Daniela Ene. Radioprotection Studies for the ESS Superconducting Linear Accelerator Preliminary Estimates

Transcription

1 !!!!!Accelerator Division!!!!!!!!!!!!!!!!!!!!!!!!!!! ESS AD Technical Note ESS/AD/0003 Daniela Ene Radioprotection Studies for the ESS Superconducting Linear Accelerator Preliminary Estimates 30 June 2010

2 ref v Radioprotection studies for the ESS superconducting linear accelerator Preliminary estimates EDMS Ref: xxxxxxx vx.y Date: Name Affiliation Author D. Ene ESS, Lund Reviewer M. Lindroos, S. Peggs Approver P. Carlsson, M. Lindroos

3 Content 1. Introduction Calculation procedure Design requirements Radiation sources Design assumptions Design approach Description of the simulations Prompt radiation simulation Geometry model Physics model Simulation technique Residual radiation simulation Activation and dose rates estimates Source terms for contaminant transport Results and discussion Shielding configurations Residual field inside the tunnel Radioactive waste. Concrete and soil shielding Sources terms required for the environmental safety assessment Soil activation and potential contamination of the groundwater Production and release of airborne radionuclides in linac tunnel Relevant requirements & policies for ESS shielding and safety linac pre-conceptual design Conclusion Acknowledgements Reference Annex Annex

4 1. Introduction. The design of the high current proton LINAC to be built in Lund, Sweden is under update phase based upon its main parameters and architecture given in (Lindroos M., 2009) and (Peggs, 2009). In the nominal design, Linac will deliver an average power of 5 MW to the target but future upgrading to higher power up to 7.5 MW is also foreseen in the design. A possible view of the ESS facility at the Lund, Sweden site is given in Figure 1 indicating the position of the Linac and its connection with the target station. Figure 1 View of ESS facility at Lund, Sweden site showing the Linac position Linac will operate at a nominal repetition rate of 20 Hz, with a peak current of 50 ma and a pulse length of 2 ms. These parameters correspond to a duty factor of 4% and 2 ma average current or * protons s -1 equivalent to 5 MW beam power at 2500 MeV. The overall architecture of ESS accelerator is schematically shown in the Figure 2, see (Eshraqi, 2010). Figure 2 Layout of ESS accelerator The electron cyclotron resonance proton source is followed by a magneto-static low energy beam transport (LEBT) system that delivers the 75 kev source beam to the radio frequency quadrupole 3

5 (RFQ). The medium energy beam transport (MEBT) that follows is matching the beam into the drift tube linac (DTL) that raises the proton beam energy to 50 MeV. After DTL a superconducting half wave spoke resonator equipped with 14 cryo-modules with a geometric of 0.45 is used in the accelerator section until 200 MeV beam energy. Further, two families of elliptical superconducting cavities are now under design optimization: i) low beta section, with = 0.63 foreseen to raise the beam energy up to 500 MeV and ii) high beta segment ( =0.75) designed to accelerate protons to the final energy. The low-beta accelerating structure consists of 10 modules of four superconducting cavities while the high-beta module might be structured into 21 of high power cryo-modules each housing eight cavities. The designs of the quadrupole doublets as well as their positions related to the cryo-modules in the superconducting elliptical cavities are presently under study. Normal conducting structures, RFQ and DTL, and superconducting structures, spokes and ellipticals, are designed to be able to upgrade the linac to a higher power of 7.5 MW at a fixed energy of 2500 MeV. The higher power can achieved by increasing the current from 50 to 75 ma after adding extra cryo-modules in the zone reserve for this purpose, see Figure 2. The beam energy characteristics relative to the linac length optimized via beam dynamics and apertures studies (Eshraqi, 2010) is given in the Figure 3. Figure 3 ESS linac beam energy along the accelerator length. The figure shows beam energy distribution vs the position along the beam line, margins of the main accelerator elements and the energy gradient achieved by the 2 elliptical accelerator sections. This energy gradient curve represents the basic information used to correlate the energy scale with the tunnel length needed to delimitate the shield blocks. The original plan (Clausen, 2003) was to place the accelerator tunnel slightly under normal ground level and covered it with the excavated soil to form a shielding hill, see Figure 4. In this scenario the klystron building is above ground and waveguides connections are done via diagonal ducts. In terms of straight radiation penetrating through the earth walls this solution has to be investigated for the new parameters of the ESS linac. Consequently to the rough approximations 4

6 accounted, the early predictions (Clausen, 2003) lead to non negligible values of the annually integrated exposure to the members of the public in the adjacent areas of the linac location. Figure 4 Cross section of the linac tunnel and klystron building for the superconducting section Therefore, additionally more detailed analysis focused upon the shielding against the skyshine will be also necessary to be done if this variant will be considered in the selection process of the civil engineering design of the ESS linac. Other variant to be accounted might be the lay-out proposed in the conceptual design of the Superconducting Proton Linac (SPL) carried out at CERN (al., 2006). Here, both linac tunnel and klystron gallery are underground. The depths in the earth (17 m below ground level for the top of the linac tunnel roof and 1 m in depth for the klystron gallery top side, SPL values) are to be optimized for ESS linac specific case. In the present studies the accelerator is considered to be installed underground on Swedish territory and klystrons will be installed in a building on the surface. As the topography of the site (Figure 1) is not yet well known it is supposed that the ground level is flat. Therefore, in this situation the location of the klystron gallery might overlap the tunnel one. For a hill like shaped ground level the klystron building will be placed on the flank of the earth wall, like in the Figure 4. In this case it is very likely that the linac tunnel will be much more lowered in the ground than in the initial design. The aims of these radioprotection-safety studies are as follows: 1 To develop a radioprotection-safety assessment approach, based on the current knowledge and understanding of the ESS accelerator. At this stage, a cautious assessment philosophy has been adopted due to limited knowledge of the machine configuration, lack of specific data for the site and provisional estimates of other parameter values. Conservative generic assumptions have been made relating to the engineering performance of the accelerator and environmental settings, since few specific data are available. 2 To assess the level of accelerator system radioprotection-safety using currently available information. The assessment relates mainly to the normal operation accelerator safety performance and addresses the released activity level required in the assessment of the potential radiological impact upon the environment during the normal running practice. It is also in the intention of this work to gather all information available upon the problem, identify 5

7 problems and to provide guidance for further more detailed studies. The present paper reports first estimates of the radiation protection shielding required for the machine and provide a preliminary characterization of the residual radiation field inside the accelerator tunnel. In this respect two scenarios were analyzed: (a) an accidental full beam loss during 1 s every day and (b) continuous beam loss of 1 W m -1, representing normal operation conditions. Representative loss positions along the accelerator at variable energies were investigated using a simplified geometry model of the linac to assess the lost proton beam prompt radiation field. Dedicated Monte Carlo (MC) simulations with PHITS and MCNPX2.6.0 computer codes were performed to analyze the propagation of neutrons through the tunnel shield wall and its surroundings. The induced radioactivity in the accelerator components, concrete walls, and air inside the tunnel were estimated using the DCHAIN SP 2001 code based on an external neutron source and spallation products derived from PHITS. Ambient dose equivalent rates due to the residual radiation were calculated with the MCNPX code using photon sources resulting from DCHAIN SP The MC estimates were complemented with the predictions of a simple line-ofsight model. A discussion of the results of the MC simulations versus the predictions of the analytical model is also provided. A preliminary shielding design is proposed and a first estimate of the induced radioactivity to be expected is also discussed. The activation of Lund specific soil was further evaluated and the subsequent contamination of ground water at a proton accelerator site is analyzed. A preliminary evaluation of the activity released in the atmosphere from the linac tunnel accounting only for the air change rate inside the entire tunnel volume was further carry out. The factor of conservativeness of the present evaluations is high. The obtained results reflect the degree of knowledge acquired so far about the linac conceptual design and have the limits of the software support resources. 6

