Experimental studies and simulations of spallation
neutron production on a thick lead target
M Majerle1 , J Adam1 , P Čaloun1 , S A Gustov2 , V Henzl1 ,
D Henzlová1 , V G Kalinnikov2 , M I Krivopustov2 , A Krása1 ,
F Křı́žek1 , A Kugler1 , I V Mirokhin2 , A A Solnyshkin2 ,
V M Tsoupko-Sitnikov2 , V Wagner1
1
2
Nuclear Physics Institute of CAS, 250 68 Řež, The Czech Republic
Joint Institute for Nuclear Research Dubna, 141980, Dubna, Moscow Region, Russia
E-mail: [email protected]
Abstract. An intensive beam of 660 MeV protons from the Dubna Phasotron was directed
towards a thick, lead target (not surrounded by shielding or neutron reflector) for 10 minutes.
Detectors and iodine samples were placed around the target. The neutron field and the
transmutation of 129 I were studied by the Activation Analysis Method.
MCNPX v.2.4.0 was used to simulate the experimental setup. The results of the simulations
were compared to the experimental values, and the influence of the setup parts to the neutron
field was explored.
1. Introduction
Accelerator Driven Systems are the combination of a classical reactor with an accelerator. The
idea of the ADS was born in Los Alamos in the 50’s, then taken up by Tolstov in the 80’s [1],
later by Bowman in 1992 [2], and Rubbia in 1993 [3]. ADS produce a large number of neutrons
in the spallation process, and introduce them into a sub-critical reactor assembly, where they
produce energy by fission of fissionable material, extra neutrons are used to produce fuel from
232 Th, and/or to transmute long-lived nuclear waste to stable isotopes, or to short-lived isotopes
which decay to stable ones.
The experiments with simplified ADS setups were performed on accelerators at the JINR
(Joint Institute for Nuclear Research, Dubna, Russia) [4]. The targets were lead, tungsten, and
uran cylinders (in some experiments surrounded with paraffin moderator or natural U blanket).
The physical aim of the experiments was to study nuclear processes in the target, to inquire into
the process of transmutation in radioactive samples, to measure the heat production, etc. Mostly,
the Activation Analysis Method was used to measure the neutron field and the production rates
of the produced isotopes.
The Phasotron experiment is an example of this work. The experiment consisted of a bare,
lead target irradiated with relativistic protons. The proton beam and the produced neutron field
were measured with the set of activation detectors in the form of thin foils. Samples of iodine
were put in the neutron field in order to study transmutation possibilities in iodine. Experimental
yields of produced elements were compared with the yields calculated with MCNPX, and causes
of discrepancies were explored.
Figure 1. The layout of the Phasotron experimental setup. Longitudinal and transverse crosssections.
2. The Phasotron experiment
The Phasotron setup consisted of a lead cylinder (radius 4.8 cm, length 4×12.1cm) at which an
intensive (1013 protons/s), high energy (660 MeV) proton beam was directed for 10 minutes. The
setup was placed in a long, narrow corridor, bounded by concrete walls (Figure 1). The corridor
was 20 m long, 2 m wide and high, and the target was placed 5 m before the back corridor
wall in the middle of the corridor, 1.2 m from the floor. Other components of the setup were
the iron table, where was placed the setup, a metal beam tube and a metal beam stopper (full,
10 cm thick, iron cylinder, placed 1 m behind the target, on the same axis as the beam tube).
To measure the proton beam, neutron field, and transmutation rates for iodine isotopes we used
the Activation Analysis method. We placed the detectors/samples in the neutron/proton field.
During the irradiation, neutrons and protons induced different nuclear reactions in them and
produced new radioactive nuclei. After the irradiation the quantities of produced isotopes were
determined from the gamma spectra of activated detectors.
2.1. The beam
The beam intensity and size were measured with two independent methods: with a wire chamber,
and with a set of Al and Cu foils. The results from both methods were in good agreement. The
wire chamber measured the flux of protons to be 1013 protons/s with the horizontal and vertical
diameters of the beam 3.2 cm and 3.8 cm. The beam profile was approximated with a Gaussian
profile, so the word diameter shall be considered as the border inside which were most protons.
