Tests of a diamond quadrant detector at Desy with Libera ...

1

Tests of a diamond quadrant detector at Hasylab (DESY) using the Libera Brilliance

J Morse European Synchrotron Radiation Facility, France

H Graafsma Hasylab, DESY, Germany

B Solar Hasylab, DESY, Germany

ESRF

B Solar Instrumentation Technologies, Slovenia

2

Acknowledgements

Eleni Berdermann GSI Darmstadt

Michal Pomorski CEA-Saclay

Harris Kagan Ohio State Unversity

Muriel Salomé ESRF Grenoble

Liam Gannon University of Bath

3

Talk Outline

Objectives:

- evaluate the feasibility of using RF readout with diamond beam position monitors;

- compare performance, practical issues… with the (usual) electrometer read-out approach.

1. X-ray Synchrotron beam monitoring requirement why diamond?

2. some background: tests at ESRF

3. the Libera Brilliance system

4. DESY F4 beamline measurements/results

4

global application scale

2009: about 50 synchrotrons in the world…

infra-red to MeV photon beams, but main interest 5 ~ 50keV

ESRF-Grenoble

4

5

‘local’ application : ESRF

European Synchrotron Radiation Facility ESRF

~5000 external user experiments / year

with high intensity, coherent

X-ray beam probes 0.5 ~ 500keV

basic and applied research in

biology (protein structures…)

materials science

chemistry, catalyisis…

(coherent) imaging

-- at micro, nano, molecular & atomicscales …

6

3rd generation synchrotrons

ESRF Ø300m

~ 50 beamlines

Beam position - intensity monitors

white/pink beam 0.2~2kW

monochromatic beam ~mW

undulator source

50 ~100m

source to end station

7

3rd generation synchrotrons

ESRF Ø300m

~ 50 beamlines

Beam position - intensity monitors

white/pink beam 0.2~2kW

monochromatic beam ~mW

undulator source

50 ~100m

source to end station

8

X-ray beamline monitoring requirements

required beam stability ~10% of beam size 0.1 ~ 50µm, nanofocusing goals 10nmmeasurement rates required dc ~ 1kHz (acoustic vibrations !)

Position

accuracy & linearity requirement ≤ 0.1%Intensity:

synchronization with optical lasers in ~psec pump probe experiments (X-ray photon bunches ~50psec at 105~108 pulses/sec

Timing:

minimal beam interference: absorption, scattering, coherence loss beamline compatibility:

package size, operation in air, dirty-vacuum, clean-UHVionizing radiation load >104 Gray/sec

device…

max. absorbed X-ray power: ≤ few mW monochromatic beamsbut ≥100W in ~mm2 ‘white’ beam applications: ONLY possible with diamond

9

why diamond ?

0 10 20 30

10

100

1000

Thic

knes

s fo

r 5%

abs

orpt

ion

(mic

rons

)

X-ray energy (keV)

Diamond (Z= 6) Silicon (Z= 14)

~practical limitssingle crystal CVD

…and short range of photoelectric- or Compton electronZ = 6 low specific X-ray absorption / beam scattering…

- ‘zero’ leakage currentcan use high E-field nsec response

- simple devices can be radiation hard

- outstanding thermal conductivity diamond 2000, cf. Si 150 (Wm-1ºK-1)

10

XBIC, Poly- and single crystal response

XBIC: signal current maps made from x, y raster scan of micron X-ray beam

Polycrystalline:grain-boundaries

trapping and local field distortions, signal response lagX-ray scattering…

Single Crystal:excellent spatial uniformity…‘unity gain’ charge collection with blocking contacts

1σ signal variation 0.103%over 100 point row

11

signal lag with fine-grain polycrystalline

10 secCharge collection increases (prompt + detrapped) with E field 1…5v/µm

beam 15 x 100µm2 ,1.3 x 1012 ph/sec at 12keV

Ralf Menk, 2006 SLS data on polycrystalline ~10µm thick (sourced by Diamond Materials??)

