The solar power you don’t hear about

Solar thermal power is back! Solar thermal gets less attention than its sexier cousin — high-tech photovoltaics — but has two big advantages. First, it is much cheaper than PV. Second, it captures energy in a form that is much easier to store — heat — typically with mirrored surfaces that concentrate sunlight onto a receiver that heats a liquid (which is then used to make steam to drive a turbine).


Back in the 1980s, Luz International was the sole commercial developer of U.S. solar thermal electric projects. The company built nine solar plants, totaling 355 MW of capacity, in California’s Mojave desert. Luz filed for bankruptcy in 1991 for a variety of reasons detailed in this Sandia report.

For 15 years, no commercial solar thermal plants have been built until the Spanish system pictured here. Technology Review has published as an advertising supplement one of the longest and most informative pieces I have seen on solar thermal, also called concentrated solar power (CSP).


California utilities are also beginning to contract for new CSP plants — “the resurrection of thermal solar arrays,” as the New York Times puts it. In July, Pacific Gas & Electric announced a plan to buy 550 megawatts of CSP in the Mojave Desert

If you want to read more about this re-emerging form of solar power, the National Renewable Energy Lab has a website with publications on the technology and potential market.

14 Responses to The solar power you don’t hear about

  1. mk says:

    So what are the disadvantages of solar thermal? Can we have a more comprehensive comparison here?

    1. Energy efficiency:

    When I look around on the web I get wildly divergent numbers for energy efficiency rates (8-40% for solar PV, 2-40% for solar thermal). I am sure it depends on how much you pay (exotic PV materials get better efficiency, believably enough).

    2. Dual forms of energy
    Is another factor that solar thermal generates electricity plus heat, whereas solar PV only generates electricity?

    3. Costs, land usage, startup vs. maintenance costs, etc…

    Lots of ways to compare technologies…

  2. Earl Killian says:

    Joe’s links mention three different California projects: PG&E’s 550 MW plant with Solel Solar Systems, and two with Stirling Energy Systems’ technology: Southern California Edison’s 500-850 MW plant and San Diego Gas & Electric (SDG&E)’s 300-900 MW plant. All three are located in Mojave desert because of its excellent insolation. There is another player in this field that is worth watching: Ausra (originally from Australia, but now in Palo Alto, CA). Ausra’s technology is interesting because they claim 24×7 generation from the sun (by storing heat overnight). At a panel discussion this year they claimed they could provide 80% of California’s power at $0.067/kWh (which is pretty darn impressive). They have plans to create a California plant “within 3 years” at the gigawatt scale on 30km^2.

    Solar power is nice, because its generates peak power at the time there is peak demand, unlike wind (which often produces more at night than during the day). Nighttime demand is about half of daytime demand. (The plug-in vehicles that Joe advocates as part of the solution to our fossil addiction are a possible source of nighttime demand.) But even so, the power industry is nervous about having too much “non-baseline” power on the grid. Estimates are that we can have between 5% and 20% of the grid being supplied by renewable energy; after that you start to run into problems. (With plug-in vehicles on the grid, a technology called V2G raises the upper limit to about 50%.) But Ausra’s technology would be “baseline” energy. Their projection of 80% of California’s supply at $0.067/kWh was based on a year of CalISO grid data, and making sure they had sufficient heat storage to ride through the rare cloudy day in the Mojave. That is why is potentially so significant.

    You’ll find more about Ausra in this video:
    Start around 32:00.

    Or watch the Vinold Khosla interview at the same URL.

    On a personal scale, there is also this possible solar thermal future product:

  3. Earl Killian says:

    To mk:

    Yes, both PV and solar thermal claim efficiencies up to 40%. However, the cost-effective efficiencies are perhaps more interesting: PV 15-20%, solar thermal 30%.

    However, even efficiency, as important as it is, is not the most important criteria. Cost is most important.

    To give you an idea, residential PV is about $9/Watt installed (about $4/W for the PV, $1/W for the inverters, $4/W for the installation). Government subsidies can bring the cost down (e.g. something like $2.60/W in California).

    At the 100MW level, solar thermal is about $1.50/Watt. This is enormously more cost effective than PV.

    If the cost of renewable energy is too high, then we will have to wait for the government to force the market to go to renewable energies (as California and other states are starting to do). If the cost is low enough, then solar thermal might replace fossil power plants for new construction even without government intervention to reflect the true cost of fossil energy in the market.

