Yang-Mills Theory on Lattice and Continuum: Spectral Gap and Existence of Limit

Yang-Mills Theory on Lattice and Continuum: Spectral Gap and Existence of Limit Christian Franchi Viceré ORCID: 0009-0001-8974-4991 WhatsApp: +44 7756 302178 December 2025 Abstract I present a unied treatment of two fundamental results for the Yang-Mills Millennium Problem. First, I prove that SU(N) Yang-Mills theory on any nite lattice Λa,L has strictly positive mass gap ∆Λ>0 . The proof uses the Perron-Frobenius theorem for transfer matrices, full support of the Haar measure, and non-constancy of the plaquette trace. No reection positivity, FKG inequalities, or smeared operators are required. Second, I establish the existence of subsequential continuum limits satisfying Osterwalder-Schrader axioms OS0OS3, which enables Hilbert space reconstruction via the OS theorem. Tightness follows from the Mitoma criterion; SO(4) invariance from the Symanzik expansion combined with regularity arguments on compact domains. The complete Millennium Problem reduces to proving that the spectral gap persists in the continuum limit. Keywords: Yang-Mills theory, lattice gauge theory, mass gap, continuum limit, OsterwalderSchrader axioms, Perron-Frobenius theorem, Mitoma criterion. MSC 2020: 81T13 (primary); 81T25, 46N50, 60B10 (secondary). Contents 1 Introduction 2 1.1 TheProblem ...................................... 2 1.2 MainResults ...................................... 2 1.3 StructureofthePaper................................. 2 2 Lattice Gauge Theory 3 2.1 CongurationSpace .................................. 3 2.2 Yang-MillsMeasure................................... 3 3 Spectral Gap on Finite Lattice 3 3.1 Strategy......................................... 3 3.2 Spectral Properties of the Transfer Matrix . . . . . . . . . . . . . . . . . . . . . . 4 3.3 Variational Characterization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4 3.4 The Single-Plaquette Operator . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4 3.5 ProofofTheorem1.1.................................. 5 3.6 QuantitativeBound .................................. 5 4 Existence of Continuum Limit 5 4.1 MomentBounds .................................... 5 4.2 SmearedOperators................................... 6 4.3 Tightness ........................................ 6 1 4.4 Existence ........................................ 6 4.5 Symmetries ....................................... 6 5 Osterwalder-Schrader Axioms 7 5.1 Verication ....................................... 7 5.2 Hilbert Space Reconstruction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7 6 What Remains Open 7 6.1 SummaryofResults .................................. 7 6.2 TheRemainingProblem................................ 7 1 Introduction 1.1 The Problem The Yang-Mills existence and mass gap problem, formulated by Jae and Witten [1] for the Clay Mathematics Institute, requires: Prove that for any compact simple gauge group G , a non-trivial quantum Yang-Mills theory exists on R4 and has a mass gap ∆>0 . The problem decomposes naturally into three independent questions: (i) Does a continuum quantum eld theory exist as a limit of lattice regularizations? (ii) Does the lattice theory have a spectral gap for every nite lattice? (iii) Does this gap persist in the continuum limit? This paper addresses (i) and (ii) completely. Question (iii) remains open and constitutes the core of the Millennium Problem. 