Solving by Evolving: Genetic Algorithm Diffusion Monte Carlo Studies of Quantum Solvation and Ro-vibrational dynamics in Helium-4 droplets

Physical Chemistry Seminar Series

Helium-4 droplets -- containing from a few to thousands of helium-4 atoms -- are increasingly being used as spectroscopic matrices. They allow very accurate studies of molecules or molecular clusters which which may be too fragile to be studied in the gas phase. Simulations of these systems are crucial for understanding how the helium matrix interacts with the molecular dopant. Quantum Monte Carlo, and, in particular, diffusion Monte Carlo (DMC), methods are the most useful in studying these systems because of the large number of particles involved. However, computations of excited states using DMC are limited by the following chicken-and-egg situation: the nodal surfaces of the target wave function must be known in advance in order to solve the Schrodinger equation using DMC. A new approach is described in which a genetic algorithm is used to compute nodal surfaces on-the-fly within the DMC computation itself. Application is made to computing excited rotational states and tunneling splittings in several systems of experimental interest.

For more information contact Prof. Angelo Bongiorno (404-385-5169)