A topic in the Open Knowledge Graph — a free, open map of 15,290 topics and the order to learn them in.
The canonical ensemble describes a system in thermal contact with a heat bath at temperature T, with fixed N and V. Microstates have probabilities proportional to exp(−E/kT), where the Boltzmann factor exp(−E/kT) is the fundamental weight. This is the most commonly used ensemble in practice.
The canonical ensemble is the statistical mechanics framework for systems that can freely exchange energy with their surroundings — the situation in nearly every real experiment. Imagine a small system (say, a gas of N molecules) in a box with thin walls in contact with a huge thermal reservoir. The reservoir is so large that heat flowing in or out doesn't change its temperature. The system's temperature T is therefore fixed, but its energy fluctuates as it constantly exchanges tiny amounts of heat with the reservoir.
The central result is the Boltzmann distribution: the probability that the system occupies a microstate with energy E_i is P_i = exp(−E_i/kT) / Z, where Z = Σ exp(−E_j/kT) is the partition function summing over all microstates. The Boltzmann factor exp(−E/kT) is the engine of this formula. It tells you that lower-energy microstates are always more probable, and the sharpness of this preference depends on temperature. At very low T, the system is almost certainly in its ground state. At very high T, the exponential suppression weakens and the system explores high-energy microstates more freely.
From your prerequisite in ensemble theory, you know that an ensemble is a conceptual collection of many identical copies of a system, each in a possible microstate. In the canonical ensemble, those copies share the same T, N, and V but can have different energies. The probability distribution over those copies is precisely the Boltzmann distribution. This is what distinguishes the canonical from the microcanonical ensemble (which fixes energy) and the grand canonical ensemble (which allows particle exchange too).
The partition function Z is more than a normalization constant — it encodes all thermodynamic information about the system. The average energy is ⟨E⟩ = −∂(ln Z)/∂β where β = 1/kT. The Helmholtz free energy is F = −kT ln Z, and from F you can derive entropy, pressure, and heat capacity. This is why mastering the canonical ensemble opens the door to computing measurable thermodynamic quantities from a microscopic model of the system.
A key intuition: the canonical ensemble is a competition between energy minimization and entropy maximization. Low-energy states are favored by the Boltzmann factor, but there may be vastly more microstates at higher energies. The balance between these two tendencies — encoded in the free energy F = U − TS — determines the equilibrium state. Thermodynamics emerges from this competition.