International Journal of Medical Physics, Clinical Engineering and Radiation Oncology, 2013, 2, 98-109 Published Online August 2013 (
Performance of Large Area Thin-Film CdTe
Detector in Diagnostic X-Ray Imaging
Diana Shvydka1, Xiance Jin1,2, E. Ishmael Parsai1
1Department of Radiation Oncology, University of Toledo, Toledo, USA
2Department of Radiotherapy and Chemotherapy, Wenzhou Medical University, Wenzhou, China
Received April 5, 2013; revised May 20, 2013; accepted June 22, 2013
Copyright © 2013 Diana Shvydka et al. This is an open access article distributed under the Creative Commons Attribution License,
which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Significant advancement in thin-film cadmium telluride (CdTe) deposition techniques in recent years has made this
material attractive for the development of low-cost large area detector. Here we evaluate the intrinsic performance of
the detector for a range of energies relevant to diagnostic imaging applications, such as fluoroscopy. The input x-ray
spectra for a set of tube potentials ranging from 70 to 140 kVp were computed with the tungsten anode spectral model
using interpolating polynomials (TASMIP) based on the measured output of our diagnostic x-ray simulator. Fre-
quency-dependent detector performance analysis was conducted through Monte Carlo simulations of energy deposition
within the detector. Intrinsic modulation transfer functions (MTF), noise power spectra (NPS), and detective quantum
efficiencies (DQE) were computed for a set of CdTe detectors of varying thickness, from 100 to 1000 μm. MTF behav-
ior at higher frequencies was affected by thickness and input energy, NPS increased with film thickness and energy, and
the resultant DQE(f) decreased with increasing the input energy, but increased with the thickness of the detector. We
found that the optimal thickness of CdTe under diagnostic x-ray beam is in the range of 300 to 600 μm. Physical prop-
erties of CdTe, such as the high atomic number and density, used in direct detection configuration, together with the
recently established thin-film manufacturing techniques makes this technology a promising photoconductor for large-
area diagnostic flat panel imaging.
Keywords: Monte Carlo; CdTe Detector; Photovoltaic; Thin Film; TASMIP; DQE; X-Ray Imaging
1. Introduction
Introduction of digital radiography systems and succes-
sive use of flat panel detectors revolutionized the field of
diagnostic imaging. Wide dynamic range, high image
quality, real-time image acquisition and processing, pre-
cise image recording, and ease of remote access are
among the most prominent improvements. One of the
decisive factors contributing to further advancements re-
mains the continuous development of different x-ray de-
tecting materials, from traditional phosphor screens in
combination with secondary photodetectors for indirect
detection to use of thin-film photoconductors in direct
detection systems. The latter approach offers a twofold
benefit: simpler device structure resulting in lower manu-
facturing cost, and a high potential of providing images
of superior contrast and sharpness due to inherently low
signal spreading within the detector.
In the direct detection approach x-rays are absorbed by
a photoconductor layer and converted to electron-hole
pairs, which are then collected as electric charges on
storage capacitors. Up to now amorphous selenium (a-Se)
is the only photoconductor developed into direct detec-
tion type commercial medical imagers, for both general
radiography and mammography applications [1-4]. De-
tectors are based on a-Se offer superior spatial resolution
due to the simple conversion process. However, low x-
ray absorption and high effective ionization energy (~50
eV) result in inadequate sensitivity, especially important
for low exposure levels of fluoroscopic mode [5].
To overcome this problem, materials of high atomic
numbers and densities, such as mercuric iodide (HgI2),
lead iodide (PbI2) [6-11], lead oxide (PbO) [12], thallium
bromide (TlBr) [13], and cadmium telluride/cadmium
zinc telluride (CdTe/CdZnTe) [14-17], have been sug-
gested to replace a-Se. All of these materials possess
effective ionization energy about 10 times lower than
that of a-Se, substantially large band gaps, necessary for
minimization of leakage currents at room temperature,
and high mobility-lifetime product, providing effective
opyright © 2013 SciRes. IJMPCERO
charge collection [9,18,19].
Due to large area requirements imposed on practical
medical detectors all of these materials are investigated
in polycrystalline (thin-film) rather than single crystal
form. This quest entails development of proper tech-
niques for thin-film deposition in order for the material to
be commercially viable. Since a-Se has been studied the
longest, by now the capabilities to manufacture high-
quality films as thick as 1 mm have been proven. For
other photoconductors this is still a subject of an ongoing
research, rendering the use of some of the materials
rather challenging. For example, strong temperature de-
pendence of TlBr’s conductivity makes it difficult to
operate at room temperature [13]; poor response time and
some limitations on film thickness are detrimental to
PbI2-based device performance [10]; chemical stability,
nonuniform sensitivity, low carrier mobilities, and high
levels of dark currents are still somewhat problematic in
HgI2 devices [10,11], although the last material appears
to be one the most promising.
Recent developments in large area photovoltaic appli-
cations of CdTe have moved this photoconductor to the
frontiers of thin-film manufacturing [20]. While the typi-
cal thickness of a solar cell is under 10 μm, the device
deposition methodologies and post-deposition treatments
for grain boundary passivation are essentially the same,
and are successfully implemented in the fabrication of x-
ray detectors up to 600 μm thick [15,16].
Even though the average atomic number of CdTe is
lower than that of HgI2, their absorption properties are
very similar over a wide range of kV x-ray energies.
Comparison of absorption coefficients [21] in Figure 1
demonstrates that for energies up to the k-edge of mer-
cury (83 keV) both materials are equally superior to a-Se.
Coincidently, even for the spectra corresponding to high-
er kV potentials (up to 140 kVp) most of x-rays have
energies in that range. Proven outstanding radiation hard-
ness of CdTe [22-26] makes it an ideal candidate for im-
aging application.
