HIROYUKI TAKAGI AND ISAO MURATA: ENERGY SPECTRUM MEASUREMENT OF 6 AND 9 MeV X-ray SOURCES

1)

pISSN 2508-1888
eISSN 2466-2461

Journal of http://dx.doi.org/10.14407/jrpr.2016.41.2.093

Radiation Protection
and Research Paper
Received July 17, 2015 / 1st Revised April 24, 2016 / Accepted June 13, 2016

Energy Spectrum Measurement of High Power and High

Energy (6 and 9 MeV) Pulsed X-ray Source for Industrial Use

Hiroyuki Takagi*, † and Isao Murata†

*Hitachi, Ltd. Power Systems Company, Ibaraki, Japan
†
Graduate School of Engineering, Osaka University, Osaka, Japan

ABSTRACT
Background: Industrial X-ray CT system is normally applied to non-destructive testing
(NDT) for industrial product made from metal. Furthermore there are some special CT
systems, which have an ability to inspect nuclear fuel assemblies or rocket motors, using
high power and high energy (more than 6 MeV) pulsed X-ray source. In these case,
pulsed X-ray are produced by the electron linear accelerator, and a huge number of pho-
tons with a wide energy spectrum are produced within a very short period. Consequently,
it is difficult to measure the X-ray energy spectrum for such accelerator-based X-ray sour-
ces using simple spectrometry. Due to this difficulty, unexpected images and artifacts
which lead to incorrect density information and dimensions of specimens cannot be avoid-
ed in CT images. For getting highly precise CT images, it is important to know the pre-
cise energy spectrum of emitted X-rays.
Materials and Methods: In order to realize it we investigated a new approach utilizing the
Bayesian estimation method combined with an attenuation curve measurement using step
shaped attenuation material. This method was validated by precise measurement of energy
spectrum from a 1 MeV electron accelerator. In this study, to extend the applicable X-ray
energy range we tried to measure energy spectra of X-ray sources from 6 and 9 MeV lin-
ear accelerators by using the recently developed method.
Results and Discussion: In this study, an attenuation curves are measured by using a
step-shaped attenuation materials of aluminum and steel individually, and the each X-ray
spectrum is reconstructed from the measured attenuation curve by the spectrum type
Correspondence to Hiroyuki Takagi Bayesian estimation method.
hiroyuki.takagi.bc@hitachi.com Conclusion: The obtained result shows good agreement with simulated spectra, and the
presently developed technique is adaptable for high energy X-ray source more than 6
MeV.

Keywords: Bayesian estimation, Energy distribution, High energy X-ray, Non-destructive

testing, Attenuation curve, accelerator

This is an Open-Access article distributed under the terms of the Creative Commons Attribution Non- Commercial License
(http://creativecommons.org/licenses/by-nc/3.0) which permits unrestricted non-commercial use, distribution, and re-
production in any medium, provided the original work is properly cited.

