Ali Alavi (Max Planck Institute for Solid State Research)
Recent developments in Full CI Quantum Monte Carlo and applications to the cuprates
2016/11/9 (Wed) 15:00-16:00 S514 Seminar Room
The Full CI quantum Monte Carlo technique is a stochastic method for solving the electronic Schrödinger equation (for its ground-state) in extremely large Hilbert spaces. We will review some recent developments in this methodology, including its extension to the CASSCF technique and multi-reference perturbation theory, as well as to low-lying excited states. As an application, I will show results on the ab initio calculation on the Heisenberg exchange parameter of various cuprates. The calculations show that the normal super-exchange mechanism is inadequate to get quantitative agreement with experiment. Instead a multi-orbital mechanism involving the 4d orbitals on Cu and the 3p orbitals on O is essential is describing the spin-gap at a quantitative level. Our findings indicate that the popular lattice models of cuprates, such as the one-band Hubbard or even 3-band Hubbard model, may be gross simplifications which miss critical physics in the description of anti-ferromagnetism and potentially superconductivity in these systems.