Nuclear Technology

ISSN: 0029-5450 (Print) 1943-7471 (Online) Journal homepage: [Link]

Testing of a High-Density Concrete as Neutron

Shielding Material

Eduardo Gallego, Alfredo Lorente & Héctor René Vega-Carrillo

To cite this article: Eduardo Gallego, Alfredo Lorente & Héctor René Vega-Carrillo (2009) Testing
of a High-Density Concrete as Neutron Shielding Material, Nuclear Technology, 168:2, 399-404,
DOI: 10.13182/NT09-A9216

To link to this article: [Link]

Published online: 10 Apr 2017.

Submit your article to this journal

View related articles

Full Terms & Conditions of access and use can be found at

[Link]

Download by: [University of Tasmania] Date: 04 September 2017, At: 04:14

 TESTING OF A HIGH-DENSITY RADIATION PROTECTION

CONCRETE AS NEUTRON KEYWORDS: neutron shielding,

high-density concrete, magnetite
SHIELDING MATERIAL concrete

EDUARDO GALLEGO* and ALFREDO LORENTE

Technical University of Madrid (UPM), Nuclear Engineering Department, Spain

HÉCTOR RENÉ VEGA-CARRILLO Universidad Autónoma de Zacatecas

Unidad Académica de Estudios Nucleares, Mexico

Received April 18, 2008

Accepted for Publication April 14, 2009
Downloaded by [University of Tasmania] at 04:14 04 September 2017

We present the testing of a high-density magnetite slabs of ordinary concrete (HA-25) used to shield med-
concrete [commercially available under the name ical accelerator facilities. In parallel to the experi-
Hormirad TM, developed by the Spanish company Con- ments, Monte Carlo calculations were performed with
strucciones Tecnicas de Radioterapia, S.L. (CT-RAD)] MCNP5, with some differences found with regard to
as neutron shielding material. The purpose of this work the experiments, attributable to uncertainties in the ele-
was to characterize the material behavior against neu- mental composition of the samples tested. Tenth-value
trons, as well as to test different mixings including boron layers have been determined for the different types
compounds in an effort to improve neutron shielding of concrete tested for the 241 Am-Be neutron source.
efficiency. Hormirad TM slabs of different thicknesses were The results show an advantageous behavior of the
exposed to a 241 Am-Be neutron source under controlled Hormirad TM when comparing it with ordinary concrete.
conditions. The original mix, which includes a high frac- Although borated concretes show a small improvement
tion of magnetite, was then modified by adding different in neutron attenuation when they are compared with
proportions of anhydrous borax (Na2B4O7). Looking for Hormirad TM alone, the resulting reduction in density
a comparison, the same experiment was repeated with and structural properties makes them less practical.

I. INTRODUCTION and neutrons and reduce the space needed for shielding.
These high-density concretes can have very different com-
Neutron shielding is of great importance in the de- positions and densities depending on the manufacturer or
sign and construction of several radiation-related facili- provider.2
ties. Practical shielding data available for the design of Optimum shielding of the radiation from particle ac-
accelerator facilities are, however, not abundant. In ad- celerators requires knowledge of the attenuation charac-
dition, with a demand for ensuring safety for public health, teristics of the shielding material. Construcciones Tecnicas
the shielding is, in general, oversized. Precise and prac- de Radioterapia, S.L. ~CT-RAD!, a commercial Spanish
tical data are thus highly important.1 company, has developed a magnetite-type concrete avail-
Of all shielding materials, concrete is probably the able under the name Hormirad TM.a The shielding fea-
most widely used because of its relatively low cost and tures against photons have been studied.3 Currently
the ease with which it can be cast into large and variously Hormirad TM is used to build bunkers, mazes, and doors
shaped results. Ordinary concrete can have different com- in medical accelerator facilities, where neutrons are also
positions in practice, usually depending on the aggre- produced.
gates used. Often concrete is made with much higher
density aggregates to increase its attenuation of photons a
U.S. and European patent applications have been presented
by the producer for the Hormirad TM “heavy mass for manu-
*E-mail: [Link]@[Link] facturing products with a high radioprotection capacity.”

