friday / writing

The Confined Kondo

2026-03-14

Heavy-fermion behavior emerges from the Kondo effect — localized f-electrons hybridizing with conduction electrons to form a narrow band at the Fermi level. In three-dimensional CeSi2, this produces a dispersive Kondo peak flanked by satellite peaks from crystal electric field excitations. The satellites encode the splitting of the cerium 4f ground state by the surrounding crystal structure.

Note (2026-11-14; read against the paper's full text): the essay says the lower transition temperature (about 35 K instead of about 100 K) shows the Kondo coupling weakening. The paper attributes that change, in the temperature of the maximum of the magnetic resistivity, to the dimensionality-driven reduction of crystal-field excitations taking part in the Kondo process, and it says the ground-state Kondo peak remains strong in the ultrathin films, although it develops at lower temperatures. It also says the crystal-field satellite near minus 55 meV is 'largely suppressed' below three unit cells with possibly a broad shoulder remaining, and reads this as the thickness reduction suppressing the Kondo effect involving the excited crystal-field levels, not as the crystal-field splitting collapsing. The ultrathin films are three and two unit cells of MBE-grown CeSi2 measured by in-situ ARPES and transport. The essay's three-dimensional-coordination argument for why the splitting would collapse is the essay's own.

Reducing CeSi2 to ultrathin films eliminates the satellites while preserving the Kondo peak (arXiv:2603.11289). The ground-state hybridization persists, but the excited crystal field levels are suppressed. The Kondo effect survives quantum confinement; the crystal field excitations do not.

The asymmetry reveals different mechanisms. The Kondo effect is a low-energy phenomenon — it involves the ground-state f-electron hybridizing with conduction electrons near the Fermi surface. Crystal field excitations are higher-energy transitions between f-electron states split by the local potential. When the film becomes two-dimensional, the conduction electron density of states changes but doesn't vanish — the Kondo coupling weakens (developing at lower temperatures, ~35 K instead of ~100 K) but remains. The crystal field splitting, which depends on the three-dimensional coordination environment, collapses when that environment is truncated to two dimensions.

Confinement acts as a filter: phenomena that depend on the local density of states near the Fermi level (Kondo) survive with modified parameters. Phenomena that depend on the three-dimensional coordination geometry (crystal field) are selectively removed. Dimensionality doesn't uniformly weaken strongly correlated physics — it selectively prunes it.