Conference Agenda

Kinetic sputtering of materials by irradiation with energetic particles is a process that is well understood for perfectly flat surfaces. In the last decades, many theoretical concepts and numerical simulation codes have been developed that can reproduce measured quantities such as the sputter yield from experiments. However, for real materials used in actual applications, such as plasma-facing components in a fusion device, the primary assumption of a perfectly flat surface often fails to account for the significant role of surface roughness.

The geometry of a rough surface can lead to various effects like e.g., redeposition of sputtered particles. In combination with the influence of locally inclined surface features on the incidence angle of incoming particles, the assessment of sputter yields for rough surfaces remains a current topic, especially in the field of plasma-wall interaction.

At TU Wien, the effect of roughness on sputtering has been successfully investigated by performing special laboratory measurements with a quasi-non-invasive low-flux ion source and a highly sensitive quartz crystal microbalance. In addition, the BCA-based ray tracing code SPRAY was developed, which allows for relatively quick calculation of sputter yields by using large-scale AFM images of rough surfaces as inputs. With information from both experiments and simulations, a better understanding of roughness effects was possible, and now the sputter yield for conventionally (Gaussian) rough surfaces can be predicted with a statistical roughness parameter [1,2]. Based on these results, more advanced types of surface roughness e.g., pyramidal or nano-columnar structures have also been investigated. In particular the latter show a significant reduction in the sputter yield, which would be especially favourable for nuclear fusion applications [3].

In this invited talk, a comprehensive overview and explanation of roughness effects on sputtering will be given, covering the latest results for conventional roughness and also for nanostructured surfaces with periodic patterns.

[1] C. Cupak, et al., Appl. Surf. Sci. 570 (2021) 151204

[2] P.S. Szabo et al., Surf. Interfaces. 30 (2022) 101924

[3] A. Lopez-Cazalilla, C. Cupak et al., Phys. Rev. Mat. 6 (2022) 075402

Erosion of the ITER main chamber first wall (FW) beryllium (Be) armour is expected to affect the lifetime of FW panel, dust formation, and release of Be impurities potentially leading to enhanced sputtering of the W divertor and tritium retention due to co-deposition. The ITER FW panels are shaped to protect leading edges and misalignments, which leads to magnetically shadowed regions, where impurity re-deposition and fuel co-deposition can occur. In addition to plasma ions, charge exchange (CX) neutrals in ITER and future reactors will play an important role on the first wall erosion and overall fuel retention, but the extent to which they contribute is still unknown.

A low-energy neutral particle analyzer (LENPA) based on the time of- flight method has been developed on EAST to measure the flux and energy of neutral particles to the first wall [1]. In the LENPA detection range, more than 85 % of neutral particles are in the energy range of 20–1000 eV. The integrated neutral flux in the energy range of 20–1000 eV increases with line-averaged density in ohmic discharges due to the increased CX reaction rates in a higher density plasma. Compare to the discharges fuelled with SMBI, the discharges without SMBI have more lower energy neutrals at the similar line-averaged density due to higher edge neutral density and proximity of fuelling. Due to the deeper penetration depth by SMBI fuelling, the neutral particles are generated closer to the core plasma and have higher energy. It was found that the neutral flux increases with heating power in all energy range. The mean energy of neutral particles increases significantly compared to the ohmic discharges. The higher ion temperature and edge density in auxiliary heated discharges result in higher neutral flux in all the energy range.

A quartz crystal microbalance (QMB) was installed together with the LENPA system on the equatorial port of EAST [2]. The neutral-induced material erosion rates and the neutral energy spectrum were measured simultaneously by the two real-time and in-situ diagnostics. The neutral-induced Al erosion rates for the 11 long-pulse full discharges are given by experimental measurement from the QMB and the theoretical calculations according to the neutral energy spectrum from the LENPA, which are consistent with each other. It is proved that higher density and heating power can increase the flux and energy of neutral particles, which results in stronger neutral-induced material erosion. The real-time Li powder injection during discharges can reduce the erosion rate due to the lower recycling and the resulting lower neutral flux.

[1] N.X. Liu, et al., Rev. Sci. Instrum. 92, 063507 (2021)

[2] Y. Zhang, et al., Nucl. Mater. Energy 26, 100877 (2021)

Material deposition, erosion and migration are important topics related to plasma-wall interaction (PWI) in tokamaks, which will directly influence the critical issues of plasma performance, material lifetime and fuel retention [1]. In EAST, the material erosion and deposition rates at the first wall has been real-time and in situ measured by the newly developed quartz crystal microbalances (QMB) diagnostic systems. And the poloidal profile of deposition on the surface of plasma facing components has been carefully characterized by different post-mortem analysis methods, such as handheld X-ray fluorescence (XRF) and inductively coupled plasma emission spectrometer (ICP), etc.. Besides, the migration of the eroded material has been simulated using DIVIMP code to better understand the PWI process in EAST.

Five QMB systems with each consisting of cooling water component and dual sensors with one closed sensor for reference have been installed at different locations in the mid-plane of EAST. A ∼80 nm aluminum (Al) film has been coated on the QMB crystal surface to measure the erosion and deposition rate. The erosion was measured during the glow discharge (GDC) with enough pumping and baking, while deposition was shown during ion cyclotron range of frequencies (ICRF) induced cleaning. It was found that the higher ICRF power was, the higher deposition rate. Besides, normal discharges mainly contribute to the total net erosion and the maximum erosion amount reaches up to 110 ng/cm². Disruptive discharges lead to a total net deposition rate of 0.48 ng/(cm²s).

After a whole experimental campaign, the thickness of deposits from the re-deposition of sputtered first wall material and the deposition of wall conditioning material can be varied from several micrometers to about 120 μm. The deposits consisted of Li, C, O, W, Mo, Fe, Cu, Ni, et., which was mainly in the form of Li₂CO₃, MoO₃ and WO₃. The total content of Li₂CO₃ was found to be higher than 90 wt.%. Poloidal distribution of redeposited Mo content indicated that in the upper inner and outer divertor region, more Mo was redeposited compared to the dome region. Similar behaviour was found with concentration of W in deposits at the lower divertor region. Near the striking points of inner lower divertor, the SiC coating on graphite surface can be eroded to about 90 μm. With increase of distance to the striking points, the thickness of eroded material decreased and then the thickness of depisits increased to about 74 μm. Deposits from the TZM first wall at high-field side exhibited double peaks of W concentration located near the upper W divertor and at the midplane tile which may be influenced by cross-field diffusion of eroded W. And the eroded W is easier to be redeposited on location closer to the upper divertor or with a smaller normalized poloidal magnetic flux.

[1] R. Yan, J. Peng, C. J. Li, et al. Nucl. Materi. Energy 30, 101103 (2022)

*Corresponding author: tel.: +86 0551 6559 3291, e-mail: yanrong@ipp.ac.cn (R. Yan)

WEST experiments are characterized by radiative power losses resulting from
tungsten (W) contamination due to wall erosion [1]. In experiments visible
spectroscopy measures the intensity of the neutral W spectral lines at the various
Plasma Facing Components (PFCs) [2]. These measurements are related to the
gross erosion flux through atomic data, while the prompt and local re-deposition at the
various PFCs is not measured in WEST. Numerical modelling is therefore required to
gain insight on the mechanisms governing W contamination in the tokamak. The
global picture of WEST W migration is analysed here with the help of simulations in a
3D realistic wall geometry with a quarter of torus symmetry using the codes
SOLEDGE [3] and ERO2.0 [4]. The study focuses on the relevance of the PFCs that
can be radially retracted in between plasma discharges (as the radio frequency
antennas), analysing the influence of their position on plasma-wall interaction and W
contamination. Simulations are performed over a range of typical WEST plasma
conditions. The WEST database is used to constrain plasma parameters using data
from embedded Langmuir probes, reflectometry, and spectroscopy. Numerical results
suggest that when the distance between the antennas and the separatrix at the outer
midplane (Radial Outer Gap, or ROG) is at its lower limit (1.5 cm), these objects
contribute the most to W contamination of the confined plasma, with 90% of W ions
inside the separatrix coming from the antennas. Moreover, when ROG is around the
most typical value (3 cm), their contribution to W contamination remains around 60%,
and thus is still predominant over the other PFCs. When ROG is equal to 1.5 cm the
antennas are radially located between the separatrix and the secondary separatrix
(magnetically connected with the upper divertor), so that the antennas partially
shadow the baffle and the upper divertor. With a larger ROG these PFCs get wet
by the plasma, increasing the heat and the particle load directed to them by a factor
up to 4. In turn, the antennas get partially shadowed. Therefore, the migration of W
particles coming from the antennas is determined by the relative position of these
components with the secondary separatrix. The fraction of W contamination caused
by the upper divertor and the baffle increases by less than a factor 2 changing ROG,
as a result of the PFCs efficient screening. Analyses suggest an optimized
configuration to minimize W contamination inside WEST operational space

[1] J. Bucalossi et al 2022 Nucl. Fusion 62 042007
[2] C. C. Klepper et al 2022 Plasma Phys. Control. Fusion 64 104008
[3] H. Bufferand et al 2021 Nucl. Fusion 61 116052
[4] J. Romazanov et al 2019 Nucl. Mat. And Energy 18 101016

Tungsten is a promising candidate material for armour components in future fusion reactors. In service these components will be exposed to intense neutron irradiation, a flux of light ions, as well as high temperatures. Knowledge of the evolution of material structure and properties caused by these conditions is essential for predicting reactor performance, safety, and longevity.

Here we consider Tungsten implanted with self-ion (to mimic neutron damage) and light ions (to mimic gas accumulation), as well as exposed to light ion plasma. Using transmission electron microscopy (TEM), high resolution transmission Kikuchi diffraction (HR-TKD) and X-ray diffraction the evolution of defect structure as a function of dose is characterised [1,2]. Our results reveal a large population of point defects that too small to be visible directly by TEM. At larger doses we observe a self-organisation of defects on length-scales far exceeding those associated with individual collision cascades. This suggests that pre-existing residual stress fields play a key role in controlling the evolution of irradiation-induced defects, consistent with our observation of extended strain patterns [3].

Nano-indentation shows a marked increase in hardness of the ion-implanted materials, combined with a pronounced increase in pileup [4,5]. Using transient grating spectroscopy, we measure elastic properties and thermal diffusivity of the ion-implanted material. Thermal diffusivity decreases sharply with implantation dose, saturating to a 50% lower value at ~0.1 dpa [6,7]. Elastic properties also evolve, with a reduction in stiffness and a small increase in elastic anisotropy [1,8].

Using density functional theory, molecular dynamics and crystal plasticity simulations, we explore the underlying mechanisms responsible for the changes observed in experiments [4,9,10]. Combining these different bits of information, a coherent picture of irradiation damage accumulation and resulting property change in tungsten begins to emerge.

[1] F. Hofmann et al., Acta Mater. 89, (2015).

[2] N. W. Phillips et al., Acta Mater. 195, 219 (2020).

[3] G. He, H. Yu, P. Karamched, J. Liu, and F. Hofmann, (2022).

[4] S. Das et al., Int. J. Plast. 135, 102817 (2020).

[5] S. Das, H. Yu, E. Tarleton, and F. Hofmann, Sci. Rep. 9, 18354 (2019).

[6] A. Reza, H. Yu, K. Mizohata, and F. Hofmann, Acta Mater. 193, 270 (2020).

[7] A. Reza et al., Acta Mater. 232, 117926 (2022).

[8] R. A. Duncan et al., Appl. Phys. Lett. 109, 151906 (2016).

[9] D. R. Mason et al., Phys. Rev. Lett. 125, 225503 (2020).

[10] D. R. Mason et al., Phys. Rev. Mater. 5, 125407 (2021).

Accurate quantification and characterization of nanoscale defects generated in materials exposed to intense radiation environments is essential for the successful development of radiation resistant materials for DEMO and beyond.

In order to make definite progress in radiation damage studies, an assessment needs to be made of the different defects types, from isolated defects to larger agglomerates, which may be present after any type of irradiation or thermal history. For larger defects, direct imaging methods like transmission electron microscopy can provide valuable information. However, for the primary damage state which consists mainly of single point defects and small clusters, experimental observation relies mostly on indirect techniques. Positron annihilation spectroscopy can be used for open-volume defects, such as vacancies and small vacancy clusters, but this technique is insensitive to self-interstitial components. The presence of both types of defects and their clusters may be inferred by measurements of the electrical resistivity. Such measurements can provide valuable information on the migration, clustering and dissociation of radiation defects, which are essential for the behaviour of materials under irradiation.

In this contribution, an overview of the different experimental methodologies will be presented and future directions will be discussed. Particular attention will be given to the application of multi-scale modelling and simulation for the interpretation of experiments and how this can leverage the quality of the obtained information.

Tungsten (W) and W alloys are considered as the most promising candidates for plasma facing materials (PFMs) in future fusion reactors. However, as a PFM, W will be irradiated by high-energy fusion neutrons causing severe displacement damage and performance degradation, which should be originated from the evolution of irradiation defects. Helium (He) and hydrogen (H), as the typical impurity elements, play a crucial role on the microstructure and mechanical properties of W. More importantly, recent experiments found that the addition of He or H has significant effects on the evolution and recombination of irradiation defects, while its underlying mechanism remains to be elucidated. Here, using the atomic simulations and Object Kinetic Monte Carlo (OKMC) methods, we systematically investigate the interaction between He/H and irradiation defects in W as well as their influences on the vacancy-interstitial recombination. We found that the small He-SIA complexes adopt a three-dimensional (3D) migration pattern owing to the strong attractive interaction of He with SIA, which is completely different from the 1D migration of SIAs in W without He. This unexpected collaborative 3D motion of He-SIA complexes increases the probability of Frenkel pairs recombination. For the case of H, it is known to segregate into vacancies to relieve local stress, which affects the spontaneous recombination radii between vacancies and SIAs in W. Such recombination radii are shown to be monotonically decrease with the increasing of H numbers and inversely proportional to the inner surface H density in vacancies, and thus reducing the recombination efficiency of irradiation defects in W. The OKMC simulations, performed based on our model, show good agreement with recent simultaneous/sequential irradiation experiments. These results improve our fundamental understanding of the influence of He and H on the evolution of irradiation defects and have great implications to estimate the performance of W-PFMs.

In a future fusion reactor neutron irradiation will create displacement damage which influences material properties, such as material strain and strength. One of the options to improve the behaviour of the material is by changing the microstructure of the material. Here we have studied how bulk microstructure, i.e. grain size in the material, influences the accumulation of radiation damage in tungsten (W). One of the hypotheses is that possibly less damage could be created due to defect annihilation at grain boundaries. To create displacement damage, we have used high energy W ions which are a good proxy for neutron irradiation, excluding transmutation, helium production and most importantly activation of the material. The amount of created defects was accessed by performing hydrogen isotope (HI) retention measurements, since lattice defects act as trapping sites for HIs with high de-trapping energy as compared to the energy of HI diffusion between solute interstitial sites. Therefore, the hydrogen isotope concentration can be treated as a measure of defect content present in the material.

The effect of microstructure on the generation of radiation damage was studied previously with W single crystal and polycrystalline W with grain sizes between 1 and 50 µm [1]. There, no significant effect was observed, with all samples showing similar D concentration in the damage zone. By using a laser-deposited, nano-crystalline W layer on a W substrate we proceeded with the study and went down with the grain size to the nanometer scale. The as-deposited layer had a grain size of few nanometers. By tempering, grain sizes of a few hundred nm up to few µm were adjusted. Samples were irradiated at 290 K by 20 MeV W-ions to create displacement damage down to 2.3 µm and with a maximum damage dose of 0.23 dpa. The concentration of defects was assessed by exposing samples to deuterium (D) atoms with an energy of 0.3 eV at 600 K and to 300 eV D ions at 450 K. In both cases D populates the created and existing defects. D retention and D depth profiles were measured by nuclear reaction analysis utilizing D(³He,p)4He nuclear reaction. In the nanograined samples D populated the damaged region more than three times faster than in samples with grain size of hundred nm and few micrometer size grains. The concentration of defects was assessed by the final D concentration in the samples. Samples with smaller grain size showed larger D concentration in the irradiated area. However, large D concentration in the non-irradiated sample showed that defect density was already high in the initial material. Samples were also analysed by transmission electron microscopy to analyse the damage distribution in the material where nanometer-size voids were observed.

[1] Pečovnik et al. J. Nucl. Mater 513, 198 (2019)

Improved diagnosis of tungsten gross and net erosion in fusion relevant experiments can be achieved with high-resolution ultraviolet spectroscopy in combination with more accurate predictions of atomic rate coefficients and time-dependent collisional radiative (CR) modeling. A mirror linked high-resolution UV imaging spectrometer providing spatial resolution down to ~40 μm has been commissioned on the linear experiment Radio Frequency Plasma Interaction Experiment (RF PIE). The imaging spectrometer and relatively simple linear geometry allows the evolution of ~50 W I and W II emission lines (200-500 nm) to be observed and compared to CR modeling using new neutral and singly ionized tungsten R-matrix excitation calculations [1,2].

Time-dependent collisional radiative modeling of tungsten spectral lines utilizing the CR solver ColRadPy [3] allows the impact of metastable states on tungsten emission and ionization to be assessed. The impact of these effects is difficult to measure on current fusion relevant experiments, however, modeling suggests metastable states potentially become important at ITER relevant divertor densities. Where a relatively large magnetic presheath creates a low density environment far from bulk conditions [4]. The relatively low electron density environment of RF PIE ~1x10¹⁷ m^-3 slows down atomic processes allowing the evolution to be resolved with the imaging spectrometer for comparison with CR modeling; the temporal evolution is directly linked to the spatial evolution of tungsten emission. A technique for measuring non-steady state (time-dependent) metastable populations will be presented using both W I and W II spectral lines. The spatially resolved evolution of singly charged tungsten emission allows for the effective ionization rate (SCD) of neutral tungsten to be inferred rather than the more widely used ionizations per photon (S/XB) (ratio of the SCD and a photon emissivity coefficient (PEC)). Multiple W I lines with different behaviors are well described with the model while W II emission does not agree with measurements using currently available neutral tungsten ionization rates (population of W²⁺ is small for RF PIE plasma conditions). Potential reasons for the discrepancy will be discussed.

[1] R. T. Smyth et al 2018 Phys. Rev. A 97 052705

[2] N. L. Dunleavy et al 2022 J. Phys. B: At. Mol. Opt. Phys. 55 175002

[3] C. A. Johnson et al 2019 Nucl. Mater. Energy 20 100579

[4] C. A. Johnson et al (2020, November 9–13) Diagnosing metastable populations in fusion edge plasmas using collisional-radiative modeling constrained by experimental observations 62nd Annual Meeting of the APS Division of Plasma Physics

*Corresponding author: tel.: +1 (320) 248-9823, e-mail: johnsonca@ornl.gov (C.A. Johnson)

Three-dimensional modeling with the plasma-fluid and kinetic neutral transport Monte Carlo code EMC3-EIRENE and the dust injection simulator (DIS) shows that local injection of solid materials in powder form leads to asymmetric and sometimes very localized modifications of the plasma distribution and plasma-material interactions under low-recycling conditions. The modeling confirms that the strategic choice and number of specific injection locations are crucial for the global control of heat fluxes, detachment, and real-time wall conditioning in next-step devices.

Low-recycling impurity powders like boron (B), lithium (Li), or boron nitride (BN) injected for wall conditioning and edge cooling cause asymmetric mitigation of the plasma-surface interaction processes according to the 3D full-torus simulations based on recent DIII-D experiments employing an impurity powder dropper (IPD) [1]. These findings also imply potential over- or underestimation of local probe data based on the toroidal location of respective diagnostics. Prior studies have shown such effects for single impurity sources in the case of N2 and B injection in Alcator C-Mod [2], DIII-D [3], and EAST [4]. In the present work, impurity powder injection is being compared for open and closed divertor configurations and variations of the injection location. 3D analysis thus far shows that asymmetries in the cooling effects and deposition of low recycling impurities on main plasma-facing components are more pronounced in closed divertor configurations such as the DIII-D small angle slot (SAS) divertor due to its enhanced impurity screening capability. In the SAS, full-torus modeling of the injection of 35 mg/s of B powder shows a reduction in the average heat fluxes and electron temperatures by 30-45% over a toroidal segment of ~130º centered around the injection location. However, a scenario with injection into the plasma crown during lower single null showed that these effects reduce to a variation of only 12-14% over the full toroidal extent in the lower open divertor. Furthermore, it is shown that impurity powders of 45-150 µm size, compared to impurity gas, penetrate deeper into the plasma due to powder particle transport and ablation effects.

[1] A. Bortolon et al 2020 Nucl. Fusion 60 126010

[2] J.D. Lore et al 2015 Phys. Plasmas 22 (5) 056106

[3] F. Effenberg et al 2021 Nucl. Mater. and Energy 26 100900

[4] Y. Luo et al 2021 Plasma Phys. Control. Fusion 63 105007

Work supported by DOE Award No. DE-AC02-09CH11466, DE-FC02-04ER54698, DE-SC0020357, DE-AC05-00OR22725 and DE-AC52-07NA27344.

Disruption-induced surface melting is foreseen as a major damage source on some of the beryllium components of the ITER first wall [1], especially during early operation phases, when practical experience with the mitigation systems is first being built up. Hydrodynamic instabilities developing in the melt layers so created may produce liquid ejecta, thereby contributing to material losses, but also acting as a potential source of solid dust, should the ejected droplets cool down below the melting point [1,2]. Although valuable information can be extracted from experimental evidence of melt splashing in fusion environments [3,4], the dynamics of the liquid pools cannot be directly observed, which makes numerical modelling an essential tool to understand the physics at play and anticipate the consequences of melt events in ITER.

The destabilization of liquid beryllium layers is susceptible to occur during both the thermal quench (TQ) and current quench (CQ) stages of plasma disruptions, albeit through different mechanisms. On the one hand, melt pools created by TQ heat loads can be subjected to strong eddy currents resulting from variations in the poloidal magnetic field, which promote Rayleigh-Taylor instabilities as the associated Lorentz forces accelerate the liquid metal towards the plasma [5]. On the other hand, the halo currents flowing into the wall during CQs generally drive the melt along the underlying solid surface [6], so that material ejection can occur in the presence of a geometric obstacle, such as the edge of a tile [4,7].

This talk is devoted to the results of multiphase Navier-Stokes simulations of beryllium melt flows in ITER disruption conditions, focusing on quantitative predictions of the liquid mass loss, along with the characteristic sizes and speeds of the ejecta. First, a concrete worst-case CQ scenario is reported, mimicking an unmitigated vertical displacement event accompanied by melt splashing at the poloidal chamfer of the ITER first wall panels [8,9]. A range of TQ-relevant parameters is then explored to provide a first assessment of the growth of Rayleigh-Taylor instabilities, accounting for the continuous production of melt by plasma heat fluxes, as well as vaporization losses at the liquid’s free surface.

[1] M. Shimada et al, J. Nucl. Mater. 438, S-996 (2013)
[2] L. Vignitchouk et al, Nucl. Fusion 58, 076008 (2018)
[3] K. Krieger et al, Nucl. Fusion 58, 026024 (2018)
[4] I. Jepu et al, Nucl. Fusion 59, 086009 (2019)
[5] M. Lehnen et al, J. Nucl. Mater. 463, 39 (2015)
[6] S. Ratynskaia et al, Nucl. Fusion 60, 104001 (2020)
[7] L. Vignitchouk et al, Nucl. Fusion 62, 036016 (2022)
[8] J. Coburn et al, Nucl. Fusion 62, 016001 (2022)
[9] L. Vignitchouk et al, Nucl. Fusion 63, 016004 (2022)

The MEMENTO (MEtalic Melt Evolution in Next-step TOkamaks) code is a new numerical implementation of the physics model originally implemented in the MEMOS-U code [1,2]. The model self-consistently describes melt formation and macroscopic motion in fusion environments and allows the assessment of metallic reactor armour damage under powerful transient plasma events [1,2]. The model has been validated against multiple dedicated EUROfusion experiments [1,2,3]. MEMENTO solves the heat and phase transform problem coupled with the incompressible Navier-Stokes equations in the shallow water approximation and with the current continuity equation on a domain with rapidly time-evolving and deforming metal-plasma interface (the free-surface).

MEMENTO utilizes non-uniform and adaptive meshing along with sub-cycling in time facilitated by the AMReX open-source framework [4,5] as well as AMReX's built-in parallelization capabilities. The code development has been guided by applications to reactor relevant melting scenarios, characterized by large wetting areas, highly deformed free surfaces and complex PFC geometries, such as leading edges, sloped surfaces or gaps. The code has been successfully employed for the modelling of two new experiments, in AUG and WEST, which realized previously unexplored regimes where the conventionally secondary thermo-capillary effects were unmasked, owing to the uncharacteristically low contribution of the typically dominant JxB force [6]. The agreement with these specially designed exposures further established the predictive power of the underlying physics model and the MEMENTO code for the assessment of gross topological erosion due to armour melting on ITER and DEMO [6].

In this presentation, we shall highlight the new features and numerical capabilities of the MEMENTO code as well as report updates of the underlying physics model that enhance its suitability in reactor relevant regimes. In particular, the latter concerns the treatment of the current that escapes from multi-emissive magnetized sheaths [7].

[1] S. Ratynskaia, E. Thoren, P. Tolias, et al., Nucl. Fusion 60, 104001 (2020).

[2] E. Thoren, S. Ratynskaia, et al., Plasma Phys. Control. Fusion 63, 035021 (2021).

[3] S. Ratynskaia, E. Thoren, P. Tolias, et al., Phys. Scr. 96, 124009 (2021).

[4] https://amrex-codes.github.io/amrex/.

[5] W. Zhang, A. Almgren, V. Beckner, et al., J. Open Source Softw. 4, 1370 (2019).

[6] S. Ratynskaia, K. Paschalidis, et al., Nucl. Mater. Energy 33, 101303 (2022).

[7] P. Tolias, M. Komm, S. Ratynskaia, et al., Nucl. Fusion 63, 026007 (2023).

*Corresponding author: e-mail: konpas@kth.se (K. Paschalidis)

A thorough analysis of transient grating spectroscopy (TGS) for penetration depths relevant to heavy-ion damage in plasma-facing materials (PFM) is presented. Recently, TGS has been used to non-destructively measure properties such as thermal diffusivity (α) and Young’s modulus (Y) within the heavy-ion damaged zone (typically ≲ 10 μm) of PFM. This work highlights several experimental considerations and provides several remedies to improve the analysis in order to accurately quantify these properties.

TGS induces a thermal grating on a sample surface with a pump laser that produces a short pulse from two crossed excitation beams. The thermal relaxation of the grating is measured with a probe laser beam that diffracts off the evolving sample surface. An analytic solution [1] to the thermoelastic equations is typically used to model TGS data. For previous experiments on heavy-ion damaged W [2-4], the conditions may not well satisfy the model’s assumptions. The analysis typically requires a very short (i.e. a δ-function) pump pulse relative to the relaxation of the grating, an assumption possibly broken when probing the length scale of the heavy-ion damage zone. The relative size of the pump to probe spots may result in sampling a significant thermal gradient along the sample surface not accounted for by the simple analytical model typically used. Lastly, the energy deposited by the crossed pump laser spots can result in a significant thermal gradient that would effectively result in a higher temperature measurement than erroneously reported.

To account for the finite pump pulse width, we use a convolution of the pump pulse with the analytic solution. The convolved solution does not have the drawback of previous work in choosing an arbitrary starting point to fit the TGS data [4]. Furthermore, we show that the TGS data is significantly altered due to the pulse width and results in a difference of at least 10-15% in the best case fitted α value. This work also highlights the importance of using a relatively large pump spot compared to the probe spot to minimize the local change in temperature. That is, preventing the measurement itself from significantly changing the property measured since α itself is temperature dependent.