    The biggest issue for solar thermal is getting the power to the users. The best places for it are in the desert southwest (e.g. the Mojave in California, and the Sonora in Arizona). The recent Stirling Energy Systems CSP farms in the Mojave required new grid infrastructure to bring the power to where people live and work in California. Even though we could in theory easily build enough CSP in the desert southwest to provide for all of the U.S. power needs, it would require massive upgrades to the U.S. power grid to bring it to the entire nation (New England is rather far from Arizona, for example).

  4. mk says:

    Ah, I see. Forgive the elementary question, but what about a solar plant based on PV? (Your figures refer to residential PV use). Is this basically never done because it is too expensive?

    I see I was mistaken earlier. I thought that parabolic mirrors could be used with PV type generation. But it seems they are only used for solar thermal power.

  5. Joe says:

    Too expensive. PV’s big advantage is modularity, the possibility of integrating a system directly into the roof of a building, for instance.

  6. Earl Killian says:

    To answer the second half of mk’s query: there is concentrated PV, but most of it is too complicated for residential and the PV costs too much for wholesale, so it isn’t much used. I think all the 40% PV efficiency numbers you hear are based multi-band-gap cells receiving concentrated sunlight, though I am not sure why concentration helps (it increases the heat, which is usually bad for PV). (One exception to what I just said: there are PV startups working on concentrators for PV that yield less than 2x concentration and require no moving parts, and so will be suitable for residential use.)

    What is annoying about the $4/W PV + $1/W inverter + $4/W installation equation is that PV at one fourth the price (i.e. $1/W, which would be a great breakthru) only reduces the residential system price by a factor of 1.5 (from $9/W to $6/W: nice, but hardly the 4x breakthru you would expect from such a PV breakthru). Where $1/W PV breakthru could make a big difference though is in wholesale electricity generation, since there one assumes the price of the inverters and installation can be reduced by the massive scale of the system.

    Of course, many off-grid homes avoid the $1/W inverter cost (and the 6-10% efficiency loss) by using DC for many of the appliances. DC lighting (e.g. LEDs) is obvious, but one can also buy extremely efficient DC refrigerators for off-grid homes. If PV comes down in price, there will be some incentive to add DC wiring to homes (in parallel with the existing AC wiring). One small advantage: all those ugly power bricks (which are AC to DC converters) plugged into the wall could go away if we had a DC standard for homes. That not is not only a convenience win, but an efficiency win.

    The other sad thing about PV: if it is 20% efficient, that means 80% of the energy is turned into heat, which is generally wasted in PV systems. One exception is the neat design by OM Solar, which puts that heat to work (e.g. for building comfort in winter).

  7. Paul K says:

    How is the cost/watt calculated? Does it mainly reflect the amortized cost of initial investment or the ongoing cost of power production?

  8. Solar for daytime peak power would no doubt work very well in the southwest, The Southeast has a lot of cloudy days, so solar woud not do so well. So what do you do for day/night base? Coal anyone? Does anyone believe that gas will be affordable? Does anyone believe that gas is really green. We need electricity to run all those hybrid and electric cars, don’t we? Anyone for sitting in the dark at night with no TV? Lets see. What do we have left? Oh yes, nuclear. Why didn’t anyone think of that before?

  9. Bernard Molloy says:

    A concentrated solar PV farm has been funded by the Australian government around 150 MW using the high efficiency 43% cells and flat mirror heliostats. The company involved Solar sytems is moving towards setting up a manufacturing plant at present. They have previously built parabloic relector concentrated PV in the outback for remote settlement.

  10. Nick says:

    For Charles Barton:
    “On sunny afternoons, those 10 plants would produce as much electricity as three nuclear reactors, but they can be built in as little as two years, compared with a decade or longer for a nuclear plant. Some of the new plants will feature systems that allow them to store heat and generate electricity for hours after sunset.” (

  11. vahila says:

    please add our website

  12. Estetik says:

    Does anyone believe that gas will be affordable? Does anyone believe that gas is really green. We need electricity to run all those hybrid and electric cars, don’t we

  13. Thermal, in conjunction with the Sterling engine may be our escape from
    oil dependency.

  14. medyum says:

    I see I was mistaken earlier. I thought that parabolic mirrors could be used with PV type generation. But it seems they are only used for solar thermal power.