Jet Reconstruction and Calibration in Athena

1
Jet Reconstruction and Calibration in Athena

• US ATLAS Software Workshop
• BNL, 27/08/03
• Ambreesh Gupta, for the JetRec Group
• University of Chicago
• Outline
• Jet Reconstruction Setup and Algorithms
• Jet Calibration
• Some Highlights from JetRec Meeting in Barcelona.

2
Jet Physics at Hadron Colliders
Final state partons produce a collimated jet of
particles.
An ideal algorithm will map jets with final state
partons.
Jet algorithms sensitive to experimental and
theoretical uncertainty.
3
The Steps of Jet Measurement
Jet Measurement can be broadly divided in three
steps
Jet Reconstruction
Energy deposit in the calorimeter is clustered by
Jet reconstruction algorithms, e.g Cone and Kt.
Energy Calibration
Energy of a Jet is calibrated for
non-compensation, dead material, magnetic field,
etc.
Flavor Identification
b-tag, ??-tag, etc.
4
Jet Algorithms
Jet algorithms have two steps Clustering
hadrons, calorimeter tower etc., for
nearness Nearness in Angle gt Cone
Algorithm Nearness in Momentum gt Kt
Algorithm Recombination Scheme is the momentum
addition rule for the set of particles that have
been identified with a jet.

• 4-vector addition
• Snow mass scheme
• . . .

5
Jet Reconstruction Packages
6
Writing Multiple Jets to the N-tuple
CaloTower ProtoJet
Truth Particle ProtoJet
Cone Algorithm Seed less Cone
Pre-Cluster
Kt Algorithm
Split-Merge
ParticleInJet
Multiple algorithm output can be stored with
different names
All algorithms can be configured to run in the
same job
CBNT
7
Jet Samples Used
The results reported use the following two
samples - Multi-jet sample produced with
different Pt cut values ( 35,70,.., 540
GeV) -SUSY data challenge sample. Run 002315.
8
Comparing Jet Multiplicity
Mean 5.7
Mean 5.6
Mean 6.0
Number of Jets
9
Comparing Jet Transverse Energy
Mean 114 GeV
Mean 118 GeV
Mean 126 GeV
Transverse Energy
10
Reconstructed jet energy In bins if true energy
at eight energy points. Good fits to gaussians.
11
Jet Energy Calibration
To account for non-compensation, dead material
etc., - H1 scheme (Implemented by Frank
Paige) - Sampling based corrections (in
development)
12
Navigating Composite Objects
Jet algorithms work with abstractions like
ProtoJet. In order to do energy correction, the
navigating the composition of Jet needed.
13
H1 Calibration

• H1 Strategy EM showers are much denser than
  hadronic ones.
• Weight high E cell with EM scale and low E
  cell with hadronic scale.
• H1 weights derived from di-jet samples. Same
  weights improve
• missing Et resolution.

14
Jet Calibration Further developments
Further studies to develop tools for calculating
and storing weights. - requires reading large
number (100K) of jets - calculate weights
(interface to a minimizer) - store the weights
to be accessed for calibration. Sampling based
weights - calculated in different eta region
of the detector. - first set of weights
obtained. Not yet in accessible in Athena.
15
Sampling based Correction
16
Evolution of JetRec Barcelona Meeting

• Since, the last JetRec working group meeting in
  Tucson (Aug
• 2001), a lot of progress made in Jet
  reconstruction software.
• Another working group meeting was held in
  Barcelona ( July
• 2003 ) to review the RTF recommendations.
  Attended by
• Present Frank Merritt, Tom LeCompte, Peter
  Loch, Srini
• Rajagopalan, Martine Bosman.
• By Phone F. Paige, D. Cavalli, S. Resconi,
  D. Rousseau, A. Gupta,
• S. George, P. Calafiura.
• Initial draft of meeting report can be found at
  JetRec meetings
• minutes web-page ( date 30th of July)

17
Barcelona Meeting Highlights

• EnergyCluster class will replace the existing
  ProtoJet class.
• The nature of abstraction remains the same, as
  far as Jet
• algorithms are concerned, but it conforms to
  the general Athena
• reconstruction setup and hence will utilize
  common services.
• The Jet Data Objects will conform to the general
  Athena scheme
• to navigate composite objects.
• Associating information with Jets, e.g, tracks
  in jet, will be
• provided by extending the present Jet class
  interface with the
• help of a templated method for the type of
  association requested.

18
Summary

• Jet Reconstruction packages are in fully
  functional state.
• There have been various user feedback using dc1
  dataset
• bugs hunted and fixed.
• With the availability of noise and pileup, a
  more realistic
• test of jet algorithms possible.
• RTF recommendations reviewed at Barcelona
  meeting.
• Structural changes proposed to conform with
  general Athena
• reconstruction setup.