New battery made of molten metals may offer low-cost, long-lasting storage for the grid

January 13, 2016 by Nancy Stauffer
Figure 1: In this liquid metal battery, the negative electrode (top) is a low-density metal called here Metal A; the positive electrode (bottom) is a higher-density metal called Metal B; and the electrolyte between them is a molten salt. During discharge (shown here), Metal A loses electrons (e-), becoming ions (A+) that travel through the electrolyte to the bottom electrode. The electrons pass through an external circuit, powering an electric load on the way. At the bottom electrode, the Metal A ions and electrons rejoin and then alloy with the Metal B electrode.

A novel rechargeable battery developed at MIT could one day play a critical role in the massive expansion of solar generation needed to mitigate climate change by midcentury. Designed to store energy on the electric grid, the high-capacity battery consists of molten metals that naturally separate to form two electrodes in layers on either side of the molten salt electrolyte between them. Tests with cells made of low-cost, Earth-abundant materials confirm that the liquid battery operates efficiently without losing significant capacity or mechanically degrading—common problems in today's batteries with solid electrodes. The MIT researchers have already demonstrated a simple, low-cost process for manufacturing prototypes of their battery, and future plans call for field tests on small-scale power grids that include intermittent generating sources such as solar and wind.

The ability to store large amounts of electricity and deliver it later when it's needed will be critical if intermittent renewable energy sources such as solar and wind are to be deployed at scales that help curtail climate change in the coming decades. Such large-scale storage would also make today's power grid more resilient and efficient, allowing operators to deliver quick supplies during outages and to meet temporary demand peaks without maintaining extra generating capacity that's expensive and rarely used.

A decade ago, the committee planning the new MIT Energy Initiative approached Donald Sadoway, MIT's John F. Elliott Professor of Materials Chemistry, to take on the challenge of grid-scale energy storage. At the time, MIT research focused on the lithium-ion —then a relatively new technology. The lithium-ion batteries being developed were small, lightweight, and short-lived—not a problem for mobile devices, which are typically upgraded every few years, but an issue for grid use.

A battery for the power grid had to be able to operate reliably for years. It could be large and stationary, but—most important—it had to be inexpensive. "The classic academic approach of inventing the coolest chemistry and then trying to reduce costs in the manufacturing stage wouldn't work," says Sadoway. "In the energy sector, you're competing against hydrocarbons, and they're deeply entrenched and heavily subsidized and tenacious." Making a dramatic shift in power production would require a different way of thinking about storage.

Figure 2: Elements highlighted in yellow are candidates for the negative electrode, while those highlighted in green are candidates for the positive electrode. The green elements are also higher density than the yellow ones, so when they’re mixed together, the green ones will naturally form a separate layer on the bottom. By choosing among the options, designers of liquid metal batteries can optimize cost, material availability, operating temperature, cell voltage, or other characteristics to suit a particular application.

Sadoway therefore turned to a process he knew well: aluminum smelting. Aluminum smelting is a huge-scale, inexpensive process conducted inside electrochemical cells that operate reliably over long periods and produce metal at very low cost while consuming large amounts of electrical energy. Sadoway thought: "Could we run the smelter in reverse so it gives back its electricity?"

Subsequent investigation led to the battery. Like a conventional battery, this one has top and bottom electrodes with an electrolyte between them (see Figure 1 in the slideshow above). During discharging and recharging, positively charged metallic ions travel from one electrode to the other through the electrolyte, and electrons make the same trip through an external circuit. In most batteries, the electrodes—and sometimes the electrolyte—are solid. But in Sadoway's battery, all three are liquid. The negative electrode—the top layer in the battery—is a low-density liquid metal that readily donates electrons. The positive electrode—the bottom layer—is a high-density liquid metal that's happy to accept those electrons. And the electrolyte—the middle layer—is a molten salt that transfers charged particles but won't mix with the materials above or below. Because of the differences in density and the immiscibility of the three materials, they naturally settle into three distinct layers and remain separate as the battery operates.

