EUROPEAN ORGANIZATION FOR NUCLEAR RESEARCH CERN SL DIVISION CERN SL-2002-007 (ECT) RADIOLOGICAL IMPACT OF THE TRIGAACCELERATOR-DRIVEN EXPERIMENT (TRADE) 1 A. Herrera-Martinez, A. Ferrari, Y. Kadi, L. Zanini, 2 G.T. Parks, 3 C. Rubbia, N. Burgio, M. Carta, A. Santagata, 4 L. Cinotti 1 CERN, Switzerland, 2 University of Cambridge, UK, 3 ENEA Italy, 4 Ansaldo Nucleare, Italy Abstract The TRADE project 1, which is part of the European Roadmap 2 towards the development of Accelerator Driven Systems (ADS), foresees the coupling of a 110 MeV, 2 ma proton cyclotron with the core of a 1 MW Triga research reactor. We performed radioprotection studies using two state-of-the-art computer code packages, FLUKA 3 and EA-MC 4. We concentrated on the calculation of the neutron and particle flux and dose rates during normal operation as well as in the case of several possible accidents, in order to assess the radiation damage and define the design of key components of the facility, such as the beam-line shielding. Both high-energy particle interactions and low-energy neutron transport are treated with a sophisticated method based on a full Monte Carlo simulation, combined with the use of modern nuclear data libraries. Presented at the 12 th Biennal RPSD Topical Meeting 2002 Santa Fe April 14-18 2002 Geneva, Switzerland May 2002
RADIOLOGICAL IMPACT OF THE TRIGA ACCELERATOR-DRIVEN EXPERIMENT (TRADE) A. Herrera-Martínez*, A. Ferrari, Y. Kadi, L. Zanini European Organization for Nuclear Research, CERN CH-1211, Geneva 23, Switzerland adonai.herrera.martinez@cern.ch office: +41 22 767 9397 * Also at the University of Cambridge C. Rubbia, N. Burgio, M. Carta, A. Santagata Ente per le Nuove Tecnologie l Energia e l Ambiente, ENEA, 00196 Rome, Italy carta@casaccia.enea.it office: +39 06 3048 3183 G.T. Parks Department of Engineering, University of Cambridge CB2 1PZ, Cambridge, United Kingdom Gtp10@cus.cam.ac.uk L. Cinotti Ansaldo Nucleare 16161 Genova, Italy cinotti@ansaldo.it office: +39 010 655 8283 for the TRADE Collaboration a SUMMARY The TRADE project 1, which is part of the European Roadmap 2 towards the development of Accelerator Driven Systems (ADS), foresees the coupling of a 110 MeV, 2 ma proton cyclotron with the core of a 1 MW Triga research reactor. We performed radioprotection studies using two state-of-the-art computer code packages, FLUKA 3 and EA-MC 4. We concentrated on the calculation of the neutron and particle flux and dose rates during normal operation as well as in the case of several possible accidents, in order to assess the radiation damage and the design of key components of the facility, such as the beam-line shielding. Both high-energy particle interactions and low-energy neutron transport are treated with a sophisticated method based on a full Monte Carlo simulation, combined with the use of modern nuclear data libraries. I. INTRODUCTION In the European Roadmap towards the construction of a full prototype of an ADS, the TRADE experiment occupies a key role due to the fact that it will constitute the first example of an accelerator-reactor coupling at power. In particular, the TRADE experiment is based on the coupling of a 110 MeV, 2 ma proton cyclotron with a 1 MW reactor of type Triga, operating at the ENEA Casaccia site (Rome). The reactor core will be modified in order to run the experiment at different levels of sub-criticality to explore the transition from an external source dominated regime to one dominated by core thermal-feedbacks. Moreover, an essential modification will consist of inserting at the center of the core a tungsten target for the production of spallation neutrons. The TRADE project will offer the possibility of observing the dynamic behaviour of an ADS, by studying for instance start up and shut down operations, the proton current power relation, neutron source importance and reactivity effects 1. An essential role in the feasibility study of the experiment is played by radioprotection calculations. In fact, such a system exhibits new characteristics with respect to a traditional reactor, due to the presence of an intense proton accelerator. Therefore one needs to perform shielding studies not only around the reactor core but also along the beam line, with particular attention paid to the permanent magnets which drive the protons towards the core of the Triga reactor. Beam losses are always present in the normal operating condition of an accelerator, and the effective dose in these conditions must be evaluated. In addition, a The working group on TRADE: TRiga Accelerator Driven Experiment (ENEA, CEA, Ansaldo): TRIGA Experiment Feasibility Report, Rome, June 2001.
