- Email: [email protected]
Neutron spectra around a tandem linear accelerator in the generation of 18F with a bonner sphere spectrometer
Applied Radiation and Isotopes 114 (2016) 154–158
Contents lists available at ScienceDirect
Applied Radiation and Isotopes journal homepage: www.elsevier.com/locate/apradiso
Neutron spectra around a tandem linear accelerator in the generation of 18F with a bonner sphere spectrometer J.I. Lagares a,n, J.E. Guerrero Araque b, R. Méndez-Villafañe b, P. Arce a, F. Sansaloni a, O. Vela a, C. Díaz c, Xandra Campo b, J.M. Pérez d a
Medical Applications Unit, CIEMAT, Av. Complutense 40, E-28040 Madrid, Spain Ionizing Radiations Laboratory, CIEMAT, Av. Complutense 40, E-28040 Madrid, Spain c Astroparticle Physics Division, Basic Research Department, CIEMAT, Av. Complutense 40, E-28040 Madrid, Spain d Technology Department, CIEMAT, Av. Complutense 40, E-28040 Madrid, Spain b
H I G H L I G H T S
A neutron spectrum has been measured in a real target used in a 18F production cyclotron. Angular dependence of neutron emission rates in p(18O,18F)n agrees with literature. Spectra obtained with Geant4 Monte Carlo using TENDL-2010 library disagree with experiment.
art ic l e i nf o
a b s t r a c t
Article history: Received 17 March 2016 Received in revised form 6 May 2016 Accepted 17 May 2016 Available online 18 May 2016
A Bonner sphere spectrometer was used to measure the neutron spectra produced at the collision of protons with an H218O target at different angles. A unique H218O target to produce 18F was designed and placed in a Tandem linear particle accelerator which produces 8.5 MeV protons. The neutron count rates measured with the Bonner spheres were unfolded with the MAXED code. With the GEANT4 Monte Carlo code the neutron spectrum induced in the (p, n) reaction was estimated, this spectrum was used as initial guess during unfolding. Although the cross section of the reaction 18 O(p,n)18F is well known, the neutron energy spectra is not correctly defined and it is necessary to verify the simulation with measurements. For this reason, the sensitivity of the unfolding method to the initial spectrum was analyzed applying small variation to the fast neutron peak. & 2016 Elsevier Ltd. All rights reserved.
Keywords: Radioisotope production PET Radiation protection Neutron spectra Bonner spheres Monte Carlo
1. Introduction The characterization of the radioisotope production targets for PET is one of the aims of an investigation project in which the Ionization Radiations Metrology Laboratory and Medical Applications Unit belonging to CIEMAT (Centro de Investigaciones Energéticas, MedioAmbientales y Tecnológicas) are involved. Construction of a new model of cyclotron for radioisotope production leads to define the proper shielding for gamma and neutron radiation which conforms to the minimum requirements of personal dose. For that reason, it is fundamental to determine through experiments the energy spectrum of the radiation produced in order to characterize the source term, and later by Monte Carlo simulations to define the optimal shielding. Other neutron n
Corresponding author. E-mail address: [email protected] (J.I. Lagares).
http://dx.doi.org/10.1016/j.apradiso.2016.05.023 0969-8043/& 2016 Elsevier Ltd. All rights reserved.
spectra studies have been done around different cyclotrons (VegaCarrillo, 2000; Hertel et al., 2004; Méndez et al., 2005; Fernández et al., 2007; Guimaraes et al., 2012). While this work was done in a big bunker with low backscatter, the measurement in those papers, in particular, the spectra around the thermal and epithermal spectra, suffer from a bigger influence from neutron scattering, due to the small size of the bunker, the large amount of iron from cyclotrons and the characteristics of their targets. The aim of the present work was to measure the neutron spectrum produced at the collision of protons with the future CIEMAT 18F production target by means a Bonner sphere spectrometer (BSS). The 18 F production target is an AL6082 (International Alloy Designation System; IADS) container specially designed in order to set it up in an accelerating proton line from a TANDEM accelerator. Inside of the container there is a niobium material (99.9% purity) with a cavity of about 0.43 cm3 filled enriched water (H218O at 99% purity) and a resistant Havars window foil of 12.5 mm thick from Goodfellow
J.I. Lagares et al. / Applied Radiation and Isotopes 114 (2016) 154–158
Fig. 1. Layout of the 18F target with its different parts: collimator, current isolator, grid, enriched water and water cooling circuit.
(Havars, 2015), which separate the vacuum of the Tandem from the enriched water, reinforced by an AL6082 grid of hexagonal holes. According to the target design, initially the proton beam pass through the AL6082 grid that reduces the beam current up to 30%, afterwards through the Havars window foil that produces a significant contribution of neutrons, and finally, the beam interacts with the water enriched that produces the radioisotope 18F through the nuclear reaction 18O (p, n) 18F, Fig. 1.