Benefits of going liquid

This novel approach provides a number of benefits. Because the components are liquid, the transfer of electrical charges and chemical constituents within each component and from one to another is ultrafast, permitting the rapid flow of large currents into and out of the battery. When the battery discharges, the top layer of molten metal gets thinner and the bottom one gets thicker. When it charges, the thicknesses reverse. There are no stresses involved, notes Sadoway. "The entire system is very pliable and just takes the shape of the container." While solid electrodes are prone to cracking and other forms of mechanical failure over time, liquid electrodes do not degrade with use.

Indeed, every time the battery is charged, ions from the top metal that have been deposited into the bottom layer are returned to the top layer, purifying the electrolyte in the process. All three components are reconstituted. In addition, because the components naturally self-segregate, there's no need for membranes or separators, which are subject to wear. The should perform many charges and discharges without losing capacity or requiring maintenance or service. And the self-segregating nature of the liquid components could facilitate simpler, less-expensive manufacturing compared to conventional batteries.

Choice of materials

For Sadoway and then-graduate student David Bradwell MEng '06, PhD '11, the challenge was to choose the best materials for the , particularly for its electrodes. Methods exist for predicting how solid metals will behave under defined conditions. But those methods "were of no value to us because we wanted to model the liquid state," says Sadoway—and nobody else was working in this area. So he had to draw on what he calls "informed intuition," based on his experience working in electrometallurgy and teaching a large freshman chemistry class.

To keep costs down, Sadoway and Bradwell needed to use electrode materials that were earth-abundant, inexpensive, and long-lived. To achieve high voltage, they had to pair a strong electron donor with a strong electron acceptor. The top electrode (the electron donor) had to be low density, and the bottom electrode (the electron acceptor) high density. "Mercifully," says Sadoway, "the way the periodic table is laid out, the strong electropositive [donor] metals are low density, and the strong electronegative [acceptor] metals are high density" (see Figure 2 in the slideshow above). And finally, all the materials had to be liquid at practical temperatures.

As their first combination, Sadoway and Bradwell chose magnesium for the top electrode, antimony for the bottom electrode, and a salt mixture containing magnesium chloride for the electrolyte. They then built prototypes of their cell—and they worked. The three liquid components self-segregated, and the battery performed as they had predicted. Spurred by their success, in 2010 they, along with Luis Ortiz SB '96, PhD '00, also a former member of Sadoway's research group, founded a company—called initially the Liquid Metal Battery Corporation and later Ambri—to continue developing and scaling up the novel technology.

Not there yet

But there was a problem. To keep the components melted, the battery had to operate at 700 degrees Celsius (1,292 degrees Farenheit). Running that hot consumed some of the electrical output of the battery and increased the rate at which secondary components, such as the cell wall, would corrode and degrade. So Sadoway, Bradwell, and their colleagues at MIT continued the search for active materials.

Early results from the magnesium and antimony cell chemistry had clearly demonstrated the viability of the liquid metal battery concept; as a result, the on-campus research effort received more than $11 million from funders including Total and the U.S. Department of Energy's ARPA–E program. The influx of research dollars enabled Sadoway to grow the research team at MIT to nearly 20 graduate and undergraduate students and postdocs who were ready to take on the challenge.

Within months, the team began to churn out new chemistry options based on various materials with lower melting points. For example, in place of the antimony, they used lead, tin, bismuth, and alloys of similar metals; and in place of the magnesium, they used sodium, lithium, and alloys of magnesium with such metals as calcium. The researchers soon realized that they were not just searching for a new battery chemistry. Instead, they had discovered a new battery "platform" from which a multitude of potentially commercially viable cell technologies with a range of attributes could spawn.

