Study of the radiation damage effect on the Inner Detector UC Berkeley, August 3 rd 2018 Supervisor: Frédéric Derue Precise measurement of the top quark mass using charmed mesons in the final state Technical supervisor: Ben Nachman Jad Zahreddine (LPNHE - Paris)
19
Embed
Study of the radiation damage effect on the Inner Detector · • Migration of the radiation damage code from release 22 to release 21 ‣ Tested different releases within the release
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
Study of the radiation damage effect on the Inner Detector
UC Berkeley, August 3rd 2018
Supervisor: Frédéric Derue
Precise measurement of the top quark mass using charmed mesons in the final state
Technical supervisor: Ben Nachman
Jad Zahreddine (LPNHE - Paris)
Pixel detector in ATLAS
The pixel detector is the innermost part of the ATLAS detectorlarge flux of particles
• 3 layers already existed since Run1‣ B-layer, Layer 1, Layer 2‣ 50 x 400 μm2
‣ Depth = 250 μm
• 1 layer added for Run 2: IBL‣ 50 x 250 μm2
‣ Depth = 200 μm‣ 3.3 cm away from the beam pipe
2
Radiation damage
Large flux of particles high dose of radiation on the pixels
Radiation damage introduces defects in the sensor bulk
‣ Charge collection efficiency decreases‣ Hit efficiency decreases‣ Spatial resolution of the track position worsens
Increases the probability of charge trapping in the sensor
Decreases the collected charge
3
Deforms the E-field
First steps and validation
• Migration of the radiation damage code from release 22 to release 21‣ Tested different releases within the release 21‣ Migration easier than we first expected
• Need to validate the outputs of the code: we firstly focused on the collected charge by the sensor after the passage of a charged particle
After irradiation, a phenomenon occurs: charge trapping
It is mainly caused by the creation of intermediate states in the gap where the charges get trapped
Expectation: collect less charges, hence a charge distribution shifted to the left smaller values in terms of dE/dx
4
Validation: Spatial resolution
Unbiased residual X gives an idea on the spatial resolution of the track position in the direction with the best resolution (φ-direction)
IBL and Layer 2 have the worst “resolution” since we rely on them to extrapolate the tracks
5
Useful information
φ
z
Track depth in a pixel
6
Cluster sizes tan θL = minimal cluster size
Fit Function:
Validation: Lorentz angle
For the same value of the bias voltage, the Lorentz angle increases with the fluence
mobility increases, E-field strength decreases
For the same value of the fluence, the Lorentz angle decreases with the bias voltage
Yet, we still need to validate directly an important input: E-field.
New idea that has never been done before in ATLAS !
CMS suggested that we could do so by measuring the dependence that the Lorentz angle has on depth. In other words, rely on the geometry of the clusters and its link with the mobility of the charges and the Lorentz angle
from B. Nachman’s talk
Link between the E-field and geometry/mobility:
8
This is feasible because: 1) The Lorentz angle and the charge trapping are independent (to a good approximation) 2) The depth dependent E-field affects the Lorentz angle in a known way
Problem: if we consider all the modules, we get a smeared distribution because of a very bad resolution for the central modules, effect of multiple scattering
Need to make some cuts:pT,trk > 20 GeV : get rid of the multiple scattering effectMake sure to consider only the far right/left modules* (also a nice way to consider high values of cluster sizes better probe the depth)
Due to lack of statistics when we make the cuts, we use particle gun events where we shoot muons at high angles and high pT,trk
* η-module = {-6, -5, 4, 5} 9
depthtrack = (xi - x0)tan (θ)
where: xi : position of a single pixel belonging to the cluster x0 : predicted track position in local coordinates θ : incidence angle
Take the absolute value of the depth, round it around the second decimal
For each range of depth, plot the transverse cluster size vs. the incidence angle and take the minimum value of the transverse cluster size.
The value of the resolution can be improved by making cuts on the pT of the tracks and considering only the central modules
18
Unbiased residual X gives an idea on the spatial resolution of the track position in the direction with the best resolution (φ-direction)
Depth dependencies
Depth should indeed go from -0.2 to 0.2 mm
• We are indeed calculating the depth using local coordinates evaluated at the level of the pixels
• z = 0 corresponds to the plane of the pixels rather than the center of the module
z=0
z=0
So we actually have the correct definition of the depth, some tests also showed that we could also take the 3 far right/left modules not just the last two (as long as | Localθ | > 0.5) 19