Most of the comments below have focused on the thermodynamics in play for this. Let me instead speak a little on the materials chemistry:
Materials that are stable liquids at relatively high temperatures and that exhibit low vapor pressure have been investigated for quite some time as heat transfer fluids. See, for instance, the Molten Salt Reactor Experiment at Oak Ridge where a lithium fluoride-beryllium fluoride mixture was used.
The size of the phase diagram space to investigate comes from the phase rule, and the conditions for equilibrium from the Gibb's criteria. The classic materials science approach (how I learned) is to try and map the phase diagram space experimentally. This can become unwieldy for many component systems.
Common chloride salts and/or nitrate salts have been investigated as potential heat transfer/storage fluids, usually in eutectic compositions to lower the melting point. The actual ionic compositions of the molten fluids can be rather involved.
The widely used austenitic stainless steels (e.g. 304, 316) are susceptible to corrosion attack from chloride ions, so understanding of coordination chemistry around chlorine is important to determine the likely stability of a containment vessel.
It would be ideal to find salts comprised of common materials that formed moderate melting temperature eutectics, have no free chlorides (or no chlorine at all), low vapor pressure, and long term stability in the presence of common stainless (or even mild) steel. Potential candidates could be nitrates or the so called deep eutectic solvents.
This is probably a dumb question, but do we need to use a liquid to do the salt's job? Might we just sidestep the corrosion question by using big-ass (solid) sheets of tungsten in lieu of a fluid?
The short version is that solids have merely ok heat capacity. The molar heat capacity of many solids at high temperatures is approximately 3R or approx. 25 J/molK. Google "Dulong-Petit Law" and "Debye Model" if you'd like to see why.
By comparison, the polar molecule water, which has extraordinarily good heat capacity, due in part to hydrogen bonding, is 75 J/molK at room temperature.
So water has almost 31x the heat capacity of tungsten by weight.
It may be I'm missing some obvious implication of something you've written that I've missed due to negligible background here, but I'm confused by the numbers you've cited here.
You list many solids as having a molar heat capacity of "25 J/molK", water as having a molar heat capacity of "75 J/molK", and conclude that water has almost 31x the heat capacity, but naively comparing the numbers cited only gets you 3x. Where does the other 10x factor come from?
Naive question maybe but why not use steam instead of salt? Well insulated pressure tank can hold a lot of steam, heat capacity is good, it's non-toxic, and it is ideal for driving turbines.
Focused sunlight is certainly capable of boiling water.
Do bis(oxalato)borates and tris(oxalato)phosphates have the requisite thermal stability? Uncommon anions, but made from common precursors.
It's my understanding that nitrates are potentially explosive, but I haven't heard of any other anions being used except for fluoride (which has its own corrosion/toxicity issues).
I don't really know anything about the oxaloborates, sorry. Can you synthesize them from the one of the borax hydrates? If so, what is the synthesis path?
Nitrates can of course be an oxidizing agent, so you'd like to keep them away from being e.g. carbothermically or aluminothermically reduced.
It's my understanding that alkali metal bisoxalatoborates can be prepared by heating the corresponding metaborates with oxalic acid ( https://www.google.com/patents/US7674911 ).
I know the lithium salt was of interest as a more thermally stable battery electrolyte (relative to LiPF6), although interest eventually shifted to lithium difluorooxalatoborate (two fluorines and one oxalate attached to boron tetrahedrally) which is still stable enough for batteries but not as stable as LiBOB. LiBOB is reported to decompose at about 290 C ( http://ma.ecsdl.org/content/MA2010-02/9/597.full.pdf ),
so I would expect the heavier analogs to go somewhat higher, but I've never heard of these salts being considered in the case of thermal storage. I do not know the thermal limit of trisoxalatophosphate salts, only that they look like oxalatoborates and share with nitrates the property of being "non-coordinating".
