Residual entropy of ordinary ice: the exact ice-rule configuration count and a precision experimental test of Pauling's estimate

Ordinary hexagonal ice (ice Ih) retains a residual entropy because a macroscopic number of proton arrangements satisfy the Bernal–Fowler ice rules. Pauling's 1935 estimate gives $S_0=R\ln(3/2)=0.8057$ cal/(mol·K); Nagle's 1966 series expansion gives the configuration count per molecule $W_1=1.50685(15)$; multicanonical Monte Carlo simulations by Berg, Muguruma and Okamoto (2007; extended 2012) gave $W_1\approx1.5071$–$1.5074$ and left open the experimental confirmation, noting a small numerical discrepancy with Nagle's series value that they did not consider worth tracing further. Later independent methods — thermodynamic integration, Wang–Landau sampling, and tensor networks — converge on $W_1\approx1.5074$–$1.5075$ ($S_0\approx0.8155$ cal/(mol·K)), and the 2012 discrepancy has been attributed to a random-number-generator bias, but no exact solution of the three-dimensional ice-rule counting problem exists, the equality of the residual entropies of hexagonal and cubic ice is unproven, the anomalous finite-size exponent $\theta\neq1$ is unexplained, and the experimental value still rests on the 1936 measurement $0.82\pm0.05$ cal/(mol·K), which cannot distinguish Pauling's estimate from the corrected value. The problem is to determine the residual entropy of ordinary ice — exactly, or with certified uncertainty — and to measure it experimentally with an order of magnitude better precision, so that the correction to Pauling's ninety-year-old estimate is finally identified.

Background

Ordinary ice (ice Ih, the hexagonal polymorph) is a proton-disordered crystal: the oxygen atoms form a rigid hexagonal lattice in which every oxygen has four tetrahedral nearest neighbours, while the hydrogen nuclei are arranged according to the Bernal–Fowler ice rules: (A) exactly one hydrogen lies on each oxygen–oxygen bond, and (B) exactly two hydrogens are close to each oxygen, so that every molecule is a neutral H$_2$O. Because a macroscopic number of proton arrangements satisfy these purely local rules, ice retains a nonzero 'residual' (zero-temperature) entropy $S_0=k_B\ln W$, where $W$ is the number of allowed configurations for $N$ molecules, conventionally written $W=(W_1)^N$ with $W_1$ the configuration count per molecule. Calorimetry by Giauque and Stout in 1936 gave $S_0=0.82\pm0.05$ cal/(mol·K), and the last careful reanalysis (Haida et al., 1974, studying the enthalpy relaxation associated with proton ordering near 85–110 K) kept that value and error bar. Pauling (1935) estimated $W_1=3/2$, i.e. $S_0=0.8057$ cal/(mol·K), by treating the molecules as independent; a lower-bound argument due to Onsager and Dupuis (documented in Nagle's 1966 paper) shows $W_1\geq3/2$ once closed loops of the lattice are accounted for. Nagle (1966) refined the count with a series expansion, $W_1=1.50685(15)$, for both hexagonal (Ih) and cubic (Ic) ice. In two dimensions the analogous problem is exactly solvable — Lieb's exact solution for square ice gives $W_1=(4/3)^{3/2}=1.5396$ — but no exact result exists for any realistic three-dimensional ice lattice.

Berg, Muguruma and Okamoto (Phys. Rev. B 75, 092202 (2007); extended in Mol. Simul. 38, 856 (2012)) attacked the 3D problem with multicanonical Monte Carlo (MUCA) simulations. A multicanonical simulation is a biased Markov-chain Monte Carlo method whose weights are tuned so that the energy histogram becomes flat, allowing the density of states — and hence the ground-state degeneracy $W$ — to be estimated with its absolute normalization, because the number of configurations at infinite temperature ($6^N$ or $2^{2N}$ for their two lattice models) is known exactly. Their estimates, $W_1=1.50738(16)$ (2007) and $W_1=1.507117(35)$ (2012, lattices up to $N=2880$), lie slightly above Nagle's series value; the 2007 comparison showed a small, statistically significant discrepancy (Gaussian difference test $Q=0.016$, i.e. two estimates with the quoted error bars would differ by this much or more with probability only 1.6%) which the authors suggested might be explained by finite-size corrections from larger lattices but considered 'somewhat academic' to trace further. Their finite-size extrapolation used $W_1(N)=W_1+a\,N^{-\theta}$ with $\theta\approx0.92\neq1$, reported with the interpretive remark that $\theta\neq1$ reflects long-range correlations between hydrogen bonds in the ground-state ensemble. Their papers closed with a challenge and an open test: (i) the experimental error bar, unchanged since 1936, is too large to distinguish Pauling's $0.8057$ from the corrected $\approx0.815$ cal/(mol·K), and an improvement by roughly one order of magnitude would unambiguously identify the correction to Pauling's value; (ii) transfer of the method to other crystal structures (other ice forms, geometrically frustrated magnets) was claimed but untested.

