AtlasPublicTopicHeader.png

Old approved Plots of the Inner Tracking Combined Performance Group

Introduction

The inner tracking combined performance plots below are approved to be shown by ATLAS speakers at conferences and similar events. However, they are superseded by newer results and should therefore only be used in exceptional cases, e.g. to show the evolution of the performance.

Please do not add figures on your own. Contact the the inner tracking combined performance coordinators (atlas-perf-idtracking-conveners@cernNOSPAMPLEASE.ch) in case of questions and/or suggestions.

Figures

Collision data (2011)

Alignment

The following plot is commented in ATLAS-COM-PHYS-2011-643, other plots from that document have been removed because they are superseded by the May 2012 alignment systematics page.

Subsystem level, “Level One”, alignment corrections performed on a run by run basis starting from a common set alignment constants. The corrections shown are for translations in the global x direction. Large movements of the detector are measured after hardware incidents. In between these periods little (<1μm) movement is observed indicating that the detector is generally very stable. At this stage run by run corrections will only be applied during data reprocessing but in the future this type of correction will be applied during the ATLAS reconstruction calibration loop. In between runs:
  • 179939 & 180149 there was a cooling system failure
  • 180481 & 180614 there was a power cut
  • 180710 & 182284 there was a LHC technical stop where the cooling & power was turned off
  • 182519 & 182726 a fire alarm went off and as a result the cooling system was required to be shut down
  • 182787 & 183003 the toroidal magnetic field was dumped

eps version of the figure

Z mass reconstruction

The following plots are commented in ATLAS-COM-PHYS-2011-865 with updates from ATL-COM-PHYS-2011-1247
They show the Z mass reconstructed with ID only tracks using the alignment prepared for the summer 2011 reprocessing, which correct for a curl weak mode observed in data.

Invariant mass distribution of Z → μμ decays, where the mass is reconstructed using track parameters from the Inner Detector track of the combined muons only, using about 702 pb-1 of data collected during spring 2011. Ideal alignment performance based on Monte Carlo is compared to observed performance of data processed with spring 2011 alignment and data processed with updated alignment constants.
eps version of the figure
Invariant mass distribution of Z → μμ decays, where the mass is reconstructed using track parameters from the Inner Detector track of the combined muons only, using about 702 pb-1 of data collected during spring 2011.
Both tracks in the barrel region |η|<1.05
Ideal alignment performance based on Monte Carlo is compared to observed performance of data processed with spring 2011 alignment and data processed with updated alignment constants.

eps version of the figure
Invariant mass distribution of Z → μμ decays, where the mass is reconstructed using track parameters from the Inner Detector track of the combined muons only, using about 702 pb-1 of data collected during spring 2011.
At least one track in the barrel region |η|<1.05 outdated plot, please use the one above with both tracks in the barrel
Ideal alignment performance based on Monte Carlo is compared to observed performance of data processed with spring 2011 alignment and data processed with updated alignment constants.

eps version of the figure
Invariant mass distribution of Z → μμ decays, where the mass is reconstructed using track parameters from the Inner Detector track of the combined muons only, using about 702 pb-1 of data collected during spring 2011.
Both tracks in the end-cap A region 1.05 < η < 2.5
Ideal alignment performance based on Monte Carlo is compared to observed performance of data processed with spring 2011 alignment and data processed with updated alignment constants.

eps version of the figure
Invariant mass distribution of Z → μμ decays, where the mass is reconstructed using track parameters from the Inner Detector track of the combined muons only, using about 702 pb-1 of data collected during spring 2011.
At least one track in the end-cap A region 1.05 < η < 2.5 outdated plot, please use the one above with both tracks in end-cap A
Ideal alignment performance based on Monte Carlo is compared to observed performance of data processed with spring 2011 alignment and data processed with updated alignment constants.

eps version of the figure
Invariant mass distribution of Z → μμ decays, where the mass is reconstructed using track parameters from the Inner Detector track of the combined muons only, using about 702 pb-1 of data collected during spring 2011.
Both tracks in the end-cap C region -2.5 < η < -1.05
Ideal alignment performance based on Monte Carlo is compared to observed performance of data processed with spring 2011 alignment and data processed with updated alignment constants.

