Comparison with simulations to experimental data for photoneutron reactions using SPring-8 Injector

. Simulations of photo-nuclear reactions by using Monte Carlo codes PHITS and FLUKA have been performed to compare to the measured data at the SPring-8 injector with 250MeV and 961MeV electrons. Measurement data of Bismuth-206 productions due to photo-nuclear reactions of 209 Bi( (cid:538) ,3n) 206 Bi and high energy neutron reactions of 209 Bi(n,4n) 206 Bi at the beam dumps have been compared with the simulations. Neutron leakage spectra outside the shield wall are also compared between experiments and simulations.


Introduction
Since 1997, the 8GeV class synchrotron radiation facility, SPring-8, has been operated successfully for the users. The shielding designs of the SPring-8 accelerators were performed by using Jenkins' formula [1] for the bulk shielding in the lateral direction and modified Swanson's formula [2] in the forward direction. To verify the verification of the shielding design, benchmark experiments were performed by using Helium-3 high sensitive neutron counter with polyethylene moderators, TLD dose meters and Polymeric track detectors. The dose distributions within the shield tunnel were also measured by using LiF Thermo-Luminesence detectors and Polymetric neutron track detectors ( CR-39 ) and the data were presented at ICRS9 [3]. The high energy neutron distributions around the beam dump of the accelerated electron are also measured by using bismuth detectors. Those experimental data were presented and compared with the empirical formula and EGS4 [4] with photo-neutron production cross section data [5].
In addition to FLUKA [6,7], the Monte Carlo code PHITS2.82 [8,9] have been improved recently to simulate the photo-nuclear reaction calculation. These simulation codes are much useful for radiation shielding design of electron accelerator machines as well as proton accelators with good usability. The neutrons and gamma dose distributions around the beam dumps of the SPring-8 injector have been simulated by using both Monte Carlo codes. However experimental data, especially photonuclear reaction data are quite insufficient to compare to the simulations.
Bismus is a useful metal to measure high energy neutrons and gamma-rays because of the neutron and gamma energy dependence of the nuclear production of the bithmus radioisotopes. At ICRS9, the data of bismus-206 due to 250MeV electrons were presented. In addition to the data produced by 250 MeV electrons, the experimental data produced by 961 MeV electrons are presented and both data have been compared to simulations. The neutron spectra outside the shield tunnel are compared between the PHITS simulations and experimental data.

Experiments
The measurements were performed in 1998. At that time, the injector of SPring-8 had two dumps, one was 250MeV (L1) dump and the other was 1.15GeV (L2) dump. The geometrical configuration is shown in Fig.1. At the experiments, the 250MeV and 961 MeV electron beam that lead to L1 dump and L2 dump were employed and the data are indicated in Table 1, respectively. The beam dumps have double cylindrical structures, made of graphite for the inner layer and iron for the outer layer to reduce photo-neutron production and shielding.
The neutron leakage spectra outside the shield tunnel were measured by using extremely high sensitive helium-3 counter with polyethylene moderators, and the data were unfolded with SAND-II code [10]. Two types (small and large) of bismuth ingot were set around the beam dumps to measure the high energy neutrons and gamma-rays  15GeV dump, respectively. Unit is mm) At the end station of the linac, accelerated electrons can be extracted to 3 ways. One is to L2 dump, one is to booster synchrotron, and the other is to small SR ring, NewSUBARU.

Simulations
Neutron and gamma effective dose (Worst geometry) simulations around the L1 dump by using FLUKA are shown in Figs

Comparison between simulations and experiments 4.1 Bismuth isotopes production around the beam dumps
Small and large size bismuth ingots were employed for activation detectors and gamma-rays from bismuth detectors were measured after irradiation. The size of small detectors is 1.12 cm and 2.9 cm for the thickness and the diameter, respectively. The large detectors are 2.3 cm thick and 8 cm in diameter. In those cases, the self-shielding effect of detectors cannot be ignored to measure the gamma-rays. Figure 6 shows the selfshielding factors depending on gamma-ray energy and the size of bismuth detectors. These are PHITS simulations on the assumption of the isotropic distribution of the radioisotopes within the detectors. The assumption of the isotropic distribution is reasonable because of not so much thick detectors for high energy photons and neutrons and it's confirmed by using PHITS simulations. The previous data [3] were without considering these self-shielding effects so that these data must be corrected, especially for large detectors. The PHITS simulations of bismuth isotopes productions have been performed using [T-Dchain] command and DCHAIN-SP [11] with weighting factor pnimul=100 of photo-nuclear reaction.

