The Lane-Emden equation is a classical nonlinear partial differential equation that shows up in astrophysics and geometry. Researchers have long studied its positive solutions, particularly whether stable ones must be radially symmetric, meaning they look the same in every direction from the origin. A conjecture by Chan and Wei said that in a certain range of parameters called the first Joseph-Lundgren interval, any stable positive solution defined on all of space must be radially symmetric. This paper disproves that conjecture by constructing explicit counterexamples.
The construction works by building a solution on a sphere that is not constant, meaning it varies depending on direction. This is done by carefully gluing together two pieces: a spherical cap region near a pole and a narrow neck region, matched smoothly at their boundary. The authors then check that this solution is genuinely stable, which requires a careful analysis of a particular eigenvalue associated with small perturbations. Showing that this eigenvalue stays strictly negative (indicating stability) requires very precise asymptotic calculations.
Once the authors have a stable but geometrically singular object called a cone, they use a rescaling argument together with a classical tool called the minimal solution to smooth out the singularity. The final result is a smooth, stable, positive solution defined on all of space that is demonstrably not radially symmetric. This is the first known example of such a solution for this type of equation, and it settles the Chan-Wei conjecture in the negative, opening new questions about what geometric or analytical conditions might actually force symmetry in related problems.