The subsequent numerical landscape consolidated: thermodynamic integration, i.e. computing the entropy change along a reversible path from a reference state of known entropy (Herrero and Ramírez), Wang–Landau sampling (a flat-histogram Monte Carlo variant) of an effective Ising representation (Li, Zhao and Yang), and tensor-network methods — projected entangled-pair states that encode the ice rule in local tensors and yield the entropy directly in the thermodynamic limit (Vanderstraeten et al.; Xu et al.) — all give $W_1\approx1.5074$–$1.5075$, i.e. $S_0\approx0.8155$ cal/(mol·K)$=3.41$ J/(mol·K), above Nagle's value at the level of the best error bars. The same body of work established that the ice ground state is a Coulomb phase, meaning the hydrogen-bond occupation variables behave like a divergence-free field whose correlations decay dipolarly as $1/r^3$ — the modern description of the long-range bond correlations. A caveat for the experimental side: the proton configurations are not exactly degenerate in real ice, and annealing (partial proton ordering) intervenes between roughly 85 and 110 K, so a residual-entropy measurement must either outrun or correctly account for that ordering.

Problem Statement

Establish the residual entropy of ordinary ice at the precision level at which the correction to Pauling's estimate becomes unambiguous. The objective is a single quantity — the residual entropy $S_0$ of hexagonal ice Ih, or equivalently the ice-rule configuration count per molecule $W_1$ with $S_0=R\ln W_1$ — approached from two linked directions:

1. (Exact theory) Determine the exact value of $W_1$ for the three-dimensional ice-rule ensemble on the ice Ih lattice, or, short of an exact solution, bound it with certified (non-statistical) error bars, settling whether the true value coincides with Nagle's series estimate $1.50685(15)$ or with the modern numerical consensus $\approx1.5074$–$1.5075$, which differ at the level of the best published error bars.

2. (Experiment) Provide a calorimetric, or equivalent thermodynamic, determination of the residual entropy of real ice Ih whose total uncertainty is roughly one order of magnitude smaller than the 1936 error bar — $\sigma\lesssim0.005$ cal/(mol·K) — sufficient to distinguish Pauling's $0.8057$ from the corrected $\approx0.8155$ cal/(mol·K) and thereby test the ice-rule prediction for a real crystal, correctly accounting for the partial proton ordering (annealing) that occurs between roughly 85 and 110 K.

The objective combines a theoretical determination of $W_1$ — exact, or with a certified bound — with an experimental determination on real ice that distinguishes Pauling's estimate from the corrected value. The opening paragraph's two further observations — that the equality of the residual entropies of hexagonal and cubic ice is unproven, and that the anomalous finite-size exponent $\theta\neq1$ is unexplained — are background context, not part of the objective.