12

operation of diamond XBPM devices

• diamond plate, thin (30…100µm) diamond with ‘X-ray transparent’ <100nm surface contacts Cr, Ti, … Ni, Al (Au, Pt, W))

• in beam, diamond bulk acts as solid state ‘ionization chamber’electron thermalization range ~few microns

• current signal readout ‘DC’ up to synchrotron RF clock frequencies possible

( ) ( )

( ) ( )DCBADCBAY

DCBADBCAX

++++−+

=

++++−+

=

YX

A BC D

position (and intensity) found with…

multiple electrodes:

exploits diffusion splitting (~10µm) of charge

e.g. simple quadrant motif

difference/sum of electrode currents A, B, C, Dgivesbeam 'centre of gravity’

sum of currents gives beam intensity

13





( ) ( )

( ) ( )DCBADCBAY

DCBADBCAX

++++−+

=

++++−+

=

YX

A BC D







Packaged device, ID09B, ID11, Desy F4 tests

14



( ) ( )

( ) ( )DCBADCBAY

DCBADBCAX

++++−+

=

++++−+

=

YX

A BC D










duo- and tetra-lateral devices

linear position response over several mm

(but less precise)

15



( ) ( )

( ) ( )DCBADCBAY

DCBADBCAX

++++−+

=

++++−+

=

YX

A BC D










duo- and tetra-lateral devices

linear position response over several mm

(but less precise)

ID06 tests– see Pomorski talk !

16

metal contacted devices, X-ray response

I-V curves under steady-state X-ray beam illumination (7.2 and 6.0 keV)

-150 -100 -50 50 100

-50

50

100

150

20010nm Ti (annealed)

130nm Au30nm Pd10nm Ni

cur

rent

(nA

)

bias (V)

0.5V

/µm

bias

333µm C*30nm Pd

130nm Au

-300 -200 -100 100 200 300

-6-5-4-3-2-1

123456

~100nm Al

curr

ent (

nA)

bias (V)

~100nm Al

0.5V

/µm

bias96µm C*

Shadow mask, sputtered contacts(GSI Darmstadt)

Lift-off litho’ evaporated contacts(Glasgow University)

current ‘gain’

Si beam flux calibration εDiamond = 13.05 +/-0.2 eV/e-h pair

(ESRF, MI-885)

Blocking contact(s) give saturated current response for >0.3Vµm-1

applied E field:

‘overbias’excellent area response uniformity

17

quadrant devices: position response

Line scan @ 7.2keV

For large beamsize (> 50µm), device ‘crossover response’ is simply the line integral across the beam intensity profile

For a small beam (< 10µm), crossover response is convolution of photoelectron thermalizationrange and lateral charge diffusion ocurring during drift

5.16 5.18 5.20 5.22 5.24 5.26 5.28 5.30 5.32 5.34

-1.6

-1.4

-1.2

-1.0

-0.8

-0.6

-0.4

-0.2

0.0

focussed X-beam 0.4 x 1.2 µm2 FWHM

sig

nal (

nA)

position (mm)

bias -40V

50%

isolation gap between quadrants ~120um

signal slope ~5% /micron

…beam focused <1um

Isolation gap ~120µm

1 2

4.9 5.0 5.1 5.2 5.3 5.4 5.5 5.6 5.7 5.8

-50

-40

-30

-20

-10

0

isolation gap betweeen quadrants ~120µm

dia

mon

d si

gnal

(nA

)

position (mm)

upper left quadrant lower left quadrant

50%

parallel X-beam through 0.2mm pinhole

signal slope ~0.5% /micron

ID21 data, beam collimated 200µm

1 2

!! This data from 0 - 10Hz bandwidth electrometer measurements, i.e. charge integral measurements…what about the time domain??

18

Vertical & horizontal position time scans

ESRFMI-885, ID21 microfocus beamline1sec/point: beam shifts

X-ray flux ~108 s-1 at 7keV~ 20fC in diamond per X ray bunch~ 10nA ‘dc equivalent’ signal current)

0 5000 10000 15000 20000 25000 30000 35000

-1.5

-1.0

-0.5

0.0

0.5

1.0

posi

tion*

(µm

)

time (sec)

vertical horizontal

*scaling 'calibration' error possibly ~10%

19

Vertical & horizontal position time scans

ESRFMI-885, ID21 microfocus beamline1sec/point: beam shifts

X-ray flux ~108 s-1 at 7keV~ 20fC in diamond per X ray bunch~ 10nA ‘dc equivalent’ signal current)

0 5000 10000 15000 20000 25000 30000 35000

-1.5

-1.0

-0.5

0.0

0.5

1.0

posi

tion*

(µm

)

time (sec)

vertical horizontal


30600 30800 31000 31200 31400 31600 31800 32000 32200 32400

-1.5

-1.0

-0.5

0.0


posi

tion*

( µm

)

time (sec)

vertical horizontal

18 x time zoom

refill30740 30760 30780 30800 30820 30840 30860 30880 30900 30920 30940

-0.15

-0.10

-0.05

0.00


180x time zoom horiz position

posi

tion

(µm

)

time (sec)

residuals sd 0.0151µm over 145 points/240secs

30740 30760 30780 30800 30820 30840 30860 30880 30900 30920 30940-1.30

-1.25

-1.20

-1.15

-1.10


180x time zoom

A

residuals sd 0.0204µm over 100 points/166secs (section A->B)

vert position

posi

tion

(µm

)

time (sec)

