The Yang-Mills Existence and Mass Gap Problem Page 1 The Yang-Mills Existence and Mass Gap Problem Statement, Evidence, and Obstacles to a Proof Includes illustrative models of runaway (finite-time blow-up) behaviour and a hypothetical form of the result Abstract This document summarizes the Yang-Mills existence and mass gap problem, one of the seven Clay Millennium Prize problems. Numerical simulation and experiment indicate that quantum Yang-Mills theory in four spacetime dimensions has a strictly positive energy gap of roughly 1,600 MeV. No rigorous proof exists. The document states the problem, reviews the supporting evidence, identifies the principal obstacle (the transition from strong to weak coupling as the lattice spacing tends to zero), and presents simplified illustrative models. These models are not physical derivations. 1. Statement of the problem Level Formulation Informal Show that the force binding atomic nuclei has a smallest nonzero energy step, and that this step cannot be made arbitrarily small. Intermediate Construct quantum Yang-Mills theory in four-dimensional spacetime with a compact simple gauge group and prove that its Hamiltonian has a strictly positive spectral gap above the vacuum. Formal Δ = E 1 − E 0 > 0, uniformly as the lattice spacing a → 0 and the volume → ∞. Here E 0 is the vacuum energy and E 1 is the lowest energy above it. The problem has two parts: existence (a mathematically rigorous quantum theory satisfying standard field-theory axioms) and the mass gap (Δ > 0). 2. Physical evidence • Lattice QCD discretizes spacetime on a grid. It places the lightest glueball, a bound state of gluons alone, at approximately 1,500 to 1,800 MeV (about 1.6 to 1.9 times the proton mass), and reproduces the proton mass to within a few percent. • Experiment has never observed an isolated quark or gluon. This confinement is consistent with a positive gap. • Measurement. On the lattice the gap is extracted from the exponential decay of a correlation function, C(t) ~ e −Δt , so that Δ = −ln C(t) / t. The Yang-Mills Existence and Mass Gap Problem Page 2 Limitation. A simulation operates at finite lattice spacing and finite volume, whereas the problem requires the limit a → 0 and volume → ∞ for all cases. Perturbation theory is an asymptotic approximation that fails at low energies, which is where the gap occurs. Numerical agreement is therefore strong evidence, but it does not constitute a proof. 3. Status by spacetime dimension Spacetime Existence Mass gap 2D (1 space + 1 time) Constructed rigorously; exactly solvable Confinement established exactly 3D (2 space + 1 time) Partially constructed; local solutions exist, but the full measure remains unconstructed Proven only at strong coupling 4D (3 space + 1 time) Open Open On a four-dimensional hypercubic lattice, a positive gap has been proven at strong coupling. Approaching the continuum, however, requires the coupling to decrease with the lattice spacing (asymptotic freedom), which moves the theory into the weak-coupling regime where no proof is known. Strong-coupling methods and weak-coupling methods each have error bounds valid only in their own regimes, and an intermediate range is covered by neither. Numerical matching across this range is accepted practice in physics but is not a rigorous argument. 