NUCLEAR TECHNOLOGY VOL. 168 NOV. 2009 399

 Gallego et al. TESTING OF NEW HIGH-DENSITY CONCRETE FOR NEUTRON SHIELDING

The purpose of this work was to test and to charac- explanation for such a large dispersion in density values
terize the Hormirad TM as neutron shielding. The basic is that thicker slabs have higher density because the thin-
Hormirad TM material was compared with conventional ner slabs need some steel mesh inside to reinforce them
concrete as well as with new mixings including boron against breaking. Added steel mesh was compensated by
compounds in an effort to improve their neutron shield- reducing the amount of magnetite that must have stones
ing efficiency. of smaller size, thus resulting in a smaller density. How-
ever, the usual shielding material is thick, and therefore,
the slabs of 5 and 10 cm are more representative of the
II. MATERIALS AND METHODS real shielding, having an observed density of 4.03 6 0.07
g{cm⫺3, which indicates a good homogeneity in density.
Slabs were exposed to a 241Am-Be neutron source in
Hormirad TM is a magnetite-type concrete, where mag- the neutron measurements laboratory of the Universidad
netite ~FeO Fe 2O 3 !, represents almost 90% of its weight. Politécnica de Madrid Nuclear Engineering Depart-
Based on information provided by the manufacturer and ment.7 A Berthold LB6411 neutron monitor 8 was used to
by the magnetite stone supplier, the elemental composi- measure the transmitted ambient dose equivalent rate,
tion is indicated in Table I, compared with the elemental H^ * ~10!. The experimental setup is shown in Fig. 1. At
composition of ordinary concrete ~taken from Ref. 4!. 65 cm from the center of the source, the H^ * ~10! of
Downloaded by [University of Tasmania] at 04:14 04 September 2017

Other magnetite concretes have typically 48 wt% of Fe 241

Am-Be is 198 6 1 mSv{h⫺1.
~Refs. 5 and 6! while Hormirad TM has about 60 wt%. A The measuring time was 400 s, with data recorded
careful manufacturing procedure is followed so that the every 20 s, so the values used were the average between
homogeneity of the resulting blocks is rather high. The at least 20 recorded values. For all the measurements, the
resulting density is on the order of 4 g{cm⫺3. distance between the source’s center and the center of the
For the tests, Hormirad TM slabs of 25 ⫻ 25 cm and monitor was 65 cm. The source and monitor are situated
different linear thicknesses ~1, 2, 5, and 10 cm! were 3 m above the ground in a large room of 9 ⫻ 15 ⫻ 8 m in
prepared by CT-RAD. Different slabs were combined to order to reduce the neutron room return 9 as much as
reach linear thicknesses between 0 and 37 cm. The di- possible. H^ * ~10! due to neutron scattering inside the room
mensions and weight of these slabs were measured with was measured using a shadow cone, resulting in a value
calibrated instruments and their density was calculated, of 6.4 6 0.1 mSv{h⫺1. This value was used to correct the
resulting in values ranging from 3.44 to 4.10 g{cm⫺3. An measured values of H * ~10!, by applying Eq. ~1!, and by
propagating uncertainties according to Eq. ~2!:
H * ~10!corrected ⫽ H * ~10!Total measured ⫺ H * ~10!dispersed
TABLE I
Elemental Composition of Hormirad TM, Based ~1!
on Information Provided by the Manufacturer, Compared
to Conventional Portland Concrete* and

Portland sH * ~10!corrected ⫽ sH2 * ~10!Total measured ⫹ sH2 * ~10!dispersed .

M
Hormirad TM Density: Concrete Density:
Element 3.44 to 4.10 g{cm⫺3 2.30 g{cm⫺3 ~2!
Fe 60.80% 1.40%
O 31.26% 52.91%
Ca 4.36% 4.40%
Si 1.87% 33.70%
H 0.44% 1.00%
Mg 0.39% 0.20%
P 0.29% —
Ti 0.19% —
Al 0.17% 3.39%
K 0.06% 1.30%
Mn 0.06% —
V 0.05% —
C 0.04% 0.10%
S 0.01% —
N 0.003% —
Na — 1.60%

*In wt%. Taken from Ref. 4. Fig. 1. Experimental setup.

