1
$\begingroup$

Hello,

Given a $\mathcal{C}^\infty$ function $\varphi$ defined on a portion of a surface $\Sigma^-$ and let $\Sigma$ be a closed surface or union of surfaces bounding a compact volume $\Omega \subset \mathbb{R}^3$ such that $\Sigma^- \subset \Sigma$. When is it possible to extend the function $\varphi$ to the whole domain $\Omega$ such that the following condition is satisfied:

$\varphi$ is $\mathcal{C}^\infty$ on $\Omega$ and $\frac{\partial \varphi}{\partial {\mathbf n}}=0$ on $\Sigma$ with ${\mathbf n}$ the normal vector to $\Sigma$?

Of course I assume that $\Sigma^-$ and $\Sigma$ are $\mathcal{C}^\infty$ 2-dim real manifolds.

$\endgroup$
3
  • $\begingroup$ The main issue is extending to $\Sigma$: if $\Sigma^-$ is relatively open in $\Sigma$, smooth functions do not necessarily extend smoothly. Once you extended to $\Sigma$ you can just take a tubular neighborhood of $\Sigma$ and smoothly cut off $\psi$ in some epsilon distance to $\Sigma$. $\endgroup$ Jan 17, 2013 at 9:18
  • $\begingroup$ @Willie: thank you for the comment. $\endgroup$
    – user16974
    Jan 17, 2013 at 9:42
  • $\begingroup$ @Willie: If we know that $\Sigma^_$ is relatively closed in $\Sigma$ does this solve the problem? $\endgroup$
    – user16974
    Jan 17, 2013 at 10:28

0

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.