We have previously explored the use of thin-film
CdTe detector in radiation therapy imaging [27]. Up to
now no systematic theoretical analysis of the thin film
CdTe detector for diagnostic x-ray imaging has been
performed. Here we investigate the theoretical perform-
ance of thin-film CdTe of thickness from 100 to 1000 μm
under a range of energies relevant to high-energy diag-
nostic imaging applications, from 70 kVp to 140 kVp,
using Monte Carlo (MC) radiation transport package
MCNP5 [28]. In this approach only energy deposition is
evaluated, with no assumptions made about the elec-
tronic properties of CdTe. Therefore our calculations
provide the upper-limit estimate for the material per-
formance in an idealized case of complete charge collec-
tion. Our goal is to find a range of optimal thickness of
Figure 1. Comparison of absorption coefficients for Se,
CdTe, and HgI2.
thin film CdTe which maximizes the detective quantum
efficiency DQE(f) for different x ray energies and com-
pare CdTe detector performance to that of other photo-
2. Methods and Materials
2.1. Input Spectra
The input x-ray spectra for the detector performance
study were obtained for a set of kV potentials ranging
from 70 to 140 kVp. For this purpose we employed a
tungsten anode spectral mode using interpolating poly-
nomials (TASMIP) [29] technique, that allows comput-
ing x-ray spectra based on the measured output of our
Ximatron clinical simulator unit (Varian, Palo Alto, CA).
The TASMIP is not a semi-empirical algorithm and uses
no physical assumptions regarding x-ray production, but
rather interpolates measured constant potential x-ray
spectra published by Fewell et al. [30]. TASMIP has
been shown to accurately reproduce both the kV-de-
pendent spectral shape and output fluence for x-ray ma-
chines employing tungsten target.
The x-ray output of the simulator was measured in the
units of mR/mAs at a distance of 100 cm from the focal
spot using an Unfors Xi External Detector. Output was
measured at 200 mA, 50 ms (10 mAs) with 0, 1, 2, 3, 4,
and 5 mm added aluminum filtration for a set of kVp
values in the range from 40 to 125, as shown in Figure 2.
The points show the actual measured values, while the
solid lines represent four-term (third order) polynomial
fits for each filtration. The attenuation curves in Figure 3
were calculated based on these fitted outputs.
Unfiltered tungsten spectra from Fewell et al. [30]
were tabulated for 70, 80, 90, 100, 120, 130, and 140
kVp and were linearly interpolated to obtain spectra with
1 keV interval. These tabulated data corresponding to the
spectra labeled EI1 though EI8 on pages 43 - 45 of [28].
The attenuation curves of these unfiltered Fewell’s spec-
tra were calculated based on the attenuation coefficients
of aluminum from Physics Laboratory of National Insti-
Copyright © 2013 SciRes. IJMPCERO
Figure 2. The simulator output for different values of added
aluminum filtration. Points show the measured values, and
the solid lines represent third-order polynomial fits for each
Figure 3. Matched attenuation curves of simulator x-ray
and Fewell’s spectra for different kVp based on the least-
square approach. The points are attenuation values calcu-
lated from the modified Fewell’s spectra, and the corre-
sponding solid lines represent attenuation profiles calcu-
lated from our simulator output.
tute of Standard and Technology (NIST). To compensate
for probable differences in the x-ray tube housing at-
tenuation values at each kVp, additional thicknesses of
aluminum were needed to compensate for the inherent
filtration. The least square approach was used to mini-
mize the percent difference in attenuation values between
our simulator’s and the Fewell’s spectra, as shown in
Figure 3. Here the points are attenuation values calcu-
lated from the modified Fewell’s spectra, and the corre-
sponding solid lines represent attenuation profiles calcu-
lated from our simulator output data. Additional alumi-
num thickness needed to match the Fewell’s spectra to
the attenuation levels of our simulator are given in Table
Once the Fewell’s spectral shapes were slightly hard-
ened to best fit our simulator attenuation values, the
number of photons for each spectrum was normalized to
the corresponding output of the simulator with no added
filtration. Based on these spectra modified with addi-
Table 1. Aluminum filtration added to match the Fewell’s
Potential (kVp)708090 100 110 120 130140
Added aluminum
thicknesses, mm0.557 0.700 0.802 1.037 0.854 0.818 0.727 0.587
tional aluminum thicknesses, we were able to calculate
the polynomial interpolating coefficients using the fol-
lowing equation
01 23
kVp kVpkVpEaEaE aEaE 
with these interpolating coefficients we were able to
compute the simulator x-ray spectra at any kVp value.
We generated simulator spectra shown in Figure 4 for 70,
80, 90, 100, 110, 120, 130, and 140 kVp values. The two
peaks present in higher kVp spectra correspond to Kα
and Kβ edges of tungsten. The average percentage devia-
tions between the modified spectra and the final interpo-
lated spectra were between 0.1% and 1.9%. These spec-
tra were utilized for source specification in our Monte
Carlo simulations.
2.2. Detector Simulation
2.2.1. Exit Spectra
During an imaging process the source beam spectrum is
modified before reaching a detector by first going through
a patient. To properly account for this effect we obtained
the exit spectra, which are the energy spectra of the pri-
mary photons transmitted through a 20 cm thick water
phantom representing a patient. Figure 5(a) shows the
MC simulation geometry used, where the source of pho-
tons having the above-calculated spectra is placed at a
distance of 100 cm from the phantom. The exit spectra
were acquired with MCNP tally F1 at the surface of thin-
film CdTe detector, located 20 cm below the water phan-
tom according to the clinical application with a source to
film distance (SFD) of 140 cm. Figures 6(a) and (b)
show the primary photon spectra before and after the
water phantom for energies of 80 kVp and 140 kVp, re-
spectively. The relative photon intensity in both graphs
was calculated based on the number of photons reaching
the surface of CdTe detector per source particle (number
of histories). Exit spectra demonstrate a two-fold effect
of the presence of the water phantom, namely reduction
in beam intensity due to absorption and scatter, and a
slight beam hardening.
Only primary photon beams were used to evaluate the
performance characteristics of the detector. In all simula-
tions we used CdTe thin-film density ρ = 5.86 g/cm3. The
electron cut-off energy (ECUT) was chosen so that the
electron range at ECUT is less than 1/3 of the smallest
dimension in the dose scoring region, 0.02 MeV for 20
Copyright © 2013 SciRes. IJMPCERO
Figure 4. Simulator spectra from 70 kVp to 140 kVp com-
puted by polynomial interpolation of modified Fewell’s
Figure 5. (a) Schematic illustration of the geometry setup
used in Monte Carlo simulations of energy deposition; (b)
Geometry setup for line spread function simulation.
micron scoring slit [31,32]. The cut-off energy for pho-
tons was set to 0.01 MeV, with coherent, photonuclear
and Doppler interactions turned off.