8 2. Calculation procedure. 2.1 Design requirements. Radiation limits imposed to comply with in the shielding design are defined by the national legislation of each country. As the ESS facility site was decided to be in Lund, Sweden these constraints were fixed in accordance with Swedish legislation (SSMFS2008:51, 2008) and IAEA recommendations (IAEA, 2004). The dose equivalent rate limit for normal operation was taken to be 0.1 Sv h -1 for public areas, 3 Sv h -1 for supervised area and 10 Sv h -1 for controlled areas. These values ensure that the annual dose equivalent limit of 1 msv for the public and 20 msv limit for the workers for an operation time of 2000 h y -1 will not be exceeded. A shielding design is proposed for each classified area mentioned above. An accident scenario considers a full beam loss (1.248*10 16 protons) at a single point that corresponds to a maximum loss of the 5 MW beam (2.5 GeV). The design requirement for a full beam loss used in CERN (Agosteo, 2001) states that a full beam loss at a localized point must not give rise to a dose equivalent rate outside the shielding higher than 100 msv h -1 ensuring that the accelerator control system is able to abort the beam in a time given rise to an integral dose that remains under permissible limit for any worker. In this stage such safe limit was fixed to 50 Sv in accordance with the regulations of the majority of countries involved in the ESS project. 2.2 Radiation sources. In the work two situations were analyzed for the shielding calculations: i) normal linac operation and ii) accidental full beam loss. For this case the expected integrated equivalent dose from full beam loss outside the predicted shields has to comply with the limits in case of a cut-off time of 50 ms (Eshraqi, 2010) At the time being no possible hot spots along the linac were specified in the design. For beam losses during routine operation the designers of the accelerator have provided as guideline value of less than 1 W m -1. These values correspond to beam losses of 6.24*10 12 /E (E in MeV) protons per meter and per second. According with (Nakashima, 2000) the beam losses during operation should keep the induced radioactivity in the machine at a level sufficiently low to permit hands-on maintenance. Until reliable information will be available it was decided to study the assumed generic routine loss scenario considering that the analysis of its consequences might indicate valuable information for feedback measures. 2.3 Design assumptions. The strategy to design the radiation shield for the linac was based on two assumptions. Firstly, in agreement with (Fasso, 1990) the uniform loss over a distance of about 10 m was assumed approximately equivalent with constant loss concentrated at a single point having 10 times more intensity. Thus the beam losses during operation were represented by point losses concentrated at defined position along the machine of 6.24*10 13 /E (E in MeV) protons s -1 in case of the beam loss scenario. Based on this first hypothesis a simple model was used to estimate the lateral shielding required for linac, see details in chapter PHITS code (Iwase, 2002) and MCNPX ( version Monte Carlo codes were used to simulate the protons transport through this model and 7

9 calculate the equivalent dose rates spatial distributions H*(10) in the specified shield geometry. The resulted neutron attenuation curves in concrete were fitted by an exponential function, approach developed in (Sullivan, 1992) for transmission of secondary neutrons generated by protons. Given beam power and energy the equivalent dose rate H*(10) was approximated by the following attenuation formula: H( ) d H(, d / ) exp( ). 1 2 r, where r and are the spherical coordinates of the current point where the dose rate is measured and d is the total distance transverse by the radiation in the shielding material. H, is the source term and the attenuation length for the large-depth exponential function which model the equilibrium status reached. Making use of Eq. 1 the mentioned above parameters were fitted ( = 90 o ) for subsequent determination of the shield thickness needed to attenuate the radiation dose by the desired factor. The second basic assumption adopted for the shielding design of the proton linac see (Agosteo, 2001) states that a shield designed for a continuous loss during the routine operation is also adequate for an accident loss of a full beam at a localized point providing that the linac cut-off time is short enough to produce an integrated dose below the acceptable limit. Consistent with the above assumption the estimation of the lateral shielding of the linac is illustrated in the Figure 5.` H*(10) (µsv h -1 ) cm public areas cm Shield depth (cm) controlled areas Figure 5 Neutron dose rate attenuation curves in concrete wall of the linac tunnel corresponding to 2.5 GeV proton beam energy (left) and full beam loss (right). Derived dry soil thicknesses required to meet the design constraints are included to demonstrate that the expected integrated equivalent dose from full beam loss outside the predicted shields complies with the limits in case of a cut-off time of 50 ms. The reference tunnel shielding, consisting in one meter of ordinary concrete and subsequent soil layer is analyzed. In the figure, with points is represented resulted PHITS attenuation profile while the continuous line is the curve fitted using the analytical formulation. Derived dry soil thicknesses required to meet the design constraints are included to demonstrate that the expected integrated equivalent dose from full beam loss outside the predicted shields complies with the limits in case of a cut-off time of 50 ms (Eshraqi, 2010). The results for the high energy zone of the linac are presented. It is apparent from the figure that for one meter concrete and subsequent derived soil shield (eight meters) required meeting the public area limit, the integral equivalent dose arising from the full beam loss in 50 ms is below the limit (50 Sv considered in this study). For the same beam H*(10) (µsv h -1 ) µsv cm µsv 6 Sv h -1 * (t cut =0.05 s) cm msv h -1 * t cut (=0.05 s) controlled areas public areas Shield depth (cm) 8

10 switched-off time, with the soil shield (5 m) derived to meet the controlled area criterion it was found prohibited integral dose in case of a full beam accident. Soil thicknesses derived to sustain the maximum beam loss are provided below in the text. 2.4 Design approach. In the Figure 5 the resulted PHITS attenuation profile was processed from the scored dose rate per incident proton 3D dose rate per incident proton matrix for a cross-section plan at around 2 m from the loss point see details in the chapter The appropriate scaling factor was further used to derive the full beam loss case. An attenuation length value of about 60 cm was estimated from the fit of the dose rate profile in the dry soil at 2.5 GeV energy beam loss. Applying the above mentioned method in case of 2.5 GeV proton energy beam loss that correspond to the end of the linac tunnel one can conceive the most conservative Linac shielding design made from one meter ordinary concrete block and further eight meters of dry soil. However according with the ALARA principle it is desirable to minimize the shielding thicknesses. To answer at this requirement besides the high energy beam loss three more beam losses simulations along the accelerator were further investigated. The corresponding energy values of these additional cases of 300, 600 and 1000 MeV were chosen. To have a picture of the shielding required for the warm accelerator zone more two simulations were performed for 20 and 60 MeV proton losses Lost beam power [W] Lost ions (* ) [# s -1 ] Full beam loss [kw] Energy (MeV) Figure 6 Strategy used in simulations (Simulation were done previously the change of the architecture of the accelerator given in Figure 2) The power of the beam loss inside the tunnel is considered constant along the beam line whereas lost beam intensity decreases progressively with the beam energy. Therefore the results found from these simulations were derived following the strategy synthetic represented in the Figure 6. In the figure the lost protons give the beam intensity scaling factors while full beam losses are used to define the accident conditions. It should takes notice that the goals of these simulations were to have a better understanding of the problem and to analyze the possibility to decrease the concrete shielding wall. The position of the machine can be only slightly changed with the earth depth. 3. Description of the simulations. 3.1 Prompt radiation simulation Geometry model. In this initial stage of the design a simplified geometry was adopted, see Figure 7. A rigorous geometry of the facility has to be modeled in the next phase, when more detailed features of the accelerator structure will be better known. The lost protons were assumed to strike a thick copper target (5 X 5 X 5) cm 3 that approximates the copper structure of the quadruples or the cavity. A parallelepiped concrete shell with cross section 9

11 of (300 X 400) cm having 100 cm thickness was used to simulate the tunnel walls. The beam axis was placed asymmetrically (100 cm from one side) inside the tunnel. A soil 300 cm thick layer was placed outside the tunnel concrete walls. The total length of the shielding structure was set to 10 m. The loss location was also asymmetrically modeled along the beam line, allowing 600 cm length in the forward direction. Figure 7 Calculation model used in simulations Figure 8 Dose rate maps in the middleplan of the tunnel model: (left) for proton energy of 300 MeV and (right) for 2.5 GeV The dose rate maps given in Figure 8 for two extreme energies accounted show that the maximum of the dose rates lay within the 10 m length of the simulated tunnel and the development of the cascade is almost contained in the geometry model. Compositional data, nuclide content and impurity elements present within the component materials have been chosen as follows: GLIDCOP Al-25 ( = 8.85 g cm -3 ) for Cu target see (Cepraga, 2007) : Cu %, Al 0.25 %, O 0.22 %, B %, Se %, Fe %, Te %, S %, Zn %, Sb %, As %, Pb %, Sn %, Mn %, P %, Bi %, Cd % (similar composition might be found in (Zucchetti, 2007)); 10