several accident cases related to the failure of the accelerator should be considered. II. RESULTS We describe the results from detailed Monte Carlo (MC) transport calculations performed in the course of this analysis: neutrons streaming in the reactor building from the vessel and analysis of their energy spectra; spallation neutrons produced in the case of the proton beam colliding against the bending magnet M4; radiation dose in the case of normal beam losses. Shielding against these contingencies was also designed and analyzed through the same methods. A. Reference Configuration The geometry of the Triga reactor was implemented in FLUKA and EA-MC. The existing reactor was reproduced in detail including the fuel pins and the other core elements (Figure 1). The model also included the new components of the TRADE experiment such as the proton beam line, the bending magnets and the modifications to the core. the order of 3 4 MeV. These neutrons are rapidly slowed down by the water moderator before reaching the core internal structures, therefore they do not represent a major problem in terms of radiation damage. It is worth noting that a few thermal neutrons, up to 10 9 n/cm 2 /s/ma, will stream along the beam tube and slightly irradiate (1.3 10-5 Gy/s/mA) the lower permanent magnet (M5), situated 3 m above the core. These detailed MC calculations (involving importance biasing techniques) show that the neutron leakage through the top of the reactor pool is negligible. φ (neutrons/cm 2 /s/ma) air reactor pool M4 beam line Figure 2. Neutron flux distribution (n/cm 2 /s/ma) following the impact of 110 MeV protons on a solid tungsten target. secondary pool core M5 concrete shielding bending magnets Figure 1. Schematic view of the implemented model. B. High-Energy Beam Target Interactions The distribution of neutrons generated in the spallation target unit (tungsten rod) and surrounding core structures is given in Figure 2. Spallation neutrons are produced in a region extending radially up to 5 cm from the center of the core, that is a few cm before reaching the fuel elements. Their average kinetic energy is of 1m However, this is not the case for photons generated by nuclear interactions which propagate mostly through the water moderator (Figure 3), although only a small fraction reach the surface of the reactor pool (10 4 γ/cm 2 /s/ma). Figure 3. Photon flux distribution (γ/cm 2 /s/ma) following the impact of 110 MeV protons on a solid tungsten target. C. Shielding Aspects of the Proton Beam Line φ (photons/cm 2 /s/ma)
One of the main worries associated with the coupling of the accelerator and the Triga reactor is the radiation produced during normal operation, due to protons escaping along the beam line. These losses should be of the order of 10-4 10-6 of the beam current, i.e. between 1 and 100 na/m, which implies, integrating over time, a large amount of radiation produced. Taking into account the maximum allowable dose rate in the accessible area of the reactor building (10 µsv/h), significant shielding will be required. Several shielding options were considered based on the nature and the thickness of the concrete used. We report here the main results from these calculations. We consider first the case in which all the protons exit the beam line interacting with the beam pipe and the shielding. This case would correspond to the failure of one of the focusing quadrupoles. In order to calculate the doses due to normal beam losses, it is sufficient to rescale the results by a conservative factor of 10-4. The distribution of the particles generated in the beam tube and the surrounding concrete shielding as a result of the interactions produced by the protons exiting the beam line is given in Figure 4. Spallation neutrons are produced in the aluminum beam tube (8 mm thick) with a yield of ~ 0.03 n/p and as deep as 1 m inside the concrete (~ 0.14 n/p, that is 82% of the neutrons produced) due mainly to high-energy neutrons interacting with the heavy nuclides present in concrete, such as barium (8% of the atoms). φ (particles/cm 2 /s/ma) Figure 4. Particle flux distribution (particle/cm 2 /s/ma) due to beam losses along the beam