Molten-Salt Batteries

a (slightly) deeper dive

I treated molten-salt batteries as a single battery type in my big battery chemistries post, but they’re actually more diverse than that, and deserve their own post.

What is a molten-salt battery anyway?

Molten-salt batteries use molten salts as an electrolyte between two electrodes.

A salt is an ionic compound between elements on opposite sides of the periodic table — in other words, something really electropositive bonded to something really electronegative. For instance, table salt, NaCl, is formed by bonding sodium, which has one electron in its outer shell, with chlorine, whose outer shell is missing one electron.

The elements that make up a salt are unstable and really “want” to get together; in less anthropomorphic terms, making a salt releases energy.

This makes salts attractive battery materials. You store energy by keeping the elements separate; you release energy by letting them come together and form ionic bonds. Usually, this reaction only happens at high temperatures, when the salt is molten.

And that gives us the basic premise of a molten-salt battery. The electrodes are made of reactive materials that “want” to form an ionic bond with each other. The electrolyte is made of some material that allows ions but not atoms to pass through. Ions flow through the electrolyte to combine into a salt, discharging energy; then, to go “uphill” you have to spend energy to separate the ions.

This is electrochemical energy storage, which is distinct from thermal storage, though you can use molten salt for both. In thermal energy storage, you simply spend energy to heat up a material, and release energy by cooling it down again (e.g. by letting hot material produce steam to spin a turbine).

Molten salt is used as short-term energy storage to help solar farms store power overnight; during the day, the sun’s heat raises the temperature in a tank of molten salt, which can then be used to produce steam to make electricity.

That’s not the kind of “molten salt battery” we’re talking about in this post. We’re looking at true batteries, where the energy is stored in the ions that make up salt being separated, rather than in the temperature of the salt.

Where would you use a molten-salt battery?

Compared to lithium-ion batteries, molten-salt batteries are usually comparable or superior in energy density, and superior in safety.

Molten-salt batteries are more expensive than lithium-ion batteries per kilowatt-hour in all cases commercialized enough to have actual list prices, but some molten-salt batteries (like the sodium-zinc-chloride ones in development by Solstice Battery) have much lower materials costs than lithium-ion, which could lead to lower prices in the long run.

They all need to operate above room temperature, spanning a wide range depending on the specific battery chemistry (from 194 degrees F to 1292 degrees F).

Most molten-salt batteries (except for the oldest type, sodium-sulfur molten-salt batteries) are nonflammable, and contain no lithium or rare minerals.

The primary intended application of molten-salt batteries to the clean energy transition is in long-term grid storage, as a complement to variable energy sources such as solar or wind.

Could molten-salt batteries also be used in electric vehicles? Some researchers are aiming there, so it’s not obviously ruled out.

(One concern I initially had, that you couldn’t put a hot battery in a vehicle, turns out to be unfounded. Ordinary internal combustion engines contain gases that reach temperatures of 2800 degrees F, above the melting point of steel. Cooling systems get the walls of the engine down to a more modest 265-465 degrees F. Putting hot things inside cars is a solved problem.)

Molten Salt Battery Types

Sodium-Sulfur

First researched by the Ford Motor Company in the 1960s, sodium-sulfur batteries were commercialized in the 1980s by the Japanese company NGK Insulators, who still produce them today.

The anode is molten sodium; the cathode is molten sulfur; the electrolyte between them is a solid called BASE that allows positive ions to flow but not electrons. During discharge, sodium ions flow through the BASE to the sulfur to make sodium sulfide, which is the “salt.”

Sodium-sulfur batteries have great energy density, at 760 Wh/kg.

The problem with sodium-sulfur batteries is that they like to explode.

ZEBRA (Sodium-nickel-chloride)

ZEBRA batteries were a successor technology to sodium-sulfur, developed in the 1980s to be cheaper and safer.

The anode is still molten sodium; the cathode is nickel chloride; the electrolyte is again a solid that allows ions but not electrons to flow through. The salt formed when ions flow is NaCl, table salt.

ZEBRA batteries were sold by GE under the brand name Durathon, for use in train locomotives and some backup power applications. But GE shut down its Schenectady manufacturing plant for ZEBRA batteries in 2015 and the project seems to be defunct today.

Some smaller companies, like Germany’s innov.energy or Switzerland’s Fiamm Sonick, sell ZEBRA batteries for backup power applications.

Sodium zinc-chloride batteries are similar to ZEBRA batteries but use cheaper zinc in place of nickel, so they are expected to have much lower materials costs. They are currently at prototype stage, and being developed by an EU industry-academic consortium under the name Solstice Battery.

