Factorization of the partition function: independent subsystems and the grand potential

Intuition Beginner

A big system made of independent parts has a partition function that is the product of the partition functions of its parts. This simple rule — multiply partition functions for independent subsystems — is one of the most powerful tools in statistical mechanics.

Why does it work? The partition function $Z$ is a sum of Boltzmann factors over all states. If the system is two independent parts $A$ and $B$, every state of the combined system is a pair (state of $A$, state of $B$). The energy of the pair is the sum: $E_{AB}=E_A+E_B$. So the Boltzmann factor splits:

$$e^{-E_{AB}/kT}=e^{-E_A/kT}\times e^{-E_B/kT}.$$

Adding up over all pairs of states gives the product of the individual sums: $Z_{AB}=Z_A\times Z_B$.

The consequence for free energy is even simpler. Since $F=-kT\ln Z$, the product of partition functions becomes a sum of free energies: $F_{AB}=F_A+F_B$. Free energy is additive for independent subsystems. Heat capacities are also additive: $C_{AB}=C_A+C_B$.

Visual Beginner

Consider a box of $N$ non-interacting molecules. Each molecule has translational, rotational, vibrational, and electronic energy levels — four "shelves" of energy, independent of each other. The total partition function of one molecule is the product of four factors:

$$Z_1=Z_{\mathrm{trans}}\times Z_{\mathrm{rot}}\times Z_{\mathrm{vib}}\times Z_{\mathrm{elec}}.$$

The partition function of the whole gas (with $N$ distinguishable molecules) would be $Z=(Z_1{)}^N$. For identical molecules, quantum indistinguishability introduces a factor of $1/N!$, giving $Z=(Z_1{)}^N/N!$.

Worked example Beginner

A diatomic molecule has a rotational partition function $Z_{\mathrm{rot}}=T/(2{\Theta }_r)$ at high temperature (where ${\Theta }_r={\hslash }^2/(2Ik)$ is the rotational temperature and $I$ is the moment of inertia). Its vibrational partition function is $Z_{\mathrm{vib}}=e^{-{\Theta }_v/(2T)}/(1-e^{-{\Theta }_v/T})$ where ${\Theta }_v=\hslash \omega /k$ is the vibrational temperature.

For nitrogen (N${}_2$): ${\Theta }_r=2.88$ K and ${\Theta }_v=3370$ K. At room temperature $T=300$ K:

• $Z_{\mathrm{rot}}=300/(2\times 2.88)=52.1$ (many rotational states are accessible).
• $Z_{\mathrm{vib}}=e^{-3370/600}/(1-e^{-3370/300})\approx 1/(1-e^{-11.2})\approx 1.00002$ (essentially frozen in the ground state).

The vibrational contribution is negligible at room temperature — this is why the heat capacity of N${}_2$ at 300 K is close to $(5/2)R$ (translation plus rotation only), not $(7/2)R$ (which would include vibration).

Check your understanding Beginner

Formal definition Intermediate+

Theorem (Factorisation of the partition function). Let a system consist of two independent subsystems $A$ and $B$, with energy spectra $\{E_a\}$ and $\{E_b\}$. The total energy of a combined state $(a,b)$ is $E_{ab}=E_a+E_b$. Then the canonical partition function of the combined system is

$$Z_{AB}(\beta )=Z_A(\beta )\cdot Z_B(\beta ).$$

Proof. By definition, $Z_{AB}={\sum }_{a,b}{e}^{-\beta E_{ab}}={\sum }_{a,b}{e}^{-\beta (E_a+E_b)}=\left({\sum }_a{e}^{-\beta E_a}\right)\left({\sum }_b{e}^{-\beta E_b}\right)=Z_A{Z}_B.$ $\square$

Corollary. For $N$ independent, identical, distinguishable subsystems: $Z_N=(Z_1{)}^N$. For $N$ identical bosons or fermions, the factorisation is modified by quantum statistics [11.05.01, 11.05.02].

Free energy additivity

Since $F=-kT\ln Z$:

$$F_{AB}=-kT\ln (Z_A{Z}_B)=-kT\ln Z_A-kT\ln Z_B=F_A+F_B.$$

All thermodynamic potentials that are derivatives of $F$ inherit this additivity:

$$U_{AB}=U_A+U_B,S_{AB}=S_A+S_B,C_{V,AB}=C_{V,A}+C_{V,B}.$$

Molecular partition function

For a single molecule with independent degrees of freedom (translation, rotation, vibration, electronic):

$$Z_1=Z_{\mathrm{trans}}\times Z_{\mathrm{rot}}\times Z_{\mathrm{vib}}\times Z_{\mathrm{elec}}.$$

The translational partition function is

$$Z_{\mathrm{trans}}=V{\left(\frac{2\pi mkT}{h^2}\right)}^{3/2}=\frac{V}{{\lambda }^3},$$

where $\lambda =h/\sqrt{2\pi mkT}$ is the thermal de Broglie wavelength.

