[PPT] - Analysis of data recorded by a GEM LPTPC Martin Ljunggren on behalf PowerPoint Presentation

SLIDE 1

Analysis of data recorded by a GEM LPTPC

Martin Ljunggren

n behalf of the LCTPC Collaboration

June 11, 2011

1 / 18

SLIDE 2

Outline

◮ The Large Prototype TPC

with GEM readout

◮ Readout electronics ◮ Track reconstruction ◮ Distortion correction ◮ Spatial resolution ◮ Momentum resolution

2 / 18

SLIDE 3

GEMs and pad plane

◮ 5152 pads, approximately 1x5 mm, organized in 28 rows ◮ A GEM-foil consists of 5 µm Cu-layers separated by 100 µm

f insulating material

◮ Hole size: 70 µm ◮ Pitch: 140 µm ◮ 360V between Cu-layers ◮ Two GEM foils give a gain of about 104 ◮ “T2K-gas”: 95% Ar, 3% CF4, 2% isobutane

3 / 18

SLIDE 4

Instrumentation

X (mm)

200
100

100 200 Y (mm)

100
50

50 100 150 200 250

Figure: The instrumented region of the pad planes (black)

4 / 18

SLIDE 5

Readout electronics

PCA16:

◮ 16 channel preamplifier and shaper ◮ Modified version of PASA-chip from ALICE. ◮ Programmable gain, shaping and decay time.

ALTRO:

◮ Originally developed for ALICE. ◮ Sampling at 20 MHz ◮ Pedestal subtraction and zero suppression ◮ Capable of storing 1024 10 bit ADC samples.

Next step: Integration of preamplifier and ADC into one chip (S-ALTRO).

5 / 18

SLIDE 6

Due to the large number of readout channels and the small space available on the pad modules, the electronics had to be connected with 30 cm long Kapton R cables.

6 / 18

SLIDE 7

Event display

Figure: Typical event without magnetic field. Drift distance: 5 cm Figure: Typical event with magnetic field. Drift distance: 10 cm

7 / 18

SLIDE 8

Track reconstruction

t [time samples] 70 80 90 100 110 120 130 140 150 Voltage [ADC units] 100 200 300 400 500

Figure: Left: Typical pulse. Right: Typical track. ◮ Time is reconstructed as the voltage weighted average of the

five samples around the peak.

◮ Adjacent pulses are grouped into clusters where coordinates

are determined by e.g. y =

P Qiyi P Qi

where Qi is the charge of the pulse and yi is the corresponding y-coordinate of the pad.

◮ For tracking, a simple track reconstruction algorithm was

used.

8 / 18

SLIDE 9

Residuals

Residuals [mm]

1.5
1
0.5

0.5 1 1.5 Entries 500 1000 1500 2000 2500 3000 3500 4000 4500

Figure: Upper: Magnified event display. Lower: Residuals integrated over the full track length from 10000 tracks with 7 cm drift length and B=0T, σ ≈ 0.31 mm for a Gaussian core accounting for 95% of the total area. Distortion correcions have been applied.

9 / 18

SLIDE 10

Distortions

If the residuals are plotted against pad row, they should line up around zero. However:

Residuals [mm]

3
2
1

1 2 3 Entries 1000 2000 3000 4000 5000 6000

Figure: Left: Residuals for 10000 tracks vs pad row for B=1T and drift length of 10

cm. Right: Residuals integrated over the full track length using 10000 tracks from the

same run

10 / 18

SLIDE 11

After corrections

Corrected using Millipede, see “A new method for the high-precision alignment of track detectors”, Volker Blobel

Residuals [mm]

1.5
1
0.5

0.5 1 1.5 Entries 1000 2000 3000 4000 5000 Residuals [mm]

1.5
1
0.5

0.5 1 1.5 Entries 1000 2000 3000 4000 5000 6000 7000

Figure: Left:Residuals for 10000 tracks vs pad row for B=0T and drift length of 5 cm (upper) and B=1T and drift length of 10 cm (lower) Right: Residuals integrated over the full track length using 10000 tracks from the same run, σ ≈ 0.16 mm (upper) and σ ≈ 0.077 mm (lower) for a Gaussian core accounting for 95% of the total area

