Track Based Alignment for the Mu3e Pixel Detector
U. Hartenstein
DPG-Fr¨uhjahrstagung 2017
For the Mu3e Collaboration
The Mu3e Experiment
Mu3e - In the Standard Model
1/8
µ
+→ e
+e
−e
+Mu3e - In the Standard Model
µ
+→ e
+e
−e
+Mu3e - In the Standard Model
1/8
µ
+→ e
+e
−e
+BR < 10
−54Beyond the Standard Model?
Motivation
• new physics?!
- predictions from SUSY, Leptoquarks, . . .
• current status (SINDRUM 1988): BR<10−12
Goal
µ+→e+e−e+
with a sensitivity ofO(10−16)
The Detector
Building the Detector
µ-beamline at PSI withO(108(9))µ/s
muon beam
target
• ”MuPix”: HighVoltage
Monolithic ActivePixelSensors
(I.Peric,Nucl.Instr.Meth.,2007, A582, 876)
• 2×2cm2 sensors
• 80×80µm2 pixels
• thinned to 50µm
impossible to have sufficient alignment after construction!
Building the Detector
3/8 muon beam
target
inner pixel layers
• ”MuPix”: HighVoltage
MonolithicActive PixelSensors
(I.Peric,Nucl.Instr.Meth.,2007, A582, 876)
• 2×2cm2 sensors
• 80×80µm2 pixels
• thinned to 50µm
impossible to have sufficient alignment after construction!
Building the Detector
outer pixel layers
muon beam
target
inner pixel layers
• ”MuPix”: HighVoltage
Monolithic ActivePixelSensors
(I.Peric,Nucl.Instr.Meth.,2007, A582, 876)
• 2×2cm2 sensors
• 80×80µm2 pixels
• thinned to 50µm
impossible to have sufficient alignment after construction!
Building the Detector
3/8
• ”MuPix”: HighVoltage
Monolithic ActivePixelSensors
(I.Peric,Nucl.Instr.Meth.,2007, A582, 876)
• 2×2cm2 sensors
• 80×80µm2 pixels
• thinned to 50µm
impossible to have sufficient alignment after construction!
Building the Detector
• ”MuPix”: HighVoltage
MonolithicActive PixelSensors
(I.Peric,Nucl.Instr.Meth.,2007, A582, 876)
• 2×2cm2 sensors
• 80×80µm2 pixels
• thinned to 50µm
impossible to have sufficient alignment after construction!
Building the Detector
3/8
• ”MuPix”: HighVoltage
MonolithicActive PixelSensors
(I.Peric,Nucl.Instr.Meth.,2007, A582, 876)
• 2×2cm2 sensors
• 80×80µm2 pixels
• thinned to 50µm alignment goal: σ ≈2µm
impossible to have sufficient alignment after construction!
Building the Detector
• ”MuPix”: HighVoltage
MonolithicActive PixelSensors
(I.Peric,Nucl.Instr.Meth.,2007, A582, 876)
• 2×2cm2 sensors
• 80×80µm2 pixels
• thinned to 50µm alignment goal: σ ≈2µm
impossible to have sufficient alignment after construction!
Building the Detector
3/8
• ”MuPix”: HighVoltage
MonolithicActive PixelSensors
(I.Peric,Nucl.Instr.Meth.,2007, A582, 876)
• 2×2cm2 sensors
• 80×80µm2 pixels
• thinned to 50µm alignment goal: σ ≈2µm
impossible to have sufficient alignment after construction!
Building the Detector
• ”MuPix”: HighVoltage
MonolithicActive PixelSensors
(I.Peric,Nucl.Instr.Meth.,2007, A582, 876)
• 2×2cm2 sensors
• 80×80µm2 pixels
• thinned to 50µm alignment goal: σ ≈2µm
impossible to have sufficient alignment after construction!
Track Based Alignment
Track Based Alignment and the Software for it
• with cosmics, lower rate, . . .
