Use of the OSCAR-Fusion V1.4.a code for a preliminary assessment of the ACP contamination within the main ITER water cooling circuit

One of the main objectives of ITER is to produce 500 MW of power from a deuterium-tritium plasma for several seconds. This goal presents two inherent challenges: firstly, in-vessel components will require active cooling to remove the heat coming from the fusion reaction (i.e., mainly fast neutrons and alpha particles). Secondly, the materials exposed to the neutron flux will yield activated corrosion products (ACPs) in all primary cooling circuits of ITER. From a safety point of view, ACPs are one of the contributors to the Occupational Radiation Exposure (ORE), they represent a source of radiological waste and also contribute to the source term for accidental scenarios involving the loss of primary confinement. Therefore, ACPs assessment is key to estimate radiological impact for nuclear workers and the public. ITER nuclear safety engineers adopted OSCAR-Fusion v1.4.a code to assess the ACPs inventory in the Integrated Blanket ELMs and Divertor (IBED) cooling loop. This paper describes the selection of input data, the modelling of the circuits and the operational scenarios used in OSCAR-Fusion calculations. This study also examines the outcomes of such calculations, notably in terms of ACPs inventory, emphasizing the impact on the ORE and highlighting its driving parameters. Additionally, the paper offers recommendations for better ACPs management in the context of the ITER project and in accordance with the ALARA principle.

Original content from this work may be used under the terms of the Creative Commons Attribution 4.0 licence.Any further distribution of this work must maintain attribution to the author(s) and the title of the work, journal citation and DOI.

Introduction
ITER D-T pulsed plasma operations imply huge fast neutrons generation 4 and active cooling of the components exposed to high and continual heat fluxes [1].Hence, the combination of activated materials and transport phenomena occurring within the water coolant leads to the accumulation of Activated Corrosion Products (ACPs) within the primary cooling systems.Among the many isotopes present in the ACPs inventory, the gamma emitters constitute not only one of the main hazards to the Occupational Radiation Exposure (ORE) [2,3], but also a radiological source term for accidental scenarios involving the loss of primary confinement [4,5].
Recent work performed in the framework of the Nuclear Integration Engineering (NIE) program [6] highlighted that the major impact on the ORE relates to activities on the primary cooling systems.Consequently, the need to update the ACPs source term for future analyses remains a crucial objective [7].The ITER Radiation, Safety and Environment (RSE) group decided to re-evaluate the inventory of ACPs.This new assessment targeted activity levels during the machine life cycle, and its distribution within the systems and buildings designed to provide the radiological confinement functions.
A precise ACPs estimation must also encompass several aspects [8]: e.g.materials composition, piping and equipment surface finishing, activation rates in the regions under neutron 4 The ITER project, www.iter.org.flux, coolant chemistry, thermal-hydraulics, geometry of the cooling loop, coolant cleaning and filters efficiency.The RSE group secured the support of CEA-Cadarache experts on ACPs and used a bespoken software developed by CEA, the OSCAR-Fusion v1.4.a code.OSCAR-Fusion [9] is a fusion adapted version of the OSCAR5 code [10] already in use in the French Nuclear Industry [11] and validated against fission operational data [12].The goal was to perform the most accurate estimate of the ACPs inventory, identify the key contributors to the loop contamination and propose dose reduction measures to limit the impact on the ORE.Good practices from the fission nuclear industry were also considered [13].
This interdisciplinary activity aims at validating the methodology for ACPs source term assessment through OSCAR-Fusion code while providing a collection of robust and comprehensive results to be used in future ITER safety analyses.The rationale for this investigation was to revise the ACPs source term used in past ITER safety analyses [14,15] and not yet fully updated nor revised.The scope of this activity is the Integrated Blanket Edge-Localized Modes (ELMs) and Divertor (IBED) loop [16], which is the main primary cooling system of ITER in terms of both coolant volume and total wetted surface.
In terms of the organization of this study, section 2 provides a brief overview of the IBED loop, its 'clients', the primary and auxiliary systems, and the operational modes.Section 3 describes the characteristics of the OSCAR-Fusion code.
Section 4 reports on the development of the OSCAR-Fusion input in the context of the ITER project, with emphasis on three aspects: neutron reaction rates, cooling circuit parameters and operational scenarios.
Section 5 provides the calculation output, focusing on both ACPs masses and activities in different regions of the circuit.
Finally, section 6 draws the overall conclusions and offers recommendations for future studies.