eps version of the figure
Invariant mass distribution of Z → μμ decays, where the mass is reconstructed using track parameters from the Inner Detector track of the combined muons only, using about 702 pb-1 of data collected during spring 2011.
At least one track in the end-cap C region -2.5 < η < -1.05 outdated plot, please use the one above with both tracks in end-cap C
Ideal alignment performance based on Monte Carlo is compared to observed performance of data processed with spring 2011 alignment and data processed with updated alignment constants.

eps version of the figure
Mean Z invariant mass versus φ for positive and negative muons, respectively. The mass distributions are fitted in RooFit using an unbinned maximum likelihood fit of a Breit-Wigner distribution, describing the intrinsic Z width, convolved with a Crystal Ball function as resolution function. Only Z candidates with a mass [71, 111] GeV are used.
Both tracks in the barrel region |η| < 1.05
Ideal alignment performance based on Monte Carlo is compared to observed performance of data processed with spring 2011 alignment and data processed with updated alignment constants.

eps version of the figure

eps version of the figure
Mean Z invariant mass versus φ for positive and negative muons, respectively. The mass distributions are fitted in RooFit using an unbinned maximum likelihood fit of a Breit-Wigner distribution, describing the intrinsic Z width, convolved with a Crystal Ball function as resolution function. Only Z candidates with a mass [71, 111] GeV are used.
Both tracks in the end-cap A region 1.05 < η < 2.5
Ideal alignment performance based on Monte Carlo is compared to observed performance of data processed with spring 2011 alignment and data processed with updated alignment constants.

eps version of the figure

eps version of the figure
Mean Z invariant mass versus φ for positive and negative muons, respectively. The mass distributions are fitted in RooFit using an unbinned maximum likelihood fit of a Breit-Wigner distribution, describing the intrinsic Z width, convolved with a Crystal Ball function as resolution function. Only Z candidates with a mass [71, 111] GeV are used.
Both tracks in the end-cap C region -2.5 < η < -1.05
Ideal alignment performance based on Monte Carlo is compared to observed performance of data processed with spring 2011 alignment and data processed with updated alignment constants.

eps version of the figure

eps version of the figure

Collision data (2010)

D* distributions

The following plots show distributions of D* and D0 candidates that are selected as follows:
  • common track requirements
    • at least 1 Pixel hit
    • at least 5 silicon hits
    • transverse impact parameter (d0) with respect to the primary vertex within 2mm
    • longitudinal impact parameter (z0*sin(theta)) with respect to the primary vertex within 2mm

  • D0 decaying into K and pi
    • pT of K and pi > 1 GeV
    • |m(Kpi)-1.865 GeV| < 20 MeV

  • D* candidates
    • pT of D* > 4.5 GeV

This plot shows the mass difference between the D* and the D0 candidates in the signal region, i.e. the region where |m(Kpi)-1.865 GeV| < 20 MeV. A fit using a data+background hypothesis yields 257 +- 36 D* events. The reconstructed mass difference of 145.56+-0.12 MeV agrees with the PDG value of 145.4 MeV. The width is 0.78+-0.15 MeV. png
eps version of the figure
This plot shows the mass difference between the D* and the D0 candidates in the sideband region, i.e. the region where 100 MeV < |m(Kpi)-1.865 GeV| < 200 MeV. png
eps version of the figure
This plot shows the invariant mass of the K and the pi for D* candidates selected via the mass difference (delta_m) between the D* and the D0: 143.9 MeV < delta_m < 146.9 MeV. A fit with a sum of a constant and a Gaussian yields 262+-40 events inside the peak. The mean of the Gaussian is 1869.2+-2.4 MeV which is close to the PDG value of 1864.8 MeV. The width is 13.9+-2.3 MeV. png
eps version of the figure
This plot shows the invariant mass of the K and the pi for D* candidates in the D* sideband region (150 MeV < delta_m < 170 MeV). png
eps version of the figure

Collision data (2009)