[1] Käding et al., Appl. Phys. A 61 (1995) 253-261

[2] F. Hoffmann et al., Sci. Rep. 5 (2015) 1-7

[3] Dennett et al., Appl. Phys. Lett. 110 (2017) 211106

[4] Reza et al., Acta Mater. 232 (2022) 117926

[5] Dennett et al., J. Appl. Phys.123 (2018) 215109

Work supported by the USDOE award, DE-SC0021656, and cooperative agreement, DE-SC0022528.

*Corresponding author: e-mail: msimmonds@eng.ucsd.edu (M. J. Simmonds)

Efforts to improve the thermomechanical properties of plasma-facing materials when exposed to high-energy particle irradiation led to a new class of materials; namely, dispersoid-strengthened W. For example, carbide-dispersion strengthened W has been shown to improve ductility, crack resistance, and radiation tolerance. Here we discuss various aspects of the W–ZrC interface structure and stability and focus on how these features impact the H/He behavior at and near the interface. Our density functional theory modeling results indicate that ZrC (111) –W (110) exhibits the highest adhesive energy, and hence forms the most stable interface. As part of our investigation of the diffusion and trapping behavior of H and He near the ZrC (111) ­­–W (110) interface, we calculate the minimum energy paths along and across the interface using the climbing image nudged elastic band (CI-NEB) method. The CI-NEB analysis demonstrates that both H and He are strongly trapped at the interface, with de-trapping energies for H on the order of 1.7 eV, irrespective of whether the diffusion is directed into the W or ZrC side of the interface. He is bound with a de-trapping energy of ~3 eV for diffusion into the ZrC and around 4 eV for diffusion into the W matrix. Both H and He can migrate along the W-ZrC interface with a preliminary calculated activation energy for the diffusion of ~1 eV. Strong trapping at the interface and the potential for diffusion along the interface suggest the possibility for the nucleation of He/H gas bubbles. This possibility will be further explored by examining how these gas atoms interact with interfacial defects and other gas atoms. The present study provides mechanistic insight towards interpreting recent experimental studies of the interface structure and the likelihood of hydrogen and helium trapping and retention in ZrC dispersed tungsten as a future plasma-facing component.

Permeation experiments can help to understand fundamental transport processes of hydrogen isotopes in plasma-facing materials of nuclear fusion devices.

In gas-driven experiments, the permeating flux depends on the concentration gradient between upstream and downstream surfaces, which are in chemical equilibrium with ambient gas and vacuum respectively. Since deuterium exhibits a high endothermic heat of solution in tungsten (1.14 eV [1]), the permeation flux (i.e. signal strength) becomes small for low temperatures due to a small upstream surface concentration.

In ion-driven experiments on the other hand, the source term of deuterium on the upstream side is temperature independent and proportional to the implanted beam flux. Thus, ion-driven permeation can be measured with reasonable signal strengths even at low temperatures. If the distribution of flux densities across the ion beam is known, ion-flux-depending processes like surface transport or the dynamics of trapping and de-trapping of hydrogen isotopes at material defects can be investigated.

In this contribution we present the well-characterized ion-driven permeation setup “TAPAS” and permeation data for deuterium in tungsten. The 100 µA beam consists of monoenergetic D₃⁺ ions and exhibits a footprint area of roughly 22 mm². We calibrated the ion current measurement by the mass loss of a sputter eroded Cu sample using sputter data from [2]. From spatially resolved height measurements across the resulting sputter crater the exact histogram of flux densities has been derived. The beam profile exhibits flux densities between 1x10¹⁹ and 3x10²⁰ D/m²s with an average of 8x10¹⁹ D/m²s. Using the histogram of flux densities, a detailed description of the transient behavior of the permeation signal is possible even in trapping dominated regimes (i.e. at low temperatures).

Deuterium permeation was measured between 600 and 1000 K in recrystallized tungsten (2000 K, 30 min) with low ion energies of 170 eV/D to suppress surface modification of the samples by sputtering. Diffusion-limited boundary conditions of the experimental conditions are shown by a variation of the sample thickness between 25 and 100 µm.

The experimental data was compared to calculations from the diffusion trapping code TESSIM-X. We considered trap density and trapping energy as fitting parameters, while deuterium diffusivity and solubility were taken from [1]. The steady state permeation signals agree well with modelling predictions for diffusion-limited boundary conditions. The best fit for trapping energies and trap densities is presented.

[1] G. Holzner et al., Phys. Scripta. 2020, 014034 (2020)

[2] W. Eckstein et al., Sputtering Data (IPP 9/82). Garching: Max-Planck-Institut für Plasmaphysik (1982)

*Corresponding author: tel.: +49 89 3299 1790, e-mail: philipp.sand@ipp.mpg.de

Experimental results from the 2022 tungsten (W)-coated Small Angle Slot (SAS-VW) divertor campaign at DIII-D coupled with interpretive 3DLIM modelling show opposing trends for core impurity content when compared to impurity deposition on far-scrape off layer (SOL) collector probes (CPs) with increasing main ion density. SAS-VW is a closed, W-coated divertor designed to more easily facilitate divertor detachment while reducing impurity leakage. An experiment was performed to run a series of upper-single-null L-mode discharges in each toroidal magnetic field (B_T) direction, with increasing main ion density (line-averaged density = 3.15-4.35e19 m^-3, a 38% increase) which approaches and slightly exceeds the divertor detachment threshold. Preliminary W deposition measurements using double-sided, graphite CPs inserted at the Midplane Materials Evaluation System (MiMES) revealed a 75% decrease over the density scan when operating in the unfavorable (ion out of the divertor) B_T direction. Notably, this trend breaks down as the density crosses the divertor detachment threshold. In constrast, soft X-ray (SXR) radiation from the same discharges is used as a proxy for W accumulation, showing core W content that increases by 77% with increasing main ion density. Similar L-mode discharges conducted in the favorable B_T direction result in significantly less deposition on CPs at both locations, MiMES and DiMES (Divertor Material Evaluation Station, located at the plasma crown). Lastly, the inner-target facing side of the DiMES probe in unfavorable B_T experiences far greater deposition than all other probe faces in either B_T direction for the detached case, indicating that W is being sourced from outside of SAS-VW or is subjected to SOL transport mechanisms yet to be understood.

An interpretive modeling workflow is presented for assessing the transport of W sputtered from the SAS-VW divertor, which is ostensibly the only source of W in the carbon-walled device. A SOL leakage coefficient and core leakage coefficient correspond to the far-SOL W deposition on CPs and core W content from SXR, respectively, each divided by the W sourcing rate obtained via emission spectroscopy to quantify impurity leakage from the divertor. Coupled DIVIMP-3DLIM modeling is presented to help elucidate the cause for opposing trends in the leakage coefficients with respect to main ion density.

Work supported by the U.S. Department of Energy, Office of Science, Office of Fusion Energy Sciences, using the DIII-D National Fusion Facility, a DOE Office of Science user facility, under Awards DE-SC0019256, DE-SC0023378, DE-FC02-04ER54698, DE-AC05-00OR22725, DE-FG02-07ER54917, and DE-NA0003525.

*Corresponding author: e-mail: smesser5@vols.utk.edu (S.H. Messer)

Due to its high melting temperature, high thermal conductivity, low sputtering yield and low hydrogen isotopes retention, tungsten has been chosen as the divertor material of International Thermonuclear Experimental Reactor (ITER) and a leading candidate of plasma-facing material (PFM) in the future fusion reactor [1]. The PFM exposed to radiation environment including high flux neutrons at high temperature always results in diverse defect microstructures, which could be observed in atomic scale by Transmission Electron Microscopy (TEM). Therefore, the defect microstructures during TEM sample preparation such as Argon ion milling or focused Ga ion beam sampling should be evaluated. In this work, Ga ion implantation induced artifact during the TEM sample preparation by focused ion beam (FIB) facility was quantified, and the HR-S/TEM examination confirmed the presence of high density of black spot defects (BSDs) after Ga ion irradiation to a fluence of 1.8×10¹⁴ Ga⁺/cm². In order to reduce or completely eliminate BSDs after Ga ion irradiation, post-processing of the TEM sample by Ar ion milling with beam energies as low as 1.0 keV were conducted. However, successive Ar ion milling was not sufficient to eliminate the FIB-induced BSDs, even though a total thickness of 20.0 nm was removed on both sides. The sampling/milling processes are also studied by molecular dynamics simulation with LAMMPS [2] and Monte-Carlo simulation with SDTrimSP Version 6.06 [3] program to understand the experimental observations.

[1] J. Knaster, A. Moeslang, T. Muroga, Nature Physics 12(5), 424-434 (2016)

[2] A. P. Thompson, H. M. Aktulga, R. Berger, et al., Comp Phys Comm, 271,10817 (2022)

[3] A. Mutzke, R. Schneider, W. Eckstein, et al., SDTrimSP Version 6.00, IPP Report 2019-02

To improve plasma confinement, interest exists in utilizing boron (B) wall conditioning of fusion tokamaks containing tungsten (W) plasma facing components [1,2]. First-principles density functional (DFT) theory calculations have been performed to model the energetics of B near W surfaces and the effect of B adsorption or segregation on hydrogen (H) behavior. The results show that B atoms are strongly adsorbed on the W surfaces and will penetrate below the W surface at temperatures higher than ~ 850 K. The B adsorption or segregation near the W surfaces influences the surface behavior of atomic H and molecular H₂. In addition, due to the possible formation of W borides in the W surface region during boronization coating [3], the energetics of intrinsic defects, H and helium (He) in bulk and/or near surfaces of W borides have been investigated using DFT caluculations. We found that the formation energy of B Frenkel pairs is lower than W Frenkel pairs, and boron point defects are more mobile than W point defects in the evaluated W borides. Notably, the diffusion activation energy of W point defects and gas atoms generally increases with B content in bulk W borides. Furthermore, the B terminated surfaces are more stable than W terminated because of the significant reconstruction of B on the surfaces of W borides. The H adsorption and diffusion near the surfaces of W borides is sensitive to both the surface termination and the surface orientation. These studies systematically investigate the influence of boronization coating on the W surfaces and the resulting effects on gas behavior at an atomistic level.

[1] A. Bortolon, R. Maingi, A. Nagy, et al., Nucl. Fusion 60,126010 (2020)

[2] F. Effenberg, A. Bortolon, H. Frerichs, et al., Nucl. Mater. Energy 26, 100900 (2021)

[3] L. Yang, K. Zhang, M. Wen, et al., Sci. Reports 7, 9353 (2017)

*Corresponding author: tel.: +1 8652968781, e-mail: liyang@utk.edu (L. Yang)

When tungsten is exposed to high temperatures (> 1200°C), it undergoes recrystallization and grain growth [1], which can degrade tungsten’s mechanical properties. As a result, there has been considerable interest in modifying tungsten’s microstructure to improve its performance in a fusion environment. Carbide Dispersion-Strengthened Tungsten (DS-W) has been considered a plasma-facing component (PFC) alternative [1]. The dispersoids can suppress recrystallization embrittlement, and the high density of grain boundaries may serve as sinks for displacement damage arising from neutron and high energy ion irradiation [2,3]. In this work, we will use W-5%TiC to explore the possible changes in surface morphology and structure of PFCs following high-temperature annealing.

Experiments were done using a polished W-5.0(a.t.%)TiC sample. Auger electron spectroscopy (AES) and scanning electron microscopy (SEM) were performed before and after 2KeV Ar+ sputter cleaning on both dispersoid and W regions for initial characterization. An effusion cell was used to anneal the sample to 1600°C over a 90 min. interval. AES was then performed in conjunction with sputtering to attain a depth profile of the sample. Ion scattering was done with a 1KeV Ne+ beam as an efficient way to recoil H isotopes and remove substrate atoms. Using scattering angles between 20-90°, the sample was then dosed with deuterium gas to observe chemisorbed H on the surface using direct recoil spectroscopy (DRS), monitoring the structural changes induced by Ne+ as a function of deuterium dose, ion mass, and ion energy.

Post-annealed SEM images suggest that dispersoids prevent significant grain growth and recrystallization of the material at up to 1600°C, consistent with previous work [3]. AES and LEIS analysis reveal a significant increase of carbon on the material’s surface, indicating carbon bonding or carbide formation from annealing. Depth profiling shows little variation in W and Ti concentrations, suggesting that the TiC dispersoids remain stable in the material, even at high temperatures. From DRS measurements, both residual H as well as chemisorbed D are detected as a result of dosing. Once sputter-cleaned, there appears to be no barrier to chemisorption of H isotopes on these surfaces. When dosing is terminated, the H signal disappears due to ion-induced desorption between the incoming Ne+ ions and the chemisorbed hydrogen isotopes. Future work will investigate H chemisorption in greater detail, as well as analyze inelastic energy losses from ion scattering.

For tungsten in divertor parts of future fusion reactors, a fine-grained microstructure is preferred to reduce its brittle-to-ductile transition temperature and minimize cracking events due to cyclic thermal and mechanical loading. Our ongoing study on warm- and cold-rolled tungsten sheets with different degrees of deformation explores the potential of potassium-doping (K-doping) to stabilize the microstructure at high operation temperatures.

Successful production of technically pure and K-doped tungsten sheets equivalently rolled up to very high logarithmic strains of 4.6 was proven in our preceding study, accompanied by in-depth analysis of the evolution of microstructure and mechanical properties after the different rolling steps. The current investigation on the same material batch analyses the microstructural changes by recovery, recrystallization and grain growth in different temperature regimes between 600 °C and 2400 °C.

The annealing studies with subsequent microhardness testing and SEM analysis reveal increasing retardation of recrystallization in K-doped tungsten with increasing rolling strain. This dependence on strain is related to an accompanying improvement of K-bubble dispersion by rolling, resulting in larger Zener pinning forces. However, abnormal grain growth at high temperatures is also enhanced in sheets with higher rolling strain. K-doped tungsten with low rolling strain, on the other hand, shows promising results especially at high temperatures without any signs of abnormal grain growth and with much less normal grain growth than its pure tungsten counterpart, leading to an order of magnitude difference (!) in mean grain size. Additionally, the irregular grain morphology observed therein could prove advantageous against crack propagation compared to the straight polygonal grain boundaries in pure tungsten.

In view of recent results from high heat flux tests on plasma facing components [1], we suggest using K-doped tungsten as armour material to minimize growth of giant grains near the plasma facing surface and risk of grain detachment.

Tungsten (W) is a strong candidate for plasma facing materials in fusion reactors. However, W surfaces form fiberform nanostructures called fuzz when exposed to helium (He) plasma under certain conditions [1]. Furthermore, it has recently been found that when a small amount of impurity is added to the He plasma or with ion energy modulation, nano-tendril bundles (NTBs) are formed [2,3]. In the actual furnace environment, W particles sputtered from the wall may be co-deposited simultaneously with particles in the plasma. In such a co-deposition environment, fuzz growth has been found to be significantly enhanced and the thickness of the fuzz layer can be on the order of millimeters [4]. However, the enhanced growth process is not fully understood. It was found that the LFN growth has always started from the edges of the samples; the condition for the origin of the LFN growth has not yet been explored. In this study, we used W meshes with different numbers of openings and W plates with NTBs used to investigate the initial growth process of LFNs.
　Different types of tungsten were used for samples: pure W plates (10x10x0.2 mm3), W-mesh, and NTB-formed W-plates. For the W meshes, three types of meshes with different mesh openings (the distance between the W wires for the mesh) were used. The samples were exposed to He plasma in the NAGDIS-II linear plasma device with W deposition by using a deeply biased W wire (-450 V) located adjacent to the sample. It was found that the larger the mesh openings, the larger the area where LFN is formed and the faster the formation tends to be. One reason for this is that the He flux becomes higher as the opening angle increases. It is suggested that the concentration of He particles is related to the growth of LFNs. On NTB samples, it was found that NTBs grew significantly when exposed to He plasma with W deposition. The growth rates were evaluated from the size changes in scanning electron microscope (SEM) micrographs, and it was found that the growth rate increased significantly when the size of the NTB reached ~0.1 mm. The results suggest that the structure larger than the sheath thickness is able to collect He ions due to the electric field formed on the fibers and result in the enhanced growth.
[1] S Kajita, W Sakaguchi, N Ohno, et al., Nuclear Fusion 49, 095005 (2009).
[2] K. Woller, D. Whyte, and G. Wright, J. Appl. Phys. 129, 115301 (2021).
[3] D. Hwangbo, S. Kajita, N. Ohno, et al., Nucl. Fusion 58, 096022 (2018).
[4] S Kajita, S Kawaguchi, N Ohno, N Yoshida, Scientific reports 8, 56 (2018).
*Corresponding author: tel.: +81 4 7136 3993, e-mail: kajita@k.u-tokyo.ac.jp (S. Kajita)

Graded tungsten/steel composite, also known as functionally graded material (FGM), may be used as an interlayer to join tungsten (W) armour and steel for the first wall (FW) of future fusion reactor. However, current thermomechanical finite element (FE) numerical simulations of the FW featuring FGM have modelled the FGM (W/steel composites) inappropriately. The mechanical modelling of the FGM has been done by considering it as elastic perfectly plastic; for this, the elastic modulus and yield strength of the composites were assumed to follow a simple linear interpolation of the properties of W and steel depending on the volume content of W, without considering the morphology of the microstructure [1–3] . This overestimates the elastic modulus and yield strength of the composites and results in erroneous stresses inside the FGM. Thus, in this work, a microstructural image-based FE simulation is utilized to study the microscopic deformation behaviour as well as to determine the homogenized macroscopic mechanical properties of the composites (FGM). These obtained homogenized macroscopic mechanical properties can be used to model the FGM appropriately. Similar approach has also been utilized previously for fusion material like W/Cu composite [4]. To accomplish this, W/steel composites of different volume contents of W were manufactured using atmospheric plasma spraying (APS) [5] and spark plasma sintering (SPS) [6] . Three compositions were prepared by APS (25, 50, 75 vol% W) and two compositions were prepared by SPS (25, 50 vol% W). After that, an open source code, called OOF2 [7], was used to transform the 2D scanning electron microscopy (SEM) micrographs of the individual composite to a 2D FE mesh. These FE meshes is then incorporated into a commercial FE solver (ANSYS) and subsequent simulations were performed. The FE simulations were carried out at various virtual temperatures between 20 °C and 700 °C. From each simulation, a global mechanical property of that particular SEM micrograph was determined. The effect of the size of different representative volume elements on the global mechanical property was also investigated. After this, the homogenized mechanical property of each composite was determined by performing several FE simulations on SEM micrographs captured at different locations (in the actual composite) by following a systematic methodology. These homogenized properties were then compared with the experimentally determined property using a miniature 4-point bending test. This microstructure modelling also predicted the deformation behaviour; the model predicted the formation of the possible crack initiation sites, preferentially initiating in steel due to localized plastic straining.

Tungsten fiber reinforced tungsten matrix(W_f/W_m)material is one of the most promising plasma facing materials(PFMs) for the future fusion reactor. Benefit by the energy dissipation via fiber debonding and sliding when the material is confronted with cracking in the matrix, the composite shows laudable ductility, which is usually called-Pseudo Toughness.

Fabricating composite with desired structure and quality is the key point to realize the Pseudo Toughness for fiber reinforced tungsten material. The most popular way to fabricate the reinforced tungsten material with large scale fibers is chemical vapor deposition (CVD) with assistance of temperature gradient control or designed sample rotation, which is facing the challenge of manufacturing efficiency as well and fiber distribution design. This work will present the new fabrication approaches developed at Southwestern Institute of Physics, which include the techniques of powder extrusion printing(PEP), injection molding, ultra-high pressure sintering. The result indicates that, injection molding is a most potential way to realize the composite with 3-dimensional reinforcement structure and the PEP is an efficient way to realize precise arrangement of flat woven structure of the reinforcement. Ultra-high-pressured sintering is a promising way to fabricate the fiber reinforced tungsten material with desired structure and quality. The compression test result indicates that, with rather limited(<10% volume) involvement of tungsten fiber, the compressive strength and the fracture energy of the W_f/W_m composite is 1.25 times and 3.57 times of the ones for the pure tungsten, respectively. The detailed information about the fabrication procedure as well as other mechanical properties of the manufactured W_f/W_m composite will be presented in the paper.

Tungsten is currently the leading candidate material for the divertor component due to its favourable thermal properties and low sputtering yield. Despite these advantages, the tungsten divertor will be subject to high fluxes of a variety of plasma species including hydrogen, helium, beryllium, and nitrogen. Tungsten exposed to these materials results in the formation of mixed materials layers, including tungsten-beryllium intermetallics [1] and tungsten nitrides [2], which will further affect the interaction of other plasma species in the divertor. Experiments have shown higher hydrogen retention [3] in both tungsten-beryllium and tungsten-nitrogen mixed materials layers. The formation of these mixed materials layers and their effects on hydrogen are not well understood. Atomistic modelling can play a key role in further understanding of mechanisms for increased hydrogen retention in these mixed materials layers. However, accurate interatomic potentials are needed for modelling these complex material systems.

In this work, we will describe the development of machine learned Spectral Neighbor Anaylsis Potentials (SNAP) for both W-Be-H and W-N-H. Machine learned interatomic potentials are trained to large datasets of atomic configurations generated using electronic structure methods like density functional theory. This results in greater flexibility and higher accuracy compared to classical potentials, which is needed for the complex chemistry occurring at the divertor surface. A description of the development and performance of these new potentials will be presented. In addition, molecular dynamics simulations of hydrogen implantation in both beryllium and nitrogen exposed tungsten will be discussed.

SNL is managed and operated by NTESS under DOE NNSA contract DE-NA0003525

[1] M.J. Baldwin, R.P. Doerner, D. Nishijima, et al., J. Nucl. Mater. 390-391, 886-890 (2009)

[2] K. Schmid, A. Manhard, Ch. Linsmeier, et al., Nucl. Fusion 50, 025006 (2010)

[3] A. Kreter, D. Nishijima, R.P. Doerner, et al., Nucl. Fusion 59, 086029 (2019)

*Corresponding author: tel.: +1 847 6825414, e-mail: mcusent@sandia.gov (M. A. Cusentino)

Plasma-facing materials in a fusion reactor have to withstand extreme conditions with high heat loads, high energy neutrons, and high particle flux. Typically, the fusion community has made use of carbon based materials or tungsten for high heat flux regions, however, neither of the materials are optimum and there is a huge scope for improvement. To advance the performance of plasma-facing armour to the level required for a fusion reactor (in terms of integrity and resistance to damage), bulk diamond stands out due to its extreme thermomechanical properties superior to any other bulk material. However, the bulk diamond could encounter significant erosion under plasma exposure (e.g. ELMy H mode plasma). Therefore, a composite material engineered from bulk diamond fabricated by the chemical vapor deposition (CVD) method and tungsten thin film has been proposed in this work. A tungsten thin film with a thickness of less than 600nm was deposited on the CVD diamond by the pulsed laser deposition (PLD) technique, which improves the erosion resistance and maintains the excellent thermomechanical properties of the diamond in the meantime. Meanwhile, after the deposition of tungsten, no graphitization process of the diamond has been observed. Microstructure and topography analysis were conducted on the diamond with and without tungsten film by scanning electron microscope (SEM), transmission electron microscopy (TEM), and atomic force microscope (AFM). The results indicate a uniform tungsten film has been deposited on CVD diamond with a roughness of less than 36nm. This work explored a new composite material engineered from CVD diamond and tungsten thin film, which is promising for application in high heat-load plasma surface interaction regions of magnetic confinement fusion devices.

Tungsten (W) is the main candidate material for the first wall of a fusion reactor, as it is resilient against erosion, has a high melting point and thermal conductivity, and shows rather benign behavior under neutron irradiation. However, the intrinsic brittleness of tungsten is a concern with respect to the fusion environment with high transient heat loads. To overcome the brittle issue of tungsten, tungsten fiber reinforced tungsten (W_f/W) composites are being developed relying on an extrinsic toughening principle. One of the crucial factors to realize the extrinsic toughening mechanism is a relatively weak interface between the fibers and the matrix.

As the interface material for tungsten fiber reinforced tungsten composites (W_f/W), yttrium oxide thin films were prepared via an electrophoretic deposition (EPD) method in this study. The yttrium oxide suspension is optimized based on its Zeta potential. The coating process was performed on the tungsten weaves (anode) and stainless steel was used as the counter electrodes (cathode). The coating structure obtained from the electrophoretic deposition (EPD) process was homogeneously distributed on all the fibers of the weave but possesses a porous structure. In addition, to optimize the coating process, the voltage influence on the coating process is demonstrated. Using the EPD prepared yttria interface, W_f/W is produced based on a chemical vapor deposition process. The initial mechanical test shows a promising property with extrinsic toughening mechanisms and high damage resilience.

EUROFER97 is a reduced activation ferritic martensitic (RAFM) steel that is proposed to be used in the structural components of the breeding blanket and the first wall of future fusion devices [1]. To meet the high-heat-flux requirement of power exhaust in the divertor of the demonstration reactor (DEMO), tungsten is considered to be an armor material [2]. Hence, accurate knowledge of the specific energy loss of the fusion plasma constituents in these materials is essential to estimate the induced damage to the plasma-facing components of fusion reactors from e.g. sputtering and other near-surface modification processes.

In this contribution, the electronic stopping cross-sections of H⁺, D⁺ and He⁺ ions in tungsten and EUROFER97 are measured using time-of-flight low-energy ion scattering (ToF-LEIS) in relative measurements for their bulks, employing Au and Cu reference samples, respectively [3]. As a complementary method, these stopping cross-sections are also deduced from ToF-LEIS measurements from sputter-deposited thin films of tungsten, and EUROFER97 [4]. The obtained results from bulk and thin film measurements are consistent. The primary energy of employed ions, in both bulk and thin film experiments, is in the range of 800 eV to 10 keV. The samples are in-situ annealed and sputter-cleaned to have the surface as clean as possible for more precise measurements. Monte Carlo simulations using the TRBS code [5] were employed to evaluate the data and subtract the nuclear stopping contributions. Discrepancies between experimental stopping cross-sections and semi-empirical models such as SRIM are observed and discussed.

[1] Pintsuk G, Diegele E, Dudarev SL, Gorley M, Henry J, Reiser J, et al. European materials development: Results and perspective. Fusion Engineering and Design. 2019;

[2] You JH, Greuner H, Böswirth B, Hunger K, Roccella S, Roche H. High-heat-flux performance limit of tungsten monoblock targets: Impact on the armor materials and implications for power exhaust capacity. Nuclear Materials and Energy. 2022 Oct 1;33.

[3] Roth D, Goebl D, Primetzhofer D, Bauer P. A procedure to determine electronic energy loss from relative measurements with TOF-LEIS. Nucl Instrum Methods Phys Res B. 2013;317(PART A):61–5.

[4] Pitthan E, Petersson P, Tran TT, Moldarev D, Kaur R, Shams-Latifi J, et al. Thin films sputter-deposited from EUROFER97 in argon and deuterium atmosphere: Material properties and deuterium retention. Accepted for publication in Nuclear Materials and Energy.

[5] Biersack JP, Steinbauer E, Bauer P. A particularly fast TRIM version for ion backscattering and high energy ion implantation. Vol. 861, Nuclear Instruments and Methods in Physics Research. 1991.

Sputtering and defect formation due to plasma-wall interactions (PWI) are key processes that must be understood for the operation of future fusion reactors with minimum maintenance [1]. In these processes, fundamental quantities such as the specific energy deposition of plasma species in wall materials or their interaction potentials with wall species play a significant role. Despite being important input variables in modelling of erosion and implantation, such quantities are often not well-known: For example, there are no available experimental datasets for electronic stopping power for slow light ions in tungsten (W). In addition, the influence of these quantities on plasma-wall related parameters such as sputtering yields still needs to be investigated, although a direct effect on the characteristic length scales of collision cascades is obvious.

