The confinement-deconfinement transition in quantum chromodynamics is a temporal event — raise the temperature past a critical point and quarks become free. In equilibrium, the transition happens everywhere simultaneously. The entire system is either confined or deconfined.
Braguta and colleagues (arXiv:2602.20970) simulate gluodynamics in Rindler spacetime — the coordinates natural to a uniformly accelerating observer. Acceleration creates a local temperature gradient via the Unruh effect: higher acceleration means higher local temperature. In this setting, the temporal phase transition becomes a spatial crossover. Different regions of space can be simultaneously confined and deconfined, separated by a boundary that follows the Tolman-Ehrenfest law.
The coexistence is stable. At appropriate combinations of bulk temperature and acceleration, a deconfined region sits near the Rindler horizon (high effective temperature) while a confined region occupies the less-accelerated space further away. The transition that is sharp in time at fixed acceleration is smooth in space at fixed time. The same physics — same critical temperature, same order parameter — manifests differently depending on which coordinate carries the variation.
Near a black hole horizon, the same effect might produce spatial stratification of QCD phases. The horizon's strong gravitational field maps to high acceleration, creating a deconfined shell around a confined exterior.
The general observation: a phase transition is a relationship between a control parameter and a response. When the control parameter is uniform (temperature everywhere the same), the transition is temporal — it happens when the parameter crosses the threshold. When the control parameter varies in space (local temperature from acceleration), the transition is spatial — it happens where the parameter crosses the threshold. The physics is identical; the geometry of the control parameter determines the geometry of the transition.
Correction (2026-11-02): “would” changed to “might”; the paper says the spatial transition “might take place” near a Schwarzschild horizon.
Further note (2026-11-02, after reading the paper body): the abstract says “might take place”; the discussion section says the results “suggest” it does and are “valid in the vicinity” of the horizon, so “might” is the cautious end of what the paper says. The simulation is of pure gluodynamics (no quarks), not full QCD as this essay's wording implies.
Amendment (2026-11-04): I checked the discussion section again. The paper says its results hold in the vicinity of the horizon at weak accelerations, and that very close to the horizon, where the local acceleration reaches the order of the inverse correlation length, “our results are not valid” and further study is needed. So my line that the horizon's strong field “maps to high acceleration, creating a deconfined shell” goes beyond the paper: the supported claim is a spatial confinement-deconfinement boundary at some distance from the horizon, with deconfinement inside it, not at the horizon itself.