Invariant mass distributions

This plot shows the invariant mass distribution of two track vertices found with the ATLAS standard V0 vertex finding code in the range 400 to 800 MeV. No mass constraint is applied during the vertex fit. The two tracks are required to have opposite charge, more than six silicon hits each and a transverse momentum greater than 100 MeV. In addition, the distance of the reconstructed secondary vertex to the primary vertex has to be larger than 0.2 mm in the transverse x-y (=R phi) plane. Furthermore, a cut is applied to require that the flight path of the V0 candidate points back to the interaction point. The selection is applied on the angle in the transverse plane between the flight direction, defined from the line from the primary vertex to the secondary vertex, and the momentum sum of the two tracks. The invariant mass is calculated assuming that both tracks are pions.
kShort.png
eps version of the figure
This plot is similar to the previous plot (one above). In addition, however, it shows the invariant mass measured in data for all two track vertex combinations (i.e. no cut on the flight distance or the angle to the flight direction of the reconstructed vertex).
kShort_overlay.png
eps version of the figure
This plot shows the invariant mass distribution of two track vertices found with the ATLAS standard V0 vertex finding code. No mass constraint is applied during the vertex fit. The two tracks are required to have opposite charge, more than six silicon hits each and a transverse momentum greater than 100 MeV. In addition, the distance of the reconstructed secondary vertex to the primary vertex has to be larger than 0.2 mm in the transverse x-y (=R phi) plane. Furthermore, a cut is applied to require that the flight path of the V0 candidate points back to the interaction point. The selection is applied on the angle in the transverse plane between the flight direction, defined from the line from the primary vertex to the secondary vertex, and the momentum sum of the two tracks. Note, this vertex selection is identical to the one used for the K shorts.

The invariant mass is calculated assuming that the positively charged track is a proton and the negatively charged track is a pion.

Lambda.png
eps version of the figure
This plot is similar to the previous plot (one above). In addition, however, it shows the invariant mass measured in data for all two track vertex combinations (i.e. no cut on the flight distance or the angle to the flight direction of the reconstructed vertex).
Lambda_overlay.png
eps version of the figure
This plot shows the invariant mass distribution of two track vertices found with the ATLAS standard V0 vertex finding code. No mass constraint is applied during the vertex fit. The two tracks are required to have opposite charge, more than six silicon hits each and a transverse momentum greater than 100 MeV. In addition, the distance of the reconstructed secondary vertex to the primary vertex has to be larger than 0.2 mm in the transverse x-y (=R phi) plane. Furthermore, a cut is applied to require that the flight path of the V0 candidate points back to the interaction point. The selection is applied on the angle in the transverse plane between the flight direction, defined from the line from the primary vertex to the secondary vertex, and the momentum sum of the two tracks. Note, this vertex selection is identical to the one used for the K shorts.

The invariant mass is calculated assuming that the positively charged track is a pion and the negatively charged track is an anti-proton.

LambdaBar.png
eps version of the figure
This plot is similar to the previous plot (one above). In addition, however, it shows the invariant mass measured in data for all two track vertex combinations (i.e. no cut on the flight distance or the angle to the flight direction of the reconstructed vertex).
LambdaBar_overlay.png
eps version of the figure

Inner Detector Material Studies

Several different geometries are used to compare simulated samples with data. The nominal sample is the default comparison, as the material model used in simulation is also used in reconstruction. The additional samples scale structures in the Inner Detector to produce roughly 10% and 20% more material in the simulation in terms of radiation length.
Fitted Kshort mean as a function of the decay radius for data and MC simulation. KshortMean_r_bar_dataMC.png
eps version of the figure
Fitted Kshort mean as a function of the decay radius for various MC simulated material descriptions. KshortMean_r_bar_MC.png
eps version of the figure

Average Number of Hits on Track in 900 GeV Data

Plots comparing the average number of silicon hits on track between 900 GeV data and Non-Diffractive Minimum Bias Monte Carlo Simulation
  • Event Selection:
    • LB with all Inner Detector DQ Flags green; BCID consistent with collisions; MBTS _1 trigger; a reconstructed primary vertex containing at least 3 tracks
  • Track Selection:
    • Combined Inner Detector tracks reconstructed using the inside-out tracking algorithm. Tracks are required to have pT>0.5 GeV; >=1 pixel hit; >=6 SCT hits and impact parameters pointing to the primary vertex: |d0| < 1.5 mm; |z0 sinθ| < 1.5 mm
  • The simulation has been reweighted to correct the beam spot z position used in simulation to that observed in data.
  • The simulation was corrected to account for the number of disabled modules in the data: as this number varied from run to run the worst case (~4% disabled Pixel modules) was applied for all data runs and the simulation
Comparison between number of pixel hits on reconstructed tracks in 900 GeV data and Non-Diffractive Minimum Bias Monte Carlo Simulatio. The plot shows the eta distribution in which the increase in the number of hits in the end cap region is clearly visible. The asymmetry in eta is due to the location of disabled modules