In this contribution, recent studies of these fundamental quantities obtained both experimentally and from calculations will be presented. More specifically, experimentally deduced energy losses and short-range repulsive potentials of impacting plasma species (H, D, and He) in candidates for plasma facing materials of next generation fusion devices (W, Fe, and EUROFER97) are evaluated. For that, ions with a wide range of energies (sub-keV to few MeV) [2] and different targets (bulk samples, pre-irradiated materials, and deposited thin films) are used. The experimental results are compared with calculations using time-dependent density functional theory and advanced molecular dynamics. The sensitivity of statistical quantities such as sputtering yields on the magnitude of the aforementioned parameters is furthermore tested with simulation codes based on the binary collision approximation. Our recent results thus not only provide necessary datasets for these fundamental quantities but also directly enhance the understanding on how these values influence ion-solid interactions relevant for future fusion reactors.

[1] S. Brezinsek et al., Nucl. Fusion. 57, 116041 (2017)

[2] P. Ström and D. Primetzhofer, J. Instrum. 17, P04011 (2022)

In this study, we perform large-scale machine learned molecular dynamics simulations to investigate the influence of tantalum concentrations on the incipient plasticity mechanisms of crystalline tungsten matrices during room temperature nanoindentation. The tabulated Gaussian approximation potential (TabGap) framework is applied to describe W-Ta interactions. We focus on (001), (011), and (111) orientations to tracking the nucleation and interaction of dislocations where results for pure tungsten predict the nucleation of twins at two different locations on the {112}<111> family, one pointing downward and the other slightly inclined with respect to the {111} indented surface orientation. These twinning systems also become active for tungsten-tantalum alloys, but the downward-pointing system becomes more predominant exhibiting defect nucleation. Our results suggest that surface energy is a key variable in the detwinning process, with higher surface energy leading to twin instability and breakdown into dislocations. However, a more detailed analysis reveals that the detwinning processes in the slightly inclined system of the W surfaces (TabGap) are as pronounced as those modeled with the Ravelo potential for Ta in the same twinning system. On the other hand, twin annihilation in the vertical system is more facilitated in W TabGap compared to Ta TabGap. This suggests that the onset of tungsten-tantalum alloys approaches the marked twin annihilation observed in pure tungsten. Our findings indicate that a complete twin nucleation criterion should consider other components of the stress tensor in addition to resolved shear stress. These findings have implications for the design of next-generation fusion machines that rely on refractory BCC metals and their alloys.

The extreme thermal energy deposited on the DEMO Plasma-Facing Components (PFC) causes melting and concomitant evaporation of the armor. The primary material for the PFC of DEMO was chosen tungsten. The development of advanced tungsten grades requires thorough testing and qualification of material in extreme fusion-relevant conditions, including both high heat and particle fluxes [1].

Pure tungsten in the recrystallized state [2] and lattice structures (AM W/WTa) fabricated with tungsten powder by additive manufacturing techniques [1, 3] were used for the experiments. The studies property of such grades of tungsten were performed at a surface heat load of energy density up to 0.75 MJ/m² (above the melting threshold) and the pulse duration of 0.25 ms within quasi-stationary plasma accelerator QSPA Kh-50 [3, 4]. Surface analysis was carried out with an SEM. The development of surface morphology of the exposed targets as well as cracking on the surface are discussed. Surface modification and crack development led to increasing in the roughness of exposed surfaces.

Analysis of surface modification of recrystallized (R) sample was performed after 100 QSPA plasma pulses causes pronounced tungsten melting. The re-solidified layer formed on the exposed surface. Erosion of Latticed AM W/WTa samples exposed within a small (up to 10) number of QSPA plasma was also evaluated. A large number of ejected particles after the first plasma impact could generate due to the removal of weakly bounded fragments from the surface. The number of ejected particles decreased with an increasing number of plasma pulses. Cracks, pores, and balls are observed on the exposed surfaces. Fine cellular structure with a typical cell size of 150…250 nm appears in the re-solidified layer as a result of plasma exposures on affected surfaces of both grades of tungsten. Intergranular cracks with widths up to μm appeared as a result of surface irradiation.

[1] J.H. You et. al. Fusion Engineering and Design 174, 112988 (2022)

[2] M. Wirtz et al. Nuclear Materials and Energy. 12, 148 (2017)

[3] N. Mantel et. al. Nucl. Fusion 62, 036017 (2022)

[4] V. A. Makhlai et. al., Phys. Scripta 196, 124043 (2021)

Performance of dispersoid-strengthened tungsten (W) alloys has been studied in the divertor of DIII-D tokamak using the Divertor Material Evaluation System (DiMES). Dispersion-strengthened W alloys are being considered as candidate PFC materials due to their enhanced thermomechanical properties compared to pure tungsten [1]. These alloys are produced using spark plasma sintering of W nanopowders alloyed with 1-10 wt.% TiC, TaC, and ZrC nanopowders. The samples segregate impurities to the dispersoid locations rather than to the W-W grain boundaries, resulting in a more stable microstructure and a higher recrystallization temperature [2]. To investigate the performance of these alloys under divertor plasma conditions and measure the impact of dispersoids on material sputtering and surface morphology, samples of W-1.1 wt.%TiC, W-1.1 wt.%TaC, and W-1.1 wt.%ZrC were exposed to DIII-D H- and L-mode plasmas using the DiMES manipulator. The H-mode plasmas had an edge localized mode (ELM) period of 10-20 ms, inter-ELM heat flux of 0.58 MWm^-2, intra-ELM heat flux of 2.57 MWm^-2, and total particle fluence of 5.8 x 10²¹ m^-2. The L-mode plasmas had a heat flux of 0.47 MWm^-2 and a total particle fluence of 6.6 x 10²¹ m^-2. The surface morphology of the samples was analyzed before and after exposure using various techniques, including optical profilometry and scanning electron microscopy (SEM). SEM micrographs of the samples showed blister-like surface features on the samples. These features were considerably larger in H-mode discharges (5-30 µm) than in L-mode discharges (200-500 nm). The blisters were largest in W-1.1 wt.% ZrC samples exposed to H-mode plasmas, where the samples were estimated to have reached 750 K. These features also showed significant surface impurity concentration (> 80% C, O, Si, Cu, and other impurities), as measured by energy-dispersive x-ray spectroscopy. Optical profilometry of the exposed samples showed depressions or “divots” at the locations of the dispersoids. This was most prominent in the W-1.1 wt.% TiC samples, in which the divots grew from <1 µm deep in the control sample, to 2.5 µm for L-mode exposure, to 3.5 µm deep for H-mode exposure. To measure the erosion yield of tungsten and the second phase dispersoids, spectroscopy data was collected for individual samples during the discharge using visible emission spectroscopy for the W-I, Ti-I, Zr-I, and Ta-I spectral lines, the results of which will be presented.

[1] E. Lang, et al., Int. J. Refract. Met. Hard Mater. 75 (2018)

[2] E. Lang, et al., J. Nucl. Mater. 545 (2021)

Work supported by DOE Award No. DE-SC0020093, DE-SC0021005, DE-FC02-04ER54698, DE-AC05-00OR22725, DE-AC52-07NA27344, DE-FG02-07ER54917, and DE-NA0003525.

To predict hydrogen isotope behavior in plasma-facing materials accurately, it is necessary to evaluate the effects of neutron-induced transmutation combined with He implantation. These effects should be studied in conjunction with high-flux D plasma exposure. In our previous studies, we showed D decorates defects produced by neutron irradiation. In such cases, the D permeation rate for damaged W was reduced, possibly because disorder in the lattice served to block D diffusion pathways. The fraction of transmuted Re in the neutron irradiated W was quite small due to low neutron fluence. Even so, it was reported that hydrogen isotope retention behavior for W-Re was quite different from that for W, where the measured thermal desorption peaks and their temperatures shifted significantly.

The results described above merit further study, as it could affect the detailed design activity for DEMO. The effect of Re on hydrogen isotope retention should be well understood under high-flux plasma exposure. Consideration of these factors, along with the effects of Re and D+He mixed-species plasma exposure conditions is being pursued under the Japan-US collaborative research, FRONTIER. The Deuterium Plasma Experiment (DPE) device, a high-flux RF-plasma source, has several unique features well-suited for this kind of work. It includes precise sample temperature control, the capability for D+He mixed plasma exposure, optical spectroscopy characterization tools, and the ability to expose multiple small specimens at once. This paper focuses on the comparison of D retention for W and W-Re materials following DPE D+He mixed plasma exposure.

The samples used in this study were stress-relieved polycrystalline W and W-10%Re. They were polished to a mirror-finish and annealed at 1173 K for 30 min. Before installation into DPE device, the samples were cleaned ultrasonically in ethanol and acetone to remove surface organic impurities. XPS and AES analyses were then applied to evaluate the surface chemical states. Thereafter, three samples with 6 mm diameter x 0.5 mm thickness, were inserted into DPE and were annealed with a resistive heater to 673 K. D and He gases were supplied by mass flow controller, producing a neutral pressure consisting of a 10:1 D to He ratio. The flux of D+He was ~1.0 X 10²⁰ m² s^-1 (measured by a reciprocating Langmuir probe), and over the length of the exposure we were able to reach a total fluence of 1.0 × 10²⁵ m².

XPS analysis showed that major chemical states of W were W-metal and W-O. Significant D desorption was observed at ~ 900 K, with the desorption peak shifting toward higher temperatures with increasing fluence. The D retention for W-Re was lower compared with pure W. This indicates that the presence of Re reduces D trapping, even in cases where there is increased damage from plasma exposure or ion exposure. In this presentation, more detail experimental results obtained by XPS, AES and TDS will be discussed to further support these findings.

Due to low sputtering yield and low fuel retention, tungsten has been used as plasma-facing material in several tokamak fusion experiments. In stellarators, it has been employed only recently for some of the plasma-facing components. However, apart from its high costs, W is difficult to machine due to its hardness and brittleness and therefore alternative materials in the form of tungsten heavy alloys are being explored and some tests have already been carried out in ASDEX Upgrade [1]. Being para-magnetic, WNiCu alloy has been envisaged for some Wendelstein 7-X (W7-X) plasma-facing components. The WNiFe materials are magnetic however, since the magnetization saturates e.g. for W97NiFe at ~ 1 Tesla [1], these will also be explored for the use in W7-X.

Small samples were prepared from pure W, W95NiCu, W97NiFe and W95NiFe alloys. After the successful tests performed in high heat flux test facility GLADIS, the samples were exposed in the stellarator Large Helical Device (LHD) during recent operation campaigns. The samples were inserted using the divertor manipulator at strike line locations under H, D and He plasma conditions. These experiments were designed to test the samples at high heat loads of about 10 MW/m² and sample temperatures at normal operating conditions i.e. < 800°C and overload conditions of > 1000°C. During the exposures at more than 1000°C, despite the overload (reaching temperatures above the melting limit and leading to segregation of Cu/Ni and partial release of alloy material) no disturbance of the plasma operation itself occurred. Melting and cracking were observed on part of the samples with weight loss after the exposures. The scanning electron microscopy with focused ion beam (SEM/FIB) and energy-dispersive X-ray spectroscopy (EDX) measurements confirmed the observed modification of the surface morphology. A detailed post-mortem analysis will be reported.

[1] R. Neu et al., Fusion Eng. Design 124 (2017) 450.

*Corresponding author: tel.: +49 3834 882770, e-mail: chandra.prakash.dhard@ipp.mpg.de (C. P. Dhard)

It has been recognized that the production and dynamics of dust in the vacuum chamber of tokamaks are important problems in the framework of the safety and tokamak performance [1]. It is expected that during plasma discharges most of the dust particles concentrate in the scrape-off layer close to the chamber walls [2]. For almost all materials the hypervelocity regime (when the speed of an impact exceeds the speed of the compression waves both in the target and in the projectile) is reached when the impact speed exceeds a few km/s; it is therefore common to consider velocities above 2–3 km/s as hypervelocity impacts [3]. In studies related with plasma facing materials (PFMs) for future nuclear fusion technology, high velocity impacts have been reported, with velocities being around 500 m/s and up to a few km/s [4, 5].

In this study we focus on understanding the fundamental characteristics of the mechanisms underlying the crater formation caused by nanoparticles impacts on PFMs. From molecular dynamics involving very large samples (up to 40 million atoms). We have determined the detailed atomistic and thermodynamic aspects of crater formation mechanism. Different stages of the penetration process are identified, and a model to understand the damage produced by hypervelocity impacts in terms of geometrically necessary dislocations, much like in classic indentation theory, will be discussed.

References

[1] Federici G. et al 2001 Nucl. Fusion 41 1967.
[2] Tsytovich V.N. and Winter J. 1998 Phys.—Usp. 41 815.
[3] Burchell M.J., Cole M.J., McDonnel J.A.M. and Zarnecki J.C. 1999 Meas. Sci. Technol. 10 41.
[4] Castaldo C, et al. 2007 Nucl. Fusion 47 L5-L9.
[5] Ratynskaia S, , et al. 2008 Nucl. Fusion 48 015006.

Plasma-facing materials (PFMs) in future fusion devices will be exposed to demanding operating conditions involving high heat fluxes, aggressive particle and neutron fluxes, and high stresses. Although tungsten has emerged as a promising candidate, unresolved issues include stability against recrystallization, plasma-induced surface damage, and degradation of bulk properties due to irradiation. Deliberate alloying of tungsten in the nanocrystalline state has been shown to enhance stability against thermal and irradiation induced microstructural changes. Similar alloying additions have also markedly improved crack mitigation in laser additively manufactured tungsten alloys. Together, these materials form a class of novel interface engineered tungsten alloys, which we show in this presentation to be stable against recrystallization and runaway grain growth up to 1500°C. Dopant species are identified through lattice Monte Carlo modelling and used to guide powder metallurgical processing of ternary tungsten alloys. Optimized chemistries containing nanoscale compositional heterogeneities, specifically titanium segregated to grain boundaries and chromium rich nanoprecipitates, are synthesized through high energy ball milling. Microstructural transitions during milling of the alloy powders are characterized through coupled transmission electron microscopy and in situ synchrotron x-ray diffraction experiments, which reveal correlations between microstructural state and thermal stability. Direct current sintering is employed to produce fully dense tungsten alloys containing a stable ultrafine grained microstructure containing engineered carbide nanoprecipitates. Synergistic doping of interfaces in the nanocrystalline state is therefore shown to provide a pathway for stabilizing tungsten above its common recrystallization point for PFM applications.

The comprehensive understanding of the interaction between plasma and plasma-facing materials constitutes a critical issue for the safe operation and the prediction of the life time of the first wall of fusion devices. From 2011 the JET tokamak, the largest fusion device with magnetic confinement, is used as a testbed for the next generation fusion device, the ITER. So, the carbon (C) wall of the JET Tokamak was replaced by beryllium (Be) in the main chamber and tungsten (W) in the divertor, a configuration which will be used in ITER (ITER - Like Wall). After the transformation, three D-D experimental campaigns were carried out in order to test the interaction between plasma and the plasma facing materials.

In the current work, tile 0 and 1 from the ILW JET Tokamak divertor exposed to the three experimental campaigns were investigated with different analytical techniques in order to assess fuel retention, material migration and deposition, and surface morphology changes. The light deposited elements are Be, C, N and O. Be is the dominant one and its content varies in the range (2 - 7.5) × 10¹⁹ at/cm² for the first 10 μm. Their deposition layer is generally thick (> 5 mm) and the areas with the highest deposition thickness is the end of tile 0 and the start of tile 1 where it extends beyond 12 μm. In these areas, Be has also the highest content. Be and C have inhomogeneous spatial distribution on the surface of both tiles. Additionally, many heavy elements were also detected: Al, P, Cl, Ti, Cr, Fe, Ni, Cu and Mo with Ni presenting the highest concentration. The origin of Be and Ni is from the main chamber, where Be is used as plasma facing material, and Ni as interlayer in order to determine the erosion of the tiles. The deuterium (fuel) content varies from 2.4 × 10¹⁸ at/cm² to 5.8 × 10¹⁸ at/cm², while no correlation between fuel retention and deposition layer thickness is observed. Melted areas and micro-cracks were detected on the surfaces of the tiles.

The more demanding Plasma Wall Interactions expected in EU-DEMO compared to ITER require examining heat exhaust alternatives to the baseline W divertor. The use of liquid metals as Plasma Facing Materials has been considered due to their low sensitivity to neutron damage and improved heat removal, especially during transients [1]. The design proposed by ENEA consists of a porous structure to support liquid tin (Sn) [2]. However, the corrosive action of this metal requires protective coatings on structural components, especially those made of CuCrZr alloy. In this respect, W-based films show promising thermo-chemical properties.

Concerning the coating fabrication method, Physical Vapor Deposition (PVD) techniques offer superior control over nanoscale properties [3]. Specifically, High Power Impulse Magnetron Sputtering (HiPIMS) applies high amplitude voltage pulses at a low duty cycle to achieve sputtering. The increased peak power improves the plasma density above the cathode and the degree of ionization of sputtered species [4]. Thus, the energy of such particles can be tuned by applying a bias voltage to the substrate. It is therefore possible to obtain smooth and dense coatings and control their properties. In addition, HiPIMS provides conformal coverage of complex substrates, an important requirement for the deposition on structural components.

Here we report on the HiPIMS deposition of compact W-based coatings on fusion-relevant CuCrZr substrates as protective barriers against liquid Sn corrosion. HiPIMS pulses (100 µs-long, duty cycle of 1.75%) were applied to the tungsten cathode at a fixed gas pressure of 0.5 Pa. We employed argon as working gas to produce W coatings, while a reactive deposition in a mixture of argon and nitrogen allowed to obtain WN_x films of varying stoichiometry. We explored different deposition conditions to tune the coatings properties. Growth, morphology, and nanostructure of the films were characterised. Lastly, we present preliminary results about liquid Sn exposure of coated copper substrates in divertor-like conditions.

[1] T.W. Morgan, P. Rindt, G.G. van Eden, et al., Plasma Phys. Control. Fusion 60, 014025 (2018)

[2] S. Roccella, G. Dose, R. de Luca, et al., J. Fusion Energy 39, 462-468 (2020)

[3] D. Dellasega, G. Merlo, C. Conti, et al., J. Appl. Phys. 112, 084328 (2012)

[4] K. Sarakinos, J. Alami, S. Konstantinidis, Surf. Coat. Technol. 204, 1661-1684 (2004)

The first wall of future fusion reactors most probably will be equipped with an armour layer consisting of a layer of a few millimetres of tungsten. Currently it is planned to apply this layer by a thermal coating technique. In order to determine whether cold spray coatings of W and Ta are suitable for plasma facing components (PFCs), high heat flux tests under low power (≤ 4 MW/m²) long pulse loads and high power (40 MW/m²) transient loads were performed on W-Ta test coatings on P92 in the high heat flux (HHF) facility GLADIS. The samples with sizes of about 80x80mm² and a layer thickness of 2mm were prepared by the company Impact Innovations. Two different W powder types were used. Powder 1 (fine) is a spherical powder with a particle size of ≤ 25 µm and powder 2 (coarse) is a polycrystalline powder with a particle size of ≤ 45 µm. The used spherical Ta powder with grain size of ≤36 µm was the same for all samples. The W content in the coatings is about 70%. In the first high heat flux tests, both test layers passed the conditions under low long pulse loads (≤ 4 MW/m²). However, the fine powder layer delaminated at the high-power loads after screening and 27 subsequent cycles of 40 MW/m². The layer made from coarse powder failed already after screening and only 3 cycles at 40 MW/m². Metallographic examinations were performed on the specimens to characterize the defects. The specimens were cut along the long side and metallographic cross sections were performed along the whole length of the sample. Subsequently, light microscopic and electron microscopic examinations were performed on the samples. The light microscope images show that the large stress cracks run from the edge through the lowest layers of the cold spray coating. Other smaller cracks were also found in subsequent layers of the coating. The bonding between the coating and the substrate was intact throughout. The cracks are due to compressive stresses in the layer. Therefore, in a 2nd experiment, a heat treatment at 950 °C for 1h on the samples was carried out before the high heat flux tests. The effect of the heat treatment is determined by measuring the bending of the samples before and after the heat treatment as well as from metallographic cross sections of witness samples. The new samples are subjected to HHF tests at GLADIS similar to the first experiment and the metallographic/microscopic investigations were performed similar as for the first samples. The contribution will report on the detailed comparison of the experiments and the consequences for further optimization and perspectives of the cold spray technique for the production of PFCs.

Tungsten is considered the material of choice for the divertor application of fusion power plants due to its intrinsic thermo-physical properties. However, high ductile-to-brittle transition temperature (DBTT) and the loss of yield strength due to recrystallization at temperatures exceeding 1000 °C are generating further research in a quest for its improvement. This work aims to improve the tungsten’s properties to sustain long-term plasma-facing conditions imposed on the surface of the divertor.

Among various particle-reinforced W-based composites, W₂C reinforced W is emerging as a promising plasma-facing material due to its microstructural stability and thermo-physical properties. W₂C particles are formed in situ during the densification of a powder mixture consisting of W and WC particles (4 at % of carbon in the form of WC nanoparticles, sample denoted as W-4WC) with a field-assisted sintering technique (FAST). Microstructural analysis revealed isotropic microstructure consisting of cubic W and hexagonal W₂C inclusions positioned at the tungsten’s grain boundaries. In addition to the microstructural and phase analysis, thermo-mechanical properties at room and elevated temperature and high-heat-flux tests were carried out. Thermo-mechanical properties measured up to 1000 °C revealed the materials’ DBTT is decreased, and ductile behaviour can be observed already at 200 °C. With the satisfying thermal conductivity, which does not drop below 100 W/m K at elevated temperatures (up to 1000 °C), retention of strength at elevated temperatures and promising HHFT behaviour, this composite makes an interesting alternative to pure tungsten.

Tungsten (W) is a promising candidate material for the plasma facing components in future fusion reactors due to its excellent properties with high temperature and plasma erosion resistance. However, it has issues regarding the intrinsic brittleness as well as operational embrittlement. Tungsten fiber reinforced tungsten composites (W_f/W) overcome these issues by using extrinsic toughening mechanisms similar to those in fiber reinforced ceramic matrix composites. The as-fabricated tungsten fiber exhibits a superior strength and ductility the same time. However, due to neutron irradiation and high temperature recrystallization during fusion operation, the ductility of the fibers may degrade during a long period of service. Therefore it is necessary to investigate the fracture behavior of W_f/W composites with brittle fibers. In the present work, the tungsten fibers were embrittled by carbonization. Subsequently, the continuous brittle fiber reinforced W_f/W composite with yttrium oxide (Y₂O₃) interface were fabricated by the field assisted sintering technology (FAST) process. The microstructure of the prepared W_f/W composites were characterized. 3-point bending test with pre-notched specimens were used to analyze the fracture process and a corresponding finite element model was established. The fracture behavior and toughening mechanisms were discussed in detail based on the experimental results and finite element simulation. The composite showed a pseudo-ductile fracture behavior. Cracks were hindered and deflected by the de-bonded fiber-matrix interface. Although the damage resilience of the W_f/W with brittle fibers is not comparable to that of the W_f/W with ductile fibers, extrinsic toughening mechanisms of interface de-bonding, crack bridging and fiber pull out still perform. This indicates the reinforcement concept still works in spite of the fact that all components in this W_f/W are brittle.

Next-generation fusion devices require advanced plasma facing material solutions. One such design incorporates a W layer armor on top of a SiC support layer, to leverage the low erosion yield and thermo-mechanical properties of W with the favorable neutronics properties of SiC [1]. Towards this end, initial W-SiC bi-layers fabricated at General Atomics suffered delamination and ablation of the W layer, during heat flux testing in the DIII-D tokamak divertor, presumably due to the high strain at the W-SiC interface [2]. To better design such armor materials, a detailed understanding of possible structures within the W-Si-C ternary phase space is needed. To model the interface structure we employed a genetic algorithm along with density functional theory (DFT) calculations to explore a very large configurational space to identify stable structures within the W-Si-C ternary phase space, for which there are no currently known compounds. In addition to ground state energy, other properties of W-Si-C phases are assessed. Subsequently, DFT calculations containing interfaces of newly identified phases were performed to assess the cohesion, stability, and strain of dissimilar phases. The results revealed previously unknown stable W-Si-C compounds for which the interfacial stability and cohesion are calculated to be much stronger, with much less strain, than the bi-layer W-SiC interface.

In order to synthesize the predicted W-SiC functionally graded coatings, a new pulsed sputtering DC process has been developed, in which the W-SiC interface is replaced by a gradient zone of varying chemical composition [3]. The gradation eliminates dissimilar interfaces that act as stress concentrators and mitigate strain-induced delamination. Compared to the W-SiC bi-layers, the functionally graded coating performed better during L-mode DIII-D plasma discharges, with no delamination and significant microstructural changes; however, the performance under more extreme conditions have yet to be determined.

Work supported by GA Corporate funds.

*Corresponding author: tel.: 858-455-3019, e-mail: bergstromz@fusion.gat.com (Z. Bergstrom)

[1] M. Tillack et al., Fusion Engineering and Design 180 (2022) 113155

[2] G. Sinclair et al., PMIF 2019 Conference

[3] Z. Lin et al., APS-DPP 2022 Conference

The efficiency of future fusion reactors depends, among other things, on materials' properties and their long-term stability. The divertor, the most thermally loaded component of the reactor chamber, consists of a sequential array of rectangular W-monoblocks connected to a copper alloy (CuCrZr) cooling tube passing through the central area of the monoblocks. Tungsten (W) is the material of choice for plasma-facing applications due to its intrinsic properties, such as its high melting point, resistance to erosion and low tritium accumulation. High expected thermal loads impose significant stress on the surface and the interface between the W-monoblock and CuCrZr heat sink.

A composite consisting of W and Cu has been proposed to mitigate the thermal stress at the interface. The composite is called a pseudo-alloy since the significant difference in melting points does not allow the creation of a strong bond at the W/Cu interface, which is an obstacle to its production. Therefore, several methods should be combined to create such material. The first step is to manufacture the W-frame, and the second is to infiltrate molten Cu. Additive manufacturing methods will make it possible to design W-frames with optimised geometry that better conducts heat.

A series of W-samples with different porosities were created to investigate the optimal infiltration conditions using the field assisted sintering technique (FAST). The variation in pore size was made by mixing two W powders. The first powder was with a particle size of approximately - 1.5 μm, and the second 30.6 μm. The average pore sizes of the sintered samples were in the range of 0.8-7.8 μm. During the subsequent molten Cu infiltration, the effects of temperature, infiltration time, atmosphere, and W pre-treatment methods on the process were evaluated. The finished W-Cu composites were characterized in terms of porosity and pore geometry, and it was found that: the concentration of Cu in the infiltrated W-frames was higher in the hydrogen atmosphere than in the argon atmosphere; increasing the infiltration time did not increase the Cu concentration in tungsten samples; if the molten Cu is filled through the frame and not the side surfaces, an uninfiltrated middle part of the frame can be avoided. The results will be applied to the infiltration of engineered W frames made by laser powder bed fusion.

Qualification of W Heavy Alloys as Plasma Facing Material

Bernd Böswirth*, H. Greuner, K. Hunger, H. Maier, R. Neu

Max Planck Institute for Plasma Physics, Boltzmannstr. 2, 85748 Garching, Germany

Tungsten (W) is a favoured plasma facing material (PFM) for the highest loaded divertor areas in fusion experiments. For excellent thermal and physical properties, the disadvantages like low fracture toughness at room temperature, high material and machining costs, are accepted. Areas with moderate steady state heat fluxes up to 10 MW/m² allow to consider W heavy alloys materials [1]. These commercially available materials allow to reduce the material and manufacturing costs. Beside the good mechanical properties, the content of Ni/Cu, respectively Ni/Fe, limits the thermal performance.

