For the M×NM\times N square lattice with M,NM,N odd, the generating function of close-packed dimer coverings leaving one prescribed site empty (a single monomer) is known in closed form only when the monomer lies on the boundary (Tzeng and Wu 2003; Wu's 2006 direct Pfaffian rederivation). Wu showed that the Pfaffian sign argument fails for a monomer in the interior, and the closed-form evaluation of the interior single-monomer generating function — together with its extension to finitely many prescribed monomers — has remained open. The subsequent literature (boundary monomers fully settled by Priezzhev-Ruelle, Giuliani-Jauslin-Lieb, and Pham; exact but numerically evaluated two-Pfaffian representations for bulk monomers by Allegra-Fortin and Allegra; ongoing computational work by Kong, Oh, and Byun-Goddard) contains no closed-form interior result. The surviving open core is a closed-form evaluation of the fixed-position monomer partition function for interior sites, with the general fixed-monomer problem as its natural extension.

Background

On an M×NM\times N square lattice, a dimer occupies an edge joining two neighboring sites, and a monomer is a site left uncovered. A close-packed dimer covering (a perfect matching of the lattice graph) covers every site exactly once. With horizontal and vertical dimers weighted xx and yy, the number of coverings is encapsulated in the dimer generating function. For MNMN even it is given by the 1961 exact solution of Kasteleyn and of Temperley and Fisher as a single finite product of cosines — when MM and NN are both even, at unit weights Z=∏m=1M/2∏n=1N/2(4cos⁡2(mπ/(M+1))+4cos⁡2(nπ/(N+1)))Z=\prod_{m=1}^{M/2}\prod_{n=1}^{N/2}\left(4\cos^{2}(m\pi/(M+1))+4\cos^{2}(n\pi/(N+1))\right), with analogous products covering the remaining parities. The method writes ZZ as a Pfaffian — the signed square root of the determinant of an antisymmetric matrix, Pf(A)2=det⁡A\mathrm{Pf}(A)^2=\det A, whose expansion sums over perfect matchings with signs — of a signed adjacency (Kasteleyn) matrix, after orienting the edges so that every face boundary carries an odd number of clockwise edges (a Kasteleyn orientation); this sign structure makes all matchings contribute with the same sign. The 1963 work of Fisher and Stephenson established that the correlation between two monomers (two vacancies) in the infinite lattice decays as r−1/2r^{-1/2}, but no closed-form defect partition function is known for the bulk.

Allowing monomers breaks the free-fermion solvability (the reducibility of the partition sum to a single Pfaffian or determinant just described): Heilmann and Lieb (1970) showed the monomer-dimer system has no phase transition, and Jerrum (1987) proved that computing the general monomer-dimer partition function, in which monomers may sit anywhere with fugacity, is #P-complete (a computationally intractable enumeration class) already on planar lattices. If the monomer positions are fixed, however, the problem maps to a close-packed dimer problem on an extended lattice obtained by attaching one new degree-1 site to each monomer site; this extended lattice is planar, so a Kasteleyn orientation exists by Kasteleyn's theorem and the fixed-position count is a single Pfaffian. The real difficulty is therefore not computability but closed form: no orientation with enough regularity to diagonalize the Kasteleyn matrix into an explicit spectral product is known once a monomer sits in the interior.

The one exact result beyond the pure dimer model concerns a single monomer on the boundary. Temperley (1974) exhibited a bijection between single-boundary-monomer dimer configurations on a planar lattice and spanning trees (connected acyclic subgraphs) on a related lattice; spanning trees are counted by the Matrix-Tree theorem as a cofactor of the graph Laplacian (the tree matrix), equivalently by the product of its nonzero eigenvalues. Using this bijection, Tzeng and Wu (2003) obtained, for M,NM,N odd and the monomer anywhere on the boundary (on the majority sublattice), the closed form

G(x,y)=x(M−1)/2 y(N−1)/2∏m=1(M−1)/2∏n=1(N−1)/2(4x2cos⁡2mπM+1+4y2cos⁡2nπN+1)G(x,y)=x^{(M-1)/2}\,y^{(N-1)/2}\prod_{m=1}^{(M-1)/2}\prod_{n=1}^{(N-1)/2}\left(4x^{2}\cos^{2}\frac{m\pi}{M+1}+4y^{2}\cos^{2}\frac{n\pi}{N+1}\right)

Wu (2006) rederived it directly by evaluating the Pfaffian of the (MN+1)(MN+1)-dimensional Kasteleyn matrix of the extended lattice: expanding along the added row and column reduces GG to a cofactor of the singular Kasteleyn matrix of the original lattice, which Wu evaluated by a pseudoinverse (ϵ\epsilon-regularization) limit, identifying the answer as (essentially) the product of the nonzero eigenvalues of the Kasteleyn matrix — mirroring the tree-matrix formula. The sign argument, however, uses that the monomer is on the boundary: for an interior monomer there exist transposition polygons (the closed loops obtained by superimposing two dimer configurations, whose signs govern Pfaffian cancellation) encircling the monomer whose sign cannot be controlled, so the Pfaffian, though evaluable, no longer equals the generating function. Wu announced a subsequent treatment of the interior case (with W. T. Lu), which was never published.

