For lattice Yang-Mills theories whose gauge group has trivial center, such as SO(3), physics predicts that Wilson's criterion for quark confinement fails: adjoint gluons screen external charges and the static quark-antiquark potential stays bounded. Nissim (arXiv:2605.16162, 2026) proved the first rigorous result of this type: for SO(3) lattice Yang-Mills on the integer lattice in dimension d at least 3, the potential is uniformly bounded for Lebesgue-almost every inverse coupling beta in a strong-coupling interval (0, beta*). That proof leaves five open issues, raised in the paper's remarks and final section and grouped here into four linked components that together constitute this problem: (1) remove the 'almost every beta' restriction by excluding complex zeros of the rescaled Wilson-loop function that approach the real interval as the loop grows, and establish a perimeter-law lower bound uniform over all rectangular loops rather than loop by loop; (2) extend deconfinement to Wilson loops in general representations satisfying the screening criterion and to the trivial-center groups SO(2N+1), N>1, where vanishing Haar moments force thicker tube clusters; (3) prove (or rigorously refute) deconfinement at weak coupling (large beta), required for any continuum-limit statement; and (4) rigorously reconcile finite-N deconfinement with Chatterjee's large-N area law. No resolution of any of these issues is known in the literature.

Background

Euclidean lattice Yang-Mills theory, introduced by Wilson [Wil74], is a probability model that discretizes quantum Yang-Mills theory. Its data are a compact Lie group GG (the gauge group), a finite subgraph Λ\Lambda of the integer lattice Zd\mathbb{Z}^d, and an inverse coupling strength β>0\beta>0. The measure places a Haar-random (uniformly distributed) group element on each oriented edge and reweights configurations by the Wilson action exp⁡(β∑ptr(Qp))\exp(\beta\sum_p \mathrm{tr}(Q_p)), where pp runs over plaquettes (unit squares), QpQ_p is the ordered product of edge variables around pp, and tr\mathrm{tr} is the normalized trace. The central observables are Wilson loops Wℓ=tr(Qℓ)W_\ell=\mathrm{tr}(Q_\ell), the normalized trace of the ordered product of edge variables around a closed loop ℓ\ell. Wilson proposed that the static potential between a quark and an antiquark separated by distance RR is V(R)=−lim⁡T→∞1Tlog⁡⟨WℓR,T⟩V(R)=-\lim_{T\to\infty}\frac{1}{T}\log\langle W_{\ell_{R,T}}\rangle, where ℓR,T\ell_{R,T} is an R×TR\times T rectangle, and his criterion for quark confinement is that V(R)→∞V(R)\to\infty as R→∞R\to\infty.

Two scaling regimes diagnose confinement. An area law, ∣⟨Wℓ⟩∣≤c1e−c2 area(ℓ)|\langle W_\ell\rangle|\le c_1e^{-c_2\,\mathrm{area}(\ell)}, implies that V(R)V(R) grows linearly (confinement); a perimeter-law lower bound, ⟨Wℓ⟩≥c1e−c2∣ℓ∣\langle W_\ell\rangle\ge c_1e^{-c_2|\ell|} with ∣ℓ∣|\ell| the loop perimeter, means the expectation decays no faster than exponentially in the perimeter, indicating screening and failure of the criterion. Osterwalder and Seiler [OS78, Sei82] proved area laws at strong coupling (small β\beta) for G=SU(N)G=SU(N), the group of unitary complex N×NN\times N matrices with determinant one, and recent work [CNS25a, CNS25b] extended the area-law regime to groups including SU(N)SU(N), U(N)U(N) (the group of all unitary complex N×NN\times N matrices), and SO(2(N−1))SO(2(N-1)), the even-dimensional rotation groups. An essential hypothesis in all of these proofs is that the center of GG — the subgroup of elements commuting with every element of GG — is nontrivial; roughly, center symmetry prevents the vacuum from screening an external charge in the fundamental representation. For groups with trivial center the physics prediction is the opposite: gluons transform in the adjoint representation (the action of the group on its own Lie algebra), and adjoint gluon pairs created from the vacuum can screen any external charge whose representation τ\tau satisfies the criterion that τ⊗An\tau\otimes A^n contains the trivial representation for some nn, where AA denotes the adjoint representation. Hence Wilson's criterion is expected to fail at all couplings in d≥3d\ge 3; this is the basis for the common statement that center symmetry is the mechanism of quark confinement [Cha21]. The simplest non-Abelian group with trivial center is SO(3)SO(3), the group of rotations of three-dimensional space.