Therefore, a series of experimental investigations on the thermomechanical and plasma physical suitability was performed on material provided by WHS Sondermetalle, Germany and Plansee SE, Austria. Adiabatically and cyclic steady-state loading of W, W-Ni-Cu and W-Ni-Fe materials have been performed at the high heat flux (HHF) test facility GLADIS using a hydrogen neutral beam.

In a first step a 15 MW/m², 1.2 s and 40 MW/m², 0.2 s adiabatic loading was performed close to the melting of Cu/Ni (T_liq~1380°C) to study the behavior under transient cyclic heat load. In a second step actively cooled small-scale flat tile mock-ups equipped with W-Ni-Cu, W-Ni-Fe and W tiles, respectively, were manufactured and tested with steady state heat loading of 10 MW/m² and 12 MW/m². The design of the CuCrZr cooling structure covered with 5 mm flat tiles ensured surface temperatures below 1100 °C. In this contribution we will present and discuss the results of the GLADIS loading and the microscopic examination with respect to the integrity of the components after the performed HHF loading.

Besides their behavior in HHF tests another important aspect for the qualification of these W heavy alloys for us a plasma facing material is fuel retention. Therefore, an examination of the deuterium (D) retention was performed for the material HPM1801 similar to the one already reported for the HPM1850 W-Ni-Fe alloy [2].

The results from the steady state HHF tests as well as the one from the complementary D retention measurements seem to confirm the previous results on the potential of W heavy alloys as plasma facing material.

[1] R. Neu et al., J. Nucl. Mater. 511 (2018) 567 - 573

[2] H. Maier et al., Nucl. Fusion 60, 126044 (2022)

*Corresponding author: tel.: +49 89 3299 1245, e-mail: bernd.boeswirth@ipp.mpg.de

Self-propagating High-temperature Synthesis (SHS) is a process used to synthesize materials using highly exothermic solid-state reactions, with a final temperature typically in the 1800 – 3500 K range. Because of these high temperatures, which allow self-purification of the product by elimination of any volatile impurity, high purity products might be obtained. By the use of specific additives, the final temperature might be controlled, and nanometric powders can be retrieved. Previously, 20g batches of pure tungsten, and of tungsten-chromium-vanadium have been synthesized1. Our goal here is to study the scaling up of the process, as significant variations in the microstructure have been observed when larger batches have been considered. Using SPS to consolidate the powders, we aim at obtaining dense bulk tiles used as plasma facing materials in the ITER project, while keeping the nanostructure of the powders, in order to obtain ductility.

1. SHS Synthesis and SPS Densification of Nanometric Tungsten - Dine - 2018 - Advanced Engineering Materials - Wiley Online Library. https://onlinelibrary.wiley.com/doi/full/10.1002/adem.201701138.

Tungsten 8 W) is one of the candidates for plasma facing materials (PMFs) in the fusion reactors due to its higher melting point and lower sputteriing yield. W will be exposed to high flux deuterium (D) and tritium (T) including helium (He) ash during operation. In addition, 14 MeV neutrons, produced by nuclear fusion reaction of D and evaluation of hydrogen isotope to the transmutation of rhenium (Re) in W. Therefore, the evaluation of hydrogen isotope permeation behavior for W-Re alloys are important.

W-10%Re alloy disks with the sizes of 6 mm diameter and 0.5 mm thickness, were used. The sample was installed in the plasma driven permeation (PDP) device at Shizuoka University and HD permeation behavior was studied at the temperature range between 723 K and 823 K. A 13.56 MHz RF power supply with a maximum output of 3 kW was used, plasma was generated at a discharge power of 550 W under a gas pressure of 1.0 Pa. The plasma flux wa calculated from the current-voltage (I-V) characteristics measured by the double Langmuir probe. The HD mixed plasma flux ws 1.0×10²¹ m^-2 s^-1, and the permeated H₂, HD, D₂, were measured with a mass spectrometer calibrated with a deutrium standard leak. The HD ratio in the plasma was determined using plasma spectroscopy. The results of HD permeation behavior for W-10%Re alloy was compared with that for W.

The diffusion coefficient for W-10%Re was higher than that for W, and the steady permeation flux was reduced, according to the experimental results of D permeation behavior in the HD (50:50) plasma irradiation experiment. This suggests that the hydrogen isotope permeation behavior was changed due to the reduction of D surface recombination coefficient and the D accumulato near the surface region by Re addition. The steady permeaton fluxes of H and D for W-10%Re at 823 K were changed. The highest permeation flux of HD was achieved at the H:D ratio of 20:80, which was quite different from that for W. This suggests that the dissolution rates of H and D on the surface of W-10%Re alloy included the isotopic effect on the permeation process for W-10%Re . In the presentation, we will discuss in the more details the results of HD mixed plasma irradiation in W and W-10%Re irradiation defects.

One of the most demanding challenges confronted by the construction of commercial nuclear fusion reactors is the development of new advanced materials that are more resistant to extreme operating conditions. Tungsten is postulated as one of the most promising candidates as a plasma-facing material (PFM) in future fusion reactors, as it satisfies most of the highly demanding requirements. However, previous works indicate that W also has major drawbacks, being a key one its deleterious tendency to easily retain light species (He and H), which causes, among other fatal effects, the formation of bubbles, cracks, and exfoliation of the material [1]. So far, numerous publications have reported that nanostructured materials are, under certain circumstances, more resistant to radiation, mainly due to the high density of grain boundaries (GB), see ref. [2] and references wherein, whose contribution shifts the radiation-damage threshold to higher values.

This implies that the overpressurization of the LIA bubbles can be delayed, thus increasing the limit of radiation damage that the material can accept without seriously compromising its performance. In an optimal scenario, the GBs would not only behave as defect sinks, but as effective diffusion channels as well, by promoting the outgassing of defects.

The present work implements a combined DFT and MD methodology to carry out an energetic and structural analysis, and to calculate migration barriers of the different defects toward and along the GB. We do it for 456 atoms in a W(110) / W(112) interface. Due to the high computational cost of DFT calculations when dealing with complex arrangements such as those tackled here, our approach has consisted of previously relaxing the system via MD simulations. This allows us to estimate initial DFT configurations which are closer to the equilibrium greatly reducing the computational time. It also allows simultaneous testing of the MD interatomic potential.

[1] I. Beyerlein et al, Prog. Mater. Sci. 74 (2015) 125.

[2] R. González-Arrabal, A. Rivera and J. M. Perlado, J Matter Radiat. Extremes 5, 055201 (2020); https://doi.org/10.1063/5.0010954

Due to its favorable properties under high heat and particle fluxes, tungsten
has been chosen as a prime candidate material for plasma facing components in
fusion devices such as ITER (divertor) and DEMO. In deuterium-tritium fusion
tungsten its subjected to 14 MeV neutrons which will cause modification crystal
structure, and accumulation of radiation-induced defects. These defects introduce new traps increasing hydrogen retention which has been identified as a
safety, operational, and economical concern. To study neutron damage protons
and heavy ions are used as a proxy.
When introducing solutes a lower production rate of high-energy defects (such
as vacancies and voids) was observed, this is argued to be driven by similar
processes to those which cause radiation-induced segregation (RIS) in fission
materials.
The research proposed aims on testing this hypothesis using self-irradiated
tungsten-alloy samples. The type and density of the defects was studied with
Positron Annihilation Lifetime Spectroscopy (PALS),Positron Annihilation Doppler
Broadening and TEM (Transmission Electron Microscopy). The comparison
between the indirect measurement (PALS) and the directs one (TEM) is performed. The samples are then exposed to deuterium plasma at the the linear
plasma generator Upgraded Pilot-PSI in DIFFER at 500K with a 10^20m2/s flux.
Deuterium retention is measured with Nuclear Reaction Analysis and Thermal
Desorption Spectroscopy.
Finally in order to verify our hypothesis we compare the experimental results
with the Perks and MIK RIS models.

As one of the promising candidates for plasma-facing materials in fusion reactors, tungsten (W) needs to withstand a constant bombardment of energetic particles, including deuterium (D). This bombardment of D can alter physical properties of W, posing threats to normal functioning of W components. In addition, hydrogen isotope retention can also be an issue for plasma-facing materials. Thus, it is of importance to develop comprehensive understanding of D behaviour in W, i.e. what are the most probable locations of implanted D within the W lattice and how the theoretical understanding agrees with the experimental observations.

Nuclear reaction analysis in channeling mode (NRA/c), provides a unique method to detect D positions. In our study, we developed a Monte Carlo simulation program for NRA/c, which was incorporated into a code called RBSADEC [1] that can take arbitrary atomic structures as input and generate signals comparable to those of experiments. We fitted experimental results [2,3], as shown in Fig.1*, using the nuclear reaction of D (³He, p) ⁴He. Results show that some D atoms are displaced from the tetrahedral interstitial sites of W, indicating a strong influence of radiation defects. Furthermore, ab initial calculations based on the density functional theory (DFT) were performed to provide most probable locations of D atoms within the W lattice to be used in the NRA/c simulations. In this process, we took into account different scenarios of D inside a defective W target, for example, single or multiple D atoms inside a mono-vacancy. This combination of NRA/c and DFT can provide a fundamental insight on the lattice location of D in different radiation conditions by connecting theoretical calculations with experimental observations.

[1] RBSADEC. Available at https://gitlab.com/xinJin/rbsadec

[2] S. Picraux, F. Vook, Phys. Rev. Lett., 33, 1216-1220 (1974)

[3] S. Picraux, F. Vook, Ion Implantation in Semiconductors, 355-360 (1975)

*: The figure can be found in the electronic copy of the abstract.

Tungsten (W) is a candidate material as a plasma facing component (PFC) in the divertor region of fusion tokamak reactors. Tungsten possesses excellent thermomechanical properties which make it suitable as a PFC including its high melting point (3422°C), low coefficient of linear thermal expansion (4.5C^-1), and high modulus of elasticity (400GPa). However, helium irradiation can alter the thermomechanical properties of tungsten resulting in embrittlement or reduction in thermal conductivity.[1] The addition of transition metal (Ta, Ti, Zr) carbide dispersoids in the range of 1-10wt% via the field-assisted sintering technique (FAST) has been shown to increase strength and ductility by increasing grain-boundary cohesion, suppress grain growth through Zenner-pinning of the carbide molecules, and increase the hardness of tungsten with no detrimental effects to hydrogen retention.[2-3] It has been observed that continuous exposure to helium plasma results in bubble formation in both ITER Grade and DS-W alloys. While helium bubbles are found in W-rich regions of the DS-W matrix, growth is suppressed within TaC regions and at the W-TaC interface. The growth of helium bubbles is not suppressed by TiC or ZrC additions.[4] Despite the promise of dispersoid-strengthened tungsten alloys (DS-W), more work is necessary to develop a comprehensive understanding of the behavior of DS-W alloys a fusion PFC. In this study, samples of pure W, ITER Grade W (IGW), and W-5TaC(wt%) fabricated by FAST are irradiated with 200eV He and 200eV Ar to facilitate the formation of W vacancies. Control samples are irradiated with 200eV He alone. The samples are analyzed via transmission electron microscopy (TEM) to investigate the effects of transition metal carbide additions on helium bubble formation.

Material issues pose a significant challenge in the design of future fusion reactors [1]. To be considered for fusion applications [2], highly integrated components are needed. Resilience against neutrons, good power exhaust, and oxidation resistance during accidental air ingress are design relevant issues while considering new material candidates. Neutron-induced effects, e.g., transmutation adding to embrittlement, retention, and changes to thermomechanical and thermo-physical properties, are crucial to material performance.

Here, the recent progress within 2019-2022 in fusion materials development for current and future fusion devices, with activities focusing on industrial upscaling and component design, is summarized. Ideally standard production technologies are used for advanced materials. However, the development of advanced materials with improved operational performance often requires the development of new methods and incorporation of (new) industrial partners to move advanced fusion materials into applications. This work is a continuation and extension of the work given by Coenen et al. [3].

In particular the industrial upscaling of self-passivating tungsten materials has taken big strides establishing the industrial production of powder together with partners at ZOZ and production of large scale samples with our international and industrial partners via the route of field-assisted sintering technology using industrial equipment (Dr. Fritsch).

On a similar path, the production of W_f/W composites has evolved to a point that mock-up production has been performed and is still ongoing following typical mono-block and flat tile designs including collaborations with Tokamak Energy and HFUT. The material has shown resilience against neutron embrittlement, however it is apparent from initial tests that joining of W/Cu and W/steel is the main challenge. The work on Yttria for the use as interfaces in W_f/W has progressed to a semi continuous process to coat tungsten weaves based on as alkoxide precursors together with ATL. Oxide ceramics such as Yttria are considered also as permeation barriers.

With additive manufacturing of W and heat sink materials, another production technique which has become more and more relevant in recent years, is under development at FZJ and its partners.

[1] Linke et al, Matter and Radiation at Extremes, AIP Publishing, 2019, 4, 056201

[2] Litnovsky et al, Encyclopedia of Nuclear Energy, Elsevier, 2021, 594-619

[3] Coenen et al, , Advanced Engineering Materials, Wiley, 22, 6 2020

We present a global sensitivity analysis framework for a coupled multi-physics model used to predict the surface response of a tungsten plasma-facing component. The multi-physics model combines a cluster dynamics simulator, used to predict the growth and evolution of subsurface helium gas bubbles, with a particle dynamics code, used to predict the helium implantation profile. In order to keep the sensitivity analysis computationally tractable, we construct a sparse, high-dimensional polynomial chaos expansion surrogate model that allows the extraction of Sobol sensitivity indices directly from the coefficients in the expansion. We use our framework in two experimental settings: ITER-like experimental conditions, and the experimental conditions of the PISCES-A linear plasma device. In both cases, we examine the effect of the model parameters on the predicted helium retention, and discuss the relevance of our results. Finally, we investigate the use of multifidelity methods to further improve the accuracy of our sensitivity results.

In future thermonuclear fusion reactors, like ITER and DEMO, plasma-facing components with tungsten armor material will have to sustain high thermal fluxes (10 MW/m² in steady state and 20 MW/m² in quasi-steady-state). For such extreme conditions, tungsten may reach temperatures higher than 1000 °C and, consequently can be prone to recrystallize, which can limit the lifetime of the divertor targets under cycling thermal loadings. Whereas recrystallization increases the material ductility, in steady-state and quasi-steady state, cracks are more prone to appear in the softened material. According to this analysis, accumulation of fatigue threshold shall be minimized to enhance the lifetime of the divertor targets made in tungsten.

A new method is proposed to perform annealing using a limited number of samples and heating conditions. The basic idea is to induce a steady-state temperature gradient in a tungsten rod by heating one side of the rod using a laser heating system. Annealing is split into two parts: a first step consists to heat the tungsten to obtain the desired temperature profile, then the temperature at each position along the bar are kept constant for a given time. During annealing, laser power was regulated thanks to a pyrometer associated with a PID system allowing to keep the temperature constant at 5 mm from the laser spot. A second pyrometer was used to scan the temperature along the rod to verify that the temperature are kept constant. Temperature recorded during annealing are consistent with the temperature profile obtained by simulation on a finite element model software (Comsol) using Tolias recommendation [1]. The uncertainties and errors associated with the temperature measurement have been quantified. The impact of the heating rate and the use of the thermal gradient on the softening of tungsten have been assessed after annealing by hardness measurements.

[1] P. Tolias, « Analytical expressions for thermophysical properties of solid and liquid tungsten relevant for fusion applications », Nuclear Materials and Energy, vol. 13, p. 42‑57, déc. 2017, doi: 10.1016/j.nme.2017.08.002.

The performance and resilience of the elements in unavoidable contact with the fusion plasma remains as one of the main issues for the development of magnetic fusion energy. The most critical part is the divertor, where the plasma exhaust is directed. ITER divertor armor consist on tungsten (with perpendicular-to-surface grains) monoblocks over a CuCrZr heat sink. These monoblocks have to allow plasma operation until its scheduled replacement in about 10 years. However, there are concerns about the resilience of these monoblocks on a real reactor plasma environment as ITER with high heat and particle fluxes. In order to test this resilience WEST tokamak was build with ITER-like monoblocks at the divertor, and with plasma configuration close to what is expected in ITER.

Those concerns have shown to be true on a critical part of the monoblocks, leading edges. In the 2018 WEST campaign poloidally-distributed cracks running perpendicular to the cooling tube axis were found on leading edges. This cracking was caused by brittle fracture after a series of disruptions at the start of the campaign [1]. Recent studies have calculated that just two typical WEST disruptions of 600 MW/m² can lead to the crack initiation [2]. But if those disruptions impact on a material which has been exposed to quasi steady state heating (from 45 to 70 MW/m²) then much less damage may be expected.

OLMAT High Heat Flux (HHF) device [3-4] has recently been upgraded with a high-energy (90J) CW laser. The laser characteristics allow heat loads simulation from steady state and slow transient simulation in a wide area (10-70 MW/m² in a 33-4 mm²), to disruption-like events up to 6.5 GW/m² (600 MW/m² in an area of 5.5 mm²). This positions OLMAT as a unique device where the WEST results may be confirmed and understood. Leading edges of ITER-like W and other candidate materials like W reinforced by W fibers will be exposed at a different number of pulses and to three different conditions: sample as received; immediately after (quasi) steady state heat loads, and heated at 500 ºC; well above the Ductile-to-Brittle Transition Temperature (DBTT). In this way, the cracking threshold will be defined, and the damage process would be better understood in order to avoid this damage in ITER.

[1] J-P. Gunn et al, Nucl. Mat. Ener. 27 (2021) 100920 https://doi.org/10.1016/j.nme.2021.100920

[2] A. Durif, et al. Phys. Scr. 97 (2022) 074004. https://doi.org/10.1088/1402-4896/ac71dc

[3] D Alegre et al. J. Fus. Ener. 39 (2020) 411-420. https://doi.org/10.1007/s10894-020-00254-5

[4] F. L. Tabarés et al., Fus. Eng. Des. 187 (2023) 113373. .https://doi.org/10.1088/0741-3335/58/1/014014

*Corresponding author: tel.: +34 913462579, e-mail: daniel.alegre@ciemat.es (D. Alegre)

Using the object kinetic Monte Carlo code KSOME (Kinetic Simulations of Microstructural Evolution) [1], displacement damage accumulation in polycrystalline tungsten was investigated to study the effect of neutron spectrum. We compared the damage caused by neutrons corresponding to the High-flux Isotope Reactor (HFIR) and 14 MeV neutrons at 1025 K as a function of dose rate (≈ dpa/s), impurity (trap) concentration, and trapping/detrapping behavior of self-interstitial atom (SIA) clusters by impurities. To simulate damage accumulation, primary cascade defects obtained from molecular dynamics simulations were inserted into the simulation domain according to the primary-knock-on atom (PKA) spectrum corresponding to each neutron spectrum. We find that damage accumulation, as a function of the parameters studied, is similar for both neutron spectra. When SIA clusters are permanently trapped, the damage increases with increasing trap concentration and is always greater than when SIA clusters are allowed to detrap. In the latter case, as the trap concentration increases, the damage initially increases, reaches a maximum and then decreases. However, there are key differences. The main difference is that the accumulated damage at a given dose is always lower for HFIR. For example, at a dose of 1.0 dpa and the highest trap concentration simulated, the density of vacancies and vacancy clusters is ≈ 2 and 1.5 times lower, respectively, in the case of HFIR

Since impurities can reduce the energy barrier for 1D diffusing SIA clusters to change their Burgers vector direction upon detrapping, we also investigated the impact of the change in the dimensionality of SIA cluster diffusion on the damage accumulation, which otherwise SIA clusters would only diffuse 1D if their size is larger than 5. After detrapping, SIA clusters can diffuse in a random 1D direction, resulting in net-3D diffusion; otherwise, they undergo a confined 1D diffusion between traps if they retain their original 1D direction. Our previous study [2] has shown that for 14 MeV, the damage is lower for the confined 1D diffusion case due to the increased probability of intracascade recombination. The difference between the two cases of SIA diffusion is sensitive to the dose rate. In the present study, we find that it is also sensitive to the PKA spectrum. We will present a detailed discussion about the comparison of damage accumulation between the two PKA spectra in relation to differences in defect production rates and trap concentration.

[1]. G. Nandipati, Kinetic Simulations of Microstructural Evolution [Computer Software] (Version: 1.0) Repo: https://doi.org/10.11578/Dc.20200826.3, Docs: https://relaxednightingale-067fd1.netlify.app (2020)

[2]. G. Nandipati et al., J. Nucl. Mater 542 (2020) 152402

*Corresponding author: tel.: +1 509 3752795, e-mail: giridhar.nandipati@pnnl.gov (G. Nandipati)

Tungsten (W) has been chosen as the divertor plasma-facing-components (PFCs) of ITER and fusion devices due to many advantages such as a high melting point and low sputtering yield. Nevertheless, the temperature of PFC will be higher and higher to achieve nuclear fusion aims. For this reason, non-contact temperature measurement diagnostics for the divertor and the first metallic wall are required to ensure safety and machine protection. Since infrared (IR) thermography, in which temperature can be determined from light emitted in the IR range, is a convenient technique for measuring surface temperature, many fusion devices are introducing IR thermography [1-4]. Emissivity, defined as the ratio between the radiation emitted by an object and the radiation emitted by a blackbody at the same temperature, is one of the primary parameters in IR thermography measurement. However, emissivity is related to various factors such as temperature, wavelength, and surface state. The surface state during plasma exposure will be modified through plasma-surface interaction leading to the change of surface morphology as erosion/deposition and blistering, thus, we will focus on the effect of surface morphology on emissivity. In this study, the spectral range of IR camera (FLIR A400) ranges from 7.5 um to 14um, and the bulk W sample with the dimension of 5 x 8 x 5.5 mm³ is utilized. In addition, the sample surface was mirror-liked polishing. Emissivity of bulk W samples exposed to deuterium (D) plasma by using applied-field magnetoplasmadynamic (AF-MPD) thruster concept [5] is investigated. The main plasma exposure conditions are as follows: incident flux at 6 x 10²² D+ m^-2s^-1 , fluence from 3.5 x 10²⁵ D m^-2 to 1.5 x 10²⁶ D m^-2 , and incident energy at 100 eV D^-1 . The pristine W emissivity was around 0.1 - 0.3, and this value was in good agreement with a previous study reported by J. Gaspar et al. [3]. Finally, we will report results from the change of W emissivity due to surface morphology after deuterium plasma exposure. This result can provide information on the surface temperature of PFCs when surface morphology changes during divertor operating conditions.

[1] M. Takeuchi et al., Plasma Fusion Res. 8 2402147 (2013).

[2] S. Shu et al., Infrared Phys. Technol. 98(2019) 1-6

[3] J. Gaspar et al., Nucl. Mater. Energy 25 (2020) 100851

[4] M.-H. Aumeunier et al., Nucl. Mater. Energy Volum 26, (2021) 100879

[5] K.B.Chai et al., Plasma Phys, Control. Fusion 63 (2021) 125020 (9pp)

We report the derivation of a relatively simple, yet physically motivated, empirical model to describe the growth dynamics of a helium bubble near a tungsten (W) surface, from nucleation until bursting, based on molecular dynamics (MD) simulations. During growth from continued helium absorption, a sub-surface bubble successively emits tungsten self-interstitial atom clusters (loop punching) with a center-of-mass motion toward the surface until it bursts. Thus, the model consists of loop-punching and bursting regimes. The loop-punching part parameterizes: 1) the critical ratio of the number of helium atoms to vacancies in the bubble (N_He/N_V) or pressure, 2) the increase in the number of vacancies (size of the emitted loop), and 3) the shift of bubble position toward the surface associated with each loop-punching event as a function of bubble size (N_V). The bursting model describes the critical N_He/N_V ratio (or pressure) and distance from the surface to initiate a bursting event, as a function of bubble size.

The bulk of the MD data have been obtained at 933 K. To incorporate temperature dependence into the model, targeted MD simulations have been performed at 500, 1500, 2000, and 2500 K. These data are then used to fit the temperature-dependent parameters in the loop-punching and bursting models. We will describe the model for W(100), (110), and (111) surface orientations, and discuss differences in the growth dynamics. Additional analyses will be presented in terms of the spatial domain of the burst hole, probability of partial bursting (where the burst hole reseals before the bubble is completely empty), and tungsten ligament thickness immediately before bursting. Results of this study are intended to inform mesoscale simulations of tungsten surface modification under helium implantation.

Building upon prior work [1] that assessed whether helium self-clustering was dominated by self-clustering or vacancy trapping, and recent experimental work [2], where the helium retention dependence on flux is studied in clean single crystal tungsten at fluxes varying in the range of 10¹⁶−10¹⁷ m⁻² s⁻¹, we used our continuum modeling tool Xolotl [3] to evaluate the helium clustering kinetics across a wide range of helium implantation fluxes. Xolotl is a spatially-dependent reaction-diffusion cluster dynamics code to simulate the divertor surface response to fusion relevant plasma exposure. The thermodynamics and kinetic parameters describing helium diffusion and clustering are based on values obtained from atomistic simulation results. In a first step, we evaluated the flux threshold under which helium self-trapping does not occur. Then, we simulated the full cycle of He ion irradiation followed by temperature programmed desorption that was performed in the experimental work of Dunand et al. [2] to study the effect of ion energy and flux on the resulting He retention and thermal desorption spectra. Finally, we added an initial concentration of pre-existing defects to understand the differences in retention between results obtained in clean single crystal versus polycrystal tungsten [4]. These benchmarks helped us refine the helium thermodynamic and kinetic parameters used in our continuum model.