400 NUCLEAR TECHNOLOGY VOL. 168 NOV. 2009

 Gallego et al. TESTING OF NEW HIGH-DENSITY CONCRETE FOR NEUTRON SHIELDING

Ambient dose equivalent H * ~10! was also calcu- Both the experimental and the calculated values were
lated using MCNP5 ~Ref. 10! and ICRP 74 ~Ref. 11! adjusted by a corrected exponential function like that in
fluence to ambient dose equivalent conversion coeffi- Eq. ~3!. We applied weighted least squares 13 using the
cients. Models of the source, room, and the experimental inverse of the variance of H * ~10! as the weighting factor:
setup were prepared as realistically as possible, both with
respect to geometry and the atomic composition of the H * ~10!~ x! ⫽ B~Sx!{H * ~10!~0!{e ⫺Sx . ~3!
tested materials, including the heterogeneity between the In this equation, H * ~10!~0! is the ambient equivalent
tested slabs. H^ * ~10! calculated values were compared dose without shielding, and the accumulation factor was
with those measured. expressed as
In a second phase, ordinary shielding concrete ~HA-
25! slabs, with a measured density of 2.20 6 0.04 g{cm⫺3, B~Sx! ⫽ 1 ⫹ bSx . ~4!
were also tested by the same procedure to have a refer-
ence to concretes usually employed in building clinical From the least-squares adjustment the values of S,
accelerator rooms. b, and H * ~10!~0! were obtained for both the experimen-
Additionally, in an attempt to find better shielding tal and the calculated values. The values obtained are
materials against neutrons, other slabs were prepared indicated in Table II. In Fig. 3 the measured and calcu-
by CT-RAD adding anhydrous borax ~Na2B4O 7 ! to the lated values together with their adjustment functions
Downloaded by [University of Tasmania] at 04:14 04 September 2017

magnetite concrete, with the following percentages in are displayed. In order to verify the validity of the ad-
weight: 1.2, 5.0, and 25.1 ~equivalent to 0.6, 1.0, and justments, we applied the x 2 statistical test. In the case
5.4 wt% in B!. of the measured data, the resulting value of x 2 is 1.2106,

III. RESULTS TABLE II

Coefficients Obtained from the Adjustment
III.A. Hormirad TM Alone of the Measured and Calculated Values of H * ~10!
with the 241Am-Be Source
The experimental values of H^ * ~10! as a function of
Coefficient Experimental Data Calculated Data
the thickness of Hormirad TM shielding can be seen in
Fig. 2. A behavior that is not purely exponential can be H * ~10!~0! ~ mSv0h! 185.7047 6 1.4823 194.2079 6 2.4713
observed. This is due to the influence of neutrons dis- S ~cm⫺1 ! 0.1300 6 0.0020 0.1220 6 0.0037
persed both in the shielding and in the air that reach the b ~no units! 0.6692 6 0.05485 0.5820 6 0.0991
detector. This effect is usually represented by an “accu-
mulation factor” 12 B~S{x!, where x is the shielding thick-
ness and S is the neutron attenuation coefficient that
depends on the neutron energy and the shielding
properties.

Fig. 3. Plot of the measured and calculated data with the

241
Am-Be source together with their adjustment func-
Fig. 2. Plot of experimental values of H^ * ~10! with regard to tions ~relative ambient dose equivalent with regard to
Hormirad TM thickness with the 241Am-Be source. Hormirad TM shielding thickness!.

NUCLEAR TECHNOLOGY VOL. 168 NOV. 2009 401

 Gallego et al. TESTING OF NEW HIGH-DENSITY CONCRETE FOR NEUTRON SHIELDING

while for the calculated data it is 4.6066. These values

are smaller than the critical value of 37.6525 for a ⫽
0.05 and 25 degrees of freedom. This means that there
are no significant statistical differences between the dis-
crete values ~measured and calculated! and the respec-
tive adjustment functions.
Using the adjusted functions the values of the half-
value layer ~HVL! x 102 and the tenth-value layer ~TVL!
x 1010 were estimated by using Eqs. ~5! and ~6!, respec-
tively. Then, changing the right-side values in the equa-
tions by _41 and _1
100 , the second HVL and TVL were obtained.
All these obtained values are indicated in Table III. As
can be seen, the most conservative values of the first
layers for the Hormirad TM would be x 102 ⫽ 10.23 6
0.16 cm and x 1010 ⫽ 27.81 6 0.85 cm, equivalent to
x 102 ⫽ 41.23 6 0.65 g{cm⫺2 and x 1010 ⫽ 112.07 6
3.43 g{cm⫺2 in terms of mass thickness:
Downloaded by [University of Tasmania] at 04:14 04 September 2017

