Path-integral computations of tunneling processes

The application of the path-integral methodology of Chandler and Wolynes [D. Chandler and P. G. Wolynes, J. Chem. Phys. 74, 4078 (1981)] to the calculation of one-electron-tunneling probabilities is revisited. We show that the evaluation of the kink free energy that is related to the tunneling splitting is associated with "polymer bead" distributions over a continuous distribution of scaled barriers, which makes both the calculation and its physical interpretation relatively difficult. In particular, we find that relative to other available techniques the method converges slowly and suffers from inaccuracies associated with the finite-temperature aspect of the calculation, and that past tentative identification of the bead distribution over the barrier with a physical picture of a "tunneling path" should be reassessed.