Let $\Omega\subseteq\mathbb{R}^n$ be a domain and $\partial\Omega\in C^2$, i.e.
- $\Omega=\overline{\Omega}^\circ$
- For all $x_0\in\partial\Omega$, there exists a neighbourhood $U\subseteq\mathbb{R}^n$ of $x_0$ and $\phi\in C^2(U)$ such that $$\overline{\Omega}\cap U=\left\{x\in U:\phi(x)\le 0\right\}\tag{1}$$ and $$\nabla\phi(x)\ne 0\;\;\;\text{for all }x\in U\tag{2}$$
I want to show, that for all $x_0\in\partial\Omega$ it holds $$\overline{B}_r(y_0)\cap\overline{\Omega}=\left\{x_0\right\}\;\;\;\text{for some }y_0\in\mathbb{R}^n\setminus\overline{\Omega}\;\text{and }r>0\tag{3}$$
This should be trivial, shouldn't it? Let $x_0\in\partial\Omega$. Since $\partial\Omega\in C^2$, there is a $y_0\in\mathbb{R}^n$ and a $r>0$ such that $(1)$ and $(2)$ hold for $U=\overline{B}_r(x_0)$.
Now, I've got two problems
- I don't see why $\phi\in C^2(U)$ (instead of $\phi\in C^1(U)$) comes into play.
- It seems like that we need $\phi^{-1}\left((-\infty,0]\right)=\left\{x_0\right\}$, but how can we achieve that?