12 ordinary concrete ( = 2.3 g cm -3 ) chemical composition of the research reactor ULYSSE (CEA Saclay) see (Aubert, 2005): O 52.9%, Si 21.8%, Ca 17.4%, C 3.7%, Al 1.5%, Fe 1%, K 0.5%, H 0.4%, Na 0.4%, Mg 0.3%, S 0.1% and other more than 70 traces up to Uranium; Air inside the tunnel ( = *10-3 g cm -3 : N %, O % and Ar %; Soil surrounding the tunnel, a composition ( = 1.6 g cm -3 ) derived from several soil-samples from Krauthausen location (North Rhine-Westphalia) see (Schlogl, 2007): O 50.1 %, Si 37.3 %, Ti 9.7 %, Fe 1.5 %, K 0.9 %, Al 0.4 %, Zn, Pb, Mn, P, Cr, Ni, Cu each less than 0.1 %. Note that for the Krauthausen soil the dry composition is given in the reference (water content is not specified). Cautious assumption of using this dry composition soil was taken for all simulations performed. Recent chemical analysis results of the soil samples taken from the Lund site (Moorman, 2010) allowed a first comparison vs. the results found by means the conservative data. Nevertheless the density of the Lund soil is still unknown so the value of = 1.6 g cm -3 was taken for this last simulation. Detailed compositions of the materials used in simulations are given in the Annex Physics model PHITS code uses an intranuclear cascade model to simulate nucleon-induced reactions and a model based on QMD theory for reactions induced by both nucleons and heavy ions. Statistical decay of compound nucleus is calculated with GEM (Furihata, 2001) extension of the evaporation model implemented in LAHET code system. The good capability of the code to predict neutron spectra from thick targets bombarded with hadrons in the intermediate range energy between 200 to 2500 MeV was previously demonstrated ( Benchmark studies against available experimental data were recently performed in order to verify the accurate description of the reaction models used in the code ( Dose rates conversion factors used by the code are given in (Sakamoto, 2001). In the present calculations default option standing for Bertini intranuclear (INC) cascade was applied. H*(10) (µ Sv h -1 ) for 10 W point loss GeV 1 GeV 600 MeV MeV Axial distance (cm) Figure 9 Neutron, photons and protons production yields as a function of the proton energies considered in this work (left). Axial distribution of H*(10) in concrete shield at various distances from the beam axis (right). 11

13 The use of the PHITS computer code is justified by the fact that at the starting moment of these simulations it was the single available computer code able to estimate the prompt mixed particle fluencies from the hadrons interactions and further ambient dose rates H*(10) by folding these fluxes with the corresponding flux to dose conversion factors. Most recent availability of MCNPX code allowed checking the PHITS estimates. From the comparison of the results was found an excellent agreement explained by the use of the same default INC model. Presently, simulations using FLUKA code (Fasso, 2000) are in progress to account for shielded proton loss (detailed RF cavity geometry accounted). From the very preliminary comparison of the production yields (see Figure 9, left side) derived using various interaction models of these codes it was found a good agreement. Note that the values in the left panel of the Figure 9 represent the production of particles from the entire problem. The axial distributions of the dose equivalent rates at several energies accounted at the concrete layer exit are shown in the right panel of the Figure 9. It is apparent from the figure that the curves are biased in the forward direction showing the cascade developing. The peak laying at around 2 m from the beam impact with the target is better observable with the thickness increasing of the concrete wall. Therefore to be on the safer side all the results presented in this paper are referring to the axial location corresponding to the maximum dose observed, around 2 m downstream from the target. One can be seen in the graph (Figure 9, right) that higher radiation levels are found for lower energies. This because at lower energy the primary protons are stopped in shorter length of the target, while above 1 GeV the development of the hadronic cascade occurs in the target leading to a more extended source distribution in the longitudinal direction. The neutron spectral fluence distributions in the concrete depths are given in the left panel of the Figure 10 for 2.5 GeV beam loss energy. The energy spectra depicted in the figure are typical for high energy proton beam shielded by concrete The shape of the spectrum is generally characterized by a high energy spallation peak with a maximum at 100 MeV, an evaporation ridge around few MeV and a closed to 1/E distribution in the intermediate energy range. As can been seen in the graph the high energy neutrons dominate the attenuation up to long distances. The shape of the neutron energy distribution remains unchanged, only the amplitude decreases in depth. Thus the high energy peak is present up to large depths inside the subsequent soil. Whereas in the low energy region concrete has a high shielding performance due to elastic scattering effect of the hydrogen. Heavier elements like iron with big inelastic cross sections at high energy neutrons are expected to be more efficient shielding materials being able to reduce the major contribution to the fluence given by high energy neutrons. 12

14 Figure 10: Neutron spectral fluences at various depths in concrete, (left) and iron ( right) for lost protons of 2.5 GeV. As resulted from the right side of the Figure 10 the major component of the spectrum shifts to lower energy due to the elastic scattering and 24 kev iron resonance. Low energy neutron flux (E<hundreds of kev) attenuates slower than high energy component being build-up from the inelastic scattering of the high energy neutrons. Typical equilibrium spectra with a shape which is independent of the shield thickness are reached with a change in amplitude of about 2 orders of magnitude with increasing of 100 cm shield thickness. Further, the intermediate energy flux is reduced quite effectively by the concrete layer placed after two meters of iron. The influence of other secondary particles produced in interactions with linac structure and concrete shield was investigated for 2.5 GeV and it was found that their contribution to the total H*(10) is negligible. For instance the ratio of H*(10) induced by photons to the total H*(10) is less than 10% showing that the neutrons are the dominant in the shielding estimates Simulation technique The splitting/importance variance technique was used for the PHITS/MCNPX simulations based on the geometry model described in chapter The nuclear data library underlying all present simulations is based mainly on the ENDF/B VII evaluated file. In order to obtain results with reasonable low uncertainty 1E+7 histories were tracked in all simulations. 3.2 Residual radiation simulation Activation and dose rates estimates For activation analysis PHITS was coupled with DCHAIN-SP (Kay T., 2001) that is using FENDL/A-2 based activation data set together with FENDL/D-1 and ENSDF decay data libraries. It calculates the nuclide inventory, radioactivity, decay heat and gamma-ray energy spectra being able to read directly the distribution of the residual nuclei produced by PHITS code. The radioactivity in the air was estimated using PHITS code option that allows the folding of particle track-length spectra with evaluated isotope production cross sections. The code contains a database with evaluated neutron and proton interaction cross sections of alpha, 14 N, 16 O targets cases that is used for the conversion of the air constituents into the radionuclides of interest. MCNPX photon transport simulations were further carried out to estimate the ambient dose equivalent rates at surface and at one meter distance from the irradiated element. 13

15 DCHAIN derived photon source implemented in the MCNPX modeling were defined in VITAMIN-J library energy group structure. ANSI/ANS conversion factors to obtain the biological equivalent dose response function were employed. In order to account for the accumulation effect an operation history covering 40 year for a duty factor of 0.7 was used in calculations. Results of these simulations are required for radioactive waste characterization. Else an operation campaign of 50 day was accounted. Subsequent cooling times up to 300 years after irradiation period were assumed. Dose rates were determined for different decay times until six months for the high energy segment of the tunnel Source terms for contaminant transport The production of radioactive nuclides in the surrounding soil was determined using the same calculation procedure as for the shielding structures. One cubic meter cell around the beam loss was used for air activation study. 14

16 4. Results and discussion. 4.1 Shielding configurations The shielding study was focused on a shielding configuration of one meter concrete tunnel wall and subsequent soil layer, which is referred in the work as the reference shielding design. From the analysis of results discussed in chapter 2.3 the additional soil layer might reduce the integral dose as follows: Limit ( Sv h -1 ) soil thickness required to meet the criterion for 10 W point beam loss (cm) Integral dose from full beam loss ( Sv) More than 590 cm thickness of the soil is required to guarantee an integrated dose in case of accident bellow the acceptable limit with sufficient margin to cut off the beam in 50 ms. For other proton beam loss energies results are shown in the left hand of the Figure 11. Estimation of the shielding size by analytical means requires knowledge of the source term and of the shielding proprieties of the material (attenuation length). Because for soil, these data are scarce in the literature the concrete material was employed instead. Using the values of the attenuation lengths and source terms from (Magistris, 2005) the calculated concrete shielding thicknesses required to reduce the ambient dose equivalent to levels below the criteria assumed in this work are given in the right panel of the Figure 11. Figure 11: Left: Neutron dose rate attenuation curves in soil shield following one meter of ordinary concrete. Right: size of the concrete shielding blocks required to reduce the dose equivalent rate below 0.1 Sv h -1 (the limit for public exposure) 3 Sv h -1 (supervised radiation area) and 10 Sv h -1 (controlled radiation area). As expected at beam loss energy above 1 GeV small difference was found between the 15

