Here is my final blogpost for Condensed Matter (there might a few more after this).
In my talk, I mentioned that to have a good thermoelectric material, you need to have high electrical conductivity and low thermal conductivity. The simplest reason I could think of is that we want to maintain the temperature difference so we do not want the temperature at both ends to equalize quickly-hence the need for low thermal conductivity.
While I was thinking about thermal conductivity, I was wondering how do they measure the thermal conductivity. It turns out that measuring thermal conductivity is not an easy as measuring electrical conductivity. Several methods are listed on wiki and one of the most widely used one was laser flash analysis whose basic working principle is basically heating up one side of the material and measure the temperature change at the other side.
This makes you think about the uncertainties of the measurements that they made of the thermoelectric efficiency which depends on how accurately you can measure electrical conductivity, thermal conductivity and thermopower (hopefully everyone memorized the formula). It was mentioned in the paper that uncertainty in each quantity may vary between 5% to 20% which means the uncertainty of zT (the maximum efficiency) might be as high as 50%! Hence, the guys measuring those stuff need to make careful with measurements so that results of high zT would be actually meaningful.
Subscribe to:
Post Comments (Atom)
I agree with your reasoning for wanting a low thermal conductivity.
ReplyDeleteThe goal of a thermoelectric material would be to have little heat, big current, so the high electrical conductivity makes sense.
That's an interesting point you bring up regarding measuring thermal conductivity. Reminds me of Andy's value for Raman coefficient of water (but I think that was his Honour's talk).
Having a read of wiki and around, there seems to be generally two methods: steady-state and transient (what you mentioned would be transient).
The steady-state method mentioned was to have two plates of known thermal conductivity: one hot one cold.
Stick the hot plate onto one side of the sample and see how hot the other plate gets when everything is back to equilibrium. This method is less popular as it requires fancy setup.