Cause of normal-phase energy overestimation relative to experiment

Ascertain the cause of the systematic overestimation of the experimental energy per particle in the normal liquid phase by path-integral Monte Carlo simulations of liquid 4He that include ab initio three-body contributions evaluated perturbatively using the pair-product approximation.

Background

While the computed energies per particle agree excellently with experiment in the superfluid phase—especially when including the three-body potential—the authors observe a systematic positive bias in the normal liquid phase. They suspect the pair-product form of the two-particle propagator may introduce an uncontrolled bias, but comparisons with the primitive approximation are computationally demanding and inconclusive.

Identifying the source of this discrepancy is crucial for achieving liquid-phase accuracies commensurate with gas-phase metrology and for guiding improvements in PIMC propagator choices and inclusion of many-body interactions.