🎲Statistical Mechanics Unit 3 Review

Statistical mechanics describes the same physical system through different "lenses" called ensembles. Each ensemble holds different macroscopic quantities fixed (energy, temperature, particle number, etc.), yet for large systems in equilibrium, they all predict the same thermodynamic properties. This is ensemble equivalence, and it's what lets you pick whichever ensemble makes your calculation easiest without worrying about getting a different answer.

The three main ensembles are the microcanonical (fixed energy), canonical (fixed temperature), and grand canonical (fixed temperature, exchanging particles). Understanding why and when these give identical results is central to connecting microscopic statistics to the macroscopic thermodynamics you measure in a lab.

Definition of Statistical Ensembles

A statistical ensemble is the collection of all microstates a system can occupy, given some set of macroscopic constraints. Which constraints you impose defines which ensemble you're working in:

Microcanonical ensemble: fixes total energy $E$ , volume $V$ , and particle number $N$
Canonical ensemble: fixes temperature $T$ , volume $V$ , and particle number $N$ (system exchanges energy with a heat bath)
Grand canonical ensemble: fixes temperature $T$ , volume $V$ , and chemical potential $\mu$ (system exchanges both energy and particles with a reservoir)

Conditions for Ensemble Equivalence

Ensemble equivalence isn't automatic. It holds when all of the following are satisfied:

The system is in thermodynamic equilibrium (no net flows of energy or matter).
The system is large enough to approach the thermodynamic limit (see below).
Interactions between particles are short-range, so distant parts of the system are effectively independent.
Thermodynamic quantities like energy and entropy are extensive, meaning they scale linearly with system size.

If any of these conditions breaks down, the ensembles can give different predictions.

Thermodynamic Limit

The thermodynamic limit is the idealization where you take $N \to \infty$ and $V \to \infty$ while keeping the density $N/V$ constant. In this limit:

Surface effects become negligible compared to bulk contributions.
Finite-size fluctuations wash out.
Intensive variables (temperature, pressure, chemical potential) converge to sharp, well-defined values.

This is the regime where ensemble equivalence holds. Real systems aren't infinite, of course, but for something like a mole of gas ( $N \sim 10^{23}$ ), the thermodynamic limit is an excellent approximation.

Major Statistical Ensembles

Microcanonical Ensemble

The microcanonical ensemble describes a completely isolated system: no exchange of energy or particles with the surroundings. Its fundamental postulate is that all accessible microstates at a given energy are equally probable.

The entropy follows directly from counting microstates:

$S = k_B \ln \Omega$

where $\Omega$ is the number of microstates with energy $E$ . This is Boltzmann's formula, and it's the most fundamental definition of entropy in statistical mechanics. The microcanonical ensemble is conceptually clean but often harder to work with in practice, since fixing energy exactly is mathematically more restrictive than fixing temperature.

Canonical Ensemble

The canonical ensemble describes a system in thermal contact with a large heat bath at temperature $T$ . Energy flows freely between system and bath, but particle number and volume stay fixed.

The probability of finding the system in microstate $i$ with energy $E_i$ is given by the Boltzmann factor:

$p_i = \frac{1}{Z} e^{-E_i / k_B T}$

where $Z$ is the canonical partition function (defined below). This is the workhorse ensemble for most practical calculations because temperature is what you actually control in most experiments.

Grand Canonical Ensemble

The grand canonical ensemble describes an open system that exchanges both energy and particles with a reservoir at temperature $T$ and chemical potential $\mu$ . The probability of a microstate with energy $E_i$ and particle number $N_i$ is:

$p_i = \frac{1}{\Xi} e^{-(E_i - \mu N_i) / k_B T}$

This ensemble is especially useful for studying phase transitions, adsorption phenomena, and any situation where particle number fluctuates.

Mathematical Foundations

Partition Functions

Partition functions are the bridge between microscopic states and macroscopic thermodynamics. Once you have the partition function, you can extract all thermodynamic quantities through derivatives.

Canonical partition function: $Z = \sum_i e^{-E_i / k_B T}$
Grand canonical partition function: $\Xi = \sum_{N=0}^{\infty} \sum_i e^{-(E_i - \mu N) / k_B T}$

For example, the average energy in the canonical ensemble is $\langle E \rangle = -\frac{\partial \ln Z}{\partial \beta}$ where $\beta = 1/k_B T$ . The Helmholtz free energy is $F = -k_B T \ln Z$ . Similar derivative relations connect $\Xi$ to pressure, average particle number, and other grand canonical quantities.

