Testbeam analysis of a single chip timepix3 ingrid Kees Ligtenberg - - PowerPoint PPT Presentation

pixel x-axis (beam direction) [mm] 2 4 6 8 10 12 14 pixel y-axis [mm] 2 4 6 8 10 12 14 pixel z-axis (drift direction) [mm] 2 4 6 8 10 12 14 16 pixel z-axis (drift direction) [mm] 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 s] µ ToT [

Timepix hits Telescope track

Testbeam analysis of a single chip timepix3 ingrid

Kees Ligtenberg

Nikhef

November 20, 2017

Kees Ligtenberg (Nikhef) Beamtest reconstruction November 20, 2017 1 / 23

Detector setup

Adapted from Thesis Pascal Wolf Bonn, 2016

Triggered by a scintillating plane 6 mimosa planes and timepix3 TPC

Kees Ligtenberg (Nikhef) Beamtest reconstruction November 20, 2017 2 / 23

FEI4 between telescope and TPC

Hit residual y [mm] 2 − 1.5 − 1 − 0.5 − 0.5 1 1.5 2 Entries 10000 20000 30000 40000 50000

Telescope track fit Timepix track fit

Telescope track has additional smearing compared to timepix3 track Possibly caused by the FEI4-chip being attached to the telescope (for timing in some setups) FEI4 is used for ATLAS IBL that has a radiation length of 1.5%. (And each mimosa plane has 0.1%) Could have caused a scatter of about ∼0.7 mRad

Kees Ligtenberg (Nikhef) Beamtest reconstruction November 20, 2017 3 / 23

Time matching

Each telescope frame (115.2 µs) can have a range of triggers Try to decode trigger number in timepix3 using rising edge only Save timepix3 tracks within 400 ns of a trigger For each frame, attempt to match all events of the timepix3 with triggers in the range

Kees Ligtenberg (Nikhef) Beamtest reconstruction November 20, 2017 4 / 23

Cluster finding in telescope and timepix3

Find clusters from hits binned by position Fill 30 × 15 (12 × 12) bins for telescope (TPC) If more than 3 (6) hits in one bin, try to merge up to 8 neighbours If cluster has more than 10 (5) hits update image: mention bins and telescope

5 10 15 20 25 30 2 4 6 8 10 12 14 graphicHistogram Entries 13 Mean x 14.92 Mean y 6.615 Std Dev x 4.358 Std Dev y 2.058 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5 graphicHistogram Entries 13 Mean x 14.92 Mean y 6.615 Std Dev x 4.358 Std Dev y 2.058

Histogram of hough transform Kees Ligtenberg (Nikhef) Beamtest reconstruction November 20, 2017 5 / 23

Fit procedure

Independently align telescope Rotate and shift Timepix3 to match telescope frame Do ToT and time walk corrections Do a double simple linear regression fit in the telescope frame

◮ Errors in directions perpendicular to beam-axis ◮ In telescope frame the tracks direction is almost parallel to beam-axis,

therefore errors along beam-axis were neglected

Match Timepix3 and telescope clusters Add final plane crossing of telescope as a point with 10 µm errors to Timepix3 fit

Kees Ligtenberg (Nikhef) Beamtest reconstruction November 20, 2017 6 / 23

Hit errors

Assign errors to hit: σ2

y = 0.0552 12

+ D2

T(z − z0)

σ2

z = (1.56 ns)vdrift 12

+ D2

L(z − z0)

where D the diffusion coefficient

Kees Ligtenberg (Nikhef) Beamtest reconstruction November 20, 2017 7 / 23

Selection cuts

Telescope At least 4 planes hit Fit and reject extreme outliers (> 700 µm) Telescope fit goes through tpc Timepix3 Hit ToT > 0.15 µs At least 20 hits Exactly one cluster Cut hit outliers (> 3σdrift, > 2σplane) Fit goes through front and back (pixel row) Matching Fits closer than 1 mm in both x and y at center of tpc A unique time match Histograms At least 30 TPC hits in fit At least 75% of total number of tpc hits in fit

