Statistical Mechanics

Bosons and Fermions

Occupation numbers in the grand canonical ensemble. A single sign separates bosons from fermions.

Updated July 2026

Identical particles in quantum mechanics are not merely alike, they are indistinguishable, and swapping two of them cannot change anything measurable. That forces the many-particle state to be either symmetric or antisymmetric under the swap. The symmetric ones are bosons, the antisymmetric ones fermions, and almost everything about how matter behaves in bulk follows from which of the two it is made of.

The cleanest way to see the consequence is to stop counting particles and start counting occupations of single-particle states, which is what the grand canonical ensemble is for.

Count states, not particles

Let the system exchange both energy and particles with a reservoir, fixed at temperature $T$ and chemical potential $μ$ . Each single-particle level of energy $ε$ can be filled independently, and the average number of particles in it is

⟨ n ⟩ = \frac{1}{e ^{β (ε - μ)} \mp 1},

the upper sign for bosons, the lower for fermions. One sign. That is the whole difference between the two families of matter, and it comes from whether the sum over occupations of a level runs over all non-negative integers or stops at one.

Definition

The two distributions

With the minus sign this is the Bose-Einstein distribution; with the plus sign, the Fermi-Dirac. The plus sign keeps $⟨ n ⟩$ between $0$ and $1$ , which is the Pauli exclusion principle appearing not as an extra rule but as a property of the counting.

What the sign does

For fermions, no level ever holds more than one particle on average. As the temperature drops, particles fill the lowest levels up to $μ$ , which at $T = 0$ is called the Fermi energy, and stack no higher. This filled sea is why metals conduct, why white dwarfs do not collapse, and why the periodic table has the shape it does.

For bosons, there is no such limit, and at low temperature they do the opposite: a macroscopic fraction of them piles into the single lowest state. That is Bose-Einstein condensation, a phase transition driven by statistics alone, with no interaction needed.

The classical limit

When the gas is dilute and hot, $e^{β (ε - μ)} ≫ 1$ , the $\mp 1$ is negligible against it, and both distributions collapse to the same thing:

⟨ n ⟩ \approx e^{- β (ε - μ)},

the Maxwell-Boltzmann result. The two quantum statistics agree with the classical one and with each other exactly when the levels are so sparsely filled that whether two particles could share a state never comes up.

Consequence

When statistics matters

The quantum effects turn on when the thermal de Broglie wavelength becomes comparable to the spacing between particles, so the wavefunctions overlap and the symmetry of the swap starts to bite. Cold and dense pushes you into the quantum regime; hot and dilute pulls you back to the classical one. Room-temperature air is firmly classical; the electrons in a metal, at the same temperature, are not.

So bosons and fermions are the same theory with one sign flipped, and that flip is enough to separate the matter that condenses from the matter that resists compression to the end.