Table of Contents
Fetching ...

A data-driven method of pile-up correction for the substructure of massive jets

Raz Alon, Ehud Duchovni, Gilad Perez, Aliaksandr P. Pranko, Pekka K. Sinervo

TL;DR

This work introduces a data-driven framework to correct for incoherent energy contributions from pile-up and multiple interactions on jet substructure observables for ultra-massive jets. By measuring the incoherent component in data via a 90° rotated cone in dijet events, the authors derive jet-variable–dependent corrections with explicit analytic forms for jet mass, angularity, and planar flow, reducing reliance on Monte Carlo models. The corrections agree with CDF data and are linked to the jet-area concept, offering a practical path to improved resolution and sensitivity in high-pT jet analyses. The method also provides insights into how jet mass area relates to observable jet shapes, facilitating broader adoption across collider environments. Overall, the approach enables more accurate, data-driven subtraction of pile-up effects in jet substructure studies with potential impact on new physics searches.

Abstract

We describe a method to measure and subtract the incoherent component of energy flow arising from multiple interactions from jet shape/substructure observables of ultra-massive jets. The amount subtracted is a function of the jet shape variable of interest and not a universal property. Such a correction is expected to significantly reduce any bias in the corresponding distributions generated by the presence of multiple interactions, and to improve measurement resolution. Since in our method the correction is obtained from the data, it is not subject to uncertainties coming from the use of theoretical calculations and/or Monte Carlo event generators. We derive our correction method for the jet mass, angularity and planar flow. We find these corrections to be in good agreement with data on massive jets observed by the CDF collaboration. Finally, we comment on the linkage with the concept of jet area and jet mass area.

A data-driven method of pile-up correction for the substructure of massive jets

TL;DR

This work introduces a data-driven framework to correct for incoherent energy contributions from pile-up and multiple interactions on jet substructure observables for ultra-massive jets. By measuring the incoherent component in data via a 90° rotated cone in dijet events, the authors derive jet-variable–dependent corrections with explicit analytic forms for jet mass, angularity, and planar flow, reducing reliance on Monte Carlo models. The corrections agree with CDF data and are linked to the jet-area concept, offering a practical path to improved resolution and sensitivity in high-pT jet analyses. The method also provides insights into how jet mass area relates to observable jet shapes, facilitating broader adoption across collider environments. Overall, the approach enables more accurate, data-driven subtraction of pile-up effects in jet substructure studies with potential impact on new physics searches.

Abstract

We describe a method to measure and subtract the incoherent component of energy flow arising from multiple interactions from jet shape/substructure observables of ultra-massive jets. The amount subtracted is a function of the jet shape variable of interest and not a universal property. Such a correction is expected to significantly reduce any bias in the corresponding distributions generated by the presence of multiple interactions, and to improve measurement resolution. Since in our method the correction is obtained from the data, it is not subject to uncertainties coming from the use of theoretical calculations and/or Monte Carlo event generators. We derive our correction method for the jet mass, angularity and planar flow. We find these corrections to be in good agreement with data on massive jets observed by the CDF collaboration. Finally, we comment on the linkage with the concept of jet area and jet mass area.

Paper Structure

This paper contains 7 sections, 25 equations, 3 figures.

Figures (3)

  • Figure 1: On the upper panel we show the CDF data and a fit based on the relation derived in Eq. \ref{['delmJ']}. The data collected had on average $\sim3$ multiple interactions per event (including the hard interaction). On the lower panel we show the corresponding MC predictions including full detector simulation CDFnew.
  • Figure 2: On the upper panel we show the CDF data and a fit based on the relation derived in Eq. \ref{['AngMI']}. On the lower panel we show the corresponding MC predictions including full detector simulation CDFnew.
  • Figure 3: On the upper panel we show the CDF data and a fit based on the relation derived in Eq. \ref{['PFMI']}. On the lower panel we show the corresponding MC predictions including full detector simulation CDFnew.