#### Alladi Ramakrishnan Hall

#### BRST v/s EPR in Maxwell

#### Ronak Soni

##### TIFR

*It has been known for a while that entanglement shouldn't be calculated*

after gauge-fixing, because gauge-fixing 'rearranges' degrees of freedom

in space. However, it is an open question whether accounting for the

entanglement of the BRST ghosts cancels this gauge-dependence. More

precisely, the question is: do two different states in the same

BRST-cohomology necessarily have the same entanglement? A piece of

evidence that the answer is yes is that the replica trick path integral is

gauge-invariant. We show, however, that the answer is no, and further that

this is completely compatible with the gauge-invariance of the replica

trick path integral. Finally, I discuss whether we can bound the

gauge-dependence in some way.

Done