Radiation Measurements xxx (xxxx) xxx–xxx

Contents lists available at ScienceDirect

Radiation Measurements
journal homepage: www.elsevier.com/locate/radmeas

Reference thermal neutron field at KRISS for calibration of neutron detectors

Yun Ho Kima, Hyeonseo Parka,∗, Yong Kyun Kimb,∗∗, Jungho Kima, Jeongsoo Kanga,1
a
Korea Research Institute of Standards and Science, Daejeon 305–340, Republic of Korea
b
Department of Nuclear Engineering, Hanyang University, Seoul 133–791, Republic of Korea

A R T I C L E I N F O A B S T R A C T

Keywords: A reference thermal neutron field has been established at the Korea Research Institute of Standards and Science
Reference thermal neutron field (KRISS) by using a 241Am-Be neutron source and a high-purity graphite pile constructed by stacking graphite
Thermal neutron fluence blocks. The properties of the graphite blocks such as impurities, density, and dimensions were studied thor-
Neutron activation analysis oughly to understand the characteristics of the generated field. The energy spectrum and thermal neutron
Thermal neutron detector calibration
fractions were simulated with the Monte Carlo N-Particle eXtended code using measured physical parameters.
The neutron effective temperature was 308 K, and the fraction of thermal neutrons was approximately 95% at
the reference position of the thermal neutron field. The thermal neutron fluence rate was determined by
adopting the Westcott convention method based on neutron activation analysis using a gold foil. The Westcott
fluence rate for thermal neutron at the reference position was (2326.7 ± 8.4) cm−2s−1. The true thermal
neutron fluence rate at the reference position in the KRISS thermal neutron field was (2700 ± 29) cm−2s−1 (at
the reference date of June 30, 2014). The response of a spherical proportional counter with He-3 (SP9 neutron
detector) was evaluated in the newly established field as (3.083 ± 0.045) cm2 for the reference calibration
condition (a parallel neutron beam with a Maxwellian energy distribution having a most probable energy of
0.025 eV).

1. Introduction incidences (Nolte et al., 2015). The calibration constant or detector

A thermal neutron field is essential for calibrating thermal neutron
detectors used for neutron monitors and dose-equivalent meters. A
thermal neutron field can be generated by moderating fast neutrons
from various neutron sources using a material such as graphite. The
neutron energy distribution and fluence rate of the resulting neutron
field should be accurately evaluated for the precise calibration of
thermal neutron detectors. Although the thermal neutron fluence can
be measured with neutron activation analysis using activation detectors
(Williams and Gilliam, 2011), the neutron energy distribution must also
be evaluated for the precise measurement of thermal neutron fluence
and for the calibration of thermal neutron detectors. However, the
measurement of the energy distribution in the thermal region is diffi-
cult. In many cases, neutron transport calculations using Monte Carlo
codes such as the Monte Carlo N-Particle eXtended code (MCNPX) are
utilized to evaluate the energy distribution.
Other difficulties in calibrating thermal neutron detectors arise from
differences among actual thermal neutron fields, such as differences in
calibration spaces, energy distributions, and the direction of neutron
response should be evaluated under pre-agreed reference conditions.
The reference conditions used at the present study are assumed to be
those of parallel-incident neutron beams with a Maxwellian energy
distribution having a most probable energy of 0.025 eV. The differences
between the reference conditions and the actual calibration conditions
should be corrected. In actual conditions, the neutrons are not parallel-
incident, and the neutrons having an energy distribution with a 1/E
shape are still included in the thermal neutron energy region owing to
incomplete thermalization. In many cases, the space for calibration is
not sufficiently large to ignore backscatter thermal neutrons.
At the Korea Research Institute of Standards and Science (KRISS), a
thermal neutron field has been established with a high-purity graphite
pile and a 241Am-Be neutron source, and various field characteristics
were studied. In the present article, we detail the facility's construction
and the characteristics of the thermal neutron field (Section 2), the
measurement system and the determination of thermal neutron fluence
based on neutron activation analysis (Section 3, 4), and the calibration
of a neutron detector using the constructed facility (Section 5).

∗
Corresponding author.
∗∗
Corresponding author.
E-mail addresses: hyeonseo@kriss.re.kr (H. Park), ykkim4@hanyang.ac.kr (Y.K. Kim).
1
Present address: Korea Institute of Nuclear Safety, Daejeon 34142, Korea.

http://dx.doi.org/10.1016/j.radmeas.2017.10.001
Received 16 February 2017; Received in revised form 12 September 2017; Accepted 5 October 2017
1350-4487/ © 2017 Elsevier Ltd. All rights reserved.

