Liquid Metal Batteries

Dr. Donald Sadoway, as chemistry professor at MIT, is developing a new type of energy storage device called a liquid metal battery for grid storage applications. The conceptual physical form of this battery is the simplest imaginable: In the charged stated it consists of three liquid layers lying on top of one another. One layer is an electropositive metal, one layer is an electrolytic salt, and one is an electronegative metal. The layers remain separate due to the different densities of the three liquids and due to inherent low solubility of the metals in the liquid salt.

The atoms of electropositive metals like to give up electrons and become positively charged ions. The atoms of electronegative metals like to gain electrons and become negatively charged. If the two metals were poured together they would form an allow give up energy in the process. In a simple mixing process this energy would take the form of heat. In the battery the energy is extracted in the form of electrical current. A conducting current collector is inserted into each metal layer. If the two collectors are connected by a conduction path atoms of the electropositive metal give up electrons and become positively charged ions. The electrons flow through the external path and are picked up by atoms of the electronegative metal which become negatively charged ions. Positive metal ions are transported through the electrolytic salt and the two metals form an alloy. Unlike a chemical compound an alloy can continuously vary the composition of its metallic components so that this process should not be thought of as two ions joining together to form a bimetallic molecule.

The volume of metal on one side of the liquid electrolyte shrinks while the volume of the alloy on the other side increases. Since all of the layers are liquid this change volume of the electrodes does not present any structural problem as often occur in batteries with solid electrodes.

This lack of a solid electrode which whose mass increases and decreases during the charge/discharge cycles avoids certain failure mechanisms associated with solid electrodes and may thus lead to a long cycle life for this battery design. This advantage is similar to that of flow batteries in which the solid electrodes are purely catalytic and do not gain or lose mass during battery operation. Flow batteries utilize an ion exchange membrane, usually nafion, the same membrane used in PEM fuel cells as a proton conductor. The fact that the liquid metal battery avoids the use of these rather expensive membranes is a potential advantage relative to flow batteries or relative to sodium sulphur batteries which use an expensive solid ceramic electrolyte.

Conceptually the liquid metal battery has as simple a design for an electrochemical cell as can well be conceived. The hope is that if one can find low cost materials for each of the three layers then the manufacturing cost of this battery could be very low.

One disadvantage of this battery design is the high operating temperature required to keep all the layers in a liquid state. In this paper Sadoway and colleagues describe a liquid metal battery based on magnesium (Mg) and antimony (Sb) as the metal electrodes and MgCl2-KCl-NaCl as the molten salt electrolyte. The operating temperature was 700C. In the liquid state at these high temperature these materials are extremely corrosive. They can potentially corrode the current collectors and the steel casing of the battery. Both long term reliability and safety are potentially compromised by this corrosivity. A leaking battery case at 700C is no laughing matter. Ask NGK Insulators whose sodium sulphur battery technology operating at a mere 350C underwent a major setback due a leak and a major battery fire.

A startup company called Ambri has been founded to develop a commercial version of the liquid metal battery design. They have abandoned Mg/Sg chemistry in favor of some other metal couple which they are keeping a secret. Ambri claims to have strong patent position since they are the only company working on this battery architecture, but apparently it is not strong enough for them to reveal any details about the chemistry of their latest design. According to this news story Ambri is targeting its first prototype deployment some time in 2014. According to their product brochure the battery cells will have a square shape with the side of the square being somewhere between 10cm and 20cm in length. The height of the cell will be significantly less than the width. A number of these cells will stacked together into a 200kWh storage unit called a core.

I am somewhat surprised at the relatively small cell size since the simplicity of this battery architecture would seem to lend itself to very large format cells. Perhaps the relatively small size is related to an aspect of cell operations that I have not seen discussed: cold starting. The cell are supposed to generate enough heat during operation to maintain themselves above the melting temperature of the battery layers, but I do not think that the cells will be shipped to their final destination at high temperature with the metals sloshing around inside. Furthermore if there is any down time at all after the batteries have been installed the battery will cool and solidify. Therefore some provision must be made for heating the battery in order to melt the constituent prior to a startup. The smaller cell size with a higher surface to volume ration may be more convenient for such a cold startup.

If Ambri really installs a working prototype next year this would represent a remarkably short development cycle for a new battery technology. However, I think it would be premature to declare victory for Ambri. They have not yet delivered a working storage installation, and even after they do so there are significant questions about the long term reliability and safety of such high temperature devices which it may take many years of successful operation to answer adequately.

Nov 23, 2013

Energy Storage News

rogerkb [at] energystoragenews [dot] com