250MeV electrons (L1 dump)
The bismuth-206 production distribution around the L1 dump was measured and compared to the Monte Carlo simulations. The bismuth detectors were set up as shown in Fig.7. In the experimental paper [3], the bismuth-206 production calculations were performed by using EGS4 with the library for 209 Bi( ,3n) 206 Bi reaction created by themselves using MCPHOTO [12] + PICA [13]. The reference paper also considered for 209 Bi(n, 4n) 206 Bi reaction by using ENDF-B/VI library.    Table 3 including PHITS simulations for D1-1 detector. In Table 3, «unknown » means that a long decay time was needed to measure the gamma-rays because of very high activities. Therefore, the information of bismuth-204 and 203 was cleared.  Fig.8. Configuration of bismuth detectors around the L2 beam dump (Vertical plane, left figure is parallel and right is perpendicular against the electron beam axis )

961MeV electrons (L2 dump)
The bismuth-206 production distribution around the L2 dump using 961 MeV electrons was measured as well as 250MeV electrons, and the configuration of bismuth detectors are as shown in Fig.8. The measurement data and simulation results are listed in Table 4. Both simulations show agreement within about 50 % as shown in Fig.9 even though simulations have large stastical errors. Another bismuth radioisotopes which were produced in the detectors of D2-2, D2-3,and D2-4 are listed in Table 5 210 Bi experimentally and directly, however it is easy to compare between the simulations. In comparison with the simulations, 210 Bi productions are listed as indicated in Table 6, and shown in good agreements.

Leakage neutron spectrum outside the roof of shield tunnel
Leakage neutron spectra have been simulated by using PHITS with the [T-Point] tally of which algorism is same as the Point Detector of MCNP5 [14] because of strong attenuation through thick shield walls. In these simulations, two points have been performed to compare the experimental data [3]. One is located directly above the L1 dump outside the roof of the tunnel with the shield thickness of 190 cm ordinary concrete, and the other is located directly above the L2 dump outside the roof of the tunnel with the shield thickness of 360 cm ordinary concrete as indicated in Figs. 1 and 8. These points are under air conditioning so that it can be avoided the change of the environmental conditions, relatively. In the simulations, the density of 2.26 g/cm 3 has been employed for ordinary concrete. Figs. 10 and 11 show the simulation results in comparison with the experimental data that were measured by using extremely high sensitive helium-3 counter with multipolyethylene moderators and unfolded the measurement data. These simulations show reasonably agree with the experimental data. However, the results of the simulation for the 250MeV electron case indicate slightly high in the region of high energy in comparison with the

ICRS-13 & RPSD-2016
7016 experimental data as shown in Fig.10. On the other hand, Fig.11 shows that the simulation results for the 961 MeV electron case indicate the slightly lower spectrum than that of experimental data in the region of intermediate   energy region. These discrepancies need future investigations as well as the improvements of the statistical errors.

Summary
The experimental data of bismuth radioisotopes distribution at around the beam dumps produced by 961 MeV electrons are presented as well as that of due to 250 MeV electrons. These data are useful to compare to the simulations. FLUKA and PHITS Monte Carlo simulations have been performed to compare to these data and the both simulations show reasonably agreements in a range of within about 50%. The PHITS simulations are performed for the leakage neutron energy spectra to compare to the experimental data, and these show slightly different shapes. These discrepancies need future investigations.  (1) 961MeV (4) Fig.11 Comparison of the leakage neutron energy spectrum at the roof outside the shield tunnel and directly above of the L2 dump. Lines indicate the same as Fig.10.
The author wishes to thank Dr. Tatsuhiko.Sato of JAEA and the developer group members for thier help and valuable discussions to use PHITS.