This is WAY over my pay scale, but may be of interest: Scientists Take First Atomic Photos of Failing Battery - The British Journal
I've had to track down dendrites in a different context. Mine were forming on the surface of a printed circuit board in near-condensing humidity, shorting to a sensitive power supply feedback control node, causing a momentary high voltage condition that blew up a separate part. The high voltage also cleared the dendrite, so the power supply returned to normal operation. Once I was able to reproduce and record some events in action in an environmental chamber, and figured out just where they must be getting triggered, the manufacturing folks were able to see them at low optical magnification. And fix them by changed the soldering process to a cleaner (though more expensive) method, leaving less residual surface contamination. But that was easy compared to what these battery guys are doing.
Maybe the challenge isn't to prevent dendrites, but... what if there was a way to use nanobots to repair dendrites automatically?
