A Muon Identification and Combined Reconstruction Procedure for the ATLAS Detector at The

A Muon Identification and Combined Reconstruction Procedure for the ATLAS Detector at The



A Muon Identification and Combined Reconstruction Procedure for the ATLAS Detector at the LHC at CERN

Th. Lagouri, D. Adams, K. Assamagan, M. Biglietti, G.Carlino, G. Cataldi, F. Conventi, A. Farilla, Y. Fisyak, S. Goldafarb, E. Gorini, K. Mair, L. Merola, A. Nairz, A. Poppleton, M. Primavera, S. Rosati, J. Shank, S. Spagnolo, L. Spogli, G. Stavropoulos, M. Verducci, T. Wenaus

Abstract—Muon identification and high momentum measurement accuracy is crucial to fully exploit the physics potential that will be accessible with ATLAS experiment at the LHC. The muon energy of physics interest ranges in a large interval from few GeV, where the b-physics studies dominate the physics program, up to the highest values that could indicate the presence of new physics. The muon detection system of the ATLAS detector is characterized by two high precision tracking systems, namely the Inner Detector and the Muon Spectrometer plus a thick calorimeter that ensures a safe hadron absorption filtering with high purity muons with energy above 3 GeV.

In order to combine the muon tracks reconstructed in the Inner Detector and the Muon Spectrometer the Muon Identification (MUID) Object-Oriented software package has been developed. The purpose of the MUID procedure is to associate tracks found in the Muon Spectrometer with the corresponding Inner Detector track and calorimeter information in order to identify muons at their production vertex with optimum parameter resolution. The performance of these two combined systems has been evaluated with single muons of fixed transverse momentum and with full physics events.

I.INTRODUCTION

The muon identification and the high momentum measurement accuracy is crucial to fully exploit the physics potential that will be accessible with ATLAS experiment at the LHC.The muon detection system of the ATLAS detector is characterized by two high precision tracking systems, namely the Inner Detector [1] and the Muon Spectrometer [2] plus a thick calorimeter that ensures a safe hadron absorption filtering with high purity muons with energy above 3 GeV.

The ATLAS Muon Spectrometer has been designed to achieve momentum measurement with high efficiency and high resolution over a wide range of transverse momentum, pseudorapidity and azimuthal angle, providing at

the same time stand-alone triggering capability. Momentum measurement is performed via magnetic deflection of muon tracks in a system of three large superconducting air-core toroid magnets instrumented with trigger chambers and high precision tracking chambers. The magnet configuration provides a field that is mostly orthogonal to the muon trajectories, while minimizing the degradation of resolution due to multiple scattering.

In order to combine the muon tracks reconstructed in the Muon Spectrometer and the Inner Detector the MUon IDentification (MUID) Object-Oriented software package has been developed.

II.Muon Combined Reconstruction: MUID

The purpose of the MUID Muon Identification package is to associate tracks found in the Muon Spectrometer (MS) with the corresponding Inner Detector (ID) track and calorimeter information in order to identify muons at their production vertex with optimum parameter resolution.

MUID is written in C++. It shares some general reconstruction classes and methods with the inner detector package iPatRec[3]. It uses currently iPatRec for ID reconstruction and accesses the results from Muon Standalone packages; MOORE [5] (in ATHENA [4] framework) (or MuonBox [6] via a C++ interface in ATRECON framework at present, although it is also possible to use MuonBox in the ATHENA framework) for the MS. The extension to other ID (xKalman++) reconstruction packages will follow-on easily since MUID has been ported into the new ATHENA architecture scheme.

MOORE (Muon Object Oriented Reconstruction) is the software package for track reconstruction in the Muon Spectrometer, developed in C++ in the ATHENA framework according to modern Object Oriented design principles. Its design was driven by the goal of performing track reconstruction in a highly modular way, with the highest possible efficiency in all the pseudorapidity range covered by

the Muon Spectrometer and with the best possible resolution needed for muon identification in ATLAS.

Detailed studies were already performed with the FORTRAN program, Muonbox for the Physics Technical Design Report [7] and have indeed shown the capability of the Muon spectrometer to reconstruct muon tracks with an efficiency 95 % for PT >10 GeV in almost all the pseudorapidity range and that the momentum resolution is better than 5 % over 80 % of the phase space for a wide range of PT (roughly from 10 to 300 GeV).

III.Track Combination Procedure

The MUID Muon Identification package combines Muon Spectrometer tracks reconstructed by MOORE (ATHENA) (or MuonBox (ATRECON)) with the Inner Detector tracks found using iPatRec. The purpose of MUID is to identify Inner Detector tracks as muons at all momenta, and provide the best possible muon parameter resolution in the vicinity of the production vertex.

