Controlled closure of the Yvon-Born-Green hierarchy for dense liquids with realistic pair potentials (surviving core of the Born-Green 1946 kinetic-theory program)

The 1946 paper by Born and Green founded a kinetic theory of liquids on an exact hierarchy of equations for the molecular distribution functions, and explicitly deferred its quantitative completion: numerical solution of its integral equation for the radial distribution function with realistic intermolecular potentials, quantitative assessment of the Kirkwood superposition approximation that closes the hierarchy at the triplet level, solution of the linearized dynamical equations for the transport coefficients, and extension to mixtures and external fields. Later literature discharged most of these deferrals: the integral equation was integrated numerically for hard spheres and the Lennard-Jones potential (1950-1952 and after) and compared with liquid-argon data; the superposition approximation was assessed against Monte Carlo data (exact at low density, unreliable at small separations and at liquid densities); rigorous viscosity and thermal-conduction expressions were derived within the series itself, and transport coefficients are now computed via time-correlation formulas with molecular dynamics; mixtures and external force fields were treated from 1949. What survives is the closure problem: for a dense liquid with a realistic pair potential, no exact or systematically convergent closure of the hierarchy is known - the bridge function that corrects the superposition approximation is known exactly only from simulation, and all practical closures are uncontrolled and mutually inconsistent across thermodynamic routes. The open target is a closure or solution scheme whose accuracy is controlled across the dense-liquid domain, from the pair potential alone, delivering the structure and the equation of state of the liquid.

Background

By the mid-1940s the statistical theory of matter had well-developed limits at either end: dilute gases were described by Boltzmann's kinetic equation (whose Chapman-Enskog solution - a systematic expansion in the ratio of the mean free path to the macroscopic gradient scale - yields the transport coefficients), and solids by lattice dynamics. The liquid state - dense, but without long-range order - had neither. Born and Green (1946) proposed to fill this gap with a kinetic theory built on a hierarchy of molecular distribution functions: $f_h$ is the probability density for $h$ specified molecules to sit at given positions and velocities, and $n_h$ is its integral over velocities. Each $f_h$ obeys an exact continuity equation, and Newton's equations with pairwise-additive forces convert these into a coupled chain in which the evolution of $f_h$ depends on $f_{h+1}$ - the equilibrium version of what is now called the Yvon-Born-Green (YBG) hierarchy (its dynamical variant is known as the Bogoliubov-Born-Green-Kirkwood-Yvon (BBGKY) hierarchy in kinetic theory). Boltzmann's equation follows in the dilute limit, and in equilibrium the hierarchy reduces to exact relations among the $n_h$; in particular it determines the radial distribution function $g(r)=n_2/n_1^2$ (the pair correlation function, measurable by X-ray and neutron scattering), from which the internal energy and, via the virial theorem, the equation of state of the liquid follow.

The hierarchy cannot be solved as posed: each level involves the next, and a macroscopic liquid has $N\sim 10^{23}$ molecules. Born and Green therefore closed it at the triplet level using the superposition approximation introduced by Kirkwood (1935), $n_3\approx n_2^{(12)}n_2^{(23)}n_2^{(31)}/n_1^3$, which approximates the triplet density as a product of pair densities. This turns the two-body equilibrium equation into a single nonlinear integral equation for $g(r)$ (their equation (5.7)), generalizing the hard-sphere equation of Kirkwood and Boggs (1942) to arbitrary pair potentials $\phi(r)$. The same triplet closure is what the paper's linearized dynamical equations - the first attempt to extend Chapman-Enskog-type transport theory toward the liquid state - require at the level of $f_3$.

The 1946 paper explicitly deferred the quantitative part of this program, and those deferrals were largely discharged over the following decades. Kirkwood's group integrated the integral equation numerically for rigid spheres and for the Lennard-Jones potential, comparing thermodynamic functions with liquid-argon experiment (Kirkwood, Maun and Alder 1950; Kirkwood, Lewinson and Alder 1952); McLellan (1952) solved it by density expansion; Broyles (1960) compared its solutions with Monte Carlo data. Monte Carlo simulation quantified the superposition approximation (Alder, Frankel and Lewinson 1955; Krumhansl and Wang 1972), establishing that it is exact at low density but unreliable at small separations and increasingly inaccurate at liquid densities. Transport theory settled on time-correlation (Green-Kubo) formulas - integral expressions giving each transport coefficient as the time integral of an equilibrium autocorrelation function of the associated microscopic current (Kubo 1957) - evaluated by molecular dynamics (Rahman 1964), and Yang (1949) extended the whole framework to liquid mixtures with external force fields.

