How do you prove basic facts on the Schrodinger equation using the spectral theorem? More precisely, here is what I have in mind.

The version of the Spectral Theorem I am familiar with is the following.

Spectral Theorem for normal operators: Let $H$ be a Hilbert space and $T\in \mathcal{B}(H)$ be a normal operator. Then there is a unique resolution $P$ of the identity of $H$ over $\sigma(T)$ (the spectrum of $T$) such that $$ T=\int_{\sigma(T)} \lambda\, dP. $$

I would like a theorem like the following.

Theorem: Let $\psi$ satisfy the 1-D Schrodinger equation $$\left(-\frac{\hbar^2}{2m} \frac{\partial^2}{\partial x^2} +u(x)\right)\psi=i\hbar \frac{\partial}{\partial t}\psi$$ where $u$ is some "nicely behaved" function, and let $f=\psi(x,0)$. Then $$(1)\qquad f=\sum_{\lambda} c_{\lambda}\phi_{\lambda}+\int_{\lambda} c_{\lambda}\phi_{\lambda}$$ where the first sum is over the discrete spectrum, the integral is over the continuous spectrum, $\phi_{\lambda}$ is the eigenfunction for $\lambda$, and $c_{\lambda}=\int f\overline{\phi_{\lambda}}\,dx$. Therefore, $$\psi=\sum_{\lambda} c_{\lambda}\phi_{\lambda}e^{-\frac{i\lambda}{\hbar}t}+\int_{\lambda} c_{\lambda}\phi_{\lambda}e^{-\frac{i\lambda}{\hbar}t}.$$

What is the mathematically correct way to state the above, and how can it be proved?

For instance, if $u_{\infty}=\lim_{x\to \pm \infty} u(x)$, for generic $u$, the discrete spectrum would consist of some finite set in $(\min u, u_{\infty})$ and the continuous spectrum would be $(u_{\infty},\infty)$.

Two issues are the following.

  1. The Spectral Theorem as stated above only works for bounded operators, and the operator is not bounded. (We probably also need growth conditions on $f$.) I know there is a Spectral Theorem for unbounded operators but it seems to require more technicalities; if that is the way to go, how are those technicalities satisfied in this case?
  2. The theorem doesn't describe what the measure $dP$ is explicitly, so how do I know that the resolution of the identity gives (1)?

closed as off-topic by Christian Remling, Matt F., RP_, Peter Humphries, Stefan Waldmann Apr 23 '17 at 17:43

This question appears to be off-topic. The users who voted to close gave this specific reason:

  • "MathOverflow is for mathematicians to ask each other questions about their research. See Math.StackExchange to ask general questions in mathematics." – Christian Remling, RP_, Peter Humphries, Stefan Waldmann
If this question can be reworded to fit the rules in the help center, please edit the question.

  • $\begingroup$ For the spectral theory of unbounded (normal) operators, the key word is "Cayley transform". For the definition of the spectral measure $dP$, the key word is "functional calculus". Both can be found in Conway's Functional Analysis. $\endgroup$ – Fan Zheng Oct 19 '15 at 2:46

There is a spectral theorem for unbounded operators. It is not simple. But if you have an unbounded operator, and want to use the spectral theorem, you have to use it:-) The difficult thing is to actually define the operator (using the boundary conditions) and prove that it is self-adjoint.

Verification of all facts you mentioned is not simple, but there are many good books on this, some of them are:

MR1186643 Berezin, F. A.; Shubin, M. A. The Schrödinger equation. Kluwer , Dordrecht, 1991. xviii+555.

Reed and Simon, Methods of modern mathematical physics.


At first you need to understand the spectral theorem for self-adjoint unbounded operators. There are many good textbooks: Yosida, Kato, Reed & Simon, Lax. Think of (1) as solution by separation of variables. But, there is no easy way to make it rigorous!


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