3
$\begingroup$

I am asking a question related to Lieb-Robinson bound and nonlocality.

As we know from Lieb-Robinson theorem (see e.g. http://arxiv.org/abs/1008.5137): Suppose a Hamiltonian system is local, i.e. $H=\sum_Z H_Z$ where $H_Z$ (with finite dimension) is supported on a set $Z$ and $||H_Z||$ decays rapidly (at least exponential) with the diameter of the set $Z$. Let $A_X$ and $B_Y$ be local operators supported on set $X$ and $Y$ with $dist(X,Y)>0$ (note $dist(X,Y)=min_{i\in X,j\in Y}(i,j)$), the one can show that there is a upper bound

$||[A_X(t),B_Y]||\le c\, e^{-a\,dist(X,Y)}\big[e^{2s|t|-1}\big]$ for some positive constant $c$, $a$, and $s$.

Then, I have two questions: 1) Suppose there is a nonlocal term in the Hamiltonian, i.e. some term $||H_X||$ decays slowly (at most power-lay) with the diameter of the set $Z$, can we show the following statement:

There exist local operators $A_X$ and $B_Y$ with $[A_X,B_Y]=0$ and $dist(X,Y)>0$, and a positive constant $t_0$ and $s>0$, such that for any $t>t_0$ we have a lower bound:

$||[A_X(t),B_Y]||\ge s$. (Answered by Norbert Schuch)

2) Suppose the Lieb Robinson bound is satisfied and one can also define the Lieb Robinson velocity for a Hamiltonian system, can we prove the Hamiltonian is local?

$\endgroup$
1
$\begingroup$

Such a lower bound will not hold. Briefly speaking, the reason is that there will always exist times $\hat t$ for which $e^{iH\hat t}$ will be the identity (or at least very close to it).

A simple example is given by $N$ spins which all interact via an Ising interaction, $H=\sum_{i,j} \sigma^z_i \sigma^z_j$. Then, the interaction does not depend on the distance, and for any time $t=n\pi$, $e^{iH\hat t}=\mathrm{Id}$. Indeed, for any finite system there will be a sequence of times $t_n$ where $e^{iHt_n}\rightarrow\mathrm{Id}$, and thus the commutator is zero.

Note that LR-bounds also hold for (sufficiently fast) decaying algebraic interactions, in which case the distance-dependent part in the bound decays algebraically (i.e. there is no light cone).

$\endgroup$
  • $\begingroup$ Thanks a lot. But very sorry. I made a mistake in formulating my question. $\endgroup$ – Del Don Aug 10 '15 at 4:35
  • $\begingroup$ @DelDon Could you maybe fix it? $\endgroup$ – Norbert Schuch Aug 10 '15 at 9:07
  • $\begingroup$ Suppose the Lieb Robinson bound is satisfied and one can also define the Lieb Robinson velocity for a Hamiltonian system, can we show the Hamiltonian is local? $\endgroup$ – Del Don Aug 18 '15 at 0:36
  • $\begingroup$ @DelDon If that is your question, maybe you should edit your question accordingly! $\endgroup$ – Norbert Schuch Aug 18 '15 at 10:53

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.