Please cite this article as: Kim, Y.H., Radiation Measurements (2017), http://dx.doi.org/10.1016/j.radmeas.2017.10.001
Y.H. Kim et al.
2. Thermal neutron field at KRISS
2.1. Construction of graphite pile
A 120 cm (width) × 120 cm (height) × 140 cm (length) graphite
pile was constructed at KRISS, as shown in Fig. 1. To understand the
thermal neutron field, the properties of the graphite pile, including its
impurities, boron concentration, density, structure, and dimensions,
were thoroughly studied. In total, 239 graphite blocks (manufactured
by Morgan Advanced Materials plc.) were stacked on an aluminum
plate positioned 60 cm above the floor to reduce the scattering of
neutrons by the floor. The density of each block was measured by
weighing to obtain the density distribution of the graphite pile. The
average density was (1.786 ± 0.015) g/cm3.
To measure the boron concentration, 16 graphite chips (20 mm in
diameter and 2 mm in thickness) were prepared from various locations
of the graphite block, and their boron concentration was measured with
a cold-neutron prompt gamma-ray activation analysis (PGAA) instru-
ment in the NG7 beam line in the NIST Center for Neutron Research
(NCNR) at NIST (the National Institute of Standards and Technology,
U.S.A. (NIST) (Arif et al., 1993). The natural boron concentration was
(0.628 ± 0.028) μg/g. Other impurities such as natural lithium and
lead were tested with glow discharge mass spectrometry (GDMS)
(SHIVA Technologies, U.S.A.) and were found to be less than 0.1 ppm in
weight.
Fig. 2 shows a schematic of the graphite pile. The y-axis denotes the
axis from the neutron source to the reference position. The z-axis and x-
axis denote the vertical direction and horizontal direction, respectively.
An X.14-type 241Am-Be radionuclide neutron source (manufactured by
Amersham) was used; its emission rate was approximately 1.2 × 107
s−1 (Park et al., 2005). The neutron source was inserted at a distance of
35 cm from the surface along the y-axis and positioned at the center in
the vertical direction. The source was not positioned at the center of the
thermal pile in order to maintain sufficient neutron moderation in the
2
Radiation Measurements xxx (xxxx) xxx–xxx
Fig. 1. Construction process of the graphite pile.
small pile. As shown in Fig. 2 (b), seven wells (filled yellow green cir-
cles) with diameters of 2.4 mm and depths of 2.1 mm were prepared on
eight graphite blocks at various positions (boxes with diagonal shading)
for installing the activation foil. The reference position of the graphite
pile for the measurement of thermal neutron fluence was 70 cm from
the source along the y-axis.
The density, boron concentrations, impurities, and detailed struc-
ture inside the constructed pile were included in the modeling of the
thermal neutron field.
2.2. Monte Carlo simulations of thermal neutron field
MCNPX 2.7.0 code (Pelowitz, 2011) was utilized to model the
thermal neutron field in the graphite pile. The neutron source was
defined as a beryllium cylinder with a density of 1.75 g/cm3 positioned
inside a stainless-steel capsule. Neutrons were generated uniformly
inside the beryllium cylinder with such an energy distribution that the
energy spectrum of neutrons from the source capsule's surface re-
produced the neutron energy spectrum of a 241Am-Be source given in
ISO8529-1 (ISO, 2001). In this manner, the anisotropic angular de-
pendence of the 241Am-Be source was well reproduced too (Park et al.,
2005). Structural details, such as the insertion hole for the neutron
source and the activation wells in the pile, were incorporated precisely
in the modeling, in which the measured boron concentration and the
density of the graphite pile were considered. To account for the effects
of the surrounding materials, structures such as the wall of the ex-
periment room, the door, and the graphite support were also con-
sidered. The ENDF/B-VII.0 cross-section library and special S(α,β) data
(at T = 293 K) were used for neutron transport (Pelowitz, 2011).
Fig. 3 shows the calculated spectral fluence at the reference posi-
tion. The dotted curve shows the neutron energy spectrum at the re-
ference position in the activation hole of a graphite block. The curve
drawn with the solid line shows the energy spectrum at the same po-
sition with a cadmium case 1 mm in thickness. The spectra were aligned
Y.H. Kim et al. Radiation Measurements xxx (xxxx) xxx–xxx

Fig. 2. Schematic diagram of the graphite pile: (a) y-z plane at x = 60 cm and (b) x-y plane at z = 60 cm.

