Quantum unique ergodicity conjecture

A wave on a connected closed surface with negative curvature can vibrate in many patterns. At very high frequency, must every sequence of increasingly fine vibration patterns spread uniformly, including rare patterns a statement about most waves could miss? The full question tracks both where the wave lies and the directions of its oscillation. This surface example belongs to a wider problem on negatively curved spaces with no boundary.[1]

Fine-scale vibration patterns on a closed negatively curved space are uniform for most high-frequency modes. Can any exceptional sequence retain a different distribution of positions or directions? Let $M$ be connected, compact, boundaryless, of dimension at least two and strictly negative sectional curvature, with $-\Delta\psi_j=\lambda_j\psi_j$, $\|\psi_j\|_2=1$ and $\lambda_j\to\infty$. Semiclassical observables resolve position and momentum at unit energy. Their subsequential limits are probability measures on the unit cotangent bundle $S^*M$, invariant under geodesic flow. QUE requires every limit to equal normalized Liouville measure; projection then gives uniform spatial mass.[1] [2]

Uniform spreading is proved for most high-frequency modes in the relevant counting sense. The unresolved issue is a thin sequence of exceptions. Existing bounds limit how strongly such waves could concentrate, but have not forced every wave sequence to become uniform.[1] [2]

Ergodicity of classical flow yields the density-one theorem, not uniqueness of all invariant quantum limits. Entropy lower bounds exclude some extreme concentration patterns. Full-support theorems on negatively curved surfaces ensure that each limit gives positive mass to every open phase-space region, but a nonuniform measure can still have full support. Neither property establishes the required unique measure. The 2026 survey also distinguishes Hecke arithmetic proofs and warns that broader chaotic billiards can fail QUE.[1] [2]

Chaotic motion can still contain a path that repeats a closed loop; a wave's vibration pattern must also fit together across the whole surface. Those are different constraints, so a picture built from paths may leave out features of individual waves. A rare family of unusual vibrations could carry structure that disappears when many patterns are averaged together. This makes the problem a test of what classical motion tells us about a single wave, beyond the average behavior of many.[1]

The conjecture probes which invariant classical distributions can actually arise from exact eigenfunctions. Classical dynamics permits more invariant measures than the uniform one; wave interference and spectral constraints may eliminate the extra possibilities. A solution would explain how those constraints convert chaotic dynamics into universal high-frequency equidistribution.[1] [2]