Density of States

The density of states $g(E)$ counts how many microstates fall within a given energy interval. For continuous systems:

$g(E) = \frac{d\Omega}{dE}$

This function is the key link between the microcanonical and canonical descriptions. The canonical partition function can be written as an integral over the density of states:

$Z = \int g(E) \, e^{-E / k_B T} \, dE$

For large systems, $g(E)$ is typically a very sharply peaked function, which is why the saddle-point (steepest descent) approximation works so well and why ensemble equivalence emerges.

Entropy Formulations

Different entropy definitions correspond to different levels of generality:

Boltzmann entropy (microcanonical): $S = k_B \ln \Omega$
Gibbs entropy (general): $S = -k_B \sum_i p_i \ln p_i$
von Neumann entropy (quantum): $S = -k_B \, \text{Tr}(\rho \ln \rho)$

The Gibbs formula reduces to the Boltzmann formula when all accessible microstates are equally probable (the microcanonical case). In the thermodynamic limit, all three formulations agree for systems that satisfy ensemble equivalence.

Equivalence in the Thermodynamic Limit

Large System Behavior

The physical reason ensembles become equivalent for large systems is that fluctuations become negligible. In the canonical ensemble, the system's energy isn't fixed, but for large $N$ it barely deviates from its average value. The system effectively "picks" a single energy, just as the microcanonical ensemble prescribes.

Quantitatively, relative fluctuations in extensive quantities shrink as:

$\frac{\Delta E}{\langle E \rangle} \propto \frac{1}{\sqrt{N}}$

For $N \sim 10^{23}$ , that's a relative fluctuation of order $10^{-12}$ . The energy distribution becomes so narrow that the canonical and microcanonical ensembles are indistinguishable.

Fluctuations vs. System Size

This scaling follows from the central limit theorem applied to weakly correlated subsystems:

Fluctuations in extensive quantities ( $E$ , $N$ ) scale as $\sqrt{N}$
Relative fluctuations of intensive quantities ( $T$ , $P$ ) decrease as $1/\sqrt{N}$
Canonical energy fluctuations: $\Delta E / \langle E \rangle \propto 1/\sqrt{N}$
Grand canonical particle number fluctuations: $\Delta N / \langle N \rangle \propto 1/\sqrt{N}$

The $1/\sqrt{N}$ suppression is the quantitative backbone of ensemble equivalence.

Extensivity and Intensivity

For ensemble equivalence to work, the system's thermodynamic potentials must be extensive (scaling linearly with $N$ ). This means you can mentally divide the system into independent subsystems whose contributions simply add up.

Extensive properties: energy, entropy, free energy, particle number
Intensive properties: temperature, pressure, chemical potential

If extensivity fails (as it does for long-range interactions), the subsystems are no longer independent, and the ensembles can diverge.

Definition of statistical ensembles, physical chemistry - Difference between microcanonical and canonical ensemble - Chemistry Stack ...

Practical Applications

Computational Methods

Ensemble equivalence is heavily exploited in simulation:

Monte Carlo simulations sample configurations in whichever ensemble is most efficient, relying on equivalence to translate results.
Molecular dynamics naturally generates microcanonical trajectories, but thermostats (Nosé-Hoover, Langevin) convert these to canonical sampling.
Density functional theory calculations often use the grand canonical ensemble because the chemical potential is a natural control variable.
Replica exchange (parallel tempering) runs copies of the system at different temperatures and swaps configurations between them, directly leveraging canonical ensemble equivalence at different $T$ .

Experimental Relevance

In the lab, you rarely have a truly isolated system (microcanonical) or a system at fixed chemical potential (grand canonical). Yet ensemble equivalence lets you analyze data with whichever framework is most convenient:

Calorimetry data (measuring heat flow) maps naturally onto the canonical ensemble.
Gas adsorption experiments, where molecules move on and off a surface, are best interpreted through the grand canonical ensemble.
The key point: for macroscopic samples, the choice of ensemble is a matter of mathematical convenience, not physical reality.

System-Specific Considerations

Solids at fixed composition are well-described by the canonical ensemble.
Fluids and gases in contact with a particle reservoir call for the grand canonical ensemble.
Biological systems often operate at constant pressure and temperature, making the isothermal-isobaric (NPT) ensemble the natural choice.

The "right" ensemble is whichever one matches your experimental control parameters.

Limitations and Exceptions

Finite-Size Effects

For small systems (nanoparticles, single biomolecules, small atomic clusters), ensemble equivalence can break down noticeably:

Surface contributions are no longer negligible compared to bulk.
Energy levels are discrete enough that the continuous density-of-states approximation fails.
Corrections to thermodynamic-limit behavior typically scale as powers of $1/N$ or $1/L$ (where $L$ is the system's linear dimension).

