Vibrational Partition Function

Heat a diatomic gas like HCl or CO to 500 K, and the molecule starts to buzz. Bonds stretch and compress with a definite quantum-mechanical rhythm. How much of that thermal energy goes into vibration, rather than translation or rotation? The answer hinges on a single sum: the vibrational partition function. This quantity matters because it converts a discrete ladder of vibrational energy levels into a continuous, temperature-dependent number. That number tells us how populated the excited vibrational states are at a given temperature. It feeds directly into predictions of heat capacity, internal energy, entropy, and even reaction rates through transition-state theory. This article derives the vibrational partition function for a harmonic oscillator using a geometric series. You will see how the choice of energy zero changes the formula's appearance. Then you will see why it never changes any measurable prediction. Two worked numerical examples ground the theory in real vibrational frequencies.

Key takeaways

The vibrational partition function sums Boltzmann factors over the harmonic oscillator ladder $v = 0, 1, 2, \ldots$
Because the levels are evenly spaced, the sum is an exact geometric series with a closed form.
Two conventions exist: measuring energy from the potential minimum (with zero-point energy included) or from the ground vibrational state itself.
At room temperature, most small-molecule vibrations have $q_{\text{vib}} \approx 1$ , meaning nearly all molecules sit in $v=0$ .
$q_{\text{vib}}$ grows toward $k_BT/h\nu$ only when $T$ is comparable to or larger than the vibrational temperature $\theta_{\text{vib}} = h\nu/k_B$ .
Key idea: the evenly-spaced harmonic ladder turns an infinite sum of Boltzmann factors into one clean closed-form expression, valid until anharmonicity or dissociation intervenes.

The core insight

Because harmonic oscillator energy levels are equally spaced by $h\nu$ , the vibrational partition function is a geometric series in $e^{-h\nu/k_BT}$ that sums exactly to a closed form; the only subtlety is deciding whether the zero of energy sits at the bottom of the potential well or at the true vibrational ground state.

The physical picture

A diatomic molecule's bond behaves, to a good first approximation, like a mass on a spring. Quantum mechanics quantizes this motion into evenly spaced energy levels,

\varepsilon_v = \left(v + \tfrac{1}{2}\right)h\nu, \quad v = 0, 1, 2, \ldots

where $\nu$ is the classical vibrational frequency set by the force constant and reduced mass. The $\tfrac{1}{2}h\nu$ term is the zero-point energy: even in the lowest state $v=0$ , the oscillator never sits still. This is a direct consequence of the uncertainty principle, and it is the same ladder structure used in the quantum treatment of vibrational spectroscopy — but here we ask a thermodynamic question, not a spectroscopic one: how are molecules distributed across these levels at temperature $T$ ? At low temperature, essentially every molecule occupies $v=0$ . As $T$ rises, higher rungs become thermally accessible, following a Boltzmann distribution. The vibrational partition function is precisely the normalizing sum over this distribution — it counts the "effective number of thermally accessible vibrational states." For most small molecules at room temperature, $h\nu \gg k_BT$ , so this number stays close to 1: vibration is the "frozen" mode compared to translation and rotation, which is why it plays only a small role in ordinary room-temperature heat capacities.

Figure

The harmonic oscillator ladder and its populations

The second figure shows how $q_{\text{vib}}$ itself behaves as a function of the dimensionless ratio $\theta_{\text{vib}}/T$ , where $\theta_{\text{vib}} = h\nu/k_B$ is the vibrational temperature. This single curve governs every harmonic oscillator, regardless of the specific molecule.

