High-Rate Capable Micromegas Detectors for ... - LMU Munich€¦ · true, simulation • reconstruction of all particles up to 10MHz = 7MHz/cm2 track nding algorithm track inclination

High-Rate Capable Micromegas Detectorsfor Ion Transmission Radiography Applications

Jona Bortfeldt et al.

LS SchaileLudwig-Maximilians-Universitat Munchen

Symposium: Advanced Semiconductor Detectors for Medical ApplicationsFebruary 13th 2015

Introduction

Katia Parodi: Motivation and Requirements

Ion Transmission Radiography

• ions with known initial energy, higher than intherapy

• residual energy measurement→ energy loss→ contrast

range telescopehead tumor

ions

430MeV/u

Present Setup at HIT

• mean particle position from steering magnets (carbon beam ∼3.4 mm FWHM)

• integrate over several 102 to 104 particles

Future

• single particle tracks from Micromegas• spatial resolution ∼0.1 mm• MHz/cm2 particle rate (multi-hit separation, signal duration)• low material budget

• single particle range/energy from suitable telescope• scintillator based• maximum rate O(MHz)

→ improve spatial resolution, decrease dose

Jona Bortfeldt (LMU Munchen) High-Rate Capable Micromegas for Ion Radiography 13/02/15 2

Introduction

Katia Parodi: Motivation and Requirements

Ion Transmission Radiography

• ions with known initial energy, higher than intherapy

• residual energy measurement→ energy loss→ contrast

range telescopehead tumor

ions

tracking Micromegas

Present Setup at HIT

• mean particle position from steering magnets (carbon beam ∼3.4 mm FWHM)

• integrate over several 102 to 104 particles

Future

• single particle tracks from Micromegas• spatial resolution ∼0.1 mm• MHz/cm2 particle rate (multi-hit separation, signal duration)• low material budget

• single particle range/energy from suitable telescope• scintillator based• maximum rate O(MHz)

→ improve spatial resolution, decrease dose


Introduction

The Micromegas Detector

cathode

mesh

copperanode strips

pillars 128μm

Ar:CO2

6mm0.5kV/cm0.047mm/ns

39kV/cm

250μm

charged particle

ioni

zatio

n

• gas amplification 103 to 104

• charge signal on stripssingle strip readout→ spatial resolution O(50µm)→ timing O(ns)

• thin amplification gap & finesegmentation→ fast drain of positive ions→ high-rate capable

• COMPASS: precision tracker, highflux

• CAST: photon detector, good energyresolution, low background

• T2K: TPC readout, large area


Floating Strip Micromegas Principles

Floating Strip Micromegas

-500V

-800Vcathode

mesh

copperanode strips

pillars 128μm

6mmAr:CO2

250μm 150μm

0.5kV/cm

39kV/cm

time [ms]-10 0 10 20 30 40 50 60 70 80 90

volta

ge [V

]

300

350

400

450

500

550

600

standard Micromegas

challenge: discharges

• charge density ≥ 2× 106 e/0.01 mm2

(Raether limit)

• conductive channel→ potentials equalize

• non-destructive,but dead time→ efficiency drop

idea: minimize the affected region

• “floating” copper strips:• strip can “float” in a discharge• individually connected to HV via 22MΩ• capacitively coupled to readout electronics

via pF HV capacitor• only two or three strips need to be

recharged

→ optimization in dedicated measurements& detailed simulation



Floating Strip Micromegas

+500V

-300Vcathode

mesh

copperanode strips

pillars 128μm

6mmAr:CO2

250μm 150μm

0.5kV/cm

39kV/cm

large R

small C

time [ms]-10 0 10 20 30 40 50 60 70 80 90

volta

ge [V

]

300

350

400

450

500

550

600

standard Micromegas

floating strip MM

challenge: discharges

• charge density ≥ 2× 106 e/0.01 mm2

(Raether limit)