→σposition≤80µm,σorientation ≤0.3◦ (from misalignment studies)
→ track based alignment
→ σ≈2µm
• Mu3e software package
• General Broken Lines (GBL)(V. Blobel, C. Kleinwort, arXiv:1201.4320v1)
• Millepede-II (MP-II) (V. Blobel, C. Kleinwort, arXiv:1103.3909v1)
Track Based Alignment and the Software for it
4/8
• with cosmics, lower rate, . . .
→σposition≤80µm,σorientation ≤0.3◦ (from misalignment studies)
→ track based alignment
→ σ≈2µm
• Mu3e software package
• General Broken Lines (GBL)(V. Blobel, C. Kleinwort, arXiv:1201.4320v1)
• Millepede-II (MP-II) (V. Blobel, C. Kleinwort, arXiv:1103.3909v1) Mu3e Track-Reconstruction:
A. Kozlinskiy - Thu,16:45 T 116.1
General Broken Lines Fit
5/8
• multiple scattering & energy loss
→ more advanced track models: e.g. GBL
• track refit to account formultiple scattering
• complete covariance matrix of all track parameters at any point
→ track based alignment withMillepede-II
V. Blobel, C. Kleinwort,
General Broken Lines Fit
5/8
• multiple scattering & energy loss
→ more advanced track models: e.g. GBL
• track refit to account formultiple scattering
• complete covariance matrix of all track parameters at any point
→ track based alignment withMillepede-II
V. Blobel, C. Kleinwort, arXiv:1201.4320v1
Millepede-II
A least squares fit with a very large number of parameters
6/8
eachtrack j has
measurements: mij ±σij and is modelled byfij(qj,p)
V. Blobel, C. Kleinwort,
Millepede-II
A least squares fit with a very large number of parameters
6/8
eachtrack j has
measurements: mij ±σij and is modelled byfij(qj,p)
χ2 =
tracks
X
j
measurements
X
i
mij −fij(qj,p) σij
2
V. Blobel, C. Kleinwort, arXiv:1103.3909v1
Millepede-II
A least squares fit with a very large number of parameters
6/8
• minimiseχ2
- 1.5 mio track parametersqj
- 45 000 alignment parametersp
→invert a 1545000×1545000 matrix - MP-II→reduction to 45000×45000
χ2 =
tracks
X
j
measurements
X
i
mij −fij(qj,p) σij
2
V. Blobel, C. Kleinwort,
Millepede-II
A least squares fit with a very large number of parameters
6/8
• minimiseχ2
- 1.5 mio track parametersqj
- 45 000 alignment parametersp
→invert a 1545000×1545000 matrix
- MP-II→reduction to 45000×45000
χ2 =
tracks
X
j
measurements
X
i
mij −fij(qj,p) σij
2
V. Blobel, C. Kleinwort, arXiv:1103.3909v1
Millepede-II
A least squares fit with a very large number of parameters
6/8
• minimiseχ2
- 1.5 mio track parametersqj
- 45 000 alignment parametersp
→invert a 1545000×1545000 matrix - MP-II→reduction to 45000×45000
χ2 =
tracks
X
j
measurements
X
i
mij −fij(qj,p) σij
2
V. Blobel, C. Kleinwort,
Surface Deformations
Surface Deformations
• 50µm chips won’t be rigid!
• Idea: align not only for rotations and shifts but also for - surface deformations
- temperature effects (∆T ≈70K)
• 3rd order polynomials for modelling sensors
50µm silicon
Surface Deformations
7/8
• 50µm chips won’t be rigid!
