Solar Energy Not Quite Ready For Prime Time

April 14, 2008 | Ted Nordhaus,

Last week, Joe Romm on his Climate Progress blog challenged Michael's assertion that solar energy technologies will need to make very significant improvements in cost and performance before they can reasonably be expected to displace large amounts of fossil fuel based energy. Michael concluded:

Like past investments in everything from dam building to microchips, any grand plan to scale up solar enough for it to constitute an emissions stabilization "wedge" will require direct federal investment.


solar thermal.jpg
Romm argued both that coal costs have significantly increased in recent years and solar technologies -- both photovoltaic and compressed solar thermal -- have come down, and concluded that,

CSP now -- and PV soon -- are in very good shape. They certainly don't need a big government spending program aimed at generating breakthroughs in order to make them cost-competitive in time to play a very large role in beating 450 ppm.... Steady increases in the future [funding for deployment of solar technologies] are certainly welcome, but I don't see the need for some new multi-billion program. Certainly more R&D funds aren't needed.


We asked Breakthrough Institute senior fellow Frank Laird to look at Romm's post and here is his response:

The resource potential for CSP in the southwest is vast, about 6,900 gigawatts peak, with conservative assumptions that excludes much land. The total installed capacity of the U.S. is about 1,000 GWp. The plants currently being installed have a few hours of storage. Basically, you run a type of oil through a tube at the focal point of parabolic mirrors (shaped like troughs) and heat up a reservoir of the oil. You then run tubes of water through the hot oil, which boils the water, and then direct the steam through a turbine, generating electricity. The oil holds lots of heat and stays hot enough to boil water for a few hours after the sun goes down. In theory, you could make the oil reservoir as large as you like and generate power all night, but the price goes up when you do that and no one I know is proposing 24/7 power with any of the existing or planned power plants.

So how much does, and will, it cost? As to the present, we've been hearing about really cheap CSP for some time now, but estimates made by people who are NOT trying to sell you a power plant have been higher (about 15 cents per kwh at the busbar) than Romm suggests. Maybe that has changed in the last few months and I'm out of date. But in fact the real question is the price the developers are bidding. Once they bid a firm price and operate the plant for 4 or 5 years without going bankrupt, then we know the present price. Until then, it's competing estimates.

As to the future, who knows? The components of CSP -- tubes, mirrors, heat exchangers, etc. -- look pretty mature to me, so I don't know why manufacturing experience would drive their prices down a great deal. On the other hand, an R&D program that could come up with radically new types of mirrors, working fluids, tracking devices, etc. might have some real potential for cost reductions.

Another problem with CSP is the grid. The best potential for CSP is the desert southwest and the particular sites will need an extension of the grid to these rather remote areas. That adds to the price.

As far as PV is concerned, this technology is very much in flux, with lots of tantalizing possibilities, from novel materials to organic PVs and self-assembling nanotech PVs. Much of the research mapping out these possibilities in fact is funded by the government, though not always by DOE. It's a great time to expand R&D, not contract it. One of the world leaders in PV is Germany, both in manufacturing and installation. The reason they install so much of it is that they subsidize it massively. The feed-in tariff for PV in Germany is more than 50 euro cents (about 75 US cents) per kwh. Electricity consumers pay the subsidy. The Germans adjust the feed-in tariff every few years, lowering it as the technology improves, but it is still vastly higher than a carbon price of $20 or $30 per tonne that Romm is discussing. Moreover, the feed-in tariff is on top of a high (by US standards) eco-tax on fossil fuels. I admire what the Germans are doing, and would like to see us do some of it. But they want cheaper PV too and strongly support R&D.


I do hope that the optimism of the solar industry reflected in Joe's post is accurate. But I think a close examination of the multiple technical obstacles associated with the actual large scale deployment of solar using present technology suggests that the real costs in the real world, as Frank suggests, are likely to be a good deal higher than the industry estimates.

To be clear, that's not to say that relative to today, where solar energy constitutes an infinitesimal fraction of U.S. energy demand, we aren't going to deploy a lot more solar in the coming decades. Given where the industry starts, solar can grow at a very vigorous rate for many decades and still only constitute a relatively small percentage of U.S. energy generation. The industry and those who run it will do well, and the VC's and other private sector investors will see healthy returns on their investment. But solar will still likely not get to a scale wherein it will displace large amounts of present electricity baseload without dramatic improvements, whether you call them breakthroughs or something else, in both the basic generation technologies and the related storage technologies. Moreover, as Romm acknowledges, bringing these technologies to scale, will also require big investments in transmission. "We will need more transmission in this country," [especially low-loss long-distance high-voltage DC lines]."

And that brings us to the larger question, which is not what it will take to see incremental improvements in the performance of particular technologies but rather what it will take to see wholesale displacement of our present fossil fuel based energy infrastructure. We can offer dueling estimates and cite dueling sources all we want, but as Frank notes, estimates are just estimates and even if we take Romm's more optimistic estimates, associated with one of the more mature clean energy technologies available, we end up with half a wedge from PV and a wedge from concentrated solar in 2050.

Romm himself has acknowledged that given the actual trajectory of global carbon emissions, we will probably need between 12 and 14 wedges to put us on a path consistent with stabilization, so this would mean that even accepting Romm's optimistic assumptions, with a $30/ton price for CO2, "massive amounts of private sector money as well as big government subsidies (not R&D) and/or mandates," renewal of the ITC tax credit, and a federal renewable portfolio standard, we get a wedge and a half. By Romm's own assessment that means we need another 21 to 25 half wedges to put us on a path consistent with stabilization. Breakthrough Senior fellow Marty Hoffert has estimated we may need as many as 18 wedges, in which case we would need to find another 33 half wedges to put us on a path consistent with stabilization.

So why exactly are we arguing about whether we need massive public investment in clean energy technology?

Romm has often reduced our advocacy of major increases in public clean energy investment to advocacy for basic and applied R&D. But in fact we have long made the case for major public investments throughout the entire technology development, adoption, and deployment process, from basic and applied R&D, to early stage demonstration, to deployment and commercialization. Whether you think we need 12 or 18 or 22 wedges to put us on a path consistent with stabilization, we will need a broad portfolio of technologies to get there. Some of those technologies, like solar thermal and solar PV, are mature enough that we ought to significantly ramp up support for deployment. Others, like carbon sequestration and storage need to be demonstrated quickly so that we can properly assess their potential to reduce emissions. Still other, like nanosolar, are in much earlier stages, but could radically change the economics of clean energy technologies.

The point is that we need to massively scale up at every stage of the process and we need to do it quickly. Carbon prices, regulatory mandates, and modest tax credits can help, and in and of themselves might get us gradual and incremental expansion of clean energy technologies, but getting to climate stabilization will require much more.