The first step (MUID StandAlone) is to re-express tracks from the Muon Spectrometer in order to have the same representation as those from the Inner Detector reconstruction. The muons are propagated through the magnetic field with energy loss and multiple scattering contributions included to obtain the parameters with their covariance at the point of closest approach to the intersection region. This is performed by applying the iPatRec track filter to aset of scattering planes representing the Calorimeter and Muon Spectrometer material, an energy loss ‘measurement’ obtained either from the observed Calorimeter energy deposition or from parametrization, and to either the MuonBox orMOORE fit parameters given at the entrance of the Muon Spectrometer.

In the next step (MUID Combined), tracks are matched by forming a chi2 with 5 degrees of freedom from the parameter differences and summed covariances. A combined fit is performed to all matches with a chi2 probability above a certain cut. When no matches satisfie this criterion, a combined fit is attempted for the best match within a road about the muon track. The combined fit is a repeat of the muon fit from the first step with the addition of the Inner Detector tracks and scattering planes assigned by iPatRec. Finally all matches to the Inner Detector giving asatisfactory combined fit are retained as identified muons.

For isolated muons the energy loss is taken from the associated Calorimeter cells. It is corrected asfunction of eta and momentum to account for the difference betweentrue and observed energy deposition obtained from thesimulation. Typically the correction adds around 7 %. The benefit of this procedure over parametrization is to correct better for Landau fluctuations, in particular at intermediate energies where the Calorimeter energy loss is significant and the Muon Spectrometer gives a more precise momentum measurement than the Inner Detector. At low momentum it is less precise than parametrization, but here the purpose of the combined fit is to identify the track as a muon rather than to improve the parameters measured by the Inner Detector. Near the threshold for penetration into the Muon Spectrometer, the Calorimeter energy deposition is greater than the remaining track momentum, thus using the measured energy provides a valuable consistency check not available from parametrization. A parametrized correction is used for non-isolated muons [7].

IV.MUID Design

The main idea of the ATHENA framework is the distinction beetween ``data'' objects and ``algorithm'' objects. Anything which is essentially a procedure, i.e. a set of rules for performing transformation on more data-like objects, or for creating new data-like objects, is an algorithm. Data objects should be simple object, without long algorithmic procedures.

The only link between algorithms are the data objects:

packages are organized in such a way that algorithms depend

on data objects but data objects do not depend on algorithms.

Data objects produced by the algorithms are posed in a common ``in memory'' store, called the Transient Data Store,

from where other modules can access them.

This model greatly reduces the coupling between the algorithms, because one algorithmic module does not need

any more to know which specific module can produce the information it needs.

MUID follows this general scheme: it contains three ATHENA algorithms MuidInit, MuidStandAlone, MuidComb that have different tasks and use the ATHENA services and a set of utility classes.

A schematic sketch of the algorithms and the objects exchanged with the Transient Event Store is shown in Figure 1.

Fig.1 MuonIdentification ATHENA algorithms (left) and data objects exchanged with the Transient Event Store (right)

V.MUID Physics Performance

The physics performances of MUID-MOORE have been estimated with Monte Carlo simulation studies, using both single tracks and physics channels. In the following some examples are reported.

A.Single Muon Studies

The reconstruction performance has been tested with single muon samples in a range of transverse momentum from 3 GeV/c to 1 TeV/c. The full reconstruction chain has been executed, namely: the reconstruction in the Muon Spectrometer alone (MOORE), the extrapolation to the vertex of the track found in the Muon Spectrometer (MUID Standalone), the reconstruction in the Inner Detector (iPatRec) and the combination of the track found in the Muon Spectrometer and in the Inner Detector (MUID Combined). The global efficiency and pT resolution as a function of pT is shown in Figure 2.

These results show rather good agreement with the performance obtained in the ATLAS Physics Technical Design Report [7].

Fig.2 Efficiency (left) and pT resolution (right) as a function of pT for MOORE, MUID Standalone, iPatRec and MUID Combined.

B.Physics studies

1)Z

The very precise measurement of the Z boson mass performed at e+e- colliders and the copious production of Z events in ATLAS provide a powerful tool to set the absolute momentum scale of the muon spectrometer.

Thanks to the abundant production of Z bosons (about 30000 events per day at low luminosity), from the known Z mass we will be able to measure other particle masses with high precision and have a cross-check between the different subdetectors, allowing the calculation of systematic uncertainties and reducing them as much as possible.