There was a problem in the y-axis labeling (shift of 0.5 in the number of hits) which has been corrected on 08/02/10. Please use only the new version!

npixvseta_pfx.png
eps version of the figure
Comparison between number of pixel hits on reconstructed tracks in 900 GeV data and Non-Diffractive Minimum Bias Monte Carlo Simulation. The plot shows the phi distribution. Again the location of the disabled modules is visible. The slight disagreement at the percent level is expected to be caused by displacement in x-y of the beam spot in simulation.

There was a problem in the y-axis labeling (shift of 0.5 in the number of hits) which has been corrected on 08/02/10. Please use only the new version!

npixvsphi_pfx.png
eps version of the figure
Comparison between number of SCT hits on reconstructed tracks in 900 GeV data and Non-Diffractive Minimum Bias Monte Carlo Simulation. The plot shows the eta distribution in which the increase in the number of hits in the end cap region is clearly visible. The structure of the SCT disks is reproduced by the simulation.

There was a problem in the y-axis labeling (shift of 0.5 in the number of hits) which has been corrected on 08/02/10. Please use only the new version!

nsctvseta_pfx.png
eps version of the figure
Comparison between number of SCT hits on reconstructed tracks in 900 GeV data and Non-Diffractive Minimum Bias Monte Carlo Simulation. The plot shows the phi distribution.

There was a problem in the y-axis labeling (shift of 0.5 in the number of hits) which has been corrected on 08/02/10. Please use only the new version!

nsctvsphi_pfx.png
eps version of the figure
Figure description ... figure

Particle Identification

Pixel dE/dx plot calculated using the Pixel cluster charge, and the track’s angle to determine the dx. The maximum charge contribution is skipped when the mean is obtained. This is to limit the Landau tail. The dE/dx is divided by the silicon density.
Pixel dE/dx plots with Lambda(bar) selection
  • Using stable-­beam sqrt(s)=900 GeV Data from 2009 (same dataset that is used for the K0s, Lambda(bar), Armenteros plots), including good Run and LB selection.
  • Apply Lambda track selection + invariant mass cut (1115.7 ± 3) MeV
  • The width cut (± 3 MeV) corresponds to about 1σ of the Lambda(bar) peaks.
  • Only the positive tracks for the Lambda decays (proton candidates) and the negative tracks for the LambdaBar decays (anti‐proton candidates) are displayed.
  • The (anti-­) proton bands are dominant; also visible background from pion tracks.
dEdx_LambdaSelection.png
eps version of the figure
Pixel dE/dx plots with Lambda(bar) sideband selection
  • Similar to previous Lambda(bar) selection plot, but the invariant mass cut selects sidebands which are ± 5σ off from the Lambda mass peak.
dEdx_LambdaSidebands.png
eps version of the figure

Cosmic ray data

The cosmic ray data plots below are a subset of the ID approved cosmic ray data plots

ID track parameter resolutions

Cosmics muons traverse the whole Inner Detector and thus leave hits in the upper and lower parts of the detector. By dividing the track to its upper and lower half according to the value of the y coordinate of hits on track and refitting both hit collections, two collision-like tracks originating from the same cosmic muon are obtained. After alignment, track parameter resolutions are studied by comparing the difference (residual) of the track parameters at the perigee point. Since both tracks have an associated error, the quoted resolution is the RMS of the residual distribution of the particular track parameter divided by square root 2. The track parameter resolutions are studied dependant on variables like the pT or the d0 of the tracks. Shifted mean values of the residual distributions can be a sign of systematic detector deformations. The following distributions show data from runs 91885,91888,91890,91891 and 91900 taken in 2008. The tracks have been refitted using Athena release 15.0.0.7 (Tier0).