Phase Transitions

Near phase transitions, ensemble equivalence requires extra care:

At critical points, correlation lengths diverge, violating the assumption that distant parts of the system are independent. Long-range correlations can make fluctuations anomalously large.
At first-order transitions, coexisting phases complicate the picture. The microcanonical ensemble can resolve phase coexistence that the canonical ensemble smooths over (the canonical ensemble enforces a Maxwell construction).
Finite-size scaling techniques are needed to extract reliable thermodynamic-limit behavior from simulations of finite systems near criticality.

Long-Range Interactions

Systems with interactions that decay slower than $r^{-d}$ (where $d$ is the spatial dimension) can violate ensemble equivalence even in the thermodynamic limit. Examples include:

Gravitational systems (stars in a galaxy, globular clusters)
Unscreened Coulomb systems (non-neutral plasmas)
Certain mean-field spin models (e.g., the Blume-Emery-Griffiths model)

A striking signature of inequivalence in these systems is negative heat capacity in the microcanonical ensemble: adding energy actually lowers the temperature. This is impossible in the canonical ensemble (where heat capacity is always positive by construction), so the two ensembles genuinely disagree.

Ensemble Inequivalence

Causes of Inequivalence

Ensemble inequivalence arises when the assumptions underlying equivalence fail. The main culprits are:

Non-additivity of energy or entropy (the system can't be cleanly divided into independent subsystems)
Long-range interactions or strong correlations that couple distant parts of the system
Phase transitions, especially first-order transitions where the entropy is non-concave as a function of energy
Finite-size effects in small systems

Non-Additive Systems

In a non-additive system, the total energy of two combined subsystems isn't simply the sum of their individual energies. Long-range interacting systems are the canonical example: every particle interacts appreciably with every other, so you can't treat subsystems as independent.

Consequences include:

The entropy $S(E)$ can be non-concave (it has convex regions), which is the mathematical hallmark of ensemble inequivalence.
The microcanonical ensemble can exhibit negative heat capacity, temperature jumps, and other features absent from the canonical description.
Specialized frameworks like Tsallis statistics or superstatistics are sometimes invoked, though their applicability remains debated.

Microcanonical vs. Canonical Differences

When inequivalence occurs, the two ensembles can predict qualitatively different physics:

The canonical ensemble always gives a concave free energy (by construction via the Legendre transform), which can erase features present in the microcanonical entropy.
A first-order transition in the canonical ensemble may appear as a continuous transition with a negative-heat-capacity region in the microcanonical ensemble.
The microcanonical phase diagram can be richer, containing phases and transitions that the canonical ensemble misses entirely.

These differences are not academic curiosities. They matter for gravitational systems, some nuclear physics problems, and certain lattice models studied in mathematical physics.

Advanced Topics

Tsallis Statistics

Tsallis statistics generalizes Boltzmann-Gibbs statistical mechanics for systems where standard extensivity assumptions fail. It replaces the exponential and logarithm with q-deformed versions:

The q-entropy is $S_q = k_B \frac{1 - \sum_i p_i^q}{q - 1}$
In the limit $q \to 1$ , this recovers the standard Gibbs entropy.
The formalism has been applied to long-range interacting systems, turbulent flows, and systems with fractal phase-space structure.

Whether Tsallis statistics represents a fundamental generalization or simply a convenient fitting tool is still a matter of active discussion.

Generalized Ensembles

Generalized ensemble methods are computational techniques designed to overcome sampling barriers in complex energy landscapes:

Multicanonical ensemble: reweights configurations to produce a flat energy histogram, enabling efficient sampling across energy barriers.
Wang-Landau sampling: iteratively estimates the density of states $g(E)$ directly, from which canonical properties at any temperature can be reconstructed.
Replica exchange (parallel tempering): runs multiple copies at different temperatures and swaps them, using ensemble equivalence to accelerate exploration of configuration space.

These methods don't change the physics but dramatically improve computational efficiency for systems like proteins, spin glasses, and polymers.

Quantum Ensemble Equivalence

Classical ensemble equivalence extends to quantum systems, but with additional structure:

Quantum statistics (Bose-Einstein for bosons, Fermi-Dirac for fermions) modify the counting of microstates and the form of partition functions.
The density matrix $\rho$ provides a unified description: $\rho = e^{-\beta H}/Z$ for the canonical ensemble, with the von Neumann entropy $S = -k_B \, \text{Tr}(\rho \ln \rho)$ as the natural entropy measure.
Quantum phase transitions (occurring at zero temperature as a parameter is varied) can exhibit ensemble-dependent behavior, particularly in finite-size systems or systems with topological order.

🎲Statistical Mechanics Unit 3 Review