17 shielding thicknesses. This is explained by both interaction cross section and the attenuation length which become approximately independent of energy above 1 GeV. Soil shielding thicknesses derived approximately by scaling the concrete results with the ratio of the two materials (2.35/1.6) are given in the Table 1. Table 1: Size of the soil shielding layers following one meter of concrete required to reduce the dose equivalent rate below 0.1 Sv h -1 (the limit for public exposure, 3 Sv h -1 (supervised radiation areas) and 10 Sv h -1 (controlled radiation area). Energy [MeV] Length [m] Public areas Supervised Areas Controlled areas [m] [m] [m] Mass (tonnes)* to be dug 4.489E E E+4 Mass (tonnes)* of removed soil 1.697E E E4 Corresponding to a tunnel with the surface of 4m X 3m The values in the table are in an overall good agreement with the previous results obtained from simulations, except for the public criterion where the dose is slightly underestimated. Note that in the analytical estimates a safety factor 3 was applied. One can roughly conclude from this comparison that the degree of the conservativeness of these preliminary estimates covers the recommended safety margin. The Table 1 provides also the minimum soil mass to be dug in case of a tunnel with a cross section of (4 m X 3 m) and a concrete wall of one meter. The last row of the table gives the mass of the soil to be removed (corresponding to the tunnel volume) and transported or used for the berm. The Figure 12 shows the effects of the soil composition upon the dose rate results in the soil shield. It is obvious than Lund soil composition (14% water content) gives less conservative results than Krauthausen soil (dry). Figure 12: Dependence of the H*(10) with the soil composition 16

18 A more complex analysis is in progress aiming to find more accurate results (use of the specific primary data and shielded beam loss) and look for alternative configurations varying the concrete shield thickness and composition as well as the water content in the soil. Simulations are foreseen also to analyze the shielding against skyshine. Dose rates profile across the tunnel in HETB zone is shown in Figure 13. In simulations a composite shield block of two meters of iron and subsequent one meter of ordinary concrete was used. In the model the shielding was placed one meter far from the loss point. Figure 13 Neutron dose rate attenuation curves in iron and concrete shielding for the HEBT zone corresponding to 10 W point proton beam loss (left) and full beam loss (right). The graph shows that for a point beam loss of 10 W (see chapter 2.3) the shield of two meters of iron is enough to decrease the dose rate to a level corresponding to the controlled area criterion. Nevertheless, the high level of the integral dose at the end of the iron shield (139 Sv) found for a cut-off time of 50 ms is not acceptable. Additional ordinary concrete layer might reduce the integral dose as follows: Limit ( Sv h -1 ) Concrete thickness required to meet the criterion for 10 W point beam loss (cm) Integral dose from full beam loss ( Sv) Corroborating the above results one can conclude that two meters of iron block followed by 40 cm of ordinary concrete layer will be required to shield the bending magnet in the linac to target connecting zone. It should be noted that additional evaluations considering backscattering of the neutrons from the target area are necessary to size correctly the magnet shielding for insuring the working conditions in the area. 17

19 4.2 Residual field inside the tunnel Activation results for copper target and for the inner wall concrete 20 cm stratum are provided in the Figure 14 as specific activities at 2.5 GeV proton beam loss energy. The dominant nuclides representing however more than 90% from the total activity are shown also. The duration of an experimental campaign, 50 days of irradiation was analyzed. High total activation value of approximate 5.2*10 7 Bq cm -3 coming from accelerator structure at 1 hour decay time is due to 64,62,61,65 Cu radionuclides while at long times 59 Fe, 60 Co, 3 H and 63 Ni are dominant. The main contributor at short times is 64 Cu produced from thermal neutron capture in 63 Cu (high activation cross section of 4.5 barn). Specific activity (Bq cm -3 ) h 1 d Total 62 Cu (T1/2 = 9.74m) 64 Cu (T1/2 = 12.7h) 61 Cu (T1/2 = 3.34h) 65 Ni (T1/2 = 2.52h) 51 Cr (T1/2 = 27.7d) 1 w 1 m 1 y Decay time (s) 10 y Specific activity (Bq cm -3 ) Fe (T1/2 = 2.73 y) Co (T1/2 = 5.27y) 63 Ni (T1/2 = 100.1y) Mn (T1/2 = 303 d) 3 H (T1/2 = 12.3y) 7 Be (T1/2 = 53.1d) h 1 d 1 w 1 m 1 y 10 y Total 45 Ca 7 Be 11 C 55 Fe 37 Ar 3 H 22 Na 56 Mn 238 U Decay time (s) Figure 14 : Obtained specific activity distributions versus decay time after 50 days of irradiation. Results correspond to 2.5 GeV proton loss energy. On the left dominant radionuclides produced in case of the copper target and for the inner 20 cm layer of concrete (right). Contribution of the tunnel concrete wall to the total activity and consequently to the H*(10) is not significant. Concrete contains 0.4% of sodium which results in thermal neutron produced 24 Na, being the major contributor to the dose rates near the walls shortly after the beam is switched off. The 37 Ar isotope issuing from Ca activation dominates the concrete activity until 6 months while for longer decay time 3 H and further 238 U together its ascendants become majority giving rise nevertheless to a small activity value. The decay time distribution of the specific activity in concrete depth until 100 cm with layers of 20 cm thickness scales approximately as is depicted in the right panel of the Figure 17, for 40 years of operation. The estimated air activity (see Figure 15) is much lesser than the activation of the materials in the tunnel thus associated exposure by air inhalation is expected to be much lower with respect to the external doses coming from accelerator structure. In the figure the results correspond to one cubic meter of the air surrounding the proton beam loss at the end segment of the accelerator (2.5 GeV). Shown estimates were derived for an irradiation history of 50 days. Radioactive isotopes in the air surrounding the beam loss are the short-lived positron emitters that are produced in oxygen and nitrogen by spallation reactions (T 1/2 few minutes), 7 Be and 3 H produced by spallation reactions as well as the 41 Ar by thermal neutron capture in the natural argon. The left panel of the Figure 15 shows the saturation effect occurring for short lived radionuclides in the air during the irradiation. Long lived nuclides 7 Be and 3 H do not reach the saturation in 50 days of operation. 18

20 h 1 d 1 w 1 m Specific activity (Bq cm -3 ) Total 7 Be(T1/2 =53.1d) 3 H(T1/2 =12.3y) 1 y 10 y 38 Cl(T1/2 =37.2h) 41 Ar(T1/2 =109.3m) 39 Cl(T1/2 =55.6m) Figure 15 Obtained specific activity of one m 3 of air surrounding the beam los inside the tunnel for 50 days of irradiation. Build-up of the radionuclide activation with the irradiation time (left) and beam-off activity distributions as a function of decay times (left). Results correspond to 2.5 GeV proton loss energy. One hour after the operation shut-down time 7 Be and 11 C are the main radionuclides produced in the air given rise to a total activation of about 1 Bq cm -3. While for longer decay times 7 Be and 3 H later on are dominants. Note the large difference of about one order of magnitude at shut down (left figure) due to the readily decaying of the short lived isotopes. In order to give a picture of the whole linac tunnel classification, activation calculations were performed for the accounted loss beam energies. Results for the copper target are presented in the left panel of the Figure Ar(T1/2 =35d) Decay time (s) 11 C(T1/2 =20.4m) Figure 16: Total specific activity in the copper target versus decay time for accounted beam loss energies (left). Right: low energy (E < 20 MeV) neutron flux distributions on top and charge residual product distributions on bottom for minimum and maximum energy of the accounted beam losses. One can be seen that the contributions of the lower energy are higher than for the high energy case. As has been discussed above (chapter 3.2) highest radiation level is expected for the lowest energy. 19

