In the lab at least, its materials are stable for over 1,000 cycles.
Current lithium batteries are based on intercalation—lithium ions squeeze into spaces within electrode materials such as graphite. As a result, most of the battery's volume and bulk is dedicated to things that don't contribute to carrying charges between the electrodes, which sets a limit on the sorts of energy densities that these technologies can reach. These limits have led to a lot of research into finding ways to get rid of one these electrode materials. People have tried pairing lithium-metal electrodes with various materials, while other efforts have tried using electrodes where lithium reacts with air to form lithium-oxygen compounds. While these worked by some measures, they tended to have problems that drastically shortened their useful lifetimes. But a recent paper describes a battery that uses lithium metal at one electrode and lithium air for the second. By some measures, the battery has decent performance out to over 1,000 charge/discharge cycles. Lots of problems The problems with lithium metal are pretty well described: It's very difficult to get the lithium to deposit evenly across the surface of the electrode. Over repeated charge/discharge cycles, things that start as subtle irregularities grow into spines called dendrites that the lithium doesn't leave in order to carry charge; eventually, the spines grow until they short the system out. The solution is generally thought to be changes to the electrolytes that the lithium ions travel through when moving between electrodes. At least…