Description of the IBED loop
ITER in-vessel components require active cooling to effectively transfer the heat generated by the products of the D-T fusion reaction i.e. mainly alpha particles and neutrons.The main in-vessel components are the First Wall-Blanket6 (FW-BLK) [17,18] and the divertor 7 (DIV) [19,20].Additional in-vessel components are the In-Vessel Coils (IVCs), mounted on the Vacuum Vessel8 (VV) surface right behind the FW-BLK, the Diagnostics and Auxiliary Heating Systems installed in the equatorial and upper ports.The Integrated Blanket ELMs and Divertor (IBED) Primary Heat Transport System (PHTS) [16,21] is designed to provide cooling to all the above mentioned in-vessel components, adequately removing the heat coming from the plasma and hence avoiding unwanted high temperature transients.This system can operate at higher pressure and temperature than those required in plasma mode [22] to enable water baking and purge the IVCs from tritium and other impurities.The baking loop foresees a heater and an heat exchanger and it is connected in parallel to the main loop.A Chemical and Volume Control System (CVCS) is connected to the PHTS through two linear headers to remove ions and particles in the coolant through resin beds and mechanical filters respectively and enable coolant chemistry control.The IBED loop considered in this work includes the above mentioned systems providing a barrier to both water and ACPs release: figure 1 below shows a simplified flow diagram of the IBED loop with the details of the ratio of the total flow rate in different branches of the circuit during both dwell/plasma burn and baking operations.Additional auxiliary systems of the IBED PHTS are the draining and drying circuit, not considered for the present study since OSCAR cannot simulate the discharge of one circuit into another.However, the transfer of ACPs to the drain tanks through the draining connections can and shall be addressed for both maintenance and accidental scenarios; the results in terms of ACPs concentration in the coolant shown in section 5 can enable such estimation.As shown in figure 1 below, the IBED loop can be sub-divided in 4 areas: -In-flux regions, i.e. the regions of the loop within the bioshield activated by the neutrons (inside the red rectangle in figure); -Out-of-flux regions, i.e. portions of the IBED loop belonging to the PHTS outside the bioshield, including piping, valves and main Heat eXchangers (HXs); -The baking circuit connected in parallel to the PHTS (inside the orange rectangle in figure 1); -The CVCS circuit connected in parallel to the PHTS (inside the green rectangle in figure 1).It is also worth recalling that IBED is the main system of the Tokamak Cooling Water System (TCWS), both in terms of coolant inventory, flow rate and power to be transferred to the heat rejection system.
The IBED is a rather large and complex assembly of piping and cooling equipment (e.g.valves, heat exchangers) containing activated coolant and corrosion products to represent an unprecedented challenge in terms of inspection and maintenance activities [8].Typical maintenance operations in ITER will be performed after 12 d from the plasma shutdownin such scenario, the ACPs represent the main source term for the ORE for the areas in which primary cooling systems are installed [2].This is due to the presence of gamma emitters radio-isotopes, such as Co-60, in the deposits of corrosion products on the inner surfaces of pipes, valves, pumps and heat exchangers.For such components, hands-on operations are considered in the ORE assessment.In addition, the ACPs contribute to both generation of radiological waste and source term for accidental scenarios (e.g.loss of coolant event).Therefore, the ACPs assessment for IBED is one of the main indicators for the overall safety performance of the ITER project.
The supply and return pipelines for the FW-BLK outside the bioshield, the so called 'Jungle Gym' constitutes a good example of an ORE hazard.There are 18 Jungle Gyms around the tokamak and each one is equipped with several valves.This complexity could potentially yield ACP-contaminated wet surface and significant exposure time to perfom inspection/maintenance tasks.
Therefore, the Jungle Gym return region is selected as one of the representative areas for the ORE impact in following section 5.

Description of OSCAR-Fusion
The OSCAR-Fusion code is the TCWS PHTS version of the OSCAR code initially developed for the PWR reactor cooling system [9].Since the coolant used in the PHTSs and PWRs is the same and the main differences concern the materials (presence of copper alloy) and the neutron spectrum and flux (see also section 4.1 and [23]), the ACP transfer modelling in the OSCAR-Fusion code is identical to that in the OSCAR code.Therefore, the OSCAR-Fusion code adopts the same control volume approach for modelling, which can be summarized as follows: -The systems are divided into discrete control volumes or regions based on their geometric, thermal-hydraulic, neutronic, material and operational characteristics; -Each control volume can encompass six media: metal, inner oxide, outer oxide/deposit9 , particles, ions and filter (including ion exchange resins and particle filters); -The code considers the following elements: Cr, Mn, Fe, Co, Ni, Zn, Zr, Ag, Sb and Cu along with their respective radioisotopes; -For each isotope (both stable and radioactive) in each medium of each region, a system of mass balance equations is solved using the following equation: transfer where m i is the mass of isotope i in a given medium (kg), t the time (s), the advection term of isotope i (kg s −1 ) and J i transfer a transfer mass rate of isotope i between 2 media or 2 isotopes (kg s −1 ).
The main transfer mechanisms accounted for in the code include corrosion-release, dissolution, precipitation, erosion, deposition, advection, purification, activation and radioactive decay (see figure 2).
The dissolution-precipitation model was enhanced in version 1.4, enabling OSCAR to more accurately calculate the incorporation of minor species (e.g. 60 Co) into oxides (see [12] for a detailed description of this new model and the other main ones).
In past ITER analyses, PACTITER [24] and OSCAR version 1.3 codes have been used, whereas IO has been using OSCAR-Fusion v1.4.a since June 2022.

Differences in the parameters of the datasets
The main difference concerns the corrosion rates of SS and Cu alloy.For the calculations performed in the past at ITER, the MOOREA law was used for both SS and Cu.The MOOREA law is an empirical model that depends on the chemistry, temperature, material, manufacturing process and operational time.It was defined under PWR conditions for SS and Ni base alloys.The model assumes an initial constant value for the first two months, followed by a gradual decrease over the next ten months, reaching a constant value thereafter.For SS, the MOOREA law depends on pH, saturation, temperature and time.For Cu alloy, it depends on saturation, temperature and time and assume the pH T to be kept constant at 7.
To reduce conservatism and to use corrosion laws defined in ITER conditions, e.g [25].OSCAR offers the possibility of using a time power law for corrosion rate that consistently decreases over time instead of the constant corrosion rate after 1 year of operation of the MOOREA law (see section 3.3).The power law in OSCAR can also depend on saturation and temperature, similar to the MOOREA law.
The differences in the other parameters of the datasets have an impact of the second order.

