Will metal salt solutions ruin Peltier (thermoelectric) modules?

[Joel Davidson]

I've got a real science question, and am hoping that it's visible amidst
the spam. :-) I'm interested in transferring heat into and out of
small amounts of aqueous solutions of various metal salts. For the
conditions I'm investigating, I think that Peltier (thermoelectric)
modules are the way to go. Obviously, this will work best when there's
as little thermal insulation as possible separating the solutions from
the modules. I could try using very thin glass plates in between the
solutions and the modules, with thermal paste. Metal is out of the
question, due to reaction with the solutes, and something like painted
metal still risks contamination via a scratch or ionic migration. I
could try coating the modules with something like a thin layer of
silicone, or some other coating, I suppose. But what would be most
efficient would be to simply put the modules directly in contact with
the solutions.

But I don't know enough about the material properties of thermoelectric
modules to be able to guess if contact with such solutions would make
them fail. Would metal ions migrate through the ceramic (?) coatings
and interfere with the semiconductor effects? Are there other potential
problems? Obviously, I'm going to be careful to prevent contact between
the solutions and the conductors that provide power to the modules.

Thanks!

[Jim Pennino]

Post by Joel Davidson
I've got a real science question, and am hoping that it's visible amidst
the spam. :-) I'm interested in transferring heat into and out of
small amounts of aqueous solutions of various metal salts. For the
conditions I'm investigating, I think that Peltier (thermoelectric)
modules are the way to go. Obviously, this will work best when there's
as little thermal insulation as possible separating the solutions from
the modules. I could try using very thin glass plates in between the
solutions and the modules, with thermal paste. Metal is out of the
question, due to reaction with the solutes, and something like painted
metal still risks contamination via a scratch or ionic migration. I
could try coating the modules with something like a thin layer of
silicone, or some other coating, I suppose. But what would be most
efficient would be to simply put the modules directly in contact with
the solutions.
But I don't know enough about the material properties of thermoelectric
modules to be able to guess if contact with such solutions would make
them fail. Would metal ions migrate through the ceramic (?) coatings
and interfere with the semiconductor effects? Are there other potential
problems? Obviously, I'm going to be careful to prevent contact between
the solutions and the conductors that provide power to the modules.
Thanks!

You will be hard pressed to find specifications for the ceramic plates
used for peltier devices and the most I have ever found is that they are
made with alumina ceramics. Note that alumina ceramics have different
properties than pure alumina.

In general, alumina ceramics are weakly attacked by strong acids and
bases and insoluable in water.

A ceramic that is 74%, which is about the cheapest stuff anyone makes,
or more alumina is concidered gas tight and will not absorb water.

I would doubt ions could migrate through alumina ceramic.

Some random thoughts on the issue.

Commonly available peltier devices are 40 X 40 mm and you will need a
heat sink on one side of it.

Unless you have a short length of glass tubing with a good, square cut
end you can somehow glue to the device, the simplest thing to do is to
themal glue a small beaker or flask to one side and the heat sink to the
other. Keep the glue layer as thin as possible by applying pressure
while the glue cures and do one side at a time.

You will want some thermal insulation around your container.

Peltier devices are current controlled, not voltage controlled, and are
best controlled with something like a cheap motor control module from
Amazon.

Switching from heat to cool is trivially done with a DPDT toggle switch.

Air cooling is a PITA so you should concider a 40 X 40 aluminum water
cooling block (about $5.00 on Amazon).

Amazon also has cheap pumps for around $15 you just put in a bucket of
water as well as pump systems with attached tanks starting at about $40.

With a thermistor thermal glued to the container and a few parts, you
can make the stuff to drive the motor controller such that the
temperature is kept to fractions of a degree C.

Such circuits can be found on the internet and the parts are, again,
available from Amazon.

[Joel Davidson]

Post by Jim Pennino
You will be hard pressed to find specifications for the ceramic plates
used for peltier devices and the most I have ever found is that they are
made with alumina ceramics. Note that alumina ceramics have different
properties than pure alumina.
In general, alumina ceramics are weakly attacked by strong acids and
bases and insoluable in water.
A ceramic that is 74%, which is about the cheapest stuff anyone makes,
or more alumina is concidered gas tight and will not absorb water.

[snip]

Many thanks! -- that's all extremely helpful. I knew about the need for
heat sinks, but you provided a lot of detail that I wasn't aware about.

I'll do a bit of reading about alumina ceramics just to be sure; more
knowledge is always good. I'm familiar with alumina from
chromatography. And from an attempt at smelting a few decades back,
when we built our little smelter from alumina insulating bricks from
a kiln. We discovered, to our chagrin, that using clay as a mortar
was NOT a good idea, as it dissolved nicely into the alumina in much
the same way that acetone dissolves styrofoam, when the setup got up
to sufficient temperatures.