Characterisation and Application of Photon Counting X-ray Detector Systems Disputation seminar

1
Characterisation and Application ofPhoton
Counting X-ray Detector SystemsDisputation
seminar
2
Disposition

• Introduction
• Motivation for research and development of X-ray
  imaging
• Short description of the Medipix project
• Applications
• Dose reduction in medical imaging
• Material recognition
• Characterisation of the Medipix system
• Charge sharing
• Conclusions

3
Section 1 Introduction

• Basics on X-ray detectors
• X-ray detectors are available on the market, why
  do any research?
• What is photon counting?

4
X-rays

• Discovered in 1895 by W. K. Röntgen
• Generated by radioactive decay
• Medical images for surgery
• Cancer therapy
• High doses
• Today the entire population is affected by X-ray
  imaging

X-ray image from Siemens
5
Negative effects of radiation

• Ionizing radiation induces cancer
• No lower limit found
• Reduction of the X-ray dose
• Reduction of the cancer frequency
• Reduction of the costs for society
• For the individual
• The risk is small compared to other cancer
  inducing factors
• Attend X-ray examinations recommended by the
  medical expertise

6
Example Mammography

• Examination on regular basis for all females
• New tumours are small and easy to treat
• Argument for short interval between examinations
• Each examination increases the lifetime dose and
  the statistical risk for cancer development
• Argument for long interval between examinations
• A compromise between risk and benefit has to be
  made
• With improved detectors the dose at each
  examination can be reduced

Mammography device from Sectra AB
7
Detector improvement

• With improved detectors the dose at each
  examination can be reduced
• The examination interval can be decreased with
  remained lifetime dose
• More cancer tumours will be discovered at an
  early stage
• More cancers will be successfully treated

Lives will be saved!
8
Readout principles

• Photons generates a charge cloud in the
  semiconductor
• Charge integrating
• Intensity equals a sum of charge
• Photon counting
• The intensity equals the number of photons
• The lowest energies must be discriminated,
  otherwise thermal noise is counted as photons
• The energy or colour of each photon can be
  measured
• Photon counting makes colour X-ray imaging
  possible

9
Illustration of photon counting
Commercials of MicroDose from Mamea imaging AB
and Spectra Imtec AB
10
Section 2 The Medipix project

• A pixellated photon counting readout chip
• One readout circuit per pixel
• Requires deep submicron CMOS processes
• Detector matrix bump bonded to the readout chip
• Detectors of silicon, CdTe and GaAs

Illustration from http//medipix.web.cern.ch/MEDIP
IX/
11
Collaboration

• The project is directed from the Cern
  microelectronics group
• 16 European institutes are participating

Institut de Física d'Altes Energies IFAE
Barcelona University of Cagliari Commissariat à
l'Energie  Atomique CEA European Organization for
Nuclear Research CERN Czech Academy of
Sciences Czech Technical University in Prague
(CTU) Friedrich-Alexander- Universität
Erlangen-Nürnberg (FAU) European Synchrotron
Radiation Facility ESRF Albert-Ludwigs-
Universität Freiburg-i.B. University of
Glasgow Medical Research Council MRC Mid-Sweden
University (Mitthögskolan) MSU Università di
Napoli Federico II National Institute for Nuclear
and High-Energy Physics NIKHEF Università di Pisa
Mittuniversitetet
Map with collaborators logotypes
12
Medipix 1

• 1 µm SACMOS technology
• 170 µm square pixels
• 64x64 pixels
• 15 bit counters
• Low energy threshold
• 3 bits individual threshold adjustment
• Operated by standard PC connected to an interface
  circuit

Medipix1 system
13
Medipix 2

• Smallest pixel size for now
• 55 µm square pixels
• 256x256 pixels (1,4x1,6 cm)
• Dead area minimized on three sides
• Chipboards with 2x4 chips exists
• Operated by a standard PC