Later work has settled the boundary case completely and has produced exact representations — but no closed forms — for bulk monomers: Priezzhev and Ruelle (2008) computed correlations of arbitrarily many boundary monomers exactly in the scaling limit, which is a free-fermion theory of central charge c=1c=1; Wu, Tzeng and Izmailian (2011) solved a single boundary monomer on a nonbipartite lattice (one whose sites cannot be two-colored so that every edge joins unlike colors); Giuliani, Jauslin and Lieb (2016) proved a rigorous Pfaffian formula for monomer-dimer partition functions on arbitrary finite planar graphs with monomer weights nonzero only on the boundary, and Pham (2018) extended it to graphs on general surfaces, again with monomers restricted to the boundary. For bulk monomers, Allegra and Fortin (2014) expressed the partition function with any finite set of monomers at fixed positions as a product of two Pfaffians (evaluated numerically in the bulk), and Allegra (2015) developed the complete fermionic (Grassmann-integral, i.e. integration over anticommuting variables) solution with monomers acting as disorder operators — insertions that create the vacancy by flipping fermion signs along a line to the boundary — stating explicitly that in the bulk an exact closed-form expression on the discrete level seems out of reach. Computational studies of single-vacancy (one interior monomer) lattices continue without closed forms (Kong 2006 and 2024; Oh 2019; Byun and Goddard 2025, who prove Kong's divisibility conjecture for one-hole tilings).

Problem Statement

Let MM and NN be odd, and let Gm0,n0(x,y)G_{m_0,n_0}(x,y) denote the dimer-monomer generating function of the M×NM\times N square lattice with horizontal and vertical dimer weights xx and yy, close-packed except for a single monomer fixed at site (m0,n0)(m_0,n_0), necessarily on the majority sublattice (that of the four corners) — equivalently, the generating polynomial of the near-perfect matchings (matchings leaving exactly one site uncovered) of the lattice graph. When the monomer lies on the boundary, GG is independent of its position and given by the Tzeng-Wu closed product displayed in the Background. The problem, posed by Wu (2006) and left unanswered since (the announced follow-up with W. T. Lu was never published), is:

  1. (Core) Determine a closed-form expression for Gm0,n0(x,y)G_{m_0,n_0}(x,y) for a monomer at an interior site, with explicit dependence on (M,N,m0,n0)(M,N,m_0,n_0), valid for all odd M,NM,N and all admissible interior sites, together with its derivation. Neither the derivation route nor the answer shape is prescribed: the direct-Pfaffian, orientation-based, and spectral-product structures discussed in the Background are context and available tools, not requirements. A bare Pfaffian, determinant, or numerical evaluation of the extended lattice — a finite standing object that merely restates the fixed-position count — is not by itself such a closed-form expression.

  2. (Generalization — optional extension, not part of the completion boundary) The same closed-form question for the close-packed partition function with finitely many prescribed monomer sites at arbitrary (including interior) positions, the number of monomers being such that the remaining sites are even in number: exact two-Pfaffian (Allegra-Fortin 2014) and disorder-operator (Allegra 2015) representations are known but admit no closed-form evaluation, and Allegra (2015) states that in the bulk an exact closed-form expression on the discrete level 'seems out of reach'.

Numerical algorithms (transfer matrices, recursions) and the mere reduction of the count to a Pfaffian or determinant of the extended lattice are known and do not by themselves constitute an answer; the completion boundary is a closed-form expression for the single interior monomer, with explicit dependence on (M,N,m0,n0)(M,N,m_0,n_0), valid for all lattice sizes and all admissible interior positions, with proof. A solution may additionally illuminate why the boundary-monomer product mirrors the spanning-tree (tree-matrix) eigenvalue formula — Wu's secondary structural question — but that interpretive question is not part of the completion boundary.

