BNL Home


Materials Design

Comscope sees its codes as an engine towards design and discovery of strongly correlated materials with desirable properties. A particular early focus of this effort has been on finding better thermoelectrics among strongly correlated semiconductors. We discuss this effort below.

Thermoelectrics: Electrons in solids carry both charge and entropy and hence conduct both electricity and heat. Thermal currents couple with electrical currents, while thermal gradients induce voltages. Engineers exploit these phenomena to design refrigerators, power generators, and temperature sensors. The thermoelectric performance of a material depends on a dimensionless parameter ZT, also known as its “figure of merit.” ZT = S2σT/κ, where S is the Seebeck coefficient or thermopower, σ is the electrical conductivity, T is the absolute temperature, and κ is the total thermal conductivity. Two factors contribute to the thermal conductivity: κ = κe + κl, where κe is the heat conductivity attributable to electrons and holes, and κl is the heat conductivity attributable to lattice vibrations (phonons). A large thermoelectric power factor, defined as S2σ, serves as a good indicator of the thermoelectric potential of a material when the lattice contribution κl dominates the thermal conductivity and where so-called phonon drag effects are not important.

FeSb2: In 2007, a German group made the surprising discovery that at low temperatures, the strongly correlated semiconductor FeSb2 with the marcasite crystal structure exhibits a giant thermoelectric power factor (S2σ) of 2300 μWK-2cm-1. Following the Dresden discovery, a major advance was carried out by the Brookhaven National Laboratories team, who succeeded in substantially improving the power factor of this material in a crystal exhibiting a metal-insulator transition.

From the standpoint of basic science, correlated materials such as FeSb2 pose many interesting questions. The mechanism for the giant thermoelectric power factor in these materials remains unknown. Generally speaking, the standard model of solid state theory fails to describe the behavior of correlated semiconductors and insulators accurately.

Comscope’s research: We are animated by the question of whether materials with even better thermoelectric properties than FeSb2 await discovery in the relatively unexplored domain of correlated semiconductors with the marcasite structure and related structures. To answer this, we are conducting an extensive study of compounds with the marcasite crystal structure and the closely related pyrite and rutile crystal structures. Martha Greenblatt (Rutgers) and Cedomir Petrovic (BNL) are actively synthesizing single crystals of these compounds. To date 

  • We have synthesized a number of compounds with the marcasite formula MX2: FeP2, FeSe2, FeTe2, CrSb2, NiSb2, CoSb2, RuSb2, MoTe2, RuSb2, OsSb2, and OsTe2.
  • We have examined ternary and quaternary structures as they offer more flexibility for tuning of physical properties. FeSbAs and FeSbP have been identified by theory as good candidates. We have managed to successfully synthesize FeSb1.97P0.03 , CoSbX (X = S, Se, Te), CoAsSb, and FeAsSe.
  • We have synthesized Fe1-xCoxSb2S4 in the family of TSb2C24 (where T=3d transition metal and C=chalcogen). Most of the family of TSb2C24 exhibits strong electronic correlations, large thermopower, and high ZT.
  • We have synthesized Fe2XS4 (X = Si, Ge) and Fe2GeSe4. These materials crystallize in the same space group as FeSb2 and are predicted to have good thermoelectric performance.

We plan to continue to use this relatively unexplored corner of strongly correlated semiconductor material space as a playground in which to search for better thermoelectrics and to correspondingly test the various capabilities of Comsuite’s codes.