In a recent paper (Phys. Rev. Lett. 102, 015301 (2009), arXiv:0810.3763) Rousseau and Denteneer claim that an unconventional "super-Mott" (SM) phase is realized by bosons trapped in an optical lattice close to a Feshbach resonance with a molecular state. The supposed SM phase, observed via quantum Monte Carlo (QMC) simulations of an atom-molecule Bose-Hubbard model, is an incompressible phase developing spontaneous atomic/molecular supercurrents which are perfectly anticorrelated. Here we show that the identification of this phase is based on a misinterpretation of the estimators of superfluidity in QMC, which break down in the presence of coherent atom/molecule conversion. Our conclusion is that the supposed SM phase is in fact a fully normal insulator.
The study of ultracold optically trapped atoms has opened new vistas in the physics of correlated quantum systems. Much attention has now turned to mixtures of bosonic and fermionic atoms. A central puzzle is the disagreement between the experimental observation of a reduced bosonic visibility
Vb, and quantum Monte Carlo (QMC) calculations which show
Vb increasing. In this paper, we present QMC simulations which evaluate the density profiles and
Vb of mixtures of bosons and fermions in one-dimensional optical lattices. We resolve the discrepancy between theory and experiment by identifying parameter regimes where
Vb is reduced, and where it is increased. We present a simple qualitative picture of the different response to the fermion admixture in terms of the superfluid and Mott-insulating domains before and after the fermions are included. Finally, we show that
Vb exhibits kinks which are tied to the domain evolution present in the pure case, and also additional structure arising from the formation of boson-fermion molecules, a prediction for future experiments.