Validation of OSCAR-Fusion
While no operating experience (OPEX) exists for fusion reactors, the OSCAR-Fusion code benefits from the validation of the OSCAR code, which is based on a unique worldwide OPEX: the EMECC expertise assessments of the PWR circuit contamination [26].To date, approximately 430 campaigns of the gamma surface activity measurements of the PWR primary and auxiliary systems have been conducted in 76 different units in France and abroad since 1971 using the EMECC system.In addition to the gamma surface activities measurements, the OSCAR results have been compared to other on-site measurements [27]: volume activities and chemical element concentrations.
The validation of the OSCAR code encompasses a wide range of conditions including water temperatures ranging from 20 • C to 350 • C, both laminar and turbulent flow regimes, reducing and oxidizing environments, as well as alkaline and acidic conditions [28].Considering the similarities between the PHTS conditions in fusion reactors and PWRs, the OSCAR-Fusion code can benefit from the validation of the OSCAR code.

Corrosion and release rates for AISI and Cu-alloy
For both materials, the corrosion rates considered are time power laws.Their time dependence is defined by Belous [25].The ratio between the release rates and the corrosion rates is considered to be between 0.25 and 0.3 (it is based on the ratio defined for the MOOREA law under PWR conditions).For AISI, to be consistent with the MOOREA corrosion rate used for PWR simulations under reducing conditions, the corrosion rate and the release rate in g•m −2 •s −1 considered for ITER simulations are: where t is the time in s, −0.548 is defined by Belous [25], f T = 124100e −51000 RT the temperature correction factor for T ⩽ 250 • C (defined for PWRs) and f sat the saturation correction factor (f satCor = 3 and f satRel = 50 defined for PWRs).Figure 3 presents the comparison of the corrosion rates for AISI steel at 250 • C under unsaturated conditions (f sat = 1) between the power law used for ITER simulations and the MOOREA law at a pH T of 7.0 used for PWR simulations.For Cu-alloy, the corrosion rate and the release rate in g•m −2 •s −1 considered for ITER simulations are: where t is the time in s, −0.721 is defined considering Belous [25], f T = 700e −28500 RT the temperature correction factor for T ⩽ 250 • C and f sat the saturation correction factor (f satCor = 3 and f satRel = 50 considering the same values as those for AISI).
The temperature correction factor, f T , is defined to have a reduction factor of 5 between 180 • C and 100 • C which is consistent with the reduction factor of 3 on the release rate between 150 • C and 100 • C measured by [29].The release rate is defined to have the same values as those measured by [29] (see figure 4).
The corrosion rate calculated by OSCAR-Fusion is compared to Belous' data [25] and Obitz' measurements [30] in figure 5. Compared to Belous' data at 100 • C and 180 • C, the corrosion rate in OSCAR-Fusion is higher by a factor of 3-7.The difference could be justified by the conditions of Belous' experiment, where the water in the autoclave is static and probably saturated.As reported by Belous et al [25], corrosion varies by a factor of 15 with changes in flow velocity from 0 to 4 m s −1 .In comparison to Obitz' data at 250 • C, the corrosion rate in OSCAR-Fusion is higher by a factor of 2-10, depending on the fluid velocity.However, the duration of the Obitz' experiment is very short (1 week), and its extrapolation to the long-term is uncertain.It should be noted that the corrosion and release rates used for AISI and Cu-alloy in this study do not depend on pH (the pH T considered is 7.0), nor on the potentially cyclic slightly oxidizing conditions due to radiolysis species not totally recombined by H 2 , nor on the fluid velocity.As reported by Belous [25], corrosion rate may vary by a factor of ∼20 depending on the water chemistry and test conditions.The current IBED design considers a pH at 25 • C varying between 7 and 9 for plasma and baking operations, respectively.The pH is controlled through the injection of chemicals into the coolant, thus preventing excessive acidity and mitigating the risk of corrosion, especially during the baking operations.

Building of OSCAR-Fusion input
As for every model, building the input for OSCAR studies implies a data selection/simplification: to ensure the validity of the results, it is therefore paramount to define which data are relevant.Therefore, it is first necessary to define what relevant means in the three main subdomains of the OSCAR input preparation, which are the neutron reaction rates, the loop modelling (geometric, thermal-hydraulic and material characteristics) and the operational scenario specification.
Among all the elements exposed to the neutron flux, only the ones transported outside of the bioshield can be defined as relevant; similarly, among the many reaction rates occurring in the under-flux materials, only the ones contributing to the activity or to the dose due to ACPs exposure can be selected.This led to the selection of seven elements: Co, Cr, Cu, Fe, Mn, Ni and Zr on the basis of the contribution to the dose rates in the regions outside the bioshield, i.e. the elements that can be transported outside the irradiated area.Table 1 shows the preliminary list of isotopes considered based on elements selection.
This list of isotopes can be further reduced in function of the study by considering additional selection criteria, such as the relatively low half-life and/or negligible contribution to the dose.Isotopes with negligible contribution to the equivalent gamma dose and/or with relatively short half-life are excluded for ORE studies.Additional criteria for accidents and waste management scenarios might apply.

