ENERGY RECOVERY OF THE CMS ELECTROMAGNETIC CALORIMETER DEAD CHANNELS Daskalakis Georgios, Geralis...

Post on 17-Jan-2016

217 views 0 download

Transcript of ENERGY RECOVERY OF THE CMS ELECTROMAGNETIC CALORIMETER DEAD CHANNELS Daskalakis Georgios, Geralis...

ENERGY RECOVERY OF THE CMS ELECTROMAGNETIC

CALORIMETER DEAD CHANNELS

Daskalakis Georgios, Geralis Theodoros, Kesisoglou Stilianos, Manolakos Ioannis, Eleni Ntomari

1

XXIX Workshop on Recent Advances in Particle Physics and Cosmology

Introduction Description of the method Position Estimation Energy Estimation Conclusions-future plans

Outline

21/04/23 Eleni Ntomari - NCSR Demokritos 2

21/04/23 Eleni Ntomari - NCSR Demokritos 3

CMS detector

ECAL One of the most accurate, distinctive and important subdetectors of the CMS experiment

Measurements of electrons and photons with an excellent energy resolution essential in the search for new physics, in particular for the postulated Higgs boson.

ECAL Endcap

ECAL Barrel

61 200 lead tungstate (PbWO4 ) crystals mounted in the central barrel

7 324 crystals in each of the two endcaps

21/04/23 Eleni Ntomari - NCSR Demokritos 4

ECAL Endcap

ECAL Barrel

21/04/23 Eleni Ntomari - NCSR Demokritos 5

The electromagnetic calorimeter is designed to perform precision measurements aiming to reach 0.5% energy resolution at high energy.

36 supermodules made of 85x20 crystals, each one divided into 4 modules.

Each Endcap is divided into 2 halves and is logically organized in 9 sectors of 40 degrees each.

A preshower detector is placed in front of the endcap crystals.

identify neutral pions in the endcaps within a fiducial region 1.653 < |η| < 2.6.

identification of electrons against minimum ionizing particles

improves the position determination of electrons and photons with high granularity.

Preshower based on Si sensors

Dead Channels-How important is it to develop a recovery algorithm?

~1% of the Electromagnetic Calorimeter Channels present problems (e.g. noisy channels, poor response) ->cannot be used for the energy estimation of the particles that "hit" near them.

21/04/23 Eleni Ntomari - NCSR Demokritos 6

Dead Channels-How important is it to develop a recovery algorithm?

~1% of the Electromagnetic Calorimeter Channels present problems (e.g. noisy channels, poor response) ->cannot be used for the energy estimation of the particles that "hit" near them.

21/04/23 Eleni Ntomari - NCSR Demokritos 7

Crystal 6

8 13 18

7 12 17

6 11 16

Dead Channels-How important is it to develop a recovery algorithm?

~1% of the Electromagnetic Calorimeter Channels present problems (e.g. noisy channels, poor response) ->cannot be used for the energy estimation of the particles that "hit" near them.

21/04/23 Eleni Ntomari - NCSR Demokritos 8

Crystal 7

8 13 18

7 12 17

6 11 16

Method description Effort to develop recovery algorithms, in order to be

able to estimate the energy of these Dead Channels, using the energy of their neighboring functioning crystals

21/04/23 Eleni Ntomari - NCSR Demokritos 9

Build position reconstruction functions using energies from all crystals

in a 5x5 or 3x3 grid, except from the missing one

Method description

21/04/23 Eleni Ntomari - NCSR Demokritos 10

Effort to develop recovery algorithms, in order to be able to estimate the energy of these Dead Channels, using the energy of their neighboring functioning crystals

Build position reconstruction functions using energies from all crystals

in a 5x5 or 3x3 grid, except from the missing one

Build energy correction functions using Monte Carlo

Energy fraction Dead Channel Energy

Method description

21/04/23 Eleni Ntomari - NCSR Demokritos 11

Effort to develop recovery algorithms, in order to be able to estimate the energy of these Dead Channels, using the energy of their neighboring functioning crystals

