Whiskers, surface growth and dendrites in lithium batteries

　　Assistant Professor Pen Bai from Washington University in St. Louis has identified three key current boundaries when it comes to creating energy-dense lithium metal batteries. According to the Assist Prof, engineers had been looking for one solution to what turns out to be three problems.

　　Figure A

　　A lithium-ion battery is made of three layers: one layer of low-voltage material (graphite) called the anode; one of high-voltage material (lithium cobalt oxide) called the cathode; and a layer of porous plastic which separates the two.

　　The separator is wetted by a liquid called an electrolyte. When the battery discharges, lithium-ions empty out of the anode, passing through the liquid electrolyte, and move into the cathode. The process is reversed as the battery charges.

　　"With half of the lithium-ion-hosting electrode materials empty at all times," Assist Prof. Bai said, "you are wasting half of your space."

　　Engineers have known that they could build a more energy-dense battery (a smaller battery with a similar output capabilities) by discarding some of the dead weight that comes with half of the host materials always being empty. They have been minimally successful by removing the graphite anode, then reducing the lithium-ions with electrons during recharge, a process which forms a thin plating of lithium metal.

　　"The problem is that the lithium metal plating is not uniform," Assist Prof. Bai said. "It can grow 'fingers.'”

　　Researchers have referred to these fingers as "dendrites." As they spread from the lithium metal plating, they can penetrate the separator in the battery, leading to a short circuit.

　　But not all fingers are the same. "If you call them all dendrites, you're looking for one solution to solve actually three problems, which is impossible," Assist Prof. Bai said. "That's why after so many years this problem has never been solved."

　　His team has identified three distinct types of fingers, or growth modes, in these lithium metal anodes. They also outline at which current each growth mode appears.

　　"If you use very high current, it builds at the tip to produce a treelike structure," Bai said. Those are "true dendrites" (see Figure A). Below the lower limit you have whiskers growing from the root (see Figure B).

　　Figure B

　　And, within those two limits there exists the dynamic transition from whiskers to dendrites, which Assist Prof. Bai calls "surface growth" (see Figure C).

　　These growths are all related to the competing reactions in the region between the liquid electrolyte and the metal deposits.

　　The study found that a nanoporous ceramic separator can block whiskers up to a certain current density, after which surface growths can slowly penetrate the separator. With a strong enough current, true dendrites form, which can easily and very quickly penetrate the separator to short the battery.

　　At this point, Assist Prof. Bai said, "Our unique transparent cell revealed that the voltage of battery could look quite normal, even though the separator has been penetrated by a lithium metal filament. Without seeing what is happening inside, you could be easily fooled by the seemingly reasonable voltage, but, really, your battery has already failed."

　　Figure C

　　In order to build a safe, efficient, reliable battery with a lithium metal anode, the three growth modes need to be controlled by three different methods.

　　This will be a challenge considering consumers want batteries that can store more energy, and at the same time want them to be charged more quickly. The combination of these two inevitably yields a higher and higher charging current, which may exceed one of the critical currents identified by Assist Prof. Bai's team.

　　Moreover, batteries can degrade, and when they do, the critical currents identified for the fresh battery no longer apply; the threshold becomes lower. At that point, given the same fast charge current, there's a higher likelihood that the battery will short.

　　"Battery operation is highly dynamic, in a very wide range of currents. Yet its disposition varies dramatically along the cycle life,” Assist Prof. Bai said. "That is why this becomes necessary."