Table 1
Measurement and simulation results for gold foils activation in the graphite pile. 100-μm-
thick gold foils with 20-mm diameter were used for activation.

Distance (cm) Saturated activity (Bq·mg−1)

Measurement Simulationa Difference

(%)

10 4.392 ± 0.011 4.471 ± 0.046 +1.8%

30 2.903 ± 0.009 2.943 ± 0.030 +1.4%
50 1.453 ± 0.006 1.487 ± 0.015 +2.3%
70 0.642 ± 0.004 0.645 ± 0.007 +0.5%
90 0.230 ± 0.002 0.227 ± 0.003 - 1.3%

a
Uncertainties of simulation results were considered for uncertainties from neutron
emission rate of the source and MCNP calculation (statistics).
Fig. 3. Calculated neutron energy spectrum for the reference position. Solid line: neutron
fluence for air with a 1-mm-thick cadmium case; solid line with square symbol: neutron
responses within a difference of 3%. These results imply that the Monte
fluence for air without the cadmium case. (B denotes the emission rate of the 241Am-Be
neutron source.)
Carlo calculations for the thermal neutron field can be used to estimate
neutron energy spectra and to understand neutron transport in the
graphite pile. Details of the measurement system and procedure are
in the intermediate- and high-energy regions. This implies that the presented in the subsequent sections.
cadmium case only affected the thermal neutron transport, as expected,
while the case had no significant effect on non-thermal neutrons.
3. Measurement system for gold foil activity
Fig. 4 presents the thermal neutron fluence fractions compared to
the total neutron fluence for various positions relative to the source; (a)
Neutron activation analysis with gold foils was adopted for mea-
shows the distance dependence, and (b) shows the horizontal dis-
suring thermal neutron fluence inside the thermal neutron field. Gold
tribution from the reference position. At the reference position (70 cm
foil is one of the most widely adopted activation detectors, and the
from the source), the thermal neutron fluence was 95% of the total
thermal-neutron-capture cross section of 197Au has been evaluated
neutron fluence. At a position 90 cm from the source, the thermal
precisely with an uncertainty of 0.14% (Young, 2011).
neutron ratio was higher than 98%, while neutron fluence decreased
A gamma detection system consisting of a 7-in well-type NaI(Tl)
significantly. The horizontal distribution of thermal neutron fraction
scintillation detector (manufactured by SCIONIX, Holland) was utilized
shows deviations less than 1.5% from the thermal neutron fraction at
for measuring the activity of the gold foil. The detection efficiency of
the center position.
the NaI(Tl) detector was calibrated by measuring gamma rays from a
To verify the calculation results, gold foils with aluminum cases
gold foil irradiated in the reference thermal neutron field of the
were activated at different distances from the 241Am-Be source. To
National Physical Laboratory (NPL), Teddington, U.K. (Thomas and
ensure a high degree of activation, 100-μm-thick gold foils were chosen,
Kolkowski, 2005). The saturated activity of the foil was 334 kBq, and its
rather than 50-μm-thick ones. Measurements of the activated foils were
uncertainty was 0.1%, as certified based on a 4πβ-counting system at
compared with the calculation results based on MCNPX, as presented in
NPL. The detection efficiency for 411.8 keV gamma rays from the 50-
Table 1. The calculation results reproduced the detector (gold foil)
μm-thick gold foil was 0.7629 ± 0.0011. Major contributions to the

Fig. 4. (a) Thermal neutron fluence fractions for different distances from the source, and (b) the horizontal distribution of the thermal neutron fluence fraction relative to the reference
position.

3
Y.H. Kim et al. Radiation Measurements xxx (xxxx) xxx–xxx

uncertainty were the saturated activity of the gold foil (0.1%), the
branching ratio of 411.8 keV gamma rays (0.08%), and counting sta-
tistics (0.04%). The calibrated detection efficiency was used for the
different-thick gold foils by correcting self-absorption in the gold foils,
based MC calculations.