21 In addition to the explanation provided above, other reason underlines this result. The power lost inside the tunnel is assumed to be the same (10 W in a point) for all energies but the intensity of the lost beam depends with the energy (see Figure 6). Therefore higher number of lost particles compensates the less number and energy of secondary particles. The right side of the Figure 16 illustrates this by comparing low energy neutron flux and the residual nuclide distributions in the copper target for low and high energy loss proton beam accounted in this work. It is apparent from these graphs that due to the highest proton beam intensity both neutron and spallation contributions (right side of the charge distribution) to the activation are higher at lower energy. Notice that the spallation products have the dominant contribution to the copper activation. It has to be stressed that activation and subsequent residual dose estimations for the accelerator structure require detailed geometry and material compositions. Second, a revision of the generic proton loss definition, of 1 W m -1 used in this analysis is required. Characterizing the beam loss in terms of the percentage from the beam loss intensity along the accelerator is more adequate; see the experience gained at existing facilities (Popova, 2004). Predicted dose rates after activation are given in the Table 2 for high energy tunnel segment. Table 2: H*(10) in Sv h -1 inside the tunnel corresponding to the high energy zone (2.5 GeV) Time 1 m in air Contact [h] Cu target Concrete wall Total Cu target Concrete wall Total E E+05 24(1 day) E E (1 week) E E (1 month) E E (6 months) E (1 year) 0.02 Residual dose levels obtained in these preliminary calculations are high. One hour after shutdown in the high energy zone of the accelerator, dose rates of few hundreds of Sv h -1 were found. These radiation levels are about four orders of magnitude less compared with the operational doses estimated one meter from the loss point (see Figure 8). As resulted from this table the residual activation field inside the tunnel is arising mainly from copper structure activation. The concrete wall contribution is negligible (< 10%). Even though the decaying of the short lived radionuclides like 61 Cu, 60 Cu, 24 Na drops the dose rate close to the target to a factor 2 after one day of cooling, due to the contribution of the other long lived relevant nuclides ( 60 Co, 54 Mn, 55 Fe, 7 Be) the dose rate remains high. In the high energy zone of the linac continuous accessibility inside the tunnel in the beam-off stage is possible after more than one month cooling time. If intervention is required previously an occupancy factor of minimum 10 hours per year allows meeting the constraint of 2 msv per person and per intervention, the criterion used in CERN for the design of the nuclear facilities (Sentis, 2006). Contact dose rate values significantly higher than the limits show that a remote system for handling 20

22 and transportation of the dismantled component might be considered. For the item extracted from the accelerator structure in the high energy zone even after six months cooling time the contact dose is higher than 1 msv h -1, the criterion used in the ITER design (Zucchetti, 2007) Consistent with the above analysis the high energy end of the linac might be classified as radiation controlled area with restricted access. The definition of the intervention inside the tunnel at the end part of the accelerator might be required. To protect the personnel during the handling operations the access and transportation paths should be set-up. It is expected that the residual dose profile to follow the trend of the activation distribution along the tunnel. Lower residual dose levels will be found if detailed primary data and less conservative assumptions will be used. Simulations rigorously modeling the geometry of linac and getting use of the real material compositions will allow assessing the self shielding of the gamma ray in the structure. The design parameter used here to derive the proton loss source is considered coarse. Therefore detailed beam loss characteristics are required to obtain more accurate residual dose levels. 21

23 4.3 Radioactive waste. Concrete and soil shielding. The induced radioactivity in the concrete wall of the linac tunnel was calculated for each 20 cm thick layers of concrete. Figure 17, left side gives the total specific activity in the five layers of the concrete wall. Specific activity (Bq g -1 ) h 1 d 1 w 1 m 1 y (0-20)cm (20-40)cm (40-60)cm (60-80)cm (80-100)cm 10 y ,300 y Decay time (s) Figure 17: Right: Residual activity as a function of the wall concrete thickness 40 y 0.7 duty cycle. Sample from high energy zone is shown. Green line stands for the exemption limit. Left: waiting time before the induced radioactivity in the shielding decreases below the exemption limit, as a function of depth in the concrete and soil. Note that for soil the 40 K activation was not accounted. The reported values refer to 40 years of continuous operation at 0.7 duty cycle followed by different cooling times until 300 years. Note the difference in shape at large cooling times against the Figure 14 due to 3 H accumulation effect. Activation in the concrete wall decreases with almost a factor 2 each 20 cm in depth. At the shut-down of the installation (40 years of irradiation) one meter layer of concrete shield will have a radioactivity exceeding the exemption limit. For these considerations it was chosen as exemption limit the value of 1 Bq g -1, a general agreed figure (IAEA, 2005) used to classify clearance of the radioactive waste. Clearance is defined as the removal of the radioactive materials or radioactive objects within the authorized practices from any further regulatory control by the regulatory body. The time required to wait until the concrete wall might be considered as exempt waste is shown in the right panel of the Figure 17. A concrete amount of about 1.496E+4 tonnes (volume of 6.365E+3 m 3 ) corresponding to the reference configuration of the tunnel (one meter thickness) will require final disposal. If a deferred strategy of the complete dismantling is chosen, the timeframes to be used for the assessment comprise minimum 85 years of active institutional control of the site, followed by an additional 100 years of passive institutional control. The acceptability of induced radioactivity in the concrete might be judged by comparing the specific activities of the produce radionuclides with the exemption limits of the Sweden legislation. Presently, there is work going on by the Swedish Radiation Safety Authority (SSM) to implement the Europen Commission and IAEA recommendations in this respect ( 22

24 Therefore the specific activities of the most important radionuclides scored in the first 20 cm of concrete at shut-down are shown comparatively with IAEA recommended exemption limits (IAEA, 1996) and (IAEA, 2005) in Table 3. It should be noted that for a mixture of radionuclides an additive weighted rule has to be applied (IAEA, 1994) and the obtained value has to comply with the 1 Bq g -1 exemption limit. Table 3 Specific saturated activities of the most important radionuclides scored in the first 20 cm of concrete wall, first 20 cm of adjacent soil and IAEA recommended exemption limits. Isotope T 1/2 Concrete Soil IAEA exemption limit (Bq g -1 ) (Bq g -1 ) (Bq g -1 ) 3 H y E E Be d E E Na 2.60 y E E Na h E E P d E E S d E E Ca d E E Sc d E E Mn d E E Fe 2.73 y E E Zn d E E Eu E E Eu E E K* Value at shut-down is reported The waste classification system proposed by (IAEA, 1994) based on clearance concept might be used for ESS facility produced radioactive waste. One example of the use of this approach is given in (Ene, 2008). The residual radioactivity of the soil layers surrounding the high energy zone of linac was scored also for the same irradiation history as above. The obtained results are shown in Figure 18. Activation concentrations scored inside first 20 cm at the entrance in each one meter soil layer are given in the figure. As resulted from the left panel of this figure activation in soil exceeds the exemption limit and remains higher for long decay times. In fact this large value is due the 40 K, whose contribution to the total activity is dominant (> 98%) after 10 years decay time, see right side of the figure. Because the 40 K activation value found in the first 20 cm of soil is below the exemption level of 10 Bq g -1 (IAEA, 2005) one can characterize the soil as radwaste only by means the other radionuclides. The Table 3 gives specific activities of radionuclides in the soil while the Figure 17, left panel shows the minimum waiting time required until the contaminated soil might be exempt (values calculated without 40 K contribution). First two meters of soil in the vicinity of the accelerator has to be treated as nuclear waste, at least for 15 years. Dismantling and disposal of this large amount of soil is more difficult than similar operations for the concrete waste. It is resulted from this analysis that with the proposed increase of the thickness of the concrete tunnel wall the contamination of the adjacent soil is not avoided and radwaste soil has to be considered in the facility decommissioning waste management. 23

25 Figure 18 Residual activity of the soil as a function of the thickness 40 y 0.7 duty cycle: total activation (right) and separate contribution of 40 K and of the other remaining radionuclides (right). Sample from high energy zone is shown. Green line stands for the exemption limit. Note the difference against the early evaluation (Clausen, 2003) which can be explained by the fact that the thickness of the concrete wall is larger in this study and the IAEA exemption limits seems to be more relaxed than the German constraints used previously. Follow-up detailed calculations of the radioactive waste arising from ESS linac operation are necessary for a more accurate estimation of the amounts of waste and their characteristics and subsequent cost estimates. It also required a ESS policy agreed by SSM for radioactive waste classification. 24

