Nope. 'Fraid not. Given the available volume in the device, the length of time the device was run, the temperatures reached, and the total power produced, there "is" no exothermic chemical reaction that will suffice
Many reactions will reach those kinds of temperatures, but they simply cannot last LONG ENOUGH. Think about a thermite demonstration...that lasts for a few seconds to maybe a few minutes....not hours on end.
Generally no, not a free running one in any case. The reactions I have seen use up their available fuel or catalyst pretty quickly in a flash in the pan effect. But that's not what we have here which is why I still consider it.
We have resistive heaters, we have a reaction that has stability in a narrow range of temperature so I can't discount a controlled exothermic reaction. And a pretty hot one at that. Either that or there is more material than we are aware of.....kinda hard to say since it's sealed, we don't know what actually was used and we don't know what was left.....all we know was the weight at the start and weight at the finish.
I can't even pretend to be comfortable with that. I actually wish I could be...