Quantum ergodicity failure in narrow spectral windows

Open problem on quantum ergodicity over small spectral windows (Remark 1.2) · arXiv:1908.03694

arXiv Problem medium confidence— first stated 2019-08-10

Authors: Noga Alon, Shirshendu Ganguly, Nikhil Srivastava

Status open high confidence

The open problem of showing that the quantum-ergodicity average in a small spectral window of width roughly 1/log(m) fails to be equidistributed on near-Ramanujan graphs remains unresolved. Theo McKenzie (2022) studied the necessity of conditions for graph quantum ergodicity and cited 1908.03694, but addressed a different question about which graph-theoretic properties individually suffice for QE. No paper was found that controls all eigenvectors in the small spectral window to prove the QE average is not equidistributed there, nor that determines a lower bound on the smallest such window.

Cited literature (1)

partial The necessity of conditions for graph quantum ergodicity and Cartesian products with an infinite graph (2022)

Theo McKenzie · Comptes Rendus Mathématique

Shows that expansion alone or high girth alone is individually insufficient for graph quantum ergodicity, citing 1908.03694 as related work, but does not address whether the QE average fails to be equidistributed in small spectral windows on near-Ramanujan graphs.

Reviewer notes. No follow-up paper was found that directly addresses the open problem of proving QE fails in small spectral windows of width ~1/log(m) on near-Ramanujan graphs. Theo McKenzie (2022) is the closest related work found but focuses on necessity of conditions for QE broadly, not the small-window question. The paper arXiv:2102.04169 (Quantum ergodicity for expanding quantum graphs in the regime of spectral delocalization) concerns metric/quantum graphs rather than combinatorial graphs and addresses when QE holds, not when it fails. The conjecture is from 2019 and no resolution was found across the web search.

Auto-reviewed 2026-05-15 with claude-sonnet-4-6 (web search enabled).

Problem. The problem of controlling the other eigenvectors well enough to say that the quantum-ergodicity average in a small spectral window $I \subset (-2\sqrt{d}, 2\sqrt{d})$ of width roughly $1/\log(m)$ is not equidistributed is not pursued here and remains open. It would also be interesting to determine whether the constructions of Theorem 1.3 can prove a lower bound on the length of the smallest such window that is possible.

Context

The works [ALM15, BLML15] prove quantum ergodicity on sequences of expander graphs when the spectral average (3) is taken over a window $I$ of width roughly $1/\log(m)$. Theorem 1.3 produces many localized eigenvectors concentrated in a very small spectral window due to high eigenvalue multiplicity, but whether this implies that the QE average itself fails to be equidistributed in that window requires controlling all other eigenvectors in the window, which the paper does not achieve.

Notes. Stated in prose inside Remark 1.2, not a labelled theorem environment. PDF source — surrounding math may be partially garbled.

Source paper

High-girth near-Ramanujan graphs with localized eigenvectors
Noga Alon, Shirshendu Ganguly, Nikhil Srivastava · 2019-08-10
https://arxiv.org/abs/1908.03694 PDF source