2.2.2. Pre-Sampling Modulation Transfer Function
Modulation transfer function is an important characteris-
tic of a detector, ultimately defining its resolution. It rep-
resents the capability of the detector to transfer signal
with minimal spreading. In our MC simulations pre-
sampling MTF was obtained in the geometry shown in
Figure 5 for the thin-film CdTe detector of different
thickness, 100, 300, 600, and 1000 μm. X-ray beams
with exit spectra having potential energies of 70 to 140
kVp were set to fall normally at the detector surface
Figure 6. Primary photon spectra before and after a 20-cm
thick water phantom for (a) 80 kVp; (b) 140kVp.
through a narrow, 2 μm wide, slit, following a typical
setup for line spread function LSF(x) measurement, as
shown in Figure 5(b) [33]. For frequency dependent
MTF(f) simulation, the detector was divided into 512
strips on each side of the slit source with a width of
10-µm, which according to the Nyquist criterion gave a
cutoff frequency of 50 mm1. Energy deposition within
each strip was acquired with *F8 tally. The MTF(f) was
calculated by performing a fast Fourier transform (FFT)
of LSF(x) with Hanning window method. The total num-
ber of points of 1024 produced smooth MTF(f) curves
with enough sampling around zero frequency.
2.2.3. Noise Power Spectrum (NPS)
Noise transfer properties of the detector were evaluated
through simulations of the frequency-dependent noise
power spectra (NPS) or Wiener spectra. The mean square
departure of signal from its average value is the variance
and the analysis of this variance into frequency compo-
nents gives the noise power spectrum [34]. The simula-
tion geometry was similar to that used to obtain MTF. A
30 × 40 cm2 photon beam was set perpendicularly inci-
dent on the X-Y plane of the detector. One-dimensional
(1-D) NPS was analyzed by a synthesized slit technique
[34,35]. The energy absorbed per scoring bin, dx, was
recorded by *F8 tally in the non-overlapping slits of 512
bins per slit. Each slit had dimensions of 0.002 × 3 cm2,
providing a Nyquist frequency of 25 mm1 in x direction.
Copyright © 2013 SciRes. IJMPCERO
The absorbed energy distributions per bin were Fourier
transformed using a 1-D FFT to yield frequency de-
pendent power spectra. Total of 420 slits were averaged
to yield the simulated NPS in units of MeV2mm2, calcu-
lated as [36,37]:
here x0 and y0 are the x and y spacing of the discrete val-
ues respectively; Nx = 1 and Ny = 512 are the number of
elements in x and y dimensions, respectively; dx,y is the
energy deposited (in MeV) within the (x, y)-th element,
,, ,
yxy xy
dd d , and ,
d was the average en-
ergy deposited per scoring bin.
The simulations of noise power spectra are quite com-
puter demanding on computer CPU time. To ensure that
the statistical error per scoring bin is below 10%, we ran
at least 108 photon histories per detector configuration.
2.2.4. Detective Quantum Efficiency (DQE)
The detective quantum efficiency (DQE) has become the
best single descriptor of radiographic detector perform-
ance. By definition, DQE is the squared ratio of the sig-
nal-to-noise (SNR) at the detector output to that at the
detector input as a function of spatial frequency f. It gives
a measure of how efficiently the imaging system makes
use of the information content of a radiation beam. For
an ideal imaging system, the DQE is equal to 1, but in
reality it is degraded by different sources of noise associ-
ated with the system.
DQE at zero frequency, DQE(0), is related to energy
absorption properties only, it is the maximum DQE of
the detector that could be achieved. When taking x-ray
quantum detection noise into account, the DQE(f) can be
expressed as [38,39]:
 
where A0 is the mean signal value at the detector output
and 0
q is the incident x-ray fluence. In our MC simula-
tions, pulse height tally *F8 was used to record the en-
ergy deposition within each detector element, averaged
into the mean signal value A0. Based on simulated MTF
and NPS we found frequency-dependent detective quan-
tum efficiencies for a set of kVp values for CdTe detec-
tors of varying thickness. We note that DQE(f) obtained
for each CdTe film thickness is a pre-sampling or intrin-
sic characteristic of the detector.
2.2.5. Conversion Gain
The next part of analysis involves assessment of the gain
associated with conversion of the energy deposited by
x-rays into electron-hole pairs in CdTe. This stage is
termed amplification stage in the multi-stage (cascaded)
linear systems theory [40]. To evaluate this gain we use
the energy required for creating one electron-hole pair in
CdTe, WCdTe ~ 5 eV for polycrystalline material [41,42]
(this value is very close to 4.43 eV measured for crystal-
line CdTe [43]). The amount of energy deposited in
CdTe per incoming X-ray is defined by the mean signal
value A0, and therefore depends on CdTe thickness. The
maximum amplification gain g can be estimated as
and represents the upper limit estimate for the number of
electron-hole pairs generated in a certain volume of
3. Results and Discussion
3.1. MTF
The modeled pre-sampling MTF of thin film CdTe with
thicknesses from 100 to 1000 μm under 80 kVp beam is
shown in Figure 7(a). As expected, MTF of CdTe de-
creases with increasing the thickness, but the decrease in
MTF becomes moderate for films thicker than 300 μm.
At the spatial frequency of 5 mm1, MTF of 300 μm de-
creases about 6.1% compared to that of 100 μm, while
the MTF of 600 μm decreases only 1.4% compared to
that of 300 μm. Even less of a relative decrease, only
0.2% compared to that of 600 μm, is observed for 1000
μm thick film. General trend of MTF reduction with the
increasing detector thickness is the result of the increase
in both the amount of secondary particle’s scatter, and
the fraction of re-absorption of k-fluorescence x-rays.
The MTF of 600 μm thick CdTe under different ener-
gies from 70 kVp to 120 kVp is shown in Figure 7(b).
For the frequency range of about 3 to 15 mm1, the low-
est energy beam produces lowest MTF. This is related to
the larger portion of incoming photons having energies
just above the k-edges of both Cd (26.7 keV) and Te
(31.8 keV). The higher absorption (see Figure 1) results
in production of a larger number of fluorescence photons,
which can be re-absorbed up to 150 μm away from the
origin. At higher frequencies (see Figure 7(b), inset), the
MTF is more degraded at the higher beam energies due
to the increase in probability of Compton interactions, re-
sulting in scattered particles depositing their energy close
to the first interaction site. The effective path lengths of
recoil electrons increase with the increasing energy, re-
sulting in increased lateral spread within the detector.