For a summary, this is an energy storage project, where the energy is stored in tanks of high and low temperature fluids. Temperature differentials can be generated by heat pumps, and can be converted back into useful energy by heat engines.
However, the theoretical maximum energy efficiency of this unit, assuming no heat losses from the tanks, will probably be poor compared to LiIon or pumped storage (of the gravity type), simply due to the fact that heat engines are strictly limited in efficiency to the Carnot efficiency (and are usually much less efficient than that).
Depending on the temperatures involved, I'd be surprised if it can return as much as half the energy that is put into it.
Also, conflation of energy and power in the article.
You're forgetting something on the Carnot cycle efficiency, which is that the efficiency of the storage cycle can be GREATER than unity. Heat pumps (of which this is an example) can move more than one joule of heat per joule of input power. I agree that the output efficiency (Joules of power : Joule of heat) is going to be low, given the relatively small temperature differential between the reservoirs, but the round-trip efficiency might not be that bad.
>However, the theoretical maximum energy efficiency of this unit, assuming no heat losses from the tanks, will probably be poor compared to LiIon or pumped storage
But Li-ion is orders of magnitude more expensive than what they're proposing (no matter how cheap the cells are getting), and couldn't possibly be deployed at this scale. It's more useful for stop-gap loads when you need instant discharge. Pumped hydro is only possible when the local geography is perfect for it and there is an abundance of water. The reality is that our future will require a host of different technologies like this working in synergy to provide grid scale energy storage. No single solution will become the standard.
EV batteries are at $227 / kwh and are on target to hit $100 per kWh in 2020. Because, costs tend to drop as production ramps up. Average american uses 35 kwh per day.
5% of US daily grid storage is an excessive goal, but assuming 100$/kwh would only cost 3,500$ / person. Assuming they last ~15 years that's under 20$/month which is clearly a deal killer, but not exactly impossible.
PS: Hydro storage is generally a much better option, but as a supplemental option batteries could be very useful.
Heat/cold storage is very efficient when you have direct consumers for heating/cooling: charge the insulated tanks in times of electrical surplus, heat/cool when needed.
Mixed application installations (e.g. campus-scale air conditioning and hot water, generously overprovisioned to allow some paid grid stabilization work on the side) might be an economic sweet spot even with terrible cycle efficiency. In any case, for grid storage, price per effective capacity will remain the key metric as long as the storage market is not saturated. Tanks scale pretty good.
Molten salt is one of the standard solar reservoir techniques. In fact there is such a system already deployed in California today in Crescent Dunes. It burns natural gas to keep the salt molten when there isn't adequate solar resource. There are some in operation in Spain as well.
It's frustrating when some reporter writes an article about one of these companies with great PR (Google, FB, Apple, Tesla et al) writing that since the reporter's never heard of this before (or of a web search either, apparently) that the company must have come up with something amazingly new.
Yes, google could cut the marginal cost of storage, but the economics of thermal storage make sense in thermal plants, not direct grid storage (I worked in thermal storage myself, though using a different technology). This is the big point, and as cheap PV has driven funding for thermal plants away it's hard to see hw this can fit in. That should have been the thrust of the article instead of breathless boosterism, especially for a magazine ostensibly aimed at the finance industry.
What is the current status of solar thermal plants in terms of commercial readiness? Are there any to pay attention to other than Ivanpah? What thermal storage technique were you working on? I may have some chemistry in this thread, but energy storage is not my day job.
There Are a few in operation in Spain and some in India that I suspect will never be operational. The economics have changed substantially just over the last five years.
I was working on supercritical (pressurized) water. Sand has also been considered.
I also think that's what they meant. Painful to read that sentence from such a large media outlet. You'd think they'd have some people on staff to help double check incorrect vocab usage...
How do they even "toss out" electricity? It's not like you can just open an overflow valve to get rid of it. Are there actually facilities for burning up useless electricity (to protect the mains frequency and thus the power plants' generators), or is the article just using strange words for "cut back production"?
