How to use solar energy at night: Concentrated solar thermal power with storage

This is a pretty good story by Scientific American on Concentrated solar thermal power Solar Baseload “” a core climate solution.  Figure from DOE.

How to Use Solar Energy at Night:  Molten salts can store the sun’s heat during the day and provide power at night

Near Granada, Spain, more than 28,000 metric tons of salt is now coursing through pipes at the Andasol 1 power plant. That salt will be used to solve a pressing if obvious problem for solar power: What do you do when the sun is not shining and at night?

The answer: store sunlight as heat energy for such a rainy day.

Part of a so-called parabolic trough solar-thermal power plant, the salts will soon help the facility light up the night””literally. Because most salts only melt at high temperatures (table salt, for example, melts at around 1472 degrees Fahrenheit, or 800 degrees Celsius) and do not turn to vapor until they get considerably hotter””they can be used to store a lot of the sun’s energy as heat. Simply use the sunlight to heat up the salts and put those molten salts in proximity to water via a heat exchanger. Hot steam can then be made to turn turbines without losing too much of the original absorbed solar energy.

The salts””a mixture of sodium and potassium nitrate, otherwise used as fertilizers””allow enough of the sun’s heat to be stored that the power plant can pump out electricity for nearly eight hours after the sun starts to set. “It’s enough for 7.5 hours to produce energy with full capacity of 50 megawatts,” says Sven Moormann, a spokesman for Solar Millennium, AG, the German solar company that developed the Andasol plant. “The hours of production are nearly double [those of a solar-thermal] power plant without storage and we have the possibility to plan our electricity production.”

Using mirrors to concentrate the sun’s energy is an old trick””the ancient Chinese and Greeks both used it to start fires””and modern power plants employing it might provide a significant source of renewable energy without any greenhouse gas emissions….

Melting salts at temperatures above 435 degrees Fahrenheit (224 degrees Celsius), however, can deliver back as much as 93 percent of the energy, plus the salts are ubiquitous because of their application as fertilizers.

“There’s a term called round-trip efficiency. Basically, it’s a measure of how much electricity is produced if the thermal energy that’s generated is first stored and then used compared to just directly taking the energy. That number is around 93 percent,” explains NREL senior engineer Greg Glatzmaier. “[For] things like compressed air and mechanical type storage, there’s more significant losses,” an average of at least 20 percent over all the various technologies.

The Andasol 1 power plant, which cost around $380 million (300 million euros) to build, is the first to actually use the technology, so it remains to be seen how it will work in commercial practice. But U.S. government laboratories””NREL as well as Sandia National Laboratory in Albuquerque, N.M.””have already proved the technology can work in demonstration projects that employed it, like the Solar Two power tower outside Barstow, Calif.

Solar Millennium is so confident the technology will work that a twin solar-thermal power plant (Andasol 2) is already near completion. “It will start operations at the beginning of summer””May or June,” Moormann says.

And Arizona Public Service Co. (APS) has contracted with Abengoa Solar to build a 280-megawatt solar thermal power plant””dubbed Solana or “sunny place”””70 miles (110 kilometers) southwest of Phoenix on nearly 2,000 acres (800 hectares) of land. “One of the great things about molten salt technology is that you can get more out of the pure solar resources, more energy out of the same facility,” says Barbara Lockwood, manager for renewable energy at APS. “It’s an alternative that provides us with additional green energy,” as much as 1,680 megawatt-hours when cloudy or after sunset.

But that extra energy comes at a cost. First, the power plant has to be enlarged so that it is both generating its full electrical capacity as well as heating up the salts. In the case of Andasol 1 that meant covering 126 acres (50 hectares) with long rows of troughs and pipe. And then there is the additional expense of the molten salt storage tanks, according to Moormann.

All told, that means thermal energy storage at Andasol 1 or power plants like it costs roughly $50 per kilowatt-hour to install, according to NREL’s Glatzmaier. But it doesn’t add much to the cost of the resulting electricity because it allows the turbines to be generating for longer periods and those costs can be spread out over more hours of electricity production. Electricity from a solar-thermal power plant costs roughly 13 cents a kilowatt-hour, according to Glatzmaier, both with and without molten salt storage systems.

That price is still nearly twice as much as electricity from a coal-fired power plant“”the current cheapest generation option if environmental costs are not taken into account. But Arizona’s APS and others can then use solar energy to meet the maximum electricity demand later in the day. “Our peak demand [for electricity] is later in the evening, once solar production is trailing off,” Lockwood says. That’s “the reason we went that direction and are so interested in storage technology.”