Build position reconstruction functions using energies from all crystals

in a 5x5 or 3x3 grid, except from the missing one

Build energy correction functions using Monte Carlo

Energy fraction Dead Channel Energy

Apply functions in areas with dead channelsTests with 2010 Collision Data

Method description

Build position reconstruction functions using energies from all crystals

in a 5x5 or 3x3 grid, except from the missing one

Build energy correction functions using Monte Carlo

Energy fraction Dead Channel Energy

Apply functions in areas with dead channelsTests with 2010 Collision Data

Data Samples

/EG/Run2010A-Sep17ReReco-v2/RECO

/Electron/Run2010B-PromptReco-v2/RECO

/EG/Run2010A-Nov4ReReco-v2/RECO

/Electron/Run2010B-Nov4ReReco_v2/RECO

21/04/23 Eleni Ntomari - NCSR Demokritos 12

Effort to develop recovery algorithms, in order to be able to estimate the energy of these Dead Channels, using the energy of their neighboring functioning crystals

8 13 18

7 12 17

6 11 16

Estimate the true position of the hit (photon or electron)

Photon: information of the supercluster

Electron/Positron: information of the supercluster or the tracker

Reconstruction of the event position:

Scurve Method :

Logarithmic weighted method:

Event position reconstruction

η

φ

ii

ii

i

w

xwestimX

8E

Ew ii

80 log

E

Eww i

i

21/04/23 Eleni Ntomari - NCSR Demokritos 13

8 13 18

7 12 17

6 11 16

Estimate the true position of the hit (photon or electron)

Photon: information of the supercluster

Electron/Positron: information of the supercluster or the tracker

Reconstruction of the event position:

Scurve Method :

Logarithmic weighted method:

Event position reconstruction

Most energetic crystal

η

φ

ii

ii

i

w

xwestimX

8E

Ew ii

80 log

E

Eww i

i

21/04/23 Eleni Ntomari - NCSR Demokritos 14

8 13 18

7 12 17

6 11 16

Estimate the true position of the hit (photon or electron)

Photon: information of the supercluster

Electron/Positron: information of the supercluster or the tracker

Reconstruction of the event position:

Scurve Method :

Logarithmic weighted method:

Most energetic crystal

Dead crystal η

φ

ii

ii

i

w

xwestimX

8E

Ew ii

80 log

E

Eww i

i

21/04/23 Eleni Ntomari - NCSR Demokritos 15

Event position reconstruction

8 13 18

7 12 17

6 11 16

Estimate the true position of the hit (photon or electron)

Photon: information of the supercluster

Electron/Positron: information of the supercluster or the tracker

Reconstruction of the event position:

Scurve Method :

Logarithmic weighted method:

Most energetic crystal

Dead crystal η

φ

EstimX [mm]

Tru

eX

-Es

tim

X [

mm

]

EstimY [mm]

Tru

eY

-Es

tim

Y[m

m]

ii

ii

i

w

xwestimX

8E

Ew ii

80 log

E

Eww i

i

21/04/23 Eleni Ntomari - NCSR Demokritos 16

Event position reconstruction

Position Resolution - Crystal 68 13 18

7 12 17

6 11 16

11/04/11 Eleni Ntomari - NCSR Demokritos 17

2010 Collision DATA

Position Resolution - Crystal 68 13 18

7 12 17

6 11 16

11/04/11 Eleni Ntomari - NCSR Demokritos 18

2010 Collision DATA

Position Resolutions - X

21/04/23 Eleni Ntomari - NCSR Demokritos 19

2010 Collision DATA

Energy Correction functions (Monte Carlo e+/e-)

21/04/23 Eleni Ntomari - NCSR Demokritos 20

The most energetic crystal (12) is split in 25 subdivisions

In most of the cases, the energy fraction follows a Gaussian distribution

The Gauss fit mean value is used to extract the constants of the formula that calculates the corrected fraction:

o f(η,φ): energy fraction (fr=Edc/Sum9)

o n, φ: hit coordinates on the crystal

o aij: constants to be defined

Fraction = f(η,φ) = Edc/sum9→Edc = (fraction x sum8 )/(1 – fraction)

