Exact solution of the two-dimensional hard-square lattice gas

The hard-square lattice gas - particles on the square lattice with nearest-neighbor exclusion, weighted by an activity $z$ - undergoes a continuous ordering transition to a checkerboard-sublattice solid that is characterized numerically to extraordinary precision ($z_c=3.79625517391234(4)$; two-dimensional Ising universality with a logarithmically divergent specific heat), yet the model has no exact solution: unlike Baxter's hard-hexagon model, the only exactly solved lattice gas with nearest-neighbor exclusion, its free energy, equation of state, and critical constants admit no known closed form. Baram and Fixman (1994) listed the unsolved analytic solution alongside then-open questions about the universality class, the solid-side specific-heat exponent, and virial-series convergence. The subsequent literature shows: the universality and exponent questions have been settled numerically, the singularity structure of the low-density series has been characterized precisely, and transfer-matrix and partition-function-root analyses have produced evidence - but no proof - of non-integrability. The surviving problem, reported as open, is to obtain the exact analytic solution of the hard-square lattice gas in the same sense as the hard-hexagon solution.

Background

The hard-square lattice gas is defined on the square lattice: each site is either occupied by a particle or empty, two nearest-neighbor sites may not both be occupied, and each particle carries a Boltzmann weight $z$, the activity (also called fugacity), so that the grand partition function sums $z^{\text{(number of particles)}}$ over all allowed configurations. In combinatorics the same object is the generating function of independent sets of the square grid (the hard-core model), and in physics it is the standard lattice model of a monolayer of atoms adsorbed on a surface with a commensurate half-filled crystalline phase.

As $z$ increases the model undergoes an ordering transition: at low activity the system is a disordered fluid, while at high activity the particles condense onto one of the two checkerboard sublattices (all sites with $x+y$ even, or all with $x+y$ odd), breaking a two-fold symmetry. The transition is continuous and occurs at $z_c=3.79625517391234(4)$ with critical density $\rho_c\approx0.368$ (Gaunt and Fisher, 1965, derived the first series; Baxter, Enting and Tsang, 1980, obtained $z_c=3.7962(1)$ from long series). The critical behavior is in the two-dimensional Ising universality class: transfer-matrix finite-size scaling (Guo and Blöte, 2002) reproduced the exact Ising correction-to-scaling spectrum and universal amplitudes to margins below $10^{-11}$, and Monte Carlo simulations (Fernandes, Arenzon and Levin, 2007) agree; the specific heat therefore diverges only logarithmically (exponent $\alpha=0$), and the order parameter vanishes with $\beta=1/8$.

The source of this problem, Baram and Fixman (1994), analyzed the fluid branch of the same model and concluded that it terminates at the same activity as the solid branch, giving a second-order transition at common $z_c$. They listed as unresolved issues the solid-side specific-heat exponent (their ratio estimate $\alpha'=0.139(50)$ was inconsistent with both a logarithmic divergence and the value $1/8$ suggested by assuming an integer $\delta$), the limited radius of convergence of the virial series (dominated by nonphysical complex singularities closer to the origin than $\rho_c$), the then-unresolved universality class (the two-fold order-parameter symmetry suggested Ising, while their exponent estimates hinted at three-state Potts behavior), and - the deepest issue - the absence of any analytic solution of the hard-square lattice gas analogous to the hard-hexagon solution.

The hard-hexagon model (the same nearest-neighbor-exclusion gas on the triangular lattice) was solved exactly by Baxter in 1980: the star-triangle relation (a functional equation on Boltzmann weights that makes row-to-row transfer matrices commute) reduces the free energy to the largest eigenvalue of a one-dimensional problem, solved through functional equations and Rogers-Ramanujan-type identities, yielding closed-form free energies, an algebraic critical activity $(11+5\sqrt5)/2$, and exact exponents. This remains the only exactly solved nearest-neighbor-exclusion lattice gas. For hard squares every analogous route has failed: vertex-model reformulations (Wood and Goldfinch, 1980) did not lead to a solution; Baxter's corner transfer matrix method (a factorization of the transfer matrix into four quadrant operators enabling extremely accurate numerics) evaluates the partition function per site at $z=1$ to 43 decimal places but produces no closed form; and the integrable manifold that Baxter's star-triangle solutions of the chiral Potts model (a solvable two-parameter generalization of the Potts spin model) provide for hard squares with additional diagonal interactions does not pass through the uncoupled point of the plain model.

