Questions about thermocouples/Seebeck Effect

[DrMrLordX]

I have some questions about thermoelectics, particularly thermocouples and how they work.

I understand the basics of how the Seebeck effect works, but the equations I see associated with the Seebeck effect only describe the voltages developed, but not the current. My confusion is probably due to a relatively weak background in physics/electrical engineering, but would anyone care to enlighten me as to what factors would determine the current generated within a thermocouple?

Also, I have been looking at commercially-available TECs to see what kind of materials are being used in such devices. I'm seeing a lot of devices with p-type and n-type bismuth telluride. I have been warned by at least one person that bismuth telluride is a toxic material, and the MSDS datasheet (http://www.espimetals.com/msds's/bismuthtelluride.pdf) doesn't really make it look like a particularly safe material either, especially when exposed to water.

How sane is it to use bismuth telluride in devices of this nature, especially when they are being used in commercial and home applications? I know the Seebeck coefficients of n-type and p-type bismuth telluride are pretty darn good, and a thermocouple made from both would easily beat a Type E Chromel/Constantan thermocouple, but the toxicity of bismuth telluride kinda creeps me out. I even found one company selling bismuth telluride-based TECs as temperature-control devices for aquariums.

Thanks in advance.

 

[PottedMeat]

Originally posted by: DrMrLordX
How sane is it to use bismuth telluride in devices of this nature, especially when they are being used in commercial and home applications? I know the Seebeck coefficients of n-type and p-type bismuth telluride are pretty darn good, and a thermocouple made from both would easily beat a Type E Chromel/Constantan thermocouple, but the toxicity of bismuth telluride kinda creeps me out. I even found one company selling bismuth telluride-based TECs as temperature-control devices for aquariums.

Thanks in advance.

I don't really see how there would be toxicicity problems with using bismuth telluride in thermoelectric cooler/heaters - there's no direct physical contact between the semiconductor and the material being cooled. As far as I can tell there's the P and N type semiconductor interconnected with metal, the metal is bonded to some ceramic slab that is in contact with the material. I don't really know how fragile they can be but one I cannibalized from a portable fridge seemed pretty sturdy and would take alot before it broke.

 

[DrMrLordX]

Hmm, a fair point. Thanks.

I still don't know how to calculate the current generated by a thermocouple subjected to a temperature differential, but it is good to know that bismuth telluride might be safe for everyday use.

 

[Paperdoc]

It's been quite a while since I dabbled in this, but I recall that simple thermocouples composed of two different metals generate basically a voltage, and that's all. With those devices it was important to measure the voltage with an extremely high impedance device (to the extrreme of a null-balancing bridge circuit) to avoid distorting the information. The voltage could be disturbed by the heat generated by current flow (changes the temperature at the junction), and by voltage losses from current flowing through the thermocouple leads themselves. So as power generators, those thermocouples were useless.

Devices made of the newer materials, I cannot comment on.

 

[DrMrLordX]

Thanks, that does help me, in a way.

 

[highwire]

A thermocouple's voltage and the loop resistance determines the current, just Ohm's law.
Of course, there are always two thermo-junctions in the loop, and these voltages oppose, so the net voltage difference between the hot and cold junctions is the effective voltage imposed upon the loop resistance producing a current.

Interesting app of a B-T thermocouple here:
http://ams.allenpress.com/perl...77%3AIIPD%3E2.0.CO%3B2

 

[DrMrLordX]

Ah, thank you for the explanation. That was simpler than I thought.