4,...,0,,),(,

jiaf j

ji

iij

Energy Correction functions (Monte Carlo e+/e-)

4 9 14 19 24

3 8 13 18 23

2 7 12 17 22

1 6 11 16 21

0 5 10 15 20

21/04/23 Eleni Ntomari - NCSR Demokritos 21

The most energetic crystal (12) is split in 25 subdivisions

In most of the cases, the energy fraction follows a Gaussian distribution

The Gauss fit mean value is used to extract the constants of the formula that calculates the corrected fraction:

o f(η,φ): energy fraction (fr=Edc/Sum9)

o n, φ: hit coordinates on the crystal

o aij: constants to be defined

Fraction = f(η,φ) = Edc/sum9→Edc = (fraction x sum8 )/(1 – fraction)

4,...,0,,),(,

jiaf j

ji

iij

Energy Correction functions (Monte Carlo e+/e-)

4 9 14 19 24

3 8 13 18 23

2 7 12 17 22

1 6 11 16 21

0 5 10 15 20

21/04/23 Eleni Ntomari - NCSR Demokritos 22

The most energetic crystal (12) is split in 25 subdivisions

In most of the cases, the energy fraction follows a Gaussian distribution

The Gauss fit mean value is used to extract the constants of the formula that calculates the corrected fraction:

o f(η,φ): energy fraction (fr=Edc/Sum9)

o n, φ: hit coordinates on the crystal

o aij: constants to be defined

Fraction = f(η,φ) = Edc/sum9→Edc = (fraction x sum8 )/(1 – fraction)

4,...,0,,),(,

jiaf j

ji

iij

Energy Resolutions

Sum8/Sum9

Sum8+Edc/Sum9

8 13 18

7 12 17

6 11 16

21/04/23 Eleni Ntomari - NCSR Demokritos 23

2010 Collis

ion DATA

Energy Resolutions

Sum8/Sum9

Sum8+Edc/Sum9

8 13 18

7 12 17

6 11 16

8 13 18

7 12 17

6 11 16

21/04/23 Eleni Ntomari - NCSR Demokritos 24

2010 Collis

ion DATA

RD: Electrons (Scurve_RD_e+e-, Spline_MC_e+e-, η>0)

11/04/11 Eleni Ntomari - NCSR Demokritos 25

2010 Collision DATA

RD: Positrons (Scurve_RD_e+e-, Spline_MC_e+e-, ceta<0,elept>30, fbrem<0.1)RD: Positrons (Scurve_RD_e+e-, Spline_MC_e+e-, η>0)

11/04/11 Eleni Ntomari - NCSR Demokritos 26

2010 Collision DATA

First analysis with Monte Carlo photons, electrons and positrons gives promising results

Conclusions

11/04/11 Eleni Ntomari - NCSR Demokritos 27

First analysis with Monte Carlo photons, electrons and positrons gives promising results

Tests of this method on Real Data appears to be quite satisfactory for both electrons and positrons, as well as EB+ and EB-

Conclusions

11/04/11 Eleni Ntomari - NCSR Demokritos 28

First analysis with Monte Carlo photons, electrons and positrons gives promising results

Tests of this method on Real Data appears to be quite satisfactory for both electrons and positrons, as well as EB+ and EB-

The correction functions estimate correctly the impact position and the missing energy of the problematic channel

Conclusions

11/04/11 Eleni Ntomari - NCSR Demokritos 29

First analysis with Monte Carlo photons, electrons and positrons gives promising results

Tests of this method on Real Data appears to be quite satisfactory for both electrons and positrons, as well as EB+ and EB-

The correction functions estimate correctly the impact position and the missing energy of the problematic channel

Studies will be extended in the ECAL endcaps

Conclusions

11/04/11 Eleni Ntomari - NCSR Demokritos 30

First analysis with Monte Carlo photons, electrons and positrons gives promising results

Tests of this method on Real Data appears to be quite satisfactory for both electrons and positrons, as well as EB+ and EB-.

The correction functions estimate correctly the impact position and the missing energy of the problematic channel.

