MathOverflow is a question and answer site for professional mathematicians. Join them; it only takes a minute:

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

Arising as the traces of $H(div; \Omega)$, I am wondering if the space $H^{-1/2}(\partial \Omega)$ has any regularity properties? (Containment in BV would be wonderful, although I doubt it holds.) It seems difficult to find anything in the literature.

share|cite|improve this question
up vote 4 down vote accepted

Of course not. For instance, it contains $L^2(\partial\Omega)$. By definition, $H^{-1/2}(\partial\Omega)$ is the dual of $H^{1/2}(\partial\Omega)$. If $\Omega$ is $n$-dimensional $n\ge3$, then $\partial\Omega$ is $(n-1)$-dimensional and $H^{1/2}(\partial\Omega)\subset L^p(\partial\Omega)$ by Sobolev injection, for every $p\le\frac{2(n-1)}{n-3}$ ($<+\infty$ if $n=3$) and therefore $H^{-1/2}(\partial\Omega)$ contains $L^{p'}$. But you cannot say much more. If $s>\frac n2$, then $H^s(\Omega)\subset C^0(\overline\Omega)$ and therefore $H^{\frac12-s}(\partial\Omega)$ is contained in the set of bounded measures. But this does not apply to $H^{-\frac12}$, even if $n=2$ (thanks to A. Rekalo). Actually, a use of the Uniform Boundedness Principle tells us that there exists an element of $H^{-\frac12}(\partial\Omega)$ that is not a bounded measure.

share|cite|improve this answer
@Andrey. I acknowledge my mistake about continuity (I'll edit). But I maintain that $H^{-1/2}$ contains $L^2$, just because $L^2\subset H^{1/2}$. – Denis Serre Dec 3 '12 at 11:41
@Denis Serre: Sorry, my reading skills are horrible today. I misread your statement (as if you wrote that $H^{-1/2}$ is contained in $L^2$). Stupid me. – Andrey Rekalo Dec 3 '12 at 12:26
no problem. This happens to me too. And I know that it is easier to see two mistakes when there is at least one for sure. – Denis Serre Dec 3 '12 at 14:32

Ok, a further question: What if we have $div \phi \equiv 1$ on $\Omega$? (Or up to a set of measure $\epsilon$.) Can we then get extra regularity for the normal trace?

share|cite|improve this answer

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

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