Statistical Thermodynamics — Partition Function Intuition
Most students can quote the partition function long before they can say what it means. The formula looks like an abstract sum; in fact it answers one very concrete question — how many energy levels are effectively available to a molecule at this temperature? Once you read it that way, every result in statistical thermodynamics becomes predictable rather than memorised. This article builds that intuition and then computes real numbers for nitrogen at 298 K.
The definition, read as a headcount
where εi is measured from the lowest level, gi is its degeneracy, and kB = 1.380649 × 10⁻²³ J K⁻¹.
Look at the two limits. A level far above the ground state, ε ≫ kBT, contributes e−large ≈ 0 — it is invisible. A level close by, ε ≪ kBT, contributes almost exactly 1. So q counts the levels that thermal energy can actually reach, with partial credit for the ones on the boundary. The ground state alone always contributes 1, so q ≥ 1 always, and q = 1 means "only the ground state is populated".
Because a molecule's energy separates into independent contributions, the partition function factorises — a sum of energies becomes a product of partition functions:
The three expressions you need
Rotation (linear molecule, T ≫ θrot): qrot = T/(σ θrot), θrot = hcB̃/kB
Vibration (harmonic, energies from the zero-point level): qvib = 1/(1 − e−θvib/T), θvib = hcν̃/kB
Conversion constant: hc/kB = 1.4388 cm K
σ is the symmetry number — the number of indistinguishable orientations reachable by rotation (1 for CO and HCl, 2 for N₂, O₂ and CO₂, 3 for NH₃, 12 for CH₄ and benzene). It exists to stop you counting the same physical arrangement twice.
Worked example 1 — translation is astronomically large
Q. Find qtr for one N₂ molecule in a 1.00 L container at 298 K.
Step 1 — molecular mass.
m = 28.0134 × 1.66054 × 10⁻²⁷ kg = 4.6517 × 10⁻²⁶ kg
Step 2 — thermal wavelength.
2πm kBT = 6.2832 × 4.6517 × 10⁻²⁶ × 1.380649 × 10⁻²³ × 298 = 1.2025 × 10⁻⁴⁵
√(2πm kBT) = 3.4677 × 10⁻²³
Λ = 6.62607 × 10⁻³⁴ / 3.4677 × 10⁻²³ = 1.911 × 10⁻¹¹ m (19.1 pm)
Step 3 — divide the box by it.
Λ³ = 6.98 × 10⁻³³ m³, V = 1.00 × 10⁻³ m³
qtr = 1.00 × 10⁻³ / 6.98 × 10⁻³³ = 1.43 × 10²⁹
Read it physically: the box holds about 10²⁹ "thermal wavelength" cells, and each is an accessible translational state. This is why translational energy behaves classically and why N! appears in the canonical partition function — with 10²⁹ states for ~10²² molecules, two molecules essentially never occupy the same state.
Worked example 2 — rotation is modest, vibration is frozen
Q. For N₂ (B̃ = 1.9987 cm⁻¹, ν̃ = 2359 cm⁻¹, σ = 2), find qrot and qvib at 298 K.
Rotation.
θrot = 1.4388 × 1.9987 = 2.876 K
qrot = T/(σθrot) = 298 / (2 × 2.876) = 298/5.752 =
51.8
About 52 rotational states are in play. The high-temperature formula is safe here because 298 K ≫ 2.876 K.
Vibration.
θvib = 1.4388 × 2359 = 3394 K
θvib/T = 3394/298 = 11.39, so e−11.39 = 1.13 × 10⁻⁵
qvib = 1/(1 − 1.13 × 10⁻⁵) = 1.0000
Effectively only the vibrational ground state is populated: about one N₂ molecule in 88 000 is vibrationally excited at room temperature. The vibration is frozen out.
Why this immediately explains heat capacity
A mode contributes to CV only if it can absorb energy, and it can absorb energy only if q for that mode is greater than 1 and growing with T. So for N₂ at 298 K:
= (5/2) × 8.314 = 20.8 J K⁻¹ mol⁻¹
The measured value for N₂ near 298 K is 20.8 J K⁻¹ mol⁻¹. Full equipartition including vibration would predict 29.1 J K⁻¹ mol⁻¹ — and the experiment says no. Statistical thermodynamics does not merely reproduce the answer; it tells you which degrees of freedom were switched off and at what temperature they switch on.
Contrast iodine, ν̃ = 214.5 cm⁻¹, a much floppier bond: θvib = 1.4388 × 214.5 = 308.6 K, so at 298 K qvib = 1/(1 − e−1.036) = 1/(1 − 0.355) = 1.55. Iodine's vibration is genuinely active at room temperature, and its molar heat capacity is correspondingly larger than nitrogen's.
From q to thermodynamics
A − A(0) = −kBT ln(qN/N!) (indistinguishable particles)
S = N kB[ln(q/N) + 1] + (U − U(0))/T
The N! is not decoration. Without it the entropy is not extensive — double the system and the entropy more than doubles, which is the Gibbs paradox. Applying the corrected expression to qtr alone gives the Sackur–Tetrode equation, whose predicted standard molar entropies for monatomic gases agree with calorimetry to within a few tenths of a J K⁻¹ mol⁻¹.
The magnitudes, side by side
| Mode | Typical θ / K | q for N₂ at 298 K | Interpretation |
|---|---|---|---|
| Translation | ~10⁻¹⁴ (effectively zero) | 1.4 × 10²⁹ (1 L) | Continuum; fully classical |
| Rotation | 1–15 | 51.8 | Many levels populated; classical above ~50 K |
| Vibration | 500–5000 | 1.0000 | Ground state only; frozen out |
| Electronic | 10⁴–10⁵ | 1 (= g₀) | Ground state only, unless g₀ > 1 |
Common mistakes
- Dropping the symmetry number σ. For N₂ or CO₂ it doubles qrot and corrupts every entropy that follows.
- Using qN instead of qN/N! for a gas. Molecules are indistinguishable; the factorial is what makes S extensive.
- Mixing the two vibrational conventions. Measuring energies from the bottom of the well gives an extra factor e−θvib/2T. Choose one origin and state it.
- Using qrot = T/σθrot for H₂ at low temperature. H₂ has θrot ≈ 88 K, so the high-temperature approximation fails badly below room temperature — sum the levels explicitly instead.
- Forgetting electronic degeneracy. For O₂ (³Σ ground state) and for many radicals and transition-metal ions, qel = g₀ ≠ 1.
Run the numbers yourself. These calculations are exponentials, cube roots and unit conversions — exactly where a careless slip hides. The ABC Chemistry Calculator Suite includes the scientific constants and thermodynamic tools to check each step.
Open the ABC Chemistry Calculator Suite →Preparing for CSIR-NET, IIT-JAM, GATE or CUET-PG? ABC Chemistry runs dedicated competitive-exam batches at the coaching centre and online across India — details at abcchemistry.in.