These are consistent with the findings of previous re-
searchers [19,44,45].
3.2. NPS
Quantum noise is an unavoidable result of statistical
fluctuations in the number of x-rays interacting within
Copyright © 2013 SciRes. IJMPCERO
Figure 7. (a) MTF of CdTe with thicknesses of 100, 300, 600,
and 1000 μm under 80 kVp beam; (b) MTF of 600 μm CdTe
under energies from 70 kVp to 120 kVp.
the detector and statistical fluctuations in the number of
electrons produced as the result of each interaction. As
shown in Figure 8(a), noise power spectrum increases
with the thickness of the detector. This is well expected
since more photons are absorbed as the detector thick-
ness increases. Comparison of Figures 8(a) and (b) in-
dicates that for the same film thickness the noise power
also increases with increasing energy, which is probably
due to increase in energy deposited per interacting pho-
Another trend evident from Figure 8 is that noise is
strongly correlated, especially for thicker CdTe films. As
x-rays traverse the detector layer they produce ionizing
particles, which deposit energy in CdTe layer through a
large number of discrete interactions along the path of
each particle, leading to correlated quantum noise. We
fitted each NPS with a Lorentzian function
yy 2
 , where w is the full width at half
maximum, and y
0 and A are fitting parameters. The fits
are shown as solid lines in Figure 8. The correlation
length in mm, reflecting the smallest feasible pixel size
for the detector, was estimated as w1. Typical values
obtained for 300 to 1000 μm thick CdTe were in the
range of 0.1 mm, which is within the range of usual pixel
sizes employed in digital imagers [5].
Figure 8. Noise power spectra under (a) 80 kVp beam and
(b) 120 kVp beam.
3.3. DQE
Shown in Figure 9 are the results for DQE(f) of 100 to
1000 μm thick CdTe. These were calculated using Equa-
tion (3), based on MTF and NPS curves shown in Fig-
ures 7 and 8, where NPS dependences were fitted with
Lorenzian functions. DQE(f) improves with increasing
thin film CdTe thickness due to increase in number of
absorbed photons. However, a thicker detector also pro-
vides longer paths for the lateral spread of secondary
electrons and photons, resulting in a loss of spatial reso-
lution and increase of noise, as shown in Figure 7 and
Figure 8. As evident from Figure 9(a) there is little im-
provement in DQE(f) as thickness increases beyond 600
μm. Effect of both the thickness and beam energy on
DQE is shown in Figure 9(b): DQE(f) decreases for
higher kVp, mainly due to decrease in interaction prob-
ability of higher energy photons, as expected from lower
absorption coefficient (Figure 1).
To summarize the trends in energy absorption with
thickness and kVp we plot in Figure 10 DQE(0) values
obtained from DQE(f) analysis. Here we clearly observe
the increase in the absorption efficiency of the detector
with the increasing thickness, becoming very moderate
after 600 μm, especially for lower energies. For example,
under 80 kVp, DQE(0) of 300 μm CdTe is about 59.7%
more of that of 100 μm, however, the DQE(0) of 1000
μm is only about 7.4% more than that of 600 μm.
Copyright © 2013 SciRes. IJMPCERO
Figure 9. Computed DQE(f). (a) Under 80 kVp; (b) Under
120 kVp.
Figure 10. DQE(0) of CdTe detector vs the film thickness in
μmfor 80 and 120 kVp x-ray spectra.
3.4. Conversion Gain
Based on the energy deposition modeled with MC and
using Equation (4) we estimate the average number of
electron-hole pairs created in the CdTe layer per 1 cm2
area of the detector. The resultant dependences of gen-
eration rates on the CdTe thickness are shown in Figure
11, where a log-log scale is used for clarity. The de-
pendence is close to (1-exp (-αt)) up to the CdTe thick-
ness t of the order of the average x-ray penetration depth,
saturating for thicker films.
Here we can also consider the noise associated with
the gain variance that can be estimated [46] based on the
number of electron-hole pairs created N and Fano factor
Figure 11. Electron-hole pair generation per incoming pho-
ton over 1 cm2 area of the CdTe detector for different di-
agnostic x-ray spectra.
F (~0.1 for most crystalline semiconductors [47]) as
. The resulting relative gain variance is
for thickness of CdTe of 100 μm and lar-
3.5. Comparison with Other Photoconductors
We compare characteristics of CdTe with other photo-
conductors, namely a-Se and HgI2. The former is used in
the only commercially available direct detection type
system, while the latter represents a more promising of
the higher atomic number materials. Due to the presence
of absorption edges a mere comparison of absorption
coefficients in Figure 1 is not sufficient for adequate
understanding of the frequency-dependent detector per-
formance at different kVp values.
MTF evaluation shown in Figure 12 illustrates this
non-trivial behavior. As expected, the lowest absorption
resulting in the lowest interaction probability is solely
responsible for the minimal signal spreading and the
highest of all three materials MTF of a-Se. However,
even though HgI2 has the highest atomic number and
therefore the highest absorption, its MTF is still higher
than that of CdTe for 80 kVp x-ray spectrum. This is re-
lated to the additional signal spreading due to the longer
path length of fluorescence photons produced from x-
rays with energies just above the k-edges of both Cd and
Te [19]. At 120 kVp, though, where more photons with
energies above HgI2 k-edge are present in the incoming
beam, this is no longer the case: MTF of CdTe is higher
than mercuric iodide’s, as evident from Figure 12(b).
The final DQE(f) calculated for the three materials are
shown in Figure 13 for photon two spectra, 80 and 120
kVp. The latter spectrum produces the highest DQE for
CdTe for frequencies f = 0.4 and above. For example, at f
= 10, still within practical interest for kV imaging appli-
cations, DQE values for CdTe, HgI2, and a-Se are 0.52,
0.45, and 0.31, respectively.
To verify our simulation results, we compared our DQE
values with those published previously. The maximum
Copyright © 2013 SciRes. IJMPCERO
Figure 12. MTF of three photoconductors with thickness of
300 μm. (a) Under 80 kVp beam; (b) Under 120 kVp beam.
Figure 13. DQE(f) of three photoconductors with thickness
of 300 μm. (a) Under 80 kVp beam; (b) Under 120 kVp beam.
measured DQE value of 0.7 for 350 μm thick CdZnTe
was obtained in the study of S. Tokuda, et al. [17] for 70
kVp beam. This is close to calculated in this work
DQE(0) = 0.76 under similar conditions (80 kVp, 300
μm thickness).