New cell chemistries began to show significant reductions in operating temperature. Cells of sodium and bismuth operated at 560 degrees Celsius. Lithium and bismuth cells operated at 550 C. And a battery with a negative electrode of lithium and a positive electrode of an antimony-lead alloy operated at 450 C.

While working with the last combination, the researchers stumbled on an unexpected electrochemical phenomenon: They found that they could maintain the high cell voltage of their original pure antimony electrode with the new antimony-lead version—even when they made the composition as much as 80 percent lead in order to lower the melting temperature by hundreds of degrees.

"To our pleasant surprise, adding more lead to the antimony didn't decrease the voltage, and now we understand why," Sadoway says. "When lithium enters into an alloy of antimony and lead, the lithium preferentially reacts with the antimony because it's a tighter bond. So when the lithium [from the top electrode] enters the bottom electrode, it ignores the lead and bonds with the antimony."

That unexpected finding reminded them how little was known in this new field of research—and also suggested new cell chemistries to explore. For example, they recently assembled a proof-of-concept cell using a positive electrode of a lead-bismuth alloy, a negative electrode of sodium metal, and a novel electrolyte of a mixed hydroxide-halide. The cell operated at just 270 C—more than 400 C lower than the initial magnesium-antimony battery while maintaining the same novel cell design of three naturally separating liquid layers.

The role of the new technology

The liquid metal battery platform offers an unusual combination of features. In general, batteries are characterized by how much energy and how much power they can provide. (Energy is the total amount of work that can be done; power is how quickly work gets done.) In general, technologies do better on one measure than the other. For example, with capacitors, fast delivery is cheap, but abundant storage is expensive. With pumped hydropower, the opposite is true.

But for grid-scale storage, both capabilities are important—and the liquid metal battery can potentially do both. It can store a lot of energy (say, enough to last through a blackout) and deliver that energy quickly (for example, to meet demand instantly when a cloud passes in front of the sun). Unlike the lithium-ion battery, it should have a long lifetime; and unlike the lead-acid battery, it will not be degraded when being completely discharged. And while it now appears more expensive than pumped hydropower, the battery has no limitation on where it can be used. With pumped hydro, water is pumped uphill to a reservoir and then released through a turbine to generate power when it's needed. Installations therefore require both a hillside and a source of water. The liquid metal battery can be installed essentially anywhere. No need for a hill or water.

Bringing it to market

Ambri has now designed and built a manufacturing plant for the liquid metal battery in Marlborough, Massachusetts. As expected, manufacturing is straightforward: Just add the electrode metals plus the electrolyte salt to a steel container and heat the can to the specified operating temperature. The materials melt into neat liquid layers to form the electrodes and electrolyte. The cell manufacturing process has been developed and implemented and will undergo continuous improvement. The next step will involve automating the processes to aggregate many cells into a large-format battery including the power electronics.

Ambri has not been public about which liquid metal battery chemistry it is commercializing, but it does say that it has been working on the same chemistry for the past four years. According to Bradwell, Ambri scientists and engineers have built more than 2,500 cells and have achieved thousands of charge-discharge cycles with negligible reduction in the amount of energy stored. Those demonstrations confirm Sadoway and Bradwell's initial thesis that an all-liquid battery would be poised to achieve better performance than solid-state alternatives and would be able to operate for decades.

Ambri researchers are now tackling one final engineering challenge: developing a low-cost, practical seal that will stop air from leaking into each individual cell, thus enabling years of high-temperature operation. Once the needed seals are developed and tested, battery production will begin. The researchers plan to deliver prototypes for field testing in several locations, including Hawaii, where sunshine is abundant but power generation still relies on burning expensive diesel fuel. One site is the Pearl Harbor naval base on Oahu. "It's unsettling that our military bases rely on the civilian power grid," says Sadoway. "If that grid goes down, the base must power up diesel generators to fill the gap. So the base can be without power for about 15 minutes, which is probably enough time for some major damage to be done." The new battery could play a key role in preventing such an outcome.

