x-ray imaging with high-z sensors for the esrf-ebs upgrade

X-ray imaging with high-Zsensors for the ESRF-EBS

Upgrade

Stergios TsigaridasPIXEL2018, TaipeiDecember 13, 2018

European Synchrotron Radiation Facility - ESRF

� ESRF is a 3rd generation synchrotronlocated in Grenoble/France

� 22 partner nations (13 Members statesand 9 Scientific Associates)

� Offering 44 beamlines in total

� More than 5000 individual user visitsper year

| December 13, 2018 | Stergios Tsigaridas

ESRF EBS Upgrade Programme

Today

EBS

� Use same tunnel - newstorage ring

� Brilliance and coherence willbe increased by a factor 100

� Upgrade programme for beaminstrumentation and detectors

e− beam properties Today EBS

Energy (GeV) 6.04 6Multi-bunch current (mA) 200 200Circumference (m) 844.39 843.98Horizontal emittance (pm ·mrad) 4000 140Vertical emittance (pm ·mrad) 4 5

Ref: ESRF Upgrade Phase II (Technical Design Study)


http://www.esrf.eu/Apache_files/Upgrade/ESRF-orange-book.pdf

Hybrid pixel detector developement at ESRF

� Need for efficient detectors at moderate andhigh x-ray energies (30-100 keV)

� Research programme focusing on high-Zsensors for direct detection

� Photon counting or charge integrating

Property Si GaAs. CdTe CZT

Z (average) 14 32 50 49.1Density (g cm−3) 2.33 5.32 5.85 5.81Energy gap (eV) 1.12 1.42 1.44 >1.6Intr. carrier ∼ 1010 2 · 106 ∼ 107 ∼ 107

conc. (cm−3)W-value (eV) 3.62 4.3 4.43 4.64

0 20 40 60 80 100Energy (keV)

0

20

40

60

80

100

Phot

oele

ctric

abs

orpt

ion

(%)

Si (500 m)

GaAs (500 m)

CdTe (1 mm)

CZT (2 mm)

Ref: Data obtained from XCOM NIST database


https://physics.nist.gov/PhysRefData/Xcom/html/xcom1.html

Hybrid pixel detectors for photon counting

� Medipix/Timepix family of chips

� 256 × 256 pixels, 55 µm pitch

� Total area of 2 cm2

� Single or multi-chip modules

� Sensor is bump bonded to chip

� Wire bonds to chip carrier board

Table : Portfolio of high-Z sensors

Sensor Thickn. Pixel Contact Vendor(mm) (µm) (type)

CdTe 1 55 Ohmic AcroradGaAs 0.5 55 Ohmic TomskCZT 2 55

Schottky RedlenCZT 2 110

.... . .

...


Maxipix readout system

Ref: C. Ponchut et al. 2011 JINST 6 C01069

� Developed by ESRF (Medipix2/Timepix)

� 1.4 kHz maximum frame rate

� 2 · 105 counts/pixel maximum count rate

� 0.29 ms dead time at 100 MHz clock

� Minimum threshold ∼ 4 keV

� 11810 pixel counter depth


http://stacks.iop.org/1748-0221/6/i=01/a=C01069

How to choose “the sensor” (?)

� Apply a common characterization procedure

� Lab/X-ray generator� Wide filtered X-ray beam� Flat field images, count rate/stability tests� Imaging performance and spatial resolution� Charge collection efficiency tests

� Synchrotron beamline� Micro-focused (5-10 µm) monochromatic beam� Sub-micron precision translation stages� Spatial characteristics of the sensor (resolution, distortions)� Single pixel spectra


Leakage current vs bias voltage

500 400 300 200 100 0vBias (V)

1.75

1.50

1.25

1.00

0.75

0.50

0.25

0.00

iBia

s (A)

1e 5 CdTe (1 mm)

Ramp upRamp down

1000 800 600 400 200 0vBias (V)

2.00

1.75

1.50

1.25

1.00

0.75

0.50

0.25

0.00

iBia

s (A)

1e 7 CZT (2 mm)

Ramp upRamp down

300 250 200 150 100 50 0vBias (V)

4

3

2

1

0

iBia

s (A)

1e 6 GaAs (500 m)

Ramp downRamp up

CdTe CZT GaAs.