2. Experimental setup The neutron fluence measurements were performed in the TANDEM linear accelerator at CMAM (Centre for Micro Analysis of
155
Material, Madrid - Spain) adjusted to 8.5 MeV protons for a beam current of 6.20 nA. The target configuration allow its using as a Faraday-Cup to measure the proton current. This data were collected continuously in order take into account the current fluctuations. The beam line includes an independent Faraday-Cup, which was used to check the target measurements at the begging and at the end of each irradiation. The differences between the current measured by the target and by the beam Faraday-Cup were less than 1%. The uncertainty in the determination of the current was around 8%. The BSS-CIEMAT consists of 12 polyethylene spheres of different diameter and a 3.3 cm-diameter spherical 3He filled (228.5 kPa) proportional counter sensitive to thermal neutrons at the centre of the moderating sphere (Méndez Villafañe et al., 2003). The measurements were made at 0° and 135° with respect to the beam axis and distances of 108 cm and 93 cm respectively from the target and counter centres (source-to-detector distance). The target was placed at the end of one lateral beam line of the accelerator bunker and 5.5 m far from the closest wall and at 1.6 m from the floor. The bunker size has around 32.5 m long, 4.5 m tall and one half has 12.5 m wide and the other half has 16.5 m. The wall has 1 m thickness. See Fig. 2 for a drawing of the hall. An integrated desktop MCA based on digital signal processing, DSA-1000 (Canberra Industries, Inc.) was used to record the count rate of each sphere. For each sphere the counting time was large enough to have an uncertainty below 1%.
3. Initial MAXED neutron spectrum calculation by Monte Carlo simulations The neutron count rates measured with the BSS were unfolded with the Maximum Entropy unfolding program (MAXED) (Reginatto et al., 1999) using the response matrix of the BSS-CIEMAT and an initial neutron spectrum calculated with GAMOS (Arce
Fig. 2. Bunker outline with the target situation inside it. Distances are written in meters The experimental hall is 4.5 m tall and the wall is 1 m thick. The TANDEM has 2 accelerator phases with a maximum nominal potential of 5 MV in each phase. Base outline courtesy of the Centre for Micro Analysis of Materials, Madrid - Spain.
156
J.I. Lagares et al. / Applied Radiation and Isotopes 114 (2016) 154–158
4. Results The results of the unfolded neutron spectra at different angles are showed in Figs. 5 and 6 (neutron energy bin structure of 20
Fig. 3. GEANT4 calculated neutron spectra produced at the collision of 8.5 MeV protons with the 18F production target at 0° and 135° (normalized to the total neutron fluence).
et al., 2008), a GEANT4 based code (Angostinelli et al., 2003). The GEANT4 simulation includes the real geometry and material composition of the 18F production target irradiated with 8.5 MeV protons. The nuclear reactions, elastic and inelastic scattering were treated with the TENDL-2010 (Talys-based Evaluated Nuclear Data Library) database of neutron and proton cross-sections (Koning and Rochman, 2010). The simulation results are the neutron fluence spectra at 0° and 135° with respect to the beam axis. The calculated neutron spectra in Fig. 3 have a neutron energy range from 2.5 keV to 8.5 MeV and a mean energy of 2.23 MeV (0°) and 1.93 MeV (135°). The peak that appears around 10 2 MeV (0°) is due to an error in the data file, and it is corrected for the final results. The neutron fluence response of the BSS cover the energy range from thermal to GeV neutrons (Alevra and Thomas, 2003), therefore, is necessary to estimate the contribution of room-return neutrons (those scattered from the walls of the facility, mostly thermal neutrons, ISO 8529-2:2000). It was performed a new simulation with MCNPX code (X-5 Monte Carlo Team. MCNP, 2008) which includes the dimensions of the irradiation room, the GEANT4 calculated neutron spectrum as source term and the ENDF/B-VII.0 (Evaluated Nuclear Data File) neutron cross-sections database (Chadwick et al., 2006). The MCNPX simulations were done under the same conditions used during the measurements. The spectra were used as initial guess in the unfolding with MAXED code (Fig. 4).
Fig. 6. MAXED unfolded neutron spectra of the 18F production target for different angles, 135° (normalized to the total neutron fluence).
Fig. 4. Calculated neutron spectra at 18F production target used as initial neutron spectra in the MAXED code (normalized to the total neutron fluence).
Fig. 7. Ratios of the measured readings, Md, and the calculated reading, Cd for MAXED unfolded neutron spectra of the 18F production target for different angles, 0 and 135°.