Early rigorous claims of a perimeter law for trivial-center-like cases by Glimm and Jaffe [GJ79] and Seiler [Sei82] contain gaps, identified in the source paper: their cluster-expansion arguments (power-series expansions of expectations in the coupling β\beta whose terms are indexed by connected clusters of plaquette excitations) keep only the leading 'tube' clusters (minimal plaquette clusters wrapping the loop), but at fixed coupling and large loop size the enormous number of small excitations of the tubes can swamp the leading term. Nissim [Nis26] filled this gap for SO(3)SO(3): for every d≥3d\ge 3 there exists β∗(d)>0\beta^*(d)>0 (the convergence radius of the cluster expansion) such that for Lebesgue-almost every β∈(0,β∗(d))\beta\in(0,\beta^*(d)), sup⁡RV(R)<∞\sup_R V(R)<\infty. The proof has three ingredients: (i) a cluster expansion in β\beta in which the first nonzero contribution to ⟨Wℓ⟩\langle W_\ell\rangle comes from 2(d−2)2(d-2) tube clusters of exactly 4∣ℓ∣−164|\ell|-16 plaquettes whose joint contribution is at least (1/9)∣ℓ∣(1/9)^{|\ell|}; (ii) a complex-analysis argument (Jensen's formula, Blaschke factorization, Harnack's inequality, and Cartan's lemma — classical results of complex analysis that bound a holomorphic function in terms of its zeros) showing that the rescaled loop function φℓ(β)=β16−4∣ℓ∣⟨Wℓ⟩\varphi_\ell(\beta)=\beta^{16-4|\ell|}\langle W_\ell\rangle, holomorphic on the disk of radius β∗\beta^*, is bounded below outside a small exceptional set of β\beta — the exceptional set exists because zeros of φℓ\varphi_\ell in the complex plane may approach the real interval as ∣ℓ∣|\ell| grows, which is the origin of the 'almost every' restriction (zeros exactly on the real interval can already be excluded by reflection positivity, a positivity property of the measure under reflection in a coordinate hyperplane); and (iii) reflection-positivity-based concavity of the potential, which passes from single-loop bounds to uniform boundedness of VV.

One further tension frames the final part of the problem: in the N→∞N\to\infty limit, Chatterjee [Cha19a] rigorously proved that SO(N)SO(N) lattice Yang-Mills satisfies an area law, while at finite NN the trivial-center groups are expected — and at strong coupling for SO(3)SO(3) now proved — to deconfine. The source paper suggests a reconciliation through the truncated model of [CNS25b], in which the exponential of the Wilson action is replaced by a product of finite-degree Taylor polynomials: it asserts, without full proof, that this truncated model obeys an area law even for gauge group SO(2N+1)SO(2N+1) (the rotation group of (2N+1)(2N+1)-dimensional space, which has trivial center for every N≥1N\ge 1), while the analogous proof for the original Wilson action would break down at the 'revival' step where center symmetry is essential.

Problem Statement

The problem is to complete the rigorous deconfinement picture for lattice Yang-Mills theories with trivial center by resolving the issues explicitly left open in Nissim's strong-coupling theorem for SO(3) [Nis26]. The common objective is to prove — or rigorously refute — that trivial-center lattice Yang-Mills theories fail Wilson's confinement criterion, without the current restrictions on coupling, loop uniformity, gauge group, and representation. Concretely, for the Wilson action with normalized trace on the lattice Zd\mathbb{Z}^d, d≥3d\ge 3, with rectangular loops ℓ\ell of perimeter ∣ℓ∣|\ell| and quark-antiquark potential Vβ(R)=−lim⁡T→∞1Tlog⁡⟨WℓR,T⟩V_\beta(R)=-\lim_{T\to\infty}\frac{1}{T}\log\langle W_{\ell_{R,T}}\rangle:

(1) All β\beta and uniformity over loops, for SO(3): prove that there exist β∗(d)>0\beta^*(d)>0 and, for each β∈(0,β∗(d))\beta\in(0,\beta^*(d)) — every β\beta, not merely almost every — constants c1(β,d),c2(β,d)>0c_1(\beta,d),c_2(\beta,d)>0 such that ⟨Wℓ⟩≥c1e−c2∣ℓ∣\langle W_\ell\rangle\ge c_1e^{-c_2|\ell|} holds for all rectangular loops ℓ\ell simultaneously; in particular sup⁡RVβ(R)<∞\sup_R V_\beta(R)<\infty for every β∈(0,β∗(d))\beta\in(0,\beta^*(d)). The author-identified obstruction is to exclude zeros of the rescaled loop functions φℓ(β)=β16−4∣ℓ∣⟨Wℓ⟩\varphi_\ell(\beta)=\beta^{16-4|\ell|}\langle W_\ell\rangle in the complex plane that approach the real interval (0,β∗)(0,\beta^*) as ∣ℓ∣→∞|\ell|\to\infty; zeros on the interval itself are already excluded by reflection positivity.