As efficient as solar-thermal power plants using parabolic troughs with molten salt storage systems like Andasol 1 or Solana are, they don’t capture as much of the sun’s heat as is possible. Above 750 degrees F (400 degrees C), the synthetic oils used to capture the sun’s heat in the troughs begin to break down, but the molten salts can take in much more heat than that.

To allow the salts to get hotter, some companies, such as SolarReserve in Santa Monica, Calif., are developing so-called power towers“”vast fields of mirrors that concentrate sunlight onto a central tower. Because of the centralized design such a structure can operate at much higher temperatures””up to 1,000 degrees F (535 degrees C)””and use molten salts directly as the fluid transferring heat in the power plant. “We are heating the salts to more than 1,000 degrees F and that results in the same inlet conditions that utilities see today on a coal-fired or nuclear power plant,” says Terry Murphy, SolarReserve’s president.

But such a power plant””and Murphy says the company has some 50 such projects in the pipeline and expects at least one (in the U.S. or Spain) to be operating by 2013″”would cost as much as $800 million for a 200-megawatt power tower. “The first molten salt power tower built is going to be a real trial,” says Thomas Mancini, manager of Sandia’s Concentrating Solar Power Program. “It’s going to take someone progressive enough to finance it or take a little more risk.”

So researchers are also looking into salts that could be used instead of the oil in parabolic trough power plants, such as those that melt at lower temperatures and therefore would not freeze as readily during cold nights, according to Hank Price, a vice president for technology development at Abengoa Solar.

Solar Millennium is working on such a salt, according to Moormann, and Sandia has developed small quantities of a new mixture of salts, including calcium nitrate and lithium nitrate, that melt below 212 degrees F (100 degrees C). “With the lithium nitrate, it’s as expensive as all the other constituents combined. Though still a lot cheaper than organic heat-transfer oils,” says chemical engineer Bob Bradshaw at Sandia in California, who is leading the research. “You don’t get something for nothing.”

And long-term research projects are looking at other thermal storage technologies, such as storing heat in sand or creating single-tank molten salt storage. “The main goal is to find a storage technology that may reduce the actual capital cost” of adding it to a power plant, says Phil Smithers, technical services leader for renewable energy at APS, which is researching those technologies under a U.S. Department of Energy grant.

Related Posts:

28 Responses to How to use solar energy at night: Concentrated solar thermal power with storage

  1. Alex Carlin says:

    Yes, and building only 100 miles by 100 miles of these installations in the USA, a small piece of Nevada, would mean we could stop burning coal right away. The same is true for China and India and Europe (in the Sahara). Science tells us we better do either this or some equivalent in the next few years or we destroy the future for our kids. I recommend we call for “100 Miles of Mirrors” because it is a catchy demand and it does the job. I also like Joe’s solution of the “wedges”, but is this demand “12 wedges” too complicated for the short time that we have?

  2. Sou says:

    Good idea. I’m betting the cost of each climate change disaster would go a long way to offset the cost of building such facilities. And if we manage to slow climate change, we might save a lot more from not having to build as many water recycling and water desalination plants.

    Queensland floods will cost Australians billions of dollars, Victorian floods will cost at least multiple millions, probably billions in lost agriculture, business interruption and infrastructure repair and replacement, losses to tourism. Costs in Brazil, Pakistan, Sri Lanka, Phillipines, Russia etc – huge. And that’s only 12 months of climate change disasters.

  3. fj3 says:

    Extremely encouraging basic technology which deployed completely with wartime speed would likely replace coal power in 5 years while providing full employment and eradicating the recession.

    Quick call the president and tell him to pull the country together and go for it. We have no time to lose!

  4. James Newberry says:

    Concerning the DOE schematic diagram, where is the salts to water heat exchanger and where is the water supply?

    RE: “All told, that means thermal energy storage at Andasol 1 or power plants like it costs roughly $50 per kilowatt-hour to install, according to NREL’s Glatzmaier. . . . Electricity from a solar-thermal power plant costs roughly 13 cents a kilowatt-hour, according to Glatzmaier, both with and without molten salt storage systems.”

    The $50 figure would seem to apply to installed kilowatt capacity cost rather than kilowatt-hour generation. Also, how can 13 cents apply irregardless of the extra storage capacity option? I think this is viable technology but somewhat deficient in reporting and diagram.

    [JR: This is just a simple diagram. The water can be closed loop.]

  5. nyc-tornado-ten says:

    I believe that PV panals are the best means of generating solar, because they can use existing buildings and parking lots, they do not consume any unused land. Also, PV generates power on the location of the end user, eliminating transimission loss. with the addition of large batteries, up to 20kw, which is becoming possible now with electric cars, we can even build in a reserve in local PV systems.