JOURNAL OF RADIATION PROTECTION AND RESEARCH, VOL.41 NO.2 JUNE 2016 93

HIROYUKI TAKAGI AND ISAO MURATA: ENERGY SPECTRUM MEASUREMENT OF 6 AND 9 MeV X-ray SOURCES

1. INTRODUCTION2) photons emitted from a linear accelerator for in-

dustrial use is more than 1.0×106 photons·mm-2 at 1
In the case of accelerator-based X-ray sources for m from the target on the beam center axis for each
industrial computed tomography systems (CTs) to in- pulse of 6 MeV electrons. This means that direct
spect metal and large materials, X-ray energy should measurement of the pulse height spectrum is almost
be higher and they are produced by electrons with en- impossible for high energy X-ray source using ordi-
ergies of more than 6 MeV. In such X-ray sources, nary X-ray spectrometry.
the peak current of the electron beam is more than To realize accurate X-ray energy spectrum measure-
100 mA and the pulse width is about 5 μs. (In this ment, we consider penetration the polychromatic
paper, X-rays over 6 MeV are called high energy X-rays with energy spectrum, φ(Ej) [j=1–M], through
X-rays.) As a result, a huge number of photons with an attenuation material, the thickness of which can be
a wide energy spectrum are produced within quite a changed gradually as ti [i=1–N]. An experimental ar-
short time. Consequently, it is difficult to measure the rangement for the attenuation curve measurement is
X-ray energy spectrum using simple spectrometry es- shown in Figure 1. An electron linear accelerator of 6
pecially for such commercially available accel- and 9 MeV electrons switchable (Linatron-M, Varian
erator-based X-ray sources. Due to this difficulty, nor- Medical Systems Inc., Palo Alto, CA), with a pulse
mally unexpected images and artifacts which lead to width of 4.2 μs, is utilized as an X-ray source. At 1
incorrect density information and dimensions of speci- m from the focal spot, the dose rate is 10 Gy·min-1
mens cannot be avoided in the CT images. For get- for 6 MeV and 30 Gy·min-1 for 9 MeV without a
ting more precise CT images, it is important to know flattening filter in both cases. In the X-ray CT sys-
the precise energy spectrum of emitted X-rays. tem, there are several hundred of solid state detectors
Several techniques have been proposed so far, in arranged within the 15° fan beam. This enables us to
which X-ray intensities are measured after penetrating measure multiple attenuation curves within the fan
through attenuation material in order to estimate the beam at the same time.
energy spectrum roughly with a wide energy bin In addition, incidence of scattered X-rays is pre-
width of around 1 MeV or to estimate quite low en- vented by a multi-hole collimator which is placed in
ergy X-rays of less than 100 keV [1-3]. In these cas- front of the detector array in order to make each de-
es, the accuracy of the estimated results is still in- tector view only the focal spot. Then, a step-shaped
sufficient from the standpoint of engineering for high attenuation material made of aluminum or steel is set
energy X-ray CT systems up to 9 MeV, because the between the collimator and the X-ray source so as to
energy resolution is insufficient, in other words, a fine cover the fan beam. On measurement of 1 MeV
energy bin width of less than 100 keV is required. In X-ray source, we previously applied the attenuation
order to know such a precise energy spectrum, we in- material of aluminum at maximum 300 mm thick
vestigated a new approach utilizing the Bayesian esti- with 10 mm step. In this study, the attenuation mate-
mation method combined with an attenuation curve rial is 580 mm thick of aluminum and 290 mm thick
measurement using step shaped attenuation material. of steel respectively at maximum on the center line
This method was validated by precise measurement of and it has 58 or 29 steps start from 10 mm to max-
energy spectrum from a 1 MeV electron accelerator imum thickness with a 10 mm step. This means that
[4]. Our purpose in this study is to extend the appli- the attenuation material is modified to relatively high-
cable X-ray energy range. Using the recently devel- er dense material or extended to relatively thicker
oped technique, we measure precise energy spectra of shape than those of 1 MeV. In the case of the alumi-
X-ray sources from 6 and 9 MeV linear accelerators num attenuation material, the obtained attenuation
with an energy bin width of 100 keV. curve has a character of slow attenuation with a lot
of measurement data. Normally, sufficient measure-
ment data will support calculating huge matrix
2. MATERIALS AND METHODS equation. On the other hand, the steel attenuation ma-
terial derives relatively large attenuation and quick at-
2.1 Attenuation curve measurement tenuation with small number measurement point.
According to results obtained with EGS5 [5], a Consequently, in this study, we can compare adapt-
Monte Carlo simulation code, the number density of ability of slow attenuation and quick attenuation.
By moving the attenuation material, step-by-step,
Corresponding author : Hiroyuki Takagi, hiroyuki.takagi.bc@hitachi.com the attenuation curve of each detector from 10 mm to
1-1, Saiwai-cho 3-chome, Hitachi-shi, Ibaraki-ken, 317-8511 Japan

94 JOURNAL OF RADIATION PROTECTION AND RESEARCH, VOL.41 NO.2 JUNE 2016 http://dx.doi.org/10.14407/jrpr.2016.41.2.093
 HIROYUKI TAKAGI AND ISAO MURATA: ENERGY SPECTRUM MEASUREMENT OF 6 AND 9 MeV X-ray SOURCES

【Top View】 【Side View】

X-ray source X-ray source

(e- accelerator) (e- accelerator)

I 0 :Primary Intensity
Appearance of
Focal Spot Fan shaped Beam (20°) Attenuate Material
tN(290 or 580㎜@θ=0º )

Attenuation

tN(290 or 580㎜ @θ=0º )

Material of
Aluminum

t1 (Un-attenuated)
ti or Steel

ti
t2(10㎜ @θ=0º )
t3(20㎜ @θ=0º )
t4(30㎜@θ=0º )