Figure

q_vib as a function of the vibrational temperature ratio

The core relations

q_{\text{vib}} = \sum_{v=0}^{\infty} e^{-\varepsilon_v / k_BT}

where: the sum runs over all vibrational quantum numbers with energies $\varepsilon_v$ measured from a chosen zero.

q_{\text{vib}}^{(0)} = \frac{e^{-h\nu/2k_BT}}{1 - e^{-h\nu/k_BT}}

where: superscript $(0)$ denotes energy measured from the bottom of the potential well, so the zero-point energy $\tfrac{1}{2}h\nu$ is retained explicitly.

q_{\text{vib}}^{(v=0)} = \frac{1}{1 - e^{-h\nu/k_BT}}

where: this form measures energy from the $v=0$ level itself, absorbing the zero-point energy into the reference.

\theta_{\text{vib}} = \frac{h\nu}{k_B}

where: $\theta_{\text{vib}}$ is the vibrational temperature, a convenient benchmark comparing $h\nu$ to thermal energy $k_BT$ .

q_{\text{vib}}^{(0)} = e^{-\theta_{\text{vib}}/2T}\, q_{\text{vib}}^{(v=0)}

where: the two conventions differ only by a temperature-independent-in-form (but $T$ -dependent through the exponential) prefactor tied to the zero-point energy.

Derivation ASumming the geometric series for the harmonic oscillator

Goal: obtain a closed-form expression for $q_{\text{vib}}$ starting from the definition of the partition function applied to evenly spaced harmonic oscillator levels.

Step 1. Write the defining sum. The partition function for a single vibrational mode is

q_{\text{vib}} = \sum_{v=0}^{\infty} e^{-\varepsilon_v/k_BT}

with $\varepsilon_v = (v + \tfrac{1}{2})h\nu$ measured from the bottom of the potential well.

Step 2. Separate the zero-point term. Because the exponent is a sum, the exponential factors:

e^{-\varepsilon_v/k_BT} = e^{-h\nu/2k_BT}\, e^{-vh\nu/k_BT}

The zero-point piece $e^{-h\nu/2k_BT}$ does not depend on $v$ , so it factors outside the sum:

q_{\text{vib}}^{(0)} = e^{-h\nu/2k_BT} \sum_{v=0}^{\infty} \left(e^{-h\nu/k_BT}\right)^{v}

Step 3. Recognize the geometric series. Let $x = e^{-h\nu/k_BT}$ . Since $h\nu > 0$ and $T>0$ , we know $0 < x < 1$ for all physical temperatures, so the series

\sum_{v=0}^{\infty} x^v = 1 + x + x^2 + \cdots

converges absolutely. This is the standard geometric series, summing to $\dfrac{1}{1-x}$ .

Step 4. Assemble the result. Substituting back,

q_{\text{vib}}^{(0)} = \frac{e^{-h\nu/2k_BT}}{1 - e^{-h\nu/k_BT}}

Step 5. Shift the energy zero to the ground vibrational state. If instead we measure energies relative to $v=0$ , so $\varepsilon_v' = vh\nu$ , the same steps (without the extra zero-point prefactor) give

q_{\text{vib}}^{(v=0)} = \sum_{v=0}^{\infty} x^v = \frac{1}{1 - e^{-h\nu/k_BT}}

Both are legitimate; they differ only by the constant energy shift $\tfrac{1}{2}h\nu$ between the two zero choices, exactly analogous to choosing a reference state for enthalpy. Physical observables built from $q_{\text{vib}}$ — internal energy differences, heat capacities, entropy — are insensitive to this choice, because they depend only on derivatives of $\ln q_{\text{vib}}$ with respect to $T$ , and an additive constant in the exponent (equivalently, a multiplicative factor independent of $T$ -derivatives beyond the shift) simply relabels the energy origin.