• Idea: align not only for rotations and shifts but also for - surface deformations
- temperature effects (∆T ≈70K)
• 3rd order polynomials for modelling sensors
x
10 5 0 5 10
y 1050510
h(x,y)
0.200.150.100.05 0.00 0.050.100.150.20
a deformed sensor
Outlook
Status & Outlook
8/8
• misalignment StudiesX
• basic software X
• MP-II testbed for the “MuPix-Telescope” X
• missing bits and pieces
• blinded tests of alignment software
Status & Outlook
• misalignment StudiesX
• basic software X
• MP-II testbed for the “MuPix-Telescope” X
• missing bits and pieces
• blinded tests of alignment software
h
Entries 986896
Mean 3.563e−05
RMS 0.02265
/ ndf
χ2 1.075e+04 / 48
Constant 1.814e+05 ± 2.355e+02 Mean 5.187e−05 ± 2.173e−05 Sigma 0.02147 ± 0.00002
−00.3 −0.2 −0.1 0 0.1 0.2 0.3
20 40 60 80 100 120 140 160 180 103
× Entries h 986896
Mean 3.563e−05
RMS 0.02265
/ ndf
χ2 1.075e+04 / 48
Constant 1.814e+05 ± 2.355e+02 Mean 5.187e−05 ± 2.173e−05 Sigma 0.02147 ± 0.00002
Residuals
pitch≈100µm σ≈22µm
Status & Outlook
8/8
• misalignment StudiesX
• basic software X
• MP-II testbed for the “MuPix-Telescope” X
• missing bits and pieces
• blinded tests of alignment software
h
Entries 986896
Mean 3.563e−05
RMS 0.02265
/ ndf
χ2 1.075e+04 / 48
Constant 1.814e+05 ± 2.355e+02 Mean 5.187e−05 ± 2.173e−05 Sigma 0.02147 ± 0.00002
residual in mm
−00.3 −0.2 −0.1 0 0.1 0.2 0.3
20 40 60 80 100 120 140 160 180 103
× Entries h 986896
Mean 3.563e−05
RMS 0.02265
/ ndf
χ2 1.075e+04 / 48
Constant 1.814e+05 ± 2.355e+02 Mean 5.187e−05 ± 2.173e−05 Sigma 0.02147 ± 0.00002
Residuals
pitch≈100µm σ≈22µm
Backup
Parametrization
• span sensors by two orthonormal vectors u and v
• use right-handed local coordinate system u,v,w
• w =w(u,v) parametrized withLegendre-polynomials and surface coefficients
x
10 5 0 5 10
y 1050510
h(x,y)
0.200.150.100.05 0.000.050.100.150.20
Figure 1: Legendre-Plane: coefficients of 0−30µm
Parametrization
• spanned by two orthonormal vectors defining the local u- &
v-coordinates
• w-coordinate defined via u×v with a value ofh(u,v)
h(x,y) =
N
X
i=0 i
X
j=0
cijPi−j(x)Pj(y), (1) withLegendre-ploynomials
Pn(x) = 2n
n
X
k=0
n k
n+k−1 2
n
xk. (2) and surface coefficientscij
General Broken Lines Fit
V. Blobel, C. Kleinwort, arXiv:1201.4320v1
Alignment Procedure
Watson
● Track (re-)fit with GBL
● Jacobian calculation
Geometry information
● Global positions & orientations
● Surface coefficients
● Temperature scaling
Hit information
● Local hit information
● Sensor ID's
MilleBinary & Steering File(s)
● Local & global derivatives
● “instructions” for pede
Interpret & update
produce
pass (not tested)
Pede
● Least squares fit
● Alignment corrections
Misalignment
• perfect alignment
• misaligned sensors
Misalignment
• perfect alignment • misaligned sensors
Misalignment Studies
• what does that mean?
→ need foralignment algorithm
• for track based alignment tracks are needed!
• “how well(mechanically) aligned to be able to align(with software)?”