For wind farms, this just involves turning the blades on their axes like "feathering" a propellor. This reduces efficiency and hence power output, potentially to zero. Turbines may also need to feather if the wind is too high to avoid exceeding max speed.
All larger systems - coal, CCGT etc - have some kind of system for throttling back input of steam or gas to turbines. Nuclear power is slightly different in that it takes a long time to cool down the reactor even once you slow the reaction, and heat is still generated from intermediate decay products. So they have "bypass" systems instead for dumping steam.
All power plants have to be able to survive the loss of their upstream output load without self-destructing. In case the wires go down.
There are better alternatives than just lithium batteries - for example liquid batteries by company with a weird name - RedFlow: https://www.youtube.com/watch?v=4OHstY_kKUY.
Edit: I also wonder how balanced heat and cold is in such system? One side effect could be cooling solar cells to improve efficiency of them. I had an idea few months back - why solar cells don't come with antifreeze that could be used to warm up your hot water in house?
I thought they did. Well, at least I know of installations that use solar power to heat water. Maybe there are combined panels that do both heat and electricity?
Fun fact, that's actually what we had in my last house. My dad wanted to heat the pool like that, so he Googled around, found a guy with a patent, and the guy helped him figure out how to build it for the house.
I see a lot of objections based on Carnot's heat cycle. What about using waste heat as the energy input? This would make the system an effectively free source of energy. Existing data centres could be retrofitted with this technology. Instead of dumping the waste heat outside with air conditioning, you store it in a fluid tank.
It's definitely possible to combine this with some sort of "co-generation" scheme, where either the heating or cooling are combined with e.g. energy flowing off of a data center.
However -- for this to work as described (hot air spins a turbine), you need to get a much higher temperature differential than would be typical for data center waste heat. You'd still need to concentrate that heat somehow into a high-temperature reservoir (think hundreds of degrees Celsius) to have it be space- and energy- efficient.
Right. Just to elaborate, there is a difference between quality and quantity of heat. A red hot nail has high quality but low quantity. A hot bath has low quality but high quantity. Data center waste heat is like bath water.
Reading this, a question comes to my mind: how cheap does an energy storage solution has to become, to finance it through speculation on energy/power exchanges (buying in times of overproduction and selling at high consumption times) ?
At least in Germany, that's exactly how most pumped hydro storage is operated, and it seems profitable. So probably about that cheap (depending on the energy market).
Construction costs for pumped hydro storage seem to be in the range of 500$/KW.
Whoever in GoogleX approved this investment wasn't aware of Carnot cycle efficiency...
Using electricity to create hot and cold reservoirs, then using the thermal difference to make electricity again cannot ever exceed ~40% efficiency. There are existing fully developed technologies which exceed that by far!
Where do I sign up to receive lots of money to try out a fanciful idea which goes against the basic laws of physics?
This is a misunderstanding of Carnot cycle efficiency. The thing that makes the Carnot cycle special is that it is reversible. So in principle, you can get 100% of the mechanical energy that you put in back out again, by running the process in reverse.
The usual efficiency calculation for a Carnot cycle tells you what fraction of heat energy that's removed from the hot reservoir can be converted to mechanical work. This is always less than 100%.
However, when running the cycle in reverse as a heat pump, it is possible to transfer more than one Joule of energy from the cold reservoir to the hot reservoir per Joule of mechanical energy used. That's why the round trip efficiency from mechanical energy back to mechanical energy can be 100%.
I have no idea what the practical efficiency of this system can be. But it's wrong to immediately reject it on theoretical grounds.
Indeed. The Carnot Cycle has nothing to do with the efficiency, since it can be reversed in what is typically referred to as a "Coefficient of Performance", which is always bigger than 1 (assuming you aren't using straight resistive/friction heating). I seem to recall from Thermos lecturers that a CoP for consumer fridges is often in the 4 - 6 range, and it increases the smaller the difference in temperature between sink and source is, inversely to the Carnot efficiency.