References

1. Linus Pauling, 'The Structure and Entropy of Ice and of Other Crystals with Some Randomness of Atomic Arrangement', Journal of the American Chemical Society 57, 2680-2684 (1935). doi:10.1021/ja01315a102, https://doi.org/10.1021/ja01315a102
2. W. F. Giauque and J. W. Stout, 'The Entropy of Water and the Third Law of Thermodynamics. The Heat Capacity of Ice from 15 to 273 K', Journal of the American Chemical Society 58, 1144-1150 (1936). doi:10.1021/ja01298a023, https://doi.org/10.1021/ja01298a023
3. O. Haida, T. Matsuo, H. Suga and S. Seki, 'Calorimetric study of the glassy state X. Enthalpy relaxation at the glass-transition temperature of hexagonal ice', Journal of Chemical Thermodynamics 6, 815-825 (1974). doi:10.1016/0021-9614(74)90227-4, https://doi.org/10.1016/0021-9614(74)90227-4
4. J. F. Nagle, 'Lattice Statistics of Hydrogen Bonded Crystals. I. The Residual Entropy of Ice', Journal of Mathematical Physics 7, 1484-1491 (1966). doi:10.1063/1.1705058, https://doi.org/10.1063/1.1705058
5. E. H. Lieb, 'Residual Entropy of Square Ice', Physical Review 162, 162-172 (1967). doi:10.1103/PhysRev.162.162, https://doi.org/10.1103/PhysRev.162.162
6. B. A. Berg, C. Muguruma and Y. Okamoto, 'Residual entropy of ordinary ice from multicanonical simulations', Physical Review B 75, 092202 (2007). doi:10.1103/PhysRevB.75.092202, https://arxiv.org/abs/cond-mat/0609211
7. B. A. Berg, C. Muguruma and Y. Okamoto, 'Residual entropy of ordinary ice calculated from multicanonical Monte Carlo simulations', Molecular Simulation 38, 856-860 (2012). doi:10.1080/08927022.2011.651140, https://doi.org/10.1080/08927022.2011.651140
8. C. P. Herrero and R. Ramirez, 'Configurational entropy of ice from thermodynamic integration', Chemical Physics Letters 568-569, 70-74 (2013). doi:10.1016/j.cplett.2013.03.010, https://arxiv.org/abs/1307.3950
9. C. P. Herrero and R. Ramirez, 'Configurational entropy of hydrogen-disordered ice polymorphs', Journal of Chemical Physics 140, 234502 (2014). doi:10.1063/1.4882650, https://arxiv.org/abs/1406.5929
10. J. Kolafa, 'Residual entropy of ices and clathrates from Monte Carlo simulation', Journal of Chemical Physics 140, 204507 (2014). doi:10.1063/1.4879061, https://doi.org/10.1063/1.4879061
11. J. Kolafa, 'Residual entropy of ice III from Monte Carlo simulation', Journal of Chemical Physics 144, 124509 (2016). doi:10.1063/1.4944612, https://doi.org/10.1063/1.4944612
12. M. V. Ferreyra and S. A. Grigera, 'Boundary conditions and the residual entropy of ice systems', Physical Review E 98, 042146 (2018). doi:10.1103/PhysRevE.98.042146, https://doi.org/10.1103/PhysRevE.98.042146
13. L. Vanderstraeten, B. Vanhecke and F. Verstraete, 'Residual entropies for three-dimensional frustrated spin systems with tensor networks', Physical Review E 98, 042145 (2018). doi:10.1103/PhysRevE.98.042145, https://arxiv.org/abs/1805.10598
14. T. Hayashi, C. Muguruma and Y. Okamoto, 'Calculation of the residual entropy of Ice Ih by Monte Carlo simulation with the combination of the replica-exchange Wang-Landau algorithm and multicanonical replica-exchange method', Journal of Chemical Physics 154, 044503 (2021). doi:10.1063/5.0038157, https://arxiv.org/abs/2009.11591
15. D.-Z. Li, Y.-J. Zhao and X.-B. Yang, 'Residual entropy of ice Ih by Wang-Landau Monte Carlo simulation of an effective Ising model', Journal of Statistical Mechanics: Theory and Experiment 2022, 103203 (2022). doi:10.1088/1742-5468/ac9470, https://doi.org/10.1088/1742-5468/ac9470
16. D.-Z. Li, W.-J. Huang, Y. Yao and X.-B. Yang, 'Exact results for the residual entropy of ice hexagonal monolayer', Physical Review E 107, 054121 (2023). doi:10.1103/PhysRevE.107.054121, https://arxiv.org/abs/2211.14679
17. D.-Z. Li, Y. Cen, X. Wang and X.-B. Yang, 'Residual entropy of ice: A study based on transfer matrices', Physical Review B 110, 195414 (2024). doi:10.1103/PhysRevB.110.195414, https://arxiv.org/abs/2404.13897
18. X.-Z. Xu, T.-Y. Lin and G.-M. Zhang, 'Equivalence of residual entropy of hexagonal and cubic ices from tensor network methods', Physical Review B 113, 214416 (2026). doi:10.1103/3myh-6pdf, https://arxiv.org/abs/2511.22477