# Conditions that ensure that the boundary of an open set has measure zero

Are there some simple conditions which would ensure that the boundary of an open set in $\mathbb{R}^n$ has measure zero? Also, is it true that the boundary of a closed set in $\mathbb{R}^n$ has measure zero?

-

The boundary of a closed or open set can have positive measure. In $\Bbb R$, for instance, let $C$ be a so-called fat Cantor set; then $C$ is closed and has empty interior, so it is its own boundary and the boundary of its complement.

I don’t immediately see a proof, but I suspect that the boundary of every regular open set in $\Bbb R^n$ has measure $0$; a set $U$ in a topological space $X$ is regular open if $U=\operatorname{int}_X\operatorname{cl}_XU$. Added: Incorrectly, as it appears; see LVK’s comment below.

-
Thanks; but are there any conditions under which the boundary does have measure zero? –  Vivek Sep 22 '12 at 4:37
A Jordan curve of positive measure divides the plane in two regular open sets. –  user31373 Sep 22 '12 at 5:03

Here is a sufficient geometric condition for a general set $E$ to have measure zero: for every $x\in E$ there is $c>0$ such that for all sufficiently small $r>0$ the $r$-neighborhood of $x$ contains a ball of radius $cr$ that is disjoint from $E$. This is a weak form of condition known as porosity. To see that it implies having measure zero, use the Lebesgue density theorem.

Porosity is easy to verify for the boundary of a given open set: it suffices to find, for every boundary point, a subset in the shape of a cone (possible twisted) with a vertex at that point. The shape and size are allowed to depend on the point. Smooth, Lipschitz and uniform domains are covered by this condition.

-