Strange hidden-charm pentaquark poles from $B^-\to J/\psi\Lambda\bar{p}$

Abstract

Recent LHCb data for $B^-\to J/\psi\Lambda\bar{p}$ show a clear peak structure at the $\Xi_c\bar{D}$ threshold in the $J/\psi\Lambda$ invariant mass ($M_{J/\psi\Lambda}$) distribution. The LHCb's amplitude analysis identified the peak with the first hidden-charm pentaquark with strangeness $P_{\psi s}^\Lambda(4338)$. We conduct a coupled-channel amplitude analysis of the LHCb data by simultaneously fitting the $M_{J/\psi\Lambda}$, $M_{J/\psi\bar{p}}$, $M_{\Lambda\bar{p}}$, and $\cos\theta_{K^*}$ distributions. Rather than the Breit-Wigner fit employed in the LHCb analysis, we consider relevant threshold effects and a unitary $\Xi_c\bar{D}$-$\Lambda_c\bar{D}_s$ coupled-channel scattering amplitude from which $P_{\psi s}^\Lambda$ poles are extracted for the first time. In our default fit, the $P_{\psi s}^\Lambda(4338)$ pole is almost a $\Xi_c \bar{D}$ bound state at $( 4338.2\pm 1.4)-( 1.9\pm 0.5 )\,i$ MeV. Our default model also fits a large fluctuation at the $\Lambda_c\bar{D}_s$ threshold, giving a $\Lambda_c\bar{D}_s$ virtual state, $P_{\psi s}^\Lambda(4255)$, at $4254.7\pm 0.4$ MeV. We also found that the $P_{\psi s}^\Lambda(4338)$ peak cannot solely be a kinematical effect, and a nearby pole is needed.