with exponential growth and tend to overestimate the impact in the short run and underestimate in the long run. While I don't think solar will contribute a sizable portion of electricity generation (>5%) until at least 2030, I would be surprised if it isn't producing a majority of the electricity of 2050.

You can also see how much difference a 5% increase in growth rate makes when maintained over a long period of time. In 2035 a 25% growth rate gets you to 13% of total electricity generation, while a 30% rate gets you to 44%.

In the current fiscal year, the Energy Department plans to spend $159 million on solar research and development. It will spend nearly double, $303 million, on nuclear energy research and development, and nearly triple, $427 million, on coal, as well as $167 million on other fossil fuel research and development.

After more than two decades in which research on converting solar power to electricity largely lapsed, the Bush administration and lawmakers in Congress are now discussing more money for the field. Dr. Orbach said the Energy Department’s proposed research plan for 2008 to 2012 includes $1.1 billion for solar advances, more than the $896 million going toward fusion.

Looking at the graph of energy research (click for a larger version), it seems like more should be given to solar. But, I wonder what the relationship between R&D spending and growth rate is? Will an extra $500 million increase the growth rate from 20 to 25%? Since the research being done likely takes 10 or more years before it gets into production, it probably has a similar lag in impact as well. Research money today likely impacts the growth rate of 2020 more than 2010. But just how much it can improve it, and how you determine whether that is a good use of tax payer money, I am not sure.

And since solar powered research helps not just the US, but all nations of the world, shouldn't all nations help to support the R&D? Right now Germany and Japan are doing a lot, but why shouldn't other rich nations help to pay the tab? Or what about companies that are making the solar cells? How much of the research should they support themselves based on the sales of their products?

The key to maintaining the high growth rates in solar is for the price to continually come down. According to Richard Swanson, solar panel prices decrease by 19% for every cumulative doubling of production. Looking at past prices, PV decrease in price by approximately 7-10% a year, which means the price is cut in 1/2 every 7 to 10 years. Solar Buzz puts the cost at 30¢ a kWh currently (I think as of 2005), so by 2015 it goes to 15¢, by 2025 7.5¢ and by 2035 3.125¢. When the price becomes competitive with coal and other forms of generation, then demand will soar.

Lately, the price of PVs hasn't been falling much, as the graph shows, due to higher prices of silicon. This is likely to be resolved soon and prices will continue on their march down. How much faster would prices fall if there was additional R&D by the US government? Once again I really have no idea of how to quantify it.

While, I'm not sure what the right amount of research funding that the US government should provide, I am confident that solar is the long term solution.

Update: While not directly comparable to R&D spending, venture capital in solar grew by 50% in 2006 to $421 million.