@font-face {font-family:"MS Mincho"; panose-1:2 2 6 9 4 2 5 8 3 4; mso-font-alt:"ＭＳ 明朝"; mso-font-charset:128; mso-generic-font-family:modern; mso-font-pitch:fixed; mso-font-signature:-536870145 1791491579 134217746 0 131231 0;}@font-face {font-family:"Cambria Math"; panose-1:2 4 5 3 5 4 6 3 2 4; mso-font-charset:0; mso-generic-font-family:roman; mso-font-pitch:variable; mso-font-signature:-536870145 1107305727 0 0 415 0;}@font-face {font-family:"@MS Mincho"; panose-1:2 2 6 9 4 2 5 8 3 4; mso-font-charset:128; mso-generic-font-family:modern; mso-font-pitch:fixed; mso-font-signature:-536870145 1791491579 134217746 0 131231 0;}p.MsoNormal, li.MsoNormal, div.MsoNormal {mso-style-unhide:no; mso-style-qformat:yes; mso-style-parent:""; margin:0in; mso-pagination:widow-orphan; mso-hyphenate:none; font-size:12.0pt; font-family:"Arial",sans-serif; mso-fareast-font-family:"MS Mincho"; mso-font-kerning:1.0pt; mso-ansi-language:DE; mso-fareast-language:ZH-CN;}p.MsoBodyText, li.MsoBodyText, div.MsoBodyText {mso-style-unhide:no; mso-style-link:"Body Text Char"; margin-top:0in; margin-right:0in; margin-bottom:6.0pt; margin-left:0in; text-align:justify; mso-pagination:widow-orphan; mso-hyphenate:none; font-size:12.0pt; font-family:"Arial",sans-serif; mso-fareast-font-family:"MS Mincho"; mso-font-kerning:1.0pt; mso-ansi-language:DE; mso-fareast-language:ZH-CN;}span.BodyTextChar {mso-style-name:"Body Text Char"; mso-style-unhide:no; mso-style-locked:yes; mso-style-link:"Body Text"; mso-ansi-font-size:12.0pt; mso-bidi-font-size:12.0pt; font-family:"Arial",sans-serif; mso-ascii-font-family:Arial; mso-fareast-font-family:"MS Mincho"; mso-hansi-font-family:Arial; mso-bidi-font-family:Arial; mso-font-kerning:1.0pt; mso-ansi-language:DE; mso-fareast-language:ZH-CN;}.MsoChpDefault {mso-style-type:export-only; mso-default-props:yes; font-size:10.0pt; mso-ansi-font-size:10.0pt; mso-bidi-font-size:10.0pt;}div.WordSection1 {page:WordSection1;}

[1] Z. Yang, S. Blondel, K.D. Hammond, et al., Fusion Science & Tech. 71 (2017)

[2] A. Dunand, M. Minissale, T. Angot, et al., Nucl. Mater. Energy 34 (2023)

[3] https://www.osti.gov/doecode/biblio/63304

[4] Y. Gasparyan, S. Ryabtsev, V. Efimov, et al., Phys. Scr. T171 (2020)

*Corresponding author: e-mail: bdwirth@utk.edu (B.D. Wirth)

@font-face {font-family:"MS Mincho"; panose-1:2 2 6 9 4 2 5 8 3 4; mso-font-alt:"ＭＳ 明朝"; mso-font-charset:128; mso-generic-font-family:modern; mso-font-pitch:fixed; mso-font-signature:-536870145 1791491579 134217746 0 131231 0;}@font-face {font-family:"Cambria Math"; panose-1:2 4 5 3 5 4 6 3 2 4; mso-font-charset:0; mso-generic-font-family:roman; mso-font-pitch:variable; mso-font-signature:-536870145 1107305727 0 0 415 0;}@font-face {font-family:"@MS Mincho"; panose-1:2 2 6 9 4 2 5 8 3 4; mso-font-charset:128; mso-generic-font-family:modern; mso-font-pitch:fixed; mso-font-signature:-536870145 1791491579 134217746 0 131231 0;}p.MsoNormal, li.MsoNormal, div.MsoNormal {mso-style-unhide:no; mso-style-qformat:yes; mso-style-parent:""; margin:0in; mso-pagination:widow-orphan; mso-hyphenate:none; font-size:12.0pt; font-family:"Arial",sans-serif; mso-fareast-font-family:"MS Mincho"; mso-font-kerning:1.0pt; mso-ansi-language:DE; mso-fareast-language:ZH-CN;}p.MsoBodyText, li.MsoBodyText, div.MsoBodyText {mso-style-unhide:no; mso-style-link:"Body Text Char"; margin-top:0in; margin-right:0in; margin-bottom:6.0pt; margin-left:0in; text-align:justify; mso-pagination:widow-orphan; mso-hyphenate:none; font-size:12.0pt; font-family:"Arial",sans-serif; mso-fareast-font-family:"MS Mincho"; mso-font-kerning:1.0pt; mso-ansi-language:DE; mso-fareast-language:ZH-CN;}a:link, span.MsoHyperlink {mso-style-unhide:no; mso-style-parent:""; color:blue; text-decoration:underline; text-underline:single;}a:visited, span.MsoHyperlinkFollowed {mso-style-noshow:yes; mso-style-priority:99; color:#954F72; mso-themecolor:followedhyperlink; text-decoration:underline; text-underline:single;}span.BodyTextChar {mso-style-name:"Body Text Char"; mso-style-unhide:no; mso-style-locked:yes; mso-style-link:"Body Text"; mso-ansi-font-size:12.0pt; mso-bidi-font-size:12.0pt; font-family:"Arial",sans-serif; mso-ascii-font-family:Arial; mso-fareast-font-family:"MS Mincho"; mso-hansi-font-family:Arial; mso-bidi-font-family:Arial; mso-font-kerning:1.0pt; mso-ansi-language:DE; mso-fareast-language:ZH-CN;}.MsoChpDefault {mso-style-type:export-only; mso-default-props:yes; font-size:10.0pt; mso-ansi-font-size:10.0pt; mso-bidi-font-size:10.0pt;}div.WordSection1 {page:WordSection1;}

The EAST tokamak aims to achieve long-pulse and high-performance plasmas under steady-state conditions. In recent EAST campaigns, the damage on the main limiter has become the main limiting factor on the operational goals, although active water cooling is used. Before 2022 campaign, the main limiter was made of graphite tiles fixed by bolts. Due to the lower heat transfer coefficient between the graphite tiles and the heat sink (around 0.96 KW/(m²·K)) and the poorer heat conduction, the graphite tile was broken and fell off the limiter. In order to prolong the lifetime of the limiter, the W/Cu flat-type limiter was used during 2022 campaign. However, it was melted seriously and caused many disruptions [1]. Observable misalignment was found at the location of the damage, which may be related to the thermal expansion. To enhance the power handling ability, a new main limiter consisting of W/Cu monoblocks with special shaping to reduce incident angle and avoid leading edges has been implemented. Although hot spots still exist, the monoblocks show longer lifetime during long-pulse operation.

A high-resolution infrared camera with a spatial resolution of 1.3 mm [2] was used to obtain the temperature distribution on the main limiter surface. The heat flux pattern has been successfully reproduced by using the PFC-Flux code and ANSYS simulations. For the discharges heated by LHW, ECRH and ICRF, the vertical heat flux on the main limiter can reach 9.82 MW/m², while the vertical heat flux was reduced to less than 1.71 MW/m² without ICRF heating. The temperature on the main limiter was significantly increased, when the ICRF power was higher than 0.6MW. The hot spots were normally observed on the ion side of the main limiter. The temperature of the hot spot increased with increasing of ICRF heating power, and decreased with higher plasma density and larger gap between plasma and the main limiter. Similar phenomena were also observed in the other devices, which indicated the hot spot may be mainly caused by the loss of fast ions [3].

[1] D.H. Zhu, C.J. Li, B.F. Gao, et al., Nucl. Fusion 62, 056004 (2022)

[2] M.W. Chen, X.F. Yang, X.Z. Gong, et al., Fusion Eng. Des. 150, 111415 (2020)

[3] Y. Ikeda, K. Tobita, K. Hamamatsu, et al., Nucl. Fusion 36, 759 (1996)

In this study, we use nonequilibrium molecular-dynamics (NEMD) simulations to analyze the transport of helium [1], intrinsic point defects and hydrogen in the presence of a thermal gradient in tungsten. For all the species examined, we find that the resulting species flux is directed opposite to the heat flux, indicating that species transport is governed by a Soret effect, namely, thermal-gradient-driven diffusion, characterized by a negative heat of transport that drives species transport uphill, i.e., from the cooler to the hot regions of the tungsten sample. The findings of our thermal and species transport analysis have been implemented in our cluster-dynamics code, Xolotl [2], which has calculated the temperature and species profiles over spatiotemporal scales representative of plasma-facing tungsten under typical reactor operating conditions, including extreme heat loads at the plasma-facing surface characteristic of plasma instabilities that induce edge-localized modes (ELMs). We demonstrate that the steady-state species profiles obtained accounting for the Soret effect vary significantly from those where temperature-gradient-driven transport is not accounted for and discuss the implications of such a Soret effect for the response to plasma exposure of plasma-facing tungsten. Most importantly, we find that accounting for the Soret effect in the species transport computations significantly reduces the retention of helium and hydrogen in PFC tungsten. Future work will expand these initial studies to include the effect of species clustering.

[1] E. Martinez, N. Mathew, D. Perez, et al., J. Appl. Phys. 130, (2021)

[2] https://www.osti.gov/doecode/biblio/63304

Tokamak limiter plays a variety of roles during its operation. Itserves primarily to protect the wall from the plasma when there are disruptions, runaway electrons, or other instabilities. EAST upgraded its main limiter from CFC into active water cooling flat-typeW/Cu limiter before 2022 EAST plasma experimental campaign,aiming for ahighthermal exhaustcapacity. However,severe W melting phenomena were observedduring plasma operations.The meltingfeaturesandits influence on plasma dischargewerecharacterized, and how the melting occurredwas also analyzed.

The W/Cu flat-type main limiter showed a good performance when the 2022 EAST spring experimental campaign first started. Then a hot spot was observed during plasma current plateu period after a periodof service. Aftermelting event of W/Cuflat-type main limiter was first observed with a CCD camera located at C sector withobvious dropletsejectionfrom limiter surface, high performanceplasma can still be obtained by adjusting the parameters. With time went by, however, melting events of main limiter were harder to avoid even after optimizing the plasma parameters and disruption due to high concentration of W impurity became frequent. Hence, the experimental campaign had to suspend for a while to replace the main limiter. Post-mortem inspection of W melting was thus performed after the replacementofW/Cu flat-type main limiter.Based on the damage process observed by CCD and post-mortem inspection, it is believed that the melting mechanism of main limitermay be like that of W/Cu flat-type lower divertor used in EAST. the joint interface would crack first because of the mismatch of thermal and mechanical properties between W and Cu. Such damage can lead to the degradation of heat transferring between Wplate and heat sink material. Thus, during the next heating and cooling process, not only the maximum temperature of main limiter would increase but also the time it took to fully cool down would be longer. And the temperature of the joint surface betweenW and Cu would no doubt increase leading to a higher thermal stress which would cause the further propagation of cracks. Such process would repeat continuously during the cyclic plasma discharge. As a result, the maximum temperature of main limiter would increase and eventually exceed the melting point of W andW melting and melting events would occur.Once melting events occurred on the W/Cu flat-type main limiter, the vast majority of cases would result in plasma disruption. And the damage of main limiter would rapidly deteriorate leading to failure of the structure.As a potentialplasma facing component, the fatigue life of flat-type structureneeds to be improvedso is the joint performance between W and Cu jointby optimization of the fabrication parameters.

Theresultsprovide important referenceforthe application and improvement of the W/Cu plasma facing componentsforhighheat flux area in fusion devices.

Microstructural evolution and hardening in lightweight multi-principal element titanium-based alloy under Ar ion irradiation

G. Tolstolutska*, M. Tikhonovsky, A. Velikodnyi, S. Karpov, V. Ruzhytskyi,

G. Tolmachova, R. Vasilenko, A. Levenets

National Science Centre ‘Kharkiv Institute of Physics and Technology’, Kharkiv, Ukraine

Nuclear fusion power reactors require improvements in high-temperature and plasma-facing materials to prevent degradation due to the neutron flux produced during deuterium and tritium fusion reactions.

Among new prospective materials multi-principal element alloys (MPEA) have attracted considerable attention in recent years due to their excellent corrosion and irradiation resistance as well as their good mechanical properties over a wide temperature range [1]. One of the most promising applications of MPEA is related to the production of high-temperature structural materials. The first developed MPEAs based on refractory metals (refractory high entropy alloys, RHEAs) demonstrated high strength up to 1600 °C, but were very heavy. Alternative MPEA alloys based on the Al-Cr-Nb-Ti-V system were created to reduce the density.

In this report, we present the results of the study of new lightweight multi-principal element titanium-based alloys 61Ti-10Cr-7Al-22V, 61Ti-10Cr-7Al-11V-11Nb and 61Ti-10Cr-7Al-22Nb (at.%) with high ductility at room and elevated temperatures. The strength characteristics of the alloys at 650°C were found to be significantly higher than industrial titanium alloys and reactor steels.

These single-phase bcc alloys were irradiated with 1.4 MeV Ar ions at room temperature and mid-range doses from 0.1 to 10 displacements per atom. Transmission electron microscopy (TEM), scanning transmission electron microscopy with energy dispersive X-ray spectrometry (STEM/EDS) and X-ray diffraction (XRD) were used to characterize the radiation defects and microstructural changes. Nanoindentation was used to measure the ion irradiation effect on hardening. In order to understand the irradiation effects in MPEA and to demonstrate their potential advantages, a comparison was performed with irradiation-induced hardening behaviour of 316 austenitic stainless steel and T91 ferritic-martensitic steel irradiated under an identical condition. It was shown that hardness increases with irradiation dose for all the materials studied, but this increase is lower in multi-principal element alloys than in stainless conventional steel.

[1] Z. Cheng, J. Sun, X. Gao et al., J. Alloy. Compd., 930 (2023) 166768

*Corresponding author: tel.: +38 (050) 109-39-16, e-mail: g.d.t@kipt.kharkov.ua (G. Tolstolutska)

The heat and hydrogenic exposure conditions on the top surface of the ITER tungsten divertor monoblocks may lead to blister formation [1] as the result of the formation and growth of hydrogen nanobubbles which may eventually lead to the development of a macroscopic crack. Simulating the nanovoid creation is thus the first stage needed for a global blistering model. These nanovoids can be formed by vacancy and vacancy cluster interactions. Several authors have proposed modelling schemes for these interactions: Ebihara et al. [2] put forward a general 1-D cluster model for iron without a thermal gradient; Oude Vrielink et al. [3] use a 3-D vacancy cluster model for their multi-scale fracture probability analysis of tungsten monoblocks, neglecting the diffusion of vacancies.

In this work, we present a general 3-D model of vacancy diffusion and clustering (up to 9 vacancies) in a non-uniform thermal field. Two sources of mono-vacancies are included in the model: (a) in the bulk, following Ebihara et al. [2], mono-vacancies relax to a temperature-dependent equilibrium value, which results from the balance between creation and annihilation [4]; (b) at the surface, mono-vacancies are created during rapid pulses of particles and heat expected during Edge Localised Mode (ELM) activity.

This model is applied to a tungsten ITER divertor monoblock submitted to a stationary heat flux of 10 MW/m² at the plasma-facing surface. Vacancy clusters are seen to align along specific isotherms, dependent on cluster size, characteristic of a thermally activated process. For instance, the divacancy concentration is localised on the 530 K isotherm, a few millimetres under the surface. Below 450 K, closer to the monoblock coolant channel, there is no clustering nor creation of new vacancies and so the monovacancy concentration remains equal to the initial one. Above 450 K, the monovacancy concentration is lower because of the competition between the divacancy creation and the annihilation/creation process linked to the thermophysical equilibrium. The trivacancy is found to be unstable, resulting in a very low overall concentration. Near the plasma-facing surface, the vacancy clusters consist predominantly of 9 vacancies, the maximum size considered in the model. To mimic the effect of the presence of hydrogen on the 3-D cluster distribution, comparative additional simulation cases are performed without monovacancy annihilation or/and diffusion.

[1] G De Temmerman et al., Plasma Phys. Control. Fusion 60, 044018 (2018)
[2] Ki Ebihara, et al., Metall Mater Trans A 52, 257–269 (2021)
[3] M A Oude Vrielink, et al., Nucl. Mater. Energy 28, 101032 (2021)
[4] Y Kraftmakher, Phys. Rep. 299, 79-188 (1998)

In fusion devices, the W/Cu monoblocks for divertor targets are subjected to both steady state heat load and extreme high pulsed heat flux, in which the transient heat flux during disruption up to thousands of MWm^-2 is expected to induce the near suface cracking even melting of W/Cu monoblocks according to the performed high heat tests experiments using electron or ions beams. The performance of such pre-damaged W/Cu monoblocks by accident transient events under steady state heat loading in subsequent normal plasma discharges is unclear so far, which is one of the main concerns for ITER operating with W divertor. Thus, the pre-damaged W/Cu monoblocks actively induced by high transient flux test using electron beams has been installed and exposed in the edge plasma of WEST, which will provide a dedicated results.

EAST, as an ITER-like device, has been installed a lots of W/Cu monoblocks for both divertor target and limiters. Before the winter plasma campaign in 2022, a new limiter mainly composed of W/Cu monoblocks was developed and installed, aiming to have a high heat exhaust capacity and a long lifetime during long pulse and high power operations. However, during the early plasma building phase without obvious auxiliary heating power, a surprised phenomenon was observed by CCD camera that a large amount runaway electrons were quickly impacted on a local surface of the limiter, resulting to the severe spraying of droplets and melting of W/Cu monoblocks, which had been confirmed by morphology inspection. The pre-melted limiter was not replaced and continued to service. Such just provided a convenience to qualitatively discuss the performance of damaged W/Cu monoblocks under normal steady state heat loading in subsequent plasma discharges.

In the following plasma discharges, the surface state of the limiter was always monitored by both CCD and IR cameras. The projected steady state heat load on limiter surface is roughly calculated up to several WMm^-2, which is just comparable with that on divertor surface. After thousands of normal plasma discharges with multiple heating powers by ICRF, ECRH, LHCD and NBI, in spite some local hot spots still existed during some plasma scenarios, no droplet was ejected from the limiter surface according to the CCD observation. Moreover, the measured temperature by IR camera can always reach to the balance state and the maximum temperature is generally within 1000 ^oC. All of those illustrate that the heat exhaust function of such pre-damaged W/Cu monoblocks induced by pulsed heat flux during disruption doesn’t degrade significantly. They may, to some extent, continue to service under steady state heat load in following normal plasma discharges. Such results provides key experiential references to the ITER and other devices.

In last December (2022) the National Ignition Facility (NIF) announced ignition with gain. A major challenge that needs to be confronted to build nuclear fusion power plants, based on the inertial confinement approach, is the development of new materials; in particular this contribution focuses on plasma facing materials (PFMs), able to withstand the combined effects of large thermal loads and radiation environments taking place in this kind of reactors.

Due to its properties, W is a prominent PFM candidate. However, previous work indicates that W has important drawbacks, such as its high ductile-to-brittle transition temperature, low yield strength, oxidation at elevated temperatures into WO₃ and recrystallization well below its melting temperature. Moreover, it has a very detrimental ability to retain light species easily, mainly hydrogen and helium, which leads, among other fatal effects, to surface blistering, cracking, and exfoliation.

In this contribution, we firstly will show the radiation environments that the PFMs would face in inertial fusion reactors emphasizing the similarities and differences with those taking place in magnetic fusion reactors. Secondly, we will discuss about the capabilities and limitations of existing experimental facilities to mimic that radiation environment. Finally, we will review some of the different approaches that are being investigated (nanostructured tungsten, engineered W surfaces, W needles and W foams), mainly focussing on the advantages and drawbacks of nanostructured W to act as PFMs in inertial fusion reactors operating in the direct drive configuration [1-6].

[1] D. Fernández-Pello, M.A. Cerdeira, J. Suárez-Recio, et al., J. Nucl. Mater., 560, 153481 (2022).

[2] R. Gonzalez-Arrabal, A. Rivera, J.M. Perlado, Matter and Radiat. at Extremes, 5, 055201 (2020).

[3] M. Panizo-Laiz, P. Díaz-Rodríguez, A. Rivera, et al., Nucl. Fusion 59, 086055 (2019)

[4] G. Valles, M. Panizo-Laiz, C. González et al., Acta Materialia, 122, 277–286 (2017).

[5] C. González, M. Panizo-Laiz, N. Gordillo, et al., Nucl. Fusion, 55, 113009 (2015).

[6] R. Gonzalez-Arrabal, M. Panizo-Laiz, N. Gordillo et al., J. Nucl. Mater., 453, 287–295 (2014).

*Corresponding author: tel.: +34910677128, e-mail: raquel.gonzalez.arrabal@upm.es (R. González-Arrabal)

For ITER operation, divertor targets plasma-facing units (PFUs), which use tungsten as armor material, have to withstand heat fluxes up to 20 MW/m² during slow transients [1]. The performance of these PFUs was demonstrated throughout a qualification program [2]. Results from this qualification program highlighted that a macro-crack (so called self-castellation) might occur on the plasma facing surface after few tens up to few hundreds cycles at 20 MW/m² [3]. This crack propagates from the loaded surface in the direction of the cooling pipe. The PFU crack evolution (depth, width) and its potential influence on the plasma operation are proposed to be studied in the WEST tokamak, by implementing a specific PFU with self-castellated monoblocks in the WEST lower divertor.
The self-castellation experiment aims at (i) studying, via dedicated WEST plasma operation, the impact of the self-castellation on the plasma operation and (ii) characterizing after the WEST operation, the macro crack evolution. This paper presents the preparation of the self-castellation experiment. The rationale for the PFU positioning on the WEST lower divertor and the position of the self-castellated monoblocks on the PFU will be presented. It is proposed to generate the self-castellation via electron beam loading in the JUDITH 2 facility on at two different radial positions in the WEST lower divertor, corresponding to two different monoblocks widths (28 mm and 30.7 mm). Results from finite elements modelling using the T-REX tool [4], performed to optimize the experimental loading protocol in JUDITH 2, are presented. It is highlighted that the monoblock width plays a major role on the crack initiation, whatever the heat flux applied. Assuming thermal cycles at 20 MW/m² (10s on / 10s off) on the two selected geometries, a macro crack should initiate after the ~45th cycle for the narrowest monoblock according to modelling, whereas no macro crack is expected before the ~20th cycle for the widest one. Given as conservative values, the number of cycles to initiate the crack depends on the ductile to brittle transition temperature (DBTT) assumed in the modelling. For the studied tungsten grade, the DBTT is not precisely known (between 250°C – 500°C). Regarding the large DBTT range, a sensitivity analysis is performed. Based on the considerations above, the pre-exposed WEST PFU characterization and the loading protocol for generating the self-castellation in JUDITH 2 are described.
[1] R.A Pitts et al, JNM 438 (2013) S48-S56
[2] T Hirai et al 2014 Phys. Scr. 2014 014006
[3] S. Panayotis et al, NME 12 (2017) 200-204
[4] A. DURIF et al, JNM 569 (2022) 153906
*Corresponding author: tel.: +33 4 25 42 33 76, e-mail: alan.durif@cea.fr (A. Durif)

Tungsten (W), the main wall material candidate for the current fusion reactor, has been used in ITER under construction and is planned to be used in future devices such as Chinese Fusion Engineering Testing Reactor (CFETR). Chemical vapor deposition (CVD) technology has recently attracted much attention as it is the main way to produce thick tungsten coatings. CVD-W has characteristics of high density, high purity and high surface coverage with a <001> oriented columnar-grain microstructure. There have studies shown that CVD-W has the advantages of thermal conductivity, thermal shock and fatigue resistance. This work evaluates the performance of CVD-W under deuterium (D) plasma irradiation to investigate the surface blistering and D retention.

Four grades of W were used including CVD-W with large-size grains (CVD-L) and small-size grains (CVD-S), rolled W (ND-W), and recrystallized W (Rec-W) as the reference material, with an average grain size of ~30.4, 10.9, 8.2, and 34.4 μm, respectively. After surface polishing, samples were irradiated to D plasma in the linear plasma device LEPS with fluences of 5×10²⁴, 1×10²⁵, and 2×10²⁵ m^-2. The ion flux was about 1.0×10²¹ D m^-2s^-1 and specimen temperature were set at about 500 K. Surface morphology changes are observed using scanning electron microscope. Thermal desorption spectroscopy was used to measure D desorption and total D retention.

Both CVD-L and CVD-S show outstanding performance in resistance to blistering and D retention compared to ND-W and Rec-W. Few blisters at low fluence and several sub-micron-diameter blisters at higher fluences are observed in both types of CVD-W. In ND-W and Rec-W, a rich modified morphology composed of various types of dense blister is observed, and the density and the size of blisters increase with the increasing fluence. Total D retention in all grades is increased with the increasing fluence. The desorption spectra show that two distinct desorption peaks at ~500 K and 600-700 K. D retention in both types of CVD-W is an order of magnitude lower than those in Rec-W and ND-W, which is mainly attributed to the lower intensity of the peak at 600-700 K. It is therefore suggested that the microstructure of columnar grain in CVD-W is of great importance in suppressing the growth of the characteristic defects related to the desorption peaks.

In this work, both types of CVD-W have shown an outstanding resistance to D-induced blistering and D retention, which indicates potential advantages as a wall material in terms of surface integrity and tritium inventory. These results provide an experimental reference for the application of CVD-W as the wall material in fusion reactor. Further investigations are undergoing to understand the mechanism determining the irradiation effect in CVD-W.

*Corresponding author: tel.: +86 13811630078, e-mail: LCheng@buaa.edu.cn (L. Cheng)

In a fusion reactor, Plasma Facing Components (PFCs) will suffer severe thermal shock, properties and behaviors of PFCs under high heat flux (HHF) loads are a key issue to realize the long-term stable operation of the further reactor. This paper investigates the thermo-mechanical behaviors of tungsten (W) armor under different steady-state heat loads by the method of finite element modeling and simulating. The temperature distribution and corresponding thermal-stress changing rule under different HHF are analyzed and deduced. The Manson-Coffin equation is employed to evaluate the fatigue lifetime (times of cyclic HHF loading) of W-armored first wall (FW) under cyclic HHF load. The results are useful for the formulation design and structural optimization of tungsten-armored PFCs for the future demonstration fusion reactor (DEMO).

Controlled melting of aluminum (Al) was performed in the divertor of the DIII-D tokamak in order to benchmark ITER-relevant melt layer dynamics with the MEMENTO code [1]. Understanding the macroscopic motion and stability of liquid beryllium (Be) is important for prediction of the consequences of possible transient melting of the ITER Be first wall during disruptions. In our experiment we used Al as a proxy for Be. Three Al blocks were exposed in the DIII-D divertor under steady L-mode discharge conditions using the Divertor Material Evaluation System (DiMES) manipulator [2]. The blocks were sized 1×1 cm in the toroidal and poloidal directions, with the top surface angled at ~32 degrees towards the incident plasma heat fluxes. During the exposure, samples were imaged by visible and infra-red (IR) cameras. Grounding of the blocks was performed via nichrome wires with resistance of ~0.15 Ω, allowing for current measurements. Heat fluxes incident on the angled surfaces were inferred from IR data aided by SMITER field line tracing.

The goal of the experiment was to obtain shallow (few hundred microns thick) melt layers on the angled surfaces for comparison with MEMENTO modeling. This was achieved on two of the blocks, while the third over-melted. Molten Al ejection was observed from two of the blocks exposed in higher power discharges with neutral beam injection heating power P_NBI = 2.4 MW and incident heat fluxes of up to 20 MW/m². Such ejection was not observed on the block exposed at P_NBI = 0.9 MW and incident heat flux of ~10 MW/m². The measured current between the block and the machine ground was up to 25 A in the high power case and up to 5 A in the low power case, with the sign consistent with electron emission from the block in both cases. Post-exposure surface deformation profiles reveal melt displacement as well as melt ejection from the block edge. MEMENTO modeling of the results will be presented.

Work supported by the US DOE under DE-FC02-04ER54698, DE-FG02-07ER54917, DE-NA0003525, DE-AC52-07NA27344, DE-AC05-00OR22725, DE-SC0019256.

[1] S. Ratynskaia, K. Paschalidis, et al., Nucl. Mater. Energy 33 (2022) 101303

[2] C.P.C. Wong et al., J. Nucl. Mater., 196-198 (1992) 871

*Corresponding author: e-mail: rudakov@fusion.gat.com (D.L. Rudakov)

The plasma-surface interaction of lithium-prefilled 304 stainless steel printed multichannel capillary porous system (CPS) based target was studied in conditions simulating steady state and transient events produced by SCU-PSI plasma linear device. The laser printed CPS (capillary size ~120 μm) revealed the excellent ability to wick liquid lithium to its surface. Liquid lithium and lithium vapor could resist incident heat load and mitigate the target temperature. In comparison to traditional woven tungsten-based meshes target, the printed steel-based CPS target seems better capacity to dissipate radiation and withstand plasma load. After repetitive steady and pulsing plasma treatments, the printed CPS structure remains macroscopicly intact except mild embrittlement. The microcosmic surface morphology of CPS after plasma irradiation showed uneven hole-like structure, mainly corresponding to its corrosion by liquid lithium, meanwhile the surface temperature exceeds 700 ℃ during irradiation. The vapor cloud morphology in front of the sample changed with time, magnetic field, plasma parameters. Under stable plasma condition, lithium radiation dissipation region gradually shift to target surface with increased magnetic field and input power. However, under pulse condition (~1 ms duration each discharge), Li I spectrum close to the surface of target was annihilated and then rekindled after a 10 ms~40 ms window period. It is inconsistent with previous enhanced Li-I emission result (3D printed lithium-filled tungsten target by P. Rindt et al), which might be attributed to dispersion of lithium atoms by high plasma pressure. Besides, obvious particles ejection were observed from the surface of target as the heat load reach 0.9 MJ/m². The size and velocity of particles were monitored by high-speed camera and calculated as up to ＜300 μm in diameter and ＜50 m/s. It reveals that laser printed CPS target could be reliable to the irradiation of plasma. With enhancement of preparation accuracy and substitute of irradiation-resistant materials, capillarity and robustness could be further improved and be considered as promising liquid lithium divertor.

