Annals of Nuclear Energy 75 (2015) 101–106

Contents lists available at ScienceDirect

Annals of Nuclear Energy

journal homepage: www.elsevier.com/locate/anucene

Technical note

Features of a subcritical nuclear reactor

Hector Rene Vega-Carrillo a,⇑, Isvi Ruben Esparza-Garcia a, Alvaro Sanchez b
a
Unidad Academica de Estudios Nucleares, Universidad Autonoma de Zacatecas, C. Cipres 10, Fracc. La Peñuela, 98068 Zacatecas, Zac., Mexico
b
Escuela Superior de Física y Matemáticas del IPN, Av. Instituto Politécnico Nacional, 07738 México, DF, Mexico

a r t i c l e i n f o a b s t r a c t

Article history: A subcritical nuclear reactor is a device where the nuclear-fission chain reaction is initiated and main-
Received 12 July 2014 tained using an external neutron source. It is a valuable educational and research tool where in a safe
Received in revised form 2 August 2014 way many reactor parameters can be measured. Here, we have used the six-factor formula to calculate
Accepted 4 August 2014
the effective multiplication factor of a subcritical nuclear reactor Nuclear Chicago model 9000. Using
Available online 24 August 2014
the MCNP5 code, a three-dimensional model of the subcritical reactor was developed to estimate the
effective multiplication factor, the neutron spectra, and the total and thermal neutron fluences along
Keywords:
the radial and axial axis. The MCNP5 results of the effective multiplication factor were compared with
Subcritical
Assembly
those obtained from the six-factor formula. The effective dose and the Ambient dose equivalent, at three
Nuclear reactor sites outside the reactor, were estimated; the Ambient dose equivalent was also measured and compared
keff with the calculated values.
Six-factors formula Ó 2014 Elsevier Ltd. All rights reserved.
Monte Carlo

1. Introduction
An important parameter in the design of a nuclear reactor is the
multiplication factor, k, that is, the amount on neutrons in any gen-
eration-to-the amount of neutrons in the previous generation ratio.
If the multiplication factor is less than one (k < 1) the nuclear reac-
tor is subcritical, which is also known as subcritical assembly. If the
reactor is infinite the k is named infinite multiplication factor, if the
reactor is finite the k is known as effective multiplication factor, keff
(Lamarsh, 1977).
Any nuclear system, like a nuclear reactor, array of nuclear fuel
during transport or storage, or even a nuclear war head, containing
nuclear materials must be analyzed to determine their criticality in
order to verify the proper safety conditions to avoid critical excur-
sions (Abderrahim et al., 2010; Pyeon et al., 2007).
Accelerator-driven subcritical reactors have been developed
and used as neutron sources, for energy production and for spent
fuel transmutation (Shahbunder et al., 2010).
The Universidad Autonoma de Zacatecas has a subcritical reac-
tor, UAZSNR, Nuclear Chicago model 9000. This is a heterogeneous
reactor, fueled with natural uranium with light water as moderator
and reflector (Maldonado et al., 2008; Papastefanou, 2004). The
system is similar to the ‘‘pickle barrel reactor’’ built at the New
York University (Dietrich, 1960).
⇑ Corresponding author.
E-mail address: fermineutron@yahoo.com (H.R. Vega-Carrillo).
To run the reactor a 0.185 GBq 239PuBe isotopic neutron (Vega-
Carrillo et al., 2009) is inserted in the center of a series of rings or
hexagonal layers of aluminum tubes. In each tube there are five
hollow rods of fuel or slugs. The main application of this type of
reactor is for teaching, training and research.
The value of the keff depends upon the amount and type of
nuclear fuel as well as the fuel geometry array and it is important
to determine its value to avoid criticality accidents (Endo et al.,
2011). The keff is the key parameter in the design or characteriza-
tion of subcritical assemblies (Papastefanou, 2004; Talamo et al.,
2011).
The aim of this work was to calculate the effective multiplica-
tion factor using the six-factor formula; as well as, to estimate
the neutron spectra along the reactor diameter and the keff using
Monte Carlo methods. The Monte Carlo methods were also used
to estimate de ambient dose equivalent and the effective dose
due to neutrons in three points around the reactor, these values
were compared with measurements done with a neutron area
monitor.
2. Materials and methods
The UAZSNR has 280 aluminum tubes which are arranged in a
hexagonal pattern. Each tube contains 5 metallic natural uranium
3 Ø  21.3 cm2 cylinders, or slugs, with 0.1 cm-thick aluminum
cladding. The reactor can be loaded with 312 aluminum tubes;
details of its design have been described in literature (Maldonado
et al., 2008; Papastefanou, 2004).

