The close-packed classical dimer model on a two-dimensional bipartite lattice is described at long distances by a Gaussian height field theory. A June 2026 derivation by S. Powell shows constructively how this theory — its stiffness, its boundary-tilt structure, and the identification of dimer observables with height operators — follows from an exact Grassmann-integral representation of the partition function, via a continuum Dirac fermion and path-integral bosonization, for both square and honeycomb lattices. The paper explicitly defers one piece to future work: monomers, the defects of the close-packing constraint, are known to correspond to topological defects (vortices) of the height field — monomer insertions are expressed through vertex operators dual to e±2πihe^{\pm 2\pi i h} — but this correspondence has never been derived constructively from the fermionic representation. The obstacle is concrete: the microscopic monomer insertion is a nonlocal Jordan–Wigner-type string, which the earlier transfer-matrix bosonization of Wilkins and Powell (2023) explicitly could not handle, forcing them to construct the vortex operator in the continuum theory and confirm its asymptotics instead. The open problem is to close this gap: bosonize the monomer insertion itself, track the spin-sector sum with its complex weights, and reproduce the exact Fisher–Stephenson monomer correlations (the r−1/2r^{-1/2} decay for opposite-sublattice pairs and the exact vanishing for same-sublattice pairs).

Background

In the classical dimer model, each link (edge) of a lattice carries either zero or one dimer, and in the close-packed ensemble every site is touched by exactly one dimer. On bipartite lattices — whose sites split into two sublattices A and B, with every link joining an A site to a B site; the square and honeycomb lattices are the standard examples — the equal-weight ensemble of close-packed configurations forms a Coulomb phase: a topologically ordered, critical phase in which correlations decay algebraically with distance. The long-wavelength description is obtained by rewriting the close-packing constraint as a lattice Gauss law for a flux carried by each dimer, and resolving it in two dimensions with a scalar height field hh, defined on the faces of the lattice so that crossing a dimer changes hh by a fixed amount. The resulting continuum theory is a Gaussian free scalar with action proportional to ∫(∇h)2\int(\nabla h)^2 and stiffness κ\kappa (equal to π\pi for the uniform model), equivalent to a two-dimensional Coulomb gas; Kenyon (2001) proved rigorously that the height is distributed as a Gaussian free field. Microscopic dimer observables map to gradients and exponentials (vertex operators) of hh.

On planar graphs the dimer model is exactly solvable: the partition function is a Pfaffian (the signed square root of a determinant of the Kasteleyn matrix), and can equivalently be written as an integral over anticommuting Grassmann variables, one per lattice site, which makes the free-fermion structure explicit. Bosonization — the exact equivalence, in two dimensions, between massless Dirac fermions and a compactified free boson, under which fermion bilinears become gradients or exponentials of the boson field — then offers a route from the exact solution to the height theory.

The height theory was historically justified as a heuristic coarse-graining, with its coefficients fixed by matching to exact correlations. Wilkins and Powell (2023) gave the first constructive derivation, by bosonizing the continuum limit of Lieb's transfer-matrix (free-fermion) solution; Powell (2026), the source of this problem, gave a more direct real-space version: expanding the Grassmann-integral representation in slowly varying fermion modes yields a massless Dirac fermion in two-dimensional Euclidean space, which path-integral bosonization maps exactly onto the height theory, with stiffness κ=π\kappa=\pi, for both square and honeycomb lattices. Sources coupling to the flux density and to the valence-bond-solid (VBS) order parameter — the standard complex, sublattice-valued symmetry-breaking order parameter of dimer models — map to the gradient and vertex operators of hh, and the sum over fermion boundary phases (spin sectors) maps to the sum over height boundary tilts.