1.2 Main Results Theorem 1.1 (Spectral Gap on Finite Lattice) . For SU(N) Yang-Mills theory on any nite lattice Λa,L with spacing a > 0 , volume L < ∞ , and inverse coupling 0< β < ∞ : ∆Λ:= inf{σ(HΛ)\{E0}}−E0>0 Theorem 1.2 (Existence of Continuum Limit) . There exist subsequential limits of lattice YangMills measures that: (i) Are probability measures on S′(R4) (ii) Satisfy Osterwalder-Schrader axioms OS0OS3 (iii) Are SO(4) -invariant and translation-invariant (iv) Support Hilbert space reconstruction with Hamiltonian H≥0 The proofs are self-contained and use only standard tools from functional analysis, spectral theory, and measure theory. 1.3 Structure of the Paper Section 2 establishes the rigorous framework for lattice gauge theory. Section 3 proves Theorem 1.1. Section 4 proves Theorem 1.2. Section 5 veries the Osterwalder-Schrader axioms and discusses Hilbert space reconstruction. Section 6 summarizes what remains open. 2 2 Lattice Gauge Theory 2.1 Conguration Space Denition 2.1 (Finite Lattice) . For a > 0 (lattice spacing) and L > 0 (half-size): Λa,L := (aZ)4∩[−L, L]4 with periodic boundary conditions. Denition 2.2 (Link Set) . L:= {(x, µ) : x∈Λ, µ ∈ {1,2,3,4}} Denition 2.3 (Conguration Space) . CΛ:= SU(N)|L| equipped with the product topology. Lemma 2.4 (Compactness) . CΛ is compact. Proof. SU(N) is compact. By Tychono's theorem, a nite product of compact spaces is compact. 2.2 Yang-Mills Measure Denition 2.5 (Wilson Action) . SW[U] := −β NX P∈P Re Tr(UP) where β= 2N/g2 is the inverse coupling and UP denotes the ordered product of link variables around plaquette P . Denition 2.6 (Yang-Mills Measure) . dµβ[U] := 1 Zβ exp β NX P Re Tr(UP)!Y ℓ∈L dµHaar(Uℓ) where Zβ is the partition function. Denition 2.7 (Kogut-Susskind Hamiltonian) . HΛ:= g2 2aX ℓ Ea ℓEa ℓ−1 g2aX P Re Tr(UP) where Ea ℓ are the electric eld operators satisfying the su(N) commutation relations. Denition 2.8 (Transfer Matrix) . TΛ:= exp(−aHΛ) 3 Spectral Gap on Finite Lattice 3.1 Strategy To prove ∆Λ>0 , I exhibit a single operator O such that the variational quantity Q[O] := ⟨0Λ|O†[HΛ, O]|0Λ⟩ is strictly positive. The operator I use is the real part of the trace of a single plaquette. 3 3.2 Spectral Properties of the Transfer Matrix Lemma 3.1 (Compactness) . TΛ:HΛ→ HΛ is a compact operator. Proof. TΛ is an integral operator with kernel K(U, U′) := Zexp(−a·Stemporal[U, U′, V ]) YdµHaar(V) Since CΛ is compact and K is continuous on CΛ× CΛ , the operator is Hilbert-Schmidt, hence compact. Lemma 3.2 (Strict Positivity of Kernel) . For all U, U′∈ CΛ : K(U, U′)>0 . Proof. The integrand exp(−a·S[···]) >0 for all congurations. The Haar measure has full support on SU(N) . The integral of a strictly positive function with respect to a measure of full support is strictly positive. Theorem 3.3 (Simplicity of Ground State) . The maximal eigenvalue λ0 of TΛ is simple. Equivalently, E0=−a−1log λ0 is a simple eigenvalue of HΛ . Proof. This follows from the Jentzsch theorem [2]: a compact integral operator with strictly positive kernel has simple maximal eigenvalue. 