Path Length

(Attenuation Material)
in Material

I:Attenuated Intensity

Vertical movement
θ
Collimator

+15° -15°
・・・ ・・・
Solid State
・・・ ・・・ Detector
Multi Channel Multi Channel
Preamp Board Preamp Board
Fig. 1. Experimental setup of array detectors and X-ray source used for the present attenuation curve measurement (The actual path length
of the attenuation material is changing according to the angle θ of incident X-rays).

maximum thickness and in addition at 0 mm (with no ation curve measurement using step shaped attenu-
attenuation material) is obtained. As a result, 59 or 30 ation material. This method was validated by precise
X-ray outputs for each detector are recorded sepa- measurement of energy spectrum from a 1 MeV elec-
rately after penetrating the 59 or 30 steps of the at- tron accelerator. The specific features of the spectrum
tenuation material by 30 or 59 X-ray irradiations type Bayesian estimation method are summarized as
(more than 10,000 pulse per irradiation), that is to follows.
say, attenuation curves yθ(ti) [θ＝0 to ±15°, i=1 to 59 (1) Its very simple principle makes for easier appli-
or 30], are obtained for all the detectors at the same cation compared to other unfolding methods.
time. (2) It never gives unfolded results having a neg-
The path length of X-rays in the attenuation materi- ative value.
al changes depending on the incident angle θ of the (3) It does not require any constraint like an initial
X-rays as shown in Fig. 1. The actual path length is guess [10].
used for ti in the calculation of energy spectra.
There should be a quantitative relation between
2.2 Spectrum type Bayesian estimation X-ray energy spectrum φ(Ej) and measured attenuation
We use the spectrum type Bayesian estimation result y(ti) for the attenuation curve. It can be de-
method [6] to obtain X-ray energy spectrum φ(Ej). scribed with a response function. The response func-
This method is based on the Bayesian theorem and it tion H consists of hij , which includes information on
has been applied to radiation measurements. For in- the attenuation coefficient at photon energy E, pene-
stance, in energy spectrum measurements of charged tration thickness ti, and other detector characteristics
particles correction of the energy loss in the sample is such as its detection efficiency. This relationship is
carried out using the spectrum type Bayesian estima- described as Equation 1.
tion method as an unfolding technique [7-9].
Furthermore, in X-ray energy spectrum measurement, y=Ｈ•φ (1)
we recently investigated an approach utilizing the In the spectrum type Bayesian estimation method,
Bayesian estimation method combined with an attenu- original X-ray spectrum is estimated iteratively carried

JOURNAL OF RADIATION PROTECTION AND RESEARCH, VOL.41 NO.2 JUNE 2016 95

HIROYUKI TAKAGI AND ISAO MURATA: ENERGY SPECTRUM MEASUREMENT OF 6 AND 9 MeV X-ray SOURCES

out based on Bayes' theorem. The spectrum type function R•Ｆ in Equation 3 in this section, focusing
Bayesian estimation method deduce the l+1 th estima- on a 6 MeV X-ray source and attenuation material of
tion estj(l+1)(i) as, aluminum as an example. The maximum energy of
the response function is therefore set to 6 MeV to
N ⎛ est (jl ) × hij ⎞ cover X-rays emitted from the 6 MeV X-ray source.
⎜ ⎟
est (jl +1) = ∑ ⎜ y (t i ) × M ⎟⎟ The column of the response function R is then div-
i =1 ⎜
⎝ ∑ j =1 est (jl ) × hij ⎠ (2) ided into 60 bins, namely they cover the X-ray spec-
trum from 0.0 to 6.0 MeV with an equal bin width of
The above formula is repeatedly used for the meas- 0.1 MeV. This means j=1–M (M=60) in Eq. 3. On
ured attenuation curve y(ti), i.e., a revised estj(l) is the other hand, the row component for the attenuation
used as the prior information in the next iteration material thickness in Eq. 3 is divided into 59, that is,
calculation. i=1–N (N=59) as mentioned later in detail. The de-
tection efficiency matrix F is evaluated by calculating
2.3 Response function evaluation the reaction rate of incident monochromatic photons
In order to estimate the source X-ray energy spec- for each energy bin in the actually used detector. The
trum, the response function H in eq. 1 is rewritten as obtained reaction rate is converted into the deposited
R•Ｆ to connect the primary X-ray energy spectrum energy in the detector, i.e., electric charge. We use
and the measured pulse height spectrum from preamp the Monte Carlo numerical simulation code EGS5 for
board as shown in Fig. 1. Matrix R of N rows and M this evaluation.
columns consists of attenuation coefficients varying The linear attenuation coefficient used in the calcu-
due to X-ray energy for the column direction and the lation of the response function R is cited from the da-
penetration thickness for the row direction. MatrixＦ ta base of the NIST .
1)