A set of 8×8 cm2 Al and Cu foils, and another segmented set of five 2×2 cm2 foils (in the
shape of the cross) were put in the beam, ca. 10 cm in front of the target. After the irradiation
we measured the yields of produced elements in the foils by means of the γ-spectroscopy. We
interpolated experimental cross-sections [10] for reactions with protons for 660 MeV protons,
and calculated the integral flux of protons that passed through the foils. Total of 1.58×1015
protons (systematic and statistic errors are 6%) passed 8×8 cm2 foils. Five 2×2 cm2 foils, with
one foil on the central axis and four others around, showed that the beam diameter was roughly
4 cm, and shifted upwards for ca. 1 cm.
2.2. The neutrons
Three types of detector foils of the square shape were used to study neutron field along the
target: Al, Au (2×2 cm2 , 50 µm), and Bi (2.5×2.5 cm2 ,1 mm) foils placed on the top of the
target along its length. In the foils neutrons produced radioactive isotopes via reactions (n, γ)
or (n, xn) (see Table 1). After the irradiation, we measured the γ-spectra of the foils and
determined the yields of produced isotopes.
Table 1. Observed reactions, their thresholds, and half-lifes of products in Al, Au, and Bi foils.
Values in the table are determined by the mass difference calculated with QTOOL [12]. For
the reaction 27 Al(n, α)24 Na the threshold value should be modified for the contribution of the
Coulomb barrier. The modified value is 5.5 MeV.
Material
Reaction
Product
27 Al
(n, α)
24 Na
(n, γ)
(n, 2n)
(n, 4n)
(n, 5n)
(n, 7n)
198 Au
(n, 4n)
(n, 5n)
(n, 6n)
(n, 6n)
(n, 7n)
(n, 8n)
206 Bi
197 Au
209 Bi
196 Au
194 Au
193 Au
191 Au
205 Bi
204 Bi
203 Bi
202 Bi
201 Bi
Threshold (MeV)
3.2
Half-life
14.9 h
0
8.1
23.2
30.2
45.7
2.7
6.2
1.6
17.6
3.2
d
d
d
h
h
22.6
29.6
38.1
45.2
54.0
61.4
6.2
15.3
11.2
11.7
1.7
1.8
d
d
h
h
h
h
From the yields of produced isotopes we calculated the numbers of produced atoms of isotope
A per 1 g of the detector material and per 1 incident proton. This value is called the production
rate B(A) [5, 6], and is defined as:
B(A) = (number of A-atoms produced)/[(1 g sensor)·(1 primary proton)].
(1)
The experimental production rates against the position along the target are plotted in Figures
2, 3, and 4 for all three sets of detector foils (errors at the graphs are only statistical errors;
systematic errors, should contribute another 4% (main part is the error in the detector efficiency
measurements). Systematic errors are the same for different foils and do not change the shape
of the distribution, they are only replacing it upwards or downwards). All graphs show a specific
shape: the maximum at around 10th cm, and the point near 30th cm, where the neutron field
starts to decrease faster - spallations stop there. The exception is the graph for 198 Au, which
is produced via (n, γ) reaction by low-energy neutrons (thermal, epithermal, and resonance
neutrons), and shows a flat distribution. The results for Bi foils are preliminary. Near the end
of the target the foils were weakly activated and the statistical errors reach 5%, anywhere else
they are lower. The other possible error comes from the integral of the proton beam, which is
determined with the 6% accuracy. Altogether, the worst results for neutron field are precise up
to 10%, but in most cases the precision is ca. 5%.
2.3. Iodine samples
High beam intensity and short irradiation time enabled us to observe and measure the production
rates of higher threshold reactions - (n, 5n), (n, 6n),... in four iodine samples. Two samples
B [g-1 proton-1]
1E-5
1E-6
Na-24
1E-7
1E-8
0
10
20
30
40
Figure 2. B-values for 24 Na in Al foils along
the target.
50
Distance along the target [cm]
1E-06
1E-07
-1
Au-198
Au-196
Au-194
Au-193
Au-191
Bi-206
Bi-205
Bi-204
Bi-203
Bi-202
Bi-201
1E-06
-1
1E-05
B [g proton ]
1E-05
-1
-1
B [g proton ]
1E-04
1E-07
1E-08
1E-08
0
10
20
30
40
50
Distance along the target [cm]
Figure 3. B-values for different isotopes in
Au foils along the target.