Sensor bias voltage (V) -500 -1000 -300Bias leakage current (µA) 20* 0.2 5Equivalent pixel leakage current (pA) 305 3 76


Count rate tests

� Detector 1.5 m away from source

� Ramp up voltage and X-rays on

� Ag Anode/100 µm filter (∼ 22 keV)

� Generator settings (35 kV/25 mA)

0 100 200 300 400time (s)

0

1

2

3

4

5

Coun

ts /

mm

2 /s

1e5 CdTe (1 mm)

0 100 200 300 400time (s)

0

1

2

3

4

5

Coun

ts /

mm

2 /s

1e5 CZT (2 mm)

0 200 400 600 800 1000time (s)

0

1

2

3

4

5

Coun

ts /

mm

2 /s

1e5 GaAs (500 m )


Leakage current stability

� Detector 1.5 m away from source

� Ramp up voltage and X-rays on

� Ag Anode/100 µm filter (∼ 22 keV)

� Generator settings (35 kV/25 mA)

0.0 0.5 1.0 1.5 2.0 2.5time (h)

10 8

10 7

10 6

10 5

10 4

iBia

s (A)

CdTe (1 mm )

0.0 0.5 1.0 1.5 2.0 2.5time (h)

10 8

10 7

10 6

10 5

10 4

iBia

s (A)

CZT (2 mm )

0.0 0.5 1.0 1.5 2.0 2.5time (h)

10 8

10 7

10 6

10 5

10 4

iBia

s (A)

GaAs (500 m )


Flat field images (1 s exposure)

� Information about the quality ofsensor/bump bonding/etc

� Tiny dots correspond to masked orunconnected pixels

� Defects appear with the form of linesor blobs

� How does this affect theperformance??

0 50 100 150 200 250pixel

0

50

100

150

200

250

pixe

l

CdTe (1 mm)

0

500

1000

1500

2000

2500

3000

0 50 100 150 200 250pixel

0

50

100

150

200

250

pixe

l

CZT (2 mm)

0

500

1000

1500

2000

2500

3000

0 50 100 150 200 250pixel

0

50

100

150

200

250

pixe

l

GaAs (500 m)

0

500

1000

1500

2000

2500

3000


About 3h later...

0 50 100 150 200 250pixel

0

50

100

150

200

250

pixe

l

CdTe (1 mm)

0

500

1000

1500

2000

2500

3000

0 50 100 150 200 250pixel

0

50

100

150

200

250

pixe

l

CZT (2 mm)

0

500

1000

1500

2000

2500

3000

0 50 100 150 200 250pixel

0

50

100

150

200

250

pixe

l

GaAs (500 m)

0

500

1000

1500

2000

2500

3000

0 50 100 150 200 250pixel

0

50

100

150

200

250

pixe

l

CdTe (1 mm)

0

500

1000

1500

2000

2500

3000

0 50 100 150 200 250pixel

0

50

100

150

200

250

pixe

lCZT (2 mm)

0

500

1000

1500

2000

2500

3000

0 50 100 150 200 250pixel

0

50

100

150

200

250

pixe

l

GaAs (500 m)

0

500

1000

1500

2000

2500

3000


Imaging performance and spatial characteristics

0 50 100 150 200 250pixel

0

50

100

150

200

250

pixe

lCZT (55 m pitch)

0

2500

5000

7500

10000

12500

15000

17500

20000

0 20 40 60 80 100 120pixel

0

20

40

60

80

100

120

pixe

l

CZT (110 m pitch)

0

10000

20000

30000

40000

50000

60000

70000

0 50 100 150 200 250pixel

0

50

100

150

200

250

pixe

l

CZT (55 m pitch)