of M rows and M columns contains photon-to-charge Finally, in regard to the scalar value of k, because
conversion factor for the specific X-ray energy in- of the dependence on the individual X-ray measure-
duced the detector. ment system, we decide the value experimentally for
When linear attenuation coefficients μj for energies the system we use.
E1,E2,E3,…,EM as representative energies for energy
bins are given, the X-ray energy spectrum φ□(Ej) af-
ter penetrating through the attenuation material of 3. RESULTS AND DISCUSSION
thickness ti is taken from the incident X-ray energy
spectrum φ(Ej) and exp(-μj·ti). The pulsed X-rays, Figure 2 shows attenuation curves in aluminum for
which penetrated through the attenuation material, are 6 and 9 MeV X-ray sources obtained at emission an-
detected and converted into a large electric charge. gles of 0° and 15°. The attenuation curves in steel are
Generally, the conversion coefficient, that is, the de- also shown in Figure 3. The attenuation curves were
tection efficiency for the X-rays has energy normalized at zero path length. Both figures show
dependence. Now, we define the detection efficiency that each curve of emission angle 0° attenuates slower
εj for X-ray energy of Ej. The output signal yi after than that of 15° similarly for both X-ray energies. In
passing through the attenuation material of thickness ti other words, the energy spectra of 0° contains higher
is proportional to the value of summarized εj ·φ□(Ej) energy photons than 15°. Figures 4 and 5 show X-ray
with energy bin j from 1 to M . energy spectra, φ(Ej), estimated from the attenuation
Applying the energy independent correction factor k curve measurements for 6 MeV and 9 MeV X-ray
to the matrix equation, which depends on the in- sources at emission angle of 0°. The spectrum estima-
dividual measurement system, we express the matrix tion was done with the spectrum type Bayesian esti-
equation to derive the attenuation curve of X-rays, mation method described in chapter 2.2 from the in-
y(ti), as in the next equations. dependent measured attenuation curve in Fig. 2 and 3
as y in Eq. 3. As an initial guess a white spectrum
⎡ y (t1 ) ⎤ ⎡ exp(− μ1 ⋅ t1 ) exp(− μ 2 ⋅ t1 ) L exp(− μ M ⋅ t1 ) ⎤ ⎡ε 1 0 L 0 ⎤ ⎡ φ (E1 ) ⎤ was used. The X-ray energy spectrum simulated by
⎢ y (t ) ⎥ ⎢ L exp(− μ M ⋅ t 2 ) ⎥⎥ ⎢⎢ 0 ε 2 0 ⎥⎥ ⎢⎢ φ (E 2 ) ⎥⎥
⎢ 2 ⎥ = k ⋅ ⎢ exp(− μ1 ⋅ t 2 ) exp(− μ 2 ⋅ t 2 ) L
⎢ M ⎥ ⎢ M M O M
⋅
⎥ ⎢M M
⋅
O M ⎥ ⎢ M ⎥ the EGS5 code is also shown. In the present simu-
⎢ ⎥ ⎢ ⎥ ⎢ ⎥ ⎢ ⎥
⎣ y (t N )⎦ ⎣exp(− μ1 ⋅ t N ) exp(− μ 2 ⋅ t N ) L exp(− μ M ⋅ t N )⎦ ⎣ 0 0 L ε M ⎦ ⎣φ (E M )⎦ lation, we assumed the tungsten target thickness was
y = k•Ｒ•Ｆ•φ (3)

We describe the evaluation process of the response 1) X-ray mass attenuation coefficients at NIST.
http://physics.nist.gov/PhysRefData/XrayMassCoef/tab4.html

96 JOURNAL OF RADIATION PROTECTION AND RESEARCH, VOL.41 NO.2 JUNE 2016

 HIROYUKI TAKAGI AND ISAO MURATA: ENERGY SPECTRUM MEASUREMENT OF 6 AND 9 MeV X-ray SOURCES

100 10-1

X-ray Flux (arb. unit)