B

σ =13.3nm rms

σ = 20.4nm rms

20

position timescan and ‘vibrations’, ID09B:

~ 14 keV beam

currents measured with Keithley 485 electrometers, (10Hz BW, mean current/electrode ~10µA

charge generated in diamond ~ 100 fC /pulse

0 5000 10000 15000

0.0

0.1

0.2

0.3

42

43

44

45

46

47

48

sum

of 4

qau

dran

t sig

nals

(µA

)

beam

pos

ition

seconds

ID09B 23 June 2008

1 =

20um

(ver

tical

)1

= 60

um (h

oriz

)

4.4 hours

FFTs using Femto DLPCA-200 current preamps (simultaneous sampling ADCs at 1ksample/sec)

0 50 100 150 200 250 300 350 400 450 5000.0000.0020.0040.0060.0080.0100.0120.0140.0160.0180.020

Ampl

itude

Frequency (Hz)

Average of 10 FFT of 1000 samples

Vertical noise amplitude

10200 10400 10600 10800

0.1

0.2

0.3

42

43

44

45

46

47

481µm

1 =

60um

(hor

iz)

1 =

20um

(ver

tical

)

sum

of 4

qau

dran

t sig

nals

(µA

)

beam

pos

ition

seconds

ID09B 23 June 2008

1µm

Machine artifacts or something upstream on beamline…

21

diamond temporal response

ESRF 4 bunch mode, ID21 beam ~108ph/sec mean flux (very weak beam intensity…)

700ns

<100pS FWHMX pulse duration

DBA-3LeCroy scope LC584A,~1GHz BW

2.3GHz, 38dB

0.5m 5m

Vb=-50V

22

diamond temporal response

ESRF 4 bunch mode, ID21 beam ~108ph/sec mean flux (very weak beam intensity…)

700ns

<100pS FWHMX pulse duration

DBA-3LeCroy scope LC584A,~1GHz BW

2.3GHz, 38dB

0.5m 5m

Vb=-50V

Signal response to crossing of one X-ray bunch

absorption of ~160 photons at 7.2keV (total ~1MeV = 12fC /pulse )

Linear fit to slope gives signal full base width ~2.5ns, e- drift velocity ~40 µm ns-1

at ~1.1 V µm-1

23

wideband position measurements, ID09B

200mV/20ns division(after 10dB attenuator)

20keV beam, incident flux ~1 x 107ph per pulse (1kHz mechanically chopped white beam)~ 5% X-ray absorption in diamond 385µm thick, ~50% photoelectric/50% Compton

~50pC/pulse in diamond (diamond electrode capacity ~ 0.5pF, bias at 500V ‘CV’ charge limit ~200pC)

X-ray beam

signal direct to DSO: poor decoupling and 50Ωmatching signal ‘bounce’

S361-1, sample

TiW contacts processed by Kagan-OSU.

Vertical beam scan 1

42mm

14

electrode signal ~60ns integrals

Qua

dran

t sig

nal

Spatial position (motor scan of diamond)

50µm

‘crossover’ response of electrodes, beam size fwhm 40µm (V), 90µm (H)

‘boxcar’ signal integration

24

i-Tech Libera Brilliance system

1234

X, Y, Σ out

over network

Signals in

~10kHz

High performance if adequate ‘tuned’ RF signal power…

but can it work with diamond signals?

!! developed for stabilization of electron beams

25 ppM

25

what’s inside? performance?

analog stage: tuned filter (352 or 500MHz)

26



-4000

-3000

-2000

-1000

0

1000

2000

3000

AD

C v

alue

isg/d-bpmlibera/1/ADCChannelA

data with attens set auto ??

0 200 400 600 800 1000 1200

ADC sample #

27



-4000

-3000

-2000

-1000

0

1000

2000

3000

0 200 400 600 800 1000 1200

ADC sample #

AD

C v

alue



-2000

0

2000

AD

C v

alue

data with Libera attens set auto??and ~40mV signal pulse amplitude in)

