Testing the Twisted-hBN Artificial-Atom Proposal by Registry-Resolved DFT: A Symmetry-Forbidden Potential Form, a Fourfold-Weaker Well, and What Survives

Preprint. I measure the moiré potential of parallel-stacked twisted bilayer hBN directly (dipole-corrected DFT at 21 interlayer registries, two-dimensional registry grid, fixed-K band-edge extraction) and re-derive the claims of the artificial-atom proposal [arXiv:2604.11360] from the measurement: the proposed cos−sin potential form is forbidden by layer-exchange symmetry (measured sine fraction 1.8e-4), the wells are fourfold shallower than claimed (first-star weight 28.3 vs 113 meV), there is no magic angle, and the high-temperature coherence claim collapses with the measured well. The qualitative architecture (AA-localized states, A1→E giant-dipole transitions, identical cells) survives. Also published on alphaXiv.