2
$\begingroup$

I read in the book the following argument. Consider $\Omega$ be an open set in $\mathbb{R}^n$ and $u$ be the solution of Schrödinger equation $iu_t+\Delta u = 0$ in $\Omega \times [0,T]$. We assume the Dirichlet boundary condition $u|_{\partial\Omega} = 0$ and $\frac{\partial u}{\partial \nu} = 0$ on $\Gamma$ where $\Gamma \subset \partial \Omega$ and $\nu$ be the unit normal vector.

Then they said that using Holmgren's uniqueness theorem and unique continuation of Schrodinger equation to imply $u=0$ on $\Omega$. I do not understand this argument. Can anyone help me to explain it?

$\endgroup$
1
$\begingroup$

I do not bother to write down all technical details. Here a sketch of the idea (at least in the static case):

You need to extend your domain $\Omega$ to a slightly bigger domain $\tilde{\Omega}$ near a point $x \in \Gamma$ and consider the zero extension of a solution $u$ in $\tilde{\Omega}$. Let us denote this extension by $\tilde{u}$. Then you have that $\tilde{u}$ solves the same equation in $\tilde{\Omega}$. (Here you apply the boundary conditions to see that $\tilde{u} \in H^2(\tilde{\Omega})$.) Clearly $\tilde{u}$ vanish in a neighbourhood contained in $\tilde{\Omega}\setminus \Omega$. This implies by the weak UCP that $\tilde{u}$ must be identically zero. Thus also $u$.

The weak UCP follows quite easily from Holmgren's uniqueness theorem by considering the set \begin{equation}\{\, x \in \Omega\,|\,u = 0 \text{ in a neighbourhood of $x$ } \}\end{equation} for a solution $u$ of $\Delta u = 0$ in $\Omega$. It is a nonempty, open and closed set. To see that it is closed you need to apply Holmgren's uniqueness theorem.

I believe that the above should also remain true for $H^1$ weak solutions. It is also valid for a quite general class of elliptic second order PDEs (in the above one could just use directly analyticity of harmonic functions!).

I guess with the time domain you extend by zero in $\tilde{\Omega} \times [0,T]$ anyway, and then use the weak UCP for $i\partial_t + \Delta$? (I must admit that I am only familiar with the UCP for elliptic second order PDEs.) Is this any helpful?

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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