Etmiss Reconstruction and Calibration

1

• Etmiss Reconstruction and Calibration
• in Athena
• Workshop on Energy
  Calibration
• of the ATLAS
  calorimeters
• July
  23, 2002

2

• Reconstruction Calibration
  
• Starting point
• ? Etmiss reconstruction from all
  calorimeter cells
• ? Comparison of reconstructed Etmiss
  quantities per
• calorimeter and total between Athena
  and Atrecon
• To have an Etmiss value usable for physics
  studies
• ? Calibration
• ? Contribution of muons and eventually
  not full-simulated
• calorimeter regions
• To evaluate Etmiss resolution
• ? MC truth Etmiss
• ? Calculation of SumEt ? ( Etmiss) ?
  SumEt
• Calibration
• ? Region Calibration , H1-Style
  Calibration
• ? Comparison with the refined calibration
  method applied in
• ATLAS Physics Technical Design
  Report, TDR Calibration

3
Etmiss reconstruction from
calorimeters

• Next slide
• Comparison of not calibrated Etmiss_calo between
  Athena and Atrecon
• ? On a sample of TDR full simulated data
  (98_2)
• bbA? ? ??? ? - had m(A? )150
  GeV , 665 evts
• ? full simulation within ? lt 3
• ? read with Atrecon release 1.3.0
  (atrecon-01-42-29)
• ? and Athena release 3.1.0, used
  RecExample
• new MissingET package (now in current
  release )
• ? No energy threshold applied (no noise
  added yet !!)

Ex(y)miss_calo - ? Ex(y) performed on calo
cells
4
Athena - Atrecon superimposed (!!) distributions
of ETmiss in calos
1
2
EM Barrel
PS Barrel
3
4
Tile
EM EndCap
5
7
HEC
PS EndCap (only Athena histo)
5

• Forward Calorimeter (Calo 6) has not been full
  simulated in this event sample
• Perfect agreement of Etmiss_calo reconstruction
  in Calorimeters between Athena and Atrecon (!!)
• PS EndCap contribution not included in Atrecon
  Etmiss routine (EMISREC) , so the total not
  calibrated reconstructed Etmiss is slightly
  different between Athena and Atrecon

Athena
Atrecon
lt gt 46.9
lt gt 47.3
Total Etmiss ?lt3
Total Etmiss ?lt3
6
Contributions implemented in the Athena MissingET
package To
obtain an Etmiss value usable for physics
analysis

• (1) Calibration a first rough method
  Region Calibration
• 7 calibration constants (
  MissingETRegionCalib class)
• ? are applied to the partial contributions from
  the calorimeters
• Ex(y)miss_reg_calib const( 1?7 )
  ? Ex(y)miss_calo ( 1?7 )
• Etmiss_reg_calib sqrt
  (Exmiss_calib ?? 2 Eymiss_calib ?? 2)
• ? have been obtained minimizing Etmiss
  resolution of bbA? ? ??
• new calibrated quantities added to output CBNT
• possibility to change default calibration
  constants in the JobOptions

7
Athena
lt gt 47.3
? 9
- - - TRUTH Etmiss
Etmiss before calib ?lt3
Etmiss Resol. before calib ?lt3
lt gt 57
? 5.4
in CBNT
Etmiss after calib ?lt3
Etmiss Resol. after calib ?lt3
8
Etmiss Resolution?lt3 ?( Ex(y)miss Rec
?lt3 - Ex(y)miss Truth ?lt3 ) Ex(y)miss
Truth - ? px(y) of stable parts except
neutrinos and ?s

• (2) Contribution of not full-simulated
  calorimeter regions
• ? when the full simulation is not
  performed till ? 5
• ? accessing MC truth info
• (3) Muon contribution
• ? accessing MC truth info (will substituted
  by
• reconstructed muons when available in
  Athena)
• (4) MC truth Etmiss
• ? to calculate directly from CBNT the
  Etmiss resolution
• (5) Calculation of SumEt ?ET calo cells
  within ?lt 5 and
• Calibration SumEt_reg_calib const(
  1?7 ) ? SumEt_calo ( 1?7 )

To evaluate Etmiss resolution

? ( Etmiss ) ? SumEt
9

• All Etmiss MC truth quantities calculated in
• CBNT_MissingETtruth class in Athena MissingEt
  package
• a separate CBNT algorithm for Etmiss MC truth
  quantities
• from the reconstructed ones
• calculation of
• ? true Etmiss from neutrinos (no ?
  cut)
• ? Etmiss seen in calorimeters
• ? from stable particles in calos ( no
  neutrinos , no ? s)
• ? for 2 ?-ranges
• ? with ?-range lt etafullsim
• ? with etafullsim lt ?-range lt 5
• etafullsim parameter ?-range
  of full simulation,
• 
  set by the user in the JobOptions
• ? truth muon contribution
• from "measurable" muons with pt gt 6
  GeV and ? lt 2.5

not yet in release
10
Etmiss quantities to be used in physics analysis
lt gt 24.4
lt gt 25.3
in CBNT
Etmiss to be used for physics
Etmiss Truth (from neutrinos)
Etmiss Calib forward (p.l.) ? (p.l.)
? 6.8
lt gt 185
Sumet to be used for physics
Final Etmiss Resolution
SumET Calib forward (p.l.)
11

• comparing with the results published in the TDR,
  that have been obtained with a more refined
  calibration method
  
• (1) TDR
  Calibration
• a slightly worse resolution has been obtained
  with this first rough calibration method
  implemented in Athena
• (2) Region
  Calibration
• some promising results with an other calibration
  approach implemented in Athena
• (3) H1-Style
  calibration
• comparison between those different approaches on
  a wider statistic (preliminary results !!)
• 
  More in details

final Etmiss Resolution ? ( Ex(y)miss Rec
?lt 5 - Ex(y)miss Truth (no ? cut) ) takes
into account also detector coverage
12