• conductive channel→ potentials equalize

• non-destructive,but dead time→ efficiency drop

idea: minimize the affected region

• “floating” copper strips:• strip can “float” in a discharge• individually connected to HV via 22MΩ• capacitively coupled to readout electronics

via pF HV capacitor• only two or three strips need to be

recharged

→ optimization in dedicated measurements& detailed simulation



Discharge Study with Floating Strip Micromegas

22M

22M

22M

22M

1k

+HV

10nF

anode strips

mesh

15pF22MΩ

cathode

• alpha source→ induces discharges

• voltage drop on one to threestrips→ recharge current

• global high voltage drop→ affects all strips

• voltage signal on sevenneighboring strips→ discharge topology



Optimization of the Floating Strip Principle

• standard Micromegas (approximate): 100 kΩ300 V drop, dead time ∼80 ms

• intermediate: 1 MΩ20 V drop, dead time ∼10 ms

• floating strip: 22 MΩ0.5 V drop → negligible

ground

signal

mesh +594V

SMD resistorcopper strips

cathode -300V

SMD capacitor 15pF

signal

measured average voltage pulse



Detailed Investigation of the Global Voltage Drop

maximum voltage drop [V]0 0.1 0.2 0.3 0.4 0.5 0.6 0.7

disc

harg

es /

1mV

0

20

40

60

80

100

120

140

160

180

200

220

• measure voltage drop of common HVpotential

• discrete structure→ probably corresponds to discharge ofone, two or three strips

• how can we show this?→ investigate discharge topology→ develop simulation→ compare predicted with measuredvoltage drop



Discharge Topology - Expected Amplitude Correlation

amplitude strip 3

am

plit

ud

e s

trip

2

only 2

only 3

4...

...1

1 2 3 4

expected correlation

• measure voltage signal on neighboringstrips

• two reasons for signals on strips:• discharge onto strip• capacitive coupling from neighboring

strips



Discharge Topology - One Strip

voltage channel 3 [V]0 0.1 0.2 0.3 0.4 0.5 0.6 0.7

volta

gech

anne

l2[V

]

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0

2

4

6

8

10

12

2L34L

34L

2L3

3L45L3L4

45L

1L23L

23L

1L2

0L12L12L

0L1

• discharges on separate stripsdistinguishable

• substructure quantitativelydescribed by simulation

amplitude strip 3

am

plit

ud

e s

trip

2

only 2

only 3

4...

...1

1 2 3 4

expected correlation



LTSpice Detector Simulation

mesh strip

mesh strip

mesh stripCcable

Casas

Casas

CmsCcoupl

Rhv

HV

Cslsl

Cslsl

Cslg

Rvd1

Rvd2

Chvgdetector HV

Chvcoupl 1MΩ

Ustrip

Uhv

• consider the involved capacitances e.g. between neighboring strips, couplingcapacitors, cable capacitance ...

• simulate discharges (blue switch)



Optimum Configuration: Global Voltage Drop

maximum voltage drop [V]0 0.1 0.2 0.3 0.4 0.5 0.6 0.7

disc

harg

es /

1mV

0

20

40

60

80

100

120

140

160

180

200

220

measurementsimulation: one strip

simulation: two strips

simulation: three strips • good agreement between simulation andmeasurement

• only two free parameters• response time of HV supply: 500 ms• voltage difference between strips at which

leakage stops: 220 V

• peaks correspond indeed to discharge ofone, two or three strip



floating strip principle works

• discharges: negligible effect on common high-voltage

• discharges are localized

measurements

• ion tracking at highest rates at HIT – Micromegas tests

• gas studies and µTPC reconstruction at Tandem/Garching

• first test of a 2d ion radiography system at Tandem/Garching


Measurement: Highest-Rate Ion Tracking

Ion Tracking with Thin Micromegas at Highest Rates @ HIT

APV25frontend

APV25frontend

scintillators + photomultipliers

particlebeam

APV25frontend

APV25frontend

scatteredparticles

Micromegas 0,one-dimensionalreadout

Micromegas 1,one-dimensionalreadout

Micromegas 2,two-dimensionalreadout

y

z

beams

• 12C @ 88 MeV/u to 430 MeV/u2MHz to 80MHz

• p @ 48 MeV to 221 MeV80MHz to 2GHz

• thanks to S. Brons and the HITaccelerator team for the support

floating strip Micromegas

• 6.4×6.4 cm2 doublet

• low material budget(FR4 + Cu ≤ 200 µm)

