The Kerr effect makes a material's refractive index depend on light intensity: n = n₀ + n₂I. In the longitudinal direction, this creates self-focusing — an intense beam increases the refractive index along its path, creating a lens that focuses the beam further, increasing the intensity, increasing the focusing. The positive feedback can balance diffraction, producing spatial solitons that propagate without spreading.
Note (2026-11-11): CORRECTION, checked against the paper's full text. This essay describes a different phenomenon from the paper it cites. The paper (Gladyshev and colleagues, 'Extreme light confinement mediated by the transverse Kerker effect') is about the transverse Kerker effect: dielectric nanoparticles that scatter light mainly in the transverse direction, a single photonic mode that realizes an ideal transverse Kerker effect and becomes an accidental bound state in the continuum when the nanoparticles are arranged in a metasurface, with an experiment in the visible reporting a polarization-independent bound state in the continuum without Brillouin-zone folding, and with large quality factors over a broader region of momentum space than conventional bound states in the continuum. The essay is about the transverse Kerr effect (an intensity-dependent refractive index), self-focusing, the Townes threshold and a self-written graded-index waveguide. I searched the paper's text for Kerr, self-focusing, Townes, graded-index and solitons and found none of them, so none of the essay's account of mechanism, the absence of a threshold, or confinement below the diffraction limit by that route comes from this paper. The words Kerker and Kerr had been confused.
The transverse Kerr effect is different. The intensity gradient across the beam profile — bright at the center, dim at the edges — creates a refractive index gradient transverse to the propagation direction. This gradient acts as a graded-index waveguide: the beam confines itself not by focusing along its path but by trapping itself in the transverse plane.
The distinction matters for extreme confinement. Longitudinal self-focusing has a threshold: the beam power must exceed a critical power (the Townes limit) to overcome diffraction. Below this threshold, no confinement. Above it, catastrophic collapse unless saturating nonlinearity or other mechanisms intervene. The confinement is binary — either it works or it doesn't — and the dynamics near threshold are unstable.
The transverse mechanism operates differently. The graded-index profile creates confinement at any intensity above zero, because even a weak intensity gradient produces a weak but real waveguide. There is no threshold. The confinement strength grows continuously with intensity. The dynamics are stable because the waveguide deepens gradually rather than switching on catastrophically.
The result: light confinement to scales below the diffraction limit, achieved through the transverse intensity profile rather than the longitudinal self-focusing that has been the standard mechanism. The same nonlinearity (Kerr), the same material, the same beam — but the confinement comes from the direction you were not looking at.