~1 ⫹ bSx 102 !e ⫺Sx 102 ⫽ ⫺21 ~5! Fig. 4. Correlation between calculated and measured H^ * ~10!
for Hormirad TM slabs of different thicknesses.
and

~1 ⫹ bSx 1010 !e ⫺Sx 1010 ⫽ _1

10 . ~6!
neutron spectra for 241Am-Be without a shield and be-
The correlation between calculated and measured hind the thicker shielding ~37.20 6 0.05 cm! are shown.
H * ~10! for Hormirad TM slabs is shown in Fig. 4. Differ-
ences were found comparing calculated with experimen- III.B. Ordinary Concrete
tal H * ~10! values since they fit reasonably well for small For ordinary concrete ~HA-25!, the resulting first
thicknesses ~higher dose rate! but deviate for thicker lay- TVL is 87.6 6 2.6 g{cm⫺2, equivalent to 39.8 6 1.2 cm,
ers ~lower dose rate!. For layers beyond 5 cm systemat- as indicated in Table IV. This is comparable to a value
ically all the calculated values with MCNP5 are smaller reported in literature 6 and larger than the Hormirad’s
than those measured, with differences tending to grow as TVL.
the thickness increases. The probable explanation for the
differences is twofold: first, the differences in the ele- III.C. Borated Mixings
mental composition between actual Hormirad TM slabs
and that utilized in the Monte Carlo calculations; second, The results of adding anhydrous borax to the
the response of LB 6411 varies with the neutron energy, Hormirad TM magnetite mixture in different percentages
and during neutron transport in the shield slabs besides in weight are also summarized in Table IV. The reduc-
the neutron attenuation, the neutron spectrum is shifted tions obtained in the first TVL with regard to the value
to epithermal and thermal energies. This is shown in for the Hormirad TM alone are indeed not significant.
Fig. 5, where the LB 6411 response and the lethargy Moreover, these compound materials have several dis-
advantages that make them not good options, like the
increase in cost and the poorer structural properties,
especially if the fraction of borax is elevated.
TABLE III
Half-Value and Tenth-Value Layers for Hormirad TM IV. CONCLUSIONS
Shielding of an 241Am-Be Source Obtained from
the Adjustment Functions of the Measured For 241Am-Be neutrons, the magnetite-based
and the Calculated Values Hormirad TM concrete has shown a clearly advanta-
geous behavior, when comparing neutron attenuation
Experimental Calculated against the thickness of the shielding with regard to
Layer ~cm! ~cm! ordinary concrete. Although borated compounds show
better neutron attenuation with respect to the mass thick-
First x 102 10.23 6 0.16 10.12 6 0.31
First x 1010 27.01 6 0.42 27.81 6 0.85
ness, the resulting reduction in density and structural
Second x 104 17.88 6 0.30 18.15 6 0.56 properties makes them less practical.
Second x 10100 48.08 6 0.75 50.19 6 1.53 However, more research seems necessary to better
characterize the elemental composition of the materials,

402 NUCLEAR TECHNOLOGY VOL. 168 NOV. 2009

 Gallego et al. TESTING OF NEW HIGH-DENSITY CONCRETE FOR NEUTRON SHIELDING
Downloaded by [University of Tasmania] at 04:14 04 September 2017

Fig. 5. LB 6411 response and lethargy fluence rate.

TABLE IV
241Am-Be
TVL Obtained for Neutrons with Ordinary Concrete and Borated
Hormirad TM Mixings of Different Characteristics

First TVL for 241Am-Be Neutrons

Average Density Mass Thickness Thickness

Shielding Material ~g{cm⫺3 ! ~g{cm⫺2 ! ~cm!

