s ummer student se s s i on cern

Primary Vertex Reconstruction

for Upgrade at LHCb

Joanna Wanczyk under supervision of Agnieszka Dziurda

Joanna Wanczyk under supervision of Agnieszka Dziurda Primary Vertex Reconstructionfor Upgrade at LHCb 0 / 8

s ummer student se s s i on cern

My origins

Joanna Wanczyk under supervision of Agnieszka Dziurda Primary Vertex Reconstructionfor Upgrade at LHCb 1 / 8

s ummer student se s s i on cern

Upgrade LHCb Detector

Main changes

• Software and hardware trigger→ full software trigger

• VELO → VeloPix detector

• TT → Upstream Tracker

• T1-T3 stations → SciFidetector

nRecPV0 5 10 15 20

0

0.1

0.2

0.3

UpgradeRun II

LHCbunofficial

Running conditions

Energy [TeV ] Luminosity [cm−2s−1] Bunchspacing [ns] Data [fb−1]

Run I (2010-2012) 7/7/8 1/3.5/4 · 1032 50 3

Run II (2015-2018) 13 4 · 1032 25 5

Upgrade (>2020) 14 2 · 1033 25 50

Joanna Wanczyk under supervision of Agnieszka Dziurda Primary Vertex Reconstructionfor Upgrade at LHCb 2 / 8

s ummer student se s s i on cern

Definition

Position of the proton-proton collision inside the VELO (VErtex LOcator) detector at LHCbis obtained from the primary vertex reconstruction (PV) algorithms which use a set of inputtracks. In order to make the best performing reconstruction a wide range of parameters isbeing studied.

Joanna Wanczyk under supervision of Agnieszka Dziurda Primary Vertex Reconstructionfor Upgrade at LHCb 3 / 8

s ummer student se s s i on cern

Quality Parameters

dz 0.5− 0 0.5

0

5000

10000 LHCbunofficial

Figure 1 – The resolution for the zcoordinate in Run II.

dz 0.5− 0 0.5

0

100

200

300

400

310×

LHCbunofficial

Figure 2 – The resolution for the zcoordinate in Upgrade.

Definition

Resolution is defined as the differencebetween the fitted and generated position ofthe PV separately for each coordinate and isobtained as the averaged width of fitted tripleGaussian :

∆i = (iMC−ifit ) [mm], where i = {x, y , z}

Table 1 – The global resolution for the primaryvertex reconstruction.

Par. Run II Upgradeσx 0.013± 0.001 0.011± 0.001σz 0.075± 0.001 0.062± 0.001

Joanna Wanczyk under supervision of Agnieszka Dziurda Primary Vertex Reconstructionfor Upgrade at LHCb 4 / 8

s ummer student se s s i on cern

Quality Parameters

pseudo pull z 4− 2− 0 2 4

0

50

100

150310×

LHCbunofficial

Figure 3 – The pseudo pull for the zcoordinate in Upgrade.

pseudo pull z 4− 2− 0 2 4

0

2000

4000

LHCbunofficial

Figure 4 – The pseudo pull for the zcoordinate in Run II.

Definition

Pseudo pulls are calculated as a resolutiondivided by reconstructed position uncertainty :

pulli =∆i

σi, where i = {x, y , z}

Table 2 – The pseudo pulls for the primary vertexalgorithms.

Par. Run II Upgradeµx −0.007± 0.003 0.002± 0.002σx 1.031± 0.023 1.039± 0.005µz 0.032± 0.002 0.016± 0.001σz 1.017± 0.016 1.077± 0.004

Outcome

Good performance & stability

Joanna Wanczyk under supervision of Agnieszka Dziurda Primary Vertex Reconstructionfor Upgrade at LHCb 5 / 8

s ummer student se s s i on cern

Radial Distance Optimization

Definition

r distance is defined as :

r =√

(x1 − x2)2 + (y1 − y2)2

where (x2, y2) is PV position and (x1, y1) is the resolverposition which for MC is equal to (0, 0)

Purpose

Optimal value of rdistance must be foundin order to minimise thefalse PV efficiency andmaximise the true PVefficiency.

radial distance r [mm]0 0.05 0.1 0.15 0.2

0

0.05

0.1

Upgrade

Run II

LHCbunofficial

Figure 5 – Radial distance distributionsfor true PV.

radial distance r [mm]0 5 10 15

0

0.1

0.2

0.3

Upgrade

Run II

LHCbunofficial

Figure 6 – Radial distance distributionsfor false PV.

Joanna Wanczyk under supervision of Agnieszka Dziurda Primary Vertex Reconstructionfor Upgrade at LHCb 6 / 8

s ummer student se s s i on cern

Merits

FOM

The maximum of thepurity ∗ true PV efficiency sets the optimalvalue of r distance to r = 0.145 mm,where in Run II it was equalr = 0.18 mm for min bias sample.

radial distance r [mm]0 0.2 0.4 0.6 0.8 1

Eff

icie

ncy

(Pur

ity)

0.96

0.965

0.97

0.975

0.98

0.985

0.99

0.995

1

true PV efficiency

purity * true PV efficiency

(true + false) PVstrue PVspurity =

LHCbunofficial

Figure 7 – FOM for upgrade

Track multiplicity20 40 60 80 100 120 140

Eff

icie

ncy

0.4

0.5

0.6

0.7

0.8

0.9

1

=4nTracks

Upgrade, min

=4nTracks

Run II, minLHCbunofficial

Figure 8 – The PVs reconstruction efficiency as thefunction of track multiplicity

Efficiency

For the upgrade we have to be able to reconstructPVs consisting more tracks, although the totalefficiency of PV’s reconstruction for Upgradewhich equals 93.88% is comparable to the Run II is96.82%.

Joanna Wanczyk under supervision of Agnieszka Dziurda Primary Vertex Reconstructionfor Upgrade at LHCb 7 / 8

Thank you foryour attention !

Any Questions ?

s ummer student se s s i on cern

BACKUP

nTracks50 100 150

0

0.01

0.02

0.03

UpgradeRun II

LHCbunofficial

Figure 9 – Number of PVs per track multiplicity.

Joanna Wanczyk under supervision of Agnieszka Dziurda Primary Vertex Reconstructionfor Upgrade at LHCb 8 / 8