0
10
20
30
40
50
Distance along the target [cm]
Figure 4. B-values for different isotopes in
Bi foils along the target (preliminary results).
contained natural 127 I isotope, and other two the mixture of 127 I and 129 I from the standard
radioactive waste. The samples were placed at the 9th cm and 21st cm along the target. The
isotopes produced in higher threshold reactions are far away from the line of stability and have
short lifetimes, and we needed to measure them in series of short measurements immediately
after the experiment. We could determine the yields of produced isotopes up to 118 I with the
accuracy of 10%. The decay products of iodine isotopes up to 116 I were detected. The yields of
produced isotopes for 129 I in the mixture of 127 I and 129 I were calculated with the substraction
of 127 I contribution, which was determined in 127 I samples.
The graphs in Figures 5 and 6 show the production rates of measured iodine isotopes at the
9th cm and the 21st cm for 127 I and 129 I. The production rates for iodine isotopes are comparable
with each other and with the production rates for other elements (10−6 g−1 proton−1 > B > 10−7
g−1 proton−1 ).
3. MCNPX simulations
Nowadays, there is a great motivation towards improving the precision of the simulation codes
which might be once used to simulate ADS systems. Experiments like the Phasotron one
can help in estimating the precision of simulations and finding the causes for discrepancies.
We simulated our experiment with MCNPX, and compared the calculations with experimental
values. Simulations offer a better insight into the experiment, a lot of things that cannot be
seen experimentally can be calculated. We used a set of simulations to see how the experimental
systematic error depends on setup uncertainties. Further, we compared the experimental results
with simulated values to test the accuracy of the MCNPX simulation code.
1E-05
1E-06
1E-06
-1
B [g proton ]
I-127
I-129
1E-07
1E-08
I-127
I-129
1E-07
-1
B [g-1 proton-1]
1E-05
1E-08
1E-09
1E-09
I-130 I-128 I-126 I-124 I-123 I-121 I-120 I-119 I-118
I-130 I-128 I-126 I-124 I-123 I-121 I-120 I-119 I-118
Isotope
Isotope
Figure 5.
in 127 I and
9th cm.
B-values for different isotopes
Samples were placed at the
129 I.
Figure 6.
in 127 I and
21st cm.
B-values for different isotopes
Samples were placed at the
129 I.
The latest production relase of MCNPX (version 2.4.0) was used to simulate the experimental
setup. Nuclear reactions of incident protons with material, the transport, and further reactions
of secondary particles are implemented in the code. The particles that cross a certain surface can
be sampled to get the neutron spectrum. This spectrum can be further convoluted with crosssections for nuclear reactions to get B-values, which we compare with the experimental values.
A satisfying agreement between experimental and simulated B-values shows that simulated
neutron spectrum corresponds to the experiment.
3.1. Simulating the setup
A simplified model of our setup - a bare, lead target with a narrow, homogenous, centrally
impinging proton beam - was first used for calculations. Most calculated values were in good
agreement with the experimental data, discrepancies were up to 20%. The simulated values
for 198 Au do not agree with the experimental flat distribution along the target, but show a
similar distribution as the neutron field for other isotopes, which is a few orders lower than
the experimental values. Isotope 198 Au is produced by low energy neutrons, and the flat
experimental distribution means that the field of low energy neutrons is homogenous along
the target. That is possible if high energy neutrons from the target are moderated in concrete
walls and partly reflected back, producing a low energy neutron field, which totally overcomes
the neutron field from the target. Another calculation with walls confirmed that this assumption
is true. The Figure 7 shows the ratio between experimental and simulated values for some of
produced isotopes, and the Figure 8 shows the calculated neutron spectrum along the target.
For these calculations, the Gaussian profile and the displacement of the beam were taken into
account, as well as concrete walls and iron components in the corridor. Generally, for most similar
experiments that we have analyzed, the MCNPX simulations underestimate the production rates
near the end of the target. This is also seen in the case of this experiment (Figure 7).
3.2. The influence of different parameters
All experimental conditions are not always precisely known (walls and metal components around
the target, beam parameters, ...). This determines the systematic error of our experimental
results, and simulations help us to estimate its value. How accurate are simulations in describing
the experiment depends on other additional factors (reactions with protons, choice of crosssection libraries, INC model, ...). Series of simulations are used to estimate the influence of
these parameters on the comparison between simulated and experimental values. Such series
are done with two identical setups but one slightly changed parameter, the results are then
2
Ratio sim/exp
1,8
Neutrons per incident proton
1
1,6
1,4
0.01
Au-198 (f4)
Au-196 (htape3x)
Au-194 (htape3x)
Na-24 (htape3x)
1,2
1
1e-04
1e-06
0
0,8
10
20
0,6
30
0,4
0
10
20
30
40
40
Distance along the target
50
50
60
100
1
1e-04
1e-06
1e-08
Neutron energy
Distance along the target [cm]
Figure 7. The comparison of simulated
production rates with experimental data for
the isotopes, produced in the Al and Au
foils. F4 and HTAPE3X are two methods of
calculating B-values.