4. Quantitative summary Δ = E 1 − E 0 ≈ 1,600 − 0 = 1,600 MeV (simulation) Required: Δ(a) ≥ c > 0 for all a, as a → 0. Equivalently, 0 < Δ < ∞ • For comparison, the proton mass is approximately 938 MeV. • The gap itself involves a single subtraction. The difficulty lies in establishing E 0 and E 1 rigorously, as the two lowest eigenvalues of an operator acting on infinitely many field degrees of freedom. • Failure mode. A collapse of the gap would occur gradually, for example 1,600, 800, 100, 1, ... → 0. Every term is positive while the limit is zero. A finite set of positive simulation values therefore cannot establish the result; a fixed lower bound c is required. • Upper bound. If the physical gap grew without limit as a → 0, the continuum theory would have no finite energy step. Simulations show a stable value, so this case is considered less likely. • Correlation length. The range over which an effect persists is 1/Δ. With ħc = 197.3 MeV·fm, 1/Δ ≈ 0.12 fm at Δ = 1,600 MeV. A vanishing gap corresponds to an unbounded range. The Yang-Mills Existence and Mass Gap Problem Page 3 5. Runaway behaviour: illustrative blow-up models To illustrate the distinction between divergence at a finite time and divergence only in an infinite limit, three simplified growth laws were examined for the spin rate ω of a rotating four-dimensional cube. These are mathematical toys and do not model Yang-Mills theory. Rule Behaviour Finite-time blow-up: dω/dt = ω 2 , with solution ω = ω 0 / (1 − ω 0 t) For ω 0 = 0.2 each doubling of ω takes half as long as the previous one (2.5 s, 1.25 s, 0.625 s, ...). The infinite sequence of doublings sums to a finite time, so ω diverges at t = 1/ω 0 = 5. Exponential growth: dω/dt proportional to ω ω becomes arbitrarily large but diverges only as t → ∞. Saturation ω approaches a finite ceiling (here 5) and never exceeds it, for any elapsed time. • No maximum finite value. Under blow-up, for any chosen speed there is a time before t = 5 at which it is exceeded. There is no largest finite speed preceding infinity. • Fluid dynamics. In viscous fluids, concentrated vortical structures can intensify through stretching, and viscosity acts as a damping mechanism. Whether damping always prevails is an open question in that field, and reported blow-up constructions remain unverified. • Connection to the mass gap. A vanishing gap corresponds to a divergent correlation length 1/Δ. In a collapse model with Δ(a) = 1,600 × (a − 0.3)/0.7, the gap reaches zero at a = 0.30. The largest finite spin rate on the sampled grid was 14, at a = 0.35, where Δ ≈ 114 MeV and the rate is 1,600/114. 6. Competing mechanisms • Tending toward a zero gap: gluons are massless at the classical level, and at fine lattice spacing the weak coupling makes them nearly free. Free massless particles have no gap. • Sustaining a positive gap: gluon self-interaction binds the field at long distances, producing confinement and a mass gap. Simulation indicates that self-interaction prevails at all accessible scales. A proof must show that it prevails at every scale down to a → 0. By contrast with the fluid case, where the claimed result is that a runaway can occur, the expected result here is that the collapse cannot occur. Both remain open pending independent verification. The Yang-Mills Existence and Mass Gap Problem Page 4 7. Claimed resolutions Several authors have posted claimed proofs of the mass gap, predominantly as self-published preprints, employing methods that include a five-dimensional orbifold regulator, tensor-network constructions, and block renormalization at very weak coupling. At least one claimed proof has been retracted. No independent verification, machine-checked formalization, or broad acceptance of any such claim was identified, and the claims are mutually inconsistent in method. 8. Requirements for a proof Component Role Status Simulation Lattice Yang-Mills at multiple spacings and large volumes, with rigorous and computer-verified error bounds Simulations exist; rigorous bounds are lacking Physics Asymptotic freedom, reflection positivity, and a confinement mechanism explain why a gap is expected Established Model A multiscale (renormalization-group) argument showing the gap persists under each halving of the lattice spacing, with errors that do not accumulate Open problem in 4D; a schematic version is given in Section 9 The model component, step by step. A complete argument of this type would proceed as follows: • Start at a coarse lattice spacing where strong-coupling methods prove a positive gap Δ 0 • Halve the lattice spacing. Show that the gap in physical units decreases by at most a fraction ε j at step j. • Repeat for every step, so that the guaranteed bound after k steps is Δ 0 × (1 − ε 1 )(1 − ε 2 )...