View
226
Download
2
Embed Size (px)
Citation preview
Standalone VeloPix Simulation
Jianchun Wang
4/30/10
Introduction
VeloPix performance after irradiation affects our current design. We want to create tools to study these effects.
As a first step I modify the standalone simulation package and look at performance of pixel detector before irradiation. More features are to be added. More sophisticated electric field based ISE-TCAD simulation.
Charge trapping.
Finite integration time.
Disclaim: Some of the electronics properties may be too good to be true. If it is need, more realistic parameters can be added to provide input to other studies.
04/30/10 Jianchun Wang 2
Besides of many interesting features, I am particular interested in: The charge collection efficiency will be reduced due to insufficient bias
and/or charge trapping, etc.
Reconstruction of angled tracks are biased due charge collection inefficiency.
Non-uniform irradiation dose on a single chip/sensor makes it difficult bias.
…
Sensor and Electronics Properties
Silicon sensor Thickness = 150 mm. Charge carrier = electron Pixel size = 55 mm x 55mm. Full depletion voltage = 30 V Bias HV = 50 V
Electronics Charge collection efficiency = 100% Noise = 100 e Gain uncertainty = 10 % Crosstalk between adjacent pixels = 0 Threshold = 1000 e Non-uniformity of thresholds = 0 % ADC (TDC) bits = 8 ADC range = 1000 – 24000 e Non-linearity = 0
04/30/10 Jianchun Wang 3M
ore
real
istic
para
met
ers
will be
add
ed w
hen
they
are
avai
labl
e
Normal Incident Tracks
04/30/10 Jianchun Wang 4
Track: 20 GeV .p Row X, Column Y With VeloPix detector row and column have same pitches.
Angle X = 0Angle Y = 0
<N> =1.55 <N> =1.26
MPV ~ 11 Ke
Above ADC range
Eta Correction
04/30/10 Jianchun Wang 5
Pixel border
Linear charge weighting
~75%, no chargesharing info available
~25%, very narrow after eta correction partially due to small portion
eta correction
Tracks At Different Angles
04/30/10 Jianchun Wang 6
Tracks at 0 degree in Y/column direction
For threshold = 1000 e, the best resolution is at 18 – 20 degree.
With Plane Tilted
04/30/10 Jianchun Wang 7
Tracks at 0 or 20 degree in Y/column direction
There are more charge sharing in column direction, thus slightly less charge sharing for normal incident track in X direction.
Different Thresholds and Noise
04/30/10 Jianchun Wang 8
Threshold = 1000 eNoise = 100 e
Threshold = 2000 eNoise = 100 e
Threshold = 1000 eNoise = 300 e
Just to show how thresholds and noise affect the resolution.
With increasing of noise, the resolution is affected the most at small angles where the shared charges are less.
Threshold affects the resolution, especially for thin detector. Trim DAC in each cell may be necessary to reduce the non-uniformity of threshold, and thus reduce the overall threshold level.
Plan
Use more realistic electronics parameters from TimePix studies, and generate inputs for other studies.
Add irradiation dose dependent effects
More sophisticated electric field based ISE-TCAD simulation.
Charge trapping.
Finite integration time.
…
May integrate it to more general simulation, depending CPU consumption ( ~10ms /hit ).
More interesting studies.
04/30/10 Jianchun Wang 9
Testbeam of Radiation
Hard Sensor
Telescope Configuration
11
TypeThickness(mm)
Size(mm2)
HV Comment
Telescope N-type Si 300 ?16 x 2432 x 16
220
DUT
sCVD 500 4x 4 250 At 0, 10, & 20, HV scan at 20
MCZ Si 300 ? 16 x 24 500 0, threshold scan
Purdue 3D 200 8x8 40 3d_4e_wb5_8, failed
200 8x8 40 3d_2e_wb216_6, HV & threshold scan
285 8x8 40 3d_2e_wb5_2, HV & threshold scan
DUTYY
120 GeV proton beam
Scint
X
Z
Y
XX XXYY
a: –22 b: +22 a: +22b: –22b: 0, –10, –20
Lab frame
04/30/10 Jianchun Wang
Diamond HV Scan at Angle What we want to extract from the testbeam for different bias HVs:
Total charge collected per particle hit in terms of MPV of the Landau distributions.
For a fixed threshold how the charge sharing information the detector can deliver, in terms of number of rows, or columns per particle hit cluster.
Spatial resolution.
Shift of spatial position measurement due to partial charge collection and tracks at angle. This can give us some ideas on effective depth, and charge trapping.
Status of each task:
Need more work on readout electronics gain and pedestal calibrations. It is difficult to compare the absolute charge before that.
Numbers of pixels per hit vs bias HV qualitatively agree with expectation. We need to obtain precise thresholds from bench test for MC simulation. Then we can have quantitative comparison to test our understanding.
Current resolution is not as good as expectation. Need more work on gain curve and telescope alignment.
Shift of center residual shows correct trend. It will be revised after the spatial measurement optimization.
04/30/10 12Jianchun Wang
Diamond Sensor Charge Sharing vs HV
Number of Columns per Cluster
Number of Rows per Cluster
HV = -250 V
Sensor rotated to ~ 20 in row direction.
More charge collected with higher bias HV till saturation.
Need more work on gain calibration to extract the absolute charge (MPV of Landau distribution).
04/30/10 13
Preliminary
Preliminary
Jianchun Wang
Diamond Sensor Residual Center vs HV
Tracks are at ~ 20 with respect to normal of sensor plane in row direction.
Use the same set of telescope spatial configuration parameters.
With low bias HV, charges generated near readout electronics have more chance to be collected, equivalent to thinner effective sensor. Thus the residual center shifts.
In extreme case, the maximum possible shift ~ tan(q)*d/2 ~ 90 mm.
04/30/10 14
Preliminary
Jianchun Wang
Diamond Sensor Charge Sharing vs Angle
Diamond sensor is biased at -250 V.
Sensor was perpendicular to beam, or rotated by ~10 & ~20 in row direction.
Gain and threshold of the electronics are different from that of HV scan.
04/30/10 15
Preliminary
Preliminary
Jianchun Wang
Charge Distributions
04/30/10 Jianchun Wang 16
Charge (Ke)
Silicon Telescope
sCVD DUT
Plane 0
Plane 8
Plane 4
Charge (Ke)
Good
Bad
Weird
Problem with Diamond Gain Calibration
04/30/10 Jianchun Wang 17
sCVD DUT
Charge (Ke)
Double peaks belong to different cluster sizes, suggesting there is offset issue.
The difference between two peaks is too big.
The offset would have to be ~ –13 in order to have the same MP. And the MP would be ~ –3. So this is not a correct hypothesis.
MP=9.8
MP=22.7
Charge (Ke)
Charge (Ke)
Diamond Residual Distribution
04/30/10 Jianchun Wang 18
Xrecon – Xtrack (mm) Yrecon – Ytrack (mm)
s = 30.5 mm s = 46.7 mm
Tracks are at ~20 in X direction wrt the diamond.
Charging sharing information is not fully used yet due to calibration issues.
In comparison 100/12 = 28.9, 150/12 =43.3.
Summary
We had tested radiation hard sensors: diamond, MCZ silicon & 3D.
Some interesting results are produced from diamond test.
Gain calibration somehow becomes bottle neck.
We provide offline analysis and alignment program for this testbeam system as our promised contribution.
We may use the telescope to test our own sensors.
04/30/10 Jianchun Wang 19