4. Determination of thermal neutron fluence rate

4.1. Neutron fluence derivation with Westcott convention method

The Westcott convention method (Westcott et al., 1958), which is

one of the most widely used convention methods, was adopted for
measuring thermal neutron fluence rate. The Westcott fluence rates
with v = v0 = 2200 m/s, which corresponds to T = 293 K, can be
Fig. 5. Fitting for the neutron energy spectrum.
obtained using the following equations, where φW , φW , th , φW , M , and
φW ,1/ E are the Westcott fluence rates for the total, thermal (below
cadmium cutoff energy), Maxwellian, and 1/E components, respec- Table 2 presents a summary of the parameters used for the evaluation of
tively (Thomas and Kolkowski, 2005): neutron fluence. Parameters related to the energy distribution of the
KRISS thermal neutron field are described in Section 4.2.
(D0 − D0 (Cd ) FCd ) R
φW = nv0 = ,
Gth gNσ0 R−1 (1)
4.2. Energy distribution of realistic thermal neutron field
(D0 − D0 (Cd ) FCd )
φW , th = nth v0 = ,
Gth gNσ0 (2) The thermal neutron fluence rate is determined as the sum of the
fluence rates for the Maxwellian component and 1/E component below
4r ⎞ the cadmium cutoff energy E(Cd). The neutron fluence rates for each
φW , M = nM v0 = φW ⎜⎛1 − ⎟,
πμ ⎠ (3) component can be determined by multiplying the average velocities of
⎝
the components of the Westcott fluence rates (Thomas and Kolkowski,
φW ,1/ E = n1/ E v0 = φW − φW , M , (4) 2005),
where vM v1/ E
φ th = φW , M + (φW , th − φW , M ) ,
v0 v0 (6)
D0 = reaction rate per mg for bare foil,
where φ th , vM , and v1/ E are the thermal neutron fluence rate and the
D0 (Cd ) = reaction rate per mg for foil with a cadmium filter,
average velocities of the Maxwellian and 1/E components, respectively.
μ = lower energy limit constant (EL = μkT) of 1/E neutrons,
The average velocities depend on μ and T, characterizing the neutron
and other parameters are summarized in Table 2.
energy distribution (Geiger and Van Der Zwan, 1966). μ and T were
estimated by fitting the whole calculated neutron energy spectrum
The cadmium ratio for gold foil can be determined by D/D(Cd), and
obtained from Monte Carlo simulations, as shown in Fig. 5. The fit
this value can be utilized to evaluate r from following equation (Geiger
function was the sum of the Maxwellian function and a 1/E function.
and Van Der Zwan, 1966).
The fit parameters were the height and the most probable temperature
D0 ⎡ Gth
+
Gr s0
⎤ (T) of the Maxwellian distribution as well as the height and the lower
RCd =
D0 (Cd )
= ⎣r T / T0 g
⎣
Gr
⎦ ⎡ FCd ( s0
g
−W )+ 1
K
⎦
⎤
(5)
energy limit (EL = μkT) of the 1/E function.μ and T at the reference
position inside the thermal neutron field were estimated to be
The detailed theory of the Westcott convention method has been (6.3 ± 0.7) and (308.0 ± 1.1) K from the fit, respectively.
presented in several other papers (Geiger and Van Der Zwan, 1966; The neutron energy distribution depends on the physical parameters
Walker et al., 1960; Westcott et al., 1958; ASTM, 2013; Boot, 1979). of the graphite moderator and neutron source. Among various

Table 2
Summary of the parameters used for the evaluation of neutron fluence.

Parameters Meaning Value Unit Remark

E(Cd) Cadmium cutoff energy 0.5 eV (Thomas and Kolkowski,

2005)
FCd Correction factor for attenuation of non-thermal neutrons in cadmium cover 1.013 MCNP modeling
g Departure of gold cross section from 1/v law in Maxwellian distribution 1.005 (Pritychenko and
Mughabghab, 2012)
Gth Thermal neutron self-shielding factor 0.934 MCNP modeling
Gr Resonance neutron self-shielding factor 0.280 (ASTM, 2013)
R Cadmium ratio for an ideal detector (σ ∼ 1/v), i.e., ratio of reaction rates of bare and cadmium- 495.6 (Geiger and Van Der Zwan,
covered foils, R= K
(r T/T0 ) , where T denotes the most probable temperature of Maxwell energy 1966)
distribution and T0 = 293 K
K πE (Cd) / E0 2.158 (Walker et al., 1960)
Coefficient for transmission of 1/E neutrons for cadmium (K= 4)
18
N Detector atoms per mg 3.056 × 10
r Relative intensity of the 1/E component 0.0043 From r T / T0 and T
s0 Westcott reduced resonance integral for gold 17.45 (Boot, 1979)
W Departure of gold cross section from 1/v law for 1/E component below E(Cd) 0.027 (Axton, 1963)
σ0 Thermal neutron cross section for gold 98.7 10−24 cm2 (Pritychenko and
Mughabghab, 2012)

