DMFT-QE Symposium: June 10

Talk 1:

Anharmonic lattice dynamics from vibrational DMFT

Timothy Berkelbach, Associate Professor, Department of Chemistry, Columbia University; Co-Director, Initiative for Computational Catalysis, Flatiron Institute

Phonons are the collective modes of atomic displacements in atoms and molecules, which emerge as noninteracting quasiparticles in the harmonic approximation to the electronic potential energy surface. Anharmonicities in the true potential cause phonon-phonon interactions, which—when strong—are hard to treat computationally, especially when nuclear quantum effects are important. In this talk, I’ll describe our recently-introduced vibrational dynamical mean-field theory, which maps the atomic dynamics of a periodic lattice onto those of a single cell coupled to a bath of harmonic oscillators. The dynamics of the impurity problem can be simulated quantum mechanically, leveraging the harmonic character of the bath, or classically, for which the impurity dynamics obey a generalized Langevin equation. I’ll present results on model systems, for which we can assess convergence with respect to iterations, cell size (as in cluster DMFT), and the use of mean-field theory + DMFT. I’ll also show results on a three-dimensional, atomistic material— namely a clathrate solid that hosts group-2 atoms, whose rattling dynamics are highly anharmonic and responsible for short phonon lifetimes and low thermal conductivities. In all cases, we show DMFT to be highly accurate and computationally cheap.

Talk 2:

Full cell DMFT for local and long-range electron correlation

Tianyu Zhu, Yale University

I will describe an ab initio formulation of dynamical mean-field theory (DMFT), called full cell DMFT, for computing charged excitations and spectra in correlated electron materials. I will first introduce the basic concept of full cell DMFT, where all orbitals in a unit cell are incorporated into the impurity problem with general bare Coulomb interactions, which is then solved by quantum chemistry impurity solvers. Next, I will discuss its application to systems with entangled local and non-local electron correlations, such as prototypical Kondo systems (magnetic impurities in bulk Cu) and LaNiO2, and demonstrate how the full cell formalism helps avoid many uncontrolled uncertainties in quantum embedding. Finally, if time permits, I will introduce a recent development in extending full cell DMFT for more systematic treatment of long-range electron correlation in periodic systems.

References:
[1] T. Zhu and G. K.-L. Chan, PRX 11, 021006 (2021)
[2] T. Zhu, L. Peng, H. Zhai, Z.-H. Cui, and G. K.-L. Chan, in preparation
[3] J. Li and T. Zhu, in preparation