0 30 60 90

ADC sample #

28



0 200 400 600 800 1000 1200-4000

-3000

-2000

-1000

0

1000

2000

3000

ADC sample #

AD

C v

alue



0 30 60 90

-2000

0

2000

AD

C v

alue

ADC sample #


FPGA and µPprocessing buffering, fast I/O

29



0 200 400 600 800 1000 1200-4000

-3000

-2000

-1000

0

1000

2000

3000

ADC sample #

AD

C v

alue



0 30 60 90

-2000

0

2000

AD

C v

alue

ADC sample #



Rok Uršič, I-Tech Dec 2004 for Libera Electron

30



0 200 400 600 800 1000 1200-4000

-3000

-2000

-1000

0

1000

2000

3000

ADC sample #

AD

C v

alue



0 30 60 90

-2000

0

2000

AD

C v

alue

ADC sample #



Rok Uršič, I-Tech Dec 2004 for Libera Electron

Guenther RehmDiamond Light Source, 2008

synchrotron circulating e- beamposition noise for Libera input signal attenuators 0-28dB

31

RF readout: Doris F4 tests May 2009

ESRF-Desy ‘DIMOX’ collaboration (readout of diamond BPMs using Libera electronics)

DORIS F4 BM white beam exit

slits20 x 20µm2

diamond BPMon motorized x-y stage

Bias (NIM unit)

Libera Brilliance

pre-amps

Si diode0.5mm

Al beam absorber plate(s)0.5…3.5mm

E6 SC diamond in ceramic mount before PCB assembly. 389µm thick, 50µm isolation cross, 3mm hole under the diamond for beam passage. ~100nm TiW contact processing: Harris Kagan, OSU

8mm

32

diamond mounting and RF signal cabling

Bias

Four quadrant Single Crystal Diamond Sensor

RF Signal Impendence Matching Circuit

LiberaBPM Electronics

X-rayBeam

Sample

Control System

Modified Brilliance: new +12dB input preamps after crossbar switch

33

Doris F4 bending magnet X-ray beam

0.0 2.0x104 4.0x104 6.0x104 8.0x104 1.0x1050.0

2.0x107

4.0x107

6.0x107

Initial bending magnet flux after 0.5mm Al and 0.5mm Si after2.0mm Al and 0.5mm Si after 3.5mm Al and 0.5mm Si

Flux

(pho

tons

/s/m

m2 /0

.1%

bw)

Energy (eV)

DesyEnergyProfile.opj

Flux incident on diamond after 0.5mm Al absorber ~1.1 x 1012 ph/sec2.9% of incident beam absorbed (photoelectric and Compton)

‘dc’ equivalent current generated in diamond ~15µA (3pC/ pulse at 5Mpps)

effect of full ‘white’ beam on PCB and diamond…Following measurements shown were made after these “accidents”

34

Three slides of results removed from this presentation (these show data that will be included in a publication in preparation)

Please contact speaker directly for these missing slides([email protected])

35

dynamic position response: jump test

Libera ADC buffer data at 130KHz sampling-average

rms position noise* vs. bandwidth

nb. noise includes real beam-sensor movements etc.

µm

µm

36

device modeling with TCAD Sentaurus

High level modeling software for semiconductor devices: 2 & 3D graphics and script input

to describe simple to complex devices.

Program solves Poisson and charge continuity (finite element methods) equations.

Simulates drift, diffusion, recombination etc. of charge carriers, and signals induced on

electrodes for various external load models

Accurate/well tested for silicon devices: input parameter and model files can easily be

configured for other semiconductor materials.

Following slides show FIRST ATTEMPTS at 2D simulations for diamond using

permittivity = 5.7 band gap = 5.47 eV electron/hole mobility = 2300/1800 (cm2/ Vs)

carrier velocity saturation model??

37

boundary conditions, field map and meshing

anode 1 anode 2200µm

cathode L Gannon, Sentaurus Device Editor

38

Signal development during charge transit

0.0 2.0x10-9 4.0x10-9 6.0x10-9 8.0x10-9

-4.0x10-8-2.0x10-8

0.02.0x10-84.0x10-86.0x10-88.0x10-81.0x10-71.2x10-71.4x10-71.6x10-71.8x10-72.0x10-72.2x10-72.4x10-7

Cur

rent

(A)

Time (s)

0.0 2.0x10-9 4.0x10-9 6.0x10-9 8.0x10-9

-4.0x10-8-2.0x10-8

0.02.0x10-84.0x10-86.0x10-88.0x10-81.0x10-71.2x10-71.4x10-71.6x10-71.8x10-72.0x10-72.2x10-72.4x10-7

Cur

rent

(A)

Time (s)

1. charge created near the cathode2. holes reach the cathode and are collected, so signal current is ~halved

3. electrons drift and diffuse across a region of homogenous electric field.

4. as electrons approach anode 1, electric field gradient increases so a rise in current is observed on this anode.

