0

Let $\mathscr{H}$ be a Hilbert space. If $\{e_{\alpha}\}_{\alpha \in I}$ is a Hilbert (orthonormal) basis, one can write every element $\psi \in \mathscr{H}$ as: \begin{eqnarray} \psi = \sum_{\alpha \in I}\langle e_{\alpha}, \psi\rangle e_{\alpha}. \tag{1}\label{1} \end{eqnarray} In quantum mechanics, one finds a more general form of this expression as follows. If $|x\rangle$ denotes an eigenvector of the position operator $\hat{x}$, then we use the set $\{|x\rangle\}_{x\in \mathbb{R}^{3}}$ as a basis to write: \begin{eqnarray} |\psi\rangle = \int dx\psi(x) |x\rangle \tag{2}\label{2} \end{eqnarray} where $\psi(x) := \langle x| \psi\rangle$. The simple generalization from (\ref{1}) to (\ref{2}) is non-rigorous, and although it works just fine, it has to be better explained in terms of mathematical rigor. However, I'm aware that expression (\ref{2}) can be obtained in rigorous mathematical terms by using either the spectral theorem for unbounded linear operators on Hilbert spaces (first developted by J. von Neumann) or by using the so-called Gelfand triple or Rigged Hilbert spaces.

Question: Can someone explain how to obtain expression (\ref{2}) rigorously using either one of the above approaches (spectral theorem or Rigged Hilbert spaces)?

MathMath
  • 1,113
  • 1
  • The expansion (2) cannot be explained by con Neumann's spectral theory. It is rigorously formalized within Gelfand's rigged Hilbert spaces theory. – Valter Moretti Jun 13 '21 at 17:52
  • @ValterMoretti could you elaborate a little more on how one can formulate (\ref{2}) within the Gelfand's rigged Hilbert space theory or point a nice reference on the subject? I know a few references but I've never found a general discussion on the topic, where (\ref{2}) is explicitly obtained; these references I know discuss the problem for particular choices of potentials. I'm a big fan of your work and your books so I'd be very happy to hear some comments from you! – MathMath Jun 13 '21 at 18:16
  • There are some volumes by Gelfand and Vilenkin, with the title "generalized functions" or something similar. In those volumes you find the theory of rigged Hilbert spaces. Some technical results therein relies on another book by Maurin on topological vector spaces. I do not like very much that approach, so it is difficult to write something about it. However you find several posts on these subjects in PSE , just look for them. – Valter Moretti Jun 13 '21 at 19:04
  • @ValterMoretti sorry for the late response. When you say "I do not like very much that approach", what exactly do you mean? You don't like the use of rigged Hilbert spaces to QM? Or the particular approach of cited authors? Is there anything wrong with the use of rigged Hilbert spaces in this context? – MathMath Jul 05 '21 at 18:15
  • 1
    @MathMath Nothing wrong, the point is that the approach based on rigged Hilbert spaces formalism is a bit complicated when one wants to use it rigorously. – Valter Moretti Jul 06 '21 at 05:39

0 Answers0