Bosons, Fermions & Quantum Matter

In everyday life no two objects are truly identical, and even two perfect copies could be told apart by following each one continuously along its path: in classical physics, a trajectory is a name tag. Quantum mechanics removes the tag. Every electron has exactly the same mass, charge and spin, and because an electron is a spreading wavefunction rather than a sharp path, once two electron wave packets overlap there is no fact of the matter about which one emerges where. Such particles are indistinguishable: not merely hard to tell apart, but impossible to tell apart even in principle.

That has a sharp mathematical consequence. Write the joint wavefunction of two identical particles as ψ(x₁, x₂), the amplitude for finding one particle at x₁ and the other at x₂. Since the labels carry no meaning, swapping them must leave every probability unchanged: |ψ(x₂, x₁)|² must equal |ψ(x₁, x₂)|². The swap can therefore only multiply the wavefunction by a constant factor of size one. In our three-dimensional world, swapping twice is the same as doing nothing, so that factor squared must equal one, leaving only plus one and minus one. (Only in strictly two-dimensional systems does this step fail, which permits exotic quasiparticles called anyons.) This is exchange symmetry: a wavefunction that keeps its sign under a swap is symmetric; one that flips its sign is antisymmetric.

Is this real, or bookkeeping? In 1987 Chung Ki Hong, Zhe-Yu Ou and Leonard Mandel at the University of Rochester sent pairs of identical photons into the two input faces of a fifty-fifty beam splitter at the same instant. Distinguishable particles would leave through different exits half the time; instead, the photons left together. The two histories that put one photon in each exit, both transmitted or both reflected, end in the same arrangement, so their amplitudes add, and for identical photons they are equal and opposite and cancel exactly. This vanishing of one-photon-in-each-exit coincidences, the Hong–Ou–Mandel dip, is now a routine test of indistinguishability.

Particles with antisymmetric joint wavefunctions are fermions, after Enrico Fermi; those with symmetric ones are bosons, after Satyendra Nath Bose. Fermions include electrons, protons, neutrons, quarks and neutrinos; bosons include the photon, the gluon, the W and Z particles and the Higgs boson. The deciding property is spin, the intrinsic angular momentum of Chapter 17: half-integer spin (one-half, three-halves and so on, in units of ħ) always means fermion, and integer spin always means boson. This spin-statistics connection was proved by Markus Fierz in 1939 and Wolfgang Pauli in 1940 from relativistic quantum field theory. Richard Feynman lamented that no elementary explanation existed, and none is universally accepted today.

To see what the sign does, put one particle in single-particle state a and the other in state b. Saying 'particle 1 is in a' would smuggle in a label, so quantum mechanics uses both orderings at once, added for bosons and subtracted for fermions, then divided by √2 for normalisation. Now set a = b. The boson combination doubles, but the fermion combination becomes something minus itself: zero, a state that cannot occur. That is the Pauli exclusion principle of Chapter 18, first stated in 1925 as a rule for atomic spectra, revealed as a direct consequence of antisymmetry. The plus sign does the reverse, making shared states more likely for bosons.

Composite particles follow the same rule through their total spin: an even number of constituent fermions makes a boson and an odd number makes a fermion, as long as nothing probes hard enough to break the composite apart. A helium-4 atom holds two protons, two neutrons and two electrons, six fermions in all, so it is a boson; helium-3, with one neutron fewer, is a fermion. The two atoms are almost exactly the same size and chemically identical, yet near absolute zero they behave utterly differently, as Part 7 will show.

Statistical physics asks how many particles, on average, occupy each single-particle state at temperature T; that average is the occupation number. For distinguishable classical particles the answer is the Maxwell–Boltzmann distribution: occupation falls by a factor of e, about 2.718, for every extra k_B T of energy above the chemical potential μ, where k_B is Boltzmann's constant and μ is a reference energy that adjusts so the occupations add up to the right number of particles. Fermions instead obey the Fermi–Dirac distribution, found independently by Fermi and by Paul Dirac in 1926: one divided by an exponential plus one. Bosons obey the Bose–Einstein distribution, the same with minus one. When occupations are small all three agree; quantum statistics shows its hand only when particles are crowded or cold.