(1 − ε k ). • Prove that the losses are summable, so that the product converges to a positive limit c as k → ∞. This is the uniform lower bound Δ(a) ≥ c > 0. Steps 1 to 3 have analogues in the schematic model of Section 9. Step 4, proving summability for the actual four-dimensional theory, is the unresolved part. A plausible research path is to examine the working 2D and 3D arguments, identify the step that fails in 4D, repair it, and verify the full chain with a proof assistant such as Lean. The missing four-dimensional idea has not been identified. The Yang-Mills Existence and Mass Gap Problem Page 5 9. Illustrative model and hypothetical result The three components were combined into a single schematic model. The coupling is g 2 = 6/(5 + 0.9k) at zoom level k, the simulated gap is 1,600 MeV with error bars ±1,600 × 0.008 × (1 + 0.12k), and the guaranteed floor is 1,600 times the product of (1 − ε j ) over zoom steps j, where ε j is the fractional loss at step j. Results for ε = 10%: Zoom step Constant losses (MeV) Shrinking losses (MeV) 5 945 1,389 10 558 1,374 20 195 1,373 30 68 1,373 Limit 0 approximately 1,373 • Constant losses: the floor decreases by a fixed fraction at every step and tends to zero, so no gap can be guaranteed. This reproduces the central difficulty. • Shrinking losses: with ε j = 0.1 × 0.6 j the losses are summable, and the floor converges to a positive value. Hypothetical result: Δ(a) ≥ c > 0 for all a, as a → 0 and the volume → ∞. In the model, c ≈ 1,373 MeV, below the simulated 1,600 MeV. The result is a lower bound, not an exact value. This result holds in the schematic model only. It has not been proven for the four-dimensional Yang-Mills theory. 10. Dimension-by-dimension comparison The table summarizes the physics, the mathematical model, and the simulation evidence in each dimension. The calculations that follow check the key claims numerically. Dim. Physics Model Simulation 2D No propagating gluon (d − 2 = 0 transverse polarizations). The coupling g 2 has dimension mass 2 Confinement follows from an exact area law. Exactly solvable (heat-kernel, Migdal-Rusakov form): Z = Σ R (dim R) 2−2h × exp[−g 2 A · C 2 (R)/2], summed over representations R on a surface of genus h and area A. Not needed for existence. The lattice result is exact: <W> = u A , with u = I 2 (β)/I 1 (β) for SU(2). The Yang-Mills Existence and Mass Gap Problem Page 6 Dim. Physics Model Simulation 3D One transverse polarization (d − 2 = 1). g 2 has dimension mass, so the theory is super-renormalizable (finitely many divergences). The gap scale is set by g 2 Stochastic quantization with regularity structures (local solutions constructed); a Hamiltonian eigenvalue approach has been outlined; a strong-coupling lattice expansion proves a gap. Lightest glueball about 4.7 √σ (SU(2)) and 4.3 √σ (SU(3)). These are approximate literature values quoted from memory and not re-verified here. 4D Two transverse polarizations. g is dimensionless but runs: asymptotic freedom, with a scale Λ generated by dimensional transmutation. Confinement is nonperturbative. Lattice theory plus multiscale renormalization-group analysis (Bałaban’s program, partly completed). A gap is proven at strong coupling; the weak-coupling extension is missing. Lightest glueball about 1,500 to 1,800 MeV; string tension √σ about 440 MeV; proton 938 MeV. 10.1 Two dimensions: exact calculation For SU(2), the plaquettes on a 2D lattice are statistically independent, so a Wilson loop