5. As electrons are collected at anode 1 the current decreases to zero (tailing caused by transit diffusion)Sentaurus Device Simulator

bias -550V

ball of charge400µm

200µm0V

20µm

1 20V

39

Signal variation with position of incident beam

200µm

400µm

0V 1 0V2

bias=-550V

column of charge

-2x10-9 0 2x10-9 4x10-9 6x10-9 8x10-9

-5.0x10-7

0.05.0x10-7

1.0x10-6

1.5x10-6

2.0x10-6

2.5x10-6

3.0x10-6

3.5x10-6

4.0x10-6

Cur

rent

(A)

Time (s)

-2x10-9 0 2x10-9 4x10-9 6x10-9 8x10-9

-5.0x10-7

0.05.0x10-7

1.0x10-6

1.5x10-6

2.0x10-6

2.5x10-6

3.0x10-6

3.5x10-6

4.0x10-6

Cur

rent

(A)

Time (s)

-2x10-9 0 2x10-9 4x10-9 6x10-9 8x10-9

-5.0x10-7

0.05.0x10-7

1.0x10-6

1.5x10-6

2.0x10-6

2.5x10-6

3.0x10-6

3.5x10-6

4.0x10-6

Cur

rent

(A)

Time (s)

-2x10-9 0 2x10-9 4x10-9 6x10-9 8x10-9

-5.0x10-7

0.05.0x10-7

1.0x10-6

1.5x10-6

2.0x10-6

2.5x10-6

3.0x10-6

3.5x10-6

4.0x10-6

Cur

rent

(A)

Time (s)

-2x10-9 0 2x10-9 4x10-9 6x10-9 8x10-9

-5.0x10-7

0.05.0x10-7

1.0x10-6

1.5x10-6

2.0x10-6

2.5x10-6

3.0x10-6

3.5x10-6

4.0x10-6

Cur

rent

(A)

Time (s)

-2x10-9 0 2x10-9 4x10-9 6x10-9 8x10-9

-5.0x10-7

0.05.0x10-7

1.0x10-6

1.5x10-6

2.0x10-6

2.5x10-6

3.0x10-6

3.5x10-6

4.0x10-6

Cur

rent

(A)

Time (s)-2x10-9 0 2x10-9 4x10-9 6x10-9 8x10-9

-5.0x10-7

0.05.0x10-7

1.0x10-6

1.5x10-6

2.0x10-6

2.5x10-6

3.0x10-6

3.5x10-6

4.0x10-6

Cur

rent

(A)

Time (s)-2x10-9 0 2x10-9 4x10-9 6x10-9 8x10-9

-5.0x10-7

0.05.0x10-7

1.0x10-6

1.5x10-6

2.0x10-6

2.5x10-6

3.0x10-6

3.5x10-6

4.0x10-6

Cur

rent

(A)

Time (s)-2x10-9 0 2x10-9 4x10-9 6x10-9 8x10-9

-5.0x10-7

0.05.0x10-7

1.0x10-6

1.5x10-6

2.0x10-6

2.5x10-6

3.0x10-6

3.5x10-6

4.0x10-6

Cur

rent

(A)

Time (s)-2x10-9 0 2x10-9 4x10-9 6x10-9 8x10-9

-5.0x10-7

0.05.0x10-7

1.0x10-6

1.5x10-6

2.0x10-6

2.5x10-6

3.0x10-6

3.5x10-6

4.0x10-6

Cur

rent

(A)

Time (s)-2x10-9 0 2x10-9 4x10-9 6x10-9 8x10-9

-5.0x10-7

0.05.0x10-7

1.0x10-6

1.5x10-6

2.0x10-6

2.5x10-6

3.0x10-6

3.5x10-6

4.0x10-6

Cur

rent

(A)

Time (s)-2x10-9 0 2x10-9 4x10-9 6x10-9 8x10-9

-5.0x10-7

0.05.0x10-7

1.0x10-6

1.5x10-6

2.0x10-6

2.5x10-6

3.0x10-6

3.5x10-6

4.0x10-6

Cur

rent

(A)

Time (s)

anode 1 anode 2

Sentaurus Device Simulator

40

Conclusions:

‘Proof of principle’ established for position readout using Libera Brilliance systemresolution < 0.1µm demonstrated, but initial DESY tests limited by (white) beam size and beam position noise

Further quantitative tests needed to directly compare narrowband RF vs. electrometer readout, especially signal/noise performance vs. absorbed beam energy

(new test in planning, will use ~10keV monochromatic X-ray beam at ESRF)

Better understanding needed of signal development in multi-electrode device coupled with response of signal processor, e.g. Libera: system ~2MHz passband at 352MHz

(modelling just started with TCAD-Sentaurus)

Tests of a diamond quadrant detector at Desy with Libera ...

Documents