Neutron reaction rates assessment
The NUCLEO section of the input file for OSCAR-Fusion code contains the data related to the interactions of the neutrons with the materials; it is broken down in three sub-sections: -Element, -Decay chain, -Reaction Rates.
The first two sub-sections are 'standard' ones, i.e. they contain isotope-specific information such as type of element, mass number, natural abundance and, for radioisotopes, decay type and decay constant.The reaction rate sub-section shall be adapted to the type of study, providing the specific reaction rates for all the relevant isotopes.In particular, the nuclides inventory variation during the entire simulation shall consider the contributions of both activation and de-activation rates, i.e. the disappearance of some relevant radionuclides by the interaction with the neutrons or by decay.To provide an adequate input for NUCLEO, the ITER RSE group performed dedicated calculations to assess the neutron spectra during FPO in the FW-BLK cooling channels and hence calculated the reaction rates in four different regions of a FW-BLK module along the radial coordinate with respect to the plasma.The neutron energy spectra (figure 6) in TRIPOLI 315 energy group structure in beryllium layer of the first wall panel and in the cooling channels of the FW-BLK outboard  The spectra were considered in the isotope inventory analysis for the ACP production in the cooling channels of FW-BLK outboard equatorial module.
FISPACT code [31] was used to provide the reaction rates for different materials in FW-BLK regions with a good level of detail in terms of both isotopes activation and spatial distribution.In addition, the disappearance of four radioactive isotopes (Co-58, Co-60, Cu-64 and Mn-56) were also investigated and the respective disappearance rates included in the reactions list.
The method of the assessment of the rates of the single-step neutron-induced nuclear reactions per atom of a target nuclide based on the use of the collapsed cross-sections-group averaged cross-sections folded with the group flux.
The nuclide balance equations after folding the crosssections with the group flux Y ik -the yield of nuclide i from the fission of nuclide k.
Regrouping the terms in (1) to separate the contributions of the single-step neutron-induced nuclear reactions with the target nuclides The last two underlined terms are considered to derive: -The rate of the specific single-step neutron-induced nuclear reaction A j (n, F)A i per atom of a target nuclide A j dNi dt -The rate of the specific single-step neutron-induced fission reaction A k (n, F)A i per atom of a target nuclide A k dNi dt The rates ϕ σ have been assessed with FISPACT-2007/EAF2007/EAF2010.
As a result, the reaction rates lists were selected for OSCAR-Fusion calculation.Simulation of activation of the materials during hydrogen/helium plasmas foreseen in the PFPO is not considered in the present work due to the expected extremely small contribution to the overall ACPs source term in comparison to FPO.However, the PFPO campaigns are simulated to evaluate the overall evolution of the mass of the CPs within the loop prior the start of the FPO.

IBED loop regions and materials
The OSCAR-Fusion input model for the IBED loop used in past calculations (i.e. with OSCAR-Fusion v1.3) has been updated on the basis of new data available from TCWS designers.The main improvements of this update consist of: -The addition of equatorial ports clients and their piping distribution at equatorial level; -The addition of the IVCs and upper port clients at upper level; -An increased detail of the HXs volumes and wet surfaces (from one region to six); -The introduction of the inlet and outlet isolation valves regions between the upper rings manifold and the jungle gym regions; -The checking and updating of the wet surfaces and hydraulic diameters on the basis of input from TCWS section for all the out-of-flux regions of the IBED loop; -Optimization of the FW-BLK regions with respect to previous studies (smaller regions number).
Because of the listed improvements, the updated model provides a general higher level of detail compared to previous analyses.The review and support from the CEA code

Operational scenario
The operational scenario for the OSCAR-Fusion model is based on the ITER Research Plan 2018 [34]: such a scenario considers two PFPO campaigns and six FPO campaigns; it also specifies that 30 d of baking operation will occur prior and after each campaign.

Pre-fusion power operation.
Two PFPO campaigns including baking are simulated in the first cycle when the plasma power and activation level are considerably lower.The purpose is to simulate the initial condition of the circuit and track the evolution of the non-ACPs prior the start of FPO.Therefore, the first cycle simulation relies on a simplified approach: constant temperature (70 • C during PFPO and 240 • C during baking) and no activation of the materials.The very first day of operation is simulated as a 'zero power' burn (i.e.no activation in the in-flux regions), followed by 29 d of baking to correctly initialize the calculation.Table 3 below summarizes the operational duration for PFPO.

Fusion power operation.
The overall FPO is simulated through six identical FPO campaigns each one generating 5 × 10 26 neutrons corresponding to an overall 4700 h of D-T plasma pulse duration and 3 × 10 27 generated neutrons [1].The sum of the neutron generation during FPO 1, 2 and 3 corresponds to 5 × 10 26 neutrons-this value corresponds to the neutron generation in each subsequent FPO campaigns (i.e.FPO 4-8) [34]; to reduce the computational time and simplify the model, the first three FPO campaigns are grouped in one.At the end of the last FPO campaign plus baking, additional 12 d of cold shutdown are simulated to track down the evolution of the ACPs due to natural decay after shutdown.Hence the overall duration of the simulation amounts to 5549 d, corresponding to approximately 15 years of continual operation.
Table 4 below summarizes the operational duration for FPO.

Period parameters used in OSCAR.
The thermal hydraulic parameters of the IBED PHTS have been modified based on input from the TCWS design team; additionally, lithium injection is simulated during FPOs and baking operation to keep the pH T at 7 (or slightly above 7 during FPOs dwell time).Table 5 below summarizes the main thermal-hydraulic parameters used in the model.