    I believe that thermal storage is great for storage of large amounts of electrical power, and this technology should also be developed. A smaller thermal storage system can be used to heat and power buildings.

  6. Theodore says:

    reply to #4

    Power can cost the same with and without storage because the plant can operate more hours per day with storage than without. This gives the opportunity to sell more electricity. The plant must be paid for at the same rate whether it is producing power and generating income or not. Storage more fully utilizes the other power plant components.

  7. Theodore says:

    It seems to me that the problem of freezing salt in long pipes could be eliminated by using a central tower design with heliostats in which the whole tower is the storage tank. The salt would not be pumped anywhere, except perhaps being stirred within the tank to keep temperatures uniform. The heat exchanger could be built into the tower or could be at the base, eliminating the need to pump it more than a very short distance.

  8. Mulga Mumblebrain says:

    sou, I’ve seen you at The Drum, but I’m blackballed, which the ABC’s anonymous thought police often do. I had twelve or so postings disappear on one of the climate change postings last week, most considered replies to denialist imbecili like ‘mick’. I see other people complaining, too.I think this is one reason why those type of blogs end up dominated by the knuckle-heads-censorship by Rightwing apparatchiki in the media.Returning to reality, we know solar thermal could power the world, but, as was said at least forty years ago to explain why solar had not been developed, ‘Standard Oil (Exxon etc) doesn’t own the sun!’.

  9. David B. Benson says:

    The solar thermal tower with some storage being built now n the Mojave has a contracted busbar price of 137 mills/kWh. That’s fairly expensive power, but does have the advantage of being load following (until the store is used up).

  10. Vic says:

    In lieu of any responsible energy plan from Australia’s government, the University of Melbourne have set up a not for profit organisation to come up with a real plan. You can find out more here,

    Their detailed, fully costed and modelled plan involves molten salt power towers and wind turbines, with biomass incineration to cover temporary shortfalls in wind and sunshine. They estimate that 3% of Australia’s GDP for ten years would cover the cost of making all of Australia’s electricity production carbon neutral, with labour being sourced from the rapidly phased out coal-fired generation industry. Sweet !

  11. David B. Benson says:

    James Newberry @4 — The storage stores energy:
    and the price is exactly as stated.

    The LCOE mostly derives from the cost of everything other than the store; with the store the price is only slightly higher for typical storage times of 4–7 hours.

  12. David B. Benson says:

    Vic @10 — Three Australains have a paper strongly recommnding adding nuclear to that mix:

    [JR: If you believe that new nuclear power plants can deliver electricity for under six cents a kilowatt hour, I have some beachfront property in Tuvalu to sell you. You can’t get a vendor to guarantee even twice that price in a PUC rate hearing.]

  13. David B. Benson says:

    According to international experience, if nuclear energy were adopted in Australia its initial cost (termed ‘first-of-a-kind’) would be about $30 per MWh higher than in the diagram, but would come down to that level as more plants were built.

  14. David B. Benson says:

    Consider using resistence heaters on the thermal storage unit driven by, say, wind turbine derived power. Estimate the cost of this power as 63 mills/kWh. Now the combination of the Rankine cycle steam generator and the thermal storage is 39% efficient, so the resulting busbar cost for power produced this way is 161.5 mills/kWh [before adding in the small cost of the resistence heaters]. So if solar thermal is acceptable at 137 mills/kWh, so might be is the additional increment produced this way.

  15. Sou says:

    Thankfully it would be extremely difficult to introduce nuclear power to Australia. As well as the cost, logistical, security and site problems, not many people want nuclear close by, despite the occasional promotion by nuclear proponents and the consequent appearance of growing support from some in the community. Other options such as the above would be preferred by more people.

    @ Mulga – It’s amazing that you’ve been blackballed on the Drum/Unleashed. You’re one of the better contributors and very popular. When I have a chance I’ll add my voice of concern.

  16. Vic says:

    David @12,

    My personal view is that in Australia at this time, renewables would be an easier sell (politically speaking) than nuclear, and given that time is of the essence, the best first step.
    Which is not to say I’m against nuclear energy. Do whatever it takes to fix this mess I say.

  17. Bill Woods says:

    In the two years since this article was published, there have been some developments. For instance, Andasol 2 is already in operation; Andasol 3 is under construction, though the price hasn’t come down.

    The Archimede plant in Italy, which uses salt in the collectors, is also in operation. In Arizona, Solana is more-or-less under construction.

    Of course, none of these provide baseload power, so they’re not replacements for coal.