References

  1. F. Y. Wu, 'Pfaffian solution of a dimer-monomer problem: Single monomer on the boundary', Physical Review E 74, 020104 (2006). DOI: 10.1103/PhysRevE.74.020104; arXiv:cond-mat/0607647. https://arxiv.org/abs/cond-mat/0607647
  2. W.-J. Tzeng and F. Y. Wu, 'Dimers on a Simple-Quartic Net with a Vacancy', Journal of Statistical Physics 110, 671-689 (2003). DOI: 10.1023/A:1022155701655. https://doi.org/10.1023/A:1022155701655
  3. P. W. Kasteleyn, 'The statistics of dimers on a lattice', Physica 27, 1209-1225 (1961). DOI: 10.1016/0031-8914(61)90063-5. https://doi.org/10.1016/0031-8914(61)90063-5
  4. H. N. V. Temperley and M. E. Fisher, 'Dimer problem in statistical mechanics—an exact result', Philosophical Magazine 6, 1061-1063 (1961). DOI: 10.1080/14786436108243366. https://doi.org/10.1080/14786436108243366
  5. H. N. V. Temperley, 'Enumeration of graphs on a large periodic lattice', in Combinatorics: Proceedings of the British Combinatorial Conference, London Mathematical Society Lecture Note Series 13 (Cambridge University Press, 1974), pp. 155-160. DOI: 10.1017/cbo9780511662072.024. https://doi.org/10.1017/cbo9780511662072.024
  6. O. J. Heilmann and E. H. Lieb, 'Monomers and Dimers', Physical Review Letters 24, 1412-1414 (1970). DOI: 10.1103/PhysRevLett.24.1412. https://doi.org/10.1103/PhysRevLett.24.1412
  7. M. E. Fisher and J. Stephenson, 'Statistical Mechanics of Dimers on a Plane Lattice. II. Dimer Correlations and Monomers', Physical Review 132, 1411-1431 (1963). DOI: 10.1103/PhysRev.132.1411. https://doi.org/10.1103/PhysRev.132.1411
  8. M. Jerrum, 'Two-dimensional monomer-dimer systems are computationally intractable', Journal of Statistical Physics 48, 121-134 (1987). DOI: 10.1007/BF01010403. https://doi.org/10.1007/BF01010403
  9. Y. Kong, 'Packing dimers on (2p+1) x (2q+1) lattices', Physical Review E 73, 016106 (2006). DOI: 10.1103/PhysRevE.73.016106. https://doi.org/10.1103/PhysRevE.73.016106
  10. J. Bouttier, M. Bowick, E. Guitter, and M. Jeng, 'Vacancy localization in the square dimer model', Physical Review E 76, 041140 (2007). DOI: 10.1103/PhysRevE.76.041140. https://doi.org/10.1103/PhysRevE.76.041140
  11. V. Priezzhev and P. Ruelle, 'Boundary monomers in the dimer model', Physical Review E 77, 061126 (2008). DOI: 10.1103/PhysRevE.77.061126; arXiv:0803.2652. https://arxiv.org/abs/0803.2652
  12. F. Y. Wu, W.-J. Tzeng, and N. S. Izmailian, 'Exact solution of a monomer-dimer problem: A single boundary monomer on a nonbipartite lattice', Physical Review E 83, 011106 (2011). DOI: 10.1103/PhysRevE.83.011106. https://doi.org/10.1103/PhysRevE.83.011106
  13. N. Allegra and J.-Y. 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
  14. N. Allegra, 'Exact solution of the 2d dimer model: Corner free energy, correlation functions and combinatorics', Nuclear Physics B 894, 685-732 (2015). DOI: 10.1016/j.nuclphysb.2015.03.022; arXiv:1410.4131. https://arxiv.org/abs/1410.4131
  15. A. Giuliani, I. Jauslin, and E. H. Lieb, 'A Pfaffian Formula for Monomer-Dimer Partition Functions', Journal of Statistical Physics 163, 211-238 (2016). DOI: 10.1007/s10955-016-1484-1; arXiv:1510.05027. https://arxiv.org/abs/1510.05027
  16. A. M. Pham, 'A Pfaffian formula for the monomer-dimer model on surface graphs', Letters in Mathematical Physics 108, 1885-1904 (2018). DOI: 10.1007/s11005-018-1059-y; arXiv:1705.00992. https://arxiv.org/abs/1705.00992
  17. S. Oh, 'State matrix recursion method and monomer-dimer problem', Discrete Mathematics 342, 1434-1445 (2019). DOI: 10.1016/j.disc.2019.01.022. https://doi.org/10.1016/j.disc.2019.01.022
  18. Y. Kong, 'Recurrence solution of monomer-polymer models on two-dimensional rectangular lattices', Physical Review E 110, 054135 (2024). DOI: 10.1103/PhysRevE.110.054135; arXiv:2405.09457. https://arxiv.org/abs/2405.09457
  19. S. H. Byun and W. Goddard, 'A Note on One-Hole Domino Tilings of Squares and Rectangles' (2025). arXiv:2502.05918. https://arxiv.org/abs/2502.05918