Keyboard shortcuts

Press ← or → to navigate between chapters

Press ? to show this help

Press Esc to hide this help


slug: posske-2026-a392387-xx-chain-nullspace-2p bibkey: posske2026a392387 doi: null url: https://oeis.org/A392387 triage: theorem motivation_gids:

  • D5/S3/Quantum/SpinChains/XXChainNullspaceCount.result

The nullspace of the periodic XX chain on twice a prime number of sites

Problem

OEIS A392387 (Thore Posske, 2026) is the nullspace dimension of the periodic spin-1/2 XX Heisenberg chain on n > 1 sites. Its COMMENTS give the count

Also, the number of subsets K of {1,…,n} such that the sum of cosines of the angles in {(2j + (1 + (-1)^|K|)/2 )*Pi/n | j in K} is zero.

and its FORMULA section records

Conjecture: a(2p) = 2(6^((p-1)/2)+1) for odd prime p.

Issue #10063 fixes the readings: the formal object is this subset count, the combinatorial description common to the entry and to Hu, Gerken and Posske (arXiv:2602.15098, appendix on the XX model); its identification with the dimension of the zero-energy subspace is the source’s Jordan–Wigner reading and is not formalized; the empty set is counted; p ranges over odd primes and n = 2p.

Motivation

The appendix of arXiv:2602.15098 rephrases the nullspace degeneracy of the XX chain as the number of choices of distinct momenta with vanishing total energy and calls the counting of such vanishing sums of roots of unity “an unsolved mathematical problem”. The frozen declaration D5/S3/Quantum/SpinChains/XXChainNullspaceCount.result settles it for chains of 2p sites: the zero-energy subspace has dimension 2(6^((p−1)/2) + 1) for every odd prime p, two states from odd fermion number and 2·6^((p−1)/2) from even fermion number.

Gap

Issue #10063 preregisters the conjecture and its literature check. The entry (revision 52, 2026-02-20) lists this formula as a conjecture next to the proved characterization a(p) = 2 exactly for primes p; the source paper gives only a brute-force table for n ≤ 22; MathDB returns no entry for “A392387”, and the repository had no declaration or dossier for it. These readings are not-found-in-searched-scope; they do not establish an exhaustive worldwide literature search or priority.

Route

Write p = 2m + 1. A subset K ⊆ {1, …, 2p} is determined by its class-state function on the residues r modulo p: which of the two elements of {1, …, 2p} congruent to r, one even and one odd, lie in K. Let z(r) be the number of taken even elements minus the number of taken odd elements of the class r, η the primitive p-th root of unity with −η = e^{iπ/p}, and ω = e^{iπ/(2p)}. The cosine sum of K is the real part of Z = Σ_r z(r) η^r when |K| is odd and of ωZ when |K| is even.

The only rational linear relation among 1, η, …, η^{p−1} is that their sum vanishes, because the cyclotomic polynomial Φ_p is the minimal polynomial of η. Hence Re Z = 0 exactly when z(r) + z(−r) = 2z(0) for all r, and Re(ωZ) = 0 exactly when z(−r) − z(r − 1) = z(0) − z(−1) for all r; since r ↦ −1 − r fixes m, the latter is the symmetry z(r) = z(−1 − r).

For odd |K| the first relation forces z to be constantly 1 or constantly −1 (if z(0) = 0, then z is odd and |K| is even): the even elements and the odd elements of {1, …, 2p}, two subsets. For even |K|, count the symmetric class-state functions over the fundamental domain 0, …, m − 1 of r ↦ −1 − r: each of the m pairs of classes has 6 states with equal z, and the parity of |K| forces the fixed class m to be empty or full, 2 states. The total is 2 + 2·6^m.

Falsifier

A zero-sum subset outside the two classes above, or a missing one, would change the count; the equivalences with the relations on z exclude both. A rational relation among 1, η, …, η^{p−2} would break the characterization; the degree p − 1 of Φ_p excludes it. A proof that a(2p) differs from 2(6^((p−1)/2) + 1) for some odd prime would contradict the kernel-checked theorem.

Evidence

Enumeration of all subsets with floating-point cosine sums (zero test |Σ| < 10⁻⁹, no sums in [10⁻⁹, 10⁻⁶)) reproduces the entry’s DATA for n = 2, …, 20 and gives a(6) = 14, a(10) = 74, a(14) = 434 and a(22) = 15554, each split as 2 odd-size and 2·6^m even-size subsets.

The canonical source is D5/S3/Quantum/SpinChains/XXChainNullspaceCount.lean. Its public declarations are nullspaceCount, claim, and result. The frozen module state has statement identity sha256:8c612fc929792c7b8f53472c4bed5c5bdc4eae4a24cc913ae117958da5dd1565. The result declaration has statement identity sha256:4cfdd0cdb4dbd00412e214229b820c9258ec4566b5cbbcdfa7bd735919c308f0. The Freeze event is sha256:edc55918d549053a26fe09fcdf107c9d474ea4d8635eedd6f73c86c299c9ba90 and has no project-level frozen prerequisites. The proof uses only the standard axioms propext, Classical.choice and Quot.sound; no sorry, native_decide, or new axiom.

Triage

theorem; resolution proved for the quoted conjecture. The public theorem has proof_shape: content; its local steps characterize the zero-sum subsets by the relations on z and count them through an explicit bijection with pairs of class states over a fundamental domain. admission_basis: open-problem-resolution under preregistration issue #10063. There is no atom and no digestion coverage edge. The result is a uniform theorem for every odd prime, not a bounded enumeration, checker, numeric reduction or certified instance, so utility: none applies.

ASSUMED-UNVERIFIED

The identification of the subset count with the dimension of the zero-energy subspace of the XX chain is the source’s Jordan–Wigner reading; the Hamiltonian, the transformation and the fermion-parity-dependent momentum quantization are not formalized. The bounded literature check does not establish exhaustive worldwide novelty, priority, or the absence of an independent proof.