Challenges in flow background removal in search for the chiral magnetic effect

We investigate the effect of resonance decays on the three-particle correlator charge separation observable in search for the chiral magnetic effect, using a simple simulation with realistic inputs. We find that resonance decays can largely account for the measured signal. We suppress the elliptic flow ($v_2$) background by using zero event-by-event $v_2$ (or via the mixed-event technique). We find that the background is suppressed, but not eliminated as naively anticipated. We identify the reason to be the non-identicalness of the resonance and final-state particle's $v_2$ and the induced correlation between the transverse momentum dependent resonance $v_2$ and decay angle. We make predictions for the charge separation signal due to resonance decays in 200~GeV Au+Au collisions.