3
$\begingroup$

I have a possibly simple question regarding estimating bounds on spectral projection subspace.

Let $H_n$ be a sequence of bounded self-adjoint operators on $\ell^2(\mathbb{Z}^2)$ converging in the strong operator topology to $H_\infty$. I assume that $\dim \big(\text{Im}\chi_{(a-\epsilon,a+\epsilon)}(H_n) \big)=m_n\in \mathbb{N}$ for all $n$. It seems that this should imply that $\dim \Big(\text{Im}\big(\chi_{[a-\epsilon,a+\epsilon]}(H_\infty) \big) \Big)\geq \liminf m_n$. I would also like to say that if $m_n\to \infty$, then $\sigma_{ess}(H_\infty)\cap [a-\epsilon,a+\epsilon]\neq \emptyset$.

Is this result or something of the sort a known result or something that anyone thinks should hold? I want to check whether I can deduce a spectral type statement about the limit in the strong operator topology.

$\endgroup$

2 Answers 2

5
$\begingroup$

Unfortunately, no. Let $H_n$ be the orthogonal projection onto $\mathbb{C}\cdot e_n$. Then $H_n \to 0$ strongly, but the spectral projection $\chi_{\{1\}}(H_n)$ has dimension $1$ for all $n$.

The second question fails too; let $H_n$ be the spectral projection onto the span of $\{e_n, \ldots, e_{2n}\}$.

$\endgroup$
4
$\begingroup$

This is false. Take $H_n=P_n$ as the projection onto $\ell^2(\{ x: |x|\le n\})$. Then $P_n\to 1$ strongly, $\dim \chi_{\{ 0\} }(P_n)=\infty$, but $\sigma (1) =\{ 1\}$.

What you have in this situation is the opposite semi-continuity: if $H_n\to H$ strongly and $\sigma(H_n)\cap (a-\epsilon,a+\epsilon)=\emptyset$, then also $a\notin\sigma(H)$.

$\endgroup$

Your Answer

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

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