🚀 go-pugleaf

In article <1992Nov10.180223.25291@galileo.cc.rochester.edu> mh001b@uhura.cc.rochester.edu (Matthew W. Hacker) writes:

>Are there any substances that conduct electricity well, but
>conduct heat poorly?  This sort of material would be pretty
>well suited [to thermocouple design...]

	Metals are good heat conductors because the electrons pick up
heat at one end, diffuse quickly through the metal, and dump the heat
wherever they interact with the colder metal elsewhere.

	Superconductors conduct electricity well, but (because
the electrons are not interacting with the lattice) are very poor
heat conductors.  It is indeed unfortunate that the temperature
range over which superconductors operate is so limited...

	The next best candidate is semiconductors: one can
create the charge carriers at some doped region, and send 'em
across an intrinsic (non-conducting) zone, while no large
population of free electrons exists except for the minimum
number of charge carriers for the applied current.  For this
reason, practical thermoelectric coolers are usually made with
semiconductors (one other reason is the large Seebeck coefficients
that are available).

	Another quaint property of semiconductors is the non-metallic
nature of the atomic bonds; semiconductors are not ductile,
and cannot be made into flexible wire.  For thermocouples
used to measure temperature, one generally just uses a very long
length of the wire, to try to minimize the heat flux through the
measuring instrument, and the errors this introduces.

	John Whitmore