Thermoelectrics Home Intro to Thermoelectricity TE Research TE Publications DiSalvo Homepage

Our Research

Here is a brief description of some of the work that we are doing in the thermoelectrics project. Our goal is to find and characterize new materials for thermoelectric applications. In particular we want to find a material that is better than the ones currently used for thermoelectric cooling at room temperature. It's not easy, but we believe that it can be done. This research is definately a high risk, high payoff venture.

 

The Challenge.

Maximizing ZT requires optimizing the Seebeck coefficient, resistivity and thermal conductivity simultaneously. This is no small feat since each of these properties depends on the electronic structure of the material. In fact, decreasing the resistivity often decreases the Seebeck coefficient. So tradeoffs must be made when trying to optimize the properties of a material, by doping or alloying for example. The best thermoelectric at room temperature is actually a heavily doped alloy of Sb2Te3 and Bi2Te3

Our research focuses mainly on searching for new compounds with enhanced thermoelectric properties, rather than modification of known materials. We try to "build in" the properties we want by using combinations of elements and starting materials that we hope will produce good thermoelectrics.

The work can be divided into two main parts, searching for new materials (synthesis) and measuring their thermoelectric properties (characterization).  Below we’ll say a little about the types of materials we have been synthesizing.  Learn about how we characterize the samples that we make by clicking here.

Synthesis of new materials.

Recently our synthetic efforts have been devoted to two main projects: Thallium containing compounds for low temperature applications, and Chevrel phase materials for high temperature power generation.

 

New Thallium Compounds

Thallium is very toxic, so it is unlikely that Tl based materials will be used in commercial applications.  However, we think Tl compounds are a good place to look for enhanced thermoelectrics, and finding one would provide proof of principle that high ZT can be achieved in bulk materials.

We have recently discovered lots of new Tl compounds (>20). Some are shown in the figure below. 

 

 

Thallium is a heavy metal, so its compounds should have low thermal conductivities.  Its solid state chemistry is similar to that of the alkali metals, but its less electropositive, so it should form compounds with smaller band gaps. Our work on Tl containing chalcogenides showed both of these expectations to be true.  Below is a graph of the measured thermal conductivities of several Tl compounds that we have discovered in our lab.  They are very low indeed! 

We have also measured the band gaps of some of our Tl compounds, and compared them to isostructural alkali metal analogues.  We found that in every case the Tl compound has a lower band gap. 

Unfortunately, the materials we have made so far do not have low enough electrical resistivities to be useful thermoelectrics.  However, the decreased band gaps and extremely low thermal conductivities that we have observed in Tl compounds leads us to believe that an advanced Tl containing thermoelectric material may be discovered.

Chevrel Phase Materials

The Chevrel phases are materials with structures composed of a three dimensional network of pseudo-cubic Mo6Q8 (Q = S, Se, Te) clusters.  The packing of the clusters leaves channels made up of interconnected cavities running throughout the structure.  This is a very large class of materials due to the versatility of the Chevrel phase structure.  Many different elements can be intercalated into the cavities, and the Mo and Q atoms can be partially substituted with other transition metals and halides, respectively.   These fillings and substitutions allow us to tune the thermoelectric properties.  Some Chevrel phases are good TE materials around 1000 oC.

We have synthesized many new Chevrel phase materials for TE testing at high temperatures with our collaborators at NASA/JPL.  In addition, we have studied the intercalation of Cu into Mo6Se8. We built a special apparatus for doing the intercalation at room temperature, using CuI dissolved in acetonitrile to move Cu ions from a bulk Cu source into a Chevrel phase powder.  This led to the discovery of a new Chevrel phase structure-type Cu4Mo6Se8, consisting of Mo6Se8 sheets that are not joined together through Mo-Se bonding, but only through Cu-Se bonding.  The well established electron counting rules for classical Chevrel phase materials predict this compound to be semiconducting.  However, DFT calculations predict metallic behavior.  We have used extended Hückel calculations to show that the separation of the Mo6Se8 sheets is responsible for the failure of the classical electron counting rules. The structure and some calculated molecular orbital diagrams and band structure are shown below.

               

Here are some other areas of interest to us ...

Collaborators