The melting crystal corner in three dimensions — the set of plane partitions, i.e. stacks of unit cubes in the positive octant that decrease weakly along all three directions, weighted by q#boxesq^{\#\text{boxes}} — admits a canonical quantum dynamics in the spirit of Rokhsar–Kivelson quantization. Dijkgraaf, Orlando and Reffert (Nucl. Phys. B 811 (2009) 463) showed that this '3D quantum crystal' Hamiltonian is an infinite system of ferromagnetic XXZ spin chains coupled by interlacing (vicious-walker) constraints, computed its ground state exactly, and found numerically that its mass gap agrees, to high precision, with the value γ=J(q+1/q−2)\gamma=J(\sqrt{q}+1/\sqrt{q}-2) that is exact in one dimension and rigorously proven in two dimensions (where the model reduces to the single XXZ chain). They conjectured that the mass gap of the quantum crystal is the same in all dimensions, and asked whether the 3D system is integrable, what its spectrum and higher eigenvectors are, and whether a quantum-group symmetry extends to it. Despite a dedicated follow-up program (Kagome-lattice reformulations, fermion–boson duality, a conjectural affine Yangian representation, and Yang–Baxter integrable subsystems), none of these questions has been settled. The problem is to prove or refute the dimension-independent mass gap conjecture and to determine the exact solvability — spectrum, eigenvectors, and symmetry algebra — of the 3D quantum crystal Hamiltonian.

Background

A 'melting crystal' in dd dimensions is the corner of a dd-dimensional lattice filled with unit cubes stacked in the positive directions, whose configurations carry the Boltzmann weight qq raised to the number of boxes. In two dimensions the configurations are Young diagrams (integer partitions) and in three dimensions plane partitions (three-dimensional Young diagrams); the three-dimensional partition function is the MacMahon function ∏k≥1(1−qk)−k\prod_{k\geq 1}(1-q^k)^{-k}, which equals the topological-string partition function on C3\mathbb{C}^3 and counts Donaldson–Thomas-type BPS states (a class of protected bound states in string-theoretic compactifications). Rokhsar and Kivelson (Phys. Rev. Lett. 61 (1988) 2376) introduced a way to quantize such classical stochastic systems: one builds a quantum Hamiltonian whose ground state is exactly the classical Boltzmann distribution (the 'RK point'), so that quantum dynamics and classical relaxation are two faces of the same generator.

Dijkgraaf, Orlando and Reffert (DOR) constructed these Hamiltonians explicitly for the crystal corner in one, two and three dimensions. Adding or removing a box is a hop with amplitude JJ, supplemented by diagonal 'growth-operator' terms that weight the number of places where a box can be added by q\sqrt q and the number of removable boxes by 1/q1/\sqrt q; at the RK point (J=VJ=V) the ground state is the qq-weighted superposition of all configurations. The one-dimensional crystal reduces to a massive lattice Laplacian and is exactly diagonalizable, with gap γ=J(q+1/q−2)\gamma=J(\sqrt q+1/\sqrt q-2). The two-dimensional crystal maps onto the half-filled sector of the ferromagnetic Heisenberg XXZ chain (a nearest-neighbor spin-1/2 chain with anisotropic zz-coupling Δ=(q+1/q)/2\Delta=(\sqrt q+1/\sqrt q)/2); it is solvable by the Bethe ansatz (the standard exact method for one-dimensional integrable quantum systems, which constructs many-body eigenfunctions from two-body scattering data), and its gap γ=2J(Δ−1)=J(q+1/q−2)\gamma=2J(\Delta-1)=J(\sqrt q+1/\sqrt q-2) is rigorously known from the XXZ chain gap theorem of Koma and Nachtergaele (Lett. Math. Phys. 40 (1997) 1-16).