The Z invariant mass has been evaluated with the reconstruction performed only in the Muon Spectrometer with MOORE (no extrapolation to the vertex, see Fig.3, upper plot) and combining the reconstruction in the Inner Detector and in the Muon Spectrometer with MUID (see Fig.3, down plot). No kinematics cuts were applied to the single muons. A Gaussian fit to the mass distribution obtained with MUID gives  = 2.78 GeV.

Fig.3 Z invariant mass obtained with MOORE (top plot) and with MUID (down plot).

2)H

The Standard Model (SM) Higgs boson decay HZZ*for mH = 130 GeV has been studied using the MOORE and MUID reconstruction software.

The applied event selection is according to the Physics TDR event selection [7]. The signal reconstruction proceeds by selecting four muons which pass the muon identification criteria followed by the following kinematic cuts:

  • Two muons with pT > 20 GeV and || < 2.5 are required for trigger
  • Two additional muons with pT > 7 GeV and || < 2.5

are required

  • One pair of muons of opposite charge is required to have an invariant mass in a window around

the Z mass, defined as mZ  m12

  • The other pair of muons is required to have an invariant mass above a certain threshold defined

as m23 threshold.

The optimised values of the m12 window and of the m34 threshold used for the Higgs-boson mass of 130 GeV are 15 GeV and 20 GeV. When only the Muon System is used (MUID Standalone) the Higgs mass resolution is = 3.12  0.07 GeV. The combination of the Muon System and Inner Detector measurements (MUID Combined) improves the

mass resolution to  = 1.86  0.03 GeV. The reconstructed mass distribution for the 130 GeV Higgs decaysusing MUID Standalone and MUID Combined is shown in Fig.4 (left) and (right) respectively. The mass resolution with MUID Combined is improved respectively to MUID Standalone by 37.5 %.

Fig.4 Higgs invariant mass obtained with MUID Standalone (left) and MUID Combined (right)

VI.Conclusions

The MUID package performs muon combined track reconstruction. Results obtained so far with MUID-MOORE, both for simulated single muon events and for some physics channels, are in good agreement with those obtained with MUID-MuonBox for the Physics TDR [7]. In addition the MUID modular design is well suited for having more than one reconstruction algorithm implemented for easy comparison. An alternative reconstruction method, based on Kalman filter technique, is presently under development in MOORE. The optimization of the low energy muon efficiency based on the Tile Calorimeter muon signal is also in progress in MUID.

VII.References

[1] ATLAS Collaboration, ATLAS Inner Detector Technical Design Report, CERN/LHCC 97-16, CERN/LHC 97-17, April 1997

[2] ATLAS Muon Collaboration, ATLAS Muon Spectrometer Technical Design Report, CERN/LHCC 97-22, May 1997

[3] R. Clifft and A. Poppleton, “iPatRec: Inner Detector pattern-recognition and tracking-fitting” ATLAS Internal Note, ATL-SOFT-94-009 (1994)

[4]

[5] J.Shank et al. “Track Reconstruction in the ATLAS Muon Spectrometer with Moore” ATLAS Internal Note, ATL-SOFT-2003-007 (28 May 2003)

[6] M. Virchaux et al. “MUONBOX a full 3D tracking program for muon reconstruction in the ATLAS Muon Spectrometer”. ATLAS Internal Note, ATL-MUON-97-198 (1997)

[7] ATLAS Detector and Physics Performanec Technical design Report, Volume I, II (25 May 1999), ATLAS TDR 14, 15 CERN/LHCC 99-14, 15

Manuscript received October 30, 2003. This work was supported in part by the EU-FP5 Programme: “Improving Human Research Potential and the Socio-economic Knowledge Base”, Marie Curie Individual Fellowship (HPMF-CT-2002-02102).

Th. Lagouri is with the Nuclear Physics Laboratory, Aristotle University of Thessaloniki, 54124, Thessaloniki, Greece (telephone: 30-2310-998058, e-mail: theodota.lagouri @cern.ch).

J.Shank is with Boston University

D.Adams, K.Assamagan, Y.Fisyak, T.Wenausare with Brookhaven National Laboratory (BNL)

K.Mair, A.Nairz, A.Poppleton, S.Rosati are with European Organization for Nuclear Research (CERN)

L.Spogli is with INFN – Laboratori Nazionali di Frascati

G.Stavropoulos is with U.C. Berkeley

G.Cataldi, E.Gorini, M.Primavera, S.Spagnolo are with INFN and University of Lecce

S.Goldfarb is with University of Michigan

M.Biglietti, G.Carlino, F.Conventi, L.Merola are with INFN and University of Napoli “Federico II”

A.Farilla, M.Verducci are with INFN and University of Roma Tre “E. Amaldi”