PowerPoint Presentation - IWORID 2002

							 Fixed Pattern Signal fluctuations
       in Si pixel detectors
                 L. Tlustos, D. Davidson, M. Campbell,
                           E. Heijne, B. Mikulec




September 2002             Lukas Tlustos, IWORID2002
                      Outline


• Flatfield studies with Medipix1 chip and
  underdepleted and depleted standard high
  resistivity Si p+ on n sensors
• “Charge sharing”




September 2002   Lukas Tlustos, IWORID2002
                   Flatfield Correction
• Additive correction
        imgcorr  img  imgdark
        imgdark  0 ...due to discriminator in photon counting


• Multiplicative correction
      imgcorr  img  corr
                       mean counts
      corr (i, j ) 
                        flood field

September 2002               Lukas Tlustos, IWORID2002
                         Motivation
• Signal to noise ratio SNR is a                            SN
                                                     SNR  
  crucial parameter for image                               N
  quality and resolution




                                    SNR
• Poisson statistics of the
  incident beam marks the upper
  limit of the achievable SNR
• SNR of Medipix1 approaches
  Poisson limit when flatfield                                   theory
  correction is applied                                          corrected
                                                                 uncorrected
• Trying to understand the
  sources of inhomogeneities in
  the raw data in order to
  improve image quality
                                                      Normalized acquisition time
   September 2002        Lukas Tlustos, IWORID2002
                 Flatfield Measurements
 Large number of flatfield acquisitions (150-1000)
      o Count rate ~10kHz/pixel
      o accumulated total counts ~ 2.5 106 counts/pixel
 with
      o 3 Medipix1 assemblies
      o Detector bias ranging from 4V-128V
 Using
      o Mo tube, 30um Mo filter, 20keV, 28kV
      o W tube, 125um Al filter, 1cm PMMA, 20kV, 30keV


September 2002         Lukas Tlustos, IWORID2002
                 Flatfield Measurements




           Spectra calculated following
                     J M Boone, A E Chavez; Med Phys 23, 1997
                     J M Boone, T R Fewell, R J Jennings; Med Phys 24, 1997
September 2002               Lukas Tlustos, IWORID2002
                 Number of counts - Vbias

 • Plateau of count rate
   is a measure for
   depletion voltage
 • Normalized number
   of counts increase
   with energy content of
   the beam → charge
   sharing



September 2002         Lukas Tlustos, IWORID2002
                 Results 10C3_G8




 •Detector bias voltage: 3,4,8,12,16,24,32,48,64,80,100,128V
 •Correction map calculated using 128V detector bias data

September 2002        Lukas Tlustos, IWORID2002
                 Results 10C3_G8

                                               Single detached pixel




September 2002     Lukas Tlustos, IWORID2002
                 Results 11B6_5J




   •Detector bias voltage: 3,4,8,12,16,24,32,48,64,80,100V
   •Correction map calculated using 100V detector bias data

September 2002        Lukas Tlustos, IWORID2002
                 Results 11B6_5J




September 2002     Lukas Tlustos, IWORID2002
          10C3_G8 map 4V corrected




September 2002   Lukas Tlustos, IWORID2002
          10C3_G8 map 8V corrected




September 2002   Lukas Tlustos, IWORID2002
        10C3_G8 map 12V corrected




September 2002   Lukas Tlustos, IWORID2002
        10C3_G8 map 16V corrected




September 2002   Lukas Tlustos, IWORID2002
        10C3_G8 map 24V corrected




September 2002   Lukas Tlustos, IWORID2002
        10C3_G8 map 32V corrected




September 2002   Lukas Tlustos, IWORID2002
September 2002   Lukas Tlustos, IWORID2002
         11B6_5J map 12V corrected




September 2002   Lukas Tlustos, IWORID2002
         11B6_5J map 16V corrected




September 2002   Lukas Tlustos, IWORID2002
         11B6_5J map 24V corrected




September 2002   Lukas Tlustos, IWORID2002
         11B6_5J map 32V corrected




September 2002   Lukas Tlustos, IWORID2002
         11B6_5J map 48V corrected




September 2002   Lukas Tlustos, IWORID2002
         11B6_5J map 64V corrected




September 2002   Lukas Tlustos, IWORID2002
September 2002   Lukas Tlustos, IWORID2002
         10C3_G8 map profile 4-16V

           Raw data                               3x3 mean filter




September 2002        Lukas Tlustos, IWORID2002
         10C3_G8 map profile 4-32V

           Raw data                               3x3 mean filter




September 2002        Lukas Tlustos, IWORID2002
      10C3_G8 map profile 32-128V

           Raw data                               3x3 mean filter




September 2002        Lukas Tlustos, IWORID2002
          11B6_5J map profile 4-32V

           Raw data                               3x3 mean filter




September 2002        Lukas Tlustos, IWORID2002
       11B6_5J map profile 32-100V

           Raw data                               3x3 mean filter




September 2002        Lukas Tlustos, IWORID2002
September 2002   Lukas Tlustos, IWORID2002
                  Conclusions (1)
• Low detector bias: wave pattern clearly visible,
  ascribed to bulk doping inhomogeneities
  causing fluctuations of the width of the
  depletion layer
• Magnitude of fluctuations decreases with
  detector bias
• But: small variations remain even at high bias
  voltages, may be due to local variations of the
  pixel geometry
 September 2002     Lukas Tlustos, IWORID2002
                 Detached Pixel 10C3_G8




September 2002         Lukas Tlustos, IWORID2002
                                       Detached Pixel 10C3_G8
                               count disconnected pixel - 10C3_G8 - Mo 20kV
                          6

                                                                                • Count decreases with
                          5
                                                                                  detector bias
Normalized count[%] [%]




                          4
                                                                                • Generated charge
   normalized count




                          3                                                       collected by adjacent
                          2
                                                                                  pixels

                          1



                          0
                           0   20      40      60           80    100     120     140
                                                 V bias [V]
                                               Vbias [V]

                      September 2002                         Lukas Tlustos, IWORID2002
                          Detached Pixel 10C3_G8
                                                           • Sum of map weights
                                                             of the 4 direct
                                                             neighboring pixels < 5
                                                           • Field not strong
weight




                                                             enough to redistribute
                                                             all the chare generated


                           Vbias [V]




         September 2002                Lukas Tlustos, IWORID2002
                  Conclusions
• Fixed pattern fluctuations are consistent with bulk
  doping inhomogeneities originating in Float Zone
  technique of crystal growth
• Inhomogeneities can influence the charge
  collection properties of sensor
• Fixed pattern fluctuations remain even in
  overdepletion and affect the obtained resolution in
  imaging applications → correction is necessary
• Implications also to optimization of resolution in
  particle tracking
• Effects could also be used to study material
  properties
September 2002     Lukas Tlustos, IWORID2002