Interesting (to me, at least) how the How Stuff Works site is using the word Thermos generically but still capitalizing the ‘T’.
I’m skeptical of this. If 1/1000 of an atmosphere can transfer energy at a certain rate, and I add a second 1/1000 of an atmosphere, both portions should still be able to transfer the same amount of energy, so twice as much would be transfered. What’s the mechanism that would change that?
Looking around, Dr. Stangelove is correct, but I haven’t found a simple reason explaining why.
Here is a first-order mathematical explanation. If I understand correctly, the short answer is that the number of molecules goes down with pressure, but the mean free path increases (so the molecules can go farther without losing energy).
I was a little bit loose with my original statement. The conductivity only goes down once you become mean free path limited. For typical scales, though (say 10 cm to 1 m range), that happens between roughly 10^-3 and 10^-6 bar.
BTW, I was totally skeptical of this as well. I was investigating it in the context of understanding whether a heating wire would overheat in a vacuum compared to air. I have a vacuum chamber which I could only prove got to 0.02 bar (I could boil water at room temperature). I thought that obviously conduction was roughly proportional to pressure, and hence this was a good experiment. Not so much, as it turns out.
Very interesting discussion. I would like to add some comments:
Hydrogen is a great heat conductor. Large electric generators producing power in the 100MW range are cooled by Pure hydrogen. Goes back to the mean free path observation earlier.
For heat transfer in gases (or liquids) the Prandtl Number is a better indicator than thermal conductivity.
Vacuum technology is taught at many universities as part of the cryogenic engineering discipline. You need really low temperatures to generate good vacuum. Vacuum technology is one of the key technologies needed for semiconductor manufacture - read gas mono layers. Basically if you cut a fresh piece of silicon wafer, you have so much time before a layer of gas (air) molecules will be formed on it. The higher the vacuum, the more time you’ll have to etch your circuit.
In nature octopuses produce some of the highest vacuum.
Lastly the phenomenon of a thin layer of vapor stopping heat transfer is known as the Leidenfrost (sp?) effect. It is also known as film boiling and was the major cause of destruction of boilers including steam engines. It is also the reason why people can walk on coal fires unharmed : the sweat layer effectively stops heat transfer. You can also heat a pan to high temperature and put a few drops of water to see them dancing around like beads - it’s film boiling again.
A sphere uniformly coated with magnets pointing at right angles to the surface will have no net magnetic field inside or outside. The argument is exactly the same as it is for the electric fields of an electrically charged shell, which is a classic college-level physics problem. The only difference is that the magnets are “dipoles” in this case, which effectively means that you have two shells of equal and opposite total charge with very close radii.
Doubling the number of molecules that are passing momentum back and forth also cuts the average distance they fly in half. This cancels out the effect of pressure, until you get to a vacuum hard enough that the molecules mostly hit the walls rather than each other. This is the beginning of the Knudsen regime in which the concept of conductivity for a gas doesn’t work right.
If we let the OP include diamagnetic materials in his construction, it can be done.
Jearl Walker’s favorite effect!
CMC fnord!