Fig. 5. MAXED unfolded neutron spectra of the 18F production target for different angles, 0° (normalized to the total neutron fluence).
J.I. Lagares et al. / Applied Radiation and Isotopes 114 (2016) 154–158
bins per decade). During unfolding the initial guess spectra are modified in shape and amplitude particularly above 2.5 keV. Mean energies above 2.5 keV were 1.57 MeV (0°) and 1.28 MeV (135°), that is a decrease by about 650 keV respect to the initial spectra. All this indicates that simulations performed do not match the reality; possibly caused by the inaccurate data of the neutron energy spectra, produced via (p,n)/(p,pn)/(p,2n) reactions, of TENDL-2010, used in GEANT4 code. Nevertheless, the ratios of the measured readings and the calculated reading of each sphere are close to unity (Fig. 7) and the values obtained of the reduced chisquared of 0.175 (0°) and 0.082 (135°) indicates that a reasonable fitting function was found by the MAXED code. The calculated reading is obtained by folding the solution spectrum with the fluence response matrix BSS-CIEMAT. Uncertainties bars shown in figure correspond to the propagation in the ratio between calculation and measurement uncertainties. The uncertainties in the unfolding process are obtained with an auxiliary code (IQU_FC) included in the UMG, 3.3 package (UMG). Only the uncertainty associated to the response matrix and the statistical uncertainty the measurement with every sphere are taken into account and they affect slightly to the uncertainty of the final spectrum obtained in the uncertainties propagation process (based on a perturbation study of both values, response matrix and count rate) (Reginatto et al., 2002; Miller et al., 2001). Nor the position of the spheres nor the proton current are considered. At energies below 2.5 keV the initial spectra were only slightly modified by the unfolding code, so the simulation and the unfolding at low energies are consistent with each other. The measured neutron fluence rates measured were 22.1 72.9% cm 2 s 1 nA 1 (0°) and 24.470.9% cm 2 s 1 nA 1 (135°). According to rule 1/4πr2 (r - source-to-detector distance), the neutron source emission rates were 3.26E6 72.9% s 1 nA 1 (0°) and 2.67E67 0.9% s 1 nA 1 (135°) (neutron fluence below 2.5 keV was not considered - neutrons scattered). Decreasing of the neutron emission with increasing angle is similar to the results obtained by Miller et al. 2001. The MAXED solution is made conditional to initial spectrum (Alevra and Thomas, 2003), but it has been shown that the calculated GEANT4 spectrum is not correctly defined. For that reason, a new MAXED solution was obtained with a new initial spectrum: a fission spectrum of peak energy of 880 keV and maximum neutron distribution of 8.5 MeV. The result at 0° in Fig. 8 shows a shifting of the peak at high energies. The solution spectra in Fig. 9, GEANT4 simulation and fission spectrum, are similar in shape and amplitude. Regardless of which of the two initial spectra is chosen, the corresponding solution spectra differ by less than 1% in their total fluence and dose values, Table 1. This means that the solution is not sensitive to the choice of one or the other initial spectrum, and that both solutions are consistent with each other.
5. Conclusion Using a Bonner sphere spectrometer the neutron spectra, to 0° and 135°, produced during the reaction 18O(p,n)18F reaction were Table 1 Neutron data about MAXED unfolded neutron spectra of the Initial spectrum
Reduced chi-squared
0.175 0.163
Fig. 8. MAXED unfolded neutron spectra of the 18F production target using a fission initial spectrum (at 0°, normalized to the total neutron fluence).
Fig. 9. MAXED unfolded neutron spectra of the 18F production target using the GEANT4 simulation and the fission spectrum as initial spectrum (at 0°, normalized to the total neutron fluence).
measured. Neutron source emission rates at two angles have been determined and results agree with other authors (Miller et al., 2001); decreasing emission with increasing angle. The neutron spectra obtained from GEANT4 differs from the experimental spectra obtained with BSS. So, it is necessary to find more accurate data for the neutron energy spectra, produced via (p,n)/(p,pn)/(p,2n) reactions, of TENDL-2010 library. Nevertheless, these neutron spectra have been used as initial spectra in the unfolding process, and the sensibility of the MAXED code to the initial spectrum has been evaluated using a fission spectrum. The corresponding solution spectra differ by less than 1% in their total fluence and dose values, so both solutions are consistent with each other.
18
F production target at 0° using different initial spectra.