4
Y.H. Kim et al. Radiation Measurements xxx (xxxx) xxx–xxx

Table 3 Table 4
Variations of the neutron effective temperature with respect to the source energy dis- Thermal neutron fluence rate in the KRISS thermal neutron field and parameters used for
tribution (252Cf, uniform energies distribution below 6 MeV and 11 MeV), source mate- evaluation.
rials (vacuum and beryllium cylinder), density ( ρ ± 3σρ) , boron concentrations
(CB ± 3σCB ), and humidity (assuming 0.1 wt% H2O) in the graphite pile. Parameter Unit Value Uncertainty (%)

Conditions T (K) Difference (%) μ Difference (%) D0 mg−1s−1 0.675 0.28

D0 (Cd) mg−1s−1 0.016 1.1
Reference parameters 308.0 0 6.33 0 T K 308 2.0
Neutron 252
Cf source 304.7 −1.07 6.16 −2.70 μ 6.3 50
sources Uniform n(E)dE 309.5 +0.50 6.86 +8.35 r 0.0043 2.5
(E < 11 MeV) RCd 41.93 1.1
Uniform n(E)dE 306.8 −0.40 5.77 −8.84 R 495.6 2.3
(E < 6 MeV) φW cm−2s−1 2331 0.36
Vacuuma 308.3 +0.12 4.74 −25.17 φW , th cm−2s−1 2327 0.36
Berylliuma 309.6 +0.53 6.87 +8.40 φW , M cm−2s−1 2322 0.37
Graphite ρ + 3σρ 306.5 −0.49 6.15 −2.89
φW ,1/ E cm−2s−1 8.88 25
proper- ρ − 3σρ 310.3 +0.74 5.43 −14.23
ties b vM / v0 1.157 1.0
CB + 3σC B 308.9 +0.28 6.75 +6.57
v1/ E / v0 3.265 14
CB − 3σCB 308.0 +0.02 6.30 −0.49
φM cm−2s−1 2686 1.1
0.1 wt% H2O 307.5 −0.17 7.29 +15.06
φ1/ E cm−2s−1 14 39
a φth cm−2s−1 2700 1.1
Neutrons with the energy distribution given in ISO8529-1 were generated uniformly
in a vacuum and with beryllium cylinders.
b
ρ and CB denote the measured density and boron concentration of graphite, respec-
tively.

parameters, the initial neutron energy distribution from the source,

neutron source structure and material, graphite density, boron con-
centration, and humidity (hydrogen contents) within the graphite pile
could change the energy distribution. MCNPX calculations with various
parameter sets were performed to obtain the energy distributions. The
temperature and lower energy limit of 1/E neutrons were obtained from
the fit and are summarized in Table 3. A maximum difference of 1.07%
in temperature was observed for neutron emissions from a 252Cf source,
and an uncertainty of 2% was applied to the neutron effective tem-
perature, considering the maximum difference. The uncertainty of μ Fig. 6. (a) Cavity of the graphite pile, and (b) the detector holder and carrier with a gold
was assumed as 50%, considering a maximum difference of 25%. activation foil.

