Two posters have suggested that the solution would be to remove the battery chargers and fly the batteries as "primary" cells that would be recharged only on the ground. That won't solve the problem.
The problem is "thermal runaway". As the battery gets hot, the internals expand, getting closer to each other, and thus increasing the likelihood of a short-circuit. The more often they are cycled, the greater the likelihood of a short-circuit.
The batteries get hot because of the current passing through them. While this happens to a certain extent during charge (at 2C which is about 130 amps for these cells) it is much worse during discharge (when the APU battery flows more than 1,000 amps for up to 30 seconds while winding up the APU). If it's going to catch fire, the problem is most likely to begin just after you "use" it. Removing the charger doesn't solve that.
The onset of thermal runaway happens at about twice the boiling point of water: this research shows these batteries are well into "coffin corner" by 200 degrees Celcius:
Modeling Thermal Runaway for Safer Lithium Ion Batteries
As Lomapaseo points out, hot fires are successfully contained by metal cans every day: in things called "engines". I suggest that Lithium Ion batteries, with a known propensity to spontaneously combust, should be contained in a fire-proof box whenever they fly, whether they are "in use" or not. We know these things are going to burn occasionally: we should ensure the tail (or the cockpit!) doesn't fall off the aircraft when they do. Either would cause the flight crew to think unkindly of the designer/manufacturer.
My greater concern is not that these batteries catch fire, but that they fail! On a fly-by-wire aircraft, that gives a whole new meaning to the term "dead stick". Yes, I know almost everything else has to fail before this matters. But consider Sully's celebrated ditching:
He lost both engines. His APU was not running. If his batteries had failed, and his aircraft had been a 787, he would have had no flight controls. At all.
I think the pilots on this forum would be more than anxious to have this possibility designed out. Before they have to fly a 787 again...
It is easy to see why Boeing chose the Lithium Ion battery. Consider the following article (especially the table it contains):
Rechargeable battery - Wikipedia, the free encyclopedia
Here, I have sorted the data by "Watt/Hours per Kilo":
Type Energy density
(Wh/kg)
Lithium-air (organic)[7] 2000
Lithium sulfur[10] 400
Lithium-ion 200
Molten salt 180
Lithium-ion polymer 165
Sodium-sulfur 150
Silver-oxide 130
Lithium iron phosphate 100
Lithium–titanate 90
Alkaline 85
Zinc bromide 80
Nickel–hydrogen 75
Nickel–zinc 60
Nickel–metal hydride 55
Nickel–iron 50
Nickel–cadmium 50
Lead–acid 35
Vanadium redox 30
Sodium-ion[13] 0
Thin film lithium 0
Sorry, the PPRuNe website doesn't like tables, but you can get the idea. Leaving aside the two outliers at the top that are not on-sale yet, Lithium Ion comes in third. Molten salt is a little aviation-unfriendly.
But Lithium iron phosphate would be my pick. Twice the energy density of NiCad, and already approved for aviation service in an application I know about. OK, they are electrically "fragile", which means you have to pay very close attention to the design of the charger: one spike and they are trash. But that's probably what happened to the 787 batteries.
The benefit of the LiFePo battery is that it does not catch fire when it fails. Another benefit is that it has a very low internal resistance, which enables it to withstand truly prodigious discharge rates (APU starting...) without suffering. And they have around 5 times the in-service life.
Just my thoughts