Preliminary studies 4.4.1. ORE contribution-'the Zirconium case'.
A preliminary calculation has been performed to estimate the contribution of the different elements to the gamma dose rate as function of the time.In particular, the Zr element contribution to the total dose rate was checked since previous OSCAR-Fusion studies for ITER did not include Zr in the list of elements undergoing neutron activation.
For such test, gamma contact isotope dose rate coefficients for a DN250 Schedule 80 (i.e. one manifold of the jungle gyms) were primarily calculated by CEA through the MERCURE code [35] and entered in the OSCAR-Fusion input dataset.
Then, the OSCAR-Fusion code calculated the dose rate generated by each isotope by multiplying the isotope dose rate coefficient by the isotope surface activity of a jungle gym pipe during the operation; the isotope dose rates at the last day of the D-T operation, or EoL are considered.Figure 8 below shows the normalized results of the isotope contributions to the maximum dose rate from one second to 100 years after the EoL and with a cut off at 10 −6 ; we can observe the constant dominant contribution of Co-60 to the overall dose rate and the absence of Zr-93 and Zr-95 from the list of main contributors.
Thus, for ORE studies: -Among all ACPs isotopes, Co-60 is the key contributor to the dose rates; -Zirconium can be excluded from the list of elements to be considered for ORE studies to optimize calculation time.
However, in the case of significant corrosion-erosion of CuCrZr, this simplification might need to be revisited.

Homogenization of the operation.
The simulation of ITER operation in OSCAR-Fusion represents a challenge in terms of computational needs.The complete sequence of operation requires many periods in the OSCAR input to negatively affect the ergonomics of the input/output (slowing down of pre and post processing actions) as well as the required time to perform a complete calculation.
To overcome this issue, the current approach simplifies the reference operational scenario by grouping different modes, significantly reducing the number of periods and hence optimizing the calculation time.
However, such a simplification implies the potential loss of information that might influence the results.Therefore, we performed a preliminary parametric study to evaluate the impact of the scenario homogenization on the results.Three operational scenarios were investigated with OSCAR: B (Brutal), C (Compromise) and D (Detailed).All the scenarios consider: -A loop 'conditioning' of 1320 d to simulate the deposit and oxide formation in a brand-new loop operating for both PFPO campaigns and baking operations; -One typical FPO campaign (32 plasma sessions corresponding to a neutron budget of 5 × 10 26 neutrons), including 30 d of baking at the end of the plasma operation -A final CS with no plasma and coolant flowing at 70 • C of 100 d.
The details of the FPO operation for the three scenarios are shown in table 6.
As shown in figures 9 and 10, compared to scenario D, both B and C show higher activity values during the plasma pulse due to the 'compression' of the burn phase at the end of campaign and sessions respectively, and hence to an underestimation of the natural decay of the isotopes.However, scenarios C and D show a very good agreement since start of the baking operation and consequent final cold shutdown phase, whereas the B scenario significantly overestimates the activity in the loop.
Hence, scenario C is selected for the simulation of the entire operational campaign corresponding to the production of 3 × 10 27 neutrons and 4700 h of D-T plasma.For completeness, figure 11 shows that in terms of mass, the three scenarios show practically perfect agreement: because the mass of CPs dominates the overall mass concerning the mass of ACPs (see also section 5.2) for both in-flux and out-of-flux regions.

OSCAR-Fusion results: ACPs
This section shows the results obtained with OSCAR-Fusion code in terms of mass and activity in the system due to the ACPs.Finally, also the contribution to the ORE for some relevant isotopes is shown as example.In the following, we will refer to EoL as the last day of D-T plasma operation, whereas EoL + Baking represent the last day of the final baking operation.

Corrosion and release rates
To illustrate the corrosion and release rates using OSCAR-Fusion (see section 3.4), figures 12 and 13 show the corrosion and release rates for Cu-alloy (HV_CuAlloy) and AISI (HV) respectively.In general Cu-alloy corrosion rate is higher than AISI one; furthermore, AISI corrosion and release rates show relatively higher fluctuations in correspondence of the plasma pulse operations, due to the temperature changes.

ACPs inventory in the IBED PHTS
The comprehensive inventory of CPs can be categorized into two distinct classes: the immobile media comprising deposit    contamination also referred to as surface activity.In the context of ORE, these layers predominantly contribute to the radiation dose originating from ACPs and to the inventory when managing radiological waste; their removal might require challenging engineering measures, such as the installation of a bespoken chemical cleaning loop.Conversely, the mobile media contribute to coolant contamination, also referred to as volume activity.Moreover, they contribute to the accumulation of radioactivity within the resins and filters present in the CVCS [36].however, its concentration is about a factor 10 3 lower than the total mass of cobalt deposit; for Co-57 and Co-58 the ratio is even higher, 6 and 7 orders of magnitude lower respectively to cobalt element.Similar ratios apply also for other types of elements and radioisotope (e.g.Fe, Cr).Thus, we can state that the overall mass of the CPs is primarily driven by the non-activated inventory; hence defining the total mass of CPs as 'ACPs mass' it is misleading from a scientific point of view and simply wrong from an administrative one (i.e.safety objectives or limits on ACPs shall refer to activity rather than mass).For this reasons, ACPs limit shall be generically defined based on overall and/or isotope specific activity.Nevertheless, the overall CPs mass is a safety relevant information, for: -some maintenance operations (e.g.frequency filter replacement and resins regeneration) as they are affected by the overall mass of CPs; -monitoring the mass transport within the circuit before nuclear operation will give relevant information on the expected level of contamination during FPO; -Enabling test and optimization of established, bespoken good practices to reduce the contamination during the D-T operation.
Figure 16 below shows the concentration of the elements as ions and particles into the coolant (the CVCS region upstream the filters and resins is chosen as representative one): Cu concentration is the highest during the PFPO and FPO operations, while Fe element dominates during the baking.The ion concentration increases during the baking due to general higher release rates-section 4.3 shows this phenomenon for Co-60.