Liquid-Metal

MIT professor Donald Sadoway has developed a range of liquid-metal batteries (magnesium-antimony, calcium-antimony, lead-antimony) where both electrodes are molten metals, with a molten salt electrolyte between them.

All-liquid batteries have several advantages: liquid electrodes aren’t subject to cracks or wear and tear, they don’t require solid separators because they naturally segregate into layers based on density, and the ions flow fast.

Sadoway has spun out a company called Ambri to commercialize the tech for grid storage, using calcium-antimony batteries.

Sodium-Ion

Sumitomo Electric has developed a sodium-ion battery intended for use in electric vehicles, containing a solid sodium anode, a sodium compound cathode, and a mixed molten salt electrolyte through which sodium ions pass.

This prototype sodium-ion battery can operate at near-ambient temperatures, uses no rare elements, is nonflammable, and has a fairly high energy density (224 Wh/kg) and cycle life (at least 1000 cycles.)

Unfortunately, the project appears to be defunct.

Aluminum-Nickel Freeze-Thaw

Researchers at the Pacific Northwest National Laboratory prototyped a molten-salt battery intended for grid storage with a “freeze-thaw” structure.

The anode is solid aluminum; the cathode is solid nickel. The salt electrolyte allows ions to move through it when molten, but traps ions when solid. So the battery can store energy for a long time, making it ideal for seasonal energy storage.

Sodium-Iodide

Researchers at the Sandia National Lab have prototyped an all-liquid molten salt battery that operates at the relatively low temperature of 230 degrees F.

The anode is molten sodium; the “catholyte” (cathode + electrolyte) is a mixture of sodium iodide and gallium chloride; during discharge, sodium ions pass through a ceramic separator to combine with iodide ions and form sodium iodide as a salt.

Gallium chloride is expensive, though, which limits the applicability of this particular chemistry.

Performance Comparisons

I compared battery chemistry types in this spreadsheet. I’ve left cells blank whenever information was unavailable.

Energy Density

Energy density, measured in watt-hours per kilogram, is especially important for electric vehicle applications; higher energy density is better because it means more storage in less weight.

Sodium-sulfur is by far the highest energy density option, but most molten-salt batteries have as least as high energy densities as the most common kinds of lithium-ion batteries.

Materials Cost

How expensive are molten-salt batteries? For types of molten-salt battery that are still at the research or startup stage, the market price isn’t known, so the best apples-to-apples comparison across batteries is the cost of materials.

The range of materials costs is huge, but sodium-zinc-chloride batteries, some liquid-metal batteries, and some others in academic development at Chinese universities are at or near the $70-80/kWh raw materials cost they’d need to compete with lithium-ion.

Cycle Life

Cycle life is the number of times a battery can be charged and discharged and still retain an (almost) full charge. This metric is especially important for grid storage batteries, which will need to be replaced infrequently.

Here, I’ve showed the longest cycle life I could find evidence for; especially for early-stage technology, just because the paper didn’t demonstrate battery cycling for very many charge-discharge cycles doesn’t mean we know the battery can’t last longer.

The battery types that have demonstrated long enough cycle life to be usable in a grid-storage context are Sadoway/Ambri’s calcium-antimony liquid metal battery, old-school ZEBRA batteries, and a prototype aluminum-graphite battery from the University of Cambridge.

Operating Temperature

Molten-salt batteries all need to be hotter than room temperature to operate, but there’s a wide range of minimum operating temperatures, from under 200 degrees to over 1200. Lower operating temperatures are better because heat management (insulation, cooling, etc) is more expensive and less reliable at high temperatures.

The low-temperature molten-salt batteries include Sumitomo’s sodium-ion battery, PNNL’s aluminum-nickel and Sandia’s sodium-iodide batteries, and several other batteries in academic development at the prototype stage.

Notably, the liquid-metal batteries are not low-temperature.

So, Is This A Thing?

It all depends on price.

Current sticker prices for the already-commercialized molten salt battery types (sodium-sulfur and ZEBRA) are way too high to compete with lithium-ion, at $600/kWh. (And lithium-ion at today’s prices is itself too expensive for grid storage. A grid storage technology probably needs to get costs down to $50/kWh to see initial deployment without reliance on subsidies).

In principle, those prices could drop with economies of scale, and in principle batteries made of cheap materials and known manufacturing processes (like sodium-zinc-chloride) should ultimately be a lot cheaper than lithium-ion.

But we don’t know if they’re going to get there until molten-salt battery manufacturers actually invest in scaling production and deployment.

Thanks to Brady Hauth for feedback.