Wednesday, June 10, 2009

How bad can it get?

Hearing some talks about properties of the metallic state of the iron pnictide superconductors underscored to me that one sees certain common features in a wide range of strongly correlated electron materials. These properties are distinctly different from electronic properties of elemental metals. These unusual properties arise from the fact that a low energy scale emerges which defines a temperature scale T0 (often in the range 10-100 K) above which quasi-particles do not exist and we have a bad metal. Signatures of this crossover from a Fermi liquid at low temperature to a bad metal are:
  • the resistivity, Hall coefficient, and thermopower can have a non-monotonic temperature dependence
  • with increasing temperature the resistivity can smoothly increase to values much larger than the Mott-Ioffe-Regel limit (h2 a/e ~ 1 mohm-cm)
  • at temperatures of order T0 the thermopower can reach values as large as kB/e ~ 50 microV/K
  • above temperatures of order T0 the Drude peak in the frequency dependent conductivity collapses and the associated spectral weight shifts to higher frequencies
  • most of the spectral weight in the frequency dependent conductivity is at high frequencies
It is important to realise that if quasi-particles dont exist the one-electron spectral function does not have dispersive features, i.e., it is completely incoherent. This means that above T0 the notion of a band structure and a Fermi surface has no meaning.

All of the above features in the transport are captured by a dynamical mean-field theory (DMFT) treatment of a range of lattice Hamiltonians such as the Hubbard model and Anderson lattice model. These issues are discussed in more detail in a paper Jaime Merino and I wrote 10 years ago. The figure below illustrates how sensistive the thermopower is to the destruction of quasi-particles.
A recent paper by Haule and Kotliar considered the coherence-incoherence crossover in the pnictides using LDA-DMFT.
Relevant papers on the thermopower and Hall coefficient of the pnictides that I became aware of this week include those of David Mandrus' group and Paul Canfield's group.

No comments:

Post a Comment