The value of DQE(0) for 300 µm thick a-Se under 80
kVp spectrum of our simulation is ~0.53. In the study of
Zhao et al. [1], the calculated DQE(0) of 300 µm under a
spectrum of 70 kVp is about 0.6 due to the Swank factor
and the broad x-ray spectrum using the cascade systems
model. These two values are quite close taking into ac-
count the differences in calculation approaches and the
input spectra. Du et al. [6] studied physical vapor deposi-
tion (PVD) polycrystalline HgI2 with thickness range
from 210 µm to 300 µm under an x-ray spectrum of 72
kVp. Their theoretical DQE(0) of 210 µm thick HgI2
prototype calculated with cascaded system is about 0.6,
taking quantum efficiency and Swank factor into account.
This is in agreement with our calculated DQE(0) = 0.78
for 300 µm HgI2 under 80 kVp, taking into account the
thickness and spectra differences.
While intrinsic imaging properties of HgI2 and CdTe
are close, comparison of their electronic properties favors
the latter material, whose carrier mobilities are at least an
order of magnitude higher [48], making it more attractive
for use in radiation detectors.
3.6. Device Operation
In order to separate and collect the charged carriers gen-
erated in the existing CdTe based radiation detectors, a
strong electric field ~103 - 104 V/cm is typically created
by application of the external biasing voltage across the
sensitive volume of the device. In this approach the value
of the noise due to leakage current associated with the
application of a large bias (~1000 V for ~1 mm thick
device) becomes the primary concern. Additionally, prob-
lems of charge trapping and polarization effect specific
to crystalline CdTe strongly contribute to drift in the de-
tector response.
With easier to grow, and typically thinner polycrystal-
line semiconductor films, it is possible to utilize a photo-
voltaic (PV) mode of operation, where a junction created
at the interface of differently doped semiconductors and/
or semiconductor-metal provides the built-in electric
field. In the photovoltaic mode the problems related to
biasing do not come into play. However, the major diffi-
culty arises from incomplete charge collection due to lo-
calization of the electric field to a region that can be
much shorter than the detector thickness.
A number of groups verified the feasibility of x-ray
detectors operating in PV mode, including those based on
Si [49], CdTe [50,51], and GaAs [52]. In the case of
GaAs it was demonstrated that the depletion region can
extend over the entire film thickness of 80 μm, resulting
Copyright © 2013 SciRes. IJMPCERO
in equally efficient charge collection of ~93% for both
PV and biased modes of operation. While it is the state of
the art GaAs structure with non-uniform doping profile
that makes such a field distribution possible, it provides a
proof of PV mode concept and points toward the need for
a stronger materials science effort in x-ray detector re-
search. So far a similar undertaking in polycrystalline
CdTe doping yielded detectors with thickness of up to
150 µm achieving full depletion at bias of <1 V [53].
At this time the deposition techniques, material proc-
essing, and device engineering approaches for CdTe
come mostly from the commercially successful field of
polycrystalline thin-film CdTe solar cells. Based on the
established results of thin-film photovoltaic applications,
the built-in field of ~104 V/cm is induced in the vicinity
of CdTe/CdS where the CdS layer can be extremely thin
(below 1 micron) and can be neglected from the point of
view of charge carrier generation. The electric field re-
gion in a much thicker (thin-film) CdTe extends over a
few micron depletion width Ld from the metallurgical
Assuming, for absorption purposes, CdTe thickness
much greater than Ld the arrangement becomes superfi-
cially analogous to that of the classical structure of crys-
talline silicon p-n junction photovoltaic device, which
remains the most established type of commercially avail-
able solar cells. In those devices charge carrier genera-
tion spans distances orders of magnitude greater than the
depletion width Ld that is on the order of microns. The
consensus for the case of Si is that the carriers generated
far from the built-in field region are collected mostly
through the diffusion currents. The key factor behind
efficient diffusion currents in crystalline Si is a signifi-
cant diffusion length D
comparable to the device
thickness, where D is the diffusion coefficient and τ is the
minority carrier lifetime. Simply stated, it is a high mate-
rial quality of crystalline Si that provides long enough τ ~
1 ms [48] and efficient diffusion (note that in some
analyses the product of mobility and τ proportional to Dτ
is taken to be a figure of merit).
Extending the diffusion current argument over to the
case of CdTe may seem unsuccessful if one use the life-
time τ ~ 1 µs measured for crystalline X-ray detectors
[48]. However, the latter quantity does not describe the
transport phenomena in polycrystalline films where the
non-equilibrium electrons and holes are spatially sepa-
rated by the built-in electric field at grain boundaries.
Such a separation can exponentially suppress recombina-
tion making charge carrier lifetimes orders of magnitude
longer. This understanding is consistent with the fact that
CdTe based PV cells retain their high efficiencies in the
thickness range of 10 µm and above [54]. Preliminary
measurements of carrier lifetime for ~100 µm thick films
have demonstrated values at least 10 times higher [53]
than those commonly accepted for the single crystal
CdTe. It is only a speculation at this time, but the effec-
tive diffusion length in polycrystalline CdTe may be
comparable to that of crystalline Si.
The latter speculation is consistent with the results of
our device operation modeling (with SCAPS software;
not presented here) based on the baseline input parame-
ters commonly accepted for CdTe based PV [55]. That
modeling shows that device voltage is a weak function of
the CdTe layer thickness, saturating for devices thicker
than 100 microns. The current, however, increases sev-
eral times when the device thickness changes from 10 to
100 microns saturating for devices thicker than ~300
microns. This is consistent indeed with the above-men-
tioned picture of crystalline Si devices with the electric
filed limited to near the junction and significant contribu-
tion of the diffusion current to the output signal.
In view of these observations, CdTe device can be op-
erated in two regimes: 1) if the output current is meas-
ured, then applying external bias is beneficial for boost-
ing signal collection; 2) in the voltage measurement
setup, carrier collection without external biasing (photo-
voltaic mode of operation) can provide strong enough
signal for a device much thicker than the depletion width.