That said, due to heat capacity of pump components and conduction losses, I would be surprised if the total cycle efficiency was above 80%, putting this well below gravity and battery storage schemes. Maybe it is just incredibly cheap? Water does have a high heat capacity...
Even I first thought of this while reading about this plan. But may be this is more about cheap and long term storage of electricity. Lithium ion batteries are expensive and also Lithium is limited. But I would definitely like to know how this idea stacks up against the idea of pumping water to a high altitude place and then using it as a battery.
I think there's a huge amount of promise in the idea of "inside out" hydro power, where power is used to pump water out of a pressure vessel deep under water (in a lake or in the ocean) and then later generated by letting water back into the vessel. All the advantages of hydro power but without the necessity of a conveniently shaped mountain range nearby.
Last I checked, a far larger percentage of livable land is next to deep water than is next to high altitude lakes.
Also, you can use existing deep holes which were dug for other reasons - for instance, open-cut mine pits tend to end up flooded once they're no longer mined.
It doesn't have a huge amount of head, but for an example of the volume of water involved, the upper reservoir at the Ludington Pumped Storage plant is about 100 billion liters (it draws from/empties into Lake Michigan).
It's pretty huge, able to produce more than a gigawatt for ~8 hours.
Does it make sense to store energy as heat differentials for long term? Insulation can only work for so long, eventually it will come back to environment temp.
You only need to hold this energy until the next peak period hits the grid. The Snowy Mountains Hydroelectric Scheme in Australia actually buys off-peak electricity to run the turbines in reverse & pump water back into reservoirs, and then generated electricity when demand is high: effectively time-shifting power generation & arbitraging the peak/off-peak price difference.
Longer term seasonal storage would be handy too in some areas, I think using excess electricity to make hydrogen/methane/ammonia is one avenue that's being explored.
There is a lot of heat in this post about "oh, it's not efficient enough why are we considering this?" Fossil fuel generated energy varies _widely_ in terms of thermal efficiency for the tech used to do it - the cost per KwH (and secondly the engineering use case) is the real deal maker/breaker.
Coal generated energy traditionally has not been very efficient, but it's cheap cheap cheap! Likewise with your home automobile: fuel is cheap enough for us to afford continuing to drive despite the "low efficiency" of an internal combustion engine.
Lithium batteries might be coming down in price, but their cost per KwH scales linearly with the battery size. You want double the capacity or halve the price? Have fun building something twice as big.
If this design can do better than linear cost scaling and bring energy storage at household electricity prices, while also using cheap materials, it's safe to say it will show up soon at a datacenter near you.
Perhaps the choice was between a 0% efficient energy store that costs $0 (dumping the excess on the floor) versus a 35% efficient store that costs $x. The economics of wastefulness can get complicated fast.
But even so, there are an awful lot of cheap energy storage schemas that are not limited by Carnot efficiency, so it's a bit baffling why anyone would try to play Maxwell's daemon with their storage facility, when they could be dumping energy into reversible chemical reactions, as one would expect inside a chemical battery. Or maybe even a flywheel system. We're getting better at materials science all the time, and we could almost certainly dump lots of energy into flywheels without exploding them.
And the article sounds a lot like they aren't just sticking a Stirling engine between a hot well and a cold well and pairing it with a motor/generator. That's about as far as I'd go in the thermal energy storage game.
There was a SoCal company that was trying to supplant battery-based UPS installations with flywheel ones. The safety equipment and siting concerns for a small UPS were crazy compared to a battery system. And the latter usually includes a Halon fire suppression system.
I assume you mean flywheels. Flywheel energy storage seemed appealing until I worked on it in an energy storage startup.
Flywheels require expensive protection against catastrophic failure. The costly part isn't fragment containment, that's affordable.
But angular momentum is conserved. That means when a flywheel crashes, it will exert a very large torque on its containment structure. If the containment breaks loose, it can do a lot of damage in moments. Containment that's safe against torque failure is very expensive.
