How computer models can lead to battery breakthroughs

Tapping into computer models that can simulate chemical reactions between materials could significantly speed up the pace of progress for battery innovation. At least that’s one of the ideas behind the early work at battery startup Pellion Technologies, which is working on developing a magnesium battery.

Pellion, backed by the Department of Energy’s ARPA-E program and Khosla Ventures, developed advanced algorithms and computer modelling that enabled it to test out 10,000 potential cathode materials to fit with its magnesium anode for its battery, the company told me at the ARPA-E conference this week. A battery is made up of an anode on one side, a cathode on the other and an electrolyte in between. Ions travel from the anode to the cathode through the electrolyte, creating a chemical reaction that allows electrons to be harvested along the way.

Pellion is an early-stage company, spun out of MIT, but the company says its magnesium batteries could have very high energy density — higher than current lithium ion batteries — which could enable it to one day disrupt the electric vehicle market and even consumer electronics and cell phones industries.

Founders of Pellion, MIT Professor Gerbrand Ceder, also helped develop The Materials Genome Project at MIT, which is a program based on using computer modelling and virtual simulations to deliver innovation in materials. The Economist once described Ceder’s work with The Materials Genome Project as “a short cut” for discovering electrodes and the interactions of inorganic chemical compounds.

Ceder described the method in a paper and presentation back in 2009 as:

With such an approach, one can, in principle, predict the behavior of novel materials without the need to synthesize them and create a virtual design laboratory. . . Using a high-throughput computational environment, coupled to a database of all known inorganic materials, basic information on all known inorganic materials and a large number of novel “designed” materials is being computed.

Image of 25,000 compounds synthesized and modeled by MIT’s Materials Genome Project.

Photo courtesy of Paul’s Lab.