The following cuts have ben applied per track (if not stated otherwise):

  • nPixelHits in the barrel >=2
  • nSCTHits in the barrel >=6
  • nTRTHits in the barrel >=25
  • |d0| < 40mm
  • pT >= 1 GeV

Additionally only events with an event phase between 5 and 30 ns are accepted to ensure proper timing of the sub-detectors. The plots show comparisons of tracks using the full Inner Detector (silicon and TRT detectors, closed triangles), only the silicon sub-detecors (open triangles) together with tracks from cosmic simulation using the full Inner Detector (stars). The requirement of at least 25 TRT hits is dropped for silicon only tracks. However, the cut on the event phase is retained to ensure proper timing and the comparability of the analyzed sets of tracks. There is no cut on the event phase for simulation events since the jitter in cosmic trigger timing is not simulated.

Transverse impact parameter resolution as a function of pT. In the low pT region, the resolution is dominated by multiple scattering effects. At higher values, the resolution is flat. Taking into account the TRT information improves the resolution. The difference to the MC curve indicates the remaining mislaignment.
eps figure
Transverse impact parameter resolution as a function of d0 itself. For this plot the d0 cut is released to 120 mm and the minum number of Pixel hits is set to one. In general the resolution for full ID tracks is better. The resolution is better in the central d0 region due to more Pixel layers crossed and less spread clusters in the Pixel detector. Dips are seen if the d0 of the tracks equal the radii of the pixel layers (indicated by dashed lines). Since the d0 is in these cases very close to a hit on a Pixel layer, the extrapolation to the perigee point is very small and the resolution improves. The MC distributions confirms the observed behaviour.
eps figure
Mean of the transverse impact parameter distribution as a function of pT. The expected value of the mean is 0 as confirmed by the MC distribution. In data a shift is seen for full ID and silicon only tracks. The shift increases with higher pT.
eps figure
Mean of the transverse impact parameter distribution as a function of d0 itself. For this plot the d0 cut is released to 120 mm and the minum number of Pixel hits is set to one. The expected value of the mean is 0 as confirmed by the MC distribution. In data a shift is seen for full ID and silicon only tracks. The shift is biggest in the central d0 region.
eps figure
Relative momentum resolution as a function of pT. The relative momentum resolution increases with higher pT due to stiffer tracks and a more difficult measurement of the sagitta. Including information from the TRT extends the lever arm and helps improving the resolution especially at high pT values. The difference to the MC curve indicates the remaining misalignment.
eps figure
Mean of the relative momentum distribution as a function of pT. The expected value of the mean is 0 as confirmed by the MC distribution. In data a shift is seen for full ID and silicon only tracks. The shift increases with higher pT.
eps figure
Resolution of the azimuthal angle as a function of pT. In the low pT region, the resolution is dominated by multiple scattering effects. At higher values, the resolution is flat. Taking into account the TRT information improves the resolution. The difference to the MC curve indicates the remaining mislaignment.
eps figure
Resolution of the polar angle as a function of eta. The resolution of the polar angle theta improves at larger eta due to broader pixel clusters that allow a more precise position measurement. Since the TRT effectively does not measure the z coordinate in the barrel region, the resolutions are equal for silicon only and full ID tracks. The difference to the MC curve indicates the remaining misalignment.
eps figure
Figure description ... figure


Major updates:
-- ChristianSchmitt - 19-Nov-2009

Responsible: ChristianSchmitt
Last reviewed by: Never reviewed

Topic attachments
I Attachment HistorySorted descending Action Size Date Who Comment
Unknown file formateps dEdx_LambdaSelection.eps r1 manage 155.5 K 2011-06-15 - 18:37 AttilioAndreazza  
PNGpng dEdx_LambdaSelection.png r1 manage 30.7 K 2011-06-15 - 18:37 AttilioAndreazza  
Unknown file formateps dEdx_LambdaSidebands.eps r1 manage 125.4 K 2011-06-15 - 18:38 AttilioAndreazza  
PNGpng dEdx_LambdaSidebands.png r1 manage 28.9 K 2011-06-15 - 18:38 AttilioAndreazza  
Edit | Attach | Watch | Print version | History: r4 < r3 < r2 < r1 | Backlinks | Raw View | WYSIWYG | More topic actions
Topic revision: r4 - 2014-09-10 - AnthonyMorley
 
    • Cern Search Icon Cern Search
    • TWiki Search Icon TWiki Search
    • Google Search Icon Google Search

    Atlas All webs login

This site is powered by the TWiki collaboration platform Powered by PerlCopyright &© 2008-2019 by the contributing authors. All material on this collaboration platform is the property of the contributing authors.
Ideas, requests, problems regarding TWiki? Send feedback