Flywheels also have to run in a vacuum. If the rotor is s composite material, there has to be an active vacuum pump because composites outgas forever. Additional system cost and failure point.
If you don't use composite rotors, the energy storage density goes down because tensile strength of metals is so low compared to composite fiber materials. Low energy density increases overall cost/MWHr.
Rotor life is limited by static fatigue and by high cycle fatigue, which have different failure mechanisms. For carbon, glass, or aramid fibers, you can count on a rotor MTTF of ~20 years if the rotor is operated so it never exceeds ~35% of the fiber's ultimate tensile strength. But that cuts into energy storage efficiency, driving the cost up more.
There are some niches that flywheels can serve with competitive economics (grid intersection phase matching, for example). But overall, it's hard to see how flywheel economics can beat the declining cost curve of electrochemistry. The foreseeable cost floor, set by some of the system issues I listed, is too high.
In-ground gives additional protection, but things can still get away.
Beacon Power is AFAICT the leader in deployed flywheel electrical energy storage. Their units are all underground.
One of their failures dislodged a heavy concrete and steel lid on the unit [1]. Part of their emergency shutdown routine is a water dump into the flywheel chamber to dissipate energy. The water flashed to steam and blew a multi-ton lid partly off the unit. The spinning rotor did not escape, so all was well.
You mean like flywheels? What's their efficiency like? I always thought that they needed magnetic bearing to get to higher (70+) efficiencies. But they seem cheap over the course of decades when compared to battery lifecycles and maintenance of other storage methods
Good luck searching for "google Malta" and getting decent product help. I just spent the last two nights battling to find online solutions for a login permissions error in their new "google backup and sync" product.
You would think Google would recognise the benifits of owning a namespace.
> Good luck searching for "google Malta" and getting decent product help.
Why would you search for “Google Malta” for product help when Malta is from X, not Google (searching for “x malta” gets more relevant results up front than using the wrong company name) and is a project, which has not yet produced a product (and any product may well have completely different branding.)
Plus, when it does result in a product, it'll be a utility-scale energy storage system, so I'd hope anyone needing product help isn't just doing a web search, anyway.
> You would think Google would recognise the benifits of owning a namespace.
X isn't Google, though they share the same corporate parent.
Hot and cold stores can be used to heat or cool a building as well as provide hot water. So, the efficiency could be better by not converting back to electricity.
This isn't really a climate change issue like the article suggest (and which doesn't sell well commercially) but an arbitrage idea. Store energy when it's cheap and use it when it's expensive.
It is a climate change issue in the sense that every green energy source we have is intermittent (except nuclear, whose non-viability has been discussed on HN today). If we want to run the whole world on sustainable energy sources, large scale energy storage must be an integral part of our grid.
Materials that are stable liquids at relatively high temperatures and that exhibit low vapor pressure have been investigated for quite some time as heat transfer fluids. See, for instance, the Molten Salt Reactor Experiment at Oak Ridge where a lithium fluoride-beryllium fluoride mixture was used.
The size of the phase diagram space to investigate comes from the phase rule, and the conditions for equilibrium from the Gibb's criteria. The classic materials science approach (how I learned) is to try and map the phase diagram space experimentally. This can become unwieldy for many component systems.
Common chloride salts and/or nitrate salts have been investigated as potential heat transfer/storage fluids, usually in eutectic compositions to lower the melting point. The actual ionic compositions of the molten fluids can be rather involved.
The widely used austenitic stainless steels (e.g. 304, 316) are susceptible to corrosion attack from chloride ions, so understanding of coordination chemistry around chlorine is important to determine the likely stability of a containment vessel.
It would be ideal to find salts comprised of common materials that formed moderate melting temperature eutectics, have no free chlorides (or no chlorine at all), low vapor pressure, and long term stability in the presence of common stainless (or even mild) steel. Potential candidates could be nitrates or the so called deep eutectic solvents.
Hopefully this is helpful detail,