Comparison between these two regimes cannot be made
based on theoretical estimates only since there are many
factors affecting the leakage current and the transport
properties in polycrystalline devices. For example, addi-
tion of CdS as buffer layer has been shown to dramati-
cally decrease the leakage current [14-17]. More experi-
mental studies on electronic properties of >100 μm thick
polycrystalline CdTe films and detectors are called upon.
4. Conclusions
We calculated x-ray spectra for our clinical diagnostic
x-ray simulator unit for tube potential voltages from 70
to 140 kVp employing a tungsten anode spectral mode
using interpolating polynomials (TASMIP) algorithm.
Using these spectra as input we studied the frequency-
dependent characteristics, MTF, NPS, and DQE, of thin
film direct detection CdTe imager by Monte Carlo simu-
lations of energy deposition. Our results showed that the
detector based on CdTe is capable of achieving a high
resolution as well as high quantum efficiency. We found
the optimal thickness of thin film CdTe for diagnostic
x-ray imaging to be in the range of 300 to 600 μm.
Utilization of high-density and high-atomic number
photoconductors is especially important for lowering the
exposure levels of fluoroscopic imaging mode, since they
are capable of not only detecting a larger fraction of the
incident radiation, but are also likely to reduce the lateral
spread of secondary electrons and photons. Physical
properties of thin-film CdTe, such as mechanical and
chemical stability, together with its outstanding radiation
Copyright © 2013 SciRes. IJMPCERO
hardness and recently developed film deposition tech-
niques, make this photoconductor an ideal choice for
large-area direct detection system employed in x-ray di-
agnostic imaging applications.
[1] W. Zhao and J. A. Rowlands, “Digital Radiology Using
Active Matrix Readout of Amorphous Selenium: Theo-
retical Analysis of Detective Quantum Efficiency,”
Medical Physics, Vol. 24, No. 12, 1997, pp. 1819-1833.
[2] W. Zhao, I. Blevis, S. Germann, J. A. Rowlands, D. Wae-
chtcher and Z. S. Huang, “Digital Radiology Using Ac-
tive Matrix Readout of Amorphous Selenium: Construc-
tion and Evaluation of a Prototype Real-Time Detector,”
Medical Physics, Vol. 24, No. 12, 1997, pp. 1834-1843.
[3] D. L. Lee, L. K. Cheung, B. G. Rodricks and G. F. Powell,
“Improved Imaging Performance of a 14×17-Inch Direct
Radiography System Using Se/TFT Detector,” Proceed-
ings of SPIE, Vol. 3336, 1998, pp. 14-23.
[4] A. Tsukamoto, S. Yamada, T. Tomisaki, M. Tanaka, T.
Sakaguchi, K. Suzuki and M. Ikeda, “Development and
Evaluation of a Large-Area Selenium-Based Flat-Panel
Detector for Real-Time Radiography and Fluoroscopy,”
Proceedings of SPIE, Vol. 3659, 1999, pp. 14-23.
[5] L. E. Antonuk, K.-W. Jee, Y. El-Mohri, M. Maolinbay, S.
Nassif, X. Rong, Q. Zhao, J. H. Siewerdsen, R. A. Street
and K. S. Shah, “Strategies to Improve the Signal and
Noise Performance of Active Matrix, Flat-Panel Imagers
for Diagnostic X-Ray Applications,” Medical Physics,
Vol. 27, No. 2, 2000, pp. 289-306. doi:10.1118/1.598831
[6] H. Du, L. E. Antonuk, Y. El-Mohri, Q. Zhao, Z. Su, J.
Yamamoto and Y. Wang, “Investigation of Signal Be-
havior of Polycrystalline HgI2 at Diagnostic Energies of
Prototype, Direct Detection, Active Matrix, Flat-Panel
Imagers,” Physics in Medicine and Biology, Vol. 53, No.
5, 2008, pp. 1325-1351. doi:10.1088/0031-9155/53/5/011
[7] K. S. Shan, P. R. Bennett, M. Klugerman, L. P. Moy, G.
Entine, D. R. Ouimette and R. Aikens, “Lead Iodide
Films for X-Ray Imaging,” Proceedings of SPIE, Vol.
3032, 1997, pp. 395-404.
[8] R. A. Streat, J. T. Rahn, S. E. Ready, K. S. Shah, P. R.
Bennett, Y. N. Dmitriyev, P. Mei, J. P. Lu, R. B. Apte, J.
Ho, K. Van Schuylenbergh, F. Lemmi, J. B. Boyce and P.
Nylen, “X-Ray Imaging Using Lead Iodide as a Semi-
conductor Detector,” Proceedings of SPIE, Vol. 3659,
1999, pp. 36-47.
[9] G. Zentai, L. Partain, R. Pavlyuchkova, G. Virshup, A.
Zuck, L. Melekhov, O. Dagan, A. Vilensky and H. Gil-
boa, “Large Area Mercuric Iodide X-Ray Imager,” Pro-
ceedings of SPIE, Vol. 4682, 2002, pp. 592-600.
[10] Y. Kang, L. E. Antonuk, Y. El-Mohri, L. Hu, Y. Li, A.
Sawant, Z. Su, Y. Wang, J. Yamamoto and Q. Zhao,
“Examination of PbI2 and HgI2 Photoconductive Materi-
als for Direct Detection, Active Matrix, Flat-Panel Imag-
ers for Diagnostic X-Ray Imaging,” IEEE Transactions
on Nuclear Science, Vol. 52, No. 1, 2005, pp. 38-45.
[11] Z. Su, L. E Antonuk, Y. El-Mohri, L. Hu, Hong Du, A.
Sawant, Y. Li, Y. Wang, J. Yamamoto and Q. Zhao,
“Systematic Investigation of the Signal Properties of
Polycrystalline HgI2 Detectors under Mammographic,
Radiographic, Fluoroscopic and Radiotherapy Irradiation
Conditions,” Physics in Medicine and Biology, Vol. 50,
No. 12, 2005, pp. 2907-2928.
[12] M. Simon, R. A. Ford, A. R. Franklin, S. P. Grabowski, B.
Menser, G. Much, A. Nascetti, M. Overdick, M. J. Po-
well and D. U. Wiechert, “PbO as Direct Conversion
X-Ray Detector Material,” Proceedings of SPIE, Vol.
5368, 2004, pp. 188-199.
[13] D. R. Ouimette, S. Nudelman and R. S. Aikens, “A New
Large-Area X-Ray Image Sensor,” Proceedings of SPIE,
Vol. 3336, 1998, pp. 470-476.
