Institut de Physique Théorique
Direction des Sciences de la Matière  - CEA-Saclay
Unité de Recherche Associée au CNRS
encart droite
Encart de droite, peut être utilisé à des fins diverses. Actuellement caché via les CSS.
Saturday, March 20, 2010

O. Parcollet's home page

 / 
Images
 / 
Pisp
 / 
 / 
 

The Dynamical Mean Field Method (DMFT) and its extensions

What is DMFT ?
Dynamical Mean Field Theory (DMFT) is a fruitful non perturbative approach to strongly correlated electrons systems. It is a local approximation in space which reduces a microscopic model on a lattice to a self-consistent quantum impurity model. It describes both the localised and the itinerant limit of electronic problems. It has already produced remarkable results and predictions on the Mott transition problem.

Why going beyond DMFT ?
However, in spite of these successes, various limitations of single site DMFT are now apparent. For example, the self-energy is independant of the momentum k by construction. Therefore, the method can not describe independant variations of quasi-particle residue and effective mass in the Fermi liquid regime. Moreover, a single site effective model preclude the treatment of more exotic orders with a order parameters which extends on a few sites, like dimerization, staggered flux, d-density wave or d-wave superconductivity which is observed experimentally in high-Tc superconductors.

Cluster methods
To overcome these limitations, we study cluster extensions of DMFT (CDMFT), using a self-consistent cluster of sites. These extensions will allow a more refined treatment of the Mott transition, in particular in its aspects depending on k, and of instabilities towards others phases (like superconductivity) and their relation with the Mott transition

Theoretical methods
The main technical difficulty is to find a efficient and reliable method to solve the self-consistent impurity problem. Even though quantum impurity models have been long studied, many useful methods are useless for DMFT problems, including integrability and Conformal Field Theory, because of the self-consistent aspect of the problem. New techniques are now being developped.

Main collaborators

References

  1. A. Georges, G. Kotliar, W. Krauth and M. J. Rozenberg
    Rev. Mod. Phys. 68 13 (1996).
Valid HTML 4.01!