Misalignment Studies
Misalignment Studies
Misalignment Studies
Misalignment Studies
Momentum Reconstruction Efficiency
Randomly Misaligned Sensors
0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
Standard Deviation of Shifts [mm]
0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4
]°Standard Deviation of Rotations [
0 0.2 0.4 0.6 0.8 1 1.2
0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
• normalised to the efficiency of a perfectly aligned detector
• efficiency plateau
Efficiency
Momentum Reconstruction Resolution
Randomly Misaligned Sensors
1.7 1.8 1.9 2 2.1 2.2
Standard Deviation of Shifts [mm]
0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4 0.45 0.5
]°Standard Deviation of Rotations [
0 0.2 0.4 0.6 0.8 1 1.2
1.7 1.8 1.9 2 2.1 2.2
• momentum resolution from RMS ofprec−pMC
• for random sensor shifts & rotations in MeV/c
Resolution
Misalignment Studies - Momentum Resolution Sigma
1.3 1.4 1.5 1.6 1.7 1.8 1.9 2 2.1 4-hit-segments
Standard Deviation of Shifts [mm]
0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4 0.45 0.5
]° Standard Deviation of Rotations [
0 0.2 0.4 0.6 0.8 1 1.2
1.3 1.4 1.5 1.6 1.7 1.8 1.9 2 2.1 4-hit-segments
0.18 0.19 0.2 0.21 0.22 0.23 0.24 0.25 6-hit-segments
Standard Deviation of Shifts [mm]
0 0.1 0.2 0.3 0.4 0.5 0.6
]° Standard Deviation of Rotations [
0 0.2 0.4 0.6 0.8 1 1.2
0.18 0.19 0.2 0.21 0.22 0.23 0.24 0.25 6-hit-segments
0.15 0.2 0.25 0.3 0.35 8-hit-segments
Standard Deviation of Shifts [mm]
0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4
]° Standard Deviation of Rotations [
0 0.2 0.4 0.6 0.8 1 1.2
0.15 0.2 0.25 0.3 0.35 8-hit-segments
Misalignment Studies - Momentum Reconstruction Efficiency (4-hit seg- ments)
shift [mm]
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8
reconstruction efficiency
0.5 0.6 0.7 0.8 0.9 1
shift [mm]
0 0.1 0.2 0.3 0.4 0.5
reconstruction efficiency
0.5 0.6 0.7 0.8 0.9 1
°] rotation [
0 0.2 0.4 0.6 0.8 1
reconstruction efficiency
0.5 0.6 0.7 0.8 0.9 1
°] rotation [
0 0.1 0.2 0.3 0.4 0.5
reconstruction efficiency
0.5 0.6 0.7 0.8 0.9 1
shift [mm]
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8
reconstruction efficiency
0.5 0.6 0.7 0.8 0.9 1
shift [mm]
0 0.1 0.2 0.3 0.4 0.5
reconstruction efficiency
0.5 0.6 0.7 0.8 0.9 1
°] rotation [
0 0.2 0.4 0.6 0.8 1
reconstruction efficiency
0.5 0.6 0.7 0.8 0.9 1
°] rotation [
0 0.1 0.2 0.3 0.4 0.5
reconstruction efficiency
0.5 0.6 0.7 0.8 0.9 1
°] torsion angle [ 0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2 2.2
reconstruction efficiency
0.5 0.6 0.7 0.8 0.9 1
shifts of layer 0 in x-direction shifts of layer 0 in z-direction rotations of layer 0 along x-axis
rotations of layer 0 along z-axis shifts of layers 0 & 1 in x-direction shifts of layers 0 & 1 in x-direction
rotations of layers 0 & 1 along x-axis rotations of layers 0 & 1 along z-axis torsion of the whole detector
Momentum Reconstruction Resolution
°] torsion angle [ 0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2 2.2
reconstruction resolution [MeV]
0 0.5 1 1.5
2 2.5
• torsion of the whole detector
• maximum rotation angle of each detector end (total: 4◦)
• fairly insensitive to torsion
Momentum Reconstruction Efficiency
shift [mm]
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8
reconstruction efficiency
0.5 0.6 0.7 0.8 0.9 1
• shifts of the innermost layer in x-direction
Momentum Reconstruction Resolution
shift [mm]
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8
momentum resolution [MeV]
0 0.5 1 1.5
2 2.5