10-2
10-1
Attenuation

10-3
Estimated Spectrum from Aluminum Attenuation
10-2
Estimated Spectrum from Steel Attenuation
Measured Attenuation Curve at 0°with 6MeV Simulated Spectrum by EGS5
Measured Attenuation Curve at 15°with 6MeV Initial Guess Spectrum
Measured Attenuation Curve at 0°with 9MeV 10-4
Measured Attenuation Curve at 15°with 9MeV 0 1 2 3 4 5 6 7 8 9
10-3 Photon Energy (MeV)
0 100 200 300 400 500 600

Path Length in Aluminum (mm)

Fig. 5. Measured X-ray energy spectra for 9 MeV X-ray source
at emission angles of 0° (normalized to unity).
Fig. 2. Measured attenuation curves in aluminum with 6 and 9
MeV X-ray source at emission angles of 0° and 15° (The
actual path length of the attenuation material is changing region compared to the experiment. It is difficult to
according to the angle θ of incident X-rays as shown in discuss the reason of this discrepancy. However, it
Fig. 1). can be said that the present unfolding method based
on the Bayes’ theorem may not reproduce spectral
100
Measured Attenuation Curve at 0°with 6MeV values having such a small likelihood. Nevertheless,
Measured Attenuation Curve at 15°with 6MeV
Measured Attenuation Curve at 0°with 9MeV these differences between estimated and simulated
10-1 Measured Attenuation Curve at 15°with 9MeV
spectra at higher energy region are not so serious be-
cause those absolute intensities are quite small.
Attenuation

10-2 Besides, in the present measurement, several hundred

attenuation curves yθ(ti) within the fan shaped beam
10-3
of 15° were measured at the same time. This means
that a lot of angle-dependent X-ray energy spectra
can be measured in the same experiment.
10-4
0 50 100 150 200 250 300 Figures 6 and 7 show measured X-ray energy spec-
Path Length in Steel (mm) tra at typical emission angles of every 5° for 6 MeV
Fig. 3. Measured attenuation curves in steel with 6 and 9 MeV and 9 MeV, respectively. These figures show the ab-
X-ray source at emission angles of 0° and 15° (The actual
solute energy spectra. The vertical axis is the number
path length of the attenuation material is changing accord-
ing to the angle θ of incident X-rays as shown in Fig. 1). of photons incident to each detector per pulse. We al-
so note from the figure that the intensity of X-rays
10-1 decreases with increase of angle θ.
Figures 8 shows the measured average energy for 6
and 9 MeV X-ray sources for each emission angle.
X-ray Flux (arb. unit)

10-2 Here, the average energy is defined as the energy

weighted with the respective normalized spectrum.
Each curve shows same trend that the average energy
10-3
Estimated Spectrum from Aluminum Attenuation decreases with increase of the emission angle θ. The
Estimated Spectrum from Steel Attenuation
Simulated Spectrum by EGS5
figure also denotes that all the angular distributions of
10-4
Initial Guess Spectrum 9 MeV source show more forward-peeked dis-
0 1 2 3 4 5 6 tributions with respect to emission angle compared to
Photon Energy (MeV)
6 MeV source. This means that higher energy photons
Fig. 4. Measured X-ray energy spectra for 6 MeV X-ray source
at emission angles of 0° (normalized to unity). are more dominant in forwarder emission angles for
both X-ray sources, and the case of 9 MeV shows it
1 mm. As shown in Fig. 4 and 5, in spite of esti- more clearly. However, each measured curve is slight-
mated from independent measured curve, the esti- 2)
ly larger than the curve of EGS5 calculation due to
mated spectra look quite a similar shape with each underestimation of EGS5 in higher energy region as
other. Furthermore, the experimental spectra are in mentioned previously.
good agreement with the simulated spectra except for
higher energy region more than 4 MeV (for 6 MeV 2) In EGS5 calculation, angular distribution was estimated with a
X-ray source) or 6 MeV (for 9 MeV X-ray source). geometry having a finite solid angle to decrease the statistical
error. The simulated angular distribution is thus plotted for a fi-
The EGS5 results show smaller values in high energy nite range of solid angle.