Activity.
Figure 17 shows the total activity (inner oxide + deposit) for the in-flux regions and out-of-flux regions: the observable fluctuations occurring in plasma operation are due to the short living radioisotopes natural decay during dwell time (i.e.mainly Cu-64 and, secondarily Mn-56); the rather stable behavior of the overall activity in the out-offlux regions is due to the contributions of medium (Co-58 and Fe-55) and long living (Co-60) isotopes.
Among all the isotopes considered in this study, the dominant contributor to the ORE is Co-60 (see also section 4.2).Therefore, it is important to understand the contribution of Co-60 loop contamination (i.e.inner oxide and deposit) to the overall activity in the out-of-flux regions.Figure 18 compares the Co-60 contribution to the overall activity (all the radioisotope considered in this study) in the out-of-flux regions for both deposit and inner oxide: in general, the deposit contribution to the activity is 2 orders of magnitude higher than the inner oxide; furthermore the Co-60 deposit to the overall activity in the out-of-flu region at EoL is significant (∼50%).
As figure 19 shows, the surface activity in the in-flux regions is obviously significantly higher than in the out-offlux regions.For the latter, a significant higher surface activity value is observed for the regions exposed to the baking operation (light green) in comparison to the bypassed ones (in blue   Table 7 shows the comparison for different out-of-flux regions of the IBED loop: even though the surface activity for the HXs regions is more than a factor of 2 lower than the corresponding Jungle Gym and in-bioshield piping for both hot and cold legs, the overall activity in the HXs regions is significantly higher (one order of magnitude) due to the huge wet surface.However, the contribution to the ORE of the HXs shall consider also the self-shielding effect due to the tubes material (i.e.stainless steel) which reduces significantly the gamma irradiation [37].
Furthermore, the difference in surface activity between HXs and Jungle Gym regions is due to the exposure of the latter to the baking temperature and consequent higher contamination because of the precipitation of ions onto the deposit.Such a phenomenon contributes significantly to the overall contamination of the piping and equipment; since the cooling trains,  and hence also the HXs, are by-passed during the baking, the absence of the precipitation of ions results in a lower surface contamination.

Coolant Activity.
The activity into the coolant, or volumetric activity, is due to the concentration of ions and particles transported by the coolant.As figure 20 shows, such a concentration is homogenous within the loop in which the coolant flows during a specific operational mode.As example, figure 16 shows the coolant activity in the CVCS piping (i.e.upstream the CVCS filters and resins); during burn mode overall activity is significantly higher (two orders of magnitude) than during baking and shutdown operations; also, the contributors to the overall activity differ for this two modes.This is mainly due to the erosion of the copper deposit during the plasma burn phases (see section 5.4) and the subsequent release of particles to the coolant.Such a phenomenon is dependent on the relatively high velocity of the coolant flowing on copper surfaces during plasma burn modes.During burn (see figure 21), Cu-64 is the dominant contributor to the volumetric activity followed by Cu-66 and Mn-56, whereas in baking and shutdown modes, Fe-55 and Mn-54 are the two most important contributors, followed by Co-60.The predominance of Fe-55 and Mn-54 during the off-plasma mode is the consequence of a large wet surface of stainless steel simulated in the model: in fact, both Fe-55 and Mn-54 are generated by the activation of stable Fe atoms.Figures 22 and 23 show the dominant contribution of Fe-55 and Co-60 to the total activity in the ion exchange resins and particle filter respectively.

Impact on the ORE
Figure 24 shows the calculated impact of the main contributors to the gamma dose rates for a study case: the Jungle Gym return region, assumed as a DN250 Schedule 80 pipe.The total dose rate increases with the performance of the FPO thus to reach a maximum at EoL of 1 mSv h −1 .Although the ORE calculation requires more detailed studies [6,38,39] encompassing the overall maintenance scenario, the layout of the surrounding system, the duration and type of maintenance operation and the position of the operators, it is possible to state that such dose rate levels represent a safety concerns.In fact, Co-60 represents 99% of the total contribution to the dose rate after the second FPO since: -short living isotopes, as Cu-64 and Fe-59, rapidly decay after the end of the plasma operation; -medium living isotopes, such as Co-58, Mn-54 and Fe-59, contribute to the total at EoL for less than factor 1/100.
Hence, after 12 d from the plasma shutdown, the dose rate for the cooling equipment outside the bio-shield due to the ACPs is such to represent a challenge in terms of ORE.As shown in figure 25, the dose rate coming from ACPs at the EoL, in the out-of-flux regions can be estimated by considering only the surface activity due to Co-60.The maximum Co-60 concentration is reached in the divertor and equatorial return manifolds, about 530 MBq m −2 , whereas the minimum is reached in the CVCS piping, 24 MBq m −2 .The surface activity of the in-flux regions is shown in black since it is generally higher than the maximum out-of-flux regions.

Contamination transfer
Departure of ions and particles from the in-flux wet surfaces leads to the spreading of contamination within the loop; three mechanisms contribute to the spreading of contamination in the circuit: particles erosion, ions dissolution (inner oxide and deposit) and metal release of ions directly into the coolant.In addition, the abrasion mechanism for components with moving parts (like valves) can also contribute to the contamination transfer; however, since the valves are not supposed to contain stellites, abrasion is not considered in the present work.Contaminating ions and particles are transported in the coolant and transferred to the out-of-flux regions through different mechanisms, such as deposition and precipitation.In the following section, we describe the contamination transfer from reference in-flux regions to out-of-flux regions focusing on Co-60 since its predominance on the ORE.