Meanwhile, back at the lab, the MIT researchers are continuing to explore other chemistries for the core of the liquid battery. Indeed, Sadoway says that his team has already developed an alternative design that offers even lower operating temperatures, more stored energy, lower cost, and a longer lifetime. Given the general lack of knowledge about the properties and potential uses of liquid metals, Sadoway believes there could still be major discoveries in the field. The results of their experiments "kicked open the doors to a whole bunch of other choices that we've made," says Sadoway. "It was really cool."

Explore further: Liquid Battery Offers Promising Solar Energy Storage Technique

More information: David J. Bradwell et al. Magnesium–Antimony Liquid Metal Battery for Stationary Energy Storage, Journal of the American Chemical Society (2012). DOI: 10.1021/ja209759s

Hojong Kim et al. Liquid Metal Batteries: Past, Present, and Future, Chemical Reviews (2013). DOI: 10.1021/cr300205k

Xiaohui Ning et al. Self-healing Li–Bi liquid metal battery for grid-scale energy storage, Journal of Power Sources (2015). DOI: 10.1016/j.jpowsour.2014.10.173

Brian L. Spatocco et al. Low-Temperature Molten Salt Electrolytes for Membrane-Free Sodium Metal Batteries, Journal of The Electrochemical Society (2015). DOI: 10.1149/2.0441514jes

Kangli Wang et al. Lithium–antimony–lead liquid metal battery for grid-level energy storage, Nature (2014). DOI: 10.1038/nature13700

Related Stories

Liquid batteries could level the load

February 14, 2012

The biggest drawback to many real or proposed sources of clean, renewable energy is their intermittency: The wind doesn’t always blow, the sun doesn’t always shine, and so the power they produce may not be available ...

Melting, coating, and all-solid-state lithium batteries

January 1, 2016

The joint research team of Prof. Yoon Seok Jung (UNIST, School of Energy and Chemical Engineering) and Prof. Seng M. Oh (Seoul National University) discovered a new way to develop all-solid-state lithium batteries without ...

A novel liquid metal battery for grid-level storage

September 4, 2014

The challenge of selling any new idea is that it has to compete with every other new idea. The process is more difficult when the idea's technology hasn't existed and addresses an issue that some industries don't see as a ...

Recommended for you

Making it easier to collaborate on code

October 26, 2016

Git is an open-source system with a polarizing reputation among programmers. It's a powerful tool to help developers track changes to code, but many view it as prohibitively difficult to use.

Dutch unveil giant vacuum to clean outside air

October 25, 2016

Dutch inventors Tuesday unveiled what they called the world's first giant outside air vacuum cleaner—a large purifying system intended to filter out toxic tiny particles from the atmosphere surrounding the machine.


Adjust slider to filter visible comments by rank

Display comments: newest first

5 / 5 (9) Jan 13, 2016
Very exciting. Physorg has been covering liquid-metal batteries for at least six years, and the technology has progressed from the conceptual stage to nearly ready for demonstration plants. I'm looking forward to the next chapter.
2 / 5 (6) Jan 13, 2016
Would it be possible to ditch the metal entirely and use salts or organic molecules in a solvent separated by their densities?

The efficiency of the device is paramount, because any energy lost multiplies the cost of the input energy. It doesn't pay to have a cheap battery that loses half the energy put through it, because then the output is twice as expensive as the input.

Another interesting piece of fact is that something around 40% of rivers have already been dammed for hydropower, leaving that approach for energy storage rather limited. One could use seawater, but most seaside is flat and low, so there isn't anywhere you can pump it.

