Particle Distributions and Correlations in ATLAS · On behalf of the ATLAS Collaboration Particle...

Cristina Oropeza Barrera, University of GlasgowOn behalf of the ATLAS Collaboration

Particle Distributions and Correlations in ATLAS

1DIS Conference - 13.April.2011

Outline

2

• Motivation

• Data Samples and Event Selection

• Correlations in ATLAS:

• Δϕ Correlations https://atlas.web.cern.ch/Atlas/GROUPS/PHYSICS/CONFNOTES/ATLAS-CONF-2010-082/

• Inclusive Two-Particle Angular Correlations

• Summary

DIS Conference - 13.April.2011

Not discussed here

https://atlas.web.cern.ch/Atlas/GROUPS/PHYSICS/CONFNOTES/ATLAS-CONF-2010-082/



Motivation

3

• Models to describe dynamics of multi-particle production are incomplete: limited explanation of emission of soft radiation.

• Study of correlations between final state particles allows us to investigate the underlying mechanisms of particle production at LHC energies.

• Identify important dynamical information that can be incorporated in models to gain a better and more global picture (tuning).

final-state particles

incoming partons

?


Data Samples and Event Selection

4

The data samples used in this analysis:

๏ ATLAS Inner Detector fully operational and solenoid at 2T,๏ triggered by a single-arm, level 1 Minimum Bias Trigger Scintillator,๏ at least one primary vertex,๏ if there is a second vertex it should not be associated to more than four tracks (to remove

events with more than one interaction per bunch crossing),๏ to contain at least two tracks in the phase-space:

๏ pT > 100 MeV ๏ |η| < 2.5

Event Requirements

Same as minimum bias analysis - H. Schulz talk

Energy Integrated Luminosity

900 GeV 7 μb-1

7 TeV 190 μb-1

Limited dataset, low luminosity, low pile-up


Pixel Detector

SCT TRT

Inclusive Two-Particle Angular Correlations

5

ATLAS Note: “Measurement of Inclusive Two-Particle Angular Correlations in Proton-Proton Collisions at √s = 900 GeV and 7 TeV” https://atlas.web.cern.ch/Atlas/GROUPS/PHYSICS/CONFNOTES/ATLAS-CONF-2011-055/




Analysis Overview

6

The inclusive two-particle angular correlation function is given by:

where〈...〉ch indicates an average over contributions from all particle multiplicities.

Correlations between emissions in a single event.Normalised by the total number of events.

Distribution of uncorrelated pairs. Particles pairs made from independent events. Normalised by its integral.

Nch is the average particle multiplicity.


Correction Procedure

7

An event weight:To account for inefficiencies in the vertex and trigger selection, the foreground and multiplicity distributions were weighted event-by-event with:

The effect of tracking inefficiencies is corrected for using a data-driven method.

Find all good tracks

Calculate the observable:

R(Δη,Δϕ) or R(Δη) or R(Δϕ)

Determine the tracking efficiency E(pT,η) for each track.

Generate a random number r and compare it to E(pT,η).

If r > E(pT,η), remove the track.



8

Each iteration corresponds to an additional application of the detector effect on the data.The -1 iteration corresponds to the observable when no detector effects are present.


-5 -4 -3 -2 -1 0 1 2 3 4 5

)R

(

-2

-1.5

-1

-0.5

0

0.5

1

1.5

2

2.5all selected tracks

1st iteration

2nd iteration

3rd iteration4th iteration

5th iteration

6th iteration

ATLAS Preliminary900 GeV Data

-5 -4 -3 -2 -1 0 1 2 3 4 5

)R

(

-2

-1.5

-1

-0.5

0

0.5

1

1.5

2

2.5

all selected tracks1st iteration2nd iteration3rd iteration4th iteration5th iteration6th iteration

ATLAS Preliminary7 TeV Data



9

The value of (each bin of) the observable is plotted as a function of the iteration number (0, 1, ..., N) and fitted using a third-degree polynomial. By extrapolating this fit to -1, an estimate of the true value can be made.

Testing method in Monte Carlo:

Number of Iterations-1 0 1 2 3 4 5 6

< 0

.2)

Valu

e of

bin

26

(0.0

<

0.5

1

1.5

2

2.5

3

ATLAS Preliminary

Iterations

Extrapolation

7 TeV Data

Number of Iterations-1 0 1 2 3 4 5 6

< 4

.8)

Valu

e of

bin

49

(4.6

<

-1.6

-1.4

-1.2

-1

-0.8

-0.6

-0.4

-0.2

0

ATLAS Preliminary

Iterations

Extrapolation

7 TeV Data

)R

(

-1.5

-1

-0.5

0

0.5

1

1.5

2

2.5

Truth MC

Corrected MC

ATLAS Preliminary7 TeV Monte Carlo (MC09)

-4 -2 0 2 4

corr

) -

R(

truth

)R

(

-0.2

-0.15

-0.1

-0.05

0

0.05

0.1

0.15

0.2

Truth MCCorrected MC

Statistical and Systematic Uncertainties

10

I. Extrapolation to N=-1The statistical error in the corrected value will be the result of propagating the statistical uncertainties in the parameters of the fit.

