Thorsten Graeve - Pusanbml.pusan.ac.kr/Publication/PDF/IntCon/042_IntConf_sorma2008_ysm.pdfThorsten Graeve Rad-icon Imaging ... RadEyeTM Phosphor screen Rad-icon RadEye1TM ... 8.8

SORMA WEST 2008

Seung Man Yun, Chang Hwy Lim, Min Kook Cho, Ho Kyung Kim*

School of Mechanical Engineering, Pusan National University, Busan 609-735, Republic of Korea

Thorsten GraeveRad-icon Imaging Corp., Belick street, Santa Clara, CA 95045-2404, USA

SORMA WEST, June 2-5, 2008, Berkeley, CA, USA

Motivation & Objective

Experimental

Image data analysis

Modeling

Monte Carlo simulation

Results

Conclusion

Contents

2


Motivation & Objective Investigated noise-power spectrum of an indirect-conversion detector

– A thin phosphor screen + CMOS photodiode array

NPS degradation due to the additional white spectral noise

– Directly absorbed x-ray photons unattenuated from the overlying scintillator???

Purpose of this study

– Measurements of signal and noise characteristics induced by direct x-ray absorption in photodiode array

3

Expected: NPS(f)

Measured

White spectral noiseExpected: NPS(f)


Experimental (1/3) Detector preparation

– Photodiode arrays manufactured by CMOS process (Rad-icon Imaging Corp., USA)

4

RadEyeTM

Phosphor screenRad-icon RadEye1TM

Array format 512 × 1024 pixels

Pixel pitch 48 mm

Field of view 25 × 50 mm2

Dynamic range 85 dB (>14 bits)

ADC bit-depth 12 bits

Pixel fill factor 0.87

Saturation 2,800,000 electrons

Dark current (each pixel) < 104 electrons/sec

Read Noise (at 1 fps) < 200 electrons


Experimental (2/3) Two phosphor screen

– Terbium-doped gadolinium oxysulfide phosphor screen (Eastman Kodak Co., USA)

5

Min-R2000TM

Composition Gd2O2S:Tb

Coverage 33.91 mg/cm2

Thickness 84 mm

Density 4.04 g/cm3

Lanex MediumTM

Composition Gd2O2S:Tb

Coverage 59.20 mg/cm2

Thickness 160 mm

Density 3.70 g/cm3

Thickness

Thickness

84 mm

160 mm

Input dose


X-ray quality in IEC standard

Technique W/Al RQA5

IEC filtration

mmAl

2.5 21

IEC HVL 0.83 6.8

Measured HVL 0.85 6.78

IEC tube voltagekVp

28 70

Adjusted voltage 28 69

Calculated q#/mm2/mGy

6,575 30,174

Measured q 6,527 29,707

Percent difference % 0.73 1.55

In order to isolate the signal and noise due to the direct x-rays from those due to the optical photons

– Insert a light-absorbing blackout material between a phosphor screen and photodiode array

Experimental (3/3)

6

IEC 61267, 62220-1-2 reports


Image data analysis Signals due to absorption of optical photons or direct x-rays are uncorrelated each other

– Mean and variance of an x-ray white image Iwhite

– Mean and variance of a direct x-ray image Idirect

– Mean and variance of a dark image Ioffset

– Magnitude of each signal and noise

7

darkdirectoptwhite SSSIMean

blackwhiteopt IMeanIMeanS

offsetblackdirect IMeanIMeanS

2222

adddarkdirectoptwhiteIVar

blackwhiteopt IVarIVar 2

offsetblackdirect IVarIVar 2

darkdirectdirect SSIMean 222

adddarkdirectdirectIVar

darkoffset SIMean 22

adddarkdirectIVar


Interaction mechanism to produce electronic signals in an indirect-conversion detect

Modeling (1/3)

8

Optical Direct

Gd2O2S:Tb screen

Photodiode array

e-

e-

x-ray photons

light photons

SiO2 passivationp-n junction

Si substrate


Modeling signal and noise characteristics using the cascaded linear-systems theory

Modeling (2/3)

9

Cascaded model due to the optical photon

Cascaded model due to the direct x-ray absorption

H. K. Kim, "Generalized cascaded model to assess noise transfer in scintillator-based x-ray imaging detectors," Appl. Phys. Lett., vol. 89, iss. 23, pp. 233504-1–233504-3, 2006.