Ordinary concrete ~HA-25! 2.20 6 0.04 87.6 6 2.6 39.8 6 1.2

Mixed Hormirad TM with 1.19% borax 3.94 6 0.07 104.7 6 3.2 26.6 6 0.8
Mixed Hormirad TM with 5% borax 3.62 6 0.07 95.9 6 2.9 26.5 6 0.8
Mixed Hormirad TM with 25.1% borax 2.68 6 0.05 79.4 6 2.4 29.6 6 0.9

and also to determine the Homirad’s neutron attenuation Accelerator-Produced Leakage Neutron and Photon Transmis-
properties for typical spectra of clinical accelerators. sion Through Concrete,” Health Phys., 84, 2, 180 ~2003!.
3. M. A. DUCH and V. PANETTIERI, Technical Reports
SPR005006 and SPR008006, Instituto de Técnicas Energéti-
ACKNOWLEDGMENTS cas, Universidad Politécnica de Catalunya ~2006! ~in Spanish!.
We wish to thank J. M. Caruncho and I. Delicado, from 4. S. M. SELTZER and M. J. BERGER, “Evaluation of the
CT-RAD, for their collaboration in supplying the shielding Collision Stopping Power of Elements and Compounds for
materials tested and data about their composition. Electrons and Positrons,” Int. J. Appl. Radiat. Isot., 33, 1189
~1982!.
5. R. G. JAEGER, E. P. BLIZARD,A. B. CHILTON, M. GRO-
REFERENCES TENHUIS, A. HÖNIG, T. A. JAEGER, and H. H. EISENLOHR,
Engineering Compendium on Radiation Shielding—Volume II:
1. J. K. SHULTIS and R. E. FAW, Radiation Shielding, Amer- Shielding Materials, Springer, Berlin0Heidelberg ~1975!.
ican Nuclear Society, La Grange Park, Illinois ~2000!.
6. “Radiation Protection for Particle Accelerator Facilities,”
2. K. R. KASE, W. R. NELSON, A. FASSO, J. C. LIU, X. NCRP Report No. 144, National Council on Radiation Protec-
MAO, T. M. JENKINS, and J. H. KLECK, “Measurements of tion and Measurements ~2005!.

NUCLEAR TECHNOLOGY VOL. 168 NOV. 2009 403

 Gallego et al. TESTING OF NEW HIGH-DENSITY CONCRETE FOR NEUTRON SHIELDING

7. E. GALLEGO, A. LORENTE, and H. R. VEGA- 10. X-5 MONTE CARLO TEAM, “MCNP—A General Monte
CARRILLO, “Characteristics of the Neutron Field of the Fa- Carlo N-Particle Transport Code, Version 5,” LA-UR-03-
cility at DIN-UPM,” Radiat. Prot. Dosim., 110, 1–4, 73 ~2004!. 1987, Los Alamos National Laboratory ~2005!.

8. B. BURGKHARDT, G. FIEG, A. KLETT, A. PLEWNIA, 11. “Conversion Coefficients for Use in Radiological Protec-
and B. R. L. SIEBERT, “The Neutron Fluence and H * ~10! tion Against External Radiation, ICRP Publication 74,” Ann.
Response of the New LB6411 Rem Counter,” Radiat. Prot. ICRP, 27, 4 ~1996!.
Dosim., 70, 361 ~1997!. 12. M. A. PRICE, C. C. HORTON, and K. T. SPINNEY, Ra-
diation Shielding, Pergamon Press, London ~1957!.
9. H. R. VEGA-CARRILLO, E. MANZANARES-ACUÑA,
M. P. IÑIGUEZ, E. GALLEGO, and A. LORENTE, “Study of 13. H. R. VEGA-CARRILLO, “Least Squares for Different
Room-Return Neutrons,” Radiat. Meas., 42, 3, 413 ~2007!. Experimental Cases,” Rev. Mex. Fis., 35, 4, 597 ~1989!.
Downloaded by [University of Tasmania] at 04:14 04 September 2017

404 NUCLEAR TECHNOLOGY VOL. 168 NOV. 2009