http://dx.doi.org/10.1016/j.anucene.2014.08.006
0306-4549/Ó 2014 Elsevier Ltd. All rights reserved.
102 H.R. Vega-Carrillo et al. / Annals of Nuclear Energy 75 (2015) 101–106
Fig. 1. View of the UAZSNR showing the lattice arrangement.
The reactor uses natural water as moderator and reflector, in
the center of the fuel array there is a tube that is used to hold
the isotopic neutron source. The fuel and the moderator are con-
tained in a 1.22 Ø  1.52 m2 stainless steel tank. In the Fig. 1 is
shown the subcritical reactor.
In this work, two methods were used to determine the keff. In
the first method, the six-factor formula was used and Monte Carlo
calculations were carried for the second method.
Also, Monte Carlo calculations were used to estimate the neu-
tron spectra in several points along the diameter in the middle
plane of the reactor. With the neutron spectra the total neutron flu-
ence and the thermal neutron fluence (E 6 0.414 eV) were also
estimated.
2.1. Using the six-factor formula to calculate the keff
The UAZSNR has 1400 fuel slugs, thus the keff was calculated
using 30, 90, 180, 300, 450, 630, 840, 1080, 1350, 1410 and 1620
fuel slugs. The effective multiplication factor, keff, was calculated
using the six factors formula (Papastefanou, 2004; Lamarsh,
1977; Vega-Carrillo, 2012) shown in Eq. (1). The calculations were
performed under MathcadÒ (Mathsoft, 1994).
keff ¼ g e p f PNL;th PNL;f
Here, g is the reproduction factor, e is the fast fission factor, p is the
resonance escape probability, f is the thermal utilization factor,
Fig. 2. Monte Carlo model of the subcritical reactor.
PNL,th and PNL,f are the thermal and the fast no leakage probabilities
respectively (Papastefanou, 2004; Lamarsh, 1977; Vega-Carrillo,
2012).
The reactor equation, shown in Eq. (2), for cylindrical geometry,
was solved using, the Wigner–Seitz method (Lamarsh, 1977),
through equivalent cells.
r2 U þ B2 U ¼ 0 ð2Þ
In this equation, U = f(r, z) is the neutron distribution along the
radius, r, and the reactor axial axis, z, and B2 is the Buckling. The
non-leakage factors were obtained using the two-group approxima-
tion (Papastefanou, 2004; Lamarsh, 1977; Vega-Carrillo, 2012).
In the assembly, neutrons are increased by the amplification
factor, l, which is calculated using Eq. (3).
1
l¼ ð3Þ
1  keff
The subcritical reactor power, P, was calculated using Eq. (4)
(Shahbunder et al., 2010; Papastefanou, 2004) where m is the aver-
age neutrons produced during the 235U fission (2.44), w is
3.124E(11) Watt-s/fission, and Q is the neutron source strength.
 
Q keff
P¼ : ð4Þ
m w 1  keff
2.2. Monte carlo calculations
A detailed model of the subcritical, as it is shown in Fig. 2, was
developed.
The MCNP 5 code with the ENDF/B-VI cross section library (X-5
Monte Carlo Team, 2003; Forster et al., 2004) was used, with the
model with 9 hexagonal rings, 270 aluminum tubes with 1350
slugs, to estimate the neutron spectra (UE(E)) and the total neutron
fluence, along the reactor diameter. Also, this model was used to
calculate the effective dose, E, and the Ambient dose equivalent,
H⁄(10), at three sites outside the reactor.
The amount of fuel slugs in the model was modified to estimate
the keff. The amount of hexagonal rings was increased from 1 up to
ð1Þ
10 hexagonal rings with 30, 90, 180, 300, 450, 630, 840, 1080,
1350, 1410 and 1620 fuel slugs, in order to compare the keff values
with those obtained with six-factor formula.
 H.R. Vega-Carrillo et al. / Annals of Nuclear Energy 75 (2015) 101–106 103

location the dose was measured 20 times and the average was cal-
culated. These values were compared with those obtained with the
Monte Carlo calculations.

