Performance of CMS Muon Reconstruction in Cosmic-Ray Events

TL;DR

This paper evaluates the CMS muon reconstruction performance using a large sample of 2008 cosmic-ray muons from the CRAFT run, validating trigger, identification, and momentum/charge measurements against Monte Carlo simulations. It details standalone and global muon reconstruction, muon-identification strategies, and specialized cosmic-muon algorithms, showing generally good data–MC agreement. The study highlights momentum resolution in the barrel—better than $1\%$ at $p_T=10\,\mathrm{GeV}/c$ and about $8\%$ at $p_T=500\,\mathrm{GeV}/c$—and low charge-misassignment rates (<$0.01\%$ at 10 GeV/$c$ rising to ~1% at 500 GeV/$c$) under CRAFT alignment. It also demonstrates near-ideal High-Level Trigger performance for Level-2 and Level-3 muon reconstruction in an offline, unbiased setting, validating the CMS muon-system design for LHC collisions. Overall, the results confirm robust muon reconstruction and identification suitable for precision physics in CMS during LHC Run 1 and beyond.

Abstract

The performance of muon reconstruction in CMS is evaluated using a large data sample of cosmic-ray muons recorded in 2008. Efficiencies of various high-level trigger, identification, and reconstruction algorithms have been measured for a broad range of muon momenta, and were found to be in good agreement with expectations from Monte Carlo simulation. The relative momentum resolution for muons crossing the barrel part of the detector is better than 1% at 10 GeV/c and is about 8% at 500 GeV/c, the latter being only a factor of two worse than expected with ideal alignment conditions. Muon charge misassignment ranges from less than 0.01% at 10 GeV/c to about 1% at 500 GeV/c.

Figures (23)

• Figure 1: Event display of a cosmic muon crossing CMS: the side view (left) and a part of the transverse view (right). "MB" and "ME" labels indicate positions of the muon barrel and the muon endcap stations, respectively. The solid blue curve represents a 1-leg global muon reconstructed using silicon tracker and muon system hits in the whole detector. Small green circles indicate hits in the silicon tracker. Short red stubs correspond to fitted track segments in the muon system; as the $z$ position is not measured in the outer barrel station, the segments in it are drawn at the $z$ center of the wheel, with their directions perpendicular to the chamber. Energy deposits in the electromagnetic and hadron calorimeters are shown as (thin) magenta and (thick) blue bars, respectively.
• Figure 2: Distributions of a) the transverse momentum $p_{\mathrm{T}}$, b) the momentum $p$, c) the azimuthal angle $\phi$, and d) the pseudorapidity $\eta$ of 2-leg global muons at the point of closest approach to the nominal beam line, for the data (points) and for the Monte Carlo simulation (histogram). The MC distributions are normalized to the number of events in the data.
• Figure 3: Distributions of the total number of hits per track for the data (points) and for the Monte Carlo simulation (histogram), for a) CosmicSTA standalone muons and b) 2-leg global muons. The MC distributions are normalized to the number of events in the data.
• Figure 4: Distributions of the $\chi^2/\text{ndf}$ of the track fit for the data (points) and for the Monte Carlo simulation (histogram), for a) CosmicSTA standalone muons and b) 2-leg global muons. The MC distributions are normalized to the number of events in the data.
• Figure 5: Distributions of residuals of the local $x$ position for the track-to-segment match in the data (points) and in the Monte Carlo simulation (histogram) for a) MB1 chambers; b) MB4 chambers; c) ME1 chambers; d) ME3 chambers. The MC distributions are normalized to the number of events in the data.