The effect of recoils on soft-drop-groomed observables in $γ$-tagged jets in a multistage approach

We investigate medium-induced modifications to jet substructure observables that characterize hard components in central Pb-Pb collisions at $\sqrt{s_{NN}}=5.02$~TeV. Using a multistage Monte Carlo simulation of in-medium jet shower evolution, we explore flavor-dependent medium effects through simulations of inclusive and $γ$-tagged jets. The results show that quark jets undergo a nonmonotonic modification compared with gluon jets in observables such as the Pb-Pb to $p$-$p$ ratio of the soft drop prong angle $r_g$, the relative prong transverse momentum $k_{T,g}$, and the groomed mass $m_g$ distributions. Due to this nonmonotonic modification, $γ$-tagged jets, enriched in quark jets, provide surprisingly clear signals of medium-induced structural modifications, distinct from effects dominated by selection bias. Further systematic studies demonstrate that these effects are dominated by recoil medium response. This work highlights the potential of hard substructures in $γ$-tagged jets as powerful tools for probing the jet-medium interactions in high-energy heavy-ion collisions, in particular by enabling detailed investigations of jet-medium parton scatterings via their associated medium response. All simulations for $γ$-tagged jet analyses carried out in this paper used triggered events containing at least one hard photon, which highlights the utility of these observables for future Bayesian analysis.

💡 Research Summary

**
This paper investigates medium‑induced modifications of jet substructure in central Pb‑Pb collisions at √sₙₙ = 5.02 TeV using a state‑of‑the‑art multistage Monte Carlo framework (JETSCAPE) that couples the MATTER and LBT modules. The authors focus on soft‑drop groomed observables—prong angular separation (r_g), relative transverse momentum (k_T,g), and groomed mass (m_g)—which characterize the hard branching reconstructed by the grooming algorithm.

The simulation chain begins with a (2 + 1)‑dimensional free‑streaming pre‑equilibrium evolution followed by viscous hydrodynamics (VISHNU) using fluctuating Trento initial conditions to generate a realistic QGP background. Hard partons are produced with PYTHIA 8, then evolved through two stages: (i) a virtuality‑ordered MATTER phase that handles high‑Q² splittings with only modest medium effects, and (ii) a low‑virtuality LBT phase that treats elastic and inelastic scatterings with medium constituents based on kinetic theory. A switching criterion Q²_sw = 4 GeV² transfers partons between the modules bidirectionally. Recoil partons generated in each scattering and the associated “hole” partons (representing the energy‑momentum deficit) are tracked throughout the LBT stage. After exiting the medium (T > T_c = 0.16 GeV), all partons are hadronized using PYTHIA 8’s Lund string model.

Two jet samples are studied: inclusive jets selected by jet p_T and γ‑tagged jets selected by a high‑p_T isolated photon. The γ‑tagged sample is dominated by quark jets, while inclusive jets contain a mixture of quark and gluon jets and suffer from a selection bias that preferentially removes jets with large‑angle hard branchings (large r_g) because such jets lose more energy outside the jet cone.

Key findings:

1. Prong angle (r_g) – The Pb‑Pb / p‑p ratio exhibits a non‑monotonic behavior: an enhancement at moderate angles (0.1 ≲ r_g ≲ 0.3) followed by suppression at larger angles. This reflects that quark jets broaden due to medium scatterings, but very wide branchings are depleted by energy loss.

2. Relative transverse momentum (k_T,g) – Medium‑induced scatterings increase k_T,g, especially in the intermediate region, yielding a ratio of 1.2–1.4. This broadening is driven by elastic scatterings and induced radiation in the LBT phase.

3. Groomed mass (m_g) – The groomed mass distribution shifts to higher values, with an average increase of ~10–15 % and a pronounced tail, consistent with simultaneous increases in r_g and k_T,g.

Systematic studies disabling the recoil component demonstrate that the observed modifications largely disappear, establishing recoil partons as the dominant source of the structural changes. Comparisons with CMS and ATLAS measurements of γ‑jet r_g and m_g show good agreement in both shape and magnitude, particularly the moderate‑r_g enhancement.

The authors argue that γ‑tagged jets provide a “golden channel” that mitigates selection bias and isolates genuine medium‑induced modifications of the hard branching. Because the photon does not interact strongly with the QGP, its p_T serves as an unbiased proxy for the initial parton energy, allowing a cleaner comparison between Pb‑Pb and p‑p. The quark‑jet enrichment further amplifies the sensitivity to medium effects, as quark jets exhibit stronger low‑virtuality modifications than gluon jets in the multistage model.

In conclusion, the study demonstrates that soft‑drop observables in γ‑tagged jets are powerful probes of jet‑medium interactions, with recoil response playing a central role. The results support the use of γ‑tagged jet substructure in future Bayesian extractions of QGP transport properties and motivate extensions of the model to include non‑on‑shell recoil dynamics and more sophisticated hadronization schemes.