Statistical Mechanics

Statistical mechanics is a branch of physics that connects the microscopic behavior of particles with the macroscopic properties of matter. It provides a microscopic foundation for thermodynamics, explaining quantities such as pressure, heat capacity, and entropy through mechanical laws and probability. Rather than tracking every constituent individually, it studies distributions over possible microscopic states. Its methods describe gases, liquids, solids, radiation, and many other systems with large numbers of interacting degrees of freedom. (damtp.cam.ac.uk)

Historical development

The subject developed from nineteenth-century attempts to explain heat and gas behavior through molecular motion. James Clerk Maxwell established a statistical description of molecular velocities, while Ludwig Boltzmann developed kinetic theory and connected entropy with the multiplicity of microscopic configurations. In 1877, Boltzmann formulated this statistical interpretation of entropy. Josiah Willard Gibbs subsequently organized statistical mechanics around ensembles; his Elementary Principles in Statistical Mechanics, published in 1902, provided a systematic framework for equilibrium theory. Quantum statistics later extended these methods to systems whose microscopic states obey quantum laws. (damtp.cam.ac.uk)

Microscopic states and macroscopic descriptions

A microstate specifies the microscopic condition of a system. In classical mechanics, this normally means the positions and momenta of its particles, represented by a point in phase space. In quantum mechanics, states and observables require a quantum description, with statistical predictions obtained from probabilities or density operators. A macrostate specifies a smaller collection of measurable quantities, such as volume, temperature, and total energy. Many different microstates can correspond to the same macrostate. (damtp.cam.ac.uk)

For equally probable accessible microstates, entropy is expressed by Boltzmann’s relation,

where $\Omega$ counts those states and $k_{\mathrm B}$ is the Boltzmann constant. For a discrete distribution with probabilities $p_i$, the Gibbs expression is

It reduces to Boltzmann’s expression when every accessible state has probability $1/\Omega$. Entropy therefore has a precise statistical meaning, rather than being merely a qualitative measure of “disorder.” (ocw.mit.edu)

Equilibrium ensembles

A statistical ensemble is a probability distribution over hypothetical copies of a system subject to specified macroscopic constraints. Three ensembles are especially important:

• Microcanonical ensemble: describes an isolated system with fixed energy, volume, and particle number. Its standard equilibrium formulation assigns equal probabilities to accessible states within a specified energy interval.
• Canonical ensemble: describes a system exchanging energy with a heat reservoir, while its volume and particle number remain fixed.
• Grand canonical ensemble: allows both energy and particles to be exchanged with a reservoir, at specified temperature, volume, and chemical potential. (damtp.cam.ac.uk)

These ensembles use different probability distributions, but often yield the same bulk thermodynamic predictions in the thermodynamic limit: particle number and volume become large at fixed density. This equivalence requires suitable conditions and is not universal, particularly for systems with long-range interactions. (damtp.cam.ac.uk)

Partition functions and observables

In the canonical ensemble, a microstate of energy $E_i$ has probability

The partition function $Z$ normalizes the probabilities and encodes equilibrium thermodynamic information. The expected value of an observable $A$, with microstate values $A_i$, is $\langle A\rangle=\sum_i p_iA_i$. Classical formulations replace state sums with appropriately normalized phase-space integrals. (damtp.cam.ac.uk)

For a temperature-independent energy spectrum, useful relations include

where $U$ is mean internal energy and $F$ is Helmholtz free energy. Energy fluctuations connect microscopic statistics to measurable response:

Thus the variance of energy determines the constant-volume heat capacity $C_V$. (damtp.cam.ac.uk)

Classical and quantum statistics

For a dilute classical gas with negligible interactions, statistical mechanics gives the equation of state $PV=Nk_{\mathrm B}T$. It also produces the Maxwell–Boltzmann velocity distribution. Interparticle forces explain deviations from ideal-gas behavior, while classical equipartition assigns an average energy $k_{\mathrm B}T/2$ to each independent quadratic term in the energy. Quantum effects limit the validity of this last result. (damtp.cam.ac.uk)

Identical quantum particles require different counting rules. Bosons obey Bose–Einstein statistics, allowing multiple particles to occupy the same single-particle state. Fermions obey Fermi–Dirac statistics and the Pauli exclusion principle. These distinctions explain phenomena including electron degeneracy, blackbody radiation, and the formation of a Bose–Einstein condensate. Both quantum distributions approach classical Maxwell–Boltzmann behavior in the dilute, nondegenerate limit. (damtp.cam.ac.uk)

Collective behavior and phase transitions

Interactions can generate collective properties absent from individual particles. Statistical mechanics describes phase transitions, including condensation and magnetic ordering, by examining equilibrium states and their free energies. The Ising model, consisting of interacting discrete spins, provides a tractable setting for studying magnetic order and critical behavior. An order parameter distinguishes phases; magnetization, for example, characterizes ferromagnetic ordering. (damtp.cam.ac.uk)

Near continuous transitions, fluctuations occur across increasingly large distances. Renormalization-group methods, closely connected with renormalization, explain how systems with different microscopic details can share critical exponents and scaling behavior. Such universality organizes transitions into classes determined by features including dimensionality, symmetry, and interaction range. (ocw.mit.edu)

Nonequilibrium and computational methods

Nonequilibrium statistical mechanics studies relaxation, transport, and driven systems. Brownian motion and diffusion connect microscopic fluctuations with observable motion and spreading. Kinetic equations, stochastic processes, and response theory describe aspects of time-dependent behavior that equilibrium ensembles alone do not determine. (ocw.mit.edu)

Exact calculations are usually restricted to relatively simple models. Molecular dynamics numerically follows particle motion, while Markov chain Monte Carlo samples statistical distributions without enumerating every state. These complementary approaches investigate interacting particles, material properties, and phase behavior when analytic state sums or integrals are impractical. (arxiv.org)