JOURNAL OF RADIATION PROTECTION AND RESEARCH, VOL.41 NO.2 JUNE 2016 97

HIROYUKI TAKAGI AND ISAO MURATA: ENERGY SPECTRUM MEASUREMENT OF 6 AND 9 MeV X-ray SOURCES

10-1 7.5×1012
Number of Photons Estimated from Aluminum Attenuation
Number of Photons Estimated from Steel Attenuation

Number of Photons (photons/pulse/sr)

Number of Calculated Photons(EGS5)
X-ray Flux (arb. unit)

Dose Rate Profile (Normalized to the Peak Value of the Estimated Photon Number(Steel))

10-2 5.0×10 12

10-3 2.5×10 12
Estimated Spectrum from Aluminum Attenuation
Estimated Spectrum from Steel Attenuation
Simulated Spectrum by EGS5
Initial Guess Spectrum
0
10-4
0 1 2 3 4 5 6 0 2 4 6 8 10 12 14 16
Emission Angle ( °)
Photon Energy (MeV) Fig. 10. Total photon number of 9MeV X-ray source for each
Fig. 6. Measured X-ray energy spectra for 6 MeV X-ray source emission angle.
at emission angles of 0° (normalized to unity).

lation. The measured values are derived by integrating

10-1 the spectra in such as Fig. 6 and 7. The EGS5 simu-
lation result agrees well with the distribution obtained
X-ray Flux (arb. unit)

by the present technique. Additionally, dose rate pro-

10-2
files from 6 and 9 MeV X-ray sources measured by a
standard air ionization chamber (CC13, IBA
10-3
Dosimetry GmbH, Schwarzenbruck, Germany) are al-
Estimated Spectrum from Aluminum Attenuation
Estimated Spectrum from Steel Attenuation
so shown in Fig. 9 and 10, respectively. Each dose
Simulated Spectrum by EGS5 rate profile is normalized to the peak value of the es-
Initial Guess Spectrum
10-4 timated photon number obtained from the case with a
0 1 2 3 4 5 6 7 8 9
Photon Energy (MeV) steel attenuation. It is notable that the distributions
Fig. 7. Measured X-ray energy spectra for 9 MeV X-ray source obtained by the ionization chamber and others are in
at emission angles of 0° (normalized to unity). good agreement, though EGS5 simulated distribution
might be supported by the distribution of ionization
2.5
chamber daringly. In any case, distributions obtained
2.0 by independent attenuation measurements with alumi-
num and steel are found to be in good agreement
Average Energy (MeV)

1.5
with each other. Besides, a forwarder-peaked dis-
1.0
Measured Average Energy of 6MeV X-ray source from Aluminum Attenuation Curve
tribution of 9 MeV source is observed compared to 6
Measured Average Energy of 6MeV X-ray source from Steel attenuation curve
Calculated Average Energy of 6 MeV X-ray source(EGS5) MeV source similar to the angular distribution of
0.5
average energy.
Measured Average Energy of 9MeV X-ray source from Aluminum Attenuation Curve
Measured Average Energy of 9MeV X-ray source from Steel Attenuation Curve
Calculated Average Enegy of 9MeV X-ray source(EGS5)
0.0
0 2 4 6 8 10 12 14 16
A discontinuity in the measured angular distribution
Emission Angle (º) is surely recognized between emission angles of 6°
Fig. 8. Average energy of X-rays for 6 and 9 MeV X-ray and 7° in Fig. 9 and 10. This is thought to be due to
source for each emission angle.
insufficient structure of the multi-hole collimator
2.5×1012 which is placed in front of the detector array, because
Number of Photons Estimated from Aluminum Attenuation
Number of Photons Estimated from Steel Attenuation it has a segment structure around there and also all
Number of Photons (photons/pulse/sr)

12 Number of Calculated Photons(EGS5)

2.0×10
Dose Rate Profile (Normalized to the Peak Value of the Estimated Photon Number(Steel)) the measured curves show discontinuity.
1.5×1012 We conclude that the angle and energy distributions
of X-rays emitted from high power and high energy
1.0×1012
pulsed X-ray sources used for industrial CT can be
5.0×1011
measured accurately with the recently developed tech-
nique by utilizing thick aluminum attenuation material
0
0 2 4 6 8 10 12 14 16
or steel attenuation material. Especially, it is a very
Emission Angle ( °)
important result for actual CT imaging that hardness
Fig. 9. Total photon number of 6 MeV X-ray source for each
emission angle. of the energy spectrum can be measured quantitatively
and it becomes dominant in forward angles in pro-
Figures 9 and 10 show the angular distributions of portion to the energy of X-ray source.
the number of photons for 6 and 9 MeV X-ray sour- This technique has a potential to become a standard
ces obtained by measurement and the EGS5 simu- method for measurement of energy spectra of in-