11 / 18

SLIDE 12

Resolution in bend plane

Drift Length [mm] 100 200 300 400 500 ]

2

[mm

2 y

σ 0.005 0.01 0.015 0.02 0.025 0.03 / ndf

2

χ 308.3 / 8 p0 4.965e-05 ± 0.00349 p1 2.634e-07 ± 1.563e-05 / ndf

2

χ 308.3 / 8 p0 4.965e-05 ± 0.00349 p1 2.634e-07 ± 1.563e-05

Figure: Measured resolution for different drift lengths. The line crosses the y-axis at 0.00349 mm2 which corresponds to an intrinsic resolution of σy(0) = 59.1 ± 0.4µm.

12 / 18

SLIDE 13

Comparison with theoretical predictions.

Figure: Predicted resolution for different magnetic field strengths and slightly different

conditions. Also shown are the points measured experimentally (shown in prev. slide)1.
1K. Ackermann et.al. Nucl.Instrum.Meth.A623:141-143,2010

13 / 18

SLIDE 14

Resolution in Z

Drift Length [mm] 50 100 150 200 250 300 ]

2

[mm

2 z

σ 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09 / ndf

2

χ 6.266 / 2 p0 0.0003385 ± 0.04144 p1 1.811e-06 ± 7.327e-05 / ndf

2

χ 6.266 / 2 p0 0.0003385 ± 0.04144 p1 1.811e-06 ± 7.327e-05

Shaping times 120 ns 90 ns 60 ns 30 ns

Figure: Measured resolution in the z-direction for different drift lengths and

shaping times. The best results are obtained with a shaping time of 60ns. Extrapolating the fitted line to half the drift length of the final TPC gives 346 ± 9µm which is well below the desired resolution of 500µm. An extrapolation to the full drift length (2.15 m) gives 446 ± 9µm, still below the goal resolution.

14 / 18

SLIDE 15

Momentum measurements

◮ p ≈ 0.3B · R ◮ σ(1/p) ≈ 9.2 · 10−3 ± 0.0002 GeV−1 ◮ The track fit includes all points along a reconstructed track. Figure: Measured track momenta (left) and 1/p-distribution (right) at a drift length

f 15 cm.

◮ The momentum resolution has been calculated from a

gaussian fit to the peak covering 42% of the total area.

◮ However, the momentum spread of the beam is ≈ 5% which

gives σ(1/p) ≈ 0.01 GeV−1 at 5 GeV, and therefore the measured width is fully consistent with the beam spread.

15 / 18

SLIDE 16

Theoretical momentum resolution

◮ Gl¨

uckstern’s formula: δ( 1

PT ) = σy 0.3L2B

720

N+4 ◮ N = 84, L ≈ 48 cm and B = 1 T. ◮ σy ≈ 76 µm (drift of 15 cm) gives σ(1/p) ≈ 3 · 10−3 GeV−1.

16 / 18

SLIDE 17

Summary

◮ Test measurements with a TPC using GEM readout have

been performed.

◮ Corrections for electric field distortions have been introduced

using the Millepede software package.

◮ Results on spatial resolution show that σy at zero drift is

59.1 ± 0.4µm and σz at zero drift is 216 ± 7µm.

◮ Result on momentum resolution is

σ(1/pt) ≈ 9.2x10−3 ± 0.0002GeV −1 at a drift length of 15

cm. The momentum spread of the beam is however non

negligible.

◮ Theoretical estimation on momentum resolution at

σy ≈ 76 µm gives σ(1/p) ≈ 3 · 10−3 GeV−1.

◮ Results on spatial resolution are consistent with the goals for

the full size ILD TPC.

17 / 18

Analysis of data recorded by a GEM LPTPC

Martin Ljunggren

June 11, 2011

Outline

◮ The Large Prototype TPC

with GEM readout

◮ Readout electronics ◮ Track reconstruction ◮ Distortion correction ◮ Spatial resolution ◮ Momentum resolution

GEMs and pad plane

◮ 5152 pads, approximately 1x5 mm, organized in 28 rows ◮ A GEM-foil consists of 5 µm Cu-layers separated by 100 µm

◮ Hole size: 70 µm ◮ Pitch: 140 µm ◮ 360V between Cu-layers ◮ Two GEM foils give a gain of about 104 ◮ “T2K-gas”: 95% Ar, 3% CF4, 2% isobutane

Instrumentation

Figure: The instrumented region of the pad planes (black)

Readout electronics

PCA16:

◮ 16 channel preamplifier and shaper ◮ Modified version of PASA-chip from ALICE. ◮ Programmable gain, shaping and decay time.