The three-dimensional crystal is the open case. Representing plane partitions as NN non-intersecting directed paths ('vicious walkers'), or equivalently as diagonally sliced interlaced integer partitions, the Hamiltonian becomes an infinite system of parallel XXZ chains in which a fermion may hop only when the blocking sites of the two neighboring chains are occupied — a three-body interlacing constraint that couples each chain to its two neighbors. Equivalently (Araujo–Orlando–Reffert) the model is a particle–hole occupancy problem on the Kagome lattice, the two-dimensional lattice of corner-sharing triangles. DOR computed the exact ground state and studied the spectrum numerically in small boxes (a 3×3×43\times3\times4 cube with 4116 configurations, and all plane partitions of integers up to 12 with 3462 configurations): the numerically extracted mass gap matches γ=J(q+1/q−2)\gamma=J(\sqrt q+1/\sqrt q-2) to high precision, the first excited level appears Z3\mathbb{Z}_3-symmetric and non-degenerate, and the second appears Z2\mathbb{Z}_2-symmetric and thrice degenerate. DOR therefore conjectured that 'the mass gap of the quantum crystal is the same in all dimensions and equals γ=−J(2−q−1/q)\gamma=-J(2-\sqrt q-1/\sqrt q)'. A heuristic explanation is that the melting dynamics is the one-particle sector of an XXZ-type Hamiltonian H=−J[△−(q1/2−1)deg⁡−(α)−(q−1/2−1)deg⁡+(α)]H=-J[\triangle-(q^{1/2}-1)\deg^-(\alpha)-(q^{-1/2}-1)\deg^+(\alpha)] on the directed acyclic configuration graph (DAG), where deg⁡−(α)\deg^-(\alpha) and deg⁡+(α)\deg^+(\alpha) are, in DOR's convention, the out- and in-degree of the configuration node α\alpha; DOR conjectured the same gap for the one-particle XXZ sector on any DAG with deg⁡+/deg⁡−→1\deg^+/\deg^-\to1, a statement proven only for trees (Handjani–Jungreis 1996). The follow-up literature (Orlando–Reffert–Reshetikhin; Araujo–Orlando–Reffert; Araujo; Wang) developed reformulations and an (unproven) affine Yangian symmetry conjecture — the affine Yangian of gl^(1)\widehat{\mathfrak{gl}}(1) being the infinite-dimensional algebra conjectured to act on the plane-partition state space — and explicitly reports that both central conjectures — the value of the mass gap and the integrability of the 3D system — remain open.

Problem Statement

Settle the two central conjectures that Dijkgraaf, Orlando and Reffert left open for the three-dimensional quantum crystal Hamiltonian H3dH_{\text{3d}} (the Rokhsar–Kivelson quantization of the melting crystal corner at J=VJ=V, equivalently the infinite system of XXZ ferromagnetic chains coupled by interlacing/vicious-walker constraints, or the particle–hole hopping model on the Kagome lattice), whose ground state is the exactly known qq-weighted superposition of plane partitions.

The overarching objective is the exact solvability of H3dH_{\text{3d}}, which raises two coupled central questions:

  1. Mass gap: prove, or refute, that the spectral gap of H3dH_{\text{3d}} above its ground state equals γ=J(q+1/q−2)\gamma=J\big(\sqrt{q}+1/\sqrt{q}-2\big) — the value exact in one dimension and rigorously proven in two — so that the gap of the quantum crystal is the same in all dimensions. A heuristic mechanism for this dimension-independence discussed by DOR is the conjectured universality that the one-particle XXZ Hamiltonian −J[△−(q1/2−1)deg⁡−(α)−(q−1/2−1)deg⁡+(α)]-J[\triangle-(q^{1/2}-1)\deg^-(\alpha)-(q^{-1/2}-1)\deg^+(\alpha)] on any directed acyclic configuration graph with deg⁡+/deg⁡−→1\deg^+/\deg^-\to1 has the same gap (rigorously known only for trees).