• shifts of the innermost layer in x-direction
Misalignment Studies - Momentum Resolution RMS (4-hit segments)
shift [mm]
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8
momentum resolution [MeV]
0 0.5 1 1.5 2 2.5
shift [mm]
0 0.1 0.2 0.3 0.4 0.5
momentum resolution [MeV]
0 0.5 1 1.5 2 2.5
°] rotation [
0 0.2 0.4 0.6 0.8 1
momentum resolution [MeV]
0 0.5 1 1.5 2 2.5
°] rotation [
0 0.1 0.2 0.3 0.4 0.5
momentum resolution [MeV]
0 0.5 1 1.5 2 2.5
shift [mm]
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8
momentum resolution [MeV]
0 0.5 1 1.5 2 2.5
shift [mm]
0 0.1 0.2 0.3 0.4 0.5
momentum resolution [MeV]
0 0.5 1 1.5 2 2.5
°] rotation [
0 0.2 0.4 0.6 0.8 1
momentum resolution [MeV]
0 0.5 1 1.5 2 2.5
°] rotation [
0 0.1 0.2 0.3 0.4 0.5
momentum resolution [MeV]
0 0.5 1 1.5 2 2.5
°] torsion angle [ 0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2 2.2
reconstruction resolution [MeV]
0 0.5 1 1.5 2 2.5
shifts of layer 0 in x-direction shifts of layer 0 in z-direction rotations of layer 0 along x-axis
rotations of layer 0 along z-axis shifts of layers 0 & 1 in x-direction shifts of layers 0 & 1 in x-direction
rotations of layers 0 & 1 along x-axis rotations of layers 0 & 1 along z-axis torsion of the whole detector
Momentum Reconstruction Efficiency
For Individual Sensors
0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
4-hit-segments
Standard Deviation of Shifts [mm]
0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4
]° Standard Deviation of Rotations [
0 0.2 0.4 0.6 0.8 1 1.2
0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
4-hit-segments
0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
6-hit-segments
Standard Deviation of Shifts [mm]
0 0.05 0.1 0.15 0.2 0.25 0.3 0.35
]° Standard Deviation of Rotations [
0 0.2 0.4 0.6 0.8 1 1.2
0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
6-hit-segments
0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
8-hit-segments
Standard Deviation of Shifts [mm]
0 0.05 0.1 0.15 0.2 0.25 0.3
]° Standard Deviation of Rotations [
0 0.2 0.4 0.6 0.8 1 1.2
0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
8-hit-segments
Momentum Resolution
For Individual Sensors
1.7 1.8 1.9 2 2.1 2.2
4-hit-segments
Standard Deviation of Shifts [mm]
0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4 0.45 0.5
]° Standard Deviation of Rotations [
0 0.2 0.4 0.6 0.8 1 1.2
1.7 1.8 1.9 2 2.1 2.2
4-hit-segments
0.5 0.52 0.54 0.56 0.58 0.6 0.62 0.64 0.66
6-hit-segments
Standard Deviation of Shifts [mm]
0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4 0.45 0.5
]° Standard Deviation of Rotations [
0 0.2 0.4 0.6 0.8 1 1.2
0.5 0.52 0.54 0.56 0.58 0.6 0.62 0.64 0.66
6-hit-segments
0.4 0.45 0.5 0.55 0.6
8-hit-segments
Standard Deviation of Shifts [mm]
0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4 0.45 0.5
]° Standard Deviation of Rotations [
0 0.2 0.4 0.6 0.8 1 1.2
0.4 0.45 0.5 0.55 0.6
8-hit-segments
The Detector
Target Inner pixel layers
Outer pixel layers Recurl pixel layers
Scintillator tiles μ Beam
110cm
18cm
• barrel detector
• two double layers of silicon sensors
• scintillating fibre tracker &
scintillating tiles (timing)
• hollow double cone target
• use re-curlers
• allow precise momentum measurements
The Phases of the Mu3e Detector
Target Inner pixel layers
Scintillating fibres Outer pixel layers μ Beam
Target Inner pixel layers
Outer pixel layers Recurl pixel layers
Scintillator tiles μ Beam
Target Inner pixel layers
Scintillating fibres Outer pixel layers Recurl pixel layers
Scintillator tiles μ Beam