Cyclic Properties of the trace in Quantum Field Theory

588 Views Asked by At

I'm trying to figure out what's going on in this paper here between lines 10 and 11. But I'll give a brief rundown of what's going on.

We are trying to compute the trace of the exponential operator $e^{A}$ where $A = a c + b c^{\dagger}$ knowing the commutation relation $[c,c^{\dagger}] = 1$.

The paper proceeds by exploiting the fact that

$$ e^{ac + bc^{\dagger}} = e^{bc^{\dagger} + ac} $$

Then using the Baker-Campbell-Hausdorff formula, it follows that: $$ e^{-\frac{1}{2} ab} e^{ac}e^{bc^{\dagger}} = e^{\frac{1}{2} ab} e^{bc^{\dagger}} e^{ac} $$

Then they consider the trace of the operators:

$$ Tr(e^{ac}e^{bc^{\dagger}}) = e^{ab}Tr(e^{bc^{\dagger}}e^{ac}) $$

Then they seem to claim that given a pair of inverse operators $UU^{-1} = 1$ it follows solely from the invariance of the trace under cyclic permutations that $Tr(AB) = Tr(UAU^{-1}B)$ so that:

$$ Tr(e^{ac}e^{bc^{\dagger}}) = e^{ab}Tr(Ue^{ac}U^{-1}e^{bc^{\dagger}}) $$

But proceeding from the third displayed equation and using the cyclic property, it seems that:

$$ e^{ab}Tr(e^{bc^{\dagger}}e^{ac}) = e^{ab}Tr(Ue^{bc^{\dagger}}U^{-1}Ue^{ac}U^{-1}) \\ = e^{ab}Tr(Ue^{ac}U^{-1}Ue^{bc^{\dagger}}U^{-1}) $$

So it seems that for the fourth displayed equation to be true, it must be that

$$ Ue^{bc^{\dagger}}U^{-1} = e^{bc^{\dagger}}$$

Which requires that $[e^{bc^{\dagger}}, U^{-1}] = 0$

But given the form of $U= e^{\beta ( c^{\dagger} c + \frac{1}{2} )}$, $[e^{bc^{\dagger}}, U] \ne 0$.

Is there something I'm missing here, or am I misinterpreting the author's justification (I've seen this justification made in several different places)?