What survived is the central difficulty of the program: closing the hierarchy with controlled accuracy. Later, more successful closures replaced rather than solved the superposition approximation: the Percus-Yevick (PY) and hypernetted-chain (HNC) equations, two classical approximate closures relating $g(r)$ to $\phi(r)$ that differ in how they discard many-body correlations (PY is exactly solvable only for hard spheres; Wertheim 1963), and interpolations between them. The exact closure is conventionally expressed through a bridge function $B(r)$, the irreducible triplet-and-higher contribution that the superposition approximation discards; today the exact $B(r)$ for Lennard-Jones-type fluids is known only from computer simulation (Kast and Tomazic 2012, who proved an exact bound on it), different closures give mutually inconsistent results through different thermodynamic routes (Tschopp, Vahid and Brader 2026), and new closure schemes remain an active research topic (Attard 1990; Baalrud and Daligault 2019).

Problem Statement

The problem, as it survives from the open questions left by Born and Green (1946), is the quantitative completion of their kinetic theory of the liquid state: to determine, from the intermolecular pair potential $\phi(r)$ alone and with controlled accuracy, the structure and thermodynamics of a dense simple liquid. Its modern form is the closure problem of the Yvon-Born-Green hierarchy. The 1946 objective covered 'all the equilibrium and dynamical properties of liquids', but the dynamical part of the program - the transport coefficients from the linearized equations of the hierarchy - was pursued within the series itself and later superseded by the Green-Kubo time-correlation route, so the surviving open target is the equilibrium closure problem.

Equilibrium closure with controlled accuracy. For a three-dimensional monatomic fluid with a realistic pair potential (for instance the Lennard-Jones 12-6 potential $\phi(r)=4\epsilon[(\sigma/r)^{12}-(\sigma/r)^6]$, the standard model pair potential of simple fluids) at dense-liquid state points, construct either (a) an exact expression for the closure (bridge) function $B(r)$ - the irreducible three-and-more-body correlation term discarded by the superposition approximation $n_3\approx n_2^{(12)}n_2^{(23)}n_2^{(31)}/n_1^3$ - or (b) a truncation, expansion, or iterative scheme for the hierarchy whose accuracy is controlled and systematically improvable across the dense-liquid regime. The construction must deliver the radial distribution function $g(r)$ and the virial-route equation of state (the pressure computed from $g(r)$ via the virial theorem) to the claimed accuracy from $\phi(r)$ and fundamental constants alone, without using simulation or experimental data for these same observables in the construction, and must establish its range of validity across the dense-liquid regime rather than leaving it to case-by-case benchmarking.

