Physical Chemistry Seminar Series: Tackling Strong Electron Correlation with Density Matrix Renormalization Group

Princeton University professor Sebastian Wouters will present this lecture.

The theory of quantum mechanics was invented nearly a century ago; in principle, it allows for accurate chemical predictions well within experimental bounds. In practice, the corresponding equations are far too complex to solve, since the size of the Hilbert space increases exponentially with the number of electrons. As a result, tremendous effort has been devoted to finding approximate, yet accurate, solutions of polynomial cost to describe low-lying potential energy surfaces.

Hartree-Fock is a well-known approximate method in electronic structure theory that is often employed as the zeroth-order reference before inducing explicit treatment of electron correlation. I will discuss one such treatment – the configuration interaction truncation scheme – in order to compare it with the density matrix renormalization group (DMRG). The latter is not biased to any particular Slater determinant, and permits one to solve the electronic structure numerically exact in large active orbital spaces, far beyond the capabilities of the traditional full configuration interaction (FCI) method.

DMRG expands the breadth of electronic structure theory, allowing for the accurate study of many biological and industrial catalysts. These systems often require a large active-space description, since they contain several transition-metal atoms or single atoms embedded in a pi-conjugated framework. I will discuss in detail how DMRG has allowed us to discover a new low-lying quintet state of the oxo-Mn(Salen) complex.

[1] S. Wouters, W. Poelmans, P.W. Ayers and D. Van Neck, Computer Physics Communications 185, 1501 (2014)

[2] S. Wouters and D. Van Neck, European Physical Journal D 68, 272 (2014)

[3] S. Wouters, T. Bogaerts, P. Van Der Voort, V. Van Speybroeck and D. Van Neck, Journal of Chemical Physics 140, 241103 (2014)