Mathematics.

statistical mechanics

Partition Functions

Mathematical Physics90 minDifficulty8 out of 10

Overview

The partition function is the central object of equilibrium statistical mechanics. For a physical system at temperature T with Hamiltonian H, the canonical partition function Z encodes all thermodynamic information: free energy, entropy, specific heat, and higher moments follow from derivatives of ln Z. Mathematically, Z is a Laplace transform of the density of states, and its analytic properties (singularities, Lee-Yang zeros) govern phase transitions. In quantum field theory, the generating functional — a continuum generalisation — plays the same role.

Intuition

Think of Z as a weighted count of all possible states a system can be in. At low temperature (large β = 1/kT), only the lowest-energy states have appreciable weight e^{-βE}; the system is 'frozen' near the ground state. At high temperature (small β), all states are nearly equally weighted and the system is disordered. The free energy F = -kT ln Z is the 'most useful' energy for predicting equilibrium: systems minimise F at fixed temperature.

Formal Definition

Definition

For a quantum system with discrete energy levels {Eₙ} or a classical system with phase space Γ, the canonical partition function at inverse temperature β = 1/(kBT) is defined as follows.

Z(β)=TreβH=neβEnZ(\beta) = \mathrm{Tr}\,e^{-\beta H} = \sum_n e^{-\beta E_n}
Canonical partition function (quantum discrete)
Z(β)=1h3NN!ΓeβH(q,p)dqdpZ(\beta) = \frac{1}{h^{3N} N!}\int_{\Gamma} e^{-\beta H(\mathbf{q},\mathbf{p})}\,d\mathbf{q}\,d\mathbf{p}
Canonical partition function (classical, N identical particles)
F=kBTlnZF = -k_B T \ln Z
Helmholtz free energy
E=lnZβ,CV=kBβ22lnZβ2\langle E \rangle = -\frac{\partial \ln Z}{\partial \beta}, \quad C_V = k_B \beta^2 \frac{\partial^2 \ln Z}{\partial \beta^2}
Mean energy and heat capacity from Z
ZGC(β,μ)=Treβ(HμN)Z_{\mathrm{GC}}(\beta, \mu) = \mathrm{Tr}\,e^{-\beta(H - \mu N)}
Grand canonical partition function

Notation

NotationMeaning
Z(β)Z(\beta)Canonical partition function at inverse temperature β
β=1/(kBT)\beta = 1/(k_B T)Inverse temperature
F=kBTlnZF = -k_B T \ln ZHelmholtz free energy
Ω(E)\Omega(E)Density of states; Z(β) = ∫Ω(E)e^{-βE} dE
ZGCZ_{\mathrm{GC}}Grand canonical partition function

Properties

Z as Laplace transform of density of states

Z(β)=0Ω(E)eβEdEZ(\beta) = \int_0^\infty \Omega(E)\, e^{-\beta E}\, dE

Thermodynamic relations from ln Z

S=kB(lnZ+βE),P=kBTlnZVS = k_B\left(\ln Z + \beta \langle E\rangle\right), \quad P = k_B T \frac{\partial \ln Z}{\partial V}

Factorisation for non-interacting subsystems

Ztot=Z1Z2ZkZ_{\mathrm{tot}} = Z_1 \cdot Z_2 \cdots Z_k

Worked Examples

  1. 1

    The two energy levels contribute one term each to the sum over states.

    Z=eβ0+eβε=1+eβεZ = e^{-\beta \cdot 0} + e^{-\beta \varepsilon} = 1 + e^{-\beta\varepsilon}
  2. 2

    The mean energy follows by differentiating.

    E=lnZβ=εeβε1+eβε=εeβε+1\langle E \rangle = -\frac{\partial \ln Z}{\partial\beta} = \frac{\varepsilon\, e^{-\beta\varepsilon}}{1 + e^{-\beta\varepsilon}} = \frac{\varepsilon}{e^{\beta\varepsilon} + 1}
  3. 3

    This is the Fermi-Dirac occupancy of a two-level quantum system (e.g. spin-1/2 in a magnetic field).

    ET00,ETε/2\langle E \rangle \xrightarrow{T\to 0} 0, \quad \langle E \rangle \xrightarrow{T\to\infty} \varepsilon/2

✓ Answer

Z = 1 + e^{-βε}; the mean energy interpolates between 0 at T=0 and ε/2 at T→∞.

Practice Problems

Mediumfree response

For a monatomic ideal gas of N particles in volume V, show that Z = (V/λ³)^N / N! where λ = h/√(2πmkBT) is the thermal de Broglie wavelength.

Hardfree response

Explain how a phase transition manifests as a non-analyticity in ln Z in the thermodynamic limit, and what the Lee-Yang theorem says about finite systems.

Common Mistakes

Common Mistake

Forgetting the factor 1/N! for identical classical particles

Identical classical particles are indistinguishable. Without the 1/N! Gibbs correction the entropy is not extensive (Gibbs paradox). The correct classical Z = Z₁^N / N!.

Common Mistake

Confusing the canonical and grand canonical ensembles

The canonical ensemble fixes N (particle number); ZC = Tr e^{-βH}. The grand canonical ensemble allows N to fluctuate; ZGC = Tr e^{-β(H-μN)}. Choose the ensemble appropriate to the physical constraint.

Quiz

The Helmholtz free energy F is related to the partition function by:
For two non-interacting subsystems with partition functions Z₁ and Z₂, the total partition function is:

Historical Background

Boltzmann introduced the statistical connection between entropy and the number of microstates in the 1870s–80s. Gibbs systematised the canonical ensemble and introduced the partition function in his 1902 monograph. The term 'partition function' (Zustandssumme in German, literally 'sum over states') reflects the summation over all configurations. Lee and Yang's 1952 work on zeros of the partition function in the complex fugacity plane gave a rigorous framework for understanding phase transitions.

  1. 1877

    Boltzmann's entropy formula S = k log W

    Ludwig Boltzmann

  2. 1902

    Gibbs introduces the canonical ensemble and partition function systematically

    Josiah Willard Gibbs

  3. 1952

    Lee and Yang study zeros of Z in the complex plane to characterise phase transitions

    Tsung-Dao Lee, Chen-Ning Yang

  4. 1965

    Ruelle proves existence of the thermodynamic limit for a broad class of systems

    David Ruelle

Summary

  • The canonical partition function Z(β) = Tr e^{-βH} is the generating function for all equilibrium thermodynamic quantities.
  • Helmholtz free energy F = -kBT ln Z; mean energy and heat capacity follow from derivatives of ln Z.
  • Z factorises for non-interacting subsystems, making analytic calculations tractable.
  • Phase transitions correspond to non-analyticities of ln Z that arise only in the thermodynamic limit N → ∞.
  • The grand canonical partition function ZGC extends the formalism to open systems with variable particle number.

References

  1. BookRuelle, D. — Statistical Mechanics: Rigorous Results (1969), W.A. Benjamin
  2. BookGibbs, J.W. — Elementary Principles in Statistical Mechanics (1902), Yale University Press
  3. BookHuang, K. — Statistical Mechanics, 2nd ed. (1987), Wiley, Chapters 7–8