5 / 5 (8) Jan 13, 2016
Excellent news - and an example of the kick into the advancement of renewable energy systems that the imperative of AGW has provided.
5 / 5 (5) Jan 13, 2016
Good stuff, but according to their own website, the seals are more of a problem than they had anticipated and commercialisation of the technolgy has been delayed for that reason.
The underlying battery chemistry is apparently sound and they're continuing to improve it.
3.3 / 5 (4) Jan 13, 2016
How much energy will be consumed in keeping the materials at the proper temperature?
5 / 5 (6) Jan 13, 2016
The efficiency of the device is paramount

There are many critical factors that are going in to the study of upcoming storage systems. Cost, efficiency, energy density, charge recharge cycles, how long the energy can be stored for etc. etc. All of these factors will determine the optimum system for any situation. A system that is only 50% as efficient - but has 10X the charge recharge cycles - may win out in the big calculation of optimal system.
5 / 5 (4) Jan 13, 2016
I suspect a lot also depends on insulating the vessels using aerogels and or vacuum. We have to petrochemical industry to thank for their development. Aerogels have been insulating LNG for decades.
For the seals it would be nice to find some regenerating/self-healing polymer between silicone rubber and glass.
3.5 / 5 (4) Jan 13, 2016
A system that is only 50% as efficient - but has 10X the charge recharge cycles - may win out in the big calculation of optimal system.

A system that is only 50% efficient may be replaced by many other things than batteries. Even compressed air can get there much cheaper than batteries.

But that was hardly the point. The optimal system is one which does not increase the cost of energy significantly, which counts the cost of the system plus the efficiency losses -combined-.

It's already hard enough to make a battery that is cheap enough to build in quantity to not add a dollar per kWh to your power bill, but if you solve that problem at the cost of halving your efficiency, you're still paying more than double the price for the electricity you actually get, and that is still not economically viable except in niche applications.

5 / 5 (6) Jan 13, 2016
A system that is only 50% efficient may be replaced by many other things than batteries.
I said a 'system', I did not say a 'battery'. Counter to your assertion - the point was very clear. The efficiency of the device is not paramount - there are many factors that will go into the decision regarding the appropriate choice of storage system for a particular situation. Lower efficiency with higher charge discharge cycles, or lower initial cost may win out in some situations. We are in interesting times.
5 / 5 (3) Jan 13, 2016
I think they could use concentrated solar heat to keep the batteries warm. Then they need not to pay for the heat.

Another pumped storage beside pumped hydro could be pumped air. Best would be to use off-shore wind to pump air into concrete domes situated at the bottom of ocean. And then let the air back up when there is no wind. If they combine that with Kinetic Power Plant (KPP) design by Rosch Innovations AG, that should have great efficiency. Underwater KPP might even work alone without that wind turbine.
5 / 5 (3) Jan 14, 2016
Just an idea, how about

Negative electrode alloy (eutectic alloy 1)
K 76.70%
Na 23.30%

Positive electrode alloy (eutectic alloy 2)
Bi 40.30%
Pb 22.20%
In 17.20%
Sn 10.70%
Cd 8.10%
Tl 1.10%

Operating temperature of about 42°C :)

I don't know how well it would cycle nor it's voltage/amperage performance but it's running temperature would offer a low parasitic load to keep it in it's liquid phase. As a bonus you could utilize a mild solar collector installation coupled to the battery array combined with passive phase change materials as a thermal bank to supply that needed energy for maintaining the liquid property of that battery system,

Might be worth a try anyway. ;)
1 / 5 (5) Jan 14, 2016
Now this is real science! Not the bogus presto-science of Global Warming computer models!
5 / 5 (1) Jan 14, 2016
What about the molten salt? Isn't that necessary to make these things chug along or could any liquid electrolyte be used at 42ºC
5 / 5 (3) Jan 14, 2016
I think that the main take-home here is that this tech is scalable, and these batteries of batteries can be built to store ever greater and greater amounts of electrical energy, as its production increases with ever more and more generating capacity ibeing brought online.

Any loss of energy associated with heating the cells will amount to only a drop in the bucket of all that cheap, renewables(mainly wind/solar) power.