26 4.4 Sources terms required for the environmental safety assessment Release of the radioactive materials into the environment is a major topic to be addressed by ESS design. The activation of the soil surrounding the accelerator tunnel and the activation of the air inside the accelerator tunnel are first steps in assessment of the environmental impact of ESS accelerator. Soil activation estimates define the potential for groundwater contamination, while the air activation in the tunnel gives input for normal airborne contamination released via the tunnel stack. Based on the source terms evaluated here further complex studies are to be fulfilled to model the migration of the contaminant through the environment and to assess the impact Soil activation and potential contamination of the groundwater The standard deviation for the flux determination in the soil beyond 1 m of concrete wall was less than 5%. The results for the ten most important isotopes from the point of view of the environmental risk assessment (Thomas R. H., 1988) are shown in the Table 6. Calculations have shown that most of these radionuclides are produced by spallation. In order to account for cumulative effect exposure periods of one year and 40 years of continuous operation with a duty factor of 0.7 were used for the activation analysis. In the table the third column shows the specific activity estimated at high energy end of the linac. Scoring was done for the first 20 cm of soil outside the concrete. Table.6 Activity concentration in first 100 cm of soil surrounding the concrete wall after 40 years of continuous operation. IAEA generic levels for liquid release and Swedish ALI values are also provided for comparison. Isotope T 1/2 Specific activity Activity ALI (Bq cm -3 ) (Bq) (Bq) IAEA water release limits 3 H y 9.67E E E+09 1.E+12 ingestion 7 Be d 9.77E E Na 2.60 y 1.54E E E+07 1.E+05 external 24 Na h 5.35E E E+08 1.E+08 external 32 P d 8.34E E E+07 1.E+06 ingestion-fish 35 S d 5.86E E E+07 1.E+09 ingestion-fish 45 Ca d 1.05E E E+07 1.E+10 ingestion 46 Sc d 2.02E E Mn d 1.59E E E Fe 2.73 y 1.94E E E Zn d 3.15E E+06 The distribution of the radionuclides in depth of the soil shield shows that the most activation occurs in the first meter of soil surrounding the concrete wall. Column four of the table gives the activation over the volume of the first soil layer calculated for the reference tunnel size (Volume = 7.07*10 9 cm 3 ) supposing that the soil is uniformly activated over the length of the accelerator as at the high energy end of the linac. More realistic is to consider the soil activation averaged over one meter layer in depth, which will reduce the values in the column 4 of the table by a factor 5. 25

27 In a conservative approach one can compare direct the radionuclide concentration in the soil with annual limits on intake (ALI) stipulated by the Swedish legislation (SSMFS2008:51, 2008) shown in the table. The results in the table indicate that for this conservative scenario only 35 S does not represent any environmental hazard. The concentrations of other nuclides are high enough to be a potential concern. Less conservative comparison with the generic clearance levels for liquid release recommended by IAEA (IAEA, 1998) given in the last column of the table together with the main exposure pathways accounted leads to the conclusion that 3 H, and 45 Ca are also unlikely to represent any off-site problem. It should be noted that the clearance levels mentioned above were derived to assure the compliance with an annual dose to the critical group of 10 Sv y -1. In CERN the constraint limit of 30 Sv y -1 used for annual release determines an annual constraint of 4.2E+11 Bq for 22 Na and 3.2E+10 for 3 H (Agosteo, 2005). These values were estimated considering the total effective doses to the critical group per unit release of 7E-16 SvBq -1 for 3 H and 9.7E-20 SvBq -1 for 22 Na. Mainly 3 H and 22 Na are a matter of concern because they are in soluble chemical form and have large T 1/2, thus they may leach into the groundwater and transfer from the site to the ground water table (Thomas R. H., 1988). The radionuclide concentrations of these two nuclides previously estimated is much bellow compared with the CERN annual release constraints. Note that a leaching fraction is associated with each radionuclide. The average value of the fraction is 100% for 3 H and 20% for 22 Na (Thomas R. H., 1988). It is clear that from this rough examination one cannot draw the conclusion that with a concrete shield of 1 m thickness the present shielding design is not adequate from the point of view of the environmental impact. A real evaluation of the environmental impact has to be done for the site of the facility. The models and scenarios for calculating the dose to the releases of the radioactivity from nuclear installations recommended in (IAEA, 2001) have to be used as guidelines for the calculation of the effective dose to the public due to releases of the activity from ESS accelerator. Results of a hydro-geological study of the construction site and site specific data will provide a realistic evaluation of the exposure pathways as well as an accurate estimation of the fraction of the ground water that might reach the public water supplies. Additionally, an ESS specific policy is needed to set the constraint limits for the site emissions. Generic contaminant transport calculations for ESS were performed in the frame of ESS-PP Deliverable 8.4 (Prolingheur, N. et al., 2009). These calculations are done for a normalized nuclide vector (the Lund soil composition and the source term were unknown at that time). In combination with the Lund soil composition data and the source term calculated here this contaminant transport approach can be used for the estimation of more realistic contamination values at the site. Further detailed studies are foreseen to estimate the degree of conservativeness of this preliminary evaluation and get more refined results and their sensitiveness against various parameters like: depth of the concrete shield, water content in the soil, more realistic irradiation history, beam loss strength, etc. Effect of the neutron streaming through the ducts housing waveguides will be further assed when the connection of the tunnel with the klystron gallery will be studied and designed. To be on the safe side the ESS accelerator design has to include protective measures to isolate the soil from groundwater in order to prevent the exchange of the contaminated water. An interface around the accelerator and the surrounding soil as in SNS design (Dole, 1998) making use of waterproof membrane and in addition a clay cap over the berm for surface water diversion are required to ensure the negligible levels of the groundwater contamination. 26

28 4.4.2 Production and release of airborne radionuclides in linac tunnel The obtained estimate of the concentration of the radioactivity in air shows high level, see (Figure 14). Estimation of the worker exposure after shut-down during an eventually required limited access time inside the tunnel assuming the air activation level found here is behind the purpose of this study. The estimated air radionuclide activities are used below for a preliminary evaluation of the activity released in the atmosphere from the linac tunnel. A rough assessment accounting only for the air change rate inside the entire tunnel volume is conservatively considered in this phase with the purpose of having a starting point for the next more detailed analysis. In fact the assessment of the released air radioactivity through the stack requires a complex study that might have various solutions and that have to involve other parameters related mainly to the transit time (time between the production until the exhausting into the environment of the radioactive air). In Sweden, the regulation of discharges is based on International Basic Safety Standards (IAEA, 2004), recommendations from ICRP 60 (ICRP60, 1991), the EU Basic Safety Standards and other directives, such as (IAEA, 2000). The annual effective dose limit to members of the public is 1 msv per year. Dose constraints set by the SSM are used and may vary between practices; for example a dose constraint of 0.1 msv per year is set for combined releases from nuclear facilities. The model for radioactive air release assessment (Sullivan, 1992) is based upon the assumption that for a given radionuclide i the presence of ventilation can be accounted by using an effective decay constant that includes the physical decay constant in addition to ventilation term, r: 2 where r =. Q is the ventilation rate in air volume per time unity and V is the volume where the mixing occurs. Thus r is the number of air exchanges per unit time. Based on this adjustment and considering the mixing of the air activation and an infinite irradiation time the released air saturation concentration is given by the equation: A i = R i T with 3 R i is the nuclide production rate by the air activation inside the tunnel and T is the operation time per year, 5000 h in this calculation. The activity of the airborne radionuclides released per year under continuous ventilation is summarized in Table 4 for three values of the air exchange rate. The values in the table were derived by using an average value of the nuclide production rate inside the tunnel model. As was stated above the reported air activation values in the chapter 4.2 were scored on a cell of one cubic meter around the beam loss. Here the air volume outside this cell was taken into account too. Additionally it was assumed that the air activation at the high energy end of the linac is uniformly extended over the entire length of the tunnel. It is apparent from the table that a smaller ventilation rate is recommended to be used to decrease the activity release and therefore the total annual dose to the critical group. The main radionuclides contributing to the released air activity and consequently to the total annual dose are 11 C and 13 N followed by 41 Ar and 14,15 O. In this discussion 7 Be and 3 H were not included because they dominate by far the total activity released in a year. The large values found, especially for 27