For a linear rotor with moment of inertia $I$ and symmetry number $\sigma$:

$$Z_{\mathrm{rot}}=\frac{1}{\sigma }\frac{T}{{\Theta }_r},{\Theta }_r=\frac{{\hslash }^2}{2Ik}.$$

For a harmonic oscillator of frequency $\omega$:

$$Z_{\mathrm{vib}}=\frac{e^{-{\Theta }_v/(2T)}}{1-e^{-{\Theta }_v/T}},{\Theta }_v=\frac{\hslash \omega }{k}.$$

For a two-level electronic system (ground state and first excited state at energy $ϵ$):

$$Z_{\mathrm{elec}}=g_0+g_1{e}^{-ϵ/kT},$$

where $g_0,g_1$ are the degeneracies.

Key derivation: Internal energy and heat capacity from factorisation Intermediate+

Proposition. The internal energy of a composite system of independent subsystems decomposes as $U={\sum }_j{U}_j$, where $U_j=-\partial \ln Z_j/\partial \beta$. The heat capacity decomposes as $C_V={\sum }_j{C}_{V,j}$.

Proof. From $Z={\prod }_j{Z}_j$:

$$\ln Z=\sum _j\ln Z_j.$$ $$U=-\frac{\partial \ln Z}{\partial \beta }=-\sum _j\frac{\partial \ln Z_j}{\partial \beta }=\sum _j{U}_j.$$ $$C_V=\frac{\partial U}{\partial T}=\sum _j\frac{\partial U_j}{\partial T}=\sum _j{C}_{V,j}.\square$$

Each degree of freedom contributes independently to the heat capacity. At sufficiently high temperature, each quadratic degree of freedom contributes $k/2$ per molecule (equipartition), but at intermediate temperatures the contributions "turn on" at different characteristic temperatures:

• Translation: always active ($T\gg 0$).
• Rotation: active for $T\gg {\Theta }_r$.
• Vibration: active for $T\gg {\Theta }_v$.

Bridge. The factorisation of the partition function builds toward the Sackur-Tetrode equation 11.04.04 and the Einstein solid 11.04.05, where the product structure of $Z$ makes the thermodynamics tractable. This is exactly the property that allows molecular thermodynamics to decompose heat capacities into translational, rotational, and vibrational contributions. The foundational reason is that independent subsystems have a Hamiltonian that splits as a sum, and the Boltzmann factor $e^{-\beta H}$ factorises into a product because $e^{-\beta (H_A+H_B)}=e^{-\beta H_A}{e}^{-\beta H_B}$. This generalises the additive structure of thermodynamics and is dual to the Legendre-transform structure that makes free energies additive. Putting these together, the partition function is the exponential generating function of thermodynamics: its logarithm is the free energy, and factorisation of $Z$ corresponds to additivity of $F$.

Exercises Intermediate+

Lean formalization Intermediate+

The factorisation theorem for partition functions is a straightforward application of the distributive law: $({\sum }_a{e}^{-\beta E_a})({\sum }_b{e}^{-\beta E_b})={\sum }_{a,b}{e}^{-\beta (E_a+E_b)}$. Formalising this requires a type for partition functions with a product operation and a proof that the product of partition functions equals the partition function of the independent composite. The mathematical content is elementary; the gap is the absence of a partition-function type in Mathlib.

Advanced results Master

Gibbs paradox and the $1/N!$ correction

If $N$ identical particles are treated as distinguishable, the partition function $Z=Z_1^N$ gives an entropy $S=Nk[\ln (V/{\lambda }^3)+3/2]$ that is not extensive — doubling $N$ and $V$ more than doubles $S$ because of the spurious $N^2$ permutations. Gibbs (1875) resolved this by inserting the factor $1/N!$, giving $Z=Z_1^N/N!$ and the correct extensive entropy $S=Nk[\ln (V/(N{\lambda }^3))+5/2]$ (the Sackur-Tetrode equation 11.04.04).

The modern resolution is that quantum indistinguishability automatically produces the $1/N!$ correction. In the classical limit ($n{\lambda }^3\ll 1$), the exact quantum partition function reduces to the Gibbs-corrected classical form. The factor $1/N!$ is not an ad hoc fix but a consequence of the quantum-mechanical treatment of identical particles.

Configurational versus kinetic contributions

For a classical system of $N$ particles with Hamiltonian $H={\sum }_i{p}_i^2/(2m)+V(\{r_i\})$, the partition function factorises into a kinetic (momentum-space) part and a configurational (position-space) part:

$$Z=\frac{1}{N!{\lambda }^{3N}}\int d^{3N}r{e}^{-\beta V(\{r_i\})}.$$

The momentum integral gives ${\lambda }^{-3N}$, and the configurational integral $Q_N=\int d^{3N}r{e}^{-\beta V}$ carries all the information about interparticle interactions. For an ideal gas, $V=0$ and $Q_N=V^N$, recovering $Z=V^N/(N!{\lambda }^{3N})$.