Neutron fluence (cm
GEANT4 Fission
157
2
s
1
nA
22.17 2.9% 21.9 7 1.2%
Neutron emission
1
)
(s
1
nA
1
)
3.26E6 7 2.9% 3.24E67 1.2%
H* (10) (ICRP74HI) (μSvh
1
nA
29.5 7 3.7% 29.6 7 1.7%
1
)
Mean energy ĒΦ (MeV) 1.57 1.49
158
J.I. Lagares et al. / Applied Radiation and Isotopes 114 (2016) 154–158
Thus, the obtained MAXED solution in Figs. 5 and 6 may be incorporate as source term into a MCNP study of shielding.
Acknowledgements This work was supported by the AMIT project, a CDTI funded project (CENIT CIN/1559/2009). The authors are also very grateful to the staff at CMAM for their valuable help in the arrangement of the experiments.
References Alevra, A.V., Thomas, D.J., 2003. Neutrón spectrometry in mixed fields: multisphere spectrometers. Radiat. Prot. Dosim. 107 (1–3), 37–72. Angostinelli, S., Allison, J., et al., 2003. GEANT4 – a simulation toolkit. Nucl. Instrum. Methods A 506, 250–303. Arce, P., Rato, P., Cañadas, M., Lagares, J.I., 2008. GAMOS: A Gean4-based easy and flexible framework for nuclear medicine applications. In: Proceeding of IEEE Nuclear Science Symposium Conference Record. pp. 3162–3168. Chadwick, M.B., Obložiniský, P., et al., 2006. ENDF/B-VII.0: next generation evaluated nuclear data library for nuclear science and technology. Nucl. Data Sheet 107, 2931–3060. Hertel, N.E., Shannon, M.P., Wang, Z.-L., Valenzano, M.P., Mengesha, W., Crowe, R.J., 2004. Nuetron measurements in the vicinity of a self-shielded PET cyclotron. Radiat. Prot. Dosim. 108 (3), 255–261. Fernández, F., Amgarou, K., Domingo, C., García, M.J., Quincoces, G., Martí-Climent, J. M., Méndez, R., Barquero, R., 2007. Neutron spectrometry in a pet cyclotron with a bonner sphere system. Radiat. Prot. Dosim. 126 (1–4), 371–375.
Guimaraes, A.M., Lacerda, M.A.S., Santos, J.A.L., Maletta, P.G.M., Rodrigues, S.L.M., Andrade, R.S., Vilela, E.C., da Silva, T.A., 2012. Use of a TLD-based multisphere spectrometry system to measure the neutron spectra around a not-self shielded PET cyclotron: preliminary results. Appl. Radiat. Isot. 71, 92–95. Koning, A. J., Rochman, D., 2010. TENDL-2010: Talys-based Evaluated Nuclear Data Library. 〈http://www.talys.eu/tendl-2010/〉. Havars window foil, 2015. 〈http://www.goodfellow.com/E/Havar-High-StrengthNon-Magnetic-Alloy.html〉 (accessed 15.09.01.). ISO 8529-2:2000. Reference neutron radiations. Part 2: Calibrations fundamentals of radiations protection devices related to the basic quantities characterizing the radiation field. Méndez Villafañe, R., Merino, Los Arcos, Gallego Díaz, J.M., Lorente Fillol, A, E., 2003. Determination of the emission rate of an Am-Be neutron source with a Bonner Sphere Spectrometer. Radiat. Prot. Dosim. 107 (1–3), 37–72. Méndez, R., Iñiguez, M.P., Martí-Climent, J.M., Peñuelas, I., Vega-Carrillo, H.R., Barquero, R., 2005. Study of the neutron field in the vicinity of an unshielded PET cyclotron. Phys. Med. Biol. 50 (21), 5141–5152. Miller, L.F., Townsend, L.W., Alvord, D.W. 2001. Characterization of neutron and photon sources from a 10.5 MeV proton beam on [18O] enriched water. In: Proceedings of AIP Conference. 576, pp. 793–798. Reginatto, M., Goldhhagen, P., 1999. MAXED, A computer code for maximun entropy deconvolution of multisphere neutron spectrometer data. Health Phys. 77, 579–583. Reginatto, M., Goldhagen, P., Neumann, S., 2002. Spectrum unfolding, sensitivity analysis and propagation of uncertainties with the maximum entropy deconvolution code MAXED. Nucl. Instrum. Methods Phys. Res. Sect. A: Accel. Spectrom. Detect. Assoc. Equip. 476, 242–246. UMG 3.3, analysis of data measured with spectrometers using unfolding techniques, 2004. 〈http://www.oecd-nea.org/tools/abstract/detail/nea-1665〉 (accessed 15.09.01.). Vega-Carrillo, H.R., 2000. Neutron energy spectra inside a PET cyclotron vault room. Nucl. Instrum. Methods Phys. Res. A 463, 375–386. X-5 Monte Carlo Team. MCNP, 2008. A General Monte Carlo N-Particle Transport Code, Version 5. Los Alamos National Laboratory. Report LA-UR-03-1987.