• Ar:CO2 93:7 gas mixture

additional detectors

• 9×9 cm2 monitoring Micromegas withx-y-readout

• trigger on secondary charged particles



Pulse Height for 88 MeV/u 12C

pulse height vs Eamp

[kV/cm]ampE25 26 27 28 29 30 31

puls

e he

ight

[adc

cha

nnel

s]

300

400

500600700

1000

2000

3000

4000

micom 0

micom 1

• exponential rise as expected(Townsend theory)

• gas gain can be selected overwide range as needed

• 30 kV/cm = 450 V

cathode

anode

Edrift

Eamp

mesh

pulse height vs Edrift

[kV/cm]driftE0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1

puls

e he

ight

[adc

cha

nnel

s]

0

200

400

600

800

1000

1200

1400

1600

1800

2000micom 0

micom 1

Edrift < 0.15 kV/cm:

• low charge separation

• low drift velocity

large Edrift > 0.5 kV/cm:

• low electron mesh transparency



Efficiency for 88 MeV/u 12C

efficiency vs Eamp

[kV/cm]ampE25 26 27 28 29 30 31

effic

ienc

y

0.84

0.86

0.88

0.9

0.92

0.94

0.96

0.98

1

micom0

micom1

efficiency vs Edrift

[kV/cm]ampE0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1

effic

ienc

y

0.84

0.86

0.88

0.9

0.92

0.94

0.96

0.98

1

micom0

micom1

optimum value:> 99% in micom0> 94% in micom 1 due to production fault



Beam Characterization

signal timing 12C, 5× 106 Hz

signal timing [25ns]0 2 4 6 8 10 12 14 16 18

sign

als

/ ns

0

100

200

300

400

500

600

bunch spacing

beam energy [MeV/u]0 50 100 150 200 250 300 350 400 450

bunc

h sp

acin

g [n

s]

100

120

140

160

180

200

220

240

260

280

300

micom 0

micom 1

expected

• good multihit resolution

• bunch spacing measureable

• bunch filling measureable



Signals at Lowest and Highest Rate

12C, E = 430 MeV/u, 5 MHz

strip20 40 60 80 100 120

time [25ns]0246810121416

0

200

400

600

800

1000

1200

1400

1600

puls

e he

ight

[arb

. uni

ts]

3 particles clearly distinguishable→ single particle tracking possible

p, E = 221 MeV, 2 GHz

strips20 40 60 80 100 120

time [25ns]0246810121416

-500

0

500

1000

1500

puls

e he

ight

[arb

. uni

ts]

integration over ∼ 800 coincident particles→ envelope of beam profile



Pulse Height & Spatial Resolution vs Rate for 88 MeV/u 12C

mean particle rate [Hz]0 10 20 30 40 50 60 70 80

610×

puls

e he

ight

[adc

cha

nnel

s]

500

1000

1500

2000

2500

1000

2000

3000

4000

5000

micom0micom1micom2

C beam12

• up to 80 MHz single particle tracks visiblebut not all of them separable

• only 20% pulse height reduction @80 MHz


610×

m]