ALTRO:

◮ Originally developed for ALICE. ◮ Sampling at 20 MHz ◮ Pedestal subtraction and zero suppression ◮ Capable of storing 1024 10 bit ADC samples.

Next step: Integration of preamplifier and ADC into one chip (S-ALTRO).

Due to the large number of readout channels and the small space available on the pad modules, the electronics had to be connected with 30 cm long Kapton R cables.

Event display

Figure: Typical event without magnetic field. Drift distance: 5 cm Figure: Typical event with magnetic field. Drift distance: 10 cm

Track reconstruction

Figure: Left: Typical pulse. Right: Typical track. ◮ Time is reconstructed as the voltage weighted average of the

five samples around the peak.

◮ Adjacent pulses are grouped into clusters where coordinates

are determined by e.g. y =

P Qiyi P Qi

where Qi is the charge of the pulse and yi is the corresponding y-coordinate of the pad.

◮ For tracking, a simple track reconstruction algorithm was

used.

Residuals

Figure: Upper: Magnified event display. Lower: Residuals integrated over the full track length from 10000 tracks with 7 cm drift length and B=0T, σ ≈ 0.31 mm for a Gaussian core accounting for 95% of the total area. Distortion correcions have been applied.

Distortions

If the residuals are plotted against pad row, they should line up around zero. However:

Figure: Left: Residuals for 10000 tracks vs pad row for B=1T and drift length of 10

same run

After corrections

Resolution in bend plane

Figure: Measured resolution for different drift lengths. The line crosses the y-axis at 0.00349 mm2 which corresponds to an intrinsic resolution of σy(0) = 59.1 ± 0.4µm.

Comparison with theoretical predictions.

Figure: Predicted resolution for different magnetic field strengths and slightly different

Resolution in Z

Figure: Measured resolution in the z-direction for different drift lengths and

Momentum measurements

◮ p ≈ 0.3B · R ◮ σ(1/p) ≈ 9.2 · 10−3 ± 0.0002 GeV−1 ◮ The track fit includes all points along a reconstructed track. Figure: Measured track momenta (left) and 1/p-distribution (right) at a drift length

◮ The momentum resolution has been calculated from a

gaussian fit to the peak covering 42% of the total area.

◮ However, the momentum spread of the beam is ≈ 5% which

gives σ(1/p) ≈ 0.01 GeV−1 at 5 GeV, and therefore the measured width is fully consistent with the beam spread.

Theoretical momentum resolution

◮ Gl¨

uckstern’s formula: δ( 1

PT ) = σy 0.3L2B

N+4 ◮ N = 84, L ≈ 48 cm and B = 1 T. ◮ σy ≈ 76 µm (drift of 15 cm) gives σ(1/p) ≈ 3 · 10−3 GeV−1.

Summary

◮ Test measurements with a TPC using GEM readout have

been performed.

◮ Corrections for electric field distortions have been introduced

using the Millepede software package.

◮ Results on spatial resolution show that σy at zero drift is

59.1 ± 0.4µm and σz at zero drift is 216 ± 7µm.

◮ Result on momentum resolution is

σ(1/pt) ≈ 9.2x10−3 ± 0.0002GeV −1 at a drift length of 15

negligible.

◮ Theoretical estimation on momentum resolution at

σy ≈ 76 µm gives σ(1/p) ≈ 3 · 10−3 GeV−1.

◮ Results on spatial resolution are consistent with the goals for

the full size ILD TPC.

Resolution

◮ Track parameters from fit gives too optimistic estimation of

the resolution.

◮ Use geometric mean of widths of the distributions with

investigated cluster included, σinc, and excluded, σexc, from fit

◮ σ = √σinc · σexc