Friday, June 12, 2009

Organic superconductor information overload?

Here is a rough summary of a few things I think I learnt (or was reminded of) this week about superconducting organic charge transfer salts.
Please post corrections and clarifications. Where there is interest I will post more details...

A long-standing mystery in the Bechgaard salts has been the presence of rapid oscillations associated with an unexpected Fermi surface reconstruction. Could this be a many-body effect as in the cuprates?

(DMe-ET)2PF6 has superconductivity next to a charge ordered insulator.

Near the Mott transition critical point the low temperature NMR relaxation rate 1/T1 T should scale with |P-Pc|^delta, where delta =2 is the same critical exponent as for the conductivity.
This is because at half filling the number of localised spins is related to the number of doublons.

Kagawa et al., PRB 78, 184402 (2008) have a very elegant way of using NMR and the DM interaction to determine the staggered magnetisation (something is normally only observable via neutron scattering) above the Neel temperature.

The spin liquid material kappa-(ET)2(CN)3 has a specific heat which is linear in temperature at low temperatures. This is in distinct contrast to ET materials which have an antiferromagnetic
ground states. Combining the specific heat coefficient gamma with the low temperature susceptibility, would give a Sommerfeld-Wilson ratio of 1-2, characteristic of a ground state with gapless fermionic excitations, as proposed by P.A. Lee and collaborators.
I would think this would also mean that the dimensionless Korringa ratio should be unity.

The Bechgaard salts (based on the TMTSF molecule) have a spin-density wave insulating state with a relatively small magnetic moment. However, one should be cautious about claiming that Mott physics is not relevant to these materials since at lower chemical pressures there are Mott insulating, antiferromagnetic and spin Peierls states. (see the phase diagram below provided by Martin Dressel). Furthermore, even in the metallic state the Drude peak has extremely small
spectral weight and only exists below temperatures of about 10 K.

It is often claimed that in these materials the role of pressure is to increase the interchain hopping integral t_b and to reduce Fermi surface nesting. However, as far as I am aware there are no band structure calculations or direct experimental measurements to back this up. [Please correct me?]
Looking at how the critical field for the magnetic field induced spin-density wave (FISDW) varies with pressure may illuminate this, since the Lebed-Gorkov theory shows how the
critical field is related to tb', the hopping integral between
second-neighbour chains. Chaikin has a nice review on the FISDW's.

In the normal metallic state Stuart Brown and collaborators find that 1/(T_1 T) has a "Curie-Weiss" form 1/(T+theta) for a range of pressures (see his chapter in the book edited by Lebed). It is found that the parameter theta increases with pressure. Theta defines the temperature scale at which there is a crossover from Korringa (1/T1 linear in T) behaviour to T1 independent of T. Such a form is predicted naturally by the Moriya-Ueda antiferromagnetic
spin fluctuation theory near a quantum critical point in two dimensions. There, theta is the temperature scale over which the antiferromagnetic correlation length varies. In contrast, the resistivity does not appear to show any such crossover. It is dominated by a linear in T term, whose magnitude decreases with increasing pressure. Understanding this difference in the charge and spin response is an important question that needs to be resolved.

1 comment:

  1. Two questions I have are:

    How convincing are the experiments showing a linear specific heat in kappa-(ET)2(CN)3? (Particularly given the relative size of the phonon that has to be subtracted.)

    What should we make of the thermal conductivity measurements, which seem to disagree with the specific heat measurements?