3. Results and discussion

Fig. 3. Locations where the doses were calculated and measured.
2.3. Calculating and measuring the dose rates
The model with 270 tubes was used to estimate the ambient
dose equivalent, H_  ð10Þ, and the anteroposterior effective dose,
E_ AP , rates at three locations, A, B, and C, around the reactor, these
locations are shown in Fig. 3. For these calculations the neutron
fluence-to-dose conversion coefficients from ICRP 74 (ICRP, 1996)
were used. The Monte Carlo results were multiplied by the source
strength, Q = 1.01E(7) s1 (Vega-Carrillo et al., 2014).
Using a neutron area monitor LB 6411 (Burgkhardt et al., 1997),
the Ambient dose equivalent rate was measured while the reactor
was on, loaded with 1350 slugs in 270 aluminum tubes. At every
In Table 1 are shown the keff values calculated for the three
cases using the six-factor formula and the Monte Carlo methods.
The six-factor formula overestimates the keff probably due the
simplifications of the method and the quality of the data.
If a 3% error is assigned to keff obtained with the six-factor for-
mula and using 3r in the keff estimated with the MCNP5 code, the
differences are not significant.
In Fig. 4 are shown the keff value, calculated with both methods,
in function of the amount of uranium.
Monte Carlo calculations are in agreement with the keff reported
by Maldonado et al. (2008).
Both methods give the same keff when the amount of fuel varies
from 800 to 1900 kg of uranium. Below 800 kg and above 1900 kg
the differences between both methods are significant, probably
due to the simplifications done with the six-factor formula.
It can be noticed that regardless the method used in the calcu-
lations, as the amount of fuel is increased the criticality increases.
However, even with the largest fuel load used in the calculations
the assembly remains subcritical.
Using the case with 1350 fuel slugs the reactor amplification
factor is 5.77 while the reactor power is 61.12 pW per neutron
emitted by the 239PuBe source.

Table 1
Values of the keff calculated by both methods with the reactor with 1350, 1410, and 1620 fuel slugs.

Al tubes Fuel slugs Fuel mass [kg] keff calculated with the six-factor formula keff calculated with the MCNP5 code
270 1350 2430 0.8498 ± 0.02549 0.8268 ± 0.0007
282 1410 2538 0.8537 ± 0.02561 0.8296 ± 0.0006
324 1620 2916 0.8653 ± 0.02596 0.8374 ± 0.0006

1.0

0.8

0.6
keff

0.4

0.2 Monte Carlo

Six factors formula

0.0
0 500 1000 1500 2000 2500 3000 3500
Uranium mass [ kg ]

Fig. 4. keff, obtained by both methods, in function of uranium mass.

104 H.R. Vega-Carrillo et al. / Annals of Nuclear Energy 75 (2015) 101–106

100

r = 2.55 cm
10-1 r = 5.10 cm
r = 10.15 cm
r = 62 (outside)
[ n/cm2 - Δ u - Q ] 10-2
Source term

10-3

10-4

10-5

10-6
E Φ E(E)

10-7

10-8

10-9

10-10
10-10 10-9 10-8 10-7 10-6 10-5 10-4 10-3 10-2 10-1 100 101 102

Neutron energy [ MeV ]

Fig. 5. Nuclear spectra in the middle plane along the UAZSNR radius.

10-1
[ n/cm -Q ]
2

10-2
Fluence per source neutron

10-3

10-4

UAZ Subcritical nuclear reactor

10-5
Total
Thermal

10-6
-60 -40 -20 0 20 40 60
Radial axis [ cm ]

Fig. 6. Total and thermal fluencies along the reactor radial axis.

In Fig. 5 are shown the neutron spectra in 3 points in the middle The total and the thermal neutron fluencies, normalized to the
239
plane along the radius of the reactor with 1350 slugs. Here, is also PuBe source strength, Q, along the reactor radius are shown in
included the neutron spectrum outside the reactor and the relative Fig. 6.
239
PuBe neutron spectrum used as source term. Here, the last hexagonal ring of fuel ends at 44.5 cm, thus the
For the point closest to the source the spectrum keeps the shape next point beyond is in the water reflector, and the point beyond
of source term neutrons, as the distance is increased, from 0.1 up to 60 cm is outside the reactor. The radial distribution is in agreement
15 MeV the fission spectrum is more clearly defined. In the spectra with measurements reported in literature (Papastefanou, 2004).
can be noticed the thermal neutron spectrum, these spectra have The total neutron fluence distribution in the middle plane of the
approximately the same shape as those reported for subcritical reactor is shown in Fig. 7, while in Fig. 8 is shown the total neutron
assemblies (Talamo et al., 2011). fluence distributions in the upper plane of the reactor.
 H.R. Vega-Carrillo et al. / Annals of Nuclear Energy 75 (2015) 101–106 105

Fig. 7. Spatial distribution of neutrons in the reactor middle plane with 1350 fuel slugs.