On the mathematical side, the transition itself is not rigorously pinned down: uniqueness of the infinite-volume Gibbs measure is proven only for $z<2.538$ and phase coexistence (long-range sublattice order) only for $z>5.3506$, bracketing - but not reaching - the physical $z_c\approx3.796$; the Ising universality and the logarithmic specific heat remain numerical facts; and even the arithmetic nature of the hard-square entropy constant $\kappa(1)\approx1.5030480824$ (algebraic or transcendental) is unknown. The low-density activity series, now known to 92 terms, is dominated by a nonphysical singularity on the negative activity axis at $z_d=-0.1193388818(6)$ with exponent $\theta=5/6$ (Dhar, 1983, via a mapping to directed-site lattice animals, i.e. connected clusters of sites built up by directed steps), which blocks direct series access to $z_c$. Systematic comparisons of partition-function roots and transfer-matrix spectra (Assis, Jacobsen, Jensen, Maillard and McCoy, 2014) found qualitative differences from every integrable model studied - suggesting, but not proving, that hard squares is genuinely non-integrable.

Problem Statement

Obtain the exact analytic solution of the two-dimensional hard-square lattice gas (nearest-neighbor exclusion on the square lattice at activity $z$, with no additional interactions), in the same sense in which Baxter solved the hard-hexagon model: closed-form expressions for the partition function per site, equivalently the free energy per site, in both the fluid phase and the sublattice-ordered solid phase; the equation of state and the order parameter as functions of $z$; and the exact critical activity $z_c$, critical density $\rho_c$, and critical exponents. Any exact solution must reproduce the pinned numerical targets: $z_c=3.79625517391234(4)$, $\rho_c\approx0.368$, Ising critical behavior with a logarithmically divergent specific heat ($\alpha=0$, $\beta=1/8$, $\nu=1$), and the published low- and high-density series coefficients.