  18. Alex Carlin says:

    Yes, solar thermal with storage can replace coal. In the transitional period we will use various other sources to plug gaps, but there is no need for killer coal. And as Wikipedia indicates:

    “Among the renewable energy sources, hydroelectric, geothermal, biogas, biomass, solar thermal with storage and ocean thermal energy conversion can provide baseload power”

    And the 7.5 hours of nighttime solar power referred to in the article gets us most of the way there.

  19. Theodore says:

    I wish a large CSP plant could be built that could produce power continuously for several days without sun. This is possible, but may not be economical, because this extra capacity would only occasionally be used. It would still be a good idea to build one, just to demonstrate that it can be done. I think many people will never understand this until they see it, and they can’t see it because it hasn’t been built. The mere fact that it is obviously possible doesn’t help. To many people, nothing is obvious.

  20. spiritkas says:


    I do find the cost argument to be a bit disengenuous. Wind and Solar are already MUCH MUCH cheaper than coal, but we don’t see that reflected in the price. The externalized costs of air pollution, lung disease, and heavy metal contamination of our food and water sources vastly tip the balance of the equation. I’m not one to do it, but a bit of research into what could be attributable direct costs of coal, not even counting climate change itself, would almost certainly show how expensive using coal really is.

    Right now the guy with lung cancer whose loosing his house to the bank and his job over health insurance companies is paying for coal. Right now a generation of children with increased rates of asthma are paying for coal. Right now the bioaccumulation of mercury and other heavy metals is disrupting our fish industry and causing untold damage to the brain development in our most vulnerable population.

    I’d like to add all those direct costs up that don’t even include flooding and hurricanes and put that into the killowatt-hour cost of using coal, then we can see what is cheaper. This isn’t a normally debatable point, it isn’t anything to do with climate change, it is the direct costs of creating giant vats of poisioned water next to coal plants and watching earthen walls burst ruining hundreds of acres of land, it is the cost of taking poison, setting it on fire, and dispersing it in the air for all to breath.

    I think to engage on a point for point ‘as is’ cost of coal vs cost of solar vs cost of wind is not a fair or even debate. Solar and wind have been cheaper for a long long time.



  21. David B. Benson says:

    Theodore @19 — Having a large thermal store might well prove economic for those regions having no hills, so no significant hydro reservoirs. Solar and wind are fine, but are not dispatchable against demand; a thermal store relieves that.

  22. David B. Benson says:

    Hopefully I did the LCOE calculations correctly.

    Poor Man’s Guide to Levelized Cost of Electricity (LCOE) for Nuclear Power Plants (NPPs)

    in which the result is busbar costs which might vary from a high of as much as 14.94 cents/kWh in the USA (maybe) to a low of 4.31 cents/kWh in China (maybe).

  23. Omega Centauri says:

    Aren’t these solar thermal plans under threat (or as least serious stress) because of the declining cost of PV, and the difficulties involved in siting large plants? That would be unfortunate, because the storage, and load following capabilities of solar thermal would be very valuable, but at current levels of solar penetration not needed.

  24. Mulga Mumblebrain says:

    sou #15, bad news! The ABC is now publishing my drivel again. I think it has something to do with the ‘moderators’, some of whom, to their credit, seem repulsed by my immoderate utterances.

  25. Bill Woods says:

    David B. Benson (#23): “Hopefully I did the LCOE calculations correctly.”

    Using NREL’s “Simple Levelized Cost of Energy Calculator”,
    , I get similar numbers, though somewhat lower. E.g.:
    40 year period, 10% discount rate, $7360/kW, 85% capacity factor, 31 mills/kWh = 27 $/kW-yr O&M, 0.01 $/kW-h fuel cost -> 11.4 ¢/kW-h

    “By way of contrast, wholesale power futures market for on-demand, firm peak power range from 27 to 33 mills/kWh in the western United States.”

    3¢/kW-h seems much too low a price for peak power, or am I misunderstanding?

  26. David B. Benson says:

    Bill Woods @26 — Thanks for the link!

    As for your question, that’s how I read the Mid-Columbia Hub figure; same for the other hubs.

  27. Bill Woods says:

    David B. Benson (#27): “Thanks for the link!”

    Here’s a roll-your-own formula:
    “Sethdayal, here’s my Excel formula for amortizing $2/watt over 40 year lifetime, 8% interest rate, producing power 90% of the time:
    =PMT(0.08/12,40*12,2/0.9)/0.72 = $0.021460.
    So I glibly say $2/watt is 2 cents/kWh, for the capital cost.”

    Plugging those numbers into NREL’s calculator gets the same result (to 2 significant figures).