(2) Other trivial-center groups and representations: extend the strong-coupling deconfinement result to the groups SO(2N+1)SO(2N+1) for N>1N>1 — where E[Q⊗k]=0\mathbb{E}[Q^{\otimes k}]=0 for Haar-distributed Q∈SO(2N+1)Q\in SO(2N+1) whenever kk is odd and k<2N+1k<2N+1, so in particular the third tensor moment vanishes and the third-order tube mechanism of SO(3)SO(3) does not contribute, and thicker tube clusters containing many copies of each plaquette are required — and to Wilson loops in general representations τ\tau satisfying the screening criterion that τ⊗An\tau\otimes A^n contains the trivial representation for some nn, with AA the adjoint representation of the gauge group.

(3) Weak coupling: prove (or refute) that Wilson's criterion also fails at weak coupling (large β\beta) for SO(3)SO(3), a regime in which cluster expansions are unavailable; the source describes this as currently out of reach, and it is required for any statement about the continuum limit of the theory.

(4) Large-N reconciliation: rigorously explain the coexistence of the area law proved for the N→∞N\to\infty limit of SO(N)SO(N) lattice Yang-Mills [Cha19a] with the deconfinement expected, and at strong coupling proved, at finite NN.

The problem is completely resolved when each of items (1)-(4) is settled by rigorous proof; a rigorous refutation of a predicted component (for example, a proof that some trivial-center theory in the stated class does satisfy Wilson's criterion) equally settles that component.

References

  1. [Nis26] Ron Nissim, Deconfinement For SO(3) Lattice Yang-Mills at Strong Coupling, arXiv:2605.16162 (2026), DOI 10.48550/arxiv.2605.16162, https://arxiv.org/abs/2605.16162
  2. [Wil74] Kenneth G. Wilson, Confinement of quarks, Physical Review D 10:2445 (1974), DOI 10.1103/physrevd.10.2445, https://doi.org/10.1103/physrevd.10.2445
  3. [OS78] K. Osterwalder and E. Seiler, Gauge field theories on a lattice, Annals of Physics 110(2):440-471 (1978), DOI 10.1016/0003-4916(78)90039-8, https://doi.org/10.1016/0003-4916(78)90039-8
  4. [Sei82] Erhard Seiler, Gauge Theories as a Problem of Constructive Quantum Field Theory and Statistical Mechanics, Lecture Notes in Physics 159, Springer (1982), DOI 10.1007/3-540-11559-5, https://doi.org/10.1007/3-540-11559-5
  5. [GJ79] James Glimm and Arthur Jaffe, Charges, vortices and confinement, Nuclear Physics B 149(1):49-60 (1979), DOI 10.1016/0550-3213(79)90156-1, https://doi.org/10.1016/0550-3213(79)90156-1
  6. [Cha19a] Sourav Chatterjee, Rigorous solution of strongly coupled SO(N) lattice gauge theory in the large N limit, Communications in Mathematical Physics 366(1):203-268 (2019), DOI 10.1007/s00220-019-03353-3, https://doi.org/10.1007/s00220-019-03353-3
  7. [Cha21] Sourav Chatterjee, A probabilistic mechanism for quark confinement, Communications in Mathematical Physics 385(2):1007-1039 (2021), DOI 10.1007/s00220-021-04086-y, https://doi.org/10.1007/s00220-021-04086-y
  8. [CNS25a] Sky Cao, Ron Nissim, and Scott Sheffield, Dynamical approach to area law for lattice Yang-Mills, arXiv:2509.04688 (2025), DOI 10.48550/arxiv.2509.04688, https://arxiv.org/abs/2509.04688
  9. [CNS25b] Sky Cao, Ron Nissim, and Scott Sheffield, Expanded regimes of area law for lattice Yang-Mills theories, arXiv:2505.16585 (2025), DOI 10.48550/arxiv.2505.16585, https://arxiv.org/abs/2505.16585