Finite $V_{\rm 2Δ}$ puzzle in low-multiplicity pp collisions from ultra-long-range azimuthal correlations in the string-shoving model

TL;DR

This work investigates ultra-long-range azimuthal correlations in pp collisions at $\sqrt{s}=13$ TeV using PYTHIA8 with the string shoving mechanism to test its ability to produce a finite $V_{2Δ}$ in low-multiplicity events. It employs a template-fit non-flow subtraction in the ultra-long-range region $|\Delta\eta|>5.0$ and compares event-activity estimators $N_{\rm ch}$, $N_{\rm mpi}$, and flattenicity to assess biases and the role of global event topology. The results show that string shoving generates a nonzero but diminishing $V_{2Δ}$ with increasing $N_{\rm ch}$, and strongest signals in low-$N_{\rm mpi}$ (dijet-like) events, while hydro models underpredict the data at low and intermediate multiplicities; flattenicity and $N_{\rm mpi}$ reduce biases relative to $N_{\rm ch}$. The study argues for a gradual transition from initial-state string dynamics to hydrodynamic behavior as multiplicity grows and advocates for topology-based event classifiers to better isolate genuine collective effects in small systems.

Abstract

Ultra-long range angular correlations have been recently reported by the ALICE collaboration in pp collisions at $\sqrt{s}=13$ TeV below ${\rm d}N_{\rm ch}/{\rm d}η=7$. The measurements have been performed as a function of the charged-particle multiplicity at midrapidity ($N_{\rm ch}$ in $|η|<0.8$), which is known to be strongly sensitive to local multiplicity fluctuations. The present work investigates the impact of the event-activity estimator on ultra-long range angular correlations. The study is conducted in the framework of PYTHIA8 with the string shoving mechanism since it gives a non-zero elliptic flow coefficient, $V_{2Δ}$. The analysis is conducted as a function of $N_{\rm ch}$, the number of parton-parton scatterings ($N_{\rm mpi}$) and flattenicity. Surprisingly, for ultra-long range correlations, pp collisions with $N_{\rm mpi}=1$ (dijets) seems to be the most sensitive to string shoving. The effect diminishes with increasing $N_{\rm mpi}$. While in data, within uncertainties, $V_{2Δ}$ exhibits a weak multiplicity dependence; the string shoving mechanism gives a $V_{2Δ}$ that decreases with the increase in $N_{\rm ch}$. The present work therefore supports the picture stating that mechanisms such as string shoving might explain the low multiplicity limit, whereas, hydro becomes relevant in high-multiplicity pp collisions. This work also suggests that flattenicity might be more effective than $N_{\rm ch}$ to better handle non-flow effects.

Figures (5)

• Figure 1: The $N_{\rm ch}$ dependence of the second-order two-particle correlation coefficient $V_{2\Delta}$ in pp collisions at $\sqrt{s}$ = 13 TeV. The black markers represent the ALICE data ALICE:2025bwp and the cyan band is for the hydro simulations Zhao:2022ugy.
• Figure 2: The $N_{\rm ch}$ dependence of the second-order two-particle correlation coefficient $V_{2\Delta}$ in pp collisions at $\sqrt{s}$ = 13 TeV using PYTHIA8 simulations. The black markers represent the ALICE data ALICE:2025bwp.
• Figure 3: The ultra-long range correlation per-triggered yield estimated as a function of $\Delta\eta$ and $\Delta\varphi$ in pp collisions at $\sqrt{s}$ = 13 TeV with PYTHIA for different estimators for small (class-I) and large event activity (class-VI).
• Figure 4: The ultra-long range correlation per-triggered yield estimated as a function of $\Delta\varphi$ in pp collisions at $\sqrt{s}$ = 13 TeV with PYTHIA for different estimators.
• Figure 5: The ultra-long range correlation per-triggered yield estimated as a function of $\Delta\varphi$ in pp collisions at $\sqrt{s}$ = 13 TeV with PYTHIA for different MPI event activity using different values of string shoving parameter.