98 JOURNAL OF RADIATION PROTECTION AND RESEARCH, VOL.41 NO.2 JUNE 2016

 HIROYUKI TAKAGI AND ISAO MURATA: ENERGY SPECTRUM MEASUREMENT OF 6 AND 9 MeV X-ray SOURCES

dustrially used X-ray sources. Finally, by utilizing in- estimation using attenuation measurement from 25
formation of the X-ray energy spectra obtained by the kVp to 18 MV. Med. Phys. 1999;26:1269-1278.
technique, the quality of obtained images will be im- 02. Iwasaki A, Kubota M, Hirota J, Fujimoto A,
proved in non-destructive industrial inspections such Suzuki K, Aoki M, Abe Y. Characteristic features
as X-ray CT. of a high-energy x-ray spectra estimation method
based on the Waggener iterative perturbation
principle. Med. Phys. 2006;33:4056-4063.
4. CONCLUSION 03. Kanno I, Imamura R, Mikami K, Uesaka A,
Hashimoto M, Ohtaka M, Ara K , Nomiya S,
A recently developed technique to measure X-ray Onabe H. A current-mode detector for unfolding
energy spectra was extended for 6 MeV and 9 MeV X-ray energy distribution. J. Nucl. Sci. Technol.
X-ray sources used for industrial CT. To extend the 2008;45(11):1165-1170.
applicable X-ray energy range we applied two differ- 04. Takagi H, Murata I. Development of precise en-
ent types of attenuation materials, one of which is rel- ergy spectrum measurement technique for high
atively thicker aluminum for slow attenuation and the power pulsed X-ray sources for industrial use. J.
other is thinner steel for quick attenuation. Nucl. Sci. Technol. 2016;53(6):766–773.
In this study, an attenuation curves are measured 05. Hirayama H, Namito Y, Bielajew AF, Wilderman
by using a step-shaped attenuation materials of alumi- SJ, Nelson WR. The EGS5 code system.
num and steel individually, and the each X-ray spec- 2006;SLAC-R-730:1-436
trum is reconstructed from the measured attenuation 06. Iwasaki S. A new approach for unfolding PHA
curve by the spectrum type Bayesian estimation problems based only on the Bayes' theorem. 9th
method. In spite of obtained from independent meas- International Symposium on Reactor Dosimetry.
urement curve with individual attenuation material, it Prague Czech. September 2-6, 1996.
show quite similarity between energy spectra from 07. Takagi H, Terada Y, Murata I, Takahashi A.
aluminum attenuation curve and energy spectra from Measurement of double differential cross section
steel attenuation curve. Comparing with simulated of charged particle emission reactions by incident
spectra by the EGS5 code, the obtained X-ray energy DT neutrons - correction for energy loss of
spectra were confirmed to have good agreement ex- charged particle in sample materials. Symposium
cept for higher energy region. Furthermore, this meas- on Nuclear Data. Tokai Japan. November 18-19,
uring technique derives measure two-dimensional 1999.
(energy and angle dependent) spectra at one time. 08. Nauchi Y, et al. Measurements of hydrogen and
Precise measurement of the X-ray energy spectrum of helium isotopes emission spectra from neutrons
high power and high energy pulsed X-ray sources induced reaction at ten’s of MeV. Symposium on
more than 6 MeV was difficult in the past, but the Nuclear Data. Tokai Japan. November 19-20,
presently expanded technique offers an easy standard 1998.
X-ray energy spectrum measurement technique partic- 09. Kondo K, Murata I, Ochiai K, Kubota N,
ularly for industrial X-ray sources. Miyamaru H, Konno C, Nishitani T. Measure-
The quality of X-ray CT image with high energy ment and analysis of neutron-induced alpha par-
X-ray sources more than 6 MeV is expected to be ticle emission double-differential cross section of
improved for artifact-less by using the information on carbon at 14.2MeV. J. Nucl. Sci. Technol.
energy spectra obtained by the technique. 2008;45: 103-115.
10. Nauchi Y, Iwasaki S. Convergence of unfolded
REFERENCES spectrum with response function for single radia-
01. Waggener RG, Blough MM, Terry JA, Chen D, tion based on Bayes' theorem. Nucl. Instrum.
Lee NE, Zhang S, McDavid WD. X-ray spectra Methods. 2014;A735:437-443.

JOURNAL OF RADIATION PROTECTION AND RESEARCH, VOL.41 NO.2 JUNE 2016 99