0.01
Figure 8. The neutron spectrum along the
target. Homogenous low energy neutron field
and high energy neutron field with the specific
shape.
compared to estimate the influence of the changed parameter.
3.2.1. The concrete walls The influence of concrete walls around the target is discussed in the
previous section. Walls produce a homogenous field of low energy neutrons without changing the
field of high energy neutrons, what was confirmed with a set of simulations with and without the
walls. Figure 9 shows that the walls have no influence on threshold reactions, and Figure 10 shows
the difference between the production rates for the non-threshold reaction 197 Au(n, γ)198 Au when
we take in account the concrete walls without them. The 2 cm thick iron table on which the
target is placed, the beam stopper made of iron after the target, and the iron beam tube were
added to the setup description, and showed to have a negligible influence on the neutron field.
1E-4
1.05
1.00
Au-196
0.95
0.90
0
10
20
30
40
Distance along the target [cm]
B [g-1 proton-1]
Ratio no_walls/walls
1.10
1E-5
walls
no walls
1E-6
1E-7
1E-8
0
10
20
30
40
50
Distance along the target [cm]
Figure 9. The ratios between the calculated
production rates for Au-196 for the cases with
and without concrete walls.
Figure 10. The calculated production rates
for Au-198 for the cases with and without
concrete walls.
3.2.2. Beam parameters The beam profile and its displacement at our experiments are never
exactly known, on the other hand these parameters influence results significantly. Accelerator
beams are usually approximated with a Gaussian profile with different FWHM in x and y
directions. We measured the beam profile and displacement (Section 2.1) with the wire chamber
and five activation detectors, the accuracy of the results is ca. 0.5 cm. To know to which point
such inaccuracy matters, we did simulations with different beam thicknesses and profiles, and
with displaced beams. The simulations with 6 mm, 6 cm thick beams and a beam with the
Gaussian profile (all beams were directed to the center of the target) showed that beam profile
cannot influence the results for more than 5%. On the other hand, the simulation with a
homogenous, for 1 cm upwards displaced beam (beam diameter was 6 cm) gave for the yields of
produced isotopes for up to 40% higher values.
3.2.3. Reactions with protons Primary and secondary protons react with our detectors and
produce small yields of the same isotopes as neutrons. MCNPX does not contain libraries for
these reactions, so we needed to convolute simulated proton spectra with cross-section values
found in other libraries. Near the 30th cm the proton influence reaches its maximum : ca. 15%
of 24 Na and 196 Au are produced with protons at this point. For 194 Au this number is 30%.
3.2.4. The choice of Intra-Nuclear Cascade model and cross-section libraries Three IntraNuclear Cascade models are implemented in our version of MCNPX: BERTINI, CEM, and
ISABEL. The results from these three models differ up to 15%. In Figures 11 and 12 are the
ratios between the calculated production rates with different models along with the statistical
errors. B-values can be calculated with built-in cross-section library (ENDF/B-VI [11]), or the
calculated neutron spectra (SSW with HTAPE3X) can be convoluted with the cross-sections
from other libraries (EXFOR [10]). The choice of the cross-section library influences B-values
for up to 20%. We cannot say, which combination of INC model and cross-section library is the
right one, together the differences from model/library choice are ca. 30%.
1,3
1,3
1,2
1,2
1,1
Au-196
Au-194
Na-24
1,0
ratio
ratio
1,1
Au-196
Au-194
Na-24
1,0
0,9
0,9
0,8
0,8
0,7
0,7
0
10
20
30
40
50
Distance along the target [cm]
Figure 11.
The ratios between the
calculated production rates with the use of
the CEM INC model and the BERTINI INC
model.
0
10
20
30
40
50
Distance along the target [cm]
Figure 12.
The ratios between the
calculated production rates with the use of
the ISABEL INC model and the BERTINI
INC model.
The cause for the experimental uncertainties are mainly not exactly known beam parameters
(this is the systematic experimental error, which does not change the shape of the distribution,
but is replacing it upwards or downwards) - ca. 20% for 0.5 cm inaccuracy. As the proton
influence can be calculated and taken into account, the differences coming from the choice of the
INC model/cross-section libraries are setting the limits for the accuracy of simulations - 30%.