\boxed{q_{\text{vib}}^{(0)} = \frac{e^{-\theta_{\text{vib}}/2T}}{1 - e^{-\theta_{\text{vib}}/T}}, \qquad q_{\text{vib}}^{(v=0)} = \frac{1}{1 - e^{-\theta_{\text{vib}}/T}}}

Worked examples

Example 1 — HCl at 298 K (foundational). HCl has a fundamental vibrational wavenumber $\tilde\nu = 2990~\text{cm}^{-1}$ . Convert to a vibrational temperature. First, find $h\nu = hc\tilde\nu$ :

h\nu = (6.626\times10^{-34}~\text{J s})(2.998\times10^{10}~\text{cm s}^{-1})(2990~\text{cm}^{-1})

h\nu = 5.941\times10^{-20}~\text{J}

Then

\theta_{\text{vib}} = \frac{h\nu}{k_B} = \frac{5.941\times10^{-20}~\text{J}}{1.381\times10^{-23}~\text{J K}^{-1}} = 4303~\text{K}

At $T = 298$ K, the ratio is

\frac{\theta_{\text{vib}}}{T} = \frac{4303}{298} = 14.44

Using the ground-state-referenced formula,

q_{\text{vib}}^{(v=0)} = \frac{1}{1 - e^{-14.44}} = \frac{1}{1 - 5.3\times10^{-7}} \approx 1.0000005

Sanity check: since $\theta_{\text{vib}} \gg T$ , we expect $q_{\text{vib}} \approx 1$ , meaning virtually all HCl molecules sit in $v=0$ at room temperature. The result confirms this: only about 1 molecule in 2 million occupies $v=1$ or higher. This is why vibrational modes contribute negligibly to $C_V$ for HCl gas near 298 K — a preview of the heat-capacity discussion to come.

Example 2 — I₂ at 500 K (applied, harder). Iodine's bond is far softer and heavier than HCl's, with $\tilde\nu = 214.5~\text{cm}^{-1}$ . We want $q_{\text{vib}}$ at 500 K, and also the fraction of molecules in $v=0$ . Step 1: compute $\theta_{\text{vib}}$ .

h\nu = (6.626\times10^{-34})(2.998\times10^{10})(214.5) = 4.257\times10^{-21}~\text{J}

\theta_{\text{vib}} = \frac{4.257\times10^{-21}}{1.381\times10^{-23}} = 308.2~\text{K}

Step 2: form the ratio and exponential.

\frac{\theta_{\text{vib}}}{T} = \frac{308.2}{500} = 0.6164

e^{-0.6164} = 0.5398

Step 3: evaluate $q_{\text{vib}}^{(v=0)}$ .

q_{\text{vib}}^{(v=0)} = \frac{1}{1 - 0.5398} = \frac{1}{0.4602} = 2.173

Step 4: find the fraction in $v=0$ using $p_0 = 1/q_{\text{vib}}^{(v=0)}$ (since the Boltzmann factor for $v=0$ referenced to itself is exactly 1).

p_0 = \frac{1}{2.173} = 0.4602 \quad (46.0\%)

Step 5: find the fraction in $v=1$ : $p_1 = e^{-\theta_{\text{vib}}/T}/q_{\text{vib}} = 0.5398/2.173 = 0.2484$ (24.8%). Sanity check: $p_0 + p_1 = 70.8\%$ , leaving about 29.2% spread over $v \ge 2$ , consistent with a slowly decaying geometric distribution rather than a sharply peaked one. Compare to HCl: because I₂'s bond is weaker and heavier ( $\theta_{\text{vib}}$ over an order of magnitude smaller), vibration is thermally "active" at a modest 500 K, giving $q_{\text{vib}} \approx 2.17$ instead of $\approx 1$ . This directly foreshadows why heavy, floppy molecules show measurable vibrational heat capacity at temperatures where stiff hydrides do not.