Contents lists available at ScienceDirect
Applied Radiation and Isotopes journal homepage: www.elsevier.com/locate/apradiso
Neutron spectra around a tandem linear accelerator in the generation of 18F with a bonner sphere spectrometer J.I. Lagares a,n, J.E. Guerrero Araque b, R. Méndez-Villafañe b, P. Arce a, F. Sansaloni a, O. Vela a, C. Díaz c, Xandra Campo b, J.M. Pérez d a
Medical Applications Unit, CIEMAT, Av. Complutense 40, E-28040 Madrid, Spain Ionizing Radiations Laboratory, CIEMAT, Av. Complutense 40, E-28040 Madrid, Spain c Astroparticle Physics Division, Basic Research Department, CIEMAT, Av. Complutense 40, E-28040 Madrid, Spain d Technology Department, CIEMAT, Av. Complutense 40, E-28040 Madrid, Spain b
H I G H L I G H T S
A neutron spectrum has been measured in a real target used in a 18F production cyclotron. Angular dependence of neutron emission rates in p(18O,18F)n agrees with literature. Spectra obtained with Geant4 Monte Carlo using TENDL-2010 library disagree with experiment.
art ic l e i nf o
a b s t r a c t
Article history: Received 17 March 2016 Received in revised form 6 May 2016 Accepted 17 May 2016 Available online 18 May 2016
A Bonner sphere spectrometer was used to measure the neutron spectra produced at the collision of protons with an H218O target at different angles. A unique H218O target to produce 18F was designed and placed in a Tandem linear particle accelerator which produces 8.5 MeV protons. The neutron count rates measured with the Bonner spheres were unfolded with the MAXED code. With the GEANT4 Monte Carlo code the neutron spectrum induced in the (p, n) reaction was estimated, this spectrum was used as initial guess during unfolding. Although the cross section of the reaction 18 O(p,n)18F is well known, the neutron energy spectra is not correctly defined and it is necessary to verify the simulation with measurements. For this reason, the sensitivity of the unfolding method to the initial spectrum was analyzed applying small variation to the fast neutron peak. & 2016 Elsevier Ltd. All rights reserved.
Keywords: Radioisotope production PET Radiation protection Neutron spectra Bonner spheres Monte Carlo
1. Introduction The characterization of the radioisotope production targets for PET is one of the aims of an investigation project in which the Ionization Radiations Metrology Laboratory and Medical Applications Unit belonging to CIEMAT (Centro de Investigaciones Energéticas, MedioAmbientales y Tecnológicas) are involved. Construction of a new model of cyclotron for radioisotope production leads to define the proper shielding for gamma and neutron radiation which conforms to the minimum requirements of personal dose. For that reason, it is fundamental to determine through experiments the energy spectrum of the radiation produced in order to characterize the source term, and later by Monte Carlo simulations to define the optimal shielding. Other neutron n
Corresponding author. E-mail address: [email protected] (J.I. Lagares).
http://dx.doi.org/10.1016/j.apradiso.2016.05.023 0969-8043/& 2016 Elsevier Ltd. All rights reserved.
spectra studies have been done around different cyclotrons (VegaCarrillo, 2000; Hertel et al., 2004; Méndez et al., 2005; Fernández et al., 2007; Guimaraes et al., 2012). While this work was done in a big bunker with low backscatter, the measurement in those papers, in particular, the spectra around the thermal and epithermal spectra, suffer from a bigger influence from neutron scattering, due to the small size of the bunker, the large amount of iron from cyclotrons and the characteristics of their targets. The aim of the present work was to measure the neutron spectrum produced at the collision of protons with the future CIEMAT 18F production target by means a Bonner sphere spectrometer (BSS). The 18 F production target is an AL6082 (International Alloy Designation System; IADS) container specially designed in order to set it up in an accelerating proton line from a TANDEM accelerator. Inside of the container there is a niobium material (99.9% purity) with a cavity of about 0.43 cm3 filled enriched water (H218O at 99% purity) and a resistant Havars window foil of 12.5 mm thick from Goodfellow
J.I. Lagares et al. / Applied Radiation and Isotopes 114 (2016) 154–158
Fig. 1. Layout of the 18F target with its different parts: collimator, current isolator, grid, enriched water and water cooling circuit.
(Havars, 2015), which separate the vacuum of the Tandem from the enriched water, reinforced by an AL6082 grid of hexagonal holes. According to the target design, initially the proton beam pass through the AL6082 grid that reduces the beam current up to 30%, afterwards through the Havars window foil that produces a significant contribution of neutrons, and finally, the beam interacts with the water enriched that produces the radioisotope 18F through the nuclear reaction 18O (p, n) 18F, Fig. 1.