3.3 Variational Characterization Lemma 3.4 (Variational Identity) . For any operator O with O|0Λ⟩ ∈ Dom(HΛ) : Q[O] = ⟨Ψ|(HΛ−E0)|Ψ⟩ where |Ψ⟩:= O|0Λ⟩ . Proof. Direct computation using HΛ|0Λ⟩=E0|0Λ⟩ . Lemma 3.5 (Characterization of Vanishing) . The following are equivalent: (i) Q[O] = 0 (ii) O|0Λ⟩ ∈ span{|0Λ⟩} (iii) Var0Λ(O)=0 Proof. The equivalence (i) ⇔ (ii) uses the simplicity of E0 (Theorem 3.3). The equivalence (ii) ⇔ (iii) is algebraic: O|0⟩=c|0⟩ i ⟨O2⟩=⟨O⟩2 . 3.4 The Single-Plaquette Operator Denition 3.6 (Plaquette Operator) . For a xed plaquette Px0 : Ox0:= Re Tr(UPx0) : CΛ→R Lemma 3.7 (Full Support) . For 0< β < ∞ , the measure dµβ has full support on CΛ . Proof. The density exp( β NPPRe Tr(UP)) is strictly positive everywhere. The product Haar measure has full support. Proposition 3.8 (Image of Plaquette Operator) . The function Re Tr : SU(N)→R has image Ncos 2π N, N In particular, it is non-constant. 4 Proof. For U∈SU(N) with eigenvalues {eiθj} satisfying Pjθj≡0 (mod 2π) : Re Tr(U) = N X j=1 cos θj The maximum N is attained at U=1 . The minimum, found by Lagrange multipliers, is Ncos(2π/N) at θj= 2π/N for all j . Continuity and connectedness of SU(N) imply the image is the full interval. Theorem 3.9 (Positive Variance) . For the plaquette operator Ox0 on any nite lattice with 0< β < ∞ : Var0Λ(Ox0)>0 Proof. Suppose Var(Ox0) = 0 . Then Ox0 is constant µβ -almost everywhere. Since µβ has full support (Lemma 3.7) and Ox0 is continuous, it must be constant everywhere. This contradicts Proposition 3.8. 3.5 Proof of Theorem 1.1 Proof of Theorem 1.1. By Theorem 3.9, Var(Ox0)>0 . By contrapositive of Lemma 3.5, Q[Ox0]> 0 .Decompose |Ψ⟩:= Ox0|0⟩=c|0⟩+|Ψ⊥⟩ with |Ψ⊥⟩ ⊥ |0⟩ . Then ∥Ψ⊥∥2= Var(Ox0) and Q[Ox0] = ⟨Ψ⊥|(H−E0)|Ψ⊥⟩ ≥ ∆Λ·Var(Ox0) Therefore: ∆Λ≥Q[Ox0] Var(Ox0)>0 3.6 Quantitative Bound Theorem 3.10 (Strong Coupling Bound) . For β≪1 : ∆Λ≥N2 aβ +O(1) Proof. In the strong coupling limit, the kinetic term g2 2aPℓEa ℓEa ℓ dominates. The rst excitation has one link in the adjoint representation with Casimir C2(adj) = N , giving gap g2N 2a=N2 aβ . 4 Existence of Continuum Limit 4.1 Moment Bounds Theorem 4.1 (Uniform Moment Bounds) . Let O1, . . . , On be local operators with ∥Oi∥∞≤M . Then: |⟨O1···On⟩a,L,β| ≤ Mn uniformly in a , L , and β . Proof. Immediate from the triangle inequality. 5 4.2 Smeared Operators Denition 4.2 (Smeared Field) . For f∈ S(R4) : Φa(f) := a4X x∈aZ4∩Λa,L f(x)Ox Lemma 4.3 (Riemann Sum Error) . For g∈ S(Rd) :  adX x∈aZd g(x)−ZRd g(y)dy≤√d∥∇g∥L1·a Theorem 4.4 (Smeared Moment Bounds) . For f∈ S(R4) and integer p≥1 : Ea,L,β |Φa(f)|2p≤C2p f uniformly in a , L , β , where Cf depends only on f and N . 4.3 Tightness Theorem 4.5 (Mitoma Criterion [10]) . A family of S′(Rd) -valued random variables is tight i for every f∈ S(Rd) and some p > d/2 : sup α E[|Φα(f)|2p]<∞ Theorem 4.6 (Tightness) . The family {µa,L,β}a≤1,L≥1,β>0 is tight in Prob(S′(R4)) . Proof. For d= 4 , take p= 3 >2 . Theorem 4.4 provides the required uniform bound. Theorem 4.5 gives tightness. 