Cascaded model due to the optical photon

Cascaded model due to the direct x-ray absorption

Modeling (3/3)

10

pdscnscnopt aqn γβα2

pdpdscndirect aqn βα)α1(2

1

βγ1γβαδ 22

scn

scnpdpdscnscnopt

Iaq

pd

pdpdscn

directI

aq 22

2β)αα(1

δ

Notations

n mean number of electrons

d2 Variance

q average incident x-ray fluence

a photodiode aperture size

a quantum absorption efficiency

b mean gain of secondary quanta

gpd quantum efficiency of photodiode

I Swank or statistical factor

scn in phosphor screen

pd in photodiode array


Monte Carlo simulation In order to avoid the contamination of the other interactions such as Compton scattered X-rays

– We removed the phosphor layer in Monte Carlo geometry when simulate the direct absorption X-rays

11


Results (1/6) Measured signal mean-variance curve due to the direct x-ray photon absorption

12


Results (2/6) Measured signal mean-variance curve due to the optical photon absorption

13


Results (3/6) Comparing with calculation and measured results (RQA5, Min-R2000TM screen)

14


Results (4/6) This discrepancy comes from Compton scattered x-ray s in p-n junction of diffusion of electrons

15

phosphor screen

SiO2 passivation

p-n junction

Si substrate

SiO2 passivation

p-n junction

Si substrate

SiO2 passivation

p-n junction

Spectrum F(E)

F(E)emLscn

F(E)emLscn


Results (5/6) Energy dependency noise power spectrum (kerma normalized)

16


Results (6/6) Screen dependency noise power spectrum (kerma normalized)

17


Conclusion

From the measurement results

– Although the magnitude of direct signal is small, but its variance is considerably large

– It gives rise to the reduction of signal-to-noise ratio

From the theoretical estimation

– The effect of direct x-ray absorption is negligible in radiographic imaging condition

Main cause of the degradation in noise characteristics and thus the reduction of SNR is not by the direct x-rays but by the other sources

– such as the Compton scattered x-rays from a phosphor screen

It should be noted that the magnitude of signal and noise due to the direct x-ray absorption is considerable in mammographic imaging condition (PE is dominant process in this condition)

18


Fin.

19


Results (1/6)

20

Measured signal and noise magnitudes with respect to the optical and direct x-ray photons

X-rayquality

Phosphorscreen

Kerma(mGy)

Signal (ADU) Variance (ADU2)

optical direct optical direct

W/Al Min-R20008.8 58.328 ± 0.448 5.495 ± 0.280 9.772 ± 2.198 23.580 ± 1.248

11.2 77.116 ± 0.502 5.649 ± 0.195 11.474 ± 1.660 30.391 ± 1.102

Slope (ADU/mGy or ADU2/mGy) 6.728 ± 0.190 0.565 ± 0.087 1.061 ± 0.057 2.686 ± 0.024

RQA5 Min-R2000

16.9 323.851 ± 8.554 45.081 ± 3.594 144.770 ± 20.694 312.464 ± 19.120

20.9 428.108 ± 5.038 57.548 ± 2.950 207.951 ± 9.270 400.229 ± 10.165

26.5 561.390 ± 15.065 76.220 ± 1.186 274.729 ± 22.456 550.722 ± 18.609

32.4 737.061 ± 8.939 95.273 ± 1.312 433.184 ± 10.615 660.602 ± 14.136


RQA5 Medium

16.9 813.002 ± 9.988 43.005 ± 0.356 80.226 ± 10.357 288.343 ± 4.900

20.9 1063.376 ± 8.598 54.812 ± 0.412 103.279 ± 13.148 372.294 ± 4.920

26.5 1378.908 ± 19.373 70.117 ± 0.681 146.894 ± 6.326 479.240 ± 11.816

32.4 1769.243 ± 13.073 87.467 ± 1.179 153.691 ± 22.018 639.821 ± 19.715



Light absorbing blackout paper

– For light trapping, baffling, beam stopping in visible ray wavelength band

100 mm thickness, Velvet like black material (Flock paper #40, Edmond Optics, USA)

Experimental

21

Visible light transparency test (Hitachi U2010, Japan)

Gd2O2S:TbEmission peak

Thorsten Graeve - Pusanbml.pusan.ac.kr/Publication/PDF/IntCon/042_IntConf_sorma2008_ysm.pdfThorsten Graeve Rad-icon Imaging ... RadEyeTM Phosphor screen Rad-icon RadEye1TM ... 8.8

Documents