Studies will be extended in the ECAL endcaps With more data, it'll be possible to built the position

corrections from data, without any usage of Monte Carlo

Conclusions

11/04/11 Eleni Ntomari - NCSR Demokritos 31

First analysis with Monte Carlo photons, electrons and positrons gives promising results

Tests of this method on Real Data appears to be quite satisfactory for both electrons and positrons, as well as EB+ and EB-.

The correction functions estimate correctly the impact position and the missing energy of the problematic channel.

Studies will be extended in the ECAL endcaps With more data, it'll be possible to built the position

corrections from data, without any usage of Monte Carlo

The ultimate goal is to pass these corrections to CMS framework

Conclusions

11/04/11 Eleni Ntomari - NCSR Demokritos 32

Back up Slides

21/04/23

33

Eleni Ntomari - NCSR Demokritos

21/04/23Eleni Ntomari - NCSR Demokritos

34

Energy Resolutions

21/04/23Eleni Ntomari - NCSR Demokritos

35

Real Data (W)ElectronsPositrons

Scurves from electrons-positrons Real Data Spline from MC electrons-positrons

21/04/23Eleni Ntomari - NCSR Demokritos

36

RD: Electrons (Scurve_RD_e+e-, Spline_MC_e+e-)

21/04/23Eleni Ntomari - NCSR Demokritos

37

RD: Electrons (Scurve_RD_e+e-, Spline_MC_e+e-, ceta>0)

21/04/23Eleni Ntomari - NCSR Demokritos38

RD: Electrons (Scurve_RD_e+e-, Spline_MC_e+e-, ceta<0)

RD: Positrons (Scurve_RD_e+e-, Spline_MC_e+e-)

21/04/23Eleni Ntomari - NCSR Demokritos39

RD: Positrons (Scurve_RD_e+e-, Spline_MC_e+e-, ceta<0,elept>30, fbrem<0.1)

21/04/23Eleni Ntomari - NCSR Demokritos

40

RD: Positrons (Scurve_RD_e+e-, Spline_MC_e+e-, ceta>0)

RD: Positrons (Scurve_RD_e+e-, Spline_MC_e+e-, ceta<0,elept>30, fbrem<0.1)

21/04/23Eleni Ntomari - NCSR Demokritos

41

RD: Positrons (Scurve_RD_e+e-, Spline_MC_e+e-, ceta<0)

Position Resolutions

21/04/23 Eleni Ntomari - NCSR Demokritos 42

Real Data (W)ElectronsPositrons

Scurves from electrons-positrons Real Data Spline from MC electrons-positrons

Real Data- Electrons X-Resolution, Scurve_RD_e+e-, Spline_MC_e+e-, ceta<0

21/04/23 Eleni Ntomari - NCSR Demokritos 43

Real Data- Positrons X-Resolution, Scurve_RD_e+e-, Spline_MC_e+e-, ceta<0

21/04/23 Eleni Ntomari - NCSR Demokritos 44

Real Data- Electrons X-Resolution, Scurve_RD_e+e-, Spline_MC_e+e-, ceta>0

21/04/23 Eleni Ntomari - NCSR Demokritos 45

Real Data- Positrons X-Resolution, Scurve_RD_e+e-, Spline_MC_e+e-, ceta>0

21/04/23 Eleni Ntomari - NCSR Demokritos 46

Real Data- Electrons Y-Resolution, Scurve_RD_e+e-, Spline_MC_e+e-, ceta<0

21/04/23 Eleni Ntomari - NCSR Demokritos 47

Real Data- Positrons Y-Resolution, Scurve_RD_e+e-, Spline_MC_e+e-, ceta<0

21/04/23 Eleni Ntomari - NCSR Demokritos 48

Real Data- Electrons Y-Resolution, Scurve_RD_e+e-, Spline_MC_e+e-, ceta>0

21/04/23 Eleni Ntomari - NCSR Demokritos 49

Real Data- Positrons Y-Resolution, Scurve_RD_e+e-, Spline_MC_e+e-, ceta>0

21/04/23 Eleni Ntomari - NCSR Demokritos 50