In ultra-peripheral relativistic heavy-ion collisions, a photon from the electromagnetic field of one nucleus can fluctuate to a quark-antiquark pair and scatter from the other nucleus, emerging as a $\rho^0$. The $\rho^0$ production occurs in two well-separated (median impact parameters of 20 and 40 fermi for the cases considered here) nuclei, so the system forms a 2-source interferometer. At low transverse momenta, the two amplitudes interfere destructively, suppressing $\rho^0$ production. Since the $\rho^0$ decay before the production amplitudes from the two sources can overlap, the two-pion system can only be described with an entangled non-local wave function, and is thus an example of the Einstein-Podolsky-Rosen paradox. We observe this suppression in 200 GeV per nucleon-pair gold-gold collisions. The interference is $87% \pm 5% {\rm (stat.)}\pm 8%$ (syst.) of the expected level. This translates into a limit on decoherence due to wave function collapse or other factors, of 23% at the 90% confidence level.
Rapidity (left) and $M_{\pi\pi}$ (right) of the $\pi^{+}\pi^{-}$ distributions for the topology (exclusive $\rho^0$, top) and MB (Coulomb breakup, bottom) samples. The points with statistical error bars are the data, and the histograms are the simulations. The ’notch’ in the topology data around y = 0 is due to the explicit rapidity cut to remove cosmic-ray backgrounds.
Rapidity (left) and $M_{\pi\pi}$ (right) of the $\pi^{+}\pi^{-}$ distributions for the topology (exclusive $\rho^0$, top) and MB (Coulomb breakup, bottom) samples. The points with statistical error bars are the data, and the histograms are the simulations. The ’notch’ in the topology data around y = 0 is due to the explicit rapidity cut to remove cosmic-ray backgrounds.
Raw (uncorrected) ρ0 $t_{\perp}$-spectrum in the range 0.0 < |y| < 0.5 for the MB data. The points are data, with statistical errors. The dashed (filled) histogram is a simulation with an interference term (“Int”), while the solid histogram is a simulation without interference (“NoInt”). The handful of events histogrammed at the bottom of the plot are the wrong-sign ($\pi^{+}\pi^{+}+\pi^{-}\pi^{-}$) events, used to estimate the combinatorial background.
Forum