line. The energy spectra of neutrons and photons escaping the beam tube and the concrete shielding is plotted in Figure 5. Flux (dn/dlne/cm^2/s/ma) 1 0 1x10 9 1x10 8 1x10 7 1x10 6 1x10 5 1x10 4 Photon flux pipe -> concrete Photon flux concrete -> air Neutron flux pipe -> concrete Neutron flux concrete -> air 1x10 3 1x10 2 1x10 3 1x10 4 1x10 5 1x10 6 1x10 7 1x10 8 1x10 9 Energy (ev) Figure 5. Energy spectra for photons and neutrons escaping the concrete shielding during normal operation. These neutrons with energy ranging from a few ev up to 50 MeV will either multiply in concrete through (n, xn) reactions or be captured in most of the cases in boron and barium (see Table 1). A small fraction will eventually escape, therefore contributing to the dose rate in the accessible area of the reactor building. Table 1. Integrated neutron flux, heating and damage (per 100 na of beam loss) in the beam transport tube and surrounding concrete shield structures due to neutrons with energy below 20 MeV Region Flux (n/cm 2.s) Heat (W/cm 3 ) DPA/yr Beam tube 7.9x10 6 Concrete 1.7x10 5 3.8x10-9 1.4x10-9 Escapes 3.8x10 2 7.1x10-20 3.2x10-12 Neutron Absorption Main Nuclear Reactions Capture 6.50 % Beam tube 0.09 % (n, xn) 1.60 % Air gap 0.02 % α prod. 88.75 % Concrete 99.89 % H prod. 3.06 % D prod. 0.06 % T prod. 0.003 %
MC calculations show that the particle flux at the surface of the concrete shielding is estimated to be φ 10 7 n/cm 2 /s/ma, consisting mostly of neutrons, φ n 8 10 6 n/cm 2 /s/ma and photons, φ γ 2 10 6 γ/cm 2 /s/ma. The effective dose to human body was calculated using Pelliccioni conversion coefficients 5, resulting in an effective dose of about 3 msv/s/ma, essentially due to neutrons (90% of the total). Assuming beam losses of 100 na (10 na/m 10 m), we obtain the corresponding dose rate of 10 3 µsv/h (at contact), which is a 100 times the allowable dose rate in the reactor building for unlimited duration exposures of personnel (10 µsv/h). This dose rate can be reduced by increasing the thickness of the concrete shielding. For instance, increasing the shielding thickness by 40 cm will reduce the dose rate by one order of magnitude (~ 100 µsv/h), as shown in Figure 6. Neutron Flux per 100 na of beam loss (n/cm2.s) 1x10-2 1x10-3 baritic concrete 1 m thick baritic concrete 1.4 m thick baritic concrete 2 m thick 1x10-2 1x10-3 1x10-4 1x10-4 1x10-2 1x10 2 1x10 3 1x10 4 1x10 5 1x10 6 1x10 7 1x10 8 Neutron Energy (ev) Figure 6. Variation of the escaping neutron flux as a function of the thickness of the shielding. Further reduction can be obtained by changing the composition of the concrete. Different types of concrete have been simulated and the flux of neutrons escaping the shielding is illustrated in Figure 7. Neutron Flux per 100 na of beam loss (n/cm2.s) 1x10 4 1x10 3 1x10 2 normal concrete normal concrete + B normal concrete + B + Ba 1x10 4 1x10 3 1x10 2 1x10-2 1x10 2 1x10 3 1x10 4 1x10 5 1x10 6 1x10 7 5x10 7 Neutron Energy (ev) Figure 7. Energy spectra of neutrons in the center of the shielding for different concrete compositions. D. M4 Bending Magnet Failure Preliminary calculations concerning neutrons streaming from the core into the reactor building in case of an M4 failure were also performed. In this situation the proton beam hits the magnet producing spallation neutrons; the neutron yield was estimated to be 0.12 neutron per proton, mostly in iron. We also calculated the flux of all particles above the reactor pool, including neutrons and photons. The neutron flux was estimated to be φ n 8 10 8 n/cm 2 /s/ma. The photon and the particle fluxes were both of the order of φ 10 10 part/cm 2 /s/ma; obviously, the particle flux is dominated by photons. The neutron, photon and particle effective dose were calculated using Pelliccioni conversion coefficients. These calculations showed that the dose is dominated by the neutrons, which are producing a dose rate of approximately 300 msv/s/ma, whereas the dose rate related to photons is only 30 msv/s/ma. These are extremely large values which should require protective measures. Figure 8 shows the effective dose due to all particles. estimated dose Figure 8. Dose rate due to all particles after a bending magnet failure (in msv/s/ma). Note that in contrast to the particle flux, the effective dose is dominated in this case by that of the neutrons. The energy spectra of neutrons and photons crossing the water surface were also calculated, as illustrated in Figure 9. The neutron spectrum shows a thermal peak due to moderation and a high-energy peak at around 40 MeV. The presence of such high-energy neutrons explains why the particle dose effective is so strongly dominated by neutrons. Dose rate (msv/s/ma)
