2
$\begingroup$

This question stems from a remark on page 39 of Folland's book "Quantum Field Theory: A Tourist Guide for Mathematicians," in which Folland writes:

"The group of unitary operators [on a Hilbert space] is connected because, by spectral theory, every unitary operator belongs to a continuous one-parameter group of unitary operators and so is connected to the identity by a continuous path."

I know by spectral theory that, as a normal operator, a unitary operator $U$ has a unique resolution of the identity $E$ such that $U = \int_{\sigma(U)}\lambda dE(\lambda)$, where $\sigma(U)$ is the spectrum of $U$. I also know that $I = \int_{\sigma(U)}dE(\lambda)$, so perhaps $U_t = \int_{\sigma(U)}\lambda^{t}dE(\lambda)$, $t \in [-1,1]$ is the one-parameter group he's talking about? How would I show that each $U_t$ is unitary and that the path $\gamma(t) = U_t$ is continuous when the unitary group is given the topology inherited from the strong operator topology on all bounded operators on a Hilbert space?

Any help is appreciated.

$\endgroup$
5
$\begingroup$

You can see this as follows: $$ \|U_tx\|^2 = \left\|\int\lambda^t\,dEx\right\|^2 = \int|\lambda|^{2t}d\|Ex\|^2 = \int\,d\|Ex\|^2 = \|E(\sigma(U))x\|^2 = \|x\|^2. $$ The continuity of $t\mapsto U_t$ follows from the uniform Lipschitz continuity of $t\mapsto\lambda^t$ for $\lambda$ on the unit circle.

But I prefer another way. One can show that $U = e^{iA}$ with a bounded selfadjoint operator $A$ and that for each bounded selfadjoint operator $B$ the operator $e^{iB}$ is unitary. Hence, $U$ belongs to the continuous one-parameter group $\{e^{itA} : t\in\mathbb R\}$.

$\endgroup$
  • 1
    $\begingroup$ The second answer is very neat! I hadn't thought to use that the spectrum of a unitary operator lies in the unit circle, so thank you. $\endgroup$ – Mathmank Oct 14 '17 at 3:17
  • $\begingroup$ It was a pleasure. :o) $\endgroup$ – amsmath Oct 14 '17 at 3:28
  • $\begingroup$ @amsmath , and how can show the group is connected, or the same the function $f(t)=e^{itA}$ is continuous $\endgroup$ – user89940 Nov 29 '18 at 1:07
  • $\begingroup$ @user89940 You can use the properties of the continuous functional calculus. For $s,t\in\mathbb R$ define the function $f(x) = e^{itx} - e^{isx}$. Then $f$ is continuous and $\|f(A)\| = \sup\{|f(x)| : x\in\sigma(A)\}\le\sup\{|f(x)| : x\in[-\pi,\pi]\}$ because $\sigma(A)\subset [-\pi,\pi]$, $\endgroup$ – amsmath Jan 3 '19 at 21:58

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.