Fig. 8. Spatial distribution of neutrons in the reactor upper plane with 1350 fuel slugs.

Table 2
Calculated and measured doses.

Location E_ AP [lSv/h] MCNP H_  ð10Þ [lSv/h] MCNP H_  ð10Þ [lSv/h] Measured _  ð10Þ
H Measured
H_  ð10Þ
MCNP

A 20.31 ± 0.61 24.29 ± 0.73 27.2 ± 1.06 1.12

B 12.84 ± 0.39 13.87 ± 0.42 14.32 ± 0.31 1.03
C 1.61 ± 0.05 1.74 ± 0.05 2.78 ± 0.10 1.60

The sharp peak in Fig. 7 is where the neutron source is located.
The neutron distribution in the middle plane is approximately five
orders of magnitude larger than the neutron distribution at the
upper plane, which is outside the reactor.
In Table 2 the calculated and measured dose rates are shown,
here the measured-to-calculated ratio of H⁄(10) is included.
Location A is above the last hexagonal fuel array, here the
neutron H_  measured is 12% larger that the H_  ð10Þ estimated
with the MCNP. At site B, located in the middle plane outside
the reactor the difference is only 3%. The largest difference,
60%, is to 1 m from the reactor container, near the middle plane.
At all locations the H_  ð10Þ measured is larger than the H_  ð10Þ
calculated; this is probably because in the MCNP model the
walls of the laboratory, and a large shield in the room, were
not included, their presence produces neutron room-return
(Vega-Carrillo et al., 2007a,b) that increases the neutron flux
and the H⁄(10). Another factor of the difference is that in the cal-
culations the fuel was modeled as fresh fuel.
106 H.R. Vega-Carrillo et al. / Annals of Nuclear Energy 75 (2015) 101–106
Using the criteria of 20 mSv/year as the maximum effective
dose for radiation workers, this means 10 lSv/h and for students
is 1 lSv/h, therefore the effective doses estimated at A, B and C
locations are used to define the working protocols around the
reactor.
4. Conclusions
The keff of the UAZSNR was calculated using the six-factor for-
mula and the Monte Carlo methods. With the Monte Carlo meth-
ods the neutron features of the reactor were estimated; also the
ambient dose equivalents, at two locations, were calculated. From
these calculations the main conclusions are:
(1) The six-factor formula tends to overestimate the keff in com-
parison to use Monte Carlo methods when the fuel is less
than 800 kg. When the amount of fuel is larger than
1900 kg the six factor formula tends to underestimate the
keff in comparison to Monte Carlo calculations.
(2) Using 1350 fuel slugs, the keff is 0.8268 ± 3%, l is 5.77. For a
239
PuBe source with a strength of 1.01E(7) s1 the reactor
power is 617 lW.
(3) Above the reactor, in the middle plane outside the reactor
and to 1 m from the vessel surface the ambient dose equiv-
alent rates are approximately 27, 14 and 3 lSv/h, while the
effective dose rates are 20, 13 and 2 lSv/h respectively.
(4) Near the reactor center the neutron spectrum above 0.5 MeV
is strongly influenced by the neutron spectrum of the source
used to run the reactor; this feature tends to vanish as the
radius increases.
(5) Inside the reactor, as the distance from the source increases
the amount of fast, epithermal and thermal neutrons
decreases, and the neutron spectra tend to be softer, where
the fast neutrons tend to diminish while the epithermal
and thermal tend to increase.
(6) The neutron distribution in the reactor middle plane is sym-
metric along the radius.
Acknowledgments
This work is part of the project Subcritical Reactors Design, par-
