Enhancement of Thermoelectric Efficiency in PbTe by Distortion of the Electronic Density of States
Joseph P. Heremans,1,2* Vladimir Jovovic,1 Eric S. Toberer,3 Ali Saramat,3 Ken Kurosaki,4 Anek Charoenphakdee,4 Shinsuke Yamanaka,4 G. Jeffrey Snyder, Science 321, 554 (2008)
Oh thermoelectrics, how I love thee. For those of you who dont know, thermoelectricity is the phenomenon of converting heat into electricity. For all our energy problems right now, this research could prove to be useful, if the zT (figure of merit for thermoelectrics) can break 2. I believe that a value of 3 or higher would increase the Carnot efficiency of thermoelectric (TE) modules for power generation, to make it a viable technology. I know the navy has a considerable amount invested as they’re planning to use it submarines. Automotive companies are also looking at it to make more efficient combustion engines, as the heat of the engine (around 600K) is the range where a good number of materials have a zT of around 1 or so.
I like the paper a lot. I respect the people who did it tremendously, but one of my problems is this. The parent compound is PbTe. Lead telluride. So many people do PbTe in search of better thermoelectrics. If you do a SciFinder scholar search, there are almost 100 or more PbTe based systems for thermoelectrics. As a synthetic chemist, I like new compounds. I personally believe, if we’re to break the zT barrier of 2 (in bulk materials) that we need to look at other compounds. For instance, beta-Zn4Sb3, also from the Snyder group is an amazing thermoelectric material as it’s an PGEC (phonon glass-electron crystal). Clathrates are also like that, and have the added bonus of ‘rattlers’ inside their cages to lower thermal conductivity.
PbTe is a cubic structure. Therefore, it’s Fermi surface is usually isotropic, which is, not that interesting. Tl-doping forces a distortion in the density of states by moving the Fermi level near a peak. Whenever Ef (the Fermi level) is in a high point in k-space, it’s bound to distort and usually a pseudo gap will appear (see Gruner or your condensed matter physics texts, as this is a known phenomenon on charge density waves and superconductors). This is what they basically did to PbTe. They made it’s Fermi surface interesting.
One interesting experiment that could come from this paper is a pair distribution function (PDF) analysis of the bulk. PbTe is cubic, like I said earlier, and electronic distortions still /should/ show up somehow. If you look at the PDF on this material, will the peaks show the same cubic ordering or will it change? Will the bond distances from Pb-Te be constant or will they have a greater range? Considering the size of Tl, I’m guessing there will be a greater range (duh), but it would be neat to model it experimentally.
More problems. How much better is Tl in PbTe for practical uses. Tl is notoriously toxic and the feasibility of actually using these in TE modules is IMHO, slim to nil. Let’s try to make something more environmentally friendly, or at least more environmentally neutral, so that when we save energy by being more efficient with thermoelectrics, we dont end up killing everyone around it.
However, I do like this paper in that they didnt resort to ‘low dimensionality’ or rather, nanostructuring to get it. As much as nano is a buzzword, I dont really like nano, so for that, this paper is awesome.