Kees Ligtenberg (Nikhef) Beamtest reconstruction November 20, 2017 8 / 23

Tests from run 347

Run 347 length 60m triggers 4 733 381 V grid 350 V E drift 280 V/cm rotation 17 degree 0 degree threshold 800e Used first 1 000 000 telescope frames:

◮ 461 426 triggers of which 330 925 with matching tracks

A drift speed of 75 µm/ns was assumed

Kees Ligtenberg (Nikhef) Beamtest reconstruction November 20, 2017 9 / 23

ToT-correction

Column 50 100 150 200 250 s] µ Average ToT [ 0.2 0.4 0.6 0.8 1 Before correction After correction

Differences in odd and even column numbers in ToT Corrected by introducing a factor for each column

Kees Ligtenberg (Nikhef) Beamtest reconstruction November 20, 2017 10 / 23

Time walk correction using ToT

x residual [mm] 1.5 − 1 − 0.5 − 0.5 1 1.5 Hits 100 200 300 400 500 600

3

10 × Without timewalk correction With timewalk correction

Timewalk corrected by fitting δxtimewalk =

c1 tToT+t0 + x0

Kees Ligtenberg (Nikhef) Beamtest reconstruction November 20, 2017 11 / 23

Find diffusion from residuals

Plot residuals in 2-dimensional histogram

Drift distance [mm] 4 6 8 10 12 14 16 18 20 x-residual [mm] 2 − 1.5 − 1 − 0.5 − 0.5 1 1.5 2 20 40 60 80 100 120 140 160 180 200 220

3

10 × Drift distance [mm] 4 6 8 10 12 14 16 18 20 y-residual [mm] 2 − 1.5 − 1 − 0.5 − 0.5 1 1.5 2 20 40 60 80 100 120 140 160 180 200 220 240

3

10 ×

Fit gauss to y-slices and put σ in graph

/ ndf

2

χ 5592 / 42

z0

σ 0.0008 ± 0.1258 D 0.00004 ± 0.07609 z0 0.03 ± 5.86 z position [mm] 4 6 8 10 12 14 16 18 20 from fit to track-residual [mm]

x

σ 0.05 0.1 0.15 0.2 0.25 0.3 / ndf

2

χ 5592 / 42

z0

σ 0.0008 ± 0.1258 D 0.00004 ± 0.07609 z0 0.03 ± 5.86 / ndf

2

χ 1.28e+04 / 42 D 0.00002 ± 0.07008 z0 0.0 ± 4.7 z position [mm] 6 8 10 12 14 16 18 20 from fit to track-residual [mm]

y

σ 0.05 0.1 0.15 0.2 0.25 / ndf

2

χ 1.28e+04 / 42 D 0.00002 ± 0.07008 z0 0.0 ± 4.7

Fit

• σ2

z0 + D2(z − z0) to graph, and get results →

Kees Ligtenberg (Nikhef) Beamtest reconstruction November 20, 2017 12 / 23

Diffusion in pixel plane perpendicular to track

/ ndf

2

χ 1.28e+04 / 42 D 0.00002 ± 0.07008 z0 0.0 ± 4.7

z position [mm] 6 8 10 12 14 16 18 20 from fit to track-residual [mm]

y

σ 0.05 0.1 0.15 0.2 0.25

/ ndf

2

χ 1.28e+04 / 42 D 0.00002 ± 0.07008 z0 0.0 ± 4.7

Fix σz0 = 0.055/ √ 12 = 0.0159mm DT = 221 µm/√cm From laser test DT = 309 µm/√cm at 330V and vdrift = 66.4µm/ns

Kees Ligtenberg (Nikhef) Beamtest reconstruction November 20, 2017 13 / 23

Diffusion in drift direction

/ ndf

2

χ 5592 / 42

z0

σ 0.0008 ± 0.1258 D 0.00004 ± 0.07609 z0 0.03 ± 5.86

z position [mm] 4 6 8 10 12 14 16 18 20 from fit to track-residual [mm]

x

σ 0.05 0.1 0.15 0.2 0.25 0.3

/ ndf

2

χ 5592 / 42

z0

σ 0.0008 ± 0.1258 D 0.00004 ± 0.07609 z0 0.03 ± 5.86

DL = 240 µm/√cm From laser test without time walk correction DL = 254 µm/√cm at 330V and vdrift = 66.4µm/ns