The photon spectrum shows some prominent peaks, amongst them a peak at 511 kev (e - e + annihilation) and at 2.2 MeV (photon emission for neutron capture in hydrogen). Previous calculations have shown that, in order to avoid such a high dose rate, a possible solution could be to raise the water level in the tank, given that an increase of 20 cm in this level results in approximately a tenfold reduction of the neutron dose. The final design of the magnet will also have an effect on the dose, given that the yoke will be finally placed above the beam tube. Flux (dn/dlne/cm^2/s/ma) 1 0 1x10 9 1x10 8 1x10 7 1x10 6 Neutron flux in water -> air photon flux water -> air 1x10 5 1x10 2 1x10 3 1x10 4 1x10 5 1x10 6 1x10 7 1x10 8 1x10 9 Energy (ev) Figure 9. Energy spectra for neutrons and photons crossing the water surface above the reactor. IV. CONCLUSIONS These calculations, which are part of a greater study performed to achieve the final licensing of the TRADE experiment, show that during normal operating conditions significant shielding is required along the beam transport line, such that the dose rate does not exceed the allowable dose rate in the reactor building (10 µsv/h). The results show the need for protective measures, especially in the case of failure of the M4 magnet, where the dose could injure seriously in less than a second anyone standing near the top of the reactor pool. The results show the clear domination of the effective dose by high-energy neutrons. The contribution of the photons (mostly produced by the interaction of the neutrons and the water) is rather small. REFERENCES 1.The TRADE Collaboration. Triga Accelerator Driven Experiment (TRADE) Feasibility Report. ENEA Report, June 2001. 2. The European Technical Working Group on ADS, A European Roadmap for Developing Accelerator Driven System (ADS) for Nuclear Waste Incineration, April 2001. 3. A. Fassò, A. Ferrari, P.R. Sala, Electron - photon transport in FLUKA status, Proceedings of the MonteCarlo 2000 Conference, Lisbon, October 23-26 2000, A. Kling, F. Barão, M. Nakagawa, L. Távora, P. Vaz eds., Springer- Verlag Berlin, p. 59-164 (2001). A. Fassò, A. Ferrari, J. Ranft, P.R. Sala, FLUKA: Status and Prospective for Hadronic Applications, Proceedings of the MonteCarlo 2000 Conference, Lisbon, October 23--26 2000, A. Kling, F. Barão, M. Nakagawa, L. Távora, P. Vaz eds., Springer-Verlag Berlin, p. 955-960 (2001). 4. Y. Kadi et al. The EA-MC code package, in Proceedings of the Fifth International Meeting on Simulating Accelerator Radiation Environment SARE 5: Models and Codes for spallation neutron sources; OECD Headquarters, Paris, France, July 17 18, 2000. 5. A. Ferrari and M. Pelliccioni, Dose equivalents for mono-energetic electrons incident on the ICRU sphere, Radiation Protection Dosimetry, 55 n 3, 207-210 (1994) (also LNF-94/005 (P) (1994)). Dose equivalents for mono-energetic positrons incident on the ICRU sphere, Radiation Protection Dosimetry, 55 n 4, 309-312 (1994). Fluence to dose equivalent conversion data and effective quality factors for high energy neutrons, Radiation Protection Dosimetry 76, n 4, 215-224 (1998). A. Ferrari, M. Pelliccioni, and M. Pillon, Fluence to effective dose and effective dose equivalent conversion coefficients for photons from 50 kev to 10 GeV, Radiation Protection Dosimetry 67, n 4, 245-251 (1996). Fluence to effective dose and effective dose equivalent conversion coefficients for electrons from 5 MeV to 10 GeV, Radiation Protection Dosimetry 69, n 2, 97-104 (1997). Fluence to effective dose conversion coefficients for protons from 5 MeV to 10 TeV, Radiation Protection Dosimetry 71, n 2, 85-91 (1997).