Quantum factorisation: the grand canonical approach

For quantum particles, the partition function does not factorise into single-particle contributions in the canonical ensemble (because entanglement correlates the particles). In the grand canonical ensemble 11.03.02, however, the grand partition function factorises into single-mode contributions $\Xi ={\prod }_k{\Xi }_k$ because the Fock-space occupation numbers $\{n_k\}$ are independent for non-interacting particles. This is the fundamental reason the grand canonical framework is preferred for quantum statistics.

Synthesis. The factorisation of the partition function builds toward the computational machinery of molecular thermodynamics 14.06.01 and the quantum statistics of ideal gases [11.05.01, 11.05.02]. The central insight is that independence at the Hamiltonian level translates directly into factorisation at the partition-function level and additivity at the free-energy level — the mathematical structure mirrors the physical decomposition. This is dual to the Legendre-transform structure of thermodynamics 11.01.02, where the different free energies are generated by swapping independent variables. The foundational reason the whole apparatus works is that the Boltzmann factor $e^{-\beta H}$ maps the additive Hamiltonian to a multiplicative weight, and the sum over states maps the product back to a sum over logarithms — this is exactly the exponential/logarithm duality that makes statistical mechanics computationally tractable. Putting these together, the partition function is the bridge between microscopic Hamiltonians and macroscopic thermodynamics, and its factorisation is the structural reason why complex systems can be decomposed into simpler parts.

Full proof set Master

Proposition. For $N$ identical indistinguishable particles in the classical limit ($n{\lambda }^3\ll 1$), the partition function is $Z=Z_1^N/N!$ and the Helmholtz free energy is $F=-kT[N\ln (Z_1/N)+N]$.

Proof. The exact quantum partition function for $N$ identical bosons or fermions involves symmetrising or antisymmetrising the $N$-particle wave function. In the classical limit, the thermal wavelength $\lambda$ is much smaller than the interparticle spacing $n^{-1/3}$, so the probability of two particles occupying the same quantum state is negligible. The symmetrisation or antisymmetrisation reduces to dividing by $N!$ (the number of permutations of $N$ particles), giving $Z=Z_1^N/N!$.

Using Stirling's approximation: $\ln N!\approx N\ln N-N$.

$\ln Z=N\ln Z_1-N\ln N+N=N\ln (Z_1/N)+N$.

$F=-kT\ln Z=-NkT[\ln (Z_1/N)+1]$.

With $Z_1=V/{\lambda }^3$: $F=-NkT[\ln (V/(N{\lambda }^3))+1]$. This is the Helmholtz free energy of the classical ideal gas, extensive and consistent with the Sackur-Tetrode equation. $\square$

Connections Master

• 11.04.04 The Sackur-Tetrode entropy is the direct application of the factorisation $Z=V^N/(N!{\lambda }^{3N})$ to compute $S=-(\partial F/\partial T{)}_V$.
• 11.04.05 The Einstein solid factorises into $N$ independent harmonic oscillators: $Z=(Z_{\mathrm{ho}}{)}^N$, giving the heat capacity of a crystalline solid.
• 11.03.02 The grand canonical factorisation $\Xi ={\prod }_k{\Xi }_k$ is the quantum analogue, where the product is over single-particle modes rather than particles.
• 11.05.01 The Bose-Einstein distribution follows from the factorisation $\Xi ={\prod }_{k}1/(1-z{e}^{-\beta {ϵ}_k})$ for non-interacting bosons.
• 11.01.02 The Legendre transform structure of thermodynamic potentials mirrors the factorisation: $F$ is additive because $\ln Z$ maps products to sums.

Historical and philosophical context Master

Gibbs recognised the need for the $1/N!$ correction in his 1875 thermodynamic work, well before quantum mechanics provided the microscopic justification. The Gibbs paradox — the observation that without the correction, entropy would increase when two volumes of the same gas at the same temperature and pressure are allowed to mix — was a conceptual puzzle that foreshadowed the quantum-mechanical principle of indistinguishability.

The factorisation of the partition function into molecular contributions (translation, rotation, vibration, electronic) is the basis of statistical thermodynamics as applied in chemistry and engineering. It allows spectroscopic data (rotational constants, vibrational frequencies, electronic excitation energies) to be converted directly into thermodynamic properties (heat capacities, equilibrium constants, Gibbs free energies of formation). This connection between spectroscopy and thermodynamics, first exploited by Tolman (1918) and developed systematically by Giauque (1930s), is one of the great successes of statistical mechanics.

Bibliography Master

@book{pathria-beale2011,
  author = {Pathria, R. K. and Beale, Paul D.},
  title = {Statistical Mechanics},
  edition = {3rd},
  publisher = {Academic Press},
  year = {2011}
}

@book{reif1965,
  author = {Reif, Frederick},
  title = {Fundamentals of Statistical and Thermal Physics},
  publisher = {McGraw-Hill},
  year = {1965}
}

@book{huang1987,
  author = {Huang, Kerson},
  title = {Statistical Mechanics},
  edition = {2nd},
  publisher = {Wiley},
  year = {1987}
}

@book{callen1985,
  author = {Callen, Herbert B.},
  title = {Thermodynamics and an Introduction to Thermostatistics},
  edition = {2nd},
  publisher = {Wiley},
  year = {1985}
}