Let us consider the unit ball $B^n$ of $\mathbf{R}^n$ and $K = B^{n-1}(0,1/2) \times \{0\}$ the ball of dimension $n-1$ and radius $1/2$ lying on the equator.

We wish to solve the Dirichlet problem (for harmonic functions) on $B^n \setminus K$ with boundary value $f = 0$ on $K$, and $f$ (let's say) is smooth on $\partial B^n$. Such a solution exists, is unique and we can find it within the class of Sobolev function.

My question is how this solution behaves near $K$. Is it Hölder continuous ? What is the best exponent of Hölder continuity one can expect ?

More generally, are there known criterias for the regularity at the boundary (not abstract continuity like with the Wiener criterion, but with a control of the modulus of continuity) ?


  • $\begingroup$ $K$ is not contained in the boundary of $B^n$; it does not even touch the boundary. $\endgroup$
    – Florian
    Nov 4, 2011 at 9:38
  • $\begingroup$ @Florian : yes, the Dirichlet problem is solved in the domain $B^n \setminus K$. The solution is is harmonic on $B^n \setminus K$ and continuous on $\overline{B^n}$. $\endgroup$
    – vizietto
    Nov 4, 2011 at 9:52

2 Answers 2


Far from the rim of the inner disk one has a good regularity. At the rim though, I think you can get the solution behaviour from explicit constructions of potentials induced by a charged ellipsoid and collapsing one of axes, e.g., as given in Kellogg's book.

  • 3
    $\begingroup$ Marius Mitrea has a bunch of papers on the regularity of the Dirichlet problem on manifolds. $\endgroup$ Nov 25, 2011 at 16:21

Grisvard's book has an extensive discussion of 2d elliptic problems with corners. Your problem is singular at the rim, and the singularity there is essentially the same one as a 2d problem with a 360 degree corner.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.