# Thermoelectrics Born Again

**Jos Heremans****Thermoelectrics Session, MIT Memorial**

Thank you, President Reif, Gene, Marianne, Paul, Eliot, and Carl.

I would like to explain a few things, and, in particular, I’d like to explain how Millie treated a problem that two navies in the world (French and US) both wanted addressed. This was a very important and difficult problem, and so she assigned it as a homework assignment to Lyndon Hicks. You may not believe that, but it’s true! This is what happened. This is how she did it.

The outcome of the assignment were these two papers — which I cite here — where she said, “Well, we should use nanomaterials for the following reasons.” I explain those reasons on the slide above. There was size quantization that happens to the electrons. There is a limitation to the lattice thermal conductivity due to the narrow structures.

I will now continue a little bit, as her collaborator, to explain how we worked on thermoelectrics outside of MIT. All that will follow, here, and I will tell you how the MIT program came about.

I was a visitor in Millie's group in 1980 and 1981. I worked with Mansour Shayegan and Boris Elman — who are both here. It’s important to remember the dates, because in 1982, this paper was published. I should add, here, that it was extraordinarily impressive. We took data at the Magnet Lab one day, and I showed it in the evening, and the next morning there was a theory by Gene and Millie that was quantitative and fitted the data exactly right, with no adjustable parameters. I just couldn't imagine! These are the kinds of things that don't happen anywhere else.

Anyway, we worked on and on as collaborators for 20 years. In this case, you'll see a 2002 paper where we worked on these nanowires, as a result of Hicks. And, of course, we have problems. So you will see there that, yes, the thermal power was enhanced, but the wires had the tendency to localize. That was the end, more or less, of the bismuth approach.

Of course, Millie was never scared of anything. She came up with the next solution, which was use bismuth-antimony alloys. I won’t get into detail on that, because it’s more important to understand what happened to the entire field. Millie had these ideas — she had fantastic ideas! And so did Gene! But she didn’t just *have* ideas, she also *seeded *ideas. That is what I would like to enhance, here.

The ideas that came from others was that perhaps we shouldn’t just nano-structure in bulk — perhaps we should change from bismuth nanowires to nanostructures in bulk. We did have other things going on, as well, in harmonicity, electron energy filtering, band engineering, and so on and so on.

Hundreds of people got involved, and dozens of new ideas came out of these papers, many of which were not Millie's, directly, but were seeded by Millie's ideas. And you can see the outcome, which is the improvement in the ZT (the thermoelectric figure of merit) over the years. You can see a clear break in the curve. The break in the curve happens with the publication of Lyndon Hicks’ homework assignment.

I also wanted to show, briefly, the new scientific thinking in the field that came out of all this. Gang and I both wanted to speak on this, so we divided that subject into electrons and phonons. I will speak about phonons, and Gang (at the end of his talk) will talk about phonons.

So, very quickly, the nanostructuring has its problems, in the sense that — for example — you need about a kilogram of material to generate a kilowatt of power. And a kilogram of nanowires is a problem, because it’s such a huge amount. So there are many other ideas that came around, to try to solve this problem. Spin-based thermoelectrics is something that *I *tried, and I'll talk more about that at MRS, tomorrow.

At the moment, people are thinking more in terms of new topological effects. Beyond the band structure, where you look at the potential, the electrostatic potential of atoms, the topological effect. There is a lot of new thought about this, and there are people in this audience who’ve started writing papers about this. It’s very interesting material.

Of course, once again, we have a paper, here, that comes from Millie. This was a paper that was with Shuang Tang, who, I think, is the last graduate student of Millie's in this topic — I'm not quite sure — where they point out the topological properties of bismuth-antimony, and that is, indeed, one of the options.

And so, with this, I will let Gang Chen tell you about phonons.

Thank you.