0

2500

5000

7500

10000

12500

15000

17500

20000

0 20 40 60 80 100 120pixel

0

20

40

60

80

100

120

pixe

l

CZT (110 m pitch)

0

10000

20000

30000

40000

50000

60000

70000

0 50 100 150 200 250pixel

0

50

100

150

200

250

pixe

l

CZT (55 m pitch)

0

2500

5000

7500

10000

12500

15000

17500

20000

0 20 40 60 80 100 120pixel

0

20

40

60

80

100

120

pixe

l

CZT (110 m pitch)

0

10000

20000

30000

40000

50000

60000

70000


Modulation tranfer function (MTF) - presampling

� Make use of slanted edge method

� Edge/line spread functions (ESF/LSF)

0 2 4 6 8 10 12 14 16 18lp/mm

0.0

0.2

0.4

0.6

0.8

1.0

MTF

CdTe (1.0 mm)GaAs (0.5 mm)CZT (2.0 mm)

� LSF (x) = ddx {ESF (x)}

� MTF (u) = |F {LSF (x)}|

0 2 4 6 8 10 12 14 16 18lp/mm

0.0

0.2

0.4

0.6

0.8

1.0

MTF

CZT (55 m)CZT (110 m)


Detective quantum efficiency (DQE)

� Measure of the image quality

� DQE = MTF 2

q ·NNPS

0 2 4 6 8 10lp/mm

0.0

0.2

0.4

0.6

0.8

1.0

DQE

CdTe (1.0 mm)GaAs (0.5 mm)CZT (2.0 mm)

� q: Incident flux per mm2

� NNPS: Normalized Noise Power Spectrum

0 2 4 6 8 10lp/mm

0.0

0.2

0.4

0.6

0.8

1.0

DQE

CZT (55 m)CZT (110 m)


Mesh scans with micro-focused beam

� Be compound refractive lenses

� Beam spot << pixel size

� Precise mapping of a sensor area

� Effective pixel shape







Page 16 | December 13, 2018 | Stergios Tsigaridas






0 5 10 15 20 25 30Scan step (a.u.)

0

5

10

15

20

25

30

Scan

step

(a.u

.)

Low threshold

16000

16500

17000

17500

18000

18500

19000

19500

0 5 10 15 20 25 30Scan step (a.u.)

0

5

10

15

20

25

30

Scan

step

(a.u

.)Medium threshold

7500

8000

8500

9000

9500

10000

10500

11000

0 5 10 15 20 25 30Scan step (a.u.)

0

5

10

15

20

25

30

Scan

step

(a.u

.)

High threshold

4500

5000

5500

6000

6500

7000

7500


Spation distortions

� Align beam on a visible defect

� Mesh scan surrounding pixels

� Set threshold close to minimum value

� Saturation when beam in inter-pixel area

0 50 100 150 200 250pixel

0

50

100

150

200

250

pixe

l

CdTe (1 mm)

0

500

1000

1500

2000

2500

3000

0 50 100 150 200 250pixel

0

50

100

150

200

250

pixe

l

CZT (2 mm)

0

2500

5000

7500

10000

12500

15000

17500

20000

0 50 100 150 200 250pixel

0

50

100

150

200

250

pixe

l

GaAs (500 m)

0

500

1000

1500

2000

2500

3000


Spation distortions






Spation distortions





0 20 40 60Scan step (a.u.)

0

10

20

30

40

50

60

Scan

step

(a.u

.)

CdTe (1 mm)

26000

28000

30000

32000

34000

36000

38000

40000

0 20 40 60Scan step (a.u.)

0

10

20

30

40

50

60

Scan

step

(a.u

.)

CZT (2 mm)

30000

32500

35000

37500

40000

42500

45000

47500

0 20 40 60Scan step (a.u.)

0

10

20

30

40

50

60

Scan

step

(a.u

.)