4.3. Determination of thermal neutron fluence rate
In order to determine the thermal neutron fluence rate in the
thermal neutron field, two gold foils were activated at the reference
position. The first foil was placed in a 1-mm-thick aluminum case,
whereas the second was placed in a 1-mm-thick cadmium case, which
had the same dimensions as the aluminum case, to obtain the cadmium
ratio. The gold foils had a diameter of 20 mm and thickness of 50 μm,
with purity higher than 99.99%. Each gold foil was irradiated for about
10 days. The activities of the bare foil and cadmium-shielded foil were
measured. The reaction rates of the bare foil (D0) and cadmium-
shielded foil D0(Cd) were (0.6749 ± 0.0019) mg−1s−1 and
(0.0161 ± 0.0002) mg−1s−1, respectively. Dominant contributions to
the uncertainty were the counting statistics (0.22% in D0 and 1.1% in
D0(Cd)) and the detection efficiency (0.14%). Uncertainties from the
neutron irradiation time, decay time, and branching ratio of 411.8 keV
gamma rays were less than 0.1%. The cadmium ratio RCd was
(41.93 ± 0.46). The cadmium ratio of the ideal 1/v detector R was
495.6, which is considered sufficiently high to reduce the influence of
non-thermal neutrons.
The Westcott fluence rates and thermal neutron fluence rate in the
established field are summarized in Table 4 together with the para-
meters used for the evaluation. The thermal neutron fluence rate was
(2700 ± 29) cm−2s−1 (at the reference date of June 30, 2014).
5. Neutron detector response evaluation
5.1. Thermal neutron fluence rate inside cavity
Graphite blocks with dimensions of 120 cm (length) × 10 cm
(width) × 10 cm (height) positioned 70 cm from the source along the y-
axis were removed to create a cavity inside the graphite pile, as shown
in Fig. 6, for the neutron detector calibration.
The thermal neutron fluence rate at the cavity center was de-
termined with the procedure described in Section 4. The value of Gth for
gold foil of 50-μm thickness (diameter of 20 mm) in the cavity was
calculated to be 0.9503. This value is different from the value of Gth at
the reference position because the gold foils are installed with different
orientations: the surface of foil in the cavity faces the neutron source,
and the surfaces of foils at the reference position face the ceiling/floor
of the experiment room. Spectral parameters T and μ were estimated by
analyzing the neutron energy distribution at the cavity center;
T = 311.3 K and μ = 6.3 were obtained. The uncertainties of T and μ
were considered to be 2% and 50%, respectively, based on the in-
vestigations of the neutron energy distribution from different parameter
sets. The other correction factors and cross-sectional parameters listed
in Table 2 were used to determine the thermal neutron fluence rate
inside the cavity.
Two gold foils of 50-μm thickness (diameter of 20 mm) were acti-
vated at the cavity center using the 1-mm-thick aluminum and cad-
mium covers for about a week, respectively. D0 = (0.604 ± 0.002)
mg−1s−1 and D0(Cd) = (0.0186 ± 0.0004) mg−1s−1 were obtained.
The major uncertainty contributions were gamma counting statistics
(0.3% in D0 and 2.0% in D0(Cd)) and the gamma detection efficiency
(0.14%). RCd was (32.45 ± 0.66), which was lower than RCd =
(41.93 ± 0.46) at the reference position.
The realistic thermal neutron fluence rate (φth ) was 2372 cm−2s−1
(at the reference date of July 15, 2015); its fraction was approximately
94% based on the simulation. The uncertainty of the thermal neutron
fluence rate was 1.1%; the major contributions were T (1%), D0

5
Y.H. Kim et al. Radiation Measurements xxx (xxxx) xxx–xxx

(0.34%), and Gth (0.14%).

5.2. Detector response and correction factors

The response of a spherical proportional counter filled with 2 atm

3
He (SP9, Centronic Ltd.) was calibrated using the established thermal
neutron field. The SP9 detector consists of a spherical part (diameter of
3.3 cm) and a cylindrical stem (length of 14 cm and diameter of 1.3 cm,
approximately). The response of the thermal neutron detector (Rth ) was
determined using
ρbare − kCd ρCd
Rth = ,
φth (7)

