Several recent experiments have observed enhanced atom losses at a scattering length where an Efimov trimer is near the atom-dimer threshold. These loss features have been attributed to an avalanche mechanism in which the loss from 3-atom recombination is amplified by secondary elastic collisions of the dimer from the recombination event. We use Monte Carlo methods to calculate the average atom loss and heating per recombination event and then calculate the atom loss rate coefficient. We find that, though the avalanche mechanism predicts significant loss enhancements, it cannot explain the narrow loss features observed in recent experiments. We have proposed a new mechanism for these loss features in BEC experiments that involves a small coexisting condensate of dimers.
Author entry (protected)