ATLAS Data Management: Status

Status/Plans for EM Calibration
S. Rajagopalan
July 13, 2006
ATLAS Week, Stockholm
Outline
 Summarize the present calibration in Athena

Cell and Cluster level corrections
 Short term updates (plan for 12.0.2)
 As-built detector simulation
 Plans for calibration data challenge.
S. Rajagopalan
ATLAS Week, Stockholm, July 2006
2
Cell level calibration
E = (ADC  mA) * (mA  GeV) * ∑ ai (ADCi – ped)
 ai = Optimal filter coefficients
 ADC  mA
ADCDAC from electronic calibration (“Ramp”)
 Mphys/Mcal correction of amplitude bias calibration-physics pulse
 DACmA injected current from calibration (~1/Rinj)

 mA  GeV derived from Test-Beam data

Includes the sampling fraction
 Can also be decomposed from fSampl*(MeV/mA)_Lar


This decomposition relies on fSampl from Geant4
For EMEC: MeV/mA = MeV/mA_0 * b /(1+a(-0) )
 a, b derived from TB data
 Current digitization/reconstruction uses sampling fraction derived from
simulation, the rest from TB02 measurements
S. Rajagopalan
ATLAS Week, Stockholm, July 2006
3
Cell Level Calibration (2)
 Calibration applied at ROD and HLT would probably be simpler.

Studying the possibility to apply “linear calibration” at the RODs and
applying a refined higher order electronics calibration and correcting for
pedestal shifts during offline reconstruction.
 HV corrections (dead on one side or non-nominal)
 Corrections for pathological cells (noisy, dead)
 Correction for temperature variation
 Correction for argon purity
 Inter-calibration between calorimeter regions obtained from Z analysis.
 Most of these have yet to be implemented in the current reconstruction.

But the software structure allows for easy plugi-in’s of correction tools
S. Rajagopalan
ATLAS Week, Stockholm, July 2006
4
Cluster Level Corrections
 Two clustering algorithms are used:

Sliding Window algorithm producing EM clusters in 5x5, 3x5, 3x7

Topological clustering with seed 6,3,0

Being studied for e-gamma reconstruction..
 Sliding Window clusters are corrected for:

Eta and phi position corrections

Energy modulations vs eta, phi

Lateral out of cone energy corrections

Longitudinal corrections including upstream matter & leakage

Gap corrections, if relevant

Correct for residual HV effects and pathological cells.

Overall energy scale
 Corrections in 12.0.1 all derived from DC2 single electrons
S. Rajagopalan
ATLAS Week, Stockholm, July 2006
5
S-shape corrections
Finite granularity of middle sampling (0.025x0.025) not small compared to shower width
Simple energy weighted position () measurement pulled toward middle of cell
Corrections derived from single electrons (Snyder)
Snyder
S. Rajagopalan
ATLAS Week, Stockholm, July 2006
6
Energy modulation
S. Snyder
Energy modulations as a function of phi
Energy modulations as a function of eta
Derived for different eta positions
0.1 to 0.2% effect
Derived for different cone sizes and eta bins
S. Rajagopalan
ATLAS Week, Stockholm, July 2006
7
Longitudinal weights
Weights extracted via a chi2 fit on single electrons (S. Paganis)
 Will be updated for 12.0.2 with latest geometry tag: DC3-05
N
2  
E
i
i
rec
i

 Etrue
2
 E2
i
true
Erec   off  w0 E0  E1  E2  w3 E3 
  2.5
 , off , w0 , w3 only functions of 
Different parametrization in gap region (gap scintillator energy is used):
E rec  par[0]EEMB  EEMEC  par[1]E gap _ sci  par[2]
S. Rajagopalan
1.3    1.6
ATLAS Week, Stockholm, July 2006
8
Longitudinal weights
A more sophisticated expression has been demonstrated to improve
linearity and good resolution at TB02 (T. Carli)
Vis
Vis
E rec  (a( E )  b( E ).EPS
 c( E )( EPS
.E1vis ) 0.5  d ( E ).i 1,3 Eicalo ).(1  f leak (depth )). f brem (E). f cell impact
E loss upstream of PS
1.5 X0, 3.6 %@10 GeV
E loss between
calo sampling
PS and calo
fraction+ lateral
0.9 X0, 4.1 %@10 GeV leakage E dependent
T. Carli
S. Rajagopalan
Longitudinal
leakage
> 30 X0, 0.3 %@10 GeV
Better than 0.1 % over 20-180 GeV:
 Done only at one  position in a
setup with less material than in
ATLAS and no B field
 Implementation of this method in
Athena is being explored by
L.Carminati, L. Mandelli, et. al.
using “calibration hits”
ATLAS Week, Stockholm, July 2006
9
Gap corrections
 Gap corrections re-derived for 12.0.0 using gap scinitillators (S.
Paganis, J. Hoffman)

