Landau's Fermi Liquid: A Field Note
The electrons in a metal want to talk to each other. Coulomb repulsion is strong. The spacing between electrons is small. The naive expectation — electrons as independent plane waves, the Sommerfeld model — should be catastrophically wrong. It works, empirically, and the reason it works is the greatest accidental insight of twentieth-century condensed matter physics.
Landau didn't discover this by calculation. He discovered it by assumption. He assumed that the interacting electron system, when you cool it down and turn off the interactions adiabatically, morphs smoothly into the non-interacting Fermi gas. No phase transition. No qualitative change. The electrons keep their identity. They just wear different clothes.
The clothes are called quasiparticles.
A quasiparticle is an electron plus its cloud of perturbations. It pushes on its neighbors, which push on theirs, and the whole collective response can be compressed into a single entity: a particle with the same charge as an electron, but with a different mass, a different g-factor, a different velocity. The effective mass $m^*$ can be 2, 10, 50 times the free electron mass. The quasiparticle is real in the sense that you can weigh it, track it, measure it. It is not real in the sense that it is an emergent object — it requires the medium to exist. Remove the lattice, remove the other electrons, and the quasiparticle dissolves.
Landau's theory rests on three postulates, deceptively simple:
Existence. An interacting Fermi system has a one-to-one correspondence with a non-interacting Fermi gas. Every eigenstate maps onto an eigenstate. The ground state is a filled Fermi sea. Excitations are particle-hole pairs near the Fermi surface.
Quasiparticle energy. The energy of a quasiparticle depends on the occupation of all other quasiparticles: $\varepsilon_{\mathbf{p}} = \varepsilon^0_{\mathbf{p}} + \sum_{\mathbf{p}'} f_{\mathbf{p}\mathbf{p}'} \delta n_{\mathbf{p}'}$. The function $f$ is the Landau interaction function — it encodes all the complexity of electron-electron interactions without ever writing down an explicit Hamiltonian.
Lifetime. Quasiparticles have a finite lifetime, but near the Fermi surface the lifetime grows as $\tau \sim 1/E^2$, where $E$ is the excitation energy above the Fermi level. At the Fermi surface itself ($E = 0$), the lifetime diverges. Quasiparticles become stable. This is why the Fermi liquid picture works at low temperature: the relevant excitations are close enough to the Fermi surface that they live long enough to be counted.
The consequence is breathtaking. A metal with strong Coulomb interactions behaves, for all thermodynamic and transport purposes, like a gas of weakly interacting particles with renormalized parameters. The specific heat coefficient $\gamma = C/T$ is enhanced by $m^*/m$. The Pauli susceptibility is enhanced by the same factor. The compressibility changes. Every measurable quantity is the free-electron answer multiplied by a Landau parameter.
But the theory has boundaries. It fails when correlations become strong enough to open a gap — the Mott insulator, the charge density wave. It fails when superconductivity emerges, because Cooper pairing rearranges the ground state entirely. It fails at the quantum critical point, where the effective mass diverges and the quasiparticle weight goes to zero.
What the Fermi liquid theory gives you is the baseline against which all anomalies are measured. It is the null hypothesis of interacting electron systems. If your material obeys Fermi liquid theory, nothing interesting is happening. If it doesn't — if $C/T$ diverges, if $\rho \sim T^{1.5}$ instead of $T^2$, if the Hall coefficient jumps discontinuously — then you have found something worth studying.
Fermi liquid theory is the art of making the complicated tractable without losing the truth.