II. Uncertainties on the EfficienciesDetermined by varying the efficiencies up or down and propagating through the analysis.

III. Non-closure in Monte Carlo

MC studies on the absolute difference between truth and corrected MC as a function of Rcorr in different models.

An absolute uncertainty of 0.05 is assigned to all bins of R in data.


corr)R(

-1.5 -1 -0.5 0 0.5 1 1.5 2 2.5

corr

) -

R(

truth

)R

(

-0.3

-0.2

-0.1

0

0.1

0.2

0.3

0.4

MC09 TuneDW TunePerugia0 TunePhojet TunePythia8 Tune

ATLAS Preliminary7 TeV Monte Carlo

Results

11

Corrected inclusive two-particle correlation functions in Δη and Δϕ. Same complex structure is seen in Monte Carlo, MC09 Tune, however the strength of the correlation seen in data is not reproduced.DIS Conference - 13.April.2011

R(Δη,Δϕ)

Results

12

Near-side correlations: integrating 0 < Δϕ < π/2.


)R

(

3

-2

-1

0

1

2

3

4

5

Data

MC09 TuneATLAS Preliminary

900 GeV

2 < 0 <

-4 -2 0 2 4

MC

) -

R(

Dat

a)

R(

-2

-1.5

-1

-0.5

0

0.5

1

1.5

2

MC09 Tune)

R(

3

-2

-1

0

1

2

3

4

5

DataMC09 TuneDW TunePhojet TunePerugia0 TunePythia8 Tune

ATLAS Preliminary

7 TeV

2 < 0 <

-4 -2 0 2 4

MC

) -

R(

Dat

a)

R(

-2

-1.5

-1

-0.5

0

0.5

1

1.5

2MC09 TuneDW TunePhojet TunePerugia0 TunePythia8 Tune

R(Δη)

Dominated by the peak at (0,0). At 7 TeV, Pythia 8 and Phojet have better agreement in the tails of the distribution while MC09 is closer in the peak.

Results

13

Away-side correlations: integrating π/2 < Δϕ < π.


R(Δη)

Dominated by the ridge structure around Δϕ=π. With the exception of DW, the tunes seem to perform better in these distributions.

)R

(

3

-2

-1

0

1

2

3

4

5

Data


900 GeV

< < 2

-4 -2 0 2 4

MC

) -

R(

Dat

a)

R(

-2

-1.5

-1

-0.5

0

0.5

1

1.5

2

MC09 Tune)

R(

3

-2

-1

0

1

2

3

4

5

Data

MC09 Tune

DW TunePhojet Tune

Perugia0 Tune

Pythia8 Tune

ATLAS Preliminary

7 TeV

< < 2

-4 -2 0 2 4

MC

) -

R(

Dat

a)

R(

-2

-1.5

-1

-0.5

0

0.5

1

1.5


Results

14

Short-range correlations: integrating 0 < Δη < 2.


R(Δϕ)

Two-peak structure. Similar to underlying event distributions. Back-to-back recoil. Most of the tunes agree well with data in a small region around Δϕ=π.

)R

(

0 5

0

0.5

1

1.5

2

2.5

3

Data


900 GeV

< 20 <

-1 0 1 2 3 4

MC

) -

R(

Dat

a)

R(

-2

-1.5

-1

-0.5

0

0.5

1

1.5

2

MC09 Tune

)R

(

0 5

0

0.5

1

1.5

2

2.5

3

DataMC09 TuneDW TunePerugia0 TunePhojet TunePythia8 Tune

ATLAS Preliminary

7 TeV

< 20 <

-1 0 1 2 3 4

MC

) -

R(

Dat

a)

R(

-2

-1.5

-1

-0.5

0

0.5

1

1.5

2

MC09 TuneDW TunePhojet TunePerugia0 TunePythia8 Tune

Results

15

Long-range correlations: integrating 2 < Δη < 5.


R(Δϕ)

Underlying structure away from the peak at (0,0). The absolute difference between data and the different models is flat across Δϕ. Pythia 8 is closest and DW (old tune) is worst.

