Multi-Collinear Splitting Kernels for Track Function Evolution

Fuente: arXiv
Salvato in:
Dettagli Bibliografici
Autori principali: Chen, Hao, Jaarsma, Max, Li, Yibei, Moult, Ian, Waalewijn, Wouter J., Zhu, Hua Xing
Natura: Preprint
Pubblicazione: 2022
Soggetti:
Accesso online:
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866929233162403840
author Chen, Hao
Jaarsma, Max
Li, Yibei
Moult, Ian
Waalewijn, Wouter J.
Zhu, Hua Xing
author_facet Chen, Hao
Jaarsma, Max
Li, Yibei
Moult, Ian
Waalewijn, Wouter J.
Zhu, Hua Xing
contents Jets and their substructure play a central role in many analyses at the Large Hadron Collider (LHC). To improve the precision of measurements, as well as to enable measurement of jet substructure at increasingly small angular scales, tracking information is often used due to its superior angular resolution and robustness to pile-up. Calculations of track-based observables involve non-perturbative track functions, that absorb infrared divergences in perturbative calculations and describe the transition to charged hadrons. The infrared divergences are directly related to the renormalization group evolution (RGE), and can be systematically computed in perturbation theory. Unlike the standard DGLAP evolution, the RGE of the track functions is non-linear, encoding correlations in the fragmentation process. We compute the next-to-leading order (NLO) evolution of the track functions, which involves in its kernel the full $1\rightarrow3$ splitting function. We discuss in detail how how we implement the evolution equation numerically, and illustrate the size of the NLO corrections. We also show that our equation can be viewed as a master equation for collinear evolution at NLO, by illustrating that by integrating out specific terms, one can derive the evolution for any $N$-hadron fragmentation function. Our results provide a crucial ingredient for obtaining track-based predictions for generic measurements at the LHC, and for improving the description of the collinear dynamics of jets.
format Preprint
id arxiv_https___arxiv_org_abs_2210_10058
institution arXiv
publishDate 2022
record_format arxiv
spellingShingle Multi-Collinear Splitting Kernels for Track Function Evolution
Chen, Hao
Jaarsma, Max
Li, Yibei
Moult, Ian
Waalewijn, Wouter J.
Zhu, Hua Xing
High Energy Physics - Phenomenology
High Energy Physics - Experiment
Nuclear Theory
Jets and their substructure play a central role in many analyses at the Large Hadron Collider (LHC). To improve the precision of measurements, as well as to enable measurement of jet substructure at increasingly small angular scales, tracking information is often used due to its superior angular resolution and robustness to pile-up. Calculations of track-based observables involve non-perturbative track functions, that absorb infrared divergences in perturbative calculations and describe the transition to charged hadrons. The infrared divergences are directly related to the renormalization group evolution (RGE), and can be systematically computed in perturbation theory. Unlike the standard DGLAP evolution, the RGE of the track functions is non-linear, encoding correlations in the fragmentation process. We compute the next-to-leading order (NLO) evolution of the track functions, which involves in its kernel the full $1\rightarrow3$ splitting function. We discuss in detail how how we implement the evolution equation numerically, and illustrate the size of the NLO corrections. We also show that our equation can be viewed as a master equation for collinear evolution at NLO, by illustrating that by integrating out specific terms, one can derive the evolution for any $N$-hadron fragmentation function. Our results provide a crucial ingredient for obtaining track-based predictions for generic measurements at the LHC, and for improving the description of the collinear dynamics of jets.
title Multi-Collinear Splitting Kernels for Track Function Evolution
topic High Energy Physics - Phenomenology
High Energy Physics - Experiment
Nuclear Theory
url https://arxiv.org/abs/2210.10058