I'm trying to prove the following theorem:
Let $H$ be a Hilbert space and $C\subseteq H$ a convex, norm-closed cone. Let $\xi \in H$. There is a unique $\eta \in C$ such that $\|\xi-\eta\| = d(\xi, C)$ (= the minimal distance of $\xi$ to $C$). Moreover, $\xi-\eta \perp \eta$.
I can show the first part of the theorem, namely that $\eta \in C$ exists with $\|\xi-\eta\| = d(\xi, C)$. I am not able to show that $\xi-\eta \perp \eta$ though, i.e. that $$0 = \langle \xi-\eta, \eta\rangle \iff \langle \xi,\eta\rangle =\|\eta\|^2.$$ Any hints or answers for this will be highly appreciated!
If you consider cones inside complex Hilbert spaces (which is often not done because of pathologies like this), then the result is not necessarily true. For example, if $H=\mathbb C$ (as complex Hilbert space with inner product $\langle z,w\rangle=\overline z w$), $C=i\mathbb R_+$ and $\xi=1+i$, then $\eta=i$ and $\langle \xi-\eta,\eta\rangle=i\neq 0$.