)R

(

2 5

-2

-1.5

-1

-0.5

0

0.5

1

Data

MC09 Tune

ATLAS Preliminary

900 GeV

< 52 <

-1 0 1 2 3 4

MC

) -

R(

Dat

a)

R(

-2

-1.5

-1

-0.5

0

0.5

1

1.5

2

MC09 Tune)

R(

2 5

-2

-1.5

-1

-0.5

0

0.5

1

DataMC09 TuneDW TunePerugia0 TunePhojet TunePythia8 Tune

ATLAS Preliminary

7 TeV

< 52 <

-1 0 1 2 3 4

MC

) -

R(

Dat

a)

R(

-2

-1.5

-1

-0.5

0

0.5

1

1.5


Summary

16

Exciting times! Many more measurements to come!


• The two-particle angular correlation function in Δη and Δϕ has been measured for pT inclusive minimum bias events in pp collisions at 900 GeV and 7 TeV.

• A complex structure was observed at both energies. It was explored in more detail by projecting the two-dimensional distribution into both Δη and Δϕ.

• The results have been compared to different Monte Carlo tunes: MC09, Phojet, DW, Perugia0 and Pythia 8 (further information on these tunes in Extra Slides).

• None of the models reproduce the strength of the correlations seen in data. The Pythia 8 tune at 7 TeV is the closest in all distributions.

Extra Slides

17DIS Conference - 13.April.2011

Monte Carlo Models - 2PC analysis

18NPPD Conference - 04.April.2011

• Pythia 6.4.21 tunes:

• MC09: produced by the ATLAS Collaboration to describe a range of minimum bias and underlying event data from the Tevatron; uses MRST LO* PDF.

• DW: older tune to CDF underlying event and Drell-Yan data; uses the older virtuality-ordered shower and non-interleaved MPI model.

• Perugia0: tuned to Tevatron; uses CTEQ 5L PDF and the new pT ordered shower and the MPI is interleaved with the initial state radiation.

• Phojet 1.12.1.35: separate hard and soft diffractive contributions; not yet tuned to recent experimental data.

• Pythia 8.130: adds to the MPI model of Pythia 6 by also interleaving the final state radiation; includes an updated model for diffraction that allows harder colour singlet exchange; uses CTEQ 5L PDF.

Total Uncertainties for 2D distributions - 2PC

19

900 GeV

Total Uncertainties for 2D distributions - 2PC

20

7 TeV

Δϕ Correlations

21

ATLAS Note: “Angular correlations between charged particles from proton-proton collisions at √s = 900 GeV and √s = 7 TeV measured with ATLAS detector” https://atlas.web.cern.ch/Atlas/GROUPS/PHYSICS/CONFNOTES/ATLAS-CONF-2010-082/




Analysis Overview

22

Some definitions:leading particle - charged particle with highest transverse momentum (pT).Δϕ - azimuthal angle difference. Unsigned angle in the transverse plane (x-y) between the leading particle and any non-leading particle.

one of the other selected tracks

ΔΦ crest shape observable

Δϕ for all non-leading particles in an event.

The minimum of the distribution is extracted from a 2nd degree-polynomial fit.

Subtract the minimum from each bin and normalise to unit area.

0 0.5 1 1.5 2 2.5 3

/50)

/ (

T N

0200400600800

100012001400160018002000

310×

|<2.5>500 MeV, |T

p=7 TeVsData,

PreliminaryATLAS

uncorrected

0 0.5 1 1.5 2 2.5 3/5

0)) /

( m

in T

- N

T(N

)/ m

in T

-N T

(N0

0.01

0.02

0.03

0.04

0.05|<2.5>500 MeV, |

Tp

=7 TeVsData,

PreliminaryATLAS

uncorrected


Analysis Overview

23

ΔΦ “same minus opposite” observable

- Divide each event into two η regions according to the η of the leading particle.- Plot Δϕ for particles with η of the same/opposite sign as the leading particle.- Subtract the “opposite region” from the “same region” and normalise.

Event-by-event, assign tracks to one of

location of leading track

0 0.5 1 1.5 2 2.5 3

/50)

/ (

T N

0

200

400

600

800

1000

1200

310×

Same region

Opposite region

|<2.5>500 MeV, |T

p=7 TeVsData,

PreliminaryATLAS

uncorrected

0 0.5 1 1.5 2 2.5 3

/50)

) / (

opp

T -

N s

ame

T(N

)/ o

pp T

- N

sam

e T

(N

0.01

0.02

0.03

0.04

0.05

0.06

0.07

0.08

|<2.5>500 MeV, |T

p=7 TeVsData, preliminaryATLAS

uncorrected



24

Tracking Efficiency in pT and η

• Loss of non-leading tracks: corrected for using a weight per entry.

