[sci.electronics] Thermoelectric coolers - how do they work - in English plea

dbell@cup.portal.com (David J Bell) (09/20/90)

Simply, the Peltier effect gives a correlation between thermoelectric
generation and electrical *pumping* of heat. Consider it analogous to
the way a DC motor can be used as a generator: with a Peltier device
(usually a P-N semiconductor junction, but I believe any two dissimilar
conductors), if you heat one junction (remember, there are always
effectively *two* junctions, but one may be made with copper wires
between the active materials), it will produce a current, just like
any thermocouple. If you drive a current into the device, though,
one junction will get hot, and the other will get cold. Reverse the
polarity of the driving current, and the hot and cold ends switch, too.

So: to cool a diode laser, bond the laser die to one junction of a
Peltier device, and bond the other junction to a heat sink or radiator.
Add current and you have instant cool. If you need more temperature
difference between "ends", stack the element with the laser attached
onto a few others (in a layer), and stack themin turn on a number
of *others* in a larger layer, on the heat sink; basically, build
a pyramid, with the coldest end at the tip, the hottest at the base.
You need the extra devices in each successive layer because they are
definately *not* 100% efficient - they produce waste heat as well as
pumping heat, just like a refrigerator...

Dave            dbell@cup.portal.com

larry@kitty.UUCP (Larry Lippman) (09/21/90)

In article <34081@cup.portal.com>, dbell@cup.portal.com (David J Bell) writes:
> Simply, the Peltier effect gives a correlation between thermoelectric
> generation and electrical *pumping* of heat. Consider it analogous to
> the way a DC motor can be used as a generator: with a Peltier device
> (usually a P-N semiconductor junction, but I believe any two dissimilar
> conductors), if you heat one junction (remember, there are always
> effectively *two* junctions, but one may be made with copper wires
> between the active materials), it will produce a current, just like
> any thermocouple. If you drive a current into the device, though,
> one junction will get hot, and the other will get cold. Reverse the
> polarity of the driving current, and the hot and cold ends switch, too.

	It is interesting to note that the Peltier effect was discovered
in 1834, and was a logical derivation of the Seebeck effect discovered
in 1821.  In 1855 William Thompson (Lord Kelvin) derived a scientific,
thermodynamic relationship explaining both effects.

	It was not until around 1910 that the German scientist Altenkirch
gave serious consideration to thermoelectric refrigeration.  However, the
low Seebeck coefficients of materials available at the time precluded the
development of any practicable devices. 

	Thermoelectric refrigeration remained a curiosity item until the
1950's, when advances in development of semiconductors created materials
having a much larger Seebeck coefficients than previously available.
The primary material used in practicable thermoelectric refrigeration
devices is bismuth telluride, often used in conjunction with bismuth
selenide or antimony telluride. 

	I have often wondered what would have happened if a greater
understanding (or "luck") in metallurgy and chemistry had existed during
the 1800's, allowing the discovery of suitable semiconductor materials
for thermoelectric refrigeration at that time.

Larry Lippman @ Recognition Research Corp.  "Have you hugged your cat today?"
VOICE: 716/688-1231   {boulder, rutgers, watmath}!ub!kitty!larry
FAX:   716/741-9635                  {utzoo, uunet}!/      \aerion!larry

mmm@cup.portal.com (Mark Robert Thorson) (09/22/90)

> 	Thermoelectric refrigeration remained a curiosity item until the
> 1950's, when advances in development of semiconductors created materials
> having a much larger Seebeck coefficients than previously available.

I recall that GE attempted to market a thermoelectric refrigerator in
the 1950's.  Anyone recall the details?  If it's anywhere near practical
to use, and if I can find one, that would be a nice historical artifact
to obtain.