A monomer is a defect of the close-packing constraint: a site left uncovered. Monomers are the fundamental fractionalized excitations of the dimer Coulomb phase: each carries an effective magnetic charge whose sign is fixed by its sublattice, opposite-sublattice pairs attract through a logarithmic entropic Coulomb interaction, and Fisher and Stephenson (1963) computed the exact leading decay Gm∼r−1/2G_m \sim r^{-1/2} of the two-monomer distribution function (the ratio of the partition functions with and without the two monomers); by sublattice charge parity, the two-monomer distribution vanishes identically when both monomers sit on the same sublattice. In the height-field description a monomer is a topological defect (vortex): a point around which the height winds, represented in the continuum theory by a magnetic vertex operator (in the source's normalization, dual to e±2πihe^{\pm 2\pi i h}) rather than by the ordinary gradient and exponential operators that represent dimer observables.

This monomer–defect correspondence is established knowledge — it follows from Coulomb-gas arguments and is confirmed by exact asymptotics — but it has never been derived constructively from the fermionic representation. In the transfer-matrix solution the microscopic monomer insertion is a nonlocal string operator (a Jordan–Wigner-type product over intervening sites), and Wilkins and Powell (2023) found that it cannot be bosonized by their approach; they instead wrote down the vortex operator in the continuum theory and checked that it reproduces the correct leading asymptotics. Powell (2026) likewise derives the full close-packed theory but contains no monomer machinery, stating that the partition function with inserted monomers can be expressed in terms of vertex-operator correlations while leaving a constructive derivation of this relationship to future work. Closing this gap is the problem posed here.

Problem Statement

Starting from the exact Grassmann-integral (equivalently, free-fermion) representation of the close-packed classical dimer model on the square and honeycomb lattices — which the source framework treats on an equal footing — with a fixed set of monomer insertions (specified sites exempted from the close-packing constraint), derive constructively, through the continuum Dirac-fermion limit and bosonization, the correspondence between monomers and topological defects of the height field. Concretely, a solution must: (i) show how monomer insertions modify the Grassmann integral, together with the treatment of the sum over fermion boundary phases (spin sectors) with their complex weights; (ii) carry out the bosonization of the resulting insertion operators so as to obtain, rather than assume, that each monomer acts as a vortex of the height field — a magnetic vertex operator dual to e±2πihe^{\pm 2\pi i h}, with charge sign fixed by the monomer's sublattice — such that the partition function with monomers is derived, within the construction itself rather than by citing the known continuum operator, as the corresponding height-theory vertex-operator correlator (at least at the level of exact long-distance asymptotics), with the normalization relating the lattice and continuum theories fixed within the derivation; and (iii) verify the result against exact lattice asymptotics: at minimum the Fisher–Stephenson decay Gm∼r−1/2G_m \sim r^{-1/2} of the two-monomer distribution function for opposite-sublattice pairs on the isotropic square lattice, together with the exact vanishing of the two-monomer distribution for same-sublattice pairs. The same Grassmann-integral method may extend to non-translation-symmetric settings (quasicrystalline tilings, spatially disordered weights) and beyond perturbation theory for interacting dimers; progress on those extensions is adjacent to, but does not by itself answer, the question posed here.