If you haven't already, it's past time to begin divesting from fossil fuels.

Also past time to stop the antiscience, denierside trolling as cover for the fear of imminent obsolescence of BigCarbon.

5 / 5 (3) Jan 15, 2016
What about the molten salt? Isn't that necessary to make these things chug along or could any liquid electrolyte be used at 42ºC


A couple examples below are proton ion transport room temperature liquid electrolytes and I'm not sure how they would (or could) perform with the alkali metal ions as I'm not qualified in that area of battery science/chemistry.

ethylamine nitrate (C₂H₈N₂O₃) is a liquid at and above 12°С
1-Butyl-3-methylimidazolium Triflimide (C₁₀H₁₅F₆N₃O₄S₂) is a liquid at and above 27°С
plus many many many many others;

Something of a bonus would be the battery's internal resistance as this would also contribute to the necessary thermal requirements to keep it in liquid phase with the possibility of no external heat source, that is to say if it's possible to operate it in the 0°С to 100/150°С regime.

not rated yet Jan 15, 2016
@Eikka ...but if you solve that problem at the cost of halving your efficiency, you're still paying more than double the price for the electricity you actually get, and that is still not economically viable except in niche applications.
The topic is about opportunity load. The energy does not cost more. it's energy that's now wasted
not rated yet Jan 16, 2016
Hook em up to the waste heat from LFTR.

Maybe even run the salt through a rig like this...though, I'm sure no one has ran molten radioelectric batteries.
not rated yet Jan 17, 2016
I wonder how sodium and mercury would work? Sodium melts at 208° F, while everyone knows that mercury is liquid at room temperature.
Just keep the water out of it. Somehow, I don't think burning sodium and vaporized mercury would be good for humans and other living things.
1 / 5 (1) Jan 17, 2016
I had to stop reading the article when the guy starts yakking about how "... you're competing against hydrocarbons, and they're deeply entrenched and heavily subsidized...."

I've seen this said before, but I can't think of how it's possible that hydrocarbons are heavily subsidized. It's just the opposite, I would think.
5 / 5 (2) Jan 18, 2016
I had to stop reading the article when the guy starts yakking about how "... you're competing against hydrocarbons, and they're deeply entrenched and heavily subsidized...."

I've seen this said before, but I can't think of how it's possible that hydrocarbons are heavily subsidized. It's just the opposite, I would think.

Think again.

Global Fossil fuel subsidies in 2013: $548bn
(For comparison: Global subisides for renewables: $88bn*)

* One could argue that fossil fuel subsides per kwH produced is lower than for renewables. What this argument is missing, though, is that fossil fuels have had more than 150 years(!) to become a self sufficient (read: no-subsidy) business - but has not managed to do so yet (and therefore likely never will)...while renewables are well on track to become so mid century.
1 / 5 (1) Jan 18, 2016
Well, I stand corrected.

However, what if we limit the subsidy discussion to the USA (which was not listed on that website's handy spreadsheet)? Would we find a net subsidy of hydrocarbons, or would we find it's had the crap taxed out of it at every level?

not rated yet Jan 19, 2016
This seems to complement some work in nuclear engineering. Pyrolysis is a way to electrowin uranium and plutonium from a bath of cesium and probably other salts. So I wonder whether designing the process around a bimetallic liquid charger could help. There is no shortage of heat in a nuclear reactor to keep up the liquid state, and spent fuel rods or semireprocessed fuel could provide heat outside the reactor block. You would half to drain out the two metal layers once in a while and possibly recover or store electricity by switching around opposing layers with different metals.

As for air sealing, that is a familiar issue with sodium cooled reactors. I also have to ask whether heat rather than electricity flow can covern at least some of the charge-discharge from the Sadoway battery.

I'd love to hear from some physicists out there.

Please sign in to add a comment. Registration is free, and takes less than a minute. Read more

Click here to reset your password.
Sign in to get notified via email when new comments are made.