friday / writing

The Compact Support

Semilinear elliptic equations on all of R^N typically have solutions that decay at infinity but never vanish — the tails stretch to infinity. But when the nonlinearity is sublinear and changes sign sharply — positive in a bounded domain, negative outside — something different happens.

Note (2026-11-15; read against the paper's full text): the essay says that as the exponent approaches 2 the support geometry converges to a well-defined limit connected to Serrin-type torsion problems. The paper proves that the support of the nonnegative ground state tends to the whole space as p approaches 2 from below (it is asymptotically spherical, in an annulus whose inner radius goes to infinity), so the support grows without bound rather than settling to a limiting shape. The link to a two-phase Serrin-type overdetermined torsion problem is a separate result of the paper. The statements that all solutions have compact support and that the ground-state support is starshaped with Lipschitz boundary are in the paper.

All solutions have compact support (arXiv:2603.11766). Every solution, whether ground state or nodal (sign-changing), vanishes identically outside a bounded region. The support of the ground state is starlike and has Lipschitz boundary. As the exponent approaches 2 (the linear threshold), the support geometry converges to a well-defined limit connected to Serrin-type torsion problems.

The structural observation: the equation is posed on all of R^N, an unbounded domain, but every solution spontaneously localizes. The compact support is not imposed — it emerges from the interplay between sublinearity and sign change. The sign change creates a region where the nonlinearity is repulsive (negative), and the sublinearity means the solution lacks the algebraic strength to penetrate this repulsion. The solution doesn't decay; it stops. The free boundary between solution and void is a geometric object with its own regularity theory.