Experimental errors are smaller or in the range of the differences due to model/library choice,
and our experimental data can be used to test the simulation code. For most produced isotopes,
discrepancies between experiment and simulation have specific trends, which we are currently
trying to explain.
3.3. Number of produced neutrons per one incident proton
An important parameter of our setup is how many neutrons are in average produced by one
incident proton. This parameter can be read from MCNPX output file. In average, per one
proton 10 neutrons are produced in spallation reactions, and 5 more with (n,xn) reactions or
are reflected from the walls. The sum is 15 neutrons per one incident proton.
4. Conclusion
The Phasotron experiment is for its simplicity an ideal example for benchmark codes testing. The
experimental results were determined by means of Activation Analysis Method with the accuracy
better than 5% (systematic error for not accurately measured position of the beam is additional
20%). Therefore, we can use the experimental values to test different models/libraries, as the
differences between simulations with different models/libraries are in this range. The production
rates for every isotope with known cross-sections of production reaction can be calculated, either
with F4 card (if cross-section libraries are included in MCNPX), or with the convolution of
the neutron spectra (HTAPE3X) with our cross-section values. For some isotopes we have
cross-section libraries for reactions with neutrons and protons, for some only for reactions with
neutrons, and for some we have none. For isotopes, produced in iodine samples, cross-sections
for reactions with neutrons or protons do not exist.
The experiences gained at Phasotron experiment are used in analyzing data and
simulating more complicated experiments (especially ”Energy plus Transmutation” setup [7],[8]).
Complicated experimental setups are not as well described by MCNPX as the simple Phasotron
experiment, however, they show the deficiencies of the code better. Analysis of the influence
of different parameters for other experiments confirms the results obtained at the Phasotron
experiment.
A very useful tool when simulating was the use of parallel processing. We found out that for
all calculations of similar experiments we can profit from a small cluster of computers described
in [9].
Concerning this experiment, we plan to evaluate all results, and to test some other calculation
codes on them - FLUKA [13], GEANT4 [14], CASCADE-04 [15]. In general, we will carry on
with experiments on various spallation targets to provide as much experimental data as possible
to help improving the accuracy of MCNPX and similar codes.
Acknowledgments
The authors are grateful to the staff of the Dubna Phasotron accelerator for providing a good
proton beam for our experiment.
These experiments were supported by the Czech Committee for collaboration with JINR
Dubna. This work was carried out partly under support of the Grant Agency of the Czech
Republic (grant No. 202/03/H043) and ASCR K1048102 (the Czech Republic).
[1] K D Tolstov 1989 Some aspects of accelerator breeding JINR preprint 18-89-778
[2] C D Bowman et al 1992 Nuclear energy generation and waste transmutation using an accelerator-driven
intense thermal neutron source Nucl. Inst. and Meth. in Phys. Res. A 320 336-367
[3] C Rubbia et al. 1995 Conceptual design of a fast neutron operated high power energy amplifier CERN preprint
AT-95-44
[4] A Krása et al 2005 JINR preprint E1-2005-46
[5] J Adam et al 2003 First Nuclear Activation Experiments Using the New Accelerator Nuclotron in Dubna
Kerntechnik 68 214
[6] W Westmeier et al 2005 Transmutation experiments on 129 I, 139 La, and 237 Np Using the Nuclotron Accelerator
Radiochimica Acta 93 65-73
[7] M I Krivopustov et al 2004 JINR preprint El-2004-79
[8] V Wagner et al 2004 Proc. Baldin ISHEPP XVII (Dubna, Russia) Experimental Studies of Spatial
Distributions of Neutrons Produced by Set-ups with Thick Lead Target Irradiated by Relativistic Protons
[9] M Majerle et al 2005 Proc. M&C 2005 (Avignon, France) MCNPX Benchmark Tests of Neutron Production
in Massive Lead Target 328
[10] http://www.nndc.bnl.gov/exfor/exfor00.htm
[11] http://www.nndc.bnl.gov/exfor/endf00.htm
[12] http://t2.lanl.gov/data/qtool.html
[13] http://www.fluka.org/
[14] http://wwwasd.web.cern.ch/wwwasd/geant4/geant4.html
[15] H Kumawat, V S Barashenkov 2005 Development of the Monte Carlo model CASCADE-2004 of high-energy
nuclear interactions Euro. Phys. Jour. A 10