Evidence of Spin-Interference Effects in Exclusive $J/ψ\to e^+e^-$ Photoproduction in Ultraperipheral Heavy-Ion Collisions

TL;DR

This paper reports the first evidence of spin-interference in exclusive $J/\psi \to e^+e^-$ photoproduction in ultraperipheral Au+Au and isobar collisions at $\sqrt{s_{NN}}=200$ GeV. The angular observable $\cos(2\phi)$ shows a negative modulation for $p_T<100$ MeV/$c$, opposite in sign to the previously measured $\rho^0 \to \pi^+\pi^-$ channel, demonstrating that the interference sign is set by the spin structure of the final-state leptons. Because the $J/\psi$ has a long lifetime, the interference persists beyond the nuclear volume, with measured amplitudes peaking near $a_2\approx -0.22$ (isobars) and $a_2\approx -0.19$ (Au+Au). The data agree qualitatively with CGC Model-I calculations that include linear polarization and two-source interference for $J/\psi$, while Model-II that adds final-state soft-photon radiation is disfavored in the measured region. This work establishes spin-dependent interference as a new tool to image gluon distributions at perturbative scales and motivates future applications to diffractive $e$+A scattering at the Electron-Ion Collider.

Abstract

We report the first evidence of spin interference in exclusive $J/ψ\to e^+e^-$ photoproduction in ultraperipheral Au+Au and isobar (Ru+Ru, Zr+Zr) collisions ($\sqrt{s_{NN}} = 200$~GeV) at STAR. A negative $\cos(2φ)$ modulation is observed for $p_T < 100$ MeV/$c$, opposite in sign to that in $ρ^{0}\!\to\!π^+π^-$ photoproduction. This establishes for the first time that the interference sign is controlled by the spin structure of the final-state daughters, resolving the ambiguity present in the all-boson $ρ^0$ channel. The compact $J/ψ$ probes gluon distributions at perturbative scales, resulting in a weaker modulation and providing stringent constraints on Color Glass Condensate calculations. These findings demonstrate that spin-dependent interference in heavy vector mesons provides a new, experimentally accessible handle on gluon structure beyond traditional cross-section measurements.

Figures (4)

• Figure 1: Schematic comparison of quantum interference in coherent vector meson production in UPCs for a given impact parameter, $\mathrm{b}$, at STAR. (a) For $\rho^{0} \to \pi^{+}\pi^{-}$, the short lifetime ($c\tau \approx 1.3$ fm [$<\mathrm{b}$]) leads to decay within the nuclear environment, limiting the spatial coherence. The spin-0 pions result in angular interference patterns that differ from those of leptonic decays. (b) In contrast, for $J/\psi \to e^{+}e^{-}$, the long lifetime ($c\tau \approx 2160$ fm [$\gg\mathrm{b}$]) allows the meson to propagate far outside the nucleus before decaying. The final-state electrons are entangled, and interference arises between production amplitudes from the two nuclei.
• Figure 2: Invariant mass $m_{ee}$ distribution for $e^+e^-$ pair candidates for pair transverse momentum $p_{T,ee}<200$ MeV/$\textit{c}$ and pair rapidity $|y_{ee}|<1$, from Ru+Ru and Zr+Zr UPCs at $\sqrt{s_{NN}}=200$ GeV. The light red and blue regions are the distributions for unlike-sign ($+-$) and like-sign ($++$, $--$) electron pairs, respectively. A template fit that includes STARlight coherent and incoherent $J/\psi$ and QED processes is shown as the green curve. The two vertical black dashed curves represent the chosen $J/\psi$ mass window, $3<m_{ee}<3.2$ GeV/$c^{2}$.
• Figure 3: The amplitude of the cos(2$\phi$) modulation, $a_2$, as a function of ${J/\psi}$$p_T$ from Au+Au (left panel) and Ru+Ru & Zr+Zr (right panel) UPCs at $\sqrt{s_{NN}}=200$ GeV. The statistical uncertainty on each data point is shown as a vertical bar, while the systematic uncertainty is shown in the shaded band. The STARlight Klein:2016yzr, CGC (Model-I) Mantysaari:2023prg and CGC (Model-II) Brandenburg:2022jgr calculations are shown with magenta line, green band, and orange line, respectively. STARlight is used here as a spin-independent baseline and not as a full theoretical prediction of the modulation. Model-I incorporates CGC calculations with linear photon polarization and interference effects, while Model-II extends this framework by including soft-photon radiation in the final state.
• Figure 4: Detector-only $\phi$ modulation from STARlight+GEANT with STARlight input $a_2^{\rm true}=0$. STARlight points are first embedded into zero-bias data, then a full GEANT3-based reconstruction is performed. We refer to this as "MC". The curves are fit to the MC points with $1+a_2^{D}\cos(2\phi)$.