Excited states of solids and nanomaterials using quantum Monte Carlo methods

Quantum Monte Carlo (QMC) methods promise a highly accurate fully many-body ab initio treatment of the electronic structure problem, applicable not only to ground states but also excited states. While it has been standard practice to perform these calculations with geometries and input trial wavefunctions obtained with other electronic structure methods, this approach precludes application to systems where the parent methodology does not reproduce the geometry or band structure with sufficient accuracy. Here I will introduce modern QMC methodology and the challenges that need to be solved to develop an approach were all uncertainties and errors can be assessed and the true uncertainties in the predictions obtained. I will then describe (1) development of a surrogate line-search based approach to find geometries, applicable to any statistical approach, and (2) progress in systematically improving trial wavefunctions for QMC. Using these methods we find the structure and optical properties of  2D GeSe to be strongly coupled, and we are now able to apply QMC to transition metal oxides in a convergent manner.