Origin of Co-60 and impact of cobalt concentration in contaminating the out-of-flux regions.
To investigate the impact of cobalt content in both in-flux and out-of-flux regions, a parametric study was carried out: the reference case has a Co concentration of 500 ppm for the in-flux regions and 2000 ppm for the out-of-flux.
This was compared to five additional cases: -  For in-flux regions, the increase in Co to 2000 ppm for both AISI and Cu-alloys (SFCo20-20) results in an increase of about 90% in the Co-60 surface activity, whereas it is only 15% for the increase in Co only for AISI (SFCo20).The decrease in Co to 200 ppm for both AISI and Cu-alloys (SFCo2-2) results in a decrease of 40%, compared to 25% for a decrease in Co only for AISI (SFCo2).
For out-of-flux regions, the change in the Co content of AISI has an unexpected impact.An increase in Co to 2000 ppm (HFCo20) results in a slight decrease of 5% in the Co-60 surface contamination, and a decrease to 200 ppm (HFCo2) results in an increase of 40%.This unexpected effect occurs during the baking phases.The Co-60 released during baking precipitates much more when the Co out-of-flux content is lower (lower Co equilibrium concentration).On the other hand, as expected, during plasma operation, the Co-60 increase slope is lower when the Co content of AISI is lower.
Thus, we can conclude that most of the Co-60 is generated by the neutron activation of the Cu-alloy material, according to two activation reactions Cu-63(n, α)Co-60 and Co-59(n, γ)Co-60 in the in-flux regions.

Co-60 departure mechanisms from in-flux regions.
Figure 27 shows the dominant phenomena for the HV region and figure 28 those for the HV-Cu alloy region.
For the stainless-steel regions under neutron flux, the metal release of ions and the deposit dissolution drive the contamination.For the copper alloys regions under flux, the erosion dominates the particle transfer to the coolant during plasma pulses, whereas during baking the metal release of ions drives the contamination, similarly to the stainless steel regions under flux.

Co-60 transfer to the out-of-flux regions.
Figure 29 shows the dominant Co-60 transfer phenomena for the Jungle-Gym return region.Co-60 contamination in the Jungle-Gym return region is mainly due to ionic precipitation onto deposit during baking; in pulsed operation, the driving mechanisms are particles deposition and ions precipitation on the inner oxide.
It is worth focusing on the ionic precipitation onto deposits, since it is the driving mechanism of the dose estimated for the         ORE.The current work is showing that precipitation occurs at the isotopic level, i.e. it is limited to Co-60, whereas the elemental Co ions concentration in the coolant never reaches the Co equilibrium concentration with respect to the deposit.Figure 30 highlights this phenomenon of isotopic precipitation in unsaturated conditions of the chemical element (isotopic dissolution-precipitation model described in [12]).
By comparing the Co-60 ion concentration and the Co-60 equilibrium concentration in the deposit of the Jungle Gym return, it is possible to observe in figure 30 that the former overcome the latter only during the baking phases: this means that most of the Co-60 contamination occurs during baking.

Conclusions and recommendations
This report provides the updated assessment of the mass and the activity of the IBED loop CPs and ACPs.These new data with regard to CPs masses and ACPs activities show the importance of specifying which portion of the CPs are activated and the necessity to review the strategy for administrative limits of the ACPs inventory.
The present work confirms the predominant contribution of Co-60 to the ORE and highlights the significant impact of the baking operation in spreading Co-60 in the circuit because of precipitation.
The focus is also on the origin of Co-60, primarily due to the presence of copper as base material for the in-vessel components, i.e. in regions under neutron flux.Therefore, for future activities, measures aimed at reducing the dose to the workers should include efforts to limit or prevent such contamination mechanisms.This entails reducing the spreading of Co-60 outside the bioshield, which can be achieved by optimizing the water baking operation in terms of frequency and duration.
The corrosion law used in this study for copper is supported by experimental data in the temperature range 110 • C-250 • C for limited duration (i.e.<200 h).We strongly recommend establishing bespoken experiments to validate the corrosion and erosion rates for Cu and Cu alloy at ITER relevant conditions, i.e., baking temperature and duration, fluid velocity, reducing, even slightly oxidizing, environment and pH T equal or higher than 7.Such experiments will yield reliable data on the loop contamination transfer mechanism due to corrosion of the Cu-alloy during baking and will enable a refinement of the OSCAR-Fusion input parameters and validation of the results.The authors believe that the success of these endeavours will depend, in part, on the ability to distribute the experimental workload through collaborative international efforts among various organizations, such as EUROfusion.
Finally, the simulated injection of Li for pH management purposes instead of ammonia represents another recommendation to operate the IBED PHTS based on the relevant experience from PWR operation.

Figure 1 .
Figure 1.IBED loop simplified flow diagram; the dotted lines represent the branches of the circuit in which 240 • C water circulates during baking.

Figure 3 .
Figure 3.Comparison of the corrosion rates for AISI between the power law used for ITER and the MOOREA law used for PWRs at T = 250 • C.

Figure 4 .
Figure 4. Comparison of the release rates for Cu-alloy between OSCAR and an experiment in the CORELE loop [29].

Figure 5 .
Figure 5.Comparison of the corrosion rates for Cu-alloy between OSCAR and experiments [25, 30].

