The Statistical Distribution

Particles in a gas do not all sit at the same energy. They spread out. Some occupy low-energy states; some climb to high-energy states. The question of how they distribute themselves among the available energy levels is the central problem of statistical mechanics. And the answer depends on what kind of particle you have.

There are only two types. The universe has been generous, in this regard. Fermions obey the Fermi-Dirac distribution. Bosons obey the Bose-Einstein distribution. That is all.

The Fermi-Dirac distribution tells you the expected number of fermions in a single-particle state of energy $\epsilon$:

$$f_{FD}(\epsilon) = \frac{1}{e^{(\epsilon - \mu)/k_B T} + 1}$$

Here $\mu$ is the chemical potential — at zero temperature, it is the Fermi energy. The denominator contains a plus one. That plus one is the fingerprint of the Pauli exclusion principle. It ensures that no state can hold more than one particle. When $T \to 0$, the distribution becomes a step function: everything below $\mu$ is occupied ($f = 1$), everything above is empty ($f = 0$). The distribution is sharp, absolute, binary. At finite temperature, the step softens. Only particles within about $k_B T$ of $\mu$ see a nonzero probability of transition. The rest remain frozen.

The Bose-Einstein distribution is formally identical but for one crucial sign:

$$f_{BE}(\epsilon) = \frac{1}{e^{(\epsilon - \mu)/k_B T} - 1}$$

The minus one in the denominator. This allows the occupation number to diverge. When $\epsilon$ approaches $\mu$ from above, the occupation goes to infinity. This is Bose-Einstein condensation: a macroscopic number of particles piling into the ground state. The bosons do not mind sharing. They pile up. They love it.

The minus one is what distinguishes bosons from fermions in every statistical prediction. It is the mathematical expression of a simple fact: bosons have no prohibition against occupying the same state. Two, a hundred, a billion — they all fit. This is why a laser works. Photons — bosons — flood the same mode, same frequency, same polarization, same direction. The Fermi-Dirac distribution would forbid this. The Bose-Einstein distribution demands it. The occupation grows because each additional boson in a mode makes it easier for the next one to join. The rate of absorption is enhanced by a factor of $(n + 1)$, where $n$ is the current occupation. Stimulated emission. The mathematics of the distribution is the mathematics of the amplification.

At high temperature and low density — when $e^{(\epsilon - \mu)/k_B T} \gg 1$ for all relevant states — the $\pm 1$ in the denominator is negligible, and both distributions converge to the Maxwell-Boltzmann distribution:

$$f_{MB}(\epsilon) \approx e^{-(\epsilon - \mu)/k_B T}$$

This is the classical limit. It is valid when the average inter-particle spacing is much larger than the thermal de Broglie wavelength. Under these conditions, the wavefunctions of individual particles barely overlap, and quantum statistics becomes irrelevant. The particles are effectively distinguishable, and classical probability rules apply.

But at low temperature and high density — conditions found in white dwarfs, in neutron stars, in liquid helium — the quantum distributions diverge dramatically from their classical cousin. Fermions fill up to the Fermi energy, creating a pressure that supports stars against gravity. Bosons condense into the ground state, flowing without friction, defying viscosity, exhibiting quantum phenomena on macroscopic scales. The same two formulas, applied under different conditions, produce entirely different universes.

The chemical potential $\mu$ is the knob that controls the distribution. For fermions at $T = 0$, $\mu$ is the Fermi energy. For photons (black-body radiation), $\mu = 0$ because photon number is not conserved. For a Bose gas in a trap, $\mu$ approaches the ground-state energy from below as the temperature drops and condensation begins. The chemical potential is the thermodynamic price of adding one more particle, and its behavior tells you whether the system is fermionic or bosonic, classical or quantum.

There is a third distribution, sometimes forgotten, that interpolates between the two. The Gentile distribution allows a state to hold up to $n$ particles, where $n$ is a parameter. When $n = 1$, it is Fermi-Dirac. When $n \to \infty$, it is Bose-Einstein. The interpolation is continuous but unphysical for $1 < n < \infty$, and yet it reveals something subtle: the distinction between fermions and bosons is not binary in the mathematics, only in the implementation of nature.