enclosing A plaquettes gives <W> = u A , with u = I 2 (β)/I 1 (β) and β = 4/(g 2 a 2 ). The lattice string tension is therefore σa 2 = −ln u. At large β, u ≈ 1 − 3/(2β), so σa 2 → 3/(2β) = (3/8) g 2 a 2 β u = I 2 /I 1 σa 2 = −ln u 3/(2β) 1 0.2402 1.4263 1.5000 2 0.4331 0.8367 0.7500 4 0.6580 0.4185 0.3750 8 0.8192 0.1994 0.1875 16 0.9078 0.0967 0.0938 32 0.9535 0.0476 0.0469 • Continuum limit. The ratio σa 2 / (a 2 ) tends to the fixed value σ = (3/8) g 2 for SU(2). In general σ = g 2 C F / 2, which gives (2/3) g 2 for SU(3), since C F = 4/3. • Gap on a circle of length L. The energies are E R = g 2 L C 2 (R)/2. For the lowest nontrivial SU(2) representation, C 2 = 3/4, so Δ = (3/8) g 2 L = σL. With g 2 = 1 and L = 4, Δ = 1.5 in units of g. The gap is positive at every L and does not vanish in the continuum limit. The Yang-Mills Existence and Mass Gap Problem Page 7 10.2 Three dimensions: scaling calculation For SU(2) the lattice coupling is β = 4/(g 2 a), and the continuum limit is β → ∞ at fixed g 2 , so a g 2 = 4/β. The literature values m/√σ ≈ 4.72 and √σ ≈ 0.335 g 2 (approximate, quoted from memory) give Δ ≈ 4.72 × 0.335 g 2 ≈ 1.58 g 2 . The gap in lattice units is then aΔ = 1.58 × 4/β. β a g 2 = 4/β aΔ = 1.58 × 4/β 4 1.000 1.580 8 0.500 0.790 16 0.250 0.395 32 0.125 0.198 Each doubling of β halves the lattice gap. The lattice gap shrinks as a power law, 1/β, while the physical gap stays fixed. 10.3 Four dimensions: scale and running calculations • String tension. With √σ ≈ 440 MeV and ħc = 197.3 MeV·fm: √σ ≈ 2.23 fm −1 , so σ ≈ 4.97 fm −2 ≈ 0.194 GeV 2 ≈ 0.98 GeV/fm. • Glueball check. A lightest-glueball ratio of about 4.0 √σ gives 4.0 × 440 = 1,760 MeV, which lies within the 1,500 to 1,800 MeV range. The value of 1,600 MeV used elsewhere in this document is a representative figure inside that range. • Running coupling (one loop, SU(3)). b 0 = 11N/(48π 2 ) = 33/(48π 2 ) = 0.0697, and with β = 6/g 2 the lattice spacing obeys aΛ = exp(−β/(12 b 0 )) = exp(−β/0.836). Halving a therefore requires Δβ = 0.836 × ln 2 = 0.579. β a relative to β = 6 Lattice gap aΔ, 4D (relative) Lattice gap aΔ, 3D (relative, 6/β) 6.0 1.000 1.000 1.000 6.5 0.550 0.550 0.923 7.0 0.302 0.302 0.857 8.0 0.091 0.091 0.750 Holding the physical gap fixed, the 4D lattice gap must fall exponentially in β, while the 3D lattice gap falls only as 1/β. This is the quantitative form of the difficulty: in 4D, a proof must keep control of a quantity that shrinks exponentially fast as the lattice spacing decreases. The one-loop formula is a simplification. The schematic model in Section 9 used 0.9 per halving instead of 0.58, which changes the numbers but not the conclusion. The Yang-Mills Existence and Mass Gap Problem Page 8 11. Conclusion Physical evidence strongly supports a positive mass gap of approximately 1,600 MeV in four-dimensional Yang-Mills theory. A rigorous proof requires a uniform lower bound c > 0 valid as a → 0, which in turn requires control across the strong-to-weak coupling transition. This control has not been established, and the problem remains open. Hypothetical Proof Framework for the Yang–Mills Mass Gap Addendum to The Yang-Mills Existence and Mass Gap Problem: Statement, Evidence, and Obstacles to a Proof Scientific status: CONDITIONAL / NOT A PROOF OF THE MILLENNIUM PROBLEM. This addendum formalizes a possible multiscale argument. Its conclusion follows only if the assumptions below are proved for the actual four-dimensional Yang–Mills theory. Those assumptions are not established here. A. Goal and notation Let a_j be a sequence of lattice spacings tending to zero, and let L_j denote the corresponding spatial volume, with the infinite-volume limit controlled as well. Let ∆ _j be the lowest non-vacuum spectral energy above the vacuum for the regulated theory at scale j. The desired result is a continuum theory