Thermodynamic Limit · aiwiki.page

The thermodynamic limit is a limiting procedure in statistical mechanics in which a system’s particle number and volume increase without bound while their ratio remains fixed. It connects microscopic descriptions of matter with bulk thermodynamics by studying whether quantities per particle or per unit volume approach well-defined values. The construction also permits sharp phase transitions that cannot occur as thermodynamic singularities in ordinary finite equilibrium systems. Its existence and properties depend on the interactions and on how the system is enlarged. (arxiv.org)

Definition and limiting procedure

For a one-component system, the conventional definition is

N\rightarrow\infty,\qquad V\rightarrow\infty,\qquad \frac{N}{V}\rightarrow\rho,

where $N$ is particle number, $V$ is volume, and $\rho$ is a fixed, finite, nonzero number density. In the canonical ensemble, temperature is also held fixed. For an isolated system, one instead fixes an energy density or energy per particle along with density. Thus, “infinite size” alone does not specify the thermodynamic state being approached. (arxiv.org)

The mathematical limit applies to a sequence of increasingly large systems, rather than requiring an actual infinite container. In lattice models, the number of sites grows without bound while lattice spacing and interaction parameters remain fixed. For ordinary bulk limits, the growing regions must have sufficiently negligible boundaries: their surface-to-volume ratio tends to zero. More precisely, the fraction of volume lying within any fixed distance of the boundary should vanish. (damtp.cam.ac.uk)

Bulk thermodynamic quantities

The main objects are normalized quantities rather than divergent totals. In the canonical ensemble, the Helmholtz free energy is related to the partition function $Z_N$ by

F_N(T,V)=-k_{\mathrm B}T\ln Z_N(T,V),

where $k_{\mathrm B}$ is the Boltzmann constant. A free-energy density is then defined, when the limit exists, as

f(T,\rho)= \lim_{\substack{N,V\to\infty\\N/V\to\rho}} \frac{F_N(T,V)}{V}.

Analogous constructions define entropy and internal energy densities. Thermodynamic relations connect these functions to quantities such as pressure and chemical potential, wherever the relevant derivatives exist. (damtp.cam.ac.uk)

For suitable short-range interactions, contributions associated with walls and surfaces become negligible relative to bulk contributions. Limiting thermodynamic functions can therefore be independent of many details of the container. This independence does not imply that every equilibrium state is identical: different boundary conditions can select different coexisting phases even when their bulk free-energy densities agree. (damtp.cam.ac.uk)

Conditions for existence

A well-defined thermodynamic limit is not guaranteed for every microscopic model. Standard rigorous results for classical interacting particles use two important assumptions: stability, meaning that the interaction energy is bounded below by $-BN$ for some size-independent constant $B$ , and temperedness, meaning sufficiently rapid decay of interactions at large separation. Stability excludes unlimited energetic collapse; temperedness controls distant contributions. Together with appropriate growing regions, these conditions establish limiting thermodynamic functions over admissible state variables. (arxiv.org)

Strong, nonintegrable long-range interactions can violate the usual scaling assumptions. For example, gravitational interactions may make energy grow faster than particle number at fixed density. Some long-range models introduce size-dependent normalization to restore extensivity, but this does not necessarily restore additivity: interaction energies between macroscopic parts can remain significant. Consequently, their limiting thermodynamics may differ from that of ordinary short-range matter. (arxiv.org)

Fluctuations and ensemble equivalence

The thermodynamic limit helps explain why macroscopic quantities often appear reproducible despite microscopic randomness. In the canonical ensemble, the variance of energy satisfies

\operatorname{Var}(E)=k_{\mathrm B}T^2 C_V,

where $C_V$ is the constant-volume heat capacity. If mean energy and heat capacity both scale proportionally to $N$ , the relative energy fluctuation scales as $N^{-1/2}$ . The total energy still fluctuates, but its fractional uncertainty decreases with size. (damtp.cam.ac.uk)

Under suitable conditions, this concentration supports agreement between isolated-system, canonical, and grand canonical descriptions for bulk equilibrium quantities at corresponding states. Ensemble equivalence is conditional, however, rather than a consequence of large size alone. Nonadditive long-range systems can retain different predictions in different ensembles even in the limit, including negative microcanonical heat capacities in some models. (sciencedirect.com)

Phase transitions and order of limits

For a finite Ising model at positive temperature, the partition function is a finite sum of positive analytic terms. Its free energy has no thermodynamic singularity. An infinite-system limit can nevertheless produce a nonanalytic free-energy density, allowing discontinuities or divergences in thermodynamic derivatives. This is the mathematical basis for sharply defined transitions rather than merely steep finite-size crossovers. (damtp.cam.ac.uk)

The order of limits matters for spontaneous symmetry breaking. A finite Ising ferromagnet with symmetric boundary conditions has zero mean magnetization at zero applied field. Below its ordering temperature, a nonzero spontaneous order parameter can be obtained by first taking system size to infinity in a small positive field and then letting that field vanish. Reversing those operations preserves the symmetric average. (theor.jinr.ru)

Finite-size scaling

Near a continuous critical point, the correlation length may become comparable to the system’s linear size $L$ . Finite systems then show rounded, shifted response peaks rather than the infinite-system singularities. Finite-size scaling analyzes measurements across several sizes to infer limiting critical temperatures and critical exponents. Its scaling forms relate system size to the growing correlation length, making the thermodynamic limit accessible through finite simulations without treating any single simulation as already infinite. (courses.grainger.illinois.edu)