Medipix2 mounted for dental imaging
14
Medipix 2

• 0.25 µm CMOS technology
• 13 bit counters
• Upper and lower threshold
• Each with 3 bits threshold adjustment
• Individual leakage current compensation (GaAs)
• Positive and negative charge signal (CdTe)

Description of the Medipix2 readout circuit for
each pixel
15
Section 3 Applications

• Dose reduction in dental imaging
• Material recognition

16
Inverval or full spectra

• The relative contrast can be improved by applying
  an energy interval in dental imaging

Relative contrasts 0.70 0.59
26 - 30 keV
4 - 70 keV
17
Tooth image for varying energy
18
Colour image of the tooth

• Colour X-ray image from RGB coding of three images

19
(No Transcript)
20
Material recognition

• Possible to distinguish between Si and Al
  although the full spectrum absorption is equal

21
Section 4 Characterisation

• Description of charge sharing
• Simulation of charge sharing
• Measurements with narrow monochrome source
• Slit measurements

22
Charge sharing

• Crosstalk between pixels

Colour X-ray image of a slit achieved with a
Medipix2 Si-detector.
23
Charge sharing

• Physical components of charge sharing
• Beam geometry and scattering
• Quantisation error
• Absorption width
• X-ray fluorescence
• Charge drift
• Back scattering
• High energy photons can be divided into several
  low energy counts (Red colour in image)

Colour X-ray image of a slit achieved with a
Medipix2 Si-detector.
24
Charge drift

• Silicon
• about 3 absorption in a 300 µm detector (40
  keV)

• CdTe
• almost 100 point absorption
• Strong X-ray flourescence

25
Flourescence

• Flourescence is a problem for CdTe detectors
• Low energies has to be discriminated, to achieve
  reasonable spatial resolution

Colour X-ray image of a slit achieved with a
Medipix2 CdTe-detector.
Colour X-ray image of a slit achieved with a
Medipix2 Si-detector.
26
Simulation of charge sharing

• Charge sharing highly distorts the measured
  spectrum (Si)
• Overdepletion supresses charge sharing slightly

27
ESRF measurements
The European Synchrotron Radiation Facility

• Narrow beam 10x10 µm
• Monochrome energy40 keV

28
CdTe point spread function

• The 10 µm wide beam is centered on a pixel
• For low energies signal is measured 165 µm away
• Flourescence

29
CdTe spectrum

• Spectrum from the pixel where the 10 µm wide beam
  is centered
• Threshold window 2 keV
• Low energy tail
• Some photons deposits a fraction of their energy
  outside the pixel

30
CdTe neighbour pixels spectra

• Neighbour pixels
• Charge sharing behaviour
• Far neighbour
• Tenfold exposure time
• Distrurbances at 24 keV and 28 keV

31
Silicon spectrum

• Cumulative spectrum on a300 µm thick detector

32
300 micron, Si, 40keV, 170 e- noise, 10 micron
std in absorption profile
Simulation versus measurements
33
700 µm thick silicon detector

• Alignment becomes important

34
Conclusions

• Photon counting X-ray systems can lead to
  significant dose reduction (paper IV)
• With the next version of Medipix the technology
  is probably mature enough to be transfered to
  product developement
• Colour imaging can be used to discern different
  materials in an object (paper III)
• Energy dependence in image correction methods
  needs to be considered (paper II)

35
Conclusions

• Charge sharing degenerates the spectral
  information
• Charge sharing corrections can be implemented
  into the readout electronics
• The 3D detector structure supresses charge
  sharing (paper I)
• CdTe and GaAs detectors are less mature than
  Silicon
• Flourescence becomes a problem
• For 1 mm thickness the charge cloud is in the
  same size as the 55 µm pixel

36
Acknowledgements

• Thanks to
• My supervisors doc. Christer Fröjdh and prof.
  Hans-Erik Nilsson
• My colleagues at the electronics design
  department
• My colleagues in the Medipix collaboration
• The Mid-Sweden University, the KK-foundation and
  the European Commission are greately acknowledged
  for their financial support
• Thanks to my family Monica, Johan and William

37
(No Transcript)