µsp

atia

l res

olut

ion

[

0

20

40

60

80

100

120

140

160

180

• highest rates: slight distortion of hitposition by hits on adjacent strips

• limited by multiple scattering

• sufficient for medical application

→ tracking of carbon ions at highest rates possible



Detection Efficiency and Up-Time

p, 221 MeV

mean particle rate [Hz]0 500 1000 1500 2000

610×

effic

ienc

y

0.7

0.75

0.8

0.85

0.9

0.95

1

proton beammicom0micom1micom2

single particle detection efficiency

detector up-time

→ no efficiency & up-time reduction in floating strip Micromegas



Rate Capability & Multi-hit Resolution

reconstructed hits per multi-event


610×

num

ber

of h

its in

det

ecto

r

1

2

3

456

10

20

30

40

reconstr., measurementreconstr., simulationtrue, simulation

• reconstruction of all particlesup to 10 MHz = 7 MHz/cm2

track finding algorithm

[rad]αtrack inclination -0.15 -0.1 -0.05 0 0.05 0.1 0.15

trac

k di

stan

ce r

[mm

]10

20

30

40

50

60

0

20

40

60

80

100

120

140

160

180

• Hough transform: d = x · cos(α) + z · sin(α)point in position space ⇔ line in Hough spaceline in position space ⇔ point in Hough space

• up to seven coincident tracks reconstructable


Mesurement: Gas Studies, µTPC and Range Telescope

23 MeV Proton Tracking at the Tandem/Garching

vacuum-flange

Kapton window

protonbeam

2 Micromegas doublets

z

x

multi-anode-

PM

Scintillator Range Telescope

absorber

goal

• further improve Micromegas high-rate capability↔ decrease signal duration→ fast Ne:CF4 gas mixtures

• investigate single plane track inclination reconstruction

• commission range telescope

• test custom amplifier electronics

floating strip Micromegas

• two 6.4×6.4 cm2 doublets,128 strips

• low material budget(FR4 + Cu ≤ 200 µm)

• APV25 based readout

range telescope

• 13 layers 1 mm scintillator

• two wavelength-shiftingfibers per layer

• read out with 64 pixelmulti-anode photomultiplier

• discrete voltage &spectroscopy amplifiers

• VME based QDC & TDCreadout system



Pulse Height and Efficiency for Ne:CF4 80:20

pulse height vs Edrift

[kV/cm]dE0 0.2 0.4 0.6 0.8 1 1.2 1.4

puls

e he

ight

[adc

cha

nnel

s]

0

1000

2000

3000

4000

5000

6000

• high gas gain

• low diffusion→ moderate decrease withincreasing drift field

cathode

anode

Edrift

Eamp

mesh

efficiency vs Edrift

[kV/cm]driftE0 0.2 0.4 0.6 0.8 1 1.2 1.4

effic

ienc

y

0.96

0.965

0.97

0.975

0.98

0.985

0.99

0.995

1

• above 96% for all drift fields

→ excellent performance with new mixture



Track Inclination Reconstruction in a Single Detector Plane

cathode

mesh

8mmϑ zcluster= t vdrift

xcluster

linear fit to data points

strip83 84 85 86 87 88 89 90

time

[25n

s]-1

0

1

2

3

4

5

6

rise time fit

time [25ns]0 2 4 6 8 10 12 14 16 18

char

ge[a

dcch

anne

ls]

0

100

200

300

400

500

600

700

800

50%

90%

10%

t0 tF

pF

simulated signals

time [s]0 0.2 0.4 0.6 0.8 1

-610×

char

ge [a

rb. u

nits

]

0

0.05

0.1

0.15

0.2strip 5strip 6strip 7strip 8strip 9strip 10strip 11

method:

• arrival time ↔ drift distance

• measure arrival time of chargecluster on strip→ signal timing t0

• linear fit to time-strip data points→ track inclination→ alternative hit position→ drift velocity

systematics:

• capacitive coupling of signals ontoneighboring strips

• simulation with parameter-freeLTSpice detector model



Track Inclination Measurement in a Single Detector Plane with Ar:CO2 93:7

]°reconstructed track inclination [15 20 25 30 35 40 45 50

entr

ies

0

50

100

150

200

250

300

°10°20°30°40

measurement

]°track inclination [

]°re

cons

truc

ted

angl

e [

20

25

30

35

40datasimulation

]°track inclination [0 5 10 15 20 25 30 35 40 45

]°an

gula

r re

solu

tion

[-10

-505

101520 measurement

• track inclination reconstruction possible for angles 20 ≤ ϑ ≤ 40

with angular resolution(

+6−4

)• systematic effect understood → calibration possible

• combined position reco possible (µTPC + centroid)



Track Inclination Measurement with the New Gas Ne:CF4

80:20 75:25

• track inclination reconstruction possible with fast Ne:CF4 gas mixture

• angular resolution(

+5−4

)for Edrift < 0.6 kV/cm



Improvement of High-Rate Capability with Ne:CF4

measured electron drift velocity

• signal duration =electron drift time + ion drift time

• electron drift time: 150 ns → 60 ns

• ion drift time: 260 ns → 85 ns

→ factor 3 improvement



Range Telescope Commissioning – Pulse Height Behavior

proton traverses layer→ constant energy loss

pulse height fiber 0 [arb. units]0 100 200 300 400 500 600

puls

e he

ight

fibe

r 1

[arb

. uni

ts]

0

100

200

300

400

500

600

700

800

0

5

10

15

20

25

30

proton stops in layer→ variable energy loss

pulse height fiber 0 [arb. units]0 100 200 300 400 500 600

puls

e he

ight

fibe

r 1

[arb

. uni

ts]

0

100

200

300

400

500

600

700

800

0

2

4

6

8

10

12

14

16

18

only 0.1% of the photons detectable→ ∼ 30± 10 photons→ ∆EFWHM/E ∼ 1

mean pulse height vs depth

• pulse height increases less thanexpected

• probably considerable quenching

→ difficult to use pulse height information→ rather use hit/miss info



Range Telescope Commissioning – One-Dimensional Position Resolution

vacuum-flange

Kapton window

protonbeam

Micromegas doublet

z

x

multi-anode-

PM

Scintillator Range Telescope

absorber

• Tandem beam not mono-energetic→ use additional collimators behindbending dipole

• mean range homogeneous over detector

• absorber edge visibletrack resolution ∼ 0.8 mm due tomultiple scattering

mean range vs position

hit position x [mm]18 20 22 24 26 28 30 32 34

mea

n pa

rtic

le r

ange

[mm

]

-1

0

1

2

3

no absorber1mm absorberdifference

ana_rt_tanaug14_0284_scaledph_perp_layphrange_rvsx_evsx_rangetelRangePos0rt0_ProjY0_has_supimp

Entries 6020Mean 48.05RMS 5.647Integral 280.7

hit position x [mm]0 10 20 30 40 50 60

mea

n pa

rtic

le r

ange

[mm

]

0

0.5

1

1.5

2

2.5

3

3.5

ana_rt_tanaug14_0284_scaledph_perp_layphrange_rvsx_evsx_rangetelRangePos0rt0_ProjY0_has_supimp

Entries 6020Mean 48.05RMS 5.647Integral 280.7



First Two-Dimensional Ion Radiography

MM doublets

phantombeam

RT

0

1

2

3

4

5

6

0

1

2

3

4

5

6

position y [mm]25 30 35 40 45 50 55 60

posi

tion

x [m

m]

25

30

35

40

45

50

55

60

0

1

2

3

4

5

6

• PLA step phantom visible

• ball point pen visible

• spatial resolution limited by multiple scattering

• resolution improvable by additional MM layers



First Two-Dimensional Ion Radiography

MM doublets

phantombeam

RT

0

1

2

3

4

5

6

0

1

2

3

4

5

6

position y [mm]16 18 20 22 24 26 28 30 32 34

posi

tion

x [m

m]

25

30

35

40

45

50

55

0

1

2

3

4

5

6

• PLA step phantom visible

• ball point pen visible

• spatial resolution limited by multiple scattering

• resolution improvable by additional MM layers


Summary

Summary

• floating strip Micromegas were optimized and work• discharges:

• behavior and topology understood• negligible influence on efficiency

• carbon ion and proton tracking at highest rates at HIT• separation of all particles at rates ≤10 MHz• spatial resolution better 180 µm at all rates ≤80 MHz• stable operation up to highest rates of 2 GHz

• 23 MeV proton tracking at Tandem/Garching• successful: fast Ne:CF4 gas mixture→ decrease signal duration by factor 3

• single plane track inclination reconstruction possible

• Micromegas + scintillator range telescope in 23 MeV proton beams• single particle range determination using hit/miss information• first 2d ion radiography successful• spatial resolution limited by multiple scattering

floating strip Micromegas:discharge tolerant, high-rate capable tracking detectors with good spatial resolution→ suitable for medical applications

Thank you!


Summary

Summary

• floating strip Micromegas were optimized and work• discharges:

• behavior and topology understood• negligible influence on efficiency

• carbon ion and proton tracking at highest rates at HIT• separation of all particles at rates ≤10 MHz• spatial resolution better 180 µm at all rates ≤80 MHz• stable operation up to highest rates of 2 GHz

• 23 MeV proton tracking at Tandem/Garching• successful: fast Ne:CF4 gas mixture→ decrease signal duration by factor 3

• single plane track inclination reconstruction possible

• Micromegas + scintillator range telescope in 23 MeV proton beams• single particle range determination using hit/miss information• first 2d ion radiography successful• spatial resolution limited by multiple scattering

floating strip Micromegas:discharge tolerant, high-rate capable tracking detectors with good spatial resolution→ suitable for medical applications

Thank you!


backup

backup – Discrete & Integrated Floating Strip Micromegas

discrete

ground

signal

mesh +500V

SMD resistor 22MΩ

copper strips

cathode -300V

6mm

128µm

SMD capacitor 15pF

• exchangable Rs and Cs→ optimization possible

• more complicated assembly→ soldering ×2 for each strip

• space requirements due to HV sustainingcomponents→ strip pitch limited to 0.5 mm

integrated

copper strips

mesh+HV

resistor 10MΩcopper strips

cathode -HV

signal

128μm

6mm

• anode strips: connected to HV viaprintable paste resistors

• readout strips: second layer of copperstripscapacitive coupling through the board,intrinsically HV sustaining


backup

backup – Track Inclination Reconstruction Systematics: LTSpice-Simulation

mesh strip

mesh strip

mesh strip

Casas

Casas

CmsCcoupl

Rhv

HV

Cslsl

Cslsl

Cslg

Chvgdetector HV

Chvcoupl 1MΩ

Ccr Ccc

signal

time

current

cathode

mesh

6 7 8 9 10 115

• use LTSpice to simulate 16 neighboring strips, readout via charge-sens.-preamps

• consider mesh-anode strip, anode strip-ground,anode strip-anode strip, coupling, stripline-striplineand stripline-ground capacitance, no free parameter

• inject time dependent current on anode strips →study signals on all other strips

time [s]0 0.2 0.4 0.6 0.8 1

-610×

char

ge [a

rb. u

nits

]

0

0.05

0.1

0.15

0.2strip 5strip 6strip 7strip 8strip 9strip 10strip 11


backup

backup – Hough Transform Based Track Building

x

z

x1

z1

point in position space (zi,xi) line in Hough space a = zib + xi

a

b

point in Hough space (bj,aj)line in position space x = -bjz + aj

x2

x3

x4

z2 z3

a1

b1

track with slope -b1 & intersect a1

• for improved stability: use Hesse normal form as transform function

• up to seven valid tracks reconstructed per event


Documents

High-Rate Capable Micromegas Detectors for ... - LMU Munich€¦ · true, simulation • reconstruction of all particles up to 10MHz = 7MHz/cm2 track nding algorithm track inclination