References

  1. Stephen Powell, 'Derivation of height field theory for the two-dimensional classical dimer model from a Grassmann-integral representation' (2026), arXiv:2606.17154, DOI 10.48550/arXiv.2606.17154, https://arxiv.org/abs/2606.17154
  2. Neil Wilkins and Stephen Powell, 'Derivation of field theory for the classical dimer model using bosonization', Physical Review E 107, 054126 (2023), DOI 10.1103/PhysRevE.107.054126, arXiv:2301.09665, https://arxiv.org/abs/2301.09665
  3. Neil Wilkins and Stephen Powell, 'Topological sectors, dimer correlations, and monomers from the transfer-matrix solution of the dimer model', Physical Review E 104, 014145 (2021), DOI 10.1103/PhysRevE.104.014145, arXiv:2103.03260, https://arxiv.org/abs/2103.03260
  4. Michael E. Fisher and John Stephenson, 'Statistical Mechanics of Dimers on a Plane Lattice. II. Dimer Correlations and Monomers', Physical Review 132, 1411 (1963), DOI 10.1103/PhysRev.132.1411, https://doi.org/10.1103/PhysRev.132.1411
  5. P. Fendley, R. Moessner, and S. L. Sondhi, 'Classical dimers on the triangular lattice', Physical Review B 66, 214513 (2002), DOI 10.1103/PhysRevB.66.214513, https://doi.org/10.1103/PhysRevB.66.214513
  6. Nicolas Allegra, 'Exact solution of the 2d dimer model: Corner free energy, correlation functions and combinatorics', Nuclear Physics B 894, 685 (2015), DOI 10.1016/j.nuclphysb.2015.03.022, arXiv:1410.4131, https://arxiv.org/abs/1410.4131
  7. Nicolas Allegra and Jean-Yves Fortin, 'Grassmannian representation of the two-dimensional monomer-dimer model', Physical Review E 89, 062107 (2014), DOI 10.1103/PhysRevE.89.062107, arXiv:1402.5512, https://arxiv.org/abs/1402.5512
  8. Fabien Alet, Yacine Ikhlef, Jesper Lykke Jacobsen, Grégoire Misguich, and Vincent Pasquier, 'Classical dimers with aligning interactions on the square lattice', Physical Review E 74, 041124 (2006), DOI 10.1103/PhysRevE.74.041124, https://doi.org/10.1103/PhysRevE.74.041124
  9. P. Falco, 'Interacting fermions picture for dimer models', Physical Review E 87, 060101 (2013), DOI 10.1103/PhysRevE.87.060101, https://doi.org/10.1103/PhysRevE.87.060101
  10. Alessandro Giuliani, Vieri Mastropietro, and Fabio Lucio Toninelli, 'Height fluctuations in non-integrable classical dimers', EPL 109, 60004 (2015), DOI 10.1209/0295-5075/109/60004, https://doi.org/10.1209/0295-5075/109/60004
  11. Alessandro Giuliani, Vieri Mastropietro, and Fabio Lucio Toninelli, 'Height fluctuations in interacting dimers', Annales de l'Institut Henri Poincaré, Probabilités et Statistiques 53, 98 (2017), DOI 10.1214/15-AIHP710, arXiv:1406.7710, https://arxiv.org/abs/1406.7710
  12. Richard Kenyon, 'Dominos and the Gaussian Free Field', The Annals of Probability 29, 1128 (2001), DOI 10.1214/aop/1015345599, https://doi.org/10.1214/aop/1015345599
  13. Felix Flicker, Steven H. Simon, and S. A. Parameswaran, 'Classical Dimers on Penrose Tilings', Physical Review X 10, 011005 (2020), DOI 10.1103/PhysRevX.10.011005, arXiv:1902.02799, https://arxiv.org/abs/1902.02799
  14. Shobhna Singh and Felix Flicker, 'Exact solution to the quantum and classical dimer models on the spectre aperiodic monotiling', Physical Review B 109, L220303 (2024), DOI 10.1103/PhysRevB.109.L220303, arXiv:2309.14447, https://arxiv.org/abs/2309.14447
  15. S. Caracciolo, R. Fabbricatore, M. Gherardi, R. Marino, G. Parisi, and G. Sicuro, 'Criticality and conformality in the random dimer model', Physical Review E 103, 042127 (2021), DOI 10.1103/PhysRevE.103.042127, https://doi.org/10.1103/PhysRevE.103.042127
  16. Michael Aizenman, Manuel Laínz Valcázar, and Simone Warzel, 'Pfaffian Correlation Functions of Planar Dimer Covers', Journal of Statistical Physics 166, 1078 (2017), DOI 10.1007/s10955-016-1684-8, arXiv:1609.02824, https://arxiv.org/abs/1609.02824
  17. E. A. Ramirez Trino, M. A. Seifi MirJafarlou, and M. A. Rajabpour, 'A Generalized Grassmann-Pfaffian Framework for Monomer-Dimer and Spanning Trees' (2025), arXiv:2511.18495, https://arxiv.org/abs/2511.18495