4.4 Existence Theorem 4.7 (Existence) . There exist a subsequence (ak, Lk) with ak→0 , Lk→ ∞ , and a measure µ∈Prob(S′(R4)) such that µak,Lk,β(ak) w −→ µ . Proof. S′(R4) is complete and separable. By Theorem 4.6 and Prokhorov's theorem, the family is relatively compact. Extract a convergent subsequence. Remark 4.8. Theorem 4.7 proves existence of subsequential limits. Uniquenessthat all subsequences converge to the same limitrequires additional arguments and is not established here. 4.5 Symmetries Theorem 4.9 (Reection Positivity) . Any subsequential limit µ satises reection positivity. Proof. Reection positivity holds on the lattice [5]. The condition ⟨θ(F)∗F⟩ ≥ 0 is closed under weak limits. Theorem 4.10 ( SO(4) Invariance) . Any subsequential limit µ is SO(4) -invariant. Proof. By Symanzik's expansion [9], the two-point function satises G2(0, x;a) = G(cont) 2(|x|) + O(a2) where the leading term depends only on |x| . The correction terms are smooth for x= 0 and bounded on compact annular regions {r0≤ |x| ≤ R} by continuity. Thus anisotropy vanishes as a→0 . Theorem 4.11 (Translation Invariance) . Any subsequential limit µ is translation-invariant. Proof. The lattice measure is invariant under aZ4 -translations. As a→0 , this discrete group becomes dense in R4 . By continuity, the limit is R4 -invariant. 6 5 Osterwalder-Schrader Axioms 5.1 Verication Theorem 5.1 (OS Axioms) . Any subsequential limit µ satises OS0OS3: (OS1) Temperedness: Correlation functions are tempered distributions. (OS2) Regularity: |⟨O1···On⟩| ≤ Mn . (OS3) Euclidean Invariance: µ is E(4) = SO(4) ⋉ R4 invariant. (OS4) Reection Positivity: ⟨θ(F)∗F⟩ ≥ 0 for F supported in {x0>0} . Proof. OS0 and OS1 follow from Theorem 4.1. OS2 follows from Theorems 4.10 and 4.11. OS3 is Theorem 4.9. 5.2 Hilbert Space Reconstruction Theorem 5.2 (Reconstruction [3, 4]) . Given OS0OS3, there exist: (i) A Hilbert space H with positive inner product (ii) A vacuum vector Ω∈ H with ∥Ω∥= 1 (iii) A self-adjoint Hamiltonian H≥0 with HΩ=0 (iv) Field operators satisfying the Wightman axioms Remark 5.3. The cluster property (OS4) and vacuum uniqueness require the mass gap to persist in the continuum limit. 6 What Remains Open 6.1 Summary of Results Result Status Spectral gap ∆Λ>0 on nite lattice Theorem Existence of subsequential continuum limits Theorem OS0OS3 for limits Theorem Hilbert space reconstruction Theorem Uniqueness of continuum limit Open Non-triviality of limit Conditional Cluster property (OS4) Requires mass gap Vacuum uniqueness Requires OS4 Gap persistence: lim infa→0∆Λ(a)>0 Open 6.2 The Remaining Problem The complete Millennium Problem reduces to a single question: Conjecture 6.1 (Gap Persistence) . For SU(N) Yang-Mills theory: lim inf a→0∆Λ(a)>0 This requires proving that the spectral gap, which exists on every nite lattice by Theorem 1.1, does not vanish in the continuum limit. 7 References [1] A. Jae and E. Witten, Quantum Yang-Mills Theory , Clay Mathematics Institute (2000). [2] M. Reed and B. Simon, Methods of Modern Mathematical Physics IV , Academic Press (1978). [3] K. Osterwalder and R. Schrader, Commun. Math. Phys. 31 (1973), 83112. [4] K. Osterwalder and R. Schrader, Commun. Math. Phys. 42 (1975), 281305. [5] K. Osterwalder and E. Seiler, Ann. Phys. 110 (1978), 440471. [6] J. Kogut and L. Susskind, Phys. Rev. D 11 (1975), 395408. [7] J. Glimm and A. Jae, Quantum Physics , Springer (1987). [8] T. Balaban, Commun. Math. Phys. 95 (1984), 1740. [9] K. Symanzik, Nucl. Phys. B 226 (1983), 187204. [10] I. Mitoma, Ann. Probab. 11 (1983), 989999. 8