where ρbare , ρCd , and kCd are the count rates from the detectors without/
with a 1-mm-thick cadmium cover and the correction factor for the
attenuation of neutrons with energies above the cadmium cutoff energy
(in an 1 mm-thick cadmium cover for the SP9 detector), respectively.
Rth describes the response of the detector only inside the specific
thermal neutron field used for the calibration. The response depends on
the shape of the neutron field. The response of the detector for reference
calibration conditions should be evaluated for the general use of the
detector. The reference conditions used at the present study presuppose
parallel-incident neutrons with a Maxwellian energy distribution (most
probable energy of 0.025 eV) (Nolte et al., 2015) and cadmium cutoff
energy of 0.5 eV. The difference of energy distribution and the incident
angles of neutrons should be considered. In many cases such as KRISS,
the calibration is performed in a limited space, rather than an open
space. Moreover, the self-absorption inside the detector changes the
neutron fluence at the reference position because it reduces the back-
scatter neutrons.
The correction factor for the difference of the energy distribution
(k1) was evaluated with the following equation (Nolte et al., 2015):
∞
∫0 (φM (E )/φM ) r (E ) dE
k1 = ,
E (Cd)
∫0 (φ (E )/ φ) r (E ) dE
where φM (E )/ φM and φ (E )/ φ are the relative spectral fluences for a
Maxwellian energy distribution with the most probable energy (E0) of
0.025 eV and for the KRISS thermal neutron field, respectively. In this
equation, r(E) is the detector's relative response (R(E)/R(E0), where R
(E) is the detector response for different neutron energies). The re-
sponses of the 2 atm SP9 detector were calculated for parallel-incident
neutrons with energies ranging from 1 meV to 100 MeV k1 = 1.026 was
obtained using the calculated responses, stemming mostly from the
neutron effective temperature (1.020) and the contribution of the 1/E
neutrons below the cadmium cutoff (others).
The correction factor for a different neutron incident angle (k2) was
obtained as follows. The SP9 detector has cylindrical symmetry with
respect to the stem axis, and the detector response significantly varies
with the neutron incident angle relative to the stem. The neutron cur-
rents with incident angles to the detector sphere at the cavity's center
were estimated to study the neutron incident angular distribution. The
detector sphere was divided by azimuthal angles relative to the stem
axis (x-axis based on the schematic in Fig. 2), and neutron currents at
the divided surfaces were calculated. The variations of neutron currents
Fig. 7. Installation of the SP9 detector for calibration in the cavity.
fluence inside the graphite pile includes the fluence of some neutrons
passing through the detection position more than once because of being
reflected by the graphite. If a detector is installed, these events are
reduced owing to interactions within the detector itself. The induced
difference of the thermal neutron fluence was estimated by calculating
the self-shielding factors of the SP9 detectors both in a vacuum and
inside the graphite pile. The ratio of the self-shielding effect in a va-
cuum to that in the graphite pile (k3) was 1.028. The uncertainties of
the normalized factors were estimated, considering the factors' varia-
tions due to different conditions, such as the neutron field character-
istics and the detector's properties.
5.3. Calibration results
The SP9 detector was inserted into the graphite pile cavity so that
the detector sphere was located in the central position of the cavity, as
(8)
shown in Fig. 7.
Count-rate measurements were performed with the bare SP9 de-
tector and the SP9 detector shielded by 1-mm-thick cadmium. Count
rates of ρbare and ρCd were (6732 ± 21) s−1 and (30.5 ± 0.2) s−1,
respectively. The uncertainties of the count rates were determined by
considering the influences from counting statistics, dead-time correc-
tions, and threshold variations. k Cd was determined to be 1.02 based on
MCNPX calculations, considering the attenuation of neutrons with
E > E(Cd). As a result, Rth = (2.825 ± 0.032) cm2 was obtained and
normalized to Rref for reference conditions by multiplying the normal-
ization factors. Rref was (3.083 ± 0.045) cm2, as listed in Table 5.
Major contributions to the uncertainty were the thermal neutron flu-
ence in the cavity (1.1%) and the normalization factors of k1 (0.50%)
and k3 (0.68%).
Table 5
Response of the 2 atm SP9 detector and parameters used for calibrationa.
Parameter Unit Value Uncertainty (%)

for different incident surfaces were within 3% relative to the average ρbare s−1
6732 0.32
neutron current. The calculation results showed that the neutrons en- ρCd s−1 30.5 0.47
tered the detector with an almost isotropic angular distribution. kCd 1.02 2.0
Therefore, a correction factor for neutron incident angles, k2, was ob- φth cm−2s−1 2372 1.1
tained by calculating the detector responses for a parallel beam (Rparallel ) Rth cm2 2.825 1.2
and an isotropic beam (Riso ). In addition, k2 (Rparallel / Riso ) = 1.034 was k1 1.026 0.50
k2 1.034 0.34
estimated to be very close to the value of 1.033, known from a study by
k3 1.028 0.68
the National Metrology Institute of Japan (NMIJ) (Nolte et al., 2015). Rref cm2 3.083 1.5
The last correction factor (k3) is for the self-absorption of the de-
tector. The effective thermal neutron fluence rate for a detector position a
k1, k2 , and k3 represent the normalization factors for the neutron energies, incident
can vary owing to self-shielding effects of the detector. Thermal neutron angles, and self-absorption of the detector.