29 these long lived isotopes suggest that the model used above seems to be too rough. Table 4 Equilibrium activity of the air released in the linac tunnel as a function of the air exchange rate ( r ). nuclide T 1/2 (s) Saturated activity (Bq cm -3 ) Annual release (Bq) r = 0.1 r = 1 r = B 2.020E E E E E B 1.740E E E E E N 4.400E E E E E+11 8 Be 6.700E E E E E-03 6 He 8.081E E E E E+13 8 Li 8.400E E E E E C 1.926E E E E E N 7.130E E E E E O 7.060E E E E E O 1.222E E E E E C 1.218E E E E E N 5.982E E E E E Ar E E E E E+15 7 Be E E E E E+16 3 H 3.89E E E E E+17 By using an additional ingredient, t transit (the travel time from the production zone to the outlet point) in the set-up of the ventilation system, the concentration activity at the released location is given by: 4 If this approach is taken, for long lived radionuclides like 7 Be and 3 H the expression in the middle of the Equation 4 is significantly less than unit. For all others, saturation is readily achieved during relatively short period of operation of only few hours (see Figure 15). Another scenario that might be accounted is to keep sealed airborne radionuclides during operation and for a time t cool following the shut-down of the beam to allow for decay before the initiation of the mixing and release. In this case the corresponding value of time t cool should be added to that of t transit in the argument of the last exponential function. The optimized solution can be found by varying the three possible parameters: ventilation rate, transit time and/or waiting time before release. For the follow-up refined analysis a more realistic geometry sizing of the tunnel is required. A common point of view related to exhausting of the air radioactivity from entire facility is also necessary. Knowledge of the position of the stack and subsequently of the airduct lengths, as well as more accurate tunnel geometry will contribute to find the best option for the air release equipment. Further detailed studies related to air activation inside the tunnel in case of the failure of the ventilation system will required a reference database for the inhalation dose conversion factors agreed by SSM Subsequent assessment of the environmental impact of airborne radionuclide will be based on the source term resulted through the analysis of the production and release of the air activation in the linac tunnel. Like for the assessment of the environmental impact consequent to the water release an agreed approach for the modeling and an ESS specific policy to set the constraint limits for the site emissions will be required. 28

30 6. Relevant requirements & policies for ESS shielding and safety linac pre-conceptual design ESS shielding design policy has to provide project-level shielding design criteria more stringent than the minimum Swedish legislation requirements. This is necessary because margins are needed for calculations and design uncertainties. Additionally, during the detailed design, provisions have to be taken to demonstrate that ESS personnel exposures will be as low as reasonably achievable (ALARA). An overview of the relevant requirements to be included in the ESS shielding and safety design policy is given in the following: Criterion Value used in this work Proposed value External exposure H*(10) stipulating the accessibility to the outer surface of the shields Limit of worker exposure during maintenance Limit of worker exposure for worst case beam control accident taking credit for automatic beam cutoff system function. Dose limits for the worst case design basis beam control faults insured by passive means (for instant thickness of the shielding). Limits for discharges of water to an unrestricted area Limits for release of the air to an unrestricted area public areas: 0.1 Sv h -1 supervised area: 3 Sv h -1 controlled areas=10 Sv h -1 Accumulated H*(10): 20 msv year -1 2 msv per person per intervention Remote handling : >1 m Sv h -1 A safety factor of 3 to be applied to above figures. Hands-on margins Hands-on limit: 10 Sv h -1 Remote handling 50 Sv (20-100) Sv ALI values taken from (SSMFS2008:51, 2008) Workers: msv Public: 10 msv Ensuring that site discharges (from all sources) are below 10% from the annual dose (1 msv y -1 ) to the critical group, consistent with (SSMFS2008:26, 2008) and (IAEA, 2000) Ensuring that site discharging (from all sources) are below 10% from the annual dose (1 msv y -1 ) to the critical group, consistent with (SSMFS2008:26, 2008) and (IAEA, 2000) Radioactive waste classification Exemption limit =1 Bq g -1 In compliance with (SSMFS2008:51, 2008) which stipulate the exemption limits at 10 * ALI for some radionuclides & accounting for IAEA recommendations based on clearance index concept (IAEA, 2004) 29

31 Additionally a reference database for conversion coefficients from activity to effective dose (Sv Bq -1 ) to be used through over the linac radioprotection assessments is necessary to be set and approved. For these coefficients the values recommended by (IAEA, 2004) are usually applied. 6. Conclusion An approach (including both methods and tools) has been developed to assess the radioprotection-safety level of the ESS accelerator. Conservative assumptions, input parameters and tools used in this approach reflect current assessment capability, as well as limited knowledge and understanding of the accelerator system, presently under preliminary design. Monte Carlo simulations complemented with analytical predictions were used to estimate the thickness of the lateral shielding of the accelerator tunnel. A tunnel configuration placed underground in the earth is proposed as reference for further detailed evaluations. An ordinary concrete shielding wall of one meter thickness was considered in the analysis with the purpose to minimize the contamination of the soil surrounding the concrete. Size of the earth shielding following one meter of concrete was derived to meet the exposure criterion for controlled, supervised and public area. It was concluded from this investigation that a shielding thickness of about six meters of earth will be required for the option designating the site as supervised area and respectively eight meters of earth for public area designation. Additionally, a preliminary estimate of the required shielding in the HEBT zone was performed. Calculations have shown that two meters of iron block followed by 40 cm of ordinary concrete layer will be required to shield the bending magnet in the linac to target connecting zone. This estimation accounts only for 1 W m -1 beam loss along the magnet and aims to reduce the prompt radiation to a level of 1 Sv h -1. Additionally, it should be noted that evaluations considering backscattering of the neutrons from the target area are necessary to size correctly the magnet shielding for insuring the working conditions in the area. The proposed thickness of the shielding guarantees an integrated dose in case of accident below the acceptable limit with sufficient margin to cut off the beam. Analysis of the possibilities to reduce the shield thicknesses in compliance with ALARA principle has to be done. Residual field inside the tunnel was further evaluated using a simple geometry to model unshielded beam loss consequences upon the machine structure, concrete wall and air inside. Residual dose levels obtained in these preliminary calculations are high. One hour after shut-down in the high energy zone of the accelerator, dose rates of few hundreds of Sv h -1 were found. Consistent with the obtained results the high energy end of the linac might be classified as radiation controlled area with restricted access. Activation of the concrete wall and adjacent soil shielding was further estimated aiming a preliminary quantitative evaluation of the radioactive waste arising during the lifetime of the facility (40 years of operation). A concrete amount of about 1.496E+4 tonnes (volume of 6.365E+3 m 3 ) corresponding to the reference configuration of the tunnel will require final disposal. If a deferred strategy of the complete dismantling will be chosen, the timeframes to be used for the assessment comprise minimum 85 years of active institutional control of the site, followed by an additional 100 years of passive institutional control. First two meters of soil in the vicinity of the accelerator has to be treated as nuclear waste, at least for 15 years. It results from this analysis that with the proposed increase of the thickness of the concrete tunnel wall the contamination of the adjacent soil is not avoided and radwaste soil has to be considered in the facility decommissioning waste management. Release of the radioactive materials into the environment was further assessed, guiding the evaluation and defining the potential sources for environmental contamination analysis, a major topic to be addressed by ESS design. Based on these estimations further complex studies are to be 30

32 fulfilled to model the migration of the contaminant through the environment and to assess the impact. Activity concentrations of radionuclides of major concern in terms of contaminant migration into the groundwater were derived for the first one meter of soil surrounding the concrete tunnel wall. The direct comparison of the obtained radionuclide concentration in the soil with annual limits for liquid release is too conservative to judge the environmental impact. Therefore a real evaluation based on site specific data will be needed. To be on the safe side the ESS accelerator design has to include protective measures to isolate the soil from groundwater in order to prevent the exchange of the contaminated water. A preliminary evaluation of the activity released in the atmosphere from the linac tunnel accounting only for the air change rate inside the entire tunnel volume was further carried out. It was found from this conservative analysis that a more realistic model will be required to set-up of the ventilation system inside the accelerator tunnel. Various scenarios of the air radioactivity release from the accelerator tunnel are to be accounted and comparatively analyzed. The main radioprotection-safety issues related to the ESS linac normal operation performance were investigated by means a conservative approach. An update of the whole content of this analysis has to be given. Detailed studies based on more realistic assumptions, input data and site specific parameters are necessary to check the degree of conservativeness of these preliminary investigations. Supplementary studies related to the shielding against skyshine as well as the analysis of the consequences of the failures and severe accident occurring have to be performed. A radioprotection-safety policy to set the criteria and guidance in agreement with SSM requirements is absolute necessary in the frame of the ESS design. Relevant requirements and policies for ESS shielding and safety linac pre-conceptual design are outlined in the chapter 6 of this study. Acknowledgements. The author thank to Dr. M. Lindroos and Dr. S. Peggs for useful discussions on technical configuration of the ESS accelerator, to M. Eshraqi for clarifications and data supplied, to K. Aquilonius (Studsvik) and M. Brandin for clarifications regarding the Swedish legislation, to H. Hahn for drawings and discussions and to M. Jarosz for postprocessing the air activation results. 31