2. Experimental setup The neutron fluence measurements were performed in the TANDEM linear accelerator at CMAM (Centre for Micro Analysis of
155
Material, Madrid - Spain) adjusted to 8.5 MeV protons for a beam current of 6.20 nA. The target configuration allow its using as a Faraday-Cup to measure the proton current. This data were collected continuously in order take into account the current fluctuations. The beam line includes an independent Faraday-Cup, which was used to check the target measurements at the begging and at the end of each irradiation. The differences between the current measured by the target and by the beam Faraday-Cup were less than 1%. The uncertainty in the determination of the current was around 8%. The BSS-CIEMAT consists of 12 polyethylene spheres of different diameter and a 3.3 cm-diameter spherical 3He filled (228.5 kPa) proportional counter sensitive to thermal neutrons at the centre of the moderating sphere (Méndez Villafañe et al., 2003). The measurements were made at 0° and 135° with respect to the beam axis and distances of 108 cm and 93 cm respectively from the target and counter centres (source-to-detector distance). The target was placed at the end of one lateral beam line of the accelerator bunker and 5.5 m far from the closest wall and at 1.6 m from the floor. The bunker size has around 32.5 m long, 4.5 m tall and one half has 12.5 m wide and the other half has 16.5 m. The wall has 1 m thickness. See Fig. 2 for a drawing of the hall. An integrated desktop MCA based on digital signal processing, DSA-1000 (Canberra Industries, Inc.) was used to record the count rate of each sphere. For each sphere the counting time was large enough to have an uncertainty below 1%.
3. Initial MAXED neutron spectrum calculation by Monte Carlo simulations The neutron count rates measured with the BSS were unfolded with the Maximum Entropy unfolding program (MAXED) (Reginatto et al., 1999) using the response matrix of the BSS-CIEMAT and an initial neutron spectrum calculated with GAMOS (Arce
Fig. 2. Bunker outline with the target situation inside it. Distances are written in meters The experimental hall is 4.5 m tall and the wall is 1 m thick. The TANDEM has 2 accelerator phases with a maximum nominal potential of 5 MV in each phase. Base outline courtesy of the Centre for Micro Analysis of Materials, Madrid - Spain.
156
J.I. Lagares et al. / Applied Radiation and Isotopes 114 (2016) 154–158
4. Results The results of the unfolded neutron spectra at different angles are showed in Figs. 5 and 6 (neutron energy bin structure of 20
Fig. 3. GEANT4 calculated neutron spectra produced at the collision of 8.5 MeV protons with the 18F production target at 0° and 135° (normalized to the total neutron fluence).
et al., 2008), a GEANT4 based code (Angostinelli et al., 2003). The GEANT4 simulation includes the real geometry and material composition of the 18F production target irradiated with 8.5 MeV protons. The nuclear reactions, elastic and inelastic scattering were treated with the TENDL-2010 (Talys-based Evaluated Nuclear Data Library) database of neutron and proton cross-sections (Koning and Rochman, 2010). The simulation results are the neutron fluence spectra at 0° and 135° with respect to the beam axis. The calculated neutron spectra in Fig. 3 have a neutron energy range from 2.5 keV to 8.5 MeV and a mean energy of 2.23 MeV (0°) and 1.93 MeV (135°). The peak that appears around 10 2 MeV (0°) is due to an error in the data file, and it is corrected for the final results. The neutron fluence response of the BSS cover the energy range from thermal to GeV neutrons (Alevra and Thomas, 2003), therefore, is necessary to estimate the contribution of room-return neutrons (those scattered from the walls of the facility, mostly thermal neutrons, ISO 8529-2:2000). It was performed a new simulation with MCNPX code (X-5 Monte Carlo Team. MCNP, 2008) which includes the dimensions of the irradiation room, the GEANT4 calculated neutron spectrum as source term and the ENDF/B-VII.0 (Evaluated Nuclear Data File) neutron cross-sections database (Chadwick et al., 2006). The MCNPX simulations were done under the same conditions used during the measurements. The spectra were used as initial guess in the unfolding with MAXED code (Fig. 4).
Fig. 6. MAXED unfolded neutron spectra of the 18F production target for different angles, 135° (normalized to the total neutron fluence).
Fig. 4. Calculated neutron spectra at 18F production target used as initial neutron spectra in the MAXED code (normalized to the total neutron fluence).
Fig. 7. Ratios of the measured readings, Md, and the calculated reading, Cd for MAXED unfolded neutron spectra of the 18F production target for different angles, 0 and 135°.
Fig. 5. MAXED unfolded neutron spectra of the 18F production target for different angles, 0° (normalized to the total neutron fluence).