Because of the plus one, Fermi–Dirac occupation can never exceed one: exclusion in statistical clothing. At absolute zero it becomes a perfect step, with every state below a sharp energy full and every state above it empty. That boundary is the Fermi energy, E_F. In copper each atom donates one electron to a shared pool, and those electrons form a Fermi sea filled to about 7.0 eV. Room temperature supplies only about 0.026 eV; k_B T would match copper's Fermi energy only near 82,000 kelvin. Electrons at the top of the sea move at about 1.6×10⁶ metres per second even at absolute zero, not because they are hot but because every slower state is taken.

Arnold Sommerfeld's free-electron model of 1927 and 1928 used this picture to solve a classical puzzle. A gas of free electrons should soak up heat, yet metals showed almost no electronic heat capacity. In the Fermi sea, an electron far below the surface cannot accept a small thermal kick, because every state the kick could reach is occupied; only electrons within a few k_B T of the Fermi energy respond, and the electronic heat capacity is about one percent of the classical prediction. The minus one of Bose–Einstein statistics does the opposite: as a state's energy approaches μ, its occupation grows without limit.

In 1924 Satyendra Nath Bose, at the University of Dacca, rederived Planck's blackbody law of Chapter 3 by treating photons of the same frequency as truly indistinguishable, so that only how many occupy each state matters, never which. He mailed the paper to Albert Einstein, who translated it into German, arranged its publication and applied the idea to atoms later that year. In a second paper, early in 1925, Einstein found that, for an ideal gas with a fixed number of bosons, below a critical temperature the excited states cannot hold all the particles, and the surplus drops into the single lowest state. This is Bose–Einstein condensation, and it needs no attractive force: it is pure statistics.

When does it happen? Every particle has a de Broglie wavelength (Chapter 6), and for an atom in a gas the typical value is the thermal wavelength, λ = h/√(2π m k_B T), which grows as the gas cools. At room temperature a rubidium atom's thermal wavelength is about 0.01 nanometres, smaller than the atom itself. Condensation begins when the wavelength becomes comparable to the spacing between atoms, so neighbouring wave packets overlap and identity takes command. Einstein's criterion for a uniform ideal gas is that a cube one thermal wavelength on a side must hold more than about 2.612 atoms.

Reaching that before the gas clumps into a liquid or solid took seventy years. Laser cooling, honoured by the 1997 Nobel Prize, slows atoms with the kicks of absorbed photons; a magnetic trap then holds them while evaporative cooling lets the most energetic atoms escape, like steam leaving coffee. In June 1995 at JILA in Boulder, Colorado, Eric Cornell and Carl Wieman saw a rubidium-87 condensate appear below about 170 nanokelvin, then isolated a pure condensate of about 2,000 atoms near 20 nanokelvin. Months later Wolfgang Ketterle's group at MIT made sodium condensates with hundreds of times more atoms, and in 1997 showed that two overlapping condensates interfere like waves. The three shared the 2001 Nobel Prize in Physics.

Photons are bosons, and the laser is boson statistics put to work. In 1917 Einstein showed that an excited atom can emit in two ways. Spontaneous emission releases a photon at a random moment in a random direction. Stimulated emission is triggered by a passing photon of the right frequency and produces an exact copy: same frequency, direction, polarisation and phase, which together define one mode of the light field. The probability of emitting into a mode already holding n photons is proportional to n + 1, the one spontaneous and the n stimulated. That is Bose enhancement: the fuller a mode, the more it attracts photons.

A laser needs three ingredients. The first is a gain medium, such as a ruby crystal, helium-neon gas or a semiconductor junction. The second is a population inversion, more atoms in the upper level than the lower, produced by pumping with light or current; without it, absorption would swallow photons faster than stimulated emission makes them. The third is an optical cavity of facing mirrors that selects the modes allowed to grow. Charles Townes and colleagues built the microwave version, the maser, in 1954, and Theodore Maiman at Hughes Research Laboratories switched on the first laser, a ruby rod glowing at 694 nm, in May 1960.

So is a laser just a bright bulb? No. A 3000 kelvin filament emits thermal light that roughly follows Planck's law, which is the Bose–Einstein distribution with μ = 0, because photons are freely created and destroyed. Its photons spread across an enormous number of modes: at 600 nm each mode holds only about 3×10⁻⁴ photons on average, and bluer modes hold fewer still. A laser crowds its photons into one or a few modes, each holding vast numbers. That is why a milliwatt pointer, far weaker than a household bulb, throws a crisp spot across a lecture hall. The difference is not total power but order: coherence, one colour, one direction.

