Boltzmann Distribution Intuition (IPhO)

The Boltzmann factor e^(-E/kT) turns “a huge messy distribution” into something you can use in two lines. IPhO problems rarely want a full partition function grind; they want smart ratios, dominant terms, and clean limiting reasoning.

Mind-stretcher feedback

• A negative absolute temperature equilibrium requires an upper-bounded energy spectrum, so population inversion can be normalised. It is hotter than any positive temperature in the thermodynamic ordering, not colder than zero kelvin.

1. Definitions (Must Know)

• Boltzmann factor: the weight e^(-E/kT) that suppresses high-energy states at temperature T.
• Degeneracy gᵢ: number of microstates that share the same energy Eᵢ. It multiplies the Boltzmann weight.
• Probability pᵢ: fraction of systems in level i at equilibrium (canonical ensemble).
• Partition function Z: normalization constant making probabilities sum to 1.
• Energy zero is arbitrary: shifting all energies by the same constant does not change ratios.
• Thermal energy scale: kT sets what counts as “cheap” or “expensive” in energy. At T = 300 K, kT ≈ 4.1 × 10⁻²¹ J ≈ 0.026 eV.

2. Key Ideas (What Earns Marks)

• Use ratios first: most questions collapse to N₂/N₁ = (g₂/g₁)e^(-Δ E/kT).
• Compare Δ E to kT: if Δ E is much larger than kT, the higher state is rare; if Δ E is much smaller than kT, energies barely matter and degeneracy dominates.
• Factor out the ground state: write Z = e^(-E₀/kT)(g₀ + g₁ e^(-Δ E₁/kT) + …) to see what terms matter.
• Take logs to linearize: ln((N₂ g₁)/(N₁ g₂)) = -(Δ E)/kT is a straight-line relationship in 1/T.
• State the equilibrium assumption: Boltzmann weights are an equilibrium result; if the problem hints “not equilibrated”, say so before applying them.

3. Detailed Explanations

A. Where the exponential comes from (intuition you can quote)

The key message: a heat bath “charges” you an entropy penalty for taking energy from it, and that turns into the exponential suppression in energy.

B. Degeneracy competes with energy

Many IPhO questions hide the real physics in gᵢ.

• Weight of a level: gᵢ e^(-Eᵢ/kT).
• You can rewrite it as an “effective energy”: gᵢ e^(-Eᵢ/kT) = e^(-(Eᵢ-kT ln gᵢ)/kT). So a large degeneracy can compensate for a moderate energy cost.

C. Dominant-term thinking (how to avoid full sums)

Suppose you need Z = ∑ᵢ gᵢ e^(-Eᵢ/kT).

1. Subtract the minimum energy E₀: Z = e^(-E₀/kT)∑ᵢ gᵢ e^(-(Eᵢ-E₀)/kT).
2. Now every exponential has a nonnegative exponent argument (Eᵢ-E₀), so you can see at a glance which terms are tiny.

D. “What changes when T changes?”

• Increasing T increases kT, so energy differences matter less, and populations spread out across more levels.
• Decreasing T makes kT smaller, so populations concentrate into the lowest-energy states (weighted by degeneracy).

E. Continuous variables: remember the density

For a continuous variable x with energy E(x), the probability is a density: p(x) ∝ g(x)e^(-E(x)/kT), where g(x) is the density of states (the “degeneracy per interval”). Forgetting this is a common source of wrong shapes.

4. Common Mistakes

• Dropping degeneracy, or adding it when the problem is already counting microstates directly.
• Using e^(+E/kT) (wrong sign) or using E₂-E₁ with the sign flipped in the ratio.
• Mixing temperature scales: using Celsius inside kT instead of absolute temperature.
• Forgetting units: Δ E and kT must be in the same units (J with J, or eV with eV).
• Treating a non-equilibrium population as if it must follow Boltzmann weights without justification.
• Normalizing incorrectly when asked for an actual probability rather than a ratio.

5. Exam Tips

• If you only need a ratio, do not compute Z.
• Move energy zero to the ground state to keep exponents small: use Eᵢ-E₀.
• If a level has a big gᵢ, check whether it dominates even if it is higher in energy.
• When fitting or extracting T, take logs and look for a straight line in 1/T.
• Always write what “level” means in the question (microstate, energy level, macrostate). Marks can hinge on that interpretation.

6. Worked Examples

Example 1: Two-level system with degeneracy

Ground level: E₀ = 0, degeneracy g₀ = 1. Excited level: E₁ = Δ, degeneracy g₁ = 3.

• Population ratio: N₁/N₀ = (g₁/g₀)e^(-Δ/kT) = 3e^(-Δ/kT).
• Excited-state probability: p₁ = N₁/(N₀ + N₁) = (3e^(-Δ/kT))/(1 + 3e^(-Δ/kT)).

Example 2: Height distribution in a uniform gravitational field (barometric idea)

For particles of mass m in a uniform field with potential E(z) = mgz, the relative density at two heights is n(z₂)/n(z₁) = e^(-(mg(z₂-z₁))/kT). This is the same ratio trick: only the energy difference mgΔ z matters.

Example 3: Extracting a temperature from a measured ratio

Two levels have energies E₂ and E₁ with degeneracies g₂ and g₁. If an experiment measures N₂/N₁ = R, then R = (g₂/g₁)e^(-(E₂-E₁)/kT) ⇒ T = (E₂-E₁)/(k ln(g₂/g₁R)). The algebra is short, but only if you start from the ratio.

7. Mind Stretchers

• Degeneracy wins: find the condition on T for N₂ > N₁ when E₂ > E₁ but g₂ is large.
• Negative temperature (concept check): what must be true about the energy spectrum of a system for population inversion to be an equilibrium state?
• Energy zero: show explicitly that replacing Eᵢ → Eᵢ + E* does not change p₂/p₁.
• Many levels: if levels are equally spaced by Δ with equal degeneracy, estimate how many levels contribute significantly when Δ/kT is small.

8. Practice