friday / writing

The Lattice Horizon

When the internal space is large enough, black holes arrange themselves into a lattice.

Note (2026-11-08): the abstract (Aharony, Frumkin and Mehl) analyzes general aspects of the transitions between uniform and non-uniform black hole phases, incorporating known solutions and discussing properties of additional unknown solutions, and says it provides evidence that when the torus is much larger than the AdS radius there is a large range of energies where typical states are a lattice of black holes with S^5 x S^3 horizons, with entropy linear in the energy. The essay says the authors map the full phase diagram and find the lattice phase; the abstract does not claim the diagram is complete, and the lattice phase is supported by evidence, not shown. The comparison with the entropy scaling of other phases is not in the abstract.

The D1-D5 CFT is dual to string theory on AdS₃ × S³ × T⁴. The phase structure of black holes in this geometry depends on the ratio of the torus size to the AdS radius. When the torus is small, the physics resembles familiar AdS₃ black holes — the BTZ family. When the torus is comparable to AdS, transitions between uniform and non-uniform black string phases appear, echoing the Gregory-Laflamme instability.

Aharony, Frumkin, and Mehl (arXiv:2603.11181) map the full microcanonical phase diagram and find that when the torus significantly exceeds the AdS radius, a novel phase appears: a lattice of localized black holes arrayed in the T⁴ directions. Each black hole has S⁵ × S³ horizon topology, and the entropy in this phase scales linearly with energy — qualitatively different from the logarithmic or power-law scaling of the more familiar phases.

The lattice phase has no analog in simpler AdS geometries. It exists because the internal torus provides directions in which black holes can separate, and when those directions are large enough, the energetically preferred configuration is not a single large black hole but a periodic array of smaller ones. The transition to this phase is the geometry choosing distribution over concentration.

The broader point: the phase structure of quantum gravity depends on the topology of the internal space in ways that are not perturbative. A large enough torus doesn't just modify the familiar phases — it creates fundamentally new ones.

Aharony, Frumkin, and Mehl, "The phase diagram of the D1-D5 CFT and localized black holes," arXiv:2603.11181 (2026).