The one-phase Bernoulli problem is a classical question in mathematics about the shape of a region where some physical quantity, like a fluid potential or temperature, satisfies certain rules inside the region and meets specific boundary conditions at its edge. The free boundary is the unknown surface separating the region where the quantity is active from the region where it is zero. A central challenge in this field is understanding what these free boundaries can look like, especially in higher dimensions, and whether stable solutions, meaning those that are locally optimal and resistant to small perturbations, must have a particularly simple geometric form.
The paper proves that in four-dimensional space, every stable global solution to this problem must be "one-dimensional," meaning it only varies in one direction. As a consequence, the free boundary, if it exists at all, can only be one or two parallel flat hyperplanes. This confirms a long-standing conjecture in the field, often called the stability conjecture for the Bernoulli problem, for dimension four. The result mirrors similar theorems that exist for related problems like the Allen-Cahn equation, and it fills in an important case in a broader program of understanding when geometric variational problems force solutions to be simple and symmetric.
Beyond settling the conjecture, the authors derive two additional results as consequences. First, they obtain sharp quantitative estimates on how curved the free boundary and the solution itself can be for stable solutions in four dimensions, which are useful tools for future analysis. Second, they show that solutions with finite Morse index, a broader class than strictly stable solutions that allows for a controlled number of unstable directions, must still be axially symmetric in four dimensions. Together, these results significantly advance the classification of solutions to the Bernoulli problem and provide new analytical tools applicable to similar free boundary questions.