References

1. A. Baram and M. Fixman, "Hard square lattice gas", J. Chem. Phys. 101, 3172-3178 (1994). DOI: 10.1063/1.467564, https://doi.org/10.1063/1.467564
2. D. S. Gaunt and M. E. Fisher, "Hard-Sphere Lattice Gases. I. Plane-Square Lattice", J. Chem. Phys. 43, 2840 (1965). DOI: 10.1063/1.1697217, https://doi.org/10.1063/1.1697217
3. R. J. Baxter, I. G. Enting and S. K. Tsang, "Hard-square lattice gas", J. Stat. Phys. 22, 465-489 (1980). DOI: 10.1007/BF01012867, https://doi.org/10.1007/BF01012867
4. R. J. Baxter, "Hard hexagons: exact solution", J. Phys. A: Math. Gen. 13, L61-L70 (1980). DOI: 10.1088/0305-4470/13/3/007, https://doi.org/10.1088/0305-4470/13/3/007
5. D. W. Wood and M. Goldfinch, "Vertex models for the hard-square and hard-hexagon gases, and critical parameters from the scaling transformation", J. Phys. A: Math. Gen. 13, 2781-2794 (1980). DOI: 10.1088/0305-4470/13/8/026, https://doi.org/10.1088/0305-4470/13/8/026
6. D. A. Huse, "Tricriticality of Interacting Hard Squares: Some Exact Results", Phys. Rev. Lett. 49, 1121-1124 (1982). DOI: 10.1103/PhysRevLett.49.1121, https://doi.org/10.1103/PhysRevLett.49.1121
7. D. Dhar, "Exact solution of a directed-site animals-enumeration problem in three dimensions", Phys. Rev. Lett. 51, 853-856 (1983). DOI: 10.1103/PhysRevLett.51.853, https://doi.org/10.1103/PhysRevLett.51.853
8. R. J. Baxter, "Planar lattice gases with nearest-neighbor exclusion", Ann. Comb. 3, 191-203 (1999). DOI: 10.1007/BF01608783, arXiv:cond-mat/9811264, https://arxiv.org/abs/cond-mat/9811264
9. W. Guo and H. W. J. Blöte, "Finite-size analysis of the hard-square lattice gas", Phys. Rev. E 66, 046140 (2002). DOI: 10.1103/PhysRevE.66.046140, https://doi.org/10.1103/PhysRevE.66.046140
10. H. C. M. Fernandes, J. J. Arenzon and Y. Levin, "Monte Carlo simulations of two-dimensional hard core lattice gases", J. Chem. Phys. 126, 114508 (2007). DOI: 10.1063/1.2539141, arXiv:cond-mat/0612372, https://arxiv.org/abs/cond-mat/0612372
11. Y.-b. Chan, "Series expansions from the corner transfer matrix renormalization group method: the hard squares model", J. Phys. A: Math. Theor. 45, 085001 (2012). DOI: 10.1088/1751-8113/45/8/085001, arXiv:1107.1677, https://arxiv.org/abs/1107.1677
12. I. Jensen, "Comment on 'Series expansions from the corner transfer matrix renormalization group method: the hard-squares model'", J. Phys. A: Math. Theor. 45, 508001 (2012). DOI: 10.1088/1751-8113/45/50/508001, arXiv:1208.1060, https://arxiv.org/abs/1208.1060
13. Y.-b. Chan, "Series expansions from the corner transfer matrix renormalization group method: II. Asymmetry and high-density hard squares", J. Phys. A: Math. Theor. 46, 125009 (2013). DOI: 10.1088/1751-8113/46/12/125009, https://doi.org/10.1088/1751-8113/46/12/125009
14. M. Assis, J. L. Jacobsen, I. Jensen, J.-M. Maillard and B. M. McCoy, "Integrability versus non-integrability: hard hexagons and hard squares compared", J. Phys. A: Math. Theor. 47, 445001 (2014). DOI: 10.1088/1751-8113/47/44/445001, arXiv:1406.5566, https://arxiv.org/abs/1406.5566
15. J. C. Vera, E. Vigoda and L. Yang, "Improved Bounds on the Phase Transition for the Hard-Core Model in 2 Dimensions", SIAM J. Discrete Math. 29, 1895-1915 (2015). DOI: 10.1137/140976923, arXiv:1306.0431, https://arxiv.org/abs/1306.0431
16. A. Blanca, Y. Chen, D. Galvin, D. Randall and P. Tetali, "Phase Coexistence for the Hard-Core Model on Z^2", Combin. Probab. Comput. 28, 1-22 (2019). DOI: 10.1017/S0963548318000238, arXiv:1611.01115, https://arxiv.org/abs/1611.01115
17. A. Sinclair, P. Srivastava, D. Štefankovič and Y. Yin, "Spatial mixing and the connective constant: optimal bounds", Probab. Theory Relat. Fields 168, 153-197 (2017). DOI: 10.1007/s00440-016-0708-2, https://doi.org/10.1007/s00440-016-0708-2
18. S. Nyckees and F. Mila, "Tensor network investigation of the hard-square model", Phys. Rev. B 106, 174433 (2022). DOI: 10.1103/PhysRevB.106.174433, arXiv:2205.02508, https://arxiv.org/abs/2205.02508
19. Y. Svoray, "The Algebraicity Problem for Hard-Core Entropy Constants on the Discrete Hypertori", arXiv:2608.28332 (2026), https://arxiv.org/abs/2608.28332