• Loss of leading tracks: bin-by-bin shape correction based on how the shape of the distribution changes when more leading tracks are removed.

0 0.5 1 1.5 2 2.5 3

/50)

/ (

T N

0

500

1000

1500

2000

2500

310×

Corrected others, corrected leadCorrected othersCorrected others, -20% lead tracksCorrected others, -40% lead tracksCorrected others, -60% lead tracksCorrected others, -80% lead tracksCorrected others, -100% lead tracks

PreliminaryATLAS

Percentage of true leading tracks0 20 40 60 80 100

Num

ber o

f tra

cks

in 7

th b

in

2000

2050

2100

2150

2200

2250

310×

PreliminaryATLAS


Statistical and Systematic Uncertainties

25

Summary of systematic uncertainties. The total uncertainty is obtained by adding in quadrature all of the contributions and the statistical uncertainty.



Results

26

0 0.5 1 1.5 2 2.5 3

/50)

) / (

min

T -

N T

(N)/

min

T -

N T

(N0

0.01

0.02

0.03

0.04

0.05

0.06|<2.5>500 MeV, |

Tp

=900 GeVsData,

Pythia Tune A

Pythia Tune P0

Pythia GAL

PreliminaryATLAS

0 0.5 1 1.5 2 2.5 3

/50)

) / (

opp

T -

N s

ame

T(N

)/ o

pp T

- N

sam

e T

(N -0.02

0

0.02

0.04

0.06

0.08

0.1

0.12

0.14

|<2.5>500 MeV, |T

p=900 GeVsData,

Pythia Tune A

Pythia Tune P0

Pythia GAL

PreliminaryATLAS

0 0.5 1 1.5 2 2.5 3

/50)

) / (

min

T -

N T

(N)/

min

T -

N T

(N

0

0.01

0.02

0.03

0.04

0.05

0.06|<2.5>500 MeV, |

Tp

=7 TeVsData,

Pythia Tune A

Pythia Tune P0

Pythia GAL

PreliminaryATLAS

0 0.5 1 1.5 2 2.5 3/5

0)) /

( o

pp T

- N

sam

e T

(N)/

opp

T -

N s

ame

T(N -0.02

0

0.02

0.04

0.06

0.08

0.1

0.12

0.14

|<2.5>500 MeV, |T

p=7 TeVsData,

Pythia Tune A

Pythia Tune P0

Pythia GAL

PreliminaryATLAS

pT > 500 MeV |η| < 2.5

Even though the MC models used in this comparison cover a large range of possible behaviours, none of them match the data well.

Results

27

pT > 500 MeV|η| < 1.0

Better agreement in the region |η| < 1.0. Not surprising as most of the tunes use CDF data as input.

0 0.5 1 1.5 2 2.5 3

/50)

) / (

min

T -

N T

(N)/

min

T -

N T

(N0

0.01

0.02

0.03

0.04

0.05

0.06

0.07|<1>500 MeV, |

Tp

=900 GeVsData,

Pythia Tune A

Pythia Tune P0

Pythia GAL PreliminaryATLAS

0 0.5 1 1.5 2 2.5 3

/50)

) / (

opp

T -

N s

ame

T(N

)/ o

pp T

- N

sam

e T

(N -0.02

0

0.02

0.04

0.06

0.08

0.1

0.12

0.14

|<1>500 MeV, |T

p=900 GeVsData,

Pythia Tune A

Pythia Tune P0

Pythia GAL

PreliminaryATLAS

0 0.5 1 1.5 2 2.5 3

/50)

) / (

min

T -

N T

(N)/

min

T -

N T

(N

0

0.02

0.04

0.06

0.08

0.1

|<1>500 MeV, |T

p=7 TeVsData,

Pythia Tune A

Pythia Tune P0

Pythia GAL

PreliminaryATLAS

0 0.5 1 1.5 2 2.5 3/5

0)) /

( o

pp T

- N

sam

e T

(N)/

opp

T -

N s

ame

T(N

-0.02

0

0.02

0.04

0.06

0.08

0.1

0.12

0.14

|<1>500 MeV, |T

p=7 TeVsData,

Pythia Tune A

Pythia Tune P0

Pythia GAL

PreliminaryATLAS


Particle Distributions and Correlations in ATLAS · On behalf of the ATLAS Collaboration Particle...

Documents

Transcript of Particle Distributions and Correlations in ATLAS · On behalf of the ATLAS Collaboration Particle...