It is widely believed that mean-field theory is exact for a wide range of classical long-range interacting systems. Is this also true once quantum fluctuations have been accounted for? As a test case we study the Hamiltonian mean-field (HMF) model for a system of bosons which is predicted (according to mean-field theory) to undergo a second-order quantum phase transition at zero temperature. The ordered phase is characterized by a spontaneously broken $\text{O}\left(2\right)$ symmetry, which, despite occurring in a one-dimensional model, is not ruled out by the Mermin-Wagner theorem due to the presence of long-range interactions. Nevertheless, a spontaneously broken symmetry implies gapless Goldstone modes whose large fluctuations can restore broken symmetries. In this work we study the influence of quantum fluctuations by projecting the Hamiltonian onto the continuous subspace of symmetry-breaking mean-field states. We find that the energetic cost of gradients in the center-of-mass wave function inhibits the breaking of the $\text{O}\left(2\right)$ symmetry, but that the energetic cost is very small, scaling as $O(1/N^2)$. Nevertheless, for any finite $N$, no matter how large, this implies that the ground state has a restored $\text{O}\left(2\right)$ symmetry. Implications for the finite-temperature phases, as well as the classical limit, of the HMF model are discussed.

• Received 11 September 2019
• Revised 3 December 2019

DOI:https://doi.org/10.1103/PhysRevE.101.012136