[14] S. Adachi, N.Hori, K. Sato, K. Uehara, Y. Izumi, H. Na-
gata, Y. Yoshimura and S. Yamada, “Experimental Eval-
uation of a-Se and CdTe Flat-Panel X-Ray Detectors for
Digital Radiography and Fluoroscopy,” Proceedings of
SPIE, Vol. 3977, 2000, pp. 38-47.
[15] S. Tokuda, S. Adachi, T. Sato, T. Yoshimuta, H. Nagata,
K. Uehara, Y. Izumi, O. Teranuma and S. Yamada, “Ex-
perimental Evaluation of a Novel CdZnTe Flat-Panel
X-Ray Detector for Digital Radiography and Fluoros-
copy,” Proceedings of SPIE, Vol. 4320, 2001, pp. 140-
[16] S. Tokuda, H. Kishihara, S. Adachi, et al., “Large-Area
Deposition of a Polycrystalline CdZnTe Film AND Its
Applicability to X-Ray Panel Detectors with Superior
Sensitivity,” Proceedings of SPIE, Vol. 4682, 2002, pp.
[17] S. Tokuda, H. Kishihara, S. Adachi and T. Sato, “Im-
provement of the Temporal Response and Output Uni-
formity of Polycrystalline CdZnTe Films for High Sensi-
tivity X-Ray Imaging,” Proceedings of SPIE, Vol. 5030,
2003, pp. 861-870.
[18] C. Choi, C. Kyun, S. Kang and S. Nam, “Comparison of
Compound Semiconductor Radiation Films Deposited by
Screen Printing Method,” Proceedings of SPIE, Vol.
6510, 2007, pp. 651042-1-651042-8.
[19] M. Hoheisel, J. Giersch and P. Bernhardt, “Intrinsic Spa-
tial Resolution of Semiconductor X-Ray Detectors: A
Simulation Study,” Nuclear Instruments and Methods in
Physics Research Section A, Vol. 531, No. 1-2, 2004, pp.
75-81. doi:10.1016/j.nima.2004.05.077
[20] B. E. McCandless and J. R. Sites, “Cadmium Telluride
Solar Cells,” In: A. Lique and S. Hegedus, Eds., Hand-
book of Photovoltaic Science and Engineering, Wiley,
New York, 2003, pp. 617-657.
[21] NIST Database, 2013.
[22] D. Shvydka, E. Parsai and J. Kang, “Radiation Hardness
Study of CdTe Thin Films for Clinical High-Energy Pho-
ton Beam Detectors,” Nuclear Instruments and Methods
in Physics Research Section A, Vol. 586, No. 2, 2008, pp.
169-173. doi:10.1016/j.nima.2007.11.017
Copyright © 2013 SciRes. IJMPCERO
[23] A. Cavallini, B. Fraboni, W. Dusi, N. Auricchio, P. Chir-
co, M. Zanarini, P. Siffert and P. Fougeres, “Radiation
Effects on II-VI Compound-Based Detectors,” Nuclear
Instruments and Methods in Physics Research Section A,
Vol. 476, No. 3, 2002, pp. 770-778.
[24] Y. Eisen, L. G. Evans, S. Floyd, C. Schlemm, R. Starr
and J. Trombka, “Radiation Damage of Schottky CdTe
Detectors Irradiated by 200 MeV Protons,” Nuclear In-
struments and Methods in Physics Research Section A,
Vol. 491, No. 1-2, 2002, pp. 176-180.
[25] D. L. Batzner, A. Romeo, M. Dobeli, K. Weinert, H.
Zogg and A. N. Tiwari, “High Energy Irradiation Proper-
ties of CdTe/CdS Solar Cells,” Proceedings of 29th IEEE
Photovoltaic Specialists Conference, New Orleans, 19-24
May 2002, pp. 982-985.
[26] E. Rossa, H. Schmickler, A. Brambilla, L. Verger, F. Mon-
gellaz, “New Development of a Radiation-Hard Poly-
crystalline CdTe Detector for LHC Luminosity Monitor-
ing,” Proceedings of DIPAC 2001—ESRF, Grenoble,
2001, pp. 94-96
[27] E. I. Parsai, D. Shvydka and J. Kang, “Design and Opti-
mization of Large Area Thin Film CdTe Detector for Ra-
diation Therapy Imaging Applications,” Medical Physics,
Vol. 37, No. 8, 2010, pp. 3980-3994.
[28] F. B. Brown, “MCNP—A General Monte Carlo N-Parti-
cle Transport Code,” Version 5, LA-UR-03-1987, Los
Alamos National Laboratory, Los Alamos, 2003.
[29] J. M. Boone and J. A. Seibert, “An Accurate Method for
Computer Generating Tungsten Anode X-Ray Spectra
from 30 to 140 kV,” Medical Physics, Vol. 24, No. 11,
1997, pp. 1661-1670. doi:10.1118/1.597953
[30] T. R. Fewell, R. E. Shuping and K. E. Healy, “Handbook
of Computed Tomography X-Ray Spectra,” HHS Publi-
cation (FDA) 81-8162, US Government Printing Office,
Washington DC, 1981.
[31] D. W. O. Rogers, “Low Energy Electron Transport with
EGS,” Nuclear Instruments and Methods, Vol. 227, No. 3,
1984, pp. 535-548. doi:10.1016/0168-9002(84)90213-4
[32] A. F. Bielajew and D. W. O. Rogers, “Experimental
Benchmarks of EGS,” In: T. M. Jenkins, W. R. Nelson, A.
Rindi, A. E. Nahum and D. W. O. Rogers, Eds., Monte
Carlo Transport of Electrons and Photons, Prenum, New
York, 1988, pp. 115-137.
[33] H. Fujita, D.-Y. Tsai, T. Itoh, K. Doi, J. Morishita, K.
Ueda and A. Ohtsuka, “A Simple Method for Determin-
ing the Modulation Transfer-Function in Digital Radiog-
raphy,” IEEE Transactions on Medical Imaging, Vol. 11,
No. 1, 1992, pp. 34-39. doi:10.1109/42.126908
[34] J. C. Dainty and R. Shaw, “Imaging Science: Principles,
Analysis and Evaluation of Photographic-Type Imaging
Processes,” Academic, London, 1974.