In 1911 Heike Kamerlingh Onnes in Leiden, the first person to liquefy helium, used it to cool mercury and watched the metal's resistance plunge to zero at about 4.2 kelvin: not small, but zero within every measurement since. In 1933 Walther Meissner and Robert Ochsenfeld found something deeper. A superconductor cooled in a weak magnetic field actively expels the field from its interior, apart from a surface layer typically tens of nanometres deep. A merely perfect conductor would instead trap any field already inside. The Meissner effect marks superconductivity as a genuinely new state of matter.

How can electrons, which are fermions, flow without friction? John Bardeen, Leon Cooper and Robert Schrieffer answered in 1957, earning the 1972 Nobel Prize: the electrons stop acting alone. A moving electron tugs nearby positive ions toward itself, leaving a fleeting ripple of positive charge that attracts a second electron. Cooper had shown in 1956 that, thanks to the filled Fermi sea beneath them, even this weak attraction binds two electrons near the Fermi energy into a Cooper pair, usually with opposite momenta and spins. With integer total spin, pairs have bosonic character, and below the critical temperature they settle collectively into one coherent quantum state. Breaking a pair costs a minimum energy, the superconducting gap, which ordinary jostling cannot pay, so current flows without dissipation.

One honest caveat: a Cooper pair in aluminium is over a micrometre across, overlapping vast numbers of other pairs, so the BCS state is a collective many-electron state rather than a textbook gas of tiny bosons. Yet pairing shows plainly in data. In 1961 Bascom Deaver and William Fairbank at Stanford, and independently Robert Doll and Martin Näbauer in Munich, found the magnetic flux trapped in tiny hollow superconducting cylinders quantised in units of h/2e, about 2.07×10⁻¹⁵ webers, as expected for carriers of charge 2e. Superconducting niobium-titanium coils now create the strong magnetic fields of most MRI scanners, and some copper oxides, a family discovered by Georg Bednorz and Alex Müller in 1986, superconduct above liquid nitrogen's 77 kelvin. As of 2026, how those copper oxides pair is still debated.

Helium stays liquid at ordinary pressure all the way to absolute zero; as Chapter 13 explained, zero-point motion keeps its light, weakly attracted atoms from freezing. Below 2.17 kelvin, the lambda point, liquid helium-4 becomes a superfluid, as found in 1937 by Pyotr Kapitsa in Moscow and independently by John Allen and Don Misener in Cambridge, whose papers appeared side by side in Nature in January 1938. A superfluid slips through channels far narrower than a hair with no measurable viscosity, creeps up and over its container's walls as a thin film, and spurts from a heated tube in the 'fountain effect'. Set rotating, it forms quantised vortices whose circulation comes only in multiples of h/m, where m is one atom's mass.

In 1938 Fritz London proposed that superfluidity is Bose–Einstein condensation in disguise: an ideal Bose gas at liquid helium-4's density would condense near 3.1 kelvin, impressively close to the real 2.17 kelvin. The atoms' strong interactions matter, though; neutron scattering indicates that even near absolute zero fewer than one atom in ten occupies the zero-momentum state, yet the whole liquid flows as a superfluid. The most telling evidence that statistics is at work is helium-3, a fermion, for which nothing happens at 2.17 kelvin. Only after its atoms pair up, much like electrons in a superconductor, does it become superfluid, at one to two and a half thousandths of a kelvin depending on pressure, as Douglas Osheroff, Robert Richardson and David Lee discovered at Cornell in 1971 and 1972, winning the 1996 Nobel Prize.

Now admire the pattern. Fermions, forbidden to share, build the scaffolding of the world: atomic shells, the Fermi sea of metals, the degeneracy pressure of white dwarfs. Bosons, encouraged to share, make a single quantum state macroscopic: the laser beam, the condensate, the superfluid. Fermions can join in by pairing. And these marvels are working technology: the superconducting magnets of CERN's Large Hadron Collider are bathed in superfluid helium at 1.9 kelvin, because it carries heat away better than any ordinary coolant.