J.I. Lagares et al. / Applied Radiation and Isotopes 114 (2016) 154–158
bins per decade). During unfolding the initial guess spectra are modified in shape and amplitude particularly above 2.5 keV. Mean energies above 2.5 keV were 1.57 MeV (0°) and 1.28 MeV (135°), that is a decrease by about 650 keV respect to the initial spectra. All this indicates that simulations performed do not match the reality; possibly caused by the inaccurate data of the neutron energy spectra, produced via (p,n)/(p,pn)/(p,2n) reactions, of TENDL-2010, used in GEANT4 code. Nevertheless, the ratios of the measured readings and the calculated reading of each sphere are close to unity (Fig. 7) and the values obtained of the reduced chisquared of 0.175 (0°) and 0.082 (135°) indicates that a reasonable fitting function was found by the MAXED code. The calculated reading is obtained by folding the solution spectrum with the fluence response matrix BSS-CIEMAT. Uncertainties bars shown in figure correspond to the propagation in the ratio between calculation and measurement uncertainties. The uncertainties in the unfolding process are obtained with an auxiliary code (IQU_FC) included in the UMG, 3.3 package (UMG). Only the uncertainty associated to the response matrix and the statistical uncertainty the measurement with every sphere are taken into account and they affect slightly to the uncertainty of the final spectrum obtained in the uncertainties propagation process (based on a perturbation study of both values, response matrix and count rate) (Reginatto et al., 2002; Miller et al., 2001). Nor the position of the spheres nor the proton current are considered. At energies below 2.5 keV the initial spectra were only slightly modified by the unfolding code, so the simulation and the unfolding at low energies are consistent with each other. The measured neutron fluence rates measured were 22.1 72.9% cm 2 s 1 nA 1 (0°) and 24.470.9% cm 2 s 1 nA 1 (135°). According to rule 1/4πr2 (r - source-to-detector distance), the neutron source emission rates were 3.26E6 72.9% s 1 nA 1 (0°) and 2.67E67 0.9% s 1 nA 1 (135°) (neutron fluence below 2.5 keV was not considered - neutrons scattered). Decreasing of the neutron emission with increasing angle is similar to the results obtained by Miller et al. 2001. The MAXED solution is made conditional to initial spectrum (Alevra and Thomas, 2003), but it has been shown that the calculated GEANT4 spectrum is not correctly defined. For that reason, a new MAXED solution was obtained with a new initial spectrum: a fission spectrum of peak energy of 880 keV and maximum neutron distribution of 8.5 MeV. The result at 0° in Fig. 8 shows a shifting of the peak at high energies. The solution spectra in Fig. 9, GEANT4 simulation and fission spectrum, are similar in shape and amplitude. Regardless of which of the two initial spectra is chosen, the corresponding solution spectra differ by less than 1% in their total fluence and dose values, Table 1. This means that the solution is not sensitive to the choice of one or the other initial spectrum, and that both solutions are consistent with each other.
5. Conclusion Using a Bonner sphere spectrometer the neutron spectra, to 0° and 135°, produced during the reaction 18O(p,n)18F reaction were Table 1 Neutron data about MAXED unfolded neutron spectra of the Initial spectrum
Reduced chi-squared
0.175 0.163
Fig. 8. MAXED unfolded neutron spectra of the 18F production target using a fission initial spectrum (at 0°, normalized to the total neutron fluence).
Fig. 9. MAXED unfolded neutron spectra of the 18F production target using the GEANT4 simulation and the fission spectrum as initial spectrum (at 0°, normalized to the total neutron fluence).
measured. Neutron source emission rates at two angles have been determined and results agree with other authors (Miller et al., 2001); decreasing emission with increasing angle. The neutron spectra obtained from GEANT4 differs from the experimental spectra obtained with BSS. So, it is necessary to find more accurate data for the neutron energy spectra, produced via (p,n)/(p,pn)/(p,2n) reactions, of TENDL-2010 library. Nevertheless, these neutron spectra have been used as initial spectra in the unfolding process, and the sensibility of the MAXED code to the initial spectrum has been evaluated using a fission spectrum. The corresponding solution spectra differ by less than 1% in their total fluence and dose values, so both solutions are consistent with each other.
18
F production target at 0° using different initial spectra.
Neutron fluence (cm
GEANT4 Fission
157
2
s
1
nA
22.17 2.9% 21.9 7 1.2%
Neutron emission
1
)
(s
1
nA
1
)
3.26E6 7 2.9% 3.24E67 1.2%
H* (10) (ICRP74HI) (μSvh
1
nA
29.5 7 3.7% 29.6 7 1.7%
1
)
Mean energy ĒΦ (MeV) 1.57 1.49
158
J.I. Lagares et al. / Applied Radiation and Isotopes 114 (2016) 154–158
Thus, the obtained MAXED solution in Figs. 5 and 6 may be incorporate as source term into a MCNP study of shielding.