GaAs (500 m)

250000

300000

350000

400000

450000

500000

550000

Ref: C. Ponchut et al., 2017 JINST 12 C12023


http://stacks.iop.org/1748-0221/12/i=12/a=C12023

Threshold scans and single pixel spectra

� Align beam at the center of a pixel

� Vary the discriminator threshold

� Measure the number of counts foreach threshold step

� The derivative of the threshold scangives the single pixel energy spectrum

� The peak should correspond with thebeam energy


Threshold scans and single pixel spectra

� Align beam at the center of a pixel

� Vary the discriminator threshold

� Measure the number of counts foreach threshold step

� The derivative of the threshold scangives the single pixel energy spectrum

� The peak should correspond with thebeam energy

10 20 30 40 50 60ethl (keV)

0

100

200

300

400

500

Coun

ts

mean = 29.98sigma = 1.29

CZT (2 mm, 110 m)


Single pixel spectra of a 3×3 area

� CZT (2 mm, 55 µm)

� Crosscheck forcalibration

10 20 30 40 50 60Energy (keV)

0

100

200

300

400

500 mean = 29.79sigma = 1.32

Pixel_9

10 20 30 40 50 60Energy (keV)

0

100

200

300

400

500 mean = 29.89sigma = 1.31

Pixel_8

10 20 30 40 50 60Energy (keV)

0

100

200

300

400

500

Coun

ts

mean = 29.59sigma = 1.28

Pixel_710 20 30 40 50 60

0

100

200

300

400

500 mean = 29.34sigma = 1.26

Pixel_6

10 20 30 40 50 600

100

200

300

400

500 mean = 29.46sigma = 1.31

Pixel_5

10 20 30 40 50 600

100

200

300

400

500

Coun

ts

mean = 29.34sigma = 1.32

Pixel_410 20 30 40 50 60

0

100

200

300

400

500 mean = 29.64sigma = 1.29

Pixel_3

10 20 30 40 50 600

100

200

300

400

500 mean = 29.98sigma = 1.29

Pixel_2

10 20 30 40 50 600

100

200

300

400

500

600

Coun

ts

mean = 27.98sigma = 1.19

Pixel_1


Outlook

� Shutdown for the EBS Upgrade (since Monday)

� Characterization of high-Z sensors at ESRF

� Preliminary results - Analysis ongoing

� More tests planned at higher energiesand with more sensors

� The choice of “the sensor” would be a compromise


Outlook

� Shutdown for the EBS Upgrade (since Monday)

� Characterization of high-Z sensors at ESRF

� Preliminary results - Analysis ongoing

� More tests planned at higher energiesand with more sensors

� The choice of “the sensor” would be a compromise

Thank you!


Backup

Backup

� Over-sample image

� Rotate

� Average slices over the edgeto obtain ESF

� Differentiate to obtain LSF

� The amplitude of Fouriertransform gives the MTF

MTF (u)=|DFT1D{LSF (x)}|

Backup

� Noise only image

ni (x ,y)=Ii (x ,y)−I (x ,y)

� Noise Power Spectrum

NPS2D(u,v)=d2p

NxNy1N

∑Ni=1|DFT2D{ni (x ,y)}|2

� Normalised Noise Power Spectrum

NNPS2D(u,v)= NPS(u,v)

〈meanpixelsignal〉2

� Averaging

NNPS2D(u,v)Averaging−−−−−−→ NNPS1D(u)

0 50 100 150 200 250

0

50

100

150

200

250

Noise only image

10080604020

020406080100

0 20 40 60 80 100 120

0

20

40

60

80

100

120

NNPS_2D

102

103

104

105

0 20 40 60 80 100 120pixel−1

1.20

1.25

1.30

1.35

1.40

NNPS

1e 7 NNPS_1D

0 1 2 3 4 5 6 7 8 9lp/mm

1.20

1.25

1.30

1.35

1.40 1e 7 NNPS_1D

x-ray imaging with high-z sensors for the esrf-ebs upgrade

Documents