33 Reference Agosteo, S, et al Nuclear Instr. & Methods B. 2007, Vol. 265, 2, p Agosteo, S, Silari, M Preliminary shielding calculations for a 2 GeV superconducting proton linac. Geneve : Technical report: TIS-RP/TM/ , CERN, Agosteo, S., et al RPD. 2005, Vols. 115, N01-4. al., Gerigk F. et Conceptual design of the SPL II. A high-power superconducting H-linac at CERN. Geneve : s.n., Aubert, F Inventaire et Compositions Chimique des Matériaux du Réacteur ULYSSE de Saclay, DAPNIA/ SENAC/E/05-193/NT Cepraga, D. G ENEA, Bologna, Italy, private communication Clausen, K. et al The ESS project Vol III, Update: Technical report, ISBN Dole, L Preliminary assesment of the nuclide migration from the activation zone around the proposed spallation neutron source facility. Oak Ridge : s.n., ORNL/TM Ene, D. et al Activation studies of the shielding structures for the EURISOL 4 MW target station. s.l. : PSI, ARIA workshop. Eshraqi, M. et al Conceptual design of the ESS Linac. Kioto, Japan : s.n., Eshraqi, M Private communication Fasso, A. et al Status and prospective for hadronic applications. Lisbone : s.n., Monte Carlo 2000 Conference. Fasso, A., et al Shielding against high energy radiation, Numerical data and functional relationships in science and technology. Berlin : Vol. 11, Ed. H. Shopper, Springer-Verlag. Furihata, S. et al The GEM Code -A simulation program for the evaporation and fission process of an excited nucleus,. s.l. : JAERI-Data Code , JAERI, Tokaimura, Japan., MCNPX Monte Carlo N-Particle Transport Code System for Multiparticle and High Energy Applications; IAEA Clearance Levels for Radio-nuclides in Solid Materials: Application of the Exemption Principles, Interim. IAEA TECDOC-855, Vienna Application of the Concept of Exclusion, Exemption and Clearance. s.l. : Safety Standards Series, No. RS-G-1.7, Classification of Radioactive Waste, A Safety Guide, IAEA Safety Series No.111-G-1.1, Vienna Clearance of materials resulting from the use of radionuclides in medicine, industry and research. Vienna : TECDOC-1000, Derivation of Activity Concentrations Values for Exclusion, Exemption and Clearance

34 Generic Models for Use in Assessing the Impact of Discharges of Radioactive Substances to the Environment. Vienna : s.n., Safet Report Series, International Basic Safety Standards for Protection against Ionizing Radiation and for the Safety of Radiation Sources Safety Series No International Basic Safety Standards for Protection against Ionizing Radiation and for the Safety of Radiation Sources. Vienna : s.n., Safety Series No Regulatory control of radioactive discharges to the envioronment. s.l. : Safet Guide SW- G-2.3, ICRP Recommendations of the International Commission on Radiological Protection, Publication 60, Pergamon Press, Oxford and New York. Iwase, H., Niita K., Nakamura T., J. Nucl. Sci. Technol. 2002, Vol. 39(11), p Kay T., Maekawa F., Kosako K., et al DCHAIN-SP 2001: High Energy Induced Radioactivity Calculation Code (in Japanese). JAERI- Data Code Lindroos M., et al The ESS Superconducting Linear accelerator. Berlin, Germany : SRF09, Magistris, M Shielding requirements and induced radioactivity in the 3.5 GeV SPL CERN-SC RP-TN. Moorman, R FZK, Julich, Germany, Private comunication Nakashima, H et al J. Nucl. Sci. Technol. Supplement1, 870. Peggs, S. et al Vancouver, Canada : PAC09, Popova, I. and Gallemeyer F Full dose radiation dose analysis for SNS accelerator Prolingheur, N. et al Activity transport in groundwater at the Lund site- A first model analysis, vol. D ESS-PP, Contribution to the environmental impact assessment of ESS. available via ftp://ess-pp@ftp.psi.ch. Sakamoto, Y. and Yamaguchi Y Dose conversion coefficients in the Shielding Design Calculation for High Energy proton Accelerator Facilities JAERI-Tech Schlogl, B. et al Calculation of the Activity Inventory of Transportable Radionuclides in Soil and Groundwater for Large Neutron Sources. s.l. : EURISOL DS/Task5/TN-07-01, Sentis, M., et. al Calculation of inetrvention doses for the CNGS facility. 2006, Vol SSMFS2008: ISSN SSMFS2008: Stralsakerthetsmyndighetens foreskrifter om grundlaggande bestammelser for skydd av arbetstagare och allmanhet vid verksamhet med joniserande stralning Swedish Radiation Safety Authority. Sullivan, A.H A Guide to Radiation and Radioactivity Levels Near High Energy Particle Accelerators. s.l. : Nuclear Technology Publishing Ashford, TN23 1JW England, Thomas R. H., Stevenson, G. R Radiological safety aspects of proton accelerators.. Technical Reports Series No. 283, Vienna, IAEA : s.n., Zucchetti, M., et al ANE. 2007, p

35 GLIDCOP Al-25 = 8.85 g cm -3 Element Wt% B O Al Cu Fe Mn P S Bi Zn Te Cd Sn Sb Se Pb As Annex 1 - Chemical compositions of materials used in simulations Isotope {Z*1000+A} # cm -3 * E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E-08 34

36 Ordinary concrete = 2.35 g cm -3 Element ppm-wt H 3500 Li 20 Be 0.95 N 0 B 16 F 0 C O Na 920 Mg 2775 Al Si P 262 S 3400 Cl 100 Ar 3.06 K 5150 Ca Sc 3.3 Ti 960 V 39 Cr 39.2 Mn 252 Fe 9640 Co 6.82 Ni 26 Cu 30.4 Zn 77 Ga 3.67 Ge 1.24 As 9 Se 1.05 Br 2.93 Rb 25.9 Sr 328 Y 9.45 Zr 54.7 Nb 2.97 Mo 1.45 Ru 2.41 Rh 65.7 Pd 16.5 Ag Cd 0.6 In 0.03 Sn 1.3 Sb 9.5 Te 1.2 I 4 Cs 2.07 Ba 132 La 9.3 Ce 16.2 Pr 2.16 Nd 8.45 Sm 1.88 Eu 0.36 Gd 1.72 Tb 0.26 Dy 1.38 Ho 0.3 Er 0.8 Tm 0.12 Yb 0.74 Lu 0.1 Hf 1.1 Ta W 2.35 Re Os 0.33 Ir Pt 2.15 Au Hg 0.12 Pb 44.4 Bi 0.2 Th 2.4 U 1.14 Annex 1 35

37 Isotope * # cm -3 * E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E-09 36

38 Isotope * Ordinary concrete -- continuation # cm -3 * E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E-09 Annex E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E-09 37

39 E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E-09 {Z*1000+A} 38

40 Soil Lund = 1.6 g cm -3 wt-% S# 1 S# 2 S# 3 Averaged H 2 O Element (ppm wt) S# 1 S# 2 S# 3 Averaged Al Ca Si Fe K Mg Na P Ti Element S*# 1 S*# 2 (ppm wt) S*# 3 Annex 1 Averaged Be V Cr Mn Co Ni Cu Zn Ga As Se Rb Sr Mo Ag Cd Te Ba Pb Bi U La Ce Pr Nd Sm Eu Gd

41 Dy Ho Er Tm Yb Lu Be V Cr Mn Co Ni Cu Zn Ga As Se S# 1, 2, 3 = Sample #1, 2, 3 40

42 Soil Lund - continuation Isotope * # cm -3 * E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E-10 Annex E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E-10 41

43 E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E-10 *{Z*1000+A} E E E E E E-10 42

44 Annex 1 Air composition = *10-3 g cm -3 Isotope * # cm -3 * E E E E E E E E E E E E E E E E E E E E E E E E E E-13 43

45 The computer diagram used for activation estimates Annex 2 Geometry and materials description MCNPX/ PHITS Neutron flux in i th cell Residues in i th cell CINDER 90/ DCHAIN-SP 2001 h Activation products & Irradiation Scheme one example for final waste assessment : 40 years irradiation at 5 MW MCNP H*(10) 44