  2. Integrability and spectrum: determine whether H3dH_{\text{3d}} is integrable in the precise sense used by the cited follow-up program — whether an RR-matrix satisfying the Yang–Baxter equation with a commuting family of transfer matrices, equivalently a Bethe-ansatz-type construction of the many-body eigenvectors (possibly an appropriate generalization of the Bethe ansatz to the constrained system of coupled chains), exists for H3dH_{\text{3d}} — and as part of an affirmative answer exhibit explicit closed-form eigenvectors for the first excited levels in terms of plane partitions (at least the reported Z3\mathbb{Z}_3-symmetric first and Z2\mathbb{Z}_2-symmetric second excited multiplets) and identify which symmetry algebra of the Hamiltonian acts on the 3D system (whether the Uq[SU(2)]U_{\sqrt{q}}[SU(2)] quantum-group symmetry of the XXZ chain (a qq-deformation of the SU(2) rotation symmetry that survives in the anisotropic chain), or the conjectured affine Yangian of gl^(1)\widehat{\mathfrak{gl}}(1) (an infinite-dimensional symmetry algebra acting on the plane-partition state space), acts on the 3D system); the analytic structure of the full spectrum is a further goal and is not a condition on resolving this question.

A rigorous proof that the 3D gap differs from γ\gamma, or that no such integrable structure exists (neither a Yang–Baxter RR-matrix with a commuting transfer-matrix family nor a Bethe-ansatz-type construction), is a negative resolution of the corresponding question and is an equally valid form of answer. The exactly solved one- and two-dimensional reductions and the published finite-size numerics (gap of the 3×3×43\times3\times4 cube; Z3\mathbb{Z}_3 first and thrice-degenerate Z2\mathbb{Z}_2 second excited levels) are the established data anchoring these questions.

References

  1. R. Dijkgraaf, D. Orlando, S. Reffert, 'Quantum crystals and spin chains', Nuclear Physics B 811 (2009) 463-490, doi:10.1016/j.nuclphysb.2008.11.027, arXiv:0803.1927, https://arxiv.org/abs/0803.1927
  2. T. Koma, B. Nachtergaele, 'The spectral gap of the ferromagnetic XXZ chain', Letters in Mathematical Physics 40 (1997) 1-16, doi:10.1023/A:1007351803403, arXiv:cond-mat/9512120, https://arxiv.org/abs/cond-mat/9512120
  3. D. Orlando, S. Reffert, N. Reshetikhin, 'On domain wall boundary conditions for the XXZ spin Hamiltonian' (2009), arXiv:0912.0348, https://arxiv.org/abs/0912.0348
  4. T. Araujo, D. Orlando, S. Reffert, 'Quantum crystals, Kagome lattice and plane partitions fermion-boson duality', Physical Review D 103 (2021) 026020, doi:10.1103/PhysRevD.103.026020, arXiv:2005.09103, https://arxiv.org/abs/2005.09103
  5. T. Araujo, 'Towards Yang-Baxter integrability of quantum crystal melting: From Kagome lattice to vertex models', Nuclear Physics B 963 (2021) 115287, doi:10.1016/j.nuclphysb.2020.115287, arXiv:2009.06812, https://arxiv.org/abs/2009.06812
  6. N. Wang, '3D fermions and affine Yangian', Nuclear Physics B 969 (2021) 115461, doi:10.1016/j.nuclphysb.2021.115461, https://doi.org/10.1016/j.nuclphysb.2021.115461
  7. A. Morozov, N. Tselousov, 'Hunt for 3-Schur polynomials', Physics Letters B 840 (2023) 137887, doi:10.1016/j.physletb.2023.137887, arXiv:2211.14956, https://arxiv.org/abs/2211.14956
  8. S. Handjani, D. Jungreis, 'Rate of convergence for shuffling cards by transpositions', Journal of Theoretical Probability 9 (1996) 983-993, doi:10.1007/BF02214260, https://doi.org/10.1007/BF02214260
  9. D. S. Rokhsar, S. A. Kivelson, 'Superconductivity and the quantum hard-core dimer gas', Physical Review Letters 61 (1988) 2376-2379, doi:10.1103/PhysRevLett.61.2376, https://doi.org/10.1103/PhysRevLett.61.2376
  10. D. Galakhov, W. Li, 'Charging solid partitions', Journal of High Energy Physics 2024 (2024) 043, doi:10.1007/JHEP01(2024)043, arXiv:2311.02751, https://arxiv.org/abs/2311.02751
  11. A. Prakash, J. S. Rao, S. A. Parameswaran, A. Lerose, 'Interface phases and dynamics in two-dimensional quantum magnets: A "holographic" approach from universality to quantum simulation' (2026), arXiv:2608.12312, https://arxiv.org/abs/2608.12312