References

1. M. Born and H. S. Green, 'A general kinetic theory of liquids I. The molecular distribution functions', Proceedings of the Royal Society of London A 188, 10-18 (1946). DOI: 10.1098/rspa.1946.0093, https://doi.org/10.1098/rspa.1946.0093
2. H. S. Green, 'A general kinetic theory of liquids II. Equilibrium properties', Proceedings of the Royal Society of London A 189, 103-117 (1947). DOI: 10.1098/rspa.1947.0031, https://doi.org/10.1098/rspa.1947.0031
3. M. Born and H. S. Green, 'A general kinetic theory of liquids III. Dynamical properties', Proceedings of the Royal Society of London A 190, 455-474 (1947). DOI: 10.1098/rspa.1947.0088, https://doi.org/10.1098/rspa.1947.0088
4. L. M. Yang, 'Kinetic theory of diffusion in gases and liquids II. General kinetic theory of liquids mixtures', Proceedings of the Royal Society of London A 198, 471-492 (1949). DOI: 10.1098/rspa.1949.0114, https://doi.org/10.1098/rspa.1949.0114
5. J. G. Kirkwood, 'Statistical Mechanics of Fluid Mixtures', The Journal of Chemical Physics 3, 300-313 (1935). DOI: 10.1063/1.1749657, https://doi.org/10.1063/1.1749657
6. J. G. Kirkwood and E. M. Boggs, 'The Radial Distribution Function in Liquids', The Journal of Chemical Physics 10, 394-402 (1942). DOI: 10.1063/1.1723737, https://doi.org/10.1063/1.1723737
7. J. G. Kirkwood, E. K. Maun and B. J. Alder, 'Radial Distribution Functions and the Equation of State of a Fluid Composed of Rigid Spherical Molecules', The Journal of Chemical Physics 18, 1040-1047 (1950). DOI: 10.1063/1.1747854, https://doi.org/10.1063/1.1747854
8. J. G. Kirkwood, V. A. Lewinson and B. J. Alder, 'Radial Distribution Functions and the Equation of State of Fluids Composed of Molecules Interacting According to the Lennard-Jones Potential', The Journal of Chemical Physics 20, 929-938 (1952). DOI: 10.1063/1.1700653, https://doi.org/10.1063/1.1700653
9. B. R. A. Nijboer and L. Van Hove, 'Radial Distribution Function of a Gas of Hard Spheres and the Superposition Approximation', Physical Review 85, 777-783 (1952). DOI: 10.1103/PhysRev.85.777, https://doi.org/10.1103/PhysRev.85.777
10. A. G. McLellan, 'A new method of solving the Born-Green equation for the radial distribution function', Proceedings of the Royal Society of London A 210, 509-517 (1952). DOI: 10.1098/rspa.1952.0016, https://doi.org/10.1098/rspa.1952.0016
11. B. J. Alder, S. P. Frankel and V. A. Lewinson, 'Radial Distribution Function Calculated by the Monte-Carlo Method for a Hard Sphere Fluid', The Journal of Chemical Physics 23, 417-419 (1955). DOI: 10.1063/1.1742004, https://doi.org/10.1063/1.1742004
12. A. A. Broyles, 'Radial Distribution Functions from the Born-Green Integral Equation', The Journal of Chemical Physics 33, 456-458 (1960). DOI: 10.1063/1.1731166, https://doi.org/10.1063/1.1731166
13. M. S. Wertheim, 'Exact Solution of the Percus-Yevick Integral Equation for Hard Spheres', Physical Review Letters 10, 321-323 (1963). DOI: 10.1103/PhysRevLett.10.321, https://doi.org/10.1103/PhysRevLett.10.321
14. R. Kubo, 'Statistical-Mechanical Theory of Irreversible Processes. I. General Theory and Simple Applications to Magnetic and Conduction Problems', Journal of the Physical Society of Japan 12, 570-586 (1957). DOI: 10.1143/jpsj.12.570, https://doi.org/10.1143/jpsj.12.570
15. A. Rahman, 'Correlations in the Motion of Atoms in Liquid Argon', Physical Review 136, A405-A411 (1964). DOI: 10.1103/PhysRev.136.A405, https://doi.org/10.1103/PhysRev.136.A405
16. J. A. Krumhansl and S.-s. Wang, 'Triplet Correlation in Liquid Argon by Monte Carlo Method: Low Densities', The Journal of Chemical Physics 56, 2034-2041 (1972). DOI: 10.1063/1.1677495, https://doi.org/10.1063/1.1677495
17. P. Attard, 'Integral equations and closure relations for the bridge function and for the triplet correlation function', The Journal of Chemical Physics 93, 7301-7309 (1990). DOI: 10.1063/1.459402, https://doi.org/10.1063/1.459402
18. H. M. Cho and J.-W. Chu, 'Inversion of radial distribution functions to pair forces by solving the Yvon-Born-Green equation iteratively', The Journal of Chemical Physics 131, 134107 (2009). DOI: 10.1063/1.3238547, https://doi.org/10.1063/1.3238547
19. S. M. Kast and D. Tomazic, 'Communication: An exact bound on the bridge function in integral equation theories', The Journal of Chemical Physics 137, 171102 (2012). DOI: 10.1063/1.4766465, https://doi.org/10.1063/1.4766465
20. S. D. Baalrud and J. Daligault, 'Mean force kinetic theory: a convergent kinetic theory for weakly and strongly coupled plasmas', Physics of Plasmas 26, 082106 (2019). DOI: 10.1063/1.5095655, https://doi.org/10.1063/1.5095655 (preprint arXiv:1904.09208, https://arxiv.org/abs/1904.09208)
21. S. M. Tschopp, H. Vahid, A. Sharma and J. M. Brader, 'Combining integral equation closures with force density functional theory for the study of inhomogeneous fluids', Soft Matter 21, 2633-2645 (2025). DOI: 10.1039/D4SM01262C, https://doi.org/10.1039/D4SM01262C (preprint arXiv:2410.18680, https://arxiv.org/abs/2410.18680)
22. S. M. Tschopp, H. Vahid and J. M. Brader, 'Routes to the density profile and structural inconsistency', The Journal of Physical Chemistry B 130, 1424-1436 (2026). DOI: 10.1021/acs.jpcb.5c07785, https://doi.org/10.1021/acs.jpcb.5c07785 (preprint arXiv:2511.10433, https://arxiv.org/abs/2511.10433)