No gap scintillators in early Rome reconstruction
 Comparison of DE/Etrue in Release 11.5.0 and 12.0.0:
12.0.1
11.5.0
Mean = -0.107±0.015
σ = 0.123±0.023
Mean = -0.015±0.007
σ = 0.066±0.009
Hoffman
S. Rajagopalan
ATLAS Week, Stockholm, July 2006
10
Performance
(using simulation from 11.0.4x)
 Slightly worse than TDR because of increased material
 All corrections need to be re-derived and optimized as well.
S. Paganis
S. Paganis
Need to use Zee for intercalibration of regions
448 regions in ATLAS (denoted by i)
Eireco = Eitrue(1+αi)
Mijreco =MZ(1+(αi+αj)/2)
fit to reference distribution abd propagate αi
S. Rajagopalan
ATLAS Week, Stockholm, July 2006
11
Short term plan
 The present 12.0.1 release have corrections that were derived
using DC2 simulation, single electrons.

These need to be re-derived as the simulation has acquired several
“as-built” realistic features.
Being re-done with 11.0.x simulation, Will be available for 12.0.x
 Will also need to be re-derived with 12.0.x (as-built simulation).


Corrections are also different for electrons & photons

Software infrastructure in place to implement different e-g
corrections, Will try to get it in 12.0.2
S. Rajagopalan
ATLAS Week, Stockholm, July 2006
12
As-built geometry in simulation
 Extensive work has gone into updating the simulation with:
 Material (M. Thoiye, V. Tsulai):

Solenoid supports like rails and titanium blocks.

Electronic crates mounted on barrel endcaps.

cables from the inner detector in the crack region.

radiation shielding in the FCAL region.

moderator and MBTS on the endcap.

Geometries have been thoroughly reviewed by sub-system experts

Some material on the endcap itself is not sufficiently declashed
(overlapping volumes in heaters, cables, crates)
 in progress for 12.0.2
S. Rajagopalan
ATLAS Week, Stockholm, July 2006
13
As-built simulation
(mis-alignments)

Misalignments for the barrel are in database (V. Tsulai, G. Unal).

Misalignments of large pieces:

Cryostats, EMBarrel (pos$neg), Solenoid

Translation and rotation
Typical misalignments introduced and tested.
 Simulation samples generated, reconstructed – being studied
(S.Laplace, R. Zitoun)



preliminary: z deplacement of cryostat 4cm reconstructed
senstivity to translation/tilt 1mm/mrad with 10k Z

Endcap misalignments pending a resolution of the cryostat issue

Sagging of electrodes and absorbers introduced (Parrour, Unal)

Establish a standard way to specify mis-alignments in DB
 Will be done for 12.0.2 (Boudreau)
Deformation (pear shape) of calorimeter NOT planned until Rel. 13

S. Rajagopalan
ATLAS Week, Stockholm, July 2006
14
As-built simulation
 Response

Decrease of the response for charge deposited close to EMEC
electrode (V. Niess, G. Unal).

HV imperfections introduced; our first use of Conditions DB in
simulation (V. Tsulai, G. Unal)
 Calibration Hits (M. Leltchouk, G. Pospelov et. al.)

Allows saving of energy deposited in active (LAr) and inactive
(absorbers) and dead material (such as upstream material) in fine
granularity.

Dead material hits produced by default in Release 12.

Broken into EM, Hadronic, invisible or escaped energy deposits.

Extensively used for EM and hadronic calibration.

EM sampling fraction & longitudinal weights being studied in detail by
using calibration hits (L. Carminati et.al.)

Local hadron calibration weights using calibration hits (S. Menke et. al.)
S. Rajagopalan
ATLAS Week, Stockholm, July 2006
15
Plans for CDC
 Re-derive the cluster corrections from as-built geometry in 12.0.x

Present correction are DC2 based and being re-derived with 11.0.42

Likely to lead to deterioration of resolution in end-cap due to
increased material. Need to handle this better in reconstruction.
 Simulate events with following conditions and develop
procedures on how to correct for them in reconstruction

Introduce realistic misalignments between large pieces.

Electrode/Absorber sagging

Introduce mis-calibration (but perhaps can be done post-digitization)

HV pathologies

Increased up-stream matter
 Effects would be studied on samples of single e, g, Zee
S. Rajagopalan
ATLAS Week, Stockholm, July 2006
16
Concluding Remark
Calorimeter Calibration Workshop, Sant Feliu, Sept. 5-8, 2006

Organized by Martine Bosman, Ilya Korolkov
 Complete review of the present calibration.

Electronics, EM, Hadronic, in-situ, Trigger


Simulation, Testbeam, commissioning
Discussion on calibration strategies during early running
 An “oversight panel” has been formed to closely look at our present
work, help promote discussions and help us plan the road ahead:

M. Cavalli-Sforza, T. Davidek, N. Ellis, D. Froidevaux, D. Lissauer, J. Stark
S. Rajagopalan
ATLAS Week, Stockholm, July 2006
17