[35] J. H. Siewerdsen, L. E. Antonuk, Y. EI-Mohri, J. York-
ston, W. Huang and I. A. Cunningham, “Signal, Noise
Power Spectrum, and Detective Quantum Efficiency of
Indirect-Detection Flat-Panel Images for Diagnostic Ra-
diology,” Medical Physics, Vol. 25, No. 5, 1998, pp.
614-628. doi:10.1118/1.598243
[36] M. B. Williams, P. A. Mangiafico and P. U. Simoni,
“Noise Power Spectra of Images from Digital Mammog-
raphy Detectors,” Medical Physics, Vol. 26, No. 7, 1999,
pp. 1279-1293. doi:10.1118/1.598623
[37] A. Sawant, L. E. Antonuk, Y. EI-Mohri, Q. Zhao, Y. Li,
Z. Su, Y. Wang, J. Yamamoto, H. Du, I. Cunningham, M.
Klugerman and K. Shah, “Segmented Crystalline Scintil-
lators: An Initial Investigation of High Quantum Effi-
ciency Detectors for Megavoltage X-Ray Imaging,”
Medical Physics, Vol. 32, No. 10, 2005, pp. 3067-3083.
[38] D. A. Jaffray, J. J. Battista, A. Fenster and P. Munro,
“Monte Carlo Studies of X-Ray Energy Absorption and
Quantum Noise in Megavoltage Transmission Radiogra-
phy,” Medical Physics, Vol. 22, No. 7, 1995, pp. 1077-
1088. doi:10.1118/1.597593
[39] I. Cunningham, “Applied Linear-Systems Theory,” In: J.
Beutel, H. L. Kundel and R. L. Van Metter, Eds., Hand-
book of Medical Imaging, SPIE Press, Bellingham, 2000,
pp. 79-159.
[40] M. J. Yaffe and R. M. Nishikawa, “X-Ray Imaging Con-
cepts: Noise, SNR and DQE,” In: J. A. Seibert, G. T.
Barnes and R. G. Gould, Eds., Specification, Acceptance
Testing and Quality Control of Diagnostic X-Ray Imaging
Equipment, Medical Physics Monograph, The American
Institute of Physics, New York, 1994, pp. 109-144.
[41] A. L. Fahrenbruch and R. H. Bube, “Fundamentals of
Solar Cells,” Academic Press, New York, 1983.
[42] R. Harju, V. G. Karpov, D. Grecu and G. Dorer, “Elec-
tron-Beam Induced Degradation in CdTe Photovoltaics,”
Journal of Applied Physics, Vol. 88, No. 4, 2000, pp.
1794-1801. doi:10.1063/1.1305857
[43] D. S. McGregor and H. Hermon, “Room-Temperature
Compound Semiconductor Radiation Detectors,” Nuclear
Instruments and Methods in Physics Research Section A,
Vol. 395, No. 1, 1997, pp. 101-124.
[44] D. L. Lee, L. K. Cheung, E. F. Palechi and L. S. Jeromin,
“A Discussion on Resolution and Dynamic Range of
Se-TFT Direct Digital Radiographic Detector,” Proceed-
ings of SPIE, 11 April 1996, Newport Beach, pp. 511-
[45] J. M. Boone, J. A. Seibert, J. M. Sabol and M. Tecotzky,
“A Monte Carlo Study of X-Ray Fluorescence in X-Ray
Detectors,” Medical Physics, Vol. 26, No. 6, 1999, pp.
905-916. doi:10.1118/1.598612
[46] G. F. Knoll, “Radiation Detection and Measurement,” 3rd
Edition, Wiley, Hoboken, 2000.
[47] N. Strokan, V. Ajdacic and B. Lalovic, “Measurements of
the Fano Factor in Germanium,” Nuclear Instruments and
Methods, Vol. 94, No. 1, 1971, pp. 147-149.
[48] A. Owens and A. Peacock, “Compound Semiconductor
Radiation Detectors,” Nuclear Instruments and Methods
in Physics Research Section A, Vol. 531, No. 1-2, 2004,
pp. 18-37. doi:10.1016/j.nima.2004.05.071
Copyright © 2013 SciRes. IJMPCERO
Copyright © 2013 SciRes. IJMPCERO
[49] K. Scharf, “Photovoltaic Effect Produced in Silicon Solar
Cells by X- and Gamma Rays,” Journal of Research of
the National Bureau of Standards, Vol. 64A, No. 4, 1960,
pp. 297-307.
[50] G. Entine, R. H. Redus, A. Feyder and P. J. Biggs, “Re-
cent Results with a CdTe Imaging Portal Scanner for Ra-
diation Therapy,” IEEE Transactions on Nuclear Science,
Vol. 40, No. 4, 1993, pp. 1012-1016.
[51] R. J. Fox and D. C. Agouridis, “CdTe Photovoltaic Gam-
ma-Ray Dosimeter,” Nuclear Instruments and Methods,
Vol. 157, No. 1, 1978, pp. 65-69.
[52] R. A. Achmadullin, V. V. Artemov, V. F. Dvoryankina,
G. G. Dvoryankina, Y. M. Dikaeva, M. G. Ermakova, O.
N. Ermakovaa, V. B. Chmil, A. G. Holodenkoc, A. A.
Kudryashova, A. I. Krikunova, A. G. Petrov, A. A.
Telegina and A. P. Vorobiev, “Photovoltaic X-Ray De-
tectors Based on Epitaxial GaAs Structures,” Nuclear In-
struments and Methods in Physics Research Section A,
Vol. 554, No. 1-3, 2005, pp. 314-319.
[53] B. N. Zaveryukhin, S. A. Mirsagatov, N. N. Zaveryu-
khina, V. V. Volodarski and E. B. Zaveryukhina, “Cad-
mium Telluride Thin-Film Detectors of Nuclear Radia-
tion,” Technical Physics Letters, Vol. 29, No. 11, 2003,
pp. 959-962. doi:10.1134/1.1631378
[54] X. Wu, “High-Efficiency Polycrystalline CdTe Thin-Film
Solar Cells,” Solar Energy, Vol. 77, No. 6, 2004, pp. 803-
814. doi:10.1016/j.solener.2004.06.006
[55] M. Burgelman, P. Nollet and S. Degrave, “Modelling
Polycrystalline Semiconductor Solar Cells,” Thin Solid
Films, Vol. 361-362, 2000, pp. 527-532.