Limiting cases & common traps

Regime	Behavior	Interpretation
$T \ll \theta_{\text{vib}}$	$q_{\text{vib}}^{(v=0)} \to 1$	Vibration frozen; only $v=0$ populated
$T \gg \theta_{\text{vib}}$	$q_{\text{vib}}^{(v=0)} \to T/\theta_{\text{vib}} = k_BT/h\nu$	Classical equipartition limit, high- $T$ expansion of $1/(1-e^{-x})\approx 1/x$
Stiff bonds / light atoms (H–X)	Large $\theta_{\text{vib}}$	$q_{\text{vib}} \approx 1$ near room T (e.g., HCl, HF)
Weak bonds / heavy atoms	Small $\theta_{\text{vib}}$	$q_{\text{vib}} > 1$ even at moderate T (e.g., I₂, Br₂)
Near dissociation, high $v$	Harmonic formula breaks down	Real potentials are anharmonic (Morse-like); the geometric series overcounts high-lying states

The chemist behind it

The vibrational partition function belongs to the broader statistical mechanical framework built by Josiah Willard Gibbs and Ludwig Boltzmann, but its molecular, quantized form owes its practical shape to the era of vibrational spectroscopy that followed the birth of quantum mechanics. Once Planck's quantization of oscillator energy and the Bohr–Sommerfeld picture matured into full quantum theory, chemists like Richard C. Tolman and later Henry Eyring applied these evenly spaced vibrational levels directly to gas-phase thermodynamics, linking spectroscopic constants (vibrational wavenumbers measured by infrared and Raman spectroscopy) to bulk heat capacities and equilibrium constants. This synthesis, crystallized in texts by Fowler and Guggenheim and later by McQuarrie, is what allows a single infrared band position to predict a macroscopic thermodynamic property with no adjustable parameters.

Origins · 1 minRichard Chace Tolman and how this was discovered

Richard Chace Tolman

1881–1948 · American physical chemist who applied statistical mechanics rigorously to molecular energy level sums, helping formalize the partition function approach used for vibrational and rotational modes

Ask the tutor: How does differentiating q_vib with temperature give the Einstein heat capacity curve for vibrations?

Why is q_vib often close to 1 at room temperature?

Most common molecules have vibrational wavenumbers of several hundred to a few thousand cm⁻¹, corresponding to vibrational temperatures of hundreds to thousands of kelvin. At 298 K, $\theta_{\text{vib}}/T$ is typically large, so $e^{-\theta_{\text{vib}}/T}$ is tiny and the geometric series barely exceeds its first term. Physically, this means nearly all molecules occupy the ground vibrational state, and excited vibrational levels are thermally inaccessible. Only soft, heavy-atom bonds like I–I or Br–Br have small enough $\theta_{\text{vib}}$ to give $q_{\text{vib}}$ noticeably above 1 near room temperature.

Does the choice of energy zero affect the calculated internal energy or heat capacity?

No. Shifting the zero of energy by a constant (here, the zero-point energy $\tfrac{1}{2}h\nu$ ) only adds a constant to $\ln q_{\text{vib}}$ 's derivative structure in a trivial, predictable way: it shifts the absolute internal energy by that same constant but leaves $C_V = (\partial U/\partial T)_V$ completely unchanged, since the constant has zero temperature derivative. Predicted heat capacities, entropies, and equilibrium constants (when computed consistently) come out identical regardless of which convention you use, as long as you are consistent throughout a calculation.

How does q_vib connect to the harmonic oscillator's heat capacity contribution?

Differentiating $\ln q_{\text{vib}}$ twice with respect to temperature produces the Einstein-type vibrational heat capacity formula, a bell-shaped function of $\theta_{\text{vib}}/T$ that rises from zero at low $T$ toward the classical value $R$ (per mode) at high $T$ . This is the direct next step built on this partition function, covered in the treatment of heat capacity contributions from molecular modes. The vibrational partition function is the necessary starting point for that derivative-based analysis.

Is the harmonic oscillator approximation ever seriously wrong for q_vib?

Yes, particularly near dissociation or for highly anharmonic, floppy modes. Real potentials flatten out at large bond extension, so high- $v$ levels crowd together rather than remaining evenly spaced, and eventually the molecule dissociates rather than continuing up an infinite ladder. The harmonic geometric-series formula formally sums to infinity and can overestimate $q_{\text{vib}}$ at very high temperatures. For most stable molecules well below their dissociation limit, though, the harmonic approximation is excellent and matches spectroscopic data closely.