Hybrid liquid lithium – porous tungsten materials are being considered as candidate plasma facing components due to their ability to maintain a continuously replenishable low-z plasma interface while tolerating both the high steady-state and transient heat fluxes in the high-duty cycle environment of a fusion reactor [1]. Porous W samples are fabricated by Spark Plasma Sintering (SPS) and result in 70% dense substrates. Samples of different pore sizes (sub-micron, 1-10 micron, and 10-20 micron) are manufactured using different sized tungsten powders (800nm, 5μm, and 15μm) during manufacturing.

Studies on the dynamics of hydrogen in hybrid liquid Li- porous tungsten substrates are being performed at the Dynamics of ION Implantation and Sputtering of Surfaces (DIONISOS) experiment at MIT. Lithiated porous tungsten samples are irradiated with deuterium using a fluence of 10²⁴/m² at solid (room temperature) and liquid (200° C) states. Virgin porous tungsten samples (e.g. without Li) are used as controls (room temperature and T = 200°C) to compare D diffusion in porous W with and without Li. Furthermore, mirror-finished highly dense tungsten samples (eg theoretical density of 98%) with the same temperature and lithium parameters as the porous W samples are also used as controls. The dense tungsten control samples aim to investigate what the effects are of sample morphology on the D diffusivity. In-situ Elastic Recoil Detection (ERD) is then used to obtain information on the depth distribution of deuterium in the samples. After ERD measurements, In-situ mass spectrometry is performed using a Residual Gas Analyzer (RGA) to quantify the deuterium retained in the samples. Results will be discussed in the context of how deuterium is retained and diffusing through the samples, and how the porous morphology of the samples affects this dynamic.

Work supported by DOE Contract DE-SC0021119

[1] Nygren, R. E., and F. L. Tabarés Nucl. Mater. Energy (2016) 

*Corresponding author: e-mail: clopezperez@psu.edu (C. López Pérez)

Divertors for fusion reactors using liquid metal (LM) based plasma facing components (PFCs) are attractive as a robust solution to the heat exhaust challenge. To reach the reactor application stage, an at-scale demonstration of performance in medium-sized tokamaks is required. This is a large step from todays small scale prototypes and basic science research. To fill this gap a new dedicated Liquid Metal Shield laboratory (LiMeS-lab) is being created where mock-up scale prototypes can be designed, manufactured and tested under relevant loading conditions, providing a key stepping stone on the path to successful LM-based PFC deployment.

LiMeS-lab is an integrated facility that enables the development and testing of LM-based PFCs from start to finish. Capillary Porous Structures (CPSs), required to restrain the LM at the plasma surface against hydrodynamic forces, will be additively manufactured as part of the PFC mock-up using a new dedicated Laser Powder Bed Fusion device specialised for tungsten printing. This approach gives wide design freedom which has been shown to lead to a hundred-fold reduction in von Mises stress compared to unshaped blocks [1].

Secondly wetting of tungsten by tin typically requires vacuum heating at high temperatures which are not compatible with multi-material PFC designs. Dedicated studies using an atom source have shown to reduce this requirement to below 500 °C. A dedicated plasma-assisted wetting setup is therefore under development as part of the LiMeS-lab setup.

Following successful wetting of the PFC mock-up, high heat and plasma flux testing will be carried out in the linear plasma device LiMeS-PSI. This device will be capable of creating continuous high-flux heat and plasma loading at DEMO relevant conditions (n_e~10²⁰ m^-3, T_e~1 to 5 eV, Γ~10²⁴ m^-2 s^-1, q>20 MW m^-2). The linear plasma generator will be specially adapted for LM research and testing and will use high-temperature cooling circuits and liquid metal supply lines providing circulating, always molten liquid metal. This enables testing of flowing concepts as well as the replenishment of static CPS systems. Furthermore, the diagnostic set will be designed to be compatible with the liquid metal vapour to enable long-term testing without unacceptable optical degradation. Lastly a dedicated and specialised post-mortem Thermal Desorption Spectroscopy setup will enable the determination of fuel retention. Progress in the research, development and construction of LiMeS-lab will be presented.

[1] P. Rindt et al 2019 Nucl. Fusion 59 05400

Silicon coated film characteristics and its effect on plasma performance with full tungsten divertor were investigated to improve plasma performance in EAST.

Siliconization using 10% SiD₄ + 90% He assisted by ion cyclotron range of frequency discharge (ICRF) or glow discharge (GD) was performed in EAST. It was found that the coated Si film on W sample at a higher baking temperature (160 °C) was smoother than that at a lower baking temperature (60 °C) [1]. An increase in the ICRF working power from 20 kW to 40 kW further increased the Si content and the film thickness by 1.5 times. The cleaning efficiency of ICRF on the surface of the W sample before siliconization was higher at 40 kW than that at 20 kW, which facilitated the removal of oxides and other compounds from the surface of W. The ratio of silicon/oxygen (Si/O) and the thickness improved considerably due to greater ionization and deposition of SiD₄. Additionally, siliconization via GD was more uniform and 1.5 nm thicker than that via ICRF, which was because of the greater working gas pressure, coating duration, and homogeneity in GD. After Si coating, the performance of silicon coating in erosion mitigation of W material is also analyzed. the W substrates could be effectively protected by Si coating from plasma erosion [2]. These results confirm Si coating served as a sacrificial protective layer and could reduce tungsten surface erosion, which provides a new reference for protecting first wall materials in future fusion devices. Further, it was found siliconization had very excellent impurity and fuel controlling capability, but the H/(H+D) ratio decreased much more slowly and impurity radiation was slightly higher with siliconization than with Li coating. In addition, we also try to perform siliconization by Si powder/SiD₄ real time injection. It was noted that ELM mitigation was observed during SiD₄ injection, but it also produced strong fueling effect and high radiation in core plasma resulting in stored energy drop with Si/SiD₄ injection [3]. Next step, we will investigate Si real-time injection with higher plasma parameter conditions in EAST.

[1] Y.H. Guan, G.Z. Zuo, X.C.Meng, et al., Nuclear Materials and Energy 34(2023)101368

[2] Z.L. Tang, G.Z. Zuo, C.L. Li, et al., Journal of Nuclear Materials 555 (2021) 153146

[3] G.Z. Zuo, Particle control for long pulse plasma operation in EAST tokamak, AAPPS-DPP2022, October 9-14, 2022, topic plenary.

Studying the physics of the lithium vapour box in the linear plasma generator Magnum-PSI

F. Romano^1,*, V.F.B. Tanke^1,2, J. Gonzalez¹, J.A. Schwartz³, S. Brons¹, R.J. Goldston³, L. Romers¹, P. de Laat¹, T.W. Morgan^1,2

¹DIFFER–Dutch Institute for Fundamental Energy Research, Eindhoven, The Netherlands

²Eindhoven University of Technology, Eindhoven, The Netherlands

³Princeton Plasma Physics Laboratory, Princeton New Jersey 08543, USA

Among the challenges to have nuclear fusion energy ready, one of the most critical is the handling of the heat flux exhaust in the divertor region of a tokamak. An innovative approach involves the use of liquid metals (LMs) such as lithium as plasma facing component (PFC). Is it therefore possible to shape the divertor as a vapour box (VB) [1] to provide lithium to the plasma, which collides with and cools the plasma, while at the same time the shaping prevents all but a small fraction from reaching the plasma core, where it would dilute the fuel. Two main steps must be taken to test and validate this approach: evaluating the lithium cooling power and understanding the Li transport mechanisms to confine it within the VB. This results in the need for experiments under relevant divertor conditions [2]. To carry this out a vapour box module (VBM) has been designed and manufactured to be operated together with the linear plasma generator Magnum-PSI. The VBM consists of one hot central box where the Li is evaporated and two cold side boxes where it recondenses. By doing so, a Li neutral vapour is provided to interact with the plasma which passes through a central channel in the VBM before striking the target surface. The investigation of the plasma parameters such as n_e and T_e before and after it enters the VBM, as well as spectroscopy and target calorimetry, gives insight into the physics of the vapour shielding (VS), the mechanism behind the plasma power dissipation due to LM. The design of the VBM was moreover focused on the reduction of the Li outflow, and SOLPS-ITER modelling predictions show an expected reduction by three and two orders of magnitude between the flux from the exit of the central box and that transported upstream and downstream respectively. Witness plates and spectroscopy are employed to monitor this as well as post-mortem analysis of the box itself. Preliminary simulations carried out with SOLPS-ITER showing the plasma power reduction and the Li confinement capabilities of the VBM [3], had to be compared with a real experimental campaign. The results of the latter and the comparison with the simulation model are shown here.

[1] R.J. Goldston, et al., Phys. Scr., T167 (2016), Article 014017

[2] J.A. Schwartz, et al., Nucl. Mater. Energy 18, 350-355, (2019)

[3] J.A. Schwartz, et al., Nucl. Mater. Energy 26, 100901, (2021)

*Corresponding author: tel.: +31682705618, e-mail: f.romano@differ.nl (F. Romano)

Due to its high melting point, low hydrogen retention and good mechanical properties, tungsten (W) has been chosen as the plasma facing material for the divertor of tokamaks like WEST or ITER. The divertor has to support high heat flux (~10 to 20 MW/m² in ITER) and high particles bombardment (~10²² m^-2.s^-1 in ITER) from the plasma. Crystalline evolution of W under helium (He) irradiation and in particular He bubbles formation are responsible for an increase of hydrogen isotopes trapping [1] and cause an alteration of physical and mechanical properties of W [2]. Hence, understanding the shape and the evolution of He bubbles are crucial points for tokamak exploitation. Transmission Electron Microscopy (TEM) is widely use to characterize He bubbles [3], but their shape is still unclear. Grazing Incidence Small Angle X-Ray Scattering (GISAXS) [4] is a non-destructive technique providing electron density fluctuation information in the sub-surface region of a macroscopic area, it is a promising technique to study He bubbles shape and growth as shown by Thompson et al. [5].

In this work, the shape and the growth kinetics of He bubbles in W have been studied in-operando during He ions bombardment of single crystalline samples by in-situ GISAXS at the European Synchrotron Radiation Facility (ESRF) in Grenoble, France and by post-mortem TEM analysis. Ions energy was held either at 400 eV (dose is 4.2x10²¹ m^-2) or at 2 keV (dose is 1.0x10²² m^-2) under constant temperature conditions (RT, 500 K, 900 K and 1200 K). 2 keV implanted samples were then annealed up to 1700 K under vacuum. In-situ GISAXS and TEM reveal the formation of facetted bubbles. We propose a shape composed of {100}, {110} and {211} facets. The growth kinetics of facets is obtained from in-situ GISAXS measurements during implantation and annealing steps.

[1] M. Ialovega et al., Phys. Scr. T171, 014066 (2020).

[2] G. De Temmerman et al., Plasma Phys. Control. Fusion 60, 044018 (2018).

[3] K. D. Hammond, Mater. Res. Express 4, 104002 (2017).

[4] G. Renaud, R. Lazzari, and F. Leroy, Surface Science Reports 64, 255 (2009).

[5] M. Thompson et al., Journal of Nuclear Materials 473, 6 (2016).

*Corresponding author: tel.: +33 695068699, e-mail: loic.corso@cea.fr (L. Corso)

We studied by means of PIC simulations the variation of a tungsten surface temperature (T_s) as a function of the properties of a facing plasma close to the edge conditions met in tokamaks in terms of particle density and temperature. The determination of T_s is achieved by solving the steady state heat equation during the plasma simulations. The temperature of the outside wall is set at 300 K, while the heat balance at the inner surface is calculated at each time iteration; it consists of the heat flux deposited by the plasma itself (kinetic energy and potential energy due to ion recombination or electron absorption), of the radiative flux proportional to T_s⁴ and of the energy lost by the surface in the plasma due to electron thermionic emission [1]. Therefore at each time step an electron current is injected at the surface according to the Richardson law, depending on T_s. These emitted electrons lower the positive space charge in the sheath, which yields a reduction of the sheath potential. Consequently, the heat flow deposited by the plasma electrons increases, as well as the surface temperature and the emitted current, until the system eventually reaches a space charge limited regime [2]. In this case, the electron current penetrating the plasma is maximum and its amplitude is regulated by the presence of a virtual cathode, ie. a potential dip, at the surface vicinity.

We present the results of 1D PIC simulations, where the plasma density, the electron temperature, as well as the thermal conductivity of the surface are varied. For all simulations, the surface temperature exhibits a bifurcation between a regime of low and high temperature (or of low and high emitted current) when the incoming plasma energy is large enough. The high temperature regime appears when the emitted current is space charge limited and its ability to cool down the surface is maximum. We also compare the 1D kinetic simulations results to a 1D fluid model, which can predict the surface temperature with a relative good agreement. We finally compare the 1D PIC simulated surface temperature to the one obtained in 2D PIC simulations, where a localized emission spot is taken into account and a return current at the surface is possible in its surrounding region, just like in unipolar arcs models [3].

[1] M.Z Tokar, A.V Nedospasov and A.V Yarochkin, Nuc. Fusion 32, 15 (1992)

[2] G.D. Hobbs and J.A. Wesson, Plasma Phys. 9, 85 (1967)

[3] A.E. Robson and P.C. Thonemann, Proc. Phys. Soc. 73, 508 (1959)

*Corresponding author: tel.: +33 669529749, e-mail: jerome.moritz@univ-lorraine.fr (Dr J. Moritz)

Recent work [1] indicates that the degradation of the thermal diffusivity in self-ion irradiated tungsten (W) depends heavily on the accumulation of small, irradiation-induced defects that are not effectively measured by existing transmission electron microscopy (TEM) techniques. Differential X-ray diffuse scattering (XRDS) measurements have been used to determine the concentrations of these small defects in other materials, and a recent study of ion-irradiated nickel employing the two techniques has demonstrated that TEM measurements undercount defects smaller than 15 Angstroms in radius [2]. However, access to the high intensity X-ray sources required to perform differential XRDS experiments is limited which prevents on-demand measurements of the defect population using this technique.

In this work, we make use of a modest intensity, laboratory X-ray source and a more efficient experimental method in order to measure the integral diffuse scattering from single crystalline W samples irradiated with both 20 and 8 MeV self-ions. The integral scattering cross sections for dislocation loops in W have been numerically calculated and are utilized to determine the concentration and size distribution of irradiation-induced loops from these measurements. These rigorous calculations are a crucial improvement to the analysis of integral scattering measurements and enable the concentrations of loops as small as 3-6 Angstroms in radius to be measured with this technique. Our measurements show that small loops with sizes below the TEM resolution limit make up a substantial portion of the defect population; and, most significantly, their concentrations in the near-bulk can be measured non-destructively using relatively accesible, fixed-anode laboratory X-ray sources.

The defect densities inferred from our integral XRDS measurements are used in conjunction with a simple kinetic model for electron transport in the presence of crystalline defects [1] to predict the associated degradation in thermal diffusivity. These predictions are compared with Transient Grating Spectroscopy (TGS) measurements of the thermal diffusivity in the damaged layer of the crystal and significantly improve upon predictions made with existing TEM measurements of self-ion-irradiated W. These results open the door to more extensive measurements of atomistic-scale defect populations induced by various irradiation conditions and the development of models to link the defect population to macroscale engineering material properties.

[1] A. Reza, Acta Materialia 193, 270-279 (2020)

[2] R. Olsen, Journal of Nuclear Materials 469, 153-161 (2016)

Despite the most recent fusion advancements, there is still a long way to go before fusion can provide clean, efficient energy. To ensure the success of the fusion devices, it is critical to determine the capacity and endurance of plasma-facing materials and components under the anticipated operating circumstances and severe neutron fluxes.

The highest energy neutrons in the EU DEMO are approximately 15 MeV. These neutrons will be utilized to produce energy and breed tritium. However, it is inevitable that neutrons will come into contact with inside-vessel elements like the divertor and breeding blanket. These values should be as low as possible since such interactions can result in material activation, nuclear heating, and contact dose rates. The Breeding Blanket (BB) component carries out a number of functions. It serves as a cooling mechanism first. The lithium in the breeding zone and the neutrons created within the plasma interact nuclearly, converting the kinetic energy of the neutrons into heat that may be eliminated. To achieve self-sufficiency, the tritium fuel must be produced using the same nuclear processes. Additionally, the breeding blanket acts as a barrier, stopping high-energy neutrons from leaving the reactor and protecting the more radiation-sensitive parts, including the superconducting magnets, from damage [1]. The divertor component also performs several functions. However, heat and helium ash extraction are the primary uses [2]. During device operation, the structure is exposed to significant thermal stresses brought on by intense particle interactions. The plasma-facing component, linear and reflector plates at the divertor, and armour of the breeding blanket should absorb the most significant amount of high-energy neutrons. These components' parts are all constructed of tungsten.

This study presents calculations of neutron transport to determine an equivalent dose rate. This job was carried out using the MCNP - 6 tool and the JEFF-3.2 nuclear data library. Moreover, contact dose rate estimates were done using the TENDL-2017 nuclear data library and the FISPACT-II tool. After radiation, the dose rates were estimated for cooling durations ranging from 0 seconds to 100 years. For the mentioned cooling periods, the major radionuclides that contributed at least 1% to the overall dose rate value were determined. Thus four EU DEMO MCNP setups were used to accomplish the computations. There are two types of HCPB and WCLL breeding blankets: heterogeneous with a layered structure divertor and homogeneous with a semi-heterogeneous divertor, which were considered in the paper.

The effects of self-ion irradiation on the hydrogen retention properties of the fusion relevant materials tungsten, molybdenum, and zirconium are investigated and compared. Materials prepared by self-ion irradiation to simulate a neutron damaged microstructure are exposed to low energy deuterium ions to infuse the microstructure with hydrogen isotopes. A Device for Exposure to Low-energy Plasma of Hydrogen Isotopes (DELPHI) facility is utilised to support this study of the dynamics of hydrogen isotope retention and hydrogen release properties as a function of irradiation-induced microstructural damage. After exposure to hydrogen isotopes in DELPHI, Thermal Desorption Spectroscopy (TDS) was used to measure the hydrogen inventory, while Positron Annihilation Spectroscopy (PAS) was used to characterise the defects present in the prepared materials. It was shown that tungsten and molybdenum exhibited comparable relative increases in hydrogen retention, and this was further supported by the observation of equivalent increases in vacancies by PAS. The results of hydrogen isotope retention and release, as well as current understandings and potential challenges will be discussed.

In future fusion reactors one of the main material candidates for the plasma-facing components is tungsten (W), as it shows favourable properties such as high melting temperature, high thermal conductivity and low intrinsic retention of hydrogen isotopes (HI). However, in a future tokamak environment the 14 MeV neutrons from the D-T fusion reaction will create defects in the crystal lattice, altering the material properties. In order to study the created defects in the tungsten material we have used Rutherford Backscattering Spectrometry in Channeling configuration (RBS-C). Ion channeling is a well-established method for studying material properties related to its crystal structure in particular lattice disorder and defects evolution induced by ion irradiation. For quantification of the disorder, the change of the ion yield of the backscattered ions is measured along a crystallographic direction.

With the aim of studying the defects evolution by RBS-C, (111) tungsten single crystals were irradiated with 10.8 MeV tungsten ions at two different doses (0.02 and 0.2 dpa) and two different temperatures (290 K and 800 K) to create different microstructures in the material [1]. Detailed Transmission Electron Microscopy (TEM) analysis of the samples was performed showing dislocation lines and loops of different size, depending on the irradiation dose and temperature. Multi-energy RBS-C analysis along the <111> direction with four He⁺⁺ beam energies of 4.5, 4.0, 3.5, and 3.0 MeV was performed. The response of the induced structural damage signal versus analysing energy gives an important information about the extension of the defects (uncorrelated or extended defects) [2,3]. For the sample with the highest damage dose the relative disorder level extracted from these measurements increases with energy. This is interpreted as extended defects such as dislocation lines which are indeed observed by TEM. In order to interpret more quantitatively the measured RBS-C spectra simulations were made by the RBSADEC code [4]. In this case realistic defects were generated from collision cascades molecular dynamics (MD) simulations. For the low damage dose sample the simulation is in good agreement with the experimental data. However, for the high damage dose it does not give the same energy dependence. The most probable reason is that the MD simulation gives only dislocation loops but no lines in contrast to the experiment. This research is performed within the EUROfusion Enabling Research project DeHydroC where one of the main goals consists of developing tools to differentiate between small and large defects and to be able to study the evolution of defects during annealing or damaging by RBS-C.

[1] Hu et al. J. Nucl. Mater. 556, 153175 (2021)

[2] L.C. Feldman et al., Academic Press, San Diego, (1982), pp. 88–116

[3] Zhang, S. et al. Phys. Rev. E 94, 043319 (2016).

*Corresponding author: tel.: +386 5885 265, e-mail: esther.punzon-quijorna@ijs.si (E. Punzon Quijorna)

Tokamak dust mobilization has been nowadays recognized as an important aspect of dust transport and survivability [1]. Recent investigations of dust collected in various tokamaks have provided evidence of the presence of a significant fraction of ferro-magnetic and strongly paramagnetic dust (TEXTOR [2], FTU [3,4,5], Alcator C-Mod [4,5], COMPASS [4,5], and DIII-D [4,5], up to 30wt% depending on machine conditions and operational time). Magnetic particulates, in stark contrast to non-magnetic ones, can be mobilized during, or even before, discharge start-up [6]. To date not enough attention has been paid to pre-plasma remobilization of magnetic dust. In the perspective of the use of stainless steel for the ITER diagnostic first wall [6] and of RAFM steel in future fusion plants [7], a fraction of magnetic dust could be an issue for the breakdown phase in these devices.

In this work, we present a combined on-line and off-line investigation of magnetic dust in ASDEX Upgrade by means of Mie-scattering diagnostics, IR-camera observations, and collection activities. The magnetic dust fraction has been characterized by optical microscope and x-ray diffraction techniques. Post mortem collection revealed similar chemical composition and morphological features compared to magnetic dust from other tokamaks, but the overall amount was much smaller (few wt%). On the other hand, the on-line investigation has excluded the presence of fly-by dust across the beginning of plasma discharges. The experimental observations are discussed in light of the small amount of magnetic dust generated in ASDEX Upgrade despite the presence of ferromagnetic P92 still covering the central column of the vessel [8]. The detailed analysis of the temporal evolution of the magnetic field suggests that magnetic dust could still be mobilized, but well before the beginning of the plasma discharges.

[1] S. Ratynskaia, L. Vignitchouk, et al., Plasma Phys. Control Fusion 64, 044004 (2022).

[2] D. Ivanova, M. Rubel, V. Philipps, et al., Phys. Scr. T138, 014025 (2009).

[3] M. De Angeli, L. Laguardia, G. Maddaluno, et al., Nucl. Fusion 55, 123005 (2015).

[4] M. De Angeli, D. Ripamonti, F. Ghezzi, et al., Fus. Eng. Des. 166, 112315 (2021).

[5] M. De Angeli, P. Tolias, C. Conti, et al., Nucl. Mater. Energy 28, 101045 (2021).

[6] R. A. Pitts, B. Bazylev, J. Linke, et al. J. Nucl. Mater. 463, 748 (2015).

[7] K. Sugiyama, K. Schmid, W. Jacob, Nucl. Mater. Energy 8, 1 (2016).

[8] I. Zammuto, L. Giannone, A. Herrmann, et al., Fus. Eng. Des. 124, 297 (2017).

*Corresponding author: tel.: +39 0266173217, e-mail: marco.deangeli@istp.cnr.it

(M. De Angeli).

Droplet generation by arcing is one of the mechanisms that can generate dust in a
fusion device. In addition, metal droplets expelled by arcs can potentially introduce
impurities into the plasma and influence operation. The diameters and velocities of
droplets determine their capability to penetrate the outer layers of the plasma and,
therefore, the measurement of their distributions is a prerequisite for prediction of
their impact on plasma operation.

In order to investigate the properties of the droplets produced by arcs a dedicated
laboratory device has been commissioned based on time-of-flight detection by light
scattering. Velocity and size distributions have been measured at different emission
angles. Recently, a high-speed camera has been added for the observation of the
flight path of expelled droplets in larger quantities than the time-of-flight system.
Tracking software obtains trajectories from video recording to complement the
statistics obtained from the time-of-flight system and allow back-tracking of the
droplets’ point of origin as well as calculating the velocity distribution. Video recording
reveals the existence of explosive droplet emission events that could potentially be
the source of larger diameter droplets, relevant for fusion devices due to their
enhanced capability to introduce impurities into the plasma. The investigations are
complemented by microscopic investigation of origin craters in order to provide
additional insight into the emission process of different fusion relevant materials
focusing on tungsten.

Although tungsten is a promising candidate for future fusion reactors, W plasma facing components may encounter undesirable surface modifications due to the interaction with H isotopes in fusion plasma1 . Depth analysis unfolds fuel retention or erosion/deposition in PFC in fusion reactors. In this work, we have studied depth profiles of WTa-D mixed coating (thickness 5 μm, different Ta and D quantities up to 20 %) by Calibration-Free Laser Induced Breakdown Spectroscopy (CF-LIBS). The samples were prepared on a Mo substrate by NILPRP, Romania, using a dual magnetron sputtering system2 . While conventional LIBS is a versatile technique for quantitative analysis, in certain case it suffers in terms of sensitivity. To overcome this difficulty associated with conventional LIBS, we are implementing Resonant- Laser Induced Breakdown Spectroscopy (RLIBS). In this technique, a single tuneable laser is used to ablate the material from the sample based on Resonant Laser Ablation (RLA) coupled with Optical Emission Spectroscopy (OES)3 . A laser beam from a tuneable Optical Parametric Oscillator (OPO) is used to ablate the material. The ablation laser wavelength is associated with excitation levels of W chosen at 255.135 nm for resonance conditions and 222.75 nm for non- resonance condition. The measurements are carried out in atmospheric pressure. OES signals from 100 subsequent laser shots in depth and repeated at several locations on the sample were recorded by an Echelle spectrometer with a resolution of 4000 (ME5000, Andor, detection range 200–856 nm) equipped with an iCCD camera (iStar's DH743) with a temporal resolution of 5 ns via a quartz optical fibre cable. The gate delay are set at 200, 300, 500 and 1000 ns to obtain optimum conditions for plasma. A smaller ablation rate and so better depth resolution was observed at resonance condition compared with higher ablation rate in non-resonance condition (using the same energy in pulse). Unlike in Resonance-Enhanced LIBS (RELIBS), where two laser beams are used, one for ablation and the other for resonant excitation, in RLIBS to maintain the simplicity of experimental setup, only one laser beam is used for ablation and simultaneous excitation of the plasma. Hence with RLIBS, better depth resolution has been observed in resonant condition when compared with the non-resonant case. The spectra are divided by sensitivity curve and the peaks are identified using NIST and Harvard databases. A comparison of the signal enhancement in resonant comparing to the nonresonant conditions has been studied and CF-LIBS analysis is carried out. Finally, the results are compared with those obtained from TOF-ERDA and GDOES.

