The Hubbard model describes strongly correlated electrons on a lattice. The cavity Hubbard model couples those electrons to a quantized electromagnetic field inside an optical cavity. The cavity photon mediates a new long-range interaction between electrons — mediated by light, not by Coulomb forces — and this interaction competes with the existing electron-electron repulsion.
The phase diagram of the coupled system contains states that neither pure electrons nor pure light would produce. A superradiant antiferromagnetic phase: the electrons develop antiferromagnetic order while simultaneously producing a macroscopic photon condensate. A chiral flux phase: the cavity photon induces circulating currents that break time-reversal symmetry, creating a state where electron motion has a preferred handedness.
Note (2026-11-10): I checked the abstract and searched the full text of arXiv 2603.13657. The phases and transitions listed match, but the essay places and explains them differently from the paper. In the paper, with no Hubbard repulsion the photon condensation is a first-order superradiant transition that induces a uniform magnetic field and an emergent chiral or pi-flux fermionic state; with repulsion a superradiant antiferromagnetic phase appears in a regime bounded by a first-order superradiant transition and a continuous antiferromagnetic transition of Gross-Neveu type (at about U = 6.4 t for a coupling of g = 6 in its finite-size scaling, using exponents taken from the established chiral Heisenberg Gross-Neveu class). The essay puts the continuous transition at a multicritical point where several phases meet, a term I did not find in the text. Its explanation, that the cavity's long-range coupling shifts the transitions into the Gross-Neveu class, is not the paper's: there the photon condensate produces the emergent flux state, and the paper cites earlier work on the Gross-Neveu class in the honeycomb-lattice Hubbard model. The method is a sign-problem-free fermion-photon quantum Monte Carlo on a two-dimensional model at half filling.
The phase transitions between these states exhibit two qualitatively different characters. The superradiant transition from the normal state to the superradiant antiferromagnet is first-order — the photon condensate appears discontinuously. But the transition at the multicritical point where several phases meet is continuous, with Gross-Neveu universality — a universality class associated with relativistic fermion field theories, not with lattice models.
The structural point: the cavity doesn't just modify the electron system — it changes which universality classes are accessible. Without the cavity, the Hubbard model's phase transitions belong to standard condensed-matter universality classes. With the cavity, the photon-mediated interaction introduces a long-range coupling that shifts the phase transitions into universality classes usually associated with high-energy physics. The light doesn't perturb the electrons; it changes the space of phases they can explore.