Molten salt storage solar power plants are not new. For example, Gemasolar is a 20 MW solar tower plant that has been operating in Spain since 2011. (
https://en.wikipedia.org/wiki/Gemasolar_Thermosolar_Plant). Gemasolar, like the SolarReserve plants, heats the salts directly from sunlight. More common are molten salt "trough" based plants, such as Andasol, also Spain, 150MW and 7.5 hours of storage (
https://en.wikipedia.org/wiki/Andasol_Solar_Power_Station) and Solana, near Pheonix AZ, USA, at 250MW x 6 hours (
https://en.wikipedia.org/wiki/Solana_Generating_Station). These trough plants use a heat transfer fluid like Therminol in the troughs, which in turn heats molten salt via a heat exchanger (hx).
Also, molten salt energy storage has been used for decades in the chemical and process industries, as well, as a medium to store useful energy from one reaction and use it in another. There were also solar thermal plants with energy storage that used other materials, such a oil for a heat transfer material and for storage. The SEGS-1 plant in Daggett, CA USA, built in 1984 had such a system. However, that machines storage system was destroyed in a fire, and it was never rebuilt. (Plant continues to operate today, without storage.) (
https://en.wikipedia.org/wiki/Solar_Energy_Generating_Systems). There are ideas to use other materials for thermal storage, too, such as ceramic or graphite honeycomb (
http://www.graphiteenergy.com/) (heat xfer fluid would be air or co2) and even sand
http://www.energymatters.com.au/renewable-news/sand-batteries-solar-em5272/).
The salts in these molten salt plants are a mixture of sodium and nitrate salts. The mix varies for different reasons; it is not necessarily a eutectic mix. The highest usable temperature is as important as the freezing temperature, and various mixes have different properties in terms of their chemical reactivity at high temps. If you can get a 100C higher with a given mix before the salts start tobreak down or react with the atmosphere, requiring that the tanks be inerted, then that might be worth a 20C higher freezing point. Of course, another consideration is that sodium and nitrate do not cost the same, and you are buying a fair bit. Here's a 2008 paper from Sandia National Lab (USA) that describes some of the considerations:
http://energy.sandia.gov/wp-content/gallery/uploads/ES2008-54174-molten-salt-for-troughs.pdfAs someone else pointed out, keeping the salts molten is not a trivial problem. Remember, these plants are built in typically desert environments, that can get quite cold at night and thermal radiation to the sky can be high. Also, during winter there can be many consecutive days without sun, even in the sunniest places on earth. Typically, all the piping needs to be heavily insulated, heat traced (that is, electrically heatable) or both, or a certain amount of hot tank energy must be reserved from energy production for the purpose of circulating through the system to keep everything molten, or a small gas-fired burner can be used to add heat to the system. A tower based plant would likely have the salt tanks at the bottom of the tower, and so the only real piping would be up and down the tower to/from the tanks, steam generator/hx on the ground, and the solar receiver on the tower. You don't see trough based systems directly heating the salt in the troughs because such machines typically have miles of receiver pipe exposed to the sky.
There are other ways to skin this cat, with tradeoffs. For example, you could built a tower based plant with a steam generating receiver at the top rather than direct heating of the salts. Then you bring steam down the tower, and use an hx to heat the salt. Later, when you want to make power from the salt storage, you use the same or a different hx to make steam, then run a turbine. The extra conversions (steam-salt-steam) costs you in efficiency when operating from storage but during the day you can operate with no conversions (steam directly to turbine), so it could be a wash, efficiency-wise depending on when you plan to make power. Direct steam also might be attractive if it simplifies the molten salt piping or if you know how to make a steam solar receiver but not a molten salt solar receiver, of course.
The economics of solar with thermal energy storage are interesting. On a straight energy basis, PV has far outpaced the cost of solar thermal energy. If you don't need or value storage (for example, your power system doesn't already have much solar on it and can manage the diurnal and weather variation just fine) then these systems are probably not interesting. On the other hand, if you already have a lot of PV on your system and your ability to deal with additional variation from a solar resource is limited, storage becomes very interesting because such machines are in theory dispatchable. That is, you run them when you need them, because power generation and solar collection are uncoupled. In that case, one might support compare PV + battery to CSP + thermal storage. In this case, the latter is still favorable, particularly at scale. Consider the 250MW x 6 hr machine mentioned above. That's 5.4 TJ. The largest battery in the US is a 32MWh (0.11 TJ) unit owned by SCE. (
http://www.greentechmedia.com/articles/read/The-Biggest-Battery-in-North-America-Gets-Unveiled-By-SCE-Today).
(By the way, someone suggested storing PV energy thermally. Of course, that would be a great waste to take a low entropy energy type like electricty, and convert it to heat, just so you can convert it back to steam and electricity again. Efficiency going from electric to heat is near 100%, but the best you could do is maybe 35% on the return trip. And that's not even counting the fact that PV is only 20%+ efficient, whereas merely collecting solar heat can be much better. On the other hand, if PV is _*REALLY*_ cheap and you have more than you know what to do with, there could be some circumstance where doing this makes some economic sense!))
All that said, in most parts of the world, it will be some time before storage is "needed," which will give the battery folks plenty of time to work their way down the price / capacity curve. Of course, the thermal energy storage people will continue to progress, too, but I strongly suspect that batteries will cross over at some point.
[ PS. I wrote my Masters thesis on CSP+TES, and worked for a CSP company for several years. ]
[ PPS. Here is a tome on the economics of CSP+TES. I am a coauthor.
http://www.csp-alliance.org/wp-content/uploads/2014/09/The_Economic_and_Reliability_Benefits_of_CSP_with_Thermal_Storage-2014_09_09-FINAL.pdf ]
-- dave