whose mass gap obeys ∆ _cont ≥ c > 0. Desired conclusion: ∆ _cont ≥ c > 0 The central difficulty is to prove a lower bound that survives every refinement and the required limits—not merely to observe positive gaps at finitely many lattice spacings. B. Explicit assumptions (the unproved part) • A1 — Starting gap: At some initial regulated scale, a rigorous estimate gives ∆ _0 ≥ d_0 > 0. • A2 — Uniform refinement estimate: For every j, including the relevant volume control, the next scale satisfies ∆ _{j+1} ≥ (1 − ε _j) ∆ _j, where 0 ≤ ε _j < 1. • A3 — Summable losses: The losses obey Σ _{j=0} ∞ ε _j < ∞ • A4 — Valid limiting theory: The regulated theories converge to a rigorously constructed continuum Yang–Mills theory, and the spectral estimates pass to the continuum and infinite-volume limits without losing the lower bound. A2–A4 are not known to hold in the required generality for four-dimensional Yang–Mills. Stating them as assumptions does not prove them. C. Conditional lemma Lemma. Suppose A1–A3 hold. Then every finite-scale gap has the lower bound ∆ _n ≥ d_0 ∏ j=0 n − 1 (1 − ε _j). Proof. For n = 1, A1 and A2 give ∆ n ≥ (1 − ε n ) ∆ n ≥ d n (1 − ε n ). Assume the bound holds at n. Applying A2 once more gives ∆ nnn ≥ (1 − ε n ) ∆ n ≥ d n ∏ j=0 n (1 − ε n ). By induction the inequality holds for every finite n. Because 0 ≤ ε _j < 1 and Σ ε _j is finite, the infinite product ∏ (1 − ε _j) converges to a strictly positive number. (Equivalently, the tail logarithms are summable; finitely many initial factors are also positive.) Define c := d_0 ∏ j=0 ∞ (1 − ε _j) > 0. Then ∆ _n ≥ c for all n. This proves a positive lower bound for the sequence of regulated gaps under A1–A3. Under the additional limiting assumption A4, the same lower bound would pass to the continuum theory. n D. What this argument would—and would not—establish If all four assumptions were rigorously proved for four-dimensional Yang–Mills, the argument would supply the core mass-gap lower bound. It would not, by itself, construct the full quantum field theory or verify every required field-theory axiom; those are part of the existence problem and must be handled explicitly. The value of c is not numerically determined by this abstract argument. In particular, the approximately 1,373 MeV value in the original document comes from its schematic model and must not be presented as a rigorous Yang–Mills bound. E. Research obligations needed to turn this into a genuine proof • Prove a starting positive gap with constants stated precisely. • Derive the refinement estimate from the actual four-dimensional Yang–Mills dynamics, uniformly in volume and gauge-field configurations—not by assuming it. • Obtain explicit ε _j bounds and prove their sum is finite in the continuum scaling regime. • Construct the continuum theory and justify convergence of the relevant Hamiltonians/spectra or correlation functions. • Show the lower bound survives the infinite-volume limit and satisfies the required axioms. • Submit the complete argument for independent expert review; formal verification could check suitable formalized steps but cannot replace missing mathematical estimates. F. Status and interpretation Item Status in this addendum Finite-product induction Proved conditionally from A1–A2 Positive infinite-product bound Proved conditionally from A1–A3 Four-dimensional refinement estimate OPEN / NOT PROVED Continuum and infinite-volume passage OPEN / NOT PROVED Yang–Mills existence and mass-gap problem NOT SOLVED BY THIS ADDENDUM Bottom line: This is a mathematically explicit conditional proof template: it identifies a sufficient route to a positive gap and proves the abstract product argument. The crucial Yang–Mills-specific estimates and the rigorous limiting construction remain unproved. It must be described as a hypothetical proof framework, not as a solution.