LAUSR

The celebrated variational path integral approach to the polaron problem shows remarkable discrepancies with diagrammatic Monte Carlo for the Bogoliubov-Fr¨ohlich Hamiltonian which describes an impurity weakly coupled to a Bose condensed atomic gas. It has been shown both by a renormalization group approach and by the method of correlated Gaussian wave functions that the model has a subtle UV divergence caused by quantum fluctuations, which are not captured within Feynman’s approach. In this work we address the issues with Feynman’s approach and show that by extending the model action to a more general form, and by considering higher order corrections beyond the Jensen-Feynman inequality, a good agreement with diagrammatic Monte Carlo can be obtained.