TY  - JOUR
AU  - HOU, Jixuan
PY  - 2026
TI  - BEYOND MEAN-FIELD: INTRODUCING BETHE-PEIERLS APPROXIMATION INTO UNDERGRADUATE STATISTICAL PHYSICS TEACHING
JO  - Physics and Engineering
SN  - 1009-7104
SP  - 52
EP  - 56
VL  - 46
IS  - 4
AB  - Conventional Weiss mean-field theory erroneously predicts a finite-temperature phase transition for the one-dimensional Ising model, giving kBTc/J=2, in contradiction with the exact result Tc=0. This inconsistency has long posed a conceptual difficulty in undergraduate statistical physics teaching. To address this problem, this paper proposes the systematic introduction of the Bethe-Peierls approximation as an effective way to resolve the “one-dimensional paradox.” In this approach, a cluster consisting of a central spin and its nearest neighbors is constructed. The interactions within the cluster are treated exactly, while the influence of the external environment is approximated by an effective mean field, thereby naturally incorporating short-range correlations. Based on the cluster partition function, the self-consistent equation is derived, and the critical temperatures of the one-, two-, and three-dimensional Ising models are obtained. The results show that the Bethe-Peierls approximation correctly gives Tc=0 in one dimension. For two and three dimensions, it predicts kBTc/J≈2.885 and 4.933, respectively, which are significantly closer to the exact or high-precision numerical values, 2.269 and 4.5115, than the conventional mean-field predictions, 4 and 6. Through systematic comparison, this work demonstrates that the Bethe-Peierls approximation has clear advantages in physical interpretation, computational feasibility, and predictive accuracy. It therefore provides a rigorous and instructive tool for teaching phase transition theory in undergraduate statistical physics.
UR  - https://doi.org/10.26599/PHYS.2026.9320409
DO  - 10.26599/PHYS.2026.9320409