1. Miyamoto, M. et al., Journal of Nuclear Materials 415, S657 (2011).

2. Dwivedi, V. et al., Eur. Phys. J. Plus 136, 1177 (2021).

3. Cleveland, D. et al., Appl Spectrosc 59, 1427 (2005).

*Corresponding author: pavel.veis@fmph.uniba.sk (P.Veis)

A semi-analytical surface model is developed to model sputtering (both physical and chemical), reflection and redeposition of impurities and is coupled with a global particle tracking code (GITRm [1] – Global Impurity Transport) to study the erosion and transport of silicon carbide (SiC) and validate it with DIII-D experimental data. It provides erosion rates and migration patterns of impurities and evolution of surface compositions in contact with the plasma. SiC represents a promising plasma-facing material for next-step fusion devices because of low hydrogenic diffusion, good mechanical and thermal properties under neutron irradiation and the benefits of a low Z impurity. The current work aims at developing and validating predictive models of SiC physical and chemical sputtering, Si and C impurity transport and redeposition in SOL plasma and of the degradation of SiC plasma-facing components in time. The proposed surface model is similar to the homogeneous mixed material model used in ERO1.0 [2]. The GITRm simulations have been run to the point where the surface concentrations of impurities attain steady-state with respect to time. The physical sputtering calculations of the surface model have been validated with a high-fidelity surface code called RustBCA [3]. We also benchmarked the coupled GITRm and surface model simulations against experiments that have been conducted to measure the gross and net erosion rates of silicon carbide coatings in the lower divertor of DIII-D under well diagnosed plasma conditions [4]. SiC material was exposed to L-mode attached plasma using the Divertor Material Evaluation System (DiMES). For low energy plasmas, chemical sputtering of SiC is shown to play a more prominent role than physical sputtering. This work showcases the recent developments and application of a framework to handle multi-species impurity tracking with the coupled GITR + surface model. Estimations of SiC PFC erosion lifetime for proposed toroidal limiters at DIII-D are also briefly discussed.

[1] Shephard et al., 63rd Annual Meeting of the APS Division of Plasma Physics, Session NM09; 2021 Nov. 13, Pittsburgh, PA. NM09.00005 (2021).

[2] Kirschner A. et al, J. Nucl. Mater. 390–391 152 (2009)

[3] Drobny et al., Journal of Open Source Software, 6(64), 3298 (2021).

[4] Rudakov, D. L., Wampler, W. R., Abrams, et. al., Phys. Scripta 014064 (2020).

*Corresponding author: tel.: +1 6128761892, e-mail: dea@fusion.gat.com (Aritra De)

This material is based upon work supported by the U.S. Department of Energy, Office of Science under Award Numbers DE-SC0018423 and DE-FC02-04ER54698.

Disclaimer: This report was prepared as an account of work sponsored by an agency of the United States Government. Neither the United States Government nor any agency thereof, nor any of their employees, makes any warranty, express or implied, or assumes any legal liability or responsibility for the accuracy, completeness, or usefulness of any information, apparatus, product, or process disclosed, or represents that its use would not infringe privately owned rights. Reference herein to any specific commercial product, process, or service by trade name, trademark, manufacturer, or otherwise does not necessarily constitute or imply its endorsement, recommendation, or favoring by the United States Government or any agency thereof. The views and opinions of authors expressed herein do not necessarily state or reflect those of the United States Government or any agency thereof.

The binary-collision approximation based code SDTrimSP is a validated Monte-Carlo code for the simulation of ion-solid interactions. In recent years the code has been successfully extended to cover also the interaction of energetic particles with arbitrary shaped, three-dimensional targets and samples exhibiting complex surface morphologies.

Up to now, the assumption of a locally amorphous target has been used in SDTrimSP. This assumption is well justified in many circumstances, since under particle bombardment often rapid amorphization processes are induced. However, there are also many cases where an underlying crystalline sample structure significantly affects quantities of interest like sputtering [1], reflection yields [2] or damage distributions. For that reason an explicit lattice module has been added to SDTrimSP, which allows to switch between simulations of amorphous and crystalline samples with a single command, keeping all other settings unchanged. This enables for example the direct investigation to which extend the sample structure influences experimentally accessible quantities like sputter yields.

In the presentation we compare the results of the lattice module with simulation results from full molecular dynamics (MD) simulations, i.e. for the case of hydrogen bombardment of tungsten and with results from the MARLOWE code [3] as well as with the corresponding results under the assumption of an amorphous target.

[1] K. Schlüter et al, Phys. Rev. Lett. 125, 225502 (2020)

[2] M. Hou and T. Robinson, Nucl. Instr. Meth. 132, 641 (1976)

[3] M. T. Robinson and I. M. Torrens, Phys. Rev. B 9(12), 5008 (1974)

Investigating helium (He) plasma and its interaction with plasma-facing materials is one of the key requirements in light of ITER pre-fusion operation phase. Experimental campaigns devoted to plasma-wall interaction studies with He plasma have been recently performed in WEST, JET-ILW and AUG tokamaks in support of the preparation of ITER operation. In addition, experiments performed in linear plasma devices (LPDs) can already reproduce some of the relevant exposure conditions expected in ITER, e.g. in terms of ion fluxes and fluences, and are routinely used to investigate He-induced materials modifications. In this framework, numerical simulations provide an essential tool for the interpretation of LPD experiments extracting information on plasma dynamics and composition, erosion and surface evolution, impurity migration and redeposition.

This contribution investigates these aspects considering a He plasma experiment in the GyM linear plasma device [1] modelled through the coupling of SOLPS-ITER [2] and ERO2.0 [3] codes. Recent work presented the coupling strategy of the two codes and applied them to study the erosion of internal walls in GyM due to He plasma, as a function of the wall material and of the bias voltage [4].

The results presented aim to evaluate the effect of impurities sputtered from the stainless-steel wall and from the sample-holder molybdenum mask onto the tungsten exposed samples. The He fluxes on the wall and on the sample-holder (approximately 10²⁰ He m^-2 s^-1) are retrieved from SOLPS-ITER simulations and benchmarked against Langmuir probes experimental measurements. The sample-holder geometry is drawn with a Computer Aided Design (CAD) tool and imported into ERO2.0 simulation volume. The SOLPS-ITER plasma background is then exploited for ERO2.0 erosion and migration studies. In particular, the contribution of wall impurities on sample erosion is estimated, identifying the main wall components which influence experimental measurements. In addition, the influence of impurities eroded from the sample-holder structures on sample erosion is investigated. Finally, the migration of particles sputtered from the samples in the GyM volume and their redeposition locations are presented.

[1] A. Uccello et al, 2020 Nucl. Mater. Energy, 25 100808

[2] X. Bonnin et al, 2016 Plasma Fusion Res. 11 1403102

[3] A. Kirschner et al, 2018 Plasma Phys. Control. Fusion 60 014041

[4] G. Alberti et al, 2023 Nucl. Fusion 63 026020

Helium (He) plays as product of Deuterium-Trtitium (DT) fusion a key role in the realisation of a nuclear fusion plant and must be finally efficiently exhausted from the confined plasma to avoid dilution of the fusion fuel mix The He exhaust takes place in the divertor where He ions are neutralised at the plasma-facing components and pumped away as neutral He gas in the divertor. However, apart from surface recombination, energetic He ions give rise to other plasma-wall interaction processes with plasma-facing materials, i.e. Beryllium (Be) and Tungsten (W) employed in JET and envisaged for ITER. Energetic He+ and He2+ ions can e.g. release Be and W by physical sputtering; with the erosion flux depending on the flux, charge state, and energy and angular distribution of the incident ions. In principle, in pure He plasmas chemical erosion processes do not occur and so that pure He plasmas can be used to measure gross Be and W sputtering yields as function of the key parameters mentioned above. Residual Hydrogen Isotopes (HI) stored in the plasma-facing materials as well as residual HI in the plasma can still induce chemical erosion, but related to the HI faction involved.

Measurements with mono-energetic and single ionic He species at pre-defined incident angles have been done in ion beam experiments and linear plasmas devices and compared with binary-collision models like implemented in TRIM_SP in previous years. For a dedicated benchmark experiment in a magnetically confined plasma device, JET operating with its Be first wall and W divertor was operated pure He plasma discharges in 2022. Focusing on the W plasma-facing components in this contribution, the conditions in JET are more challenging compared with the mentioned experiments in ion beam and linear plasma devices as the tokamak conditions include at once an energy distribution, an impact angle distribution, a flux distribution, and a combination of different ionisation stages of ions as well as a mixture of Be and He ions varying along the interaction zone.

A series of JET L-mode discharges in low triangularity shape with the outer strike-line positioned on the bulk W-divertor have been carried out at a plasma current Ip of 2.0MA and toroidal magnetic field Bt of 2.0T. The L-mode discharges varied in input power Paux (1, 2 and 5 MW) by either He neutral beam injection or radio frequency heating of the H minority in the plasma. The impact energy of the impinging ions was varied by application of He gas puff ramps in the divertor varying the divertor conditions from ionisation to complete detachment or in other words from hot to cold divertor conditions. This allowed an assessment of both the W erosion yield and the energy threshold for W sputtering in JET. Optical emission spectroscopy (OES) of different W I lines (including WI at 400.9nm and 429.5nm), was applied to quantify the W gross erosion source, OES of BeI-IV and HeI-II lines giving complementary the impurity composition, and Langmuir Probes, were used to measure the total impinging ion flux, thus providing together quantitative information about the W sputtering behaviour under these plasma conditions. In general two characteristic energy thresholds for W erosion can be observed in all discharges carried out reflecting the W sputtering by Be ions and He ions, respectively. The highest threshold is energy is related to the main plasma species He whereas the lowest threshold is caused by Be ions and other trace impurities present in the plasma (e.g. oxygen and carbon). The absolute magnitude of gross erosion flux depends on the impact energy and total impinging flux in the different discharges. Additional observation of WII emission lines at e.g. 434.8nm permitted access to the lowest ionic W species allowed to determine the flux ratio of W0 and W+, thus providing information on the local re-deposition of W, or comparing with the gross erosion, the final net erosion flux of W in these L-mode plasmas without transients. The OES was performed by a direct imaging optical system equipped with three spectrometers covering a large spectral interval in the UV and VIS range so that multiple WI and WII lines from different ground state levels could be recorded. This permits a) a benchmark of the collisional radiative model for W in the ADAS database for both ionisation stages and b) improved modelling of global W erosion with plasma-wall interaction codes like ERO using the atomic and molecular data from ADAS as input.

Erosion of plasma-facing components affects their lifetime and other plasma-material interaction (PMI) issues important for ITER. Microscale morphology is shown to have a significant effect on surface sputtering properties [1], thus influencing the erosion-deposition pattern in tokamaks. Linear plasma devices (LPDs) are a perfect testbed for the investigation of this topic due to their cost-effectiveness and well-controlled exposure conditions. Modelling of the experiments with PMI codes, like ERO2.0 [1-2], is then highly recommended to gain insight into relevant processes.

Present work reports on the investigation of the role of roughness in the sputtering process of tungsten (W) by helium (He) plasma of the linear device GyM (B≅80 mT). Helium is of great interest when studying PMI since it will be present in a fusion plasma as an intrinsic impurity and it will also be the main plasma species during ITER pre-fusion power operation. W coatings deposited on: silicon (Si) substrates with pyramids on the surface and four different average roughnesses (R_a≅3, 300, 600, 900 nm), and graphite substrates with irregular surface and three different R_a (≅7, 90, 280 nm), as well as reference polished bulk W samples (R_a≅10 nm), have been exposed in GyM changing the incident He⁺ energy (E_He+) between one experiment and the other in 30 – 350 eV range (i.e. by applying different bias voltage values to the samples), for a fluence of 4.0e24 He⁺m^-2. Net erosion of the samples has been estimated from mass loss data. Morphology modifications have been investigated by scanning electron and atomic force microscopies. Experimental outcomes have been finally compared to ERO2.0 results.

Considering W/Si samples, surface modifications were limited to the formation of ripples at the nanoscale for E_He+≥250 eV. This allowed to evaluate the quasi-static effective sputtering yield (Y_W|_GyM) from mass loss data, on the one hand, and run single time step ERO2.0 simulations, on the other hand. For E_He+≤200 eV, Y_W|_GyM is negligible. For higher energies, Y_W|_GyM decreases by increasing the mean value of the surface inclination angle distribution (δ_m), in agreement with ERO2.0 results. δ_m is thus the key-parameter determining the erosion of the samples rather than R_a, as also pointed out in [3]. Moreover, Y_W|_GyM is about one order of magnitude lower than that from ion beam experiments and binary collision approximation (BCA) calculations, confirming what was observed in other LPDs [4]. Since the energy and angular distributions of sputtered particles in ERO2.0 were provided by the BCA SDTrimSP code, the effective Y_W from the code is also higher than Y_W|_GyM. Calibration of ERO2.0 input, as the sputtering distributions and the incoming plasma flux, was necessary to improve the quantitative agreement with experimental data.

The exposures of W coatings deposited on graphite substrates and polished bulk W samples are currently ongoing and the results will be presented during the Conference.

[1] A. Eksaeva, et al., Phys. Scr. T171, 014057 (2020)

[2] G. Alberti, et al., Nucl. Fusion 61, 066039 (2021)

[3] C. Cupak, et al., Appl. Surf. Sci. 570, 151204 (2021)

[4] R.P. Doerner, Scr. Mater. 143, 137-141 (2018)

*Corresponding author: tel.: +39 02 66173453, e-mail: andrea.uccello@istp.cnr.it

Plasma facing material sputtering under ion bombardment is a major topic for plasma fusion as it can limit these components lifetime, induces contamination issues in plasma core, and, in the end reduces fusion process efficiency.

Until now, electrons contribution to surface erosion has been neglected. However it has been experimentally proven that a combine electron beam to the surface can significantly increase ion sputtering process by partially ionizing material surface [1]–[3]. This phenomenon called synergic erosion is defined as the sputtering of a surface under a conjugated beam of ions and electrons. It could have a major impact on material ageing in tokamaks where the mean energy of the electrons impacting the walls is close to 100 eV [4].

This phenomenon has been brought to light experimentally on an ONERA irradiation set-up [5]. Researches that are more recent have attempted to give quantitative results of erosion yield for two materials (SiO₂ and BN) under combined electron and xenon (Xe⁺) ion impact to compare with erosion yield for Xe⁺ ion impact only. For now, our experimental set-up is limited to noble gas plasma only (argon, krypton, xenon, etc.). Although this is not representative of a tokamak wall ion sputtering, as xenon is not chemically reactive, it allows distinguishing between purely kinetic and chemical components of ion sputtering. The development in the future of new experimental set-up in ONERA will nonetheless allow creating chemically reactive plasma (such as hydrogen, oxygen, etc.).

In this paper, we will present the experimental methodology and set-up used to carry out a parametric study of erosion under Xe⁺ ion beam with and without added electron beam. This work present the obtained results, analyse and compare them to tokamak operating conditions. This paper will also present the physical processes steering these synergetic effect.

[1] J. Ahn, C. R. Perleberg, D. L. Wilcox, J. W. Coburn, and H. F. Winters, J. Appl. Phys., vol. 46, pp. 4581–4583, Oct. 1975.

[2] B. Lang, Appl. Surf. Sci., vol. 37, no. 1, pp. 63–77, Jan. 1989.

[3] B. Carrière and B. Lang, 1977.

[4] A. Eksaeva et al., Phys. Scr., vol. 2020, no. T171, 2020.

[5] P. Sarrailh et al., presented at the International Electric Propulsion Conference, IEPC-2017, Atlanta (GA - USA), Oct. 2017.

Understanding impurity transport is critical for minimizing net erosion of plasma-facing surfaces in a tokamak, and for minimizing core contamination by impurities. Preliminary impurity transport modeling indicates that the majority of eroded tungsten in the DIII-D V-shaped small angle slot, tungsten-coated (SAS-VW) divertor is re-deposited within the divertor, closer to the slot vertex. However, higher net erosion near the outboard slot entrance may contribute to elevated tungsten leakage upstream and into the core.

The Global Impurity Transport Code (GITR) is a 3D fully kinetic Monte Carlo code that tracks impurity particle motion, capturing full gyro-motion, prompt re-deposition, and long-range impurity migration. The plasma profiles in the scrape-off layer (SOL) simulated using SOLPS-ITER were used to inform the GITR simulations [1,2].

GITR has been validated in a linear device [3], although it has yet to be validated in a tokamak environment. GITR simulations of tungsten transport in the DIII-D SAS-VW divertor, in combination with experimental analysis [4], provide a validation opportunity for GITR predictions in a tokamak. This validation will allow GITR to inform and optimize future designs of closed slot-like divertors to further minimize net erosion and the leakage of tungsten impurities into the core.

The SAS-VW divertor is designed to promote divertor detachment, thus reducing net erosion and providing divertor concepts that lengthen plasma-facing component lifetimes. Net erosion is difficult to measure experimentally using existing diagnostics, but GITR provides a high-fidelity physics model for tracking gross erosion, prompt re-deposition, and net erosion rates of tungsten, in addition to the tungsten impurity migration along the SOL. Predicted gross erosion rates and SOL impurity concentrations will be compared against experimental spectroscopic measurements.

The gross erosion rate of tungsten at 3 locations along the SAS-VW divertor was experimentally measured using a multi-chordal divertor spectrometer (MDS) to quantify characteristic W-I 400.9 nm photon emission. The S/XB method is used to calculate the experimental average gross erosion flux for direct comparison to the gross erosion flux predicted using GITR. Leakage from the SOL into the core is experimentally characterized using X-ray and UV emission spectroscopys. Equivalent synthetic diagnostics in GITR provide the computational prediction for comparison.

[1] R. Maurizio et al., Nucl. Fusion 61, 116042 (2021)

[2] G. Sinclair et al., Nucl. Fusion 62, 106024 (2022)

[3] T.R. Younkin et al., Nucl. Fusion 62, 056007 (2022)

[4] T. Abrams et al., Phys. Scr. 96, 124073 (2021)

Funding statement: This work contributes to the Plasma Surface Interactions 2 project, which is part of the Scientific Discovery through Advanced Computing (SciDAC) program, and is jointly sponsored by the Fusion Energy Sciences (FES) and Advanced Scientific Computing Research (ASCR) programs within the U.S. Department of Energy Office of Science. This material is based upon work supported by the U.S. Department of Energy, Office of Science, Office of Fusion Energy Sciences, using the DIII-D National Fusion Facility, a DOE Office of Science user facility, under Awards DE-FC02-04ER54698, DE-AC05-00OR22725, and DE-SC0018423.

Disclaimer: This report was prepared as an account of work sponsored by an agency of the United States Government. Neither the United States Government nor any agency thereof, nor any of their employees, makes any warranty, express or implied, or assumes any legal liability or responsibility for the accuracy, completeness, or usefulness of any information, apparatus, product, or process disclosed, or represents that its use would not infringe privately owned rights. Reference herein to any specific commercial product, process, or service by trade name, trademark, manufacturer, or otherwise does not necessarily constitute or imply its endorsement, recommendation, or favoring by the United States Government or any agency thereof. The views and opinions of authors expressed herein do not necessarily state or reflect those of the United States Government or any agency thereof.

Tungsten fiber-reinforced tungsten composites (W_f/W) have already shown that the mechanical properties of tungsten can be significantly improved by incorporating fibers into the metal matrix. This matrix is usually produced by powder metallurgical processes or by chemically vapor deposited tungsten (CVD-W). By developing novel yarn-based textile preforms and distributing them homogeneously in the
CVD-W matrix by a layer-by-layer process, new KLST-type specimens were cut by electrical discharge machining for three-point bending tests according to DIN EN ISO 179-1:2000. The material properties of the fabricated specimens showed beside very high rel. densities of more than 97 % and fiber volume fractions over
14 % such reproducible force-displacement curves at room temperature that a cyclic loading test was performed on the W_f/W specimens with over 10,000 cycles between 50-90 % of the determined relative maximum load. These cyclic loading tests showed a clear trend of each local peak of the sinusoids and showed predictable fatigue behavior up to the point of critical displacement. Thus, the pseudo-ductility of W_f/W composites has reached a new benchmark by which fatigue behavior can be extrapolated from cyclic loading tests and conclusions can be made about the point at which the material will exhibit fatigue failure under cyclic mechanical loading.

The interaction between the plasma and the plasma-facing materials (PFMs) is crucial for achieving the optimal performance of the fusion devices. The fusionrelevant materials have been intensely studied by depth profile analysis using LIBS [1-2] and Calibration-Free (CF)-LIBS [3-5]. The aim of this work is the LIBS depth profile analysis of Be-based samples containing variable levels of trace elements originating from the fuel or seeding gases in the plasma or as other impurities to study the possible co-deposits and if and how they affect the fuel retention. A set of Be-based samples with varying amount of typical fuel and impurity elements (N, Ne, He, D) were produced to simulate the co-deposits on the PFMs. Three samples from inner and outer JET limiter tiles were also analyzed. The measurements were performed using a setup already described elsewhere [1] at two different pressures: 2 mbar and 100 mbar of argon. Relevant He spectral lines could be unambiguously distinguished but the samples containing N and Ne up to few percent did not show the corresponding spectral lines. As SIMS and TOF-ERDA observed these two elements, further analysis is necessary to establish the limit of detection of these elements for LIBS. In all the analysed samples D/H was observed. The signal of D was the highest on the surface and then reduced to a lower level. In lab-made codeposits, D/He is uniform throughout the rest of the sample and this is also revealed by LIBS. On JET samples, the main D peak is at the surface, however, LIBS gives indications of deeper deposition which is confirmed by SIMS.

[1] J. Karhunen, et al., J. Nucl. Mater. 463, 931 (2015) [2] M. Suchoňová, et al., Nucl. Mater. and Energy 12, 611 (2017) [3] P. Veis, et al., Phys. Scr. T171, 014073 (2020) [4] V. Dwivedi et al., Nuclear Materials and Energy 27, 100990 (2021) [5] H. van der Meiden et al., Nucl. Fusion 61 (2021) *Corresponding author: roldan2@uniba.sk (A. Marín Roldán) † See the author list of “Overview of JET results for optimising ITER operation” by J. Mailloux et al 2022 Nucl. Fusion 62 042026

Future fusion reactors will need to control and contain their tritium inventory to comply with safety and environmental regulations. In particular, permeation requires a close monitoring: through this mechanism, tritium penetrates the plasma-facing components (PFCs), gets trapped in the defects of the materials, and contaminates the coolant. To predict these phenomena, the transport and trapping parameters of tritium in these materials need to be evaluated and used in simulation codes.

Several nuances of steel are relevant for fusion, among which Eurofer97. This low-activation steel was studied with deuterium TDS experiments that revealed the presence of three trapping sites in this material, two of which have high detrapping energies (1.27 eV and 1.65 eV). Concurrent gas-driven permeation experiments performed with Hypertomate (HYdrogen PERmeation in Tokamak-relevant MATErials) between 200°C and 450°C with hydrogen highlighted the importance of trapping for permeation in this temperature range and yielded transport parameters such as diffusivity and solubility.

The results of these experiments were processed with the MHIMS reaction-diffusion code. This analysis showed that usual modelling parameters such as effective diffusivity and effective solubility over-simplify the complex dynamics of trapping, which can lead to an underestimation of the tritium inventory in PFCs [1] and to an inaccurate estimation of the permeation dynamics.

The influence of neutron damaging on retention was then investigated: to mimic such defects, pristine Eurofer97 samples were self-damaged with Fe ions in the JANNuS facility in Orsay (France). The impact on trapping parameters was studied through TDS experiments and the Eurofer97 model was adapted. The consequences of this self-damaging on retention will be discussed.

This analysis was complemented with gas-driven tritium experiments performed at room temperature. These experiments investigate the difference between permeation through Eurofer97 when the downstream medium is water or water and air. These measurements were performed with a new setup called Wapiti (WAter-interface Permeation In Tritium-exposed materIals) consisting in several permeation cells that expose thin Eurofer97 membranes to gaseous tritium.

The presentation covers both experimental and modelling results, as well as the adaptation of this method to other fusion-relevant materials such as 316L steel.

[1] F. Montupet-Leblond et al, Influence of traps reversibility on hydrogen permeation and retention in Eurofer97, Nuclear Fusion 62, (2022)

The estimation of fuel losses in a fusion reactor is very important in order to guarantee a safe reactor operation and a sustainable fuel cycle. Therefore, the hydrogen isotope permeation through first wall and breeding blanket materials has to be determined. In ITER, W and Be will be used as plasma-facing materials, Cu and Cu alloys (such as CuCrZr) as heat sink materials and 316L(N)-IG (IG: ITER grade) steel as structural material. By measuring the gas-driven hydrogen isotope permeation in these wall materials, general physical parameters are obtained, with which the estimation of fuel losses can be improved.
Deuterium permeation measurements through different ITER and future fusion device materials were performed. Specially fabricated test samples of combined wall materials were studied, in order to enable the estimation of the permeation flux through a whole plasma-facing component. The sample preparation, characterization and measuring cycle were kept identical for all samples in order to provide a reliable comparison between the samples. The crystal phase was analysed by x-ray diffraction and the microstructure was investigated by scanning electron microscopy. The deuterium permeation flux was measured between 300°C to 550°C at different applied pressures (range: 25 mbar to 800 mbar).
For the study of the bulk properties, 316L(N)-IG, Eurofer97, Cu and CuCrZr-IG substrates [1,2] were cut to disks, polished and analysed. The combined material systems were fabricated by magnetron sputter deposition. On the polished substrates, thin W or Cu layers were deposited. The material systems Cu on 316L(N)-IG [2], W on CuCrZr-IG and W on Eurofer97 [3] were studied. For the comparison of different coatings, the layer permeability [3] was calculated. In order to determine the influence of interfaces between the different materials, the thickness of the Cu layer was varied in the Cu on 316L(N)-IG system.
In this presentation, an overview of the permeability of bulk and combined systems will be given and discussed. The permeabilities of CuCrZr-IG and bulk Cu sample are very similar. The layer permeability of a Cu layer is more than one order of magnitude larger than the W layer permeability. The influence of the microstructure on the permeability is large compared to the influence of interfaces on the permeability in the combined systems.
[1] A. Houben, J. Engels, M. Rasiński, and Ch. Linsmeier, NME 19, 55 (2019)
[2] A. Houben, M. Rasiński, S. Brezinsek, and Ch. Linsmeier, NME 33 (2022), 101256
[3] A. Houben, M. Rasiński, L. Gao, and Ch. Linsmeier, NME 24 (2020), 100752
* Corresponding author: tel.: +49 2461 615102, e-mail: an.houben@fz-juelich.de (A. Houben)

For estimating tritium retention in fusion reactor wall materials, and for extrapolating
tritium losses through plasma-facing components as well as cooling structures, it is
important to understand the diffusion and permeation behaviour of hydrogen
isotopes. Usually, e.g., for computer simulations, it is assumed that hydrogen
isotopes permeate homogeneously through a given material, which justifies 1-D
simulation domains. However, it has been shown in several documented cases that,
e.g., grain boundaries, which typically form a 3-D percolating network, can have an
influence on H isotope diffusion, e.g., in Ni [1, 2]. In addition, multi-phase materials
such as tungsten heavy alloys inherently have a 3-D structure.
Using a novel application of hydrogenography, i.e., the change of the optical
properties of certain indicator materials such as yttrium if these are exposed to
hydrogen (see, e.g., [3]), we directly visualize the influence of grain boundaries in
recrystallized tungsten on D permeation, as well as of the percolating matrix in the
tungsten heavy alloy HPM-1850. The recently patented method [4] makes use of
structured yttrium thin films applied to the exit side of permeation samples to enable
permeation experiments with 2-D spatial resolution.
We believe our findings give a strong incentive to check whether 1-D diffusion
simulations are generally sufficient to describe H isotope diffusion in solids, or if 3-D
percolating structures in materials give rise to new effects. Tentative evidence for
such an effect has already been found during a study of D retention in HPM-1850.
Here, D retention in the 3-D structured heavy alloy differed substantially from that of
pure W for large D plasma fluences [5].
Furthermore, the hydrogenography method shown in our contribution may be useful
to detect and localize potential hydrogen leakage in future infrastructure distributing
hydrogen as fuel, as well as in the tritium plant of fusion reactors.
[1] A.M. Brass, A. Chanfreau, Acta Mater. 44, 3823-3831 (1996)
[2] A. Oudriss et al., Scripta Mater. 66, 37-40 (2012)
[3] F.J.A. den Broeder et al., Nature 394, 656-658 (1998)
[4] A. Manhard, U. von Toussaint, M. Stienecker, German Patent 10 2020 111 959.3
(2021) and International Patent PCT/EP2021/061391 (2021)
[5] H. Maier et al., Nucl. Fusion 60, 126044 (2020)

One of the main goals of DEMO (and ITER) will be to demonstrate tritium self-sufficiency which poses a critical engineering challenge: It requires efficient breeding and tritium extraction systems as well as minimizing losses of tritium to structural and plasma-facing components. These components will act as tritium sinks, because it can be trapped in the bulk of these components at defects. These defects are both intrinsic in nature but are also generated by energetic particle bombardment, mainly through displacement damage by the high energy fusion neutrons. To overcome these engineering challenges at design time both, efficient and accurate modeling tools and validated material parameter databases are needed. In the current DEMO design the majority of the structural components are made of the reduced activation ferritc martensitic steel EUROFER, whereas the plasma-facing armor elements will be made of Tungsten (W). While the material database for W has evolved quite significantly over the past ten years, the situation for model parameters for predicting tritium transport and retention in EUROFER is not satisfactory. In particular information about the trap site generation by displacement damage and its subsequent annealing at elevated temperatures is still lacking. Also the binding energy of tritium to these defects is not well known.
In this work the potential influence of HIs present during displacement damage on the resulting trap site concentration is investigated. EUROFER samples are exposed to multiple sequences of displacement damage by MeV W-ions and subsequent loading by deuterium (D)-ions from a low temperature plasma. After each of these damage/loading sequences the D depth profile in the surfaces is measured by nuclear reaction analysis (NRA) using a 3He beam. These depth profiles show a clear increase in the trap site concentration when the displacement damaging step is performed on a sample already containing trapped D. The presence of He in analysis spots from previous NRA-depth profiling also results in a strong increase of local retention, due to traps formed by the implanted ³He. These results suggest that, similar to W, the presence of D and He in EUROFER will result in increased retention compared to pure EUROFER under displacement damage. Different from displacement damaged W, thermal desorption spectra show that in EUROFER, as long as displacement damage is limited to the near surface, retention is dominated by intrinsic defects in the bulk.
By modeling the D loading of the damaged samples and the subsequent TDS measurements by state of the art diffusion-trapping codes, the trap types present in the material and the binding energy of HIs to these traps are determined.

