Robert - thank you for you excellent job of dissecting the valve, and the great pictures that show us what you see. Once at home (and the kids in bed) I pulled out an (almost empty) MSR cannister...and at least looking at if from the top, your pictures seem to make great sense. Contrary to what might seem to "make sense"....it would not appear that the valves have any safety shutoff mechanism at all. Check me on this: my interpretation of what I see is that the inner blue plastic sleeve acts simply as a retainer/tensioner for the spring, which in turn loads the red valve plunger, which in turn loads the valve gasket against the top of the threaded shaft. Put on the stove burner, the red plunger gets depressed against the spring, the gasket becomes unloaded, and gas gets past the whole thing.
If it gets too hot, the gasket could deform and loose the seal (though the material is probably cross-linked, and would simply burn, not just deform), the red plunger could warp, and again you loose the seal, or the blue sleeve could deform and/or break, in which case the loading from the spring is lost, and you get catastrophic loss of the seal. In the end, once you take the burner head off, you leak out all your remaining fuel (either slow, or really, really fast!).
Does this look correct to you? Do you see anything that looks like it might act as a safety "shut-off?"
If what I said above is correct, then we would need to look at the thermal behavior of the red and blue plastic parts, and also the seal, and see which degrades at the lowest temperatures (for the techno-geeks out there - look the the glass-transition temperatures of the red and blue parts, and the decomposition temp. for the gasket....).
I think I will go let that almost empty MSR can leak out its remaining fuel and liberate those parts. Maybe next week I can get some time on the machine I need and do that thermal testing.....please bear with me though if it takes longer - the next several weeks are either busy, or have me out of town.....
Thanks again Robert - excellent work!
added in edit: oh, by the way, pretty much the highest temperature you can expect plastics to not become soft at is about 200C (though this is really high-end)...but that at least puts one boundary on the quesion.