X-Message-Number: 2058
Date: Sat, 3 Apr 93 18:51:15 CST
From: Brian Wowk <>
Subject: CRYONICS Ethyl Chloride
Steve Harris:
> Look, I originally proposed the design with heat conductor at
> liquid nitrogen temp, separated from the -135 C space by
> insulating foam to get the proper thermal gradient, in order to
> simplify things, and allow liquid nitrogen to be used as primary
> coolant. If, however, we are complicate things and have a bunch
> of Peltier effect coolers* keep our liquid nitrogen liquid, then
> let me point out that we might as well use the coolers to
> primarily keep our ethyl chloride vat slurried at the higher temp
> of -135, and forget the liquid nitrogen completely!
Agreed. By the way, my intention was to use Peltier effect
coolers to cool the room at -130'C not the LN2 at -196'C. LN2 would
then be conserved by turning off current between Seebeck (inverse
Peltier) modules in the LN2 and the room. Let me explain.
A typical single-element semiconductor thermoelectric cooling
module works as follows:
----------------------------------
| cold metal |
----------------------------------
| | | | | ^
current | | P type | | N type | | current
| |semiconductor| |semiconductor| |
V | | | | |
apply ------------------ ----------------- apply
negative warm metal | | warm metal positive
potential ------------------ ----------------- potential
here here
The pattern of temperature-dependent contact potentials that the
electric current passes through results in heat being pumped from the
cold side to the warm side (i.e. refrigeration). If the applied
potentials are removed, the current will reverse, and you will have a
thermoelectric generator instead of a cooler. If you then short out the
two terminals on the warm side, heat will flow like crazy from the warm
side to the cold side (which is the Seebeck effect).
Now what I had imagined was something like this:
----------------------------------
| cold metal (in LN2) |
----------------------------------
| | | |
| P type | | N type |
--------------- ---------------
| |
| <- long wires -> |
| |
| |
--------------- ---------------
| P type | | N type |
| | | |
---------------------------------------
| warm metal (in -130'C room) |
---------------------------------------
This is a Seebeck effect device that is intended to passively transport
heat to the LN2 from any point in the room by using electricity instead
of temperature gradients (the LN2 is heavily insulated, and gains no
heat from ordinary thermal conduction). Many such devices placed at
various points around the room could maintain a uniform cool temperature
even if the air circulation system fails completely.
The Seebeck effect device described above would work just fine
if the thermocouples were metal, and the connecting wires were
ridiculously thick. (Yes, I know thermocouples can be connected in
series, but this just means many small wires instead of one big one.)
The wire thicknesses would be reasonable for semiconductor
thermocouples, IF THE ABOVE DEVICE WOULD WORK. I have not seen this
configuration in my research thus far, and don't know if this design
will work with semiconductor junctions. I will advise when I find out.
In any case it may be far simpler to just use slurried ethyl
chloride as thermal balast as Steve Harris suggests. Offhand, I can
think of the following advantages:
1) Little or no insulation would be required, saving space.
2) No need for replenishment.
3) No need for venting vapor.
4) Advantages 1-3 mean small amounts of ethyl chloride could be
put in multiple reservoirs, including all around the outer
walls in contact with cooling thermopiles.
I like it! Steve, could you please post the density, cost, and heat of
fusion of ethyl chloride?
> Of course, all this may be academic. Peltier-effect cooling
> modules are available to cool down to -50 C (perhaps an 80 C
> differential) but I've never seen one that will do to -135 C, let
> alone -196 C. Nor is it clear to me that you can simply stack
> them-- although this should work in theory, in practice there may
> well be absolute material limitations.
Single-stage bismuth telluride cooling modules can operate
between a 70'C temperature difference. And yes these modules are
routinely stacked ("cascaded") to operate between even greater
temperature differences. A cursory search of engineering abstracts
turned up a system that operated between +50'C and -60'C.
I have a line on some thermoelectric suppliers, and will be
calling them next week.
--- Brian Wowk
Rate This Message: http://www.cryonet.org/cgi-bin/rate.cgi?msg=2058