• (1) TDR Calibration
• Results published in the ATLAS Physics Technical
  Design Report (TDR) have been obtained using
  different set of calibration constants for
  hadronic cluster cells, em cluster cells and
  cells outside clusters in each calorimeter region
• This means to
• ? start from the identified objects (jets, e/?
  clusters, tau jets), and calibrate all
  associated cells ,
• ? calibrate cells outside clusters (using
  minimum bias events)
• ? sum all contributions at cell level to
  calculate final Etmiss.
• ? If the identified objects are already
  correctly calibrated, in principle, the final
  Etmiss is automatically calibrated too, only
  cells outside clusters contribution should be
  corrected
• This refined calibration method can be applied
  only after cluster identification and
  calibrationmore work needed to implement all
  that in Athena

13
Published in ATLAS physics TDR - 1999 Etmiss
Resolution 0.46 ? SumEt
bbA? ? ?? evts Noise
added Ecell gt 1.5 ?(Noise)
Etmiss Resolution ? ( Ex(y)miss Rec ?lt 5 -
Ex(y)miss Truth (no ? cut) ) SumEt
?ET calo cells within ?lt 5
14

• (2) Region Calibration in Athena
• is an approximation of the complete Etmiss
  calibration
• the present constants are average constants
  for cells in em clusters, hadronic
  clusters and outside clusters
• calibration constants depend on physics event
  type and energy (!!) 3 sets of calibration
  constants have been provided for low, middle and
  high range of A? masses
• users can calculate calibration constants for
  their own physics events using Etmiss truth
  quantities from CBNT
• This is a rough calibration method that has been
  implemented in Athena as first step to provide
  soon a usable Etmiss for
• physics

15

• (3) H1-Style Calibration in Athena
• First attempt to calibrate Etmiss with H1-Style
  calibration method in Athena
• Simple approach
• ? used available weights tuned for jets
  (see next slide)
• New calibration class in MissingET package
• (MissingETH1Calib class )
• ? apply H1 weights to all calorimeters
  cells entering in the Etmiss
• calculation
• ? 4 new calibrated quantities in output
  CBNT
• Ex(y)miss_H1_calib, Etmiss_H1_calib,
  SumEt_H1_calib

not yet in release
16

• (3) H1-Style Calibration in Athena
• H1 weights
• ? from Table 1 of note H1-Style Jet
  Calibration in Athena ,
• by H. Ma, F.E. Paige, S. Rajagopalan
  ( 3 May 2002 )
• ? tuned for jets (JET35 sample, TDR
  geometry)
• ? depend on cell energy and
  Calorimeter region
• ? ? dependency is not yet taken into
  account, as first
• approximation
• ? for the moment hardwired in the code
  (still to think how to
• make them accessible from the
  jobOptions)
• Comparison between Etmiss Region Calibration and
  Etmiss H1-Style Calibration on 2 mass points
• m(A? )150, 800 GeV
• 
  ? preliminary results...

17
m(A?)150 GeV
Etmiss ltREG/truthgt 0.997
lt gt 57
? 5.4
- - - TRUTH Etmiss
Etmiss after REG calib ?lt3
Etmiss Resol. After REG calib ?lt3
lt gt 55.5
? 4.5
Etmiss ltH1/truthgt 0.965
Etmiss after H1 calib ?lt3
Etmiss Resol. After H1 calib ?lt3
18
m(A?)800 GeV
Etmiss ltREG/truthgt 1.040
lt gt 103.2
? 13.2
- - - TRUTH Etmiss
Etmiss after REG calib ?lt3
Etmiss Resol. after REG calib ?lt3
lt gt 99.72
? 11.4
Etmiss ltH1/truthgt 1.004
Etmiss after H1 calib ?lt3
Etmiss Resol. after H1 calib ?lt3
19

• Event samples
• ? physics channels bbA? ? ??? ? - had,
  
• 
  bbA? ? ??? had - had
  
• ? A? mass points m(A? )150, 300,
  450, 600, 800 GeV
  
• ? total number of evts 5650
• TDR full simulated data (98_2) within ? lt
  3, read with
• Athena release 3.2.0, used RecExample
  new
• MissingET package

Preliminary comparison of these 2 new
implemented calibration methods in Athena with
Atlas physics TDR results on a wider statistic
20
REGION TDR calibration
H1-Style TDR calibration
No Noise addition in Athena results (not yet
available) !!
Etmiss Resolution ? ( Ex(y)miss Rec ?lt 5 -
Ex(y)miss Truth (no ? cut) ) SumEt
?ET calo cells within ?lt 5
preliminary !!
21
Last Minute Results...
H1-Style
H1-Style
H1-Style TDR calibration
Added an event sample with ?? e decays ?
physics channels bbA? ? ??? e - had
? total number of evts 2500 (low
statistic)
Red curve all decay channels weighted with
the right decay BR BR(bbA? ? ??? e - X) 30
22
Conclusions

• ? H1-Style Calibration seems a very
• powerful and promising method to
• calibrate Etmiss
• Etmiss resolution seems in
• nice agreement with TDR results
• (but...no noise added !)
• Etmiss mean value in agreement
• with the Truth mean value within
• few
• H1-Style Calibration results to be refined
• ? on more statistic
• ? with noise addition (when it will be
  available in Athena)
• expected a worsening of the results
  (!!)
• ? re-tuning H1 weights (?)
  
• ? TDR Calibration to be implemented ...
  
  
• 
  WORK IN PROGRESS...

H1-Style TDR calibration
23
From talk by Srini, ATLAS Software Week, May 30
2002