Acknowledgements This work was supported by the AMIT project, a CDTI funded project (CENIT CIN/1559/2009). The authors are also very grateful to the staff at CMAM for their valuable help in the arrangement of the experiments.
References Alevra, A.V., Thomas, D.J., 2003. Neutrón spectrometry in mixed fields: multisphere spectrometers. Radiat. Prot. Dosim. 107 (1–3), 37–72. Angostinelli, S., Allison, J., et al., 2003. GEANT4 – a simulation toolkit. Nucl. Instrum. Methods A 506, 250–303. Arce, P., Rato, P., Cañadas, M., Lagares, J.I., 2008. GAMOS: A Gean4-based easy and flexible framework for nuclear medicine applications. In: Proceeding of IEEE Nuclear Science Symposium Conference Record. pp. 3162–3168. Chadwick, M.B., Obložiniský, P., et al., 2006. ENDF/B-VII.0: next generation evaluated nuclear data library for nuclear science and technology. Nucl. Data Sheet 107, 2931–3060. Hertel, N.E., Shannon, M.P., Wang, Z.-L., Valenzano, M.P., Mengesha, W., Crowe, R.J., 2004. Nuetron measurements in the vicinity of a self-shielded PET cyclotron. Radiat. Prot. Dosim. 108 (3), 255–261. Fernández, F., Amgarou, K., Domingo, C., García, M.J., Quincoces, G., Martí-Climent, J. M., Méndez, R., Barquero, R., 2007. Neutron spectrometry in a pet cyclotron with a bonner sphere system. Radiat. Prot. Dosim. 126 (1–4), 371–375.
Guimaraes, A.M., Lacerda, M.A.S., Santos, J.A.L., Maletta, P.G.M., Rodrigues, S.L.M., Andrade, R.S., Vilela, E.C., da Silva, T.A., 2012. Use of a TLD-based multisphere spectrometry system to measure the neutron spectra around a not-self shielded PET cyclotron: preliminary results. Appl. Radiat. Isot. 71, 92–95. Koning, A. J., Rochman, D., 2010. TENDL-2010: Talys-based Evaluated Nuclear Data Library. 〈http://www.talys.eu/tendl-2010/〉. Havars window foil, 2015. 〈http://www.goodfellow.com/E/Havar-High-StrengthNon-Magnetic-Alloy.html〉 (accessed 15.09.01.). ISO 8529-2:2000. Reference neutron radiations. Part 2: Calibrations fundamentals of radiations protection devices related to the basic quantities characterizing the radiation field. Méndez Villafañe, R., Merino, Los Arcos, Gallego Díaz, J.M., Lorente Fillol, A, E., 2003. Determination of the emission rate of an Am-Be neutron source with a Bonner Sphere Spectrometer. Radiat. Prot. Dosim. 107 (1–3), 37–72. Méndez, R., Iñiguez, M.P., Martí-Climent, J.M., Peñuelas, I., Vega-Carrillo, H.R., Barquero, R., 2005. Study of the neutron field in the vicinity of an unshielded PET cyclotron. Phys. Med. Biol. 50 (21), 5141–5152. Miller, L.F., Townsend, L.W., Alvord, D.W. 2001. Characterization of neutron and photon sources from a 10.5 MeV proton beam on [18O] enriched water. In: Proceedings of AIP Conference. 576, pp. 793–798. Reginatto, M., Goldhhagen, P., 1999. MAXED, A computer code for maximun entropy deconvolution of multisphere neutron spectrometer data. Health Phys. 77, 579–583. Reginatto, M., Goldhagen, P., Neumann, S., 2002. Spectrum unfolding, sensitivity analysis and propagation of uncertainties with the maximum entropy deconvolution code MAXED. Nucl. Instrum. Methods Phys. Res. Sect. A: Accel. Spectrom. Detect. Assoc. Equip. 476, 242–246. UMG 3.3, analysis of data measured with spectrometers using unfolding techniques, 2004. 〈http://www.oecd-nea.org/tools/abstract/detail/nea-1665〉 (accessed 15.09.01.). Vega-Carrillo, H.R., 2000. Neutron energy spectra inside a PET cyclotron vault room. Nucl. Instrum. Methods Phys. Res. A 463, 375–386. X-5 Monte Carlo Team. MCNP, 2008. A General Monte Carlo N-Particle Transport Code, Version 5. Los Alamos National Laboratory. Report LA-UR-03-1987.