6
Y.H. Kim et al.
6. Summary and conclusions
A reference thermal neutron field using a high-purity graphite pile
and a 241Am-Be source has been established at KRISS. A graphite pile
with dimensions of 120 cm (width) × 140 cm (length) × 120 cm
(height) was constructed by stacking graphite blocks. Monte Carlo si-
mulations were performed in light of the detailed structure and
knowledge of the impurities and densities of the graphite blocks to
study the characteristics of the KRISS thermal neutron field.
The thermal neutron fluence rate at the reference position was de-
termined by adopting the Westcott convention method with gold foil.
The Westcott fluence rate for thermal neutron was (2326.7 ± 8.4)
cm−2s−1. The neutron effective temperature and thermal neutron
fraction were 308 K and 95%, respectively. The realistic determined
thermal neutron fluence rate was (2700 ± 29) cm−2s−1.
The response of a 2 atm SP9 neutron detector was calibrated by
measuring both thermal neutron fluence and count rates from the de-
tector inside the graphite pile cavity. At the center of the cavity, the
neutron effective temperature and thermal neutron fraction were 311 K
and 94%, respectively. The realistic thermal neutron fluence rate was
(2372 ± 26) cm−2s−1. The detector responses for the KRISS thermal
neutron field and the reference calibration condition were determined
as (2.825 ± 0.032) cm2 and (3.083 ± 0.045) cm2, respectively.
Acknowledgements
This work was supported by the Korea Research Institute of
Standards and Science under the project “Development of measurement
standards for radiation” [17011059].
Radiation Measurements xxx (xxxx) xxx–xxx
References
Arif, M., Dewey, M.S., Greene, G.L., Snow, W.M., 1993. Facilities for fundamental neutron
physics research at the NIST cold neutron research facility. J. Res. Natl. Inst. Stand.
Technol. 98, 135–144.
ASTM, 2013. Standard Test Method for Determining Thermal Neutron Reaction Rates and
Thermal Neutron Fluence Rates by Radioactivation Techniques. http://dx.doi.org/
10.1520/E0262-13.
Axton, E.J., 1963. An absolute calibration of the National Bureau of Standards thermal
neutron flux. J. Res. NBS. A. Phys. Ch. 67A (3), 215–221.
Boot, S.J., 1979. Calibration of reference thermal and epithermal neutron flux densities
from the GLEEP reactor at AERE Harwell. Nucl. Instrum. Meth. 164, 513–519.
Geiger, K.W., Van Der Zwan, L., 1966. Slowing down spectrum and neutron temperature
in a thermal neutron flux density standard. Metrologia 2 (1), 1–5.
ISO, 2001. Neutron Reference Radiations for Calibrating Neutron Measuring Devices Used
for Radiation Protection Purposes and for Determining Their Response as a Function
of Neutron Energy. ISO 8529–1.
Nolte, R., Böttger, R., Chen, J., Harano, H., Thomas, D.J., 2015. International key com-
parison of thermal neutron fluence measurement—CCRI(III)-K8. Metrologia 52.
Park, H., Choi, K.-O., Lee, J.-M., Lee, K.B., Hahn, M.S., Kralik, M., 2005. Absolute mea-
surement of the neutron emission rate with a manganese sulphate bath system. J.
Korean Phys. Soc. 47 (4), 603–609.
Pelowitz, D.B., 2011. In: MCNPX User's Manual, Version 2.7.0, LA-CP-11–00438.
Pritychenko, B., Mughabghab, S.F., 2012. Neutron thermal cross sections, Westcott fac-
tors, resonance integrals, Maxwellian averaged cross sections and astrophysical re-
action rates calculated from the ENDF/B-VII.1, JEFF-3.1.2, JENDL-4.0, ROSFOND-
2010, CENDL-3.1 and EAF-2010 evaluated data libraries. Nucl. Data Sheets 113,
3120–3144.
Thomas, J.D., Kolkowski, P., 2005. Thermal fluence and Dose Equivalent Standards at
NPL. DQL RN008.
Walker, W.H., Westcott, C.H., Alexander, T.K., 1960. Measurement of radiative capture
resonance integrals in a thermal reactor spectrum, and the thermal cross section of
Pu-240. Can. J. Phys. 38, 58–79.
Williams, J.G., Gilliam, D.M., 2011. Thermal neutron standards. Metrologia 48, 254–262.
Westcott, C.H., Walker, W.H., Alexander, T.K., 1958. Effective cross sections and cad-
mium ratios for the neutron spectra of thermal reactors. In: Proc. Second Intern. Conf.
Peaceful Uses at. Energy, Geneva. a/Conf. 15/P/202.
Young, P. G., 2011. ENDF/B-VII.1. Data for 197Au, https://www-nds.iaea.org/exfor/
endf.htm.