Kees Ligtenberg (Nikhef) Beamtest reconstruction November 20, 2017 14 / 23

Hitmap

50 100 150 200 250 50 100 150 200 250 200 400 600 800 1000 1200 1400

Note some defects

Kees Ligtenberg (Nikhef) Beamtest reconstruction November 20, 2017 15 / 23

Deformations in pixel plane perpendicular to track

Column 50 100 150 200 250 Row 50 100 150 200 250 0.1 − 0.08 − 0.06 − 0.04 − 0.02 − 0.02 0.04 0.06 0.08 0.1

Deformatiosn in mm Residuals are filled at expected row and column

Kees Ligtenberg (Nikhef) Beamtest reconstruction November 20, 2017 16 / 23

Deformations in drift direction

Column 50 100 150 200 250 Row 50 100 150 200 250 0.1 − 0.08 − 0.06 − 0.04 − 0.02 − 0.02 0.04 0.06 0.08 0.1

Deformations in mm Residuals are filled at expected row and column

Kees Ligtenberg (Nikhef) Beamtest reconstruction November 20, 2017 17 / 23

Conclusions

Telescope track does not seem te be usable, possible due to material between telescope and Timepix3. A single point was used ToT was corrected and used for a time walk correction The diffusion coefficient fitted from data is DT = 70 µm/√mm and DL = 76 µm/√mm Deformations in the pixel plane are found to be small (< 20 µm)

Kees Ligtenberg (Nikhef) Beamtest reconstruction November 20, 2017 18 / 23

Next: looking at cross-talk at different voltages

Number of timepix hits 50 100 150 200 250 300 350 400 normalised entries 0.01 0.02 0.03 0.04 0.05 0.06 0.07 run333 (330 V) run340 (335 V) run341 (340 V) run342 (345 V) run347 (350 V)

Number of hits at different grid-voltages

Kees Ligtenberg (Nikhef) Beamtest reconstruction November 20, 2017 19 / 23

Indication of cut effects

less than 4 planes hit in telescope Successful All telescope clusters failed fit Telescope and tpc fits do not match More than one cluster in tpc tpc fit leaves or enters on the side Tpc entry already has a matching cluster Less than 20 hits in tpc Telescope fit missed tpc all tpc clusters failed fit

50 100 150 200 250 300

3

10 ×

status of timepixTrigger

Cuts are roughly in order presented before

Kees Ligtenberg (Nikhef) Beamtest reconstruction November 20, 2017 20 / 23

Align telescope planes

z

Align with 3 degrees of freedom: x, y shifts and rotation around z Fix z position and assume all detectors perpendicular to the z-axis Find corrections from residuals Find mean of residuals using gaus fit Find rotation using histogram of ∆φ = (yrx − xry)/(x2 + y2), where x, y are the hit coordinates with respect to the average hit position and r is the residual, histogram is weighted by

• x2 + y2

Kees Ligtenberg (Nikhef) Beamtest reconstruction November 20, 2017 21 / 23

Telescope alignment procedure

1 Fit through points in plane 2 and 5

shift planes 1,3,4,5,6 in x and y

2 Fit through points in plane 2 and 5

rotate plane 5 around its average hit postition to match plane 2

3 Fit through points in plane 2 and 5

rotate all planes around their average hit position

4 Fit through points in all planes

check if converged.

Kees Ligtenberg (Nikhef) Beamtest reconstruction November 20, 2017 22 / 23

Mimosa telescope

MIMOSA26 User Manual, 2011

Detector with digital silicon pixels Rolling shutter readout with 115.2 µs per frame

Kees Ligtenberg (Nikhef) Beamtest reconstruction November 20, 2017 23 / 23