Figure 6 .
Figure 6.Neutron spectra in TRIPOLI 315 energy group structure in the cooling channels of the blanket module BLK15_S02 at the flattop of the nominal plasma shot of the fusion power of 500 MW; the spectrum in the beryllium layer is added as a reference.

Figure 7 .
Figure 7. IBED PHTS OSCAR-Fusion input model: the orange rectangle frames the baking loop regions, the green rectangle frames the CVCS regions and the red rectangle frames the in-flux regions.

4. 2
.1.Regions.The IBED loop can be broken down in 4 areas: -In-flux regions, i.e. the regions of the loop within the bioshield activated by the neutrons; -Out-of-flux regions, i.e. portions of the IBED loop belonging to the PHTS outside the bioshield; -The baking circuit connected in parallel to the PHTS; -The CVCS circuit connected in parallel to the PHTS.Each area encompasses several regions, each region being defined in function of geometric (e.g.wet surface), thermalhydraulic (e.g.temperature) and material (e.g.AISI316 or Copper, roughness) parameters.The current version of the input has 60 regions: 28 for the out-of-flux, 26 for the in-flux, 3 for CVCS and 3 for the baking circuit.Figure 7 below gives an overall view of the IBED loop model elaborated through the OSCAR GUI.

Figure 10 .
Figure 10.Out-of-flux activity (TBq)-comparison of scenarios B, C and D.

Figure 11 .
Figure 11.Corrosion products mass (kg)-comparison of scenarios B, C and D.

Figure 14
shows the results in terms of total mass of inner oxide and deposit of CPs (metallic elements) for the in-flux and out-of-flux regions, as well as the total mass of CPs trapped by the CVCS resins and filters.At the very beginning of the operation, there is a sharp increase in the mass of both in-flux and out-of-flux regions: this is due to the 'passivation' of the circuit, meaning the formation of the inner oxide and deposit.During baking operation, the thickness of the deposits increases significantly in the in-flux

Figure 14 .
Figure 14.Corrosion products mass (kg) in different regions of the IBED loop.

Figure 15 .
Figure 15.Mass comparison (g) between co element (total) and its stable and radioactive isotopes in the out-of-flux regions.

Figure 17 .
Figure 17.Total activity (TBq) for the in-flux and out-of-flux regions.

Figure 18 .
Figure 18.Activity comparison in the out of flux regions-All isotopes vs Co-60.
and light blue) (NB: the regions in light grey represent the CVCS filter and resins).

Figure 19 .
Figure 19.Surface activity in the IBED loop at EoL.

Figure 20 .
Figure 20.Coolant volumetric activity 12 d after the EoL plus 30 d of baking.

Figure 21 .
Figure 21.Coolant activity-detail of main contributors to the total activity (MBq/t)-Detail of the last FPO campaign.

Figure 26
Figure26shows the results of such comparison in terms of Co-60 contamination of the Jungle-Gym return region.For in-flux regions, the increase in Co to 2000 ppm for both AISI and Cu-alloys (SFCo20-20) results in an increase of about 90% in the Co-60 surface activity, whereas it is only 15% for the increase in Co only for AISI (SFCo20).The decrease in Co to 200 ppm for both AISI and Cu-alloys (SFCo2-2) results in a decrease of 40%, compared to 25% for a decrease in Co only for AISI (SFCo2).For out-of-flux regions, the change in the Co content of AISI has an unexpected impact.An increase in Co to 2000 ppm (HFCo20) results in a slight decrease of 5% in the Co-60 surface contamination, and a decrease to 200 ppm (HFCo2) results in an increase of 40%.This unexpected effect occurs during the baking phases.The Co-60 released during baking precipitates much more when the Co out-of-flux content is lower (lower Co equilibrium concentration).On the other hand, as expected, during plasma operation, the Co-60 increase slope is lower when the Co content of AISI is lower.Thus, we can conclude that most of the Co-60 is generated by the neutron activation of the Cu-alloy material, according to two activation reactions Cu-63(n, α)Co-60 and Co-59(n, γ)Co-60 in the in-flux regions.

Figure 23 .
Figure 23.Activity (TBq) in the particles trapped in the CVCS filters.

Figure 25 .
Figure 25.Co-60 surface activity at EoL + Baking-the regions in black and grey represent loop regions with values above and below the selected color scale, respectively.

Figure 26 .
Figure 26.Impact of cobalt concentration in the material for both in-flux and out-of-flux regions on the Co-60 contamination in the Jungle Gym return.

Figure 30 .
Figure 30.Comparison of the Co and Co-60 ion concentrations with Co and Co-60 deposit equilibrium concentrations.

Table 2 .
Material composition and properties used in IBED model.
[32]2.IBED materials.The coolant circulating in the IBED loop interacts with different materials[32]: AISI316, AISI304, oxygen free copper and copper alloy.To better simulate the variety of materials used in different regions of the loop, it was chosen to simulate 4 types of AISI316, one type of copper alloy and one type of pure copper.AISI304 (typically used for the out-of-flux piping) has not been simulated since its very similar material composition to AISI316; therefore, all the regions with stainless steel are simulated as AISI316.Materials composition and properties are summarized in table 2.

Table 5 .
Thermal-hydraulic parameters during the different operational phases.
Figure 8. ACPs normalized contribution to the contact dose rate of a jungle primary pipe.

Table 6 .
Detail of the operational scenarios to study the impact of the period homogenization.

Table 7 .
Surface and overall activity at EoL + Baking comparison for hot and cold leg regions.