tially supported by the COZCyT. The second author thanks to
CONACyT for the scholarship to obtain the MSc degree.
References
Abderrahim, H.A., Baeten, P., De Bruyn, D., Heyse, J., Schuumans, P., Wagemans, J.,
2010. MYRRA, a multipurpose hybrid research reactor for high-end
applications. Nucl. Phys. News 20, 24–28.
Burgkhardt, B., Fieg, G., Klett, A., Plewnia, A., Siebert, B.R.L., 1997. The neutron
fluence and H⁄(10) response of the new LB 6411 rem counter. Radiat. Prot.
Dosim. 70, 361–364.
Dietrich, F.S., 1960. Subcritical reactor as a College facility. Am. J. Phys. 28, 309–314.
Endo, T., Yamamoto, A., Yamane, Y., 2011. Detected-neutron multiplication factor
measured by neutron source multiplication method. Ann. Nucl. Energy 38,
2417–2427.
Forster, R.A., Cox, L.J., Barrett, R.F., Booth, T.E., Briesmeister, J.F., Brown, F.B., Bull, J.S.,
Geisler, G.C., Goorley, J.T., Mosteller, R.D., Post, S.E., Prael, R.E., Selcow, E.C., Sood,
A., 2004. MCNP™ version 5. Nucl. Instrum. Methods Phys. Res., Sect. B 213, 82–
86.
ICRP, 1996. Conversion coefficients for use in radiological protection against
external radiation. International Commission on Radiation Protection
publication 74. Ann. ICRP 26, 199–200.
Lamarsh, J.R., 1977. Introduction to Nuclear Engineering. In: Reading Mass.
Addison-Wesley, USA.
Maldonado, G.I., Xoubi, N., Zhao, Z., 2008. Enhancement of a subcritical
experimental facility via MCNP simulations. Ann. Nucl. Energy 35, 263–268.
MathSoft, 1994. Mathcad Plus 5.0 User’s Guide. MathSoft Inc., Cambridge Mass,
USA.
Papastefanou, C., 2004. Measurement of neutron flux and albedo of water for
thermal neutrons with foils of indium in a subcritical nuclear reactor. J.
Radioanal. Nucl. Chem. 261, 671–678.
Pyeon, C.H., Hirano, Y., Misawa, T., Unesaki, H., Ichihara, Ch., Iwasaki, T., Shiroya, S.,
2007. Preliminary experiments on accelerator-driven subcritical reactor with
pulsed neutron generator in Kyoto University critical assembly. J. Nucl. Sci.
Technol. 44, 1368–1378.
Shahbunder, H., Ho Pyeon, Ch., Misawa, T., Lim, J.-Y., Shiroya, S., 2010. Subcritical
multiplication factor and source efficiency in accelerator-driven system. Ann.
Nucl. Energy 37, 1214–1222.
Talamo, A., Gohar, Y., Aliberti, G., Cao, Y., Smith, D., Zhong, Z., Kiyavitskaya, H.,
Bournos, V., Fokov, Y., Routkovskaya, C., Serafimovich, I., 2011. MCNPX, MONK,
and ERANOS analyses of the YALINA booster subcritical assembly. Nucl. Eng.
Des. 241, 1606–1615.
Vega-Carrillo, H.R., Manzanares, E., Iñiguez, M.P., Gallego, E., Lorente, A., 2007a.
Study of room-return neutrons. Radiat. Meas. 42, 413–419.
Vega-Carrillo, H.R., Manzanares, E., Iñiguez, M.P., Gallego, E., Lorente, A., 2007b.
Spectrum of isotopic neutron sources inside concrete walls spherical cavities.
Radiat. Meas. 42, 1373–1379.
Vega-Carrillo, H.R., Manzanares, E., Hernandez-Davila, V.M., Ramirez-Gonzalez,
Hernandez-Villasana, J., Ruiz-Chacon, A., 2009. Spectrometry and dosimetry of a
neutron source. Radiat. Eff. Defects Solids 164, 218–223.
Vega-Carrillo, H.R., 2012. Teoría de Reactores Nucleares. Editorial Académica
Española, Saarbrucken, Germany.
Vega-Carrillo, H.R., Hernández-Dávila, V.M., Rivera, T., Sánchez, A., 2014. Nuclear
and dosimetric features of an isotopic neutron source. Radiat. Phys. Chem. 95,
122–124.
X-5 Monte Carlo Team, 2003. MCNP-A General Monte Carlo N-Particle Transport
Code, Version 5, Los Alamos National Laboratory Report LA-CP-03-0245.