The ITER blanket system is unique to the world. Fundamentally, the size required to enshroud the ITER plasma necessitates a scale of manufacturing never before encountered in fusion technology. This industrial production, combined with exposure to heat fluxes and particle fluences on an unprecedented scale, pushes the limits of industrial capabilities to fabricate blanket components.

Iterative collaboration with industry is essential for the qualification and demonstration of the many difficult-to-manufacture aspects of the blanket final design. Though semi-prototypes have previously demonstrated proof-of-principle of design, full-scale prototypes must ensure that the challenging sub-millimetric tolerances specified on the plasma-facing surfaces can be achieved in order to meet the operational power handling loads.

While we qualify and subsequently manufacture these complex systems, the design of many interfacing systems (e.g. diagnostics and remote handling) remains only partially complete. Navigating this parallel development in a tightly integrated tokamak requires progressive design handover to manufacturing suppliers of the many unique blanket module variants. This task is now nearing completion thanks to the development of several new design and analysis tools (especially in Field-Line Tracing, Electromagnetic analysis, 3D neutronics, and fracture mechanics to name a few). Additionally, effort is on-going with R&D to resolve issues in assembly and remote-handling, with maintenance strategies (including the provision of spares), and to sufficiently accommodate geometric tolerance issues with the ITER vacuum vessel.

Today, the blanket system is progressing on several parallel fronts. Several components are in series production, including the blanket shield blocks (more than half complete) and some of the blanket connections. Other systems, including the First Wall panels, the blanket manifold, and the module connections, are in a mix of qualification and manufacturing. Lastly, the ITER team is finalizing all First Wall and Shield Block variants for final handover to suppliers.

As with any first-of-a-kind system, several difficulties have arisen upon entering the full-scale production phase (remote handling compatibility, assembly welding reliability, plasma-facing surface tolerance capability and joint performance, customization limits, among others). These are being dealt with in turn, all while pushing forward towards completion of the series production of the complete ITER blanket system.

Water-cooled ceramic breeding blanket (WCCB) is the core component of CFETR, which is responsible for tritium breeding, neutron shielding and heat extraction. In order to realize these functions, in WCCB engineering design, mixed ball bed composed of tritium breeding materials Li2TiO3 and Be12Ti neutron multiplying materials were filled into the box body made of RAFM steel plates. In RAFM steel plate, close-lined square water-cooled flow channels were installed, and under normal operating conditions, 15.5 MPa pressurized water with inlet/outlet temperatures of 285◦C/325◦C is used to cool the blanket. At the same time, in order to improve the thermal and particle impact resistance of WCCB components, tungsten is selected as the plasma-facing material. Therefore, in the manufacturing of blanket components, it involves the connection of tungsten /RAFM steel dissimilar material, the connection of RAFM steel and RAFM steel and the forming of RAFM steel internal cooling channels. On the other hand, the WCCB using modular design, the size of each module is relatively large, about 1.2x1x1 m^3, resulting in tungsten area coated on RAFM steel is greater than 1 m², the length of the channel nearly 3 m, which increases the difficulty of module machining and manufacturing. To solve these issues, ASIPP sets up a special engineering research team to carry out the development and research of the manufacturing technology for CFETR WCCB module.

In this report, we will provide a brief overview of recent research progress in the fabrication and performance testing of full-size tungsten-coating first-wall and multi-bent double-wall coolant tubes, as well as the development of the final assembly process for full-size WCCB modules. The development of blanket first wall components is relatively difficult, so we adopted a step by step strategy. In the past five years, we have completed the development of 600mm straight mockup, small-scale U-shaped mockup and one-ninth U-shaped first wall mockup, thus breaking through the issues of large-area tungsten/steel diffusion connection, hot isostatic pressing forming of square flow channel and electron beam welding of bending welds with long of 3m, and then first completed the manufacturing of full-size first-wall components with W coating. Following the researches of the selection of intermediate layer, deformation control of circle tube and heat treatment process, we carried out the development of batch production of full-size multi-bending double-wall cooling tube. Finally, the engineering team figured out the assembly process of multiple steel parts of full-size WCCB module, and determined the machining technology and welding structure of each step. These studies have laid a solid foundation for the application of full-size WCCB module in CFETR.

Deuterium (D) plasma exposure of tungsten (W) can cause the formation of a D supersaturated surface layer (DSSL) with a thickness of ~10 nm, in which a significant amount of static D retention (up to ~10 at.%) has been detected [1]. In our previous work [2], in-operando LIBS (laser-induced breakdown spectroscopy) measurements were conducted during D plasma exposure in the PISCES-A linear plasma device to explore dynamic, as well as static, D retention in W. Concerning the incident ion energy, E_i, dependence, it was found that both dynamic and static retention stayed constant in a range of E_i ~ 45-175 eV. This is in contrast to [1], where a threshold for the DSSL formation around 100 eV was proposed.

To address this discrepancy, we have exposed W samples to D plasmas in PISCES-A at three different E_i(D⁺) ~ 75, 125, and 175 eV by controlling the bias voltage applied to the sample target. Other plasma exposure parameters were fixed: sample temperature T_s ~ 423 K, ion flux G_i ~ 1.3×10²¹ m^-2s^-1, and fluence f_i ~ 3.0×10²⁴ m^-2. The ion composition is estimated to be (D⁺, D₂⁺, D₃⁺) ≈ (0.58, 0.32, 0.10) using an updated 0-D model [3]. After the plasma exposures, thinned samples were created from the W samples using FIB (focused ion beam) for cross-sectional TEM (transmission electron microscopy) observations of the near-surface regions. It is revealed that the plasma-induced damaged layer, i.e. DSSL, is formed even at the lowest E_i ~ 75 eV. The layer thickness is found to slowly increase with increasing E_i from ~7-8 nm at 75 eV to ~11-12 nm at 175 eV. Furthermore, high-density cavities, most probably D bubbles, are clearly seen inside the DSSL. The present result is consistent with our previous experiment [2]. Although [1] speculated that the DSSL would not form below an ion energy of 100 eV, our results may be due to the higher G_i (x10) in our experiments.

The total static D retention in the W sample exposed at E_i ~ 75 eV was quantified with TDS (thermal desorption spectroscopy) to be ~4.7×10¹⁹ D m^-2. To further confirm the existence of D in the DSSL, the depth profile through the DSSL will be measured with a high depth resolution using GDOES (glow discharge optical emission spectroscopy), the signal of which will be calibrated with NRA (nuclear reaction analysis). Detailed analysis of the D depth profile will be reported at the conference.

[1] L. Gao, W. Jacob, U. von Toussaint et al., Nuclear Fusion 57, 016026 (2017)

[2] D. Nishijima, R.P. Doerner, M.J. Baldwin et al., Nuclear Fusion 61, 116028 (2021)

[3] D. Nishijima, E.M. Hollmann, M.J. Baldwin et al., presented at HTPD 2022.

Work supported by US-DOE cooperative agreement, DE-SC00225258.

*Corresponding author: tel.: +1-858-245-8107, e-mail: dnishijima@eng.ucsd.edu (D. Nishijima)

Controlled nuclear fusion research is approaching the point to where the development of a deuterium–tritium (D–T) ‘burning plasma’ fusion pilot plant (FPP) [1,2], as a prototype modest net fusion power reactor, is being contemplated. However, numerous scientific questions [1,2] and technology challenges that surround the choice and performance of the plasma facing materials, must still be solved before any credible next-step fusion facility can be realized. A significant issue identified [1], is the potential for synergistic impact on materials performance under simultaneous plasma and fusion neutron loading, which will not only introduce in the material surface, a non-equilibrium gas concentration due to plasma implantation, but also simultaneously occurring atomic displacement damage. These combined actions promote questions about material ‘in-service‘ performance (E.g. erosion, retention, thermal properties ...) under burning plasma relevant materials interaction (BPMI), since dissolved gaseous species can act to stabilize displacement damage, as has been observed in recent experiments, E.g. [3], and calculations [4].

To address these BPMI issues, PISCES linear plasma device experiments have been strategically refocused to contribute in several new ways. We will report on scientific and engineering results from the newly upgraded PISCES-Rf facility, which is a steady state 5-30 kW range helicon plasma device for plasma-materials experiments. This new PMI device can deliver, low-impurity, high-atomic-ion fluxes (D⁺, He⁺) at the material target, from dense, n_e ~10¹⁹ m^-3, D₂ and He plasmas of temperature, T_e in the range ~0.1-10 eV. We will also report on the current state of development of a new 3 MV accelerator that will be integrated with PISCES-Rf and provide an energetic heavy-ion beam that will induce displacement damage in PISCES-Rf targets (E.g. 10^-8 ­–10^-2 dpa /s for 20 MeV W⁶⁺ on W) as a surrogate for neutron damage. Results from in-situ/in-operando near-surface PMI diagnostics and first materials experiments will also be discussed.

Work supported by the USDOE award, DE-SC0021656, and cooperative agreement, DE-SC0022528.

[1] T Carter (Chair) FESAC Report to the US-DOE, https://science.osti.gov/-/media/fes/fesac/pdf/2020/202012/FESAC_Report_2020_Powering_the_Future.pdf (2020)

[2] R. Hazeltine (Chair), ReNeW Report to US-DOE, http://burningplasma.org/web/ReNeW/ReNeW.report.web412.pdf (2009)

[3] S. Markelj, Nuclear Fusioin 59 8 086050 (2019)

[4] D. Kato, H. Iwakiri, Y. Watanabe, K. Morishita and T. Muroga, Nuclear Fusion 55 5 083019 (2015)

*Corresponding author: tel.: +1 858 534 1655, e-mail: m1baldwin@ucsd.edu (M.J. Baldwin)

As part of the EUROfusion programmatic objective to fulfil Mission 3 (“Neutron tolerant materials”) of the European Roadmap to Fusion Electricity, neutron tolerant materials for in-vessel components are investigated within the Work Package Materials as a follow up of the former EFDA program. This comprises structural, heat sink, and plasma facing materials for the breeding blanket and the divertor as well as optical and dielectric functional materials for diagnostics and heat and current drive systems. Besides the assessment of the three baseline materials, i.e. EUROFER97 steel, tungsten and CuCrZr alloy, a multitude of newly developed as well as industrially available advanced materials were characterized in the non-irradiated and neutron irradiated condition.

By assessment of the material technology readiness level for each individual material with a focus not only on performance but also on industrially available production technologies and viability of component manufacturing, a down-selection process has been performed. The remaining material options being further investigated as part of the conceptual design phase study for DEMO as well as the rationale behind this down-selection is presented. The primary aim of the material’s program is to provide alternative, viable material options during the conceptual and engineering design phase of DEMO – during which the components’ design as well as the operational boundary conditions are still in development – resulting in a risk mitigation by extension of the operational performance or the accessibility of advanced component designs.

Assessing the performance of the ITER design for the tungsten (W) divertor Plasma Facing Units (PFUs) in a tokamak environment is a high priority issue to ensure efficient plasma operation. This contribution reviews the most recent results from experiments and post-mortem analysis of the ITER-like PFUs exposed in the WEST tokamak and the associated modelling, with a focus on understanding damage evolution. Several shaping options, sharp or chamfered leading edge (LE), unshaped or shaped blocks with a toroidal bevel as foreseen for ITER, were investigated, under steady state heat fluxes of up to 120 MW.m^‑2 and 6 MW.m^-2 on the sharp LE and top surface, respectively, and under fast transients due to disruptions. A very high spatial resolution (VHR) IR camera (0.1 mm/pixel) was used to derive the temperature and heat load distributions on the individual blocks during steady state plasma exposure, with a particular focus on the LEs. The main findings are summarized below:

• For the misaligned/unshaped PFUs within the range of specified assembly tolerances (±0.3 mm), regularly spaced cracking was observed on LEs. A dedicated modelling tool was developed to assess crack initiation under WEST conditions, showing that the observed crack pattern can be reproduced assuming brittle failure during disruptions.
• A block with 2.3 mm overexposed sharp LE was installed for a dedicated melting experiment. Crack formation on the LE has been detected for the first time in situ with the VHR IR system. The measurement gives a LE temperature of ≈2500°C for several seconds, possibly outlining the impact of W recrystallization on the process. W-melting occurred under higher power loads. The observed melting onset and displacement were successfully modelled with the MEMENTO code.
• On the top surface of the unshaped blocks, post-mortem analysis did not show any evidence of cracking, except for one pre-damaged block featuring a micro-crack network pattern (created in a high heat flux facility before exposure in WEST), where broadening of the cracks and new cracks were observed after exposure in WEST.
• In the last experimental campaign, WEST was equipped with shaped ITER-like PFUs. The impact of the toroidal bevel to shadow the poloidal LEs is clearly confirmed. First observations on the optical hot spots, predicted to occur in ITER due to plasma impact on PFU poloidal edges through toroidal gaps, will also be reported.

One of the most important issues for ITER operation concerns the potential damaging of divertor plasma facing units (PFUs), in particular tungsten (W) cracking, induced by plasma exposure. Pre-mature cracking and melting of the leading edge (LE) of actively cooled PFUs have already been observed in WEST experiments, running with moderate radio frequency (RF) heating power [1]. Post-exposure analysis and numerical simulation [2] have suggested that cracks initiates due to transient events such as disruptions, in which brittle failure is the most credible mechanism. Dedicated experiments were recently conducted in WEST using higher RF heating power to melt tungsten on the LE. A 2 x 9 mm² groove (depth, width) was machined on the surface of a upstream monoblock to overexpose the sharp leading edge of the adjacent downstream vertically misaligned (+0.3mm) monoblock. The surface temperature of the monoblock was monitored with a very high spatial resolution infrared (IR) camera (0.1 mm/pixel). Cracking of the poloidal LE was observed during the power increase over successive plasma discharges. The aim of this contribution is to report on the crack detection during plasma experiment and on the mechanisms leading to the crack initiation. The study will firstly present the IR data analysis and heat flux calculations to derive the peak heat flux and heat flux decay length, secondly the thermo-mechanical simulations (stress, strain fields), and thirdly the post-exposure analysis to characterize the cracks.

Five mains cracks were observed for the first time in real time via the very high spatial resolution IR camera, when the temperature of the tungsten was over 2500 °C on the LE (higher than both, the ductile to brittle transition temperature and the recrystallization thresholds). These cracks are highlighted on the top surface thanks to a local enhancement of the apparent temperature due to the cavity effect. Cracks initiates when the parallel heat flux exceeds 90 MW/m² (corresponding to injected RF power P_inj = 4.3 MW), before the discharge during which W melting occurred (parallel heat flux above 120 MW/m², P_inj = 5.5 MW). Follow-up of the IR crack detection has been performed during progressive increase of power up to the melting point. In this temperature range (from crack initiation » 2500°C, up to melting point »3400°C), W softening due to recovery and recrystallization is expected. Post-exposure analysis reveal both horizontal and vertical cracks on the poloidal LE with high crack density (<1mm between vertical cracks in poloidal direction. The vertical cracks are 28 to 77 μm wide and spread over 0.3 to 2.5mm on the top surface of the MB. Thermo-mechanical simulations is ongoing to study the failure mechanism.

[1] M.Diez and al, Nucl. Fusion 61 (2021), https://doi.org/10.1088/1741-4326/ac1dc6

[2] A.Durif and al, FED 188 (2023) 113441, https://doi.org/10.1016/j.fusengdes.2023

*Corresponding author: Quentin.TICHIT@cea.fr (Q. TICHIT)

Dedicated experiments have been carried out in the DIII-D tokamak using the Divertor Material Evaluation System (DiMES) [1] to evaluate selected high-performing advanced tungsten composites against the performance of ITER-grade tungsten (ITER-W) [2] under tokamak divertor plasmas. For the first time, samples of tungsten fiber-reinforced tungsten (Wf/W) composites [3] and micro-structured tungsten (micro-W) have been exposed in low power L-mode discharges, together with specimens of ITER-W, to study their gross erosion, surface changes and hydrogenic retention. Combined SIERRA [4] thermal modeling and IRTV measurements were used to estimate heat flux and monitor sample temperatures to avoid melting of the materials during repeat discharges and remain within the desired temperature range for retention in W. The materials were exposed to seven discharges with ~3 s dwell time and heat flux of ~0.8 MW/m² on flat samples and ~6 MW/m² on angled samples. Estimates of sample gross erosion from UV spectroscopy indicate several times lower erosion rates for Wf/W and micro-W than for the ITER-W sample. Surface changes and composition are investigated by pre- and post-exposure Scanning Electron Microscopy with Focused Ion Beam and Auger electron spectroscopy. The sample of micro-W appears to have an excellent response with minimal damage, while Wf/W sample is found to have a significant deposition of carbon on the surface, requiring further investigation. Results on hydrogenic retention measurements using thermal desorption spectroscopy (TDS) are also presented. Design of a dedicated H-mode experiment motivated by this study will be discussed for further evaluation of advanced tungsten materials.

Work supported by the US DOE under DE-FC02-04ER54698 and DE-NA0003525

[1] C.P.C. Wong et al., J. Nucl. Mater., 196–8 871 (1992)

[2] ITER Materials Assessment Report, ITER Doc. G74 MA 10 01-07-11 W 0.2, 2001

[3] J.W. Coenen et al., Phys. Scr. 96 (2021) 124063

[4] SIERRA Multimechanics Module: Aria Thermal Theory Manual - Version 5.0 (2021) doi:10.2172/1777076.

*Corresponding author: tel.: +1 858 2638438, e-mail: popovicz@fusion.gat.com (Z. Popovic)

DISCLAIMER

This report was prepared as an account of work sponsored by an agency of the United States Government. Neither the United States Government nor any agency thereof, nor any of their employees, makes any warranty, express or implied, or assumes any legal liability or responsibility for the accuracy, completeness, or usefulness of any information, apparatus, product, or process disclosed, or represents that its use would not infringe privately owned rights. Reference herein to any specific commercial product, process, or service by trade name, trademark, manufacturer, or otherwise, does not necessarily constitute or imply its endorsement, recommendation, or favoring by the United States Government or any agency thereof. The views and opinions of authors expressed herein do not necessarily state or reflect those of the United States Government or any agency thereof.

In the ITER DT phase, tungsten plasma facing material will be exposed to the burning plasma with high density helium, produced by fusion reactions, besides hydrogen isotopes. Many previous studies have shown that retention properties of hydrogen isotope in tungsten are significantly affected by helium pre-irradiation, suggesting the contribution of helium bubbles [1-3]. However, direct evidence of the effect of helium bubbles on hydrogen retention remains elusive. In this study, the effects of helium irradiation on hydrogen isotope retention in tungsten were precisely evaluated from a microscopic viewpoint using two types of high function transmission electron microscopes (TEM): in-situ ion irradiation TEM with a high-resolution quadrupole mass spectrometer (In-situ TEM-QMS) at Shimane University, and an aberration-corrected scanning transmission electron microscope combined with electron energy-loss spectroscopy (STEM-EELS) at Kyoto University.

In-situ TEM-QMS enabled quantitative analysis of gas species released from a Φ3 mm samples with irradiated area of Φ2 mm and simultaneous observation of microstructural evolution under annealing. It was shown that helium pre-irradiation causes a significant increase in deuterium retention and that most of the deuterium is desorbed without any apparent change in the bubbles up to an annealing temperature of ~800 K. STEM-EELS observation also revealed that the most of deuterium atoms are locally trapped inside the bubbles for the sample pre-irradiated with helium and post-irradiated with deuterium. In addition, deuterium dissociation from the inside of the bubble and the helium accumulation into the bubble were observed under the annealing up to ~500 K.

Quantitative data on hydrogen and helium retention in tungsten will be presented and discussed. The results clearly indicate the direct impact of helium bubbles on the hydrogen isotope retention behavior in tungsten.

[1] M. Miyamoto, Nucl. Fusion, 49, 065035 (2009)

[2] Y. Sakoi et al., J. Nucl. Mater, 442, S715 (2013)

[3] M. Ialovega et al., Phys. Scripta 2020, 014066 (2020)

In a future nuclear fusion device helium (He) will not only impinge onto the plasma-facing surfaces from the plasma but will be also created in the bulk by nuclear reactions and tritium decay. For tungsten (W), helium is known to create clusters and nano-size bubbles modifying its thermophysical properties severely as well as influencing the interaction with hydrogen isotopes (HI). This study summarizes very recent investigations on the uptake, transport, and retention of deuterium (D) in He-containing W, conducted to get a quantitative insight on the phenomena. Implications for the loss of tritium in the wall of a future DEMO reactor as well as for ITER (where a He campaign is planned, preceding the nuclear phase) will be discussed.

To quantify the influence on D uptake at the surface, He was implanted close to the surface with keV ions with different fluences and at different temperatures. Samples were then exposed to a low flux, low energy D ion beam. 20 MeV W irradiation was performed before He implantation to create defects within the first 2.3 mm. These defects trap penetrating D and make it hence possible to quantify transport with e.g. ion beam analysis methods. ³He nuclear reaction analysis (NRA) and elastic recoil detection analysis show that D gets preferentially retained where He is implanted with concentrations of up to 10 at.%. At the same time D transport beyond the He zone is vastly reduced. E.g. compared to a He-free W sample, implantation of 2×10²⁰ He/m² reduces the D uptake by a factor of 15.

To determine transport parameters of D in the bulk of He-containing W, several experimental series were conducted. In one, He was either implanted with 0.3 MeV (0.5 mm deep), with 4.25 MeV (7mm deep) or with both energies. Afterwards, samples were heated to 2000 K to anneal the displacement damage and facilitate He bubble growth. All samples were exposed simultaneously to a low flux D plasma at 370 K to decorate the He-created traps. ³He NRA depth profiling shows again, that D gets only retained where He was implanted. Thermal desorption spectroscopy (TDS) with a slow ramp of 3 K/min up to 1000 K shows two D release peaks between 400 and 800 K. Macroscopic rate equation modelling with the TESSIM-X code of these desorption spectra together with the measured D depth profiles gives de-trapping energies between 1.28 and 1.81 eV.

Similarly, He was implanted with nine different energies ranging from 0.3 to 2.75 MeV to achieve a constant He concentration between 0.5 and 4.0 µm. Samples with six different He concentrations of 10, 50, 100, 500, 1000 and 5000 appm were prepared. ³He NRA was applied between the exposures to study D transport and retention. D uptake proceeds slower into depth, the higher the implanted He concentration. The fraction of retained D per He is 7.5 at 10 ppm He and drops continuously to 0.75 at 5000 ppm He. Compared to a self-damaged W reference sample, the diffusion front is much more smeared out. Transmission electron microscopy shows nanometer-sized voids in this case whose existence can explain this observation.

Conventional fuzz formed on bulk tungsten (W) exposed to helium (He) plasma at temperatures above 1000 K increases in thickness with He fluence but saturates to a thickness of ~7 μm [1]. The fuzz layer thickness is enhanced when exposures are conducted in the presence of a W deposition source [2-3]. A 7.5 µm-thick fuzz layer developed at only 9x10²⁴ m^-2 (W/He ratio = 0.009) for deposition normal to the W surface [3], while mm-thick fuzz developed on the sample leading edge for deposition parallel to the sample surface [2]. Since W deposition fluxes of 10¹⁷ m^-2s^-1 were measured in JET-ILW [4], this work investigates fuzz layers formed under ITER-like conditions (e.g., W/He ratio ~ 10^-4).

Bulk W samples were exposed to pure He plasma at 9x10²² m^-2s^-1, 76 eV, and 1123 K in the presence of a W mesh (72% open area) positioned 25 mm upstream. With the mesh biased negatively to achieve a He incident energy of 200 eV and a W/He ratio of ~1x10^‑4 at the sample surface, the He fluence was varied from 0.1-1.7x10²⁶ m^-2. Note that W flux was calculated from mass gain on the sample assuming no re-erosion of deposited W, and He flux was measured with a Langmuir probe. The fuzz thickness was measured by cross-sectional SEM to increase linearly from 0.2 to 14 µm with increasing He fluence, in agreement with fuzz formed in Magnum-PSI by a deposition source with W/He ratio = 0.8x10^-4 [5]. Alternatively, with the He fluence kept at ~2x10²⁶ m^-2, the mesh bias was varied such that the W/He ratio at the sample varied from 0.44-6.0x10^-4. The fuzz thickness increased from 7.8 to 115 µm with increasing W/He ratio. Furthermore, a sample exposed to a mixed 0.1He-0.9D plasma at ~0.1x10²⁶ He·m^-2, 76 eV, 1123 K, and W/He ratio = 4.4x10^-4, formed a 2.4 µm-thick fuzz layer. As expected, this is much larger than the 0.2 µm-thick layer formed with a He fluence of 0.1x10²⁶ m^-2 but a smaller W/He ratio of ~1x10^-4.

In summary, fuzz up to 115 µm thick was formed on W exposed to He plasma when a W deposition source was present, much larger than the thickness of conventional fuzz. Additionally, the thickness increased linearly with He fluence and showed no indication of saturation at fluences up to 1.7x10²⁶ m^-2, in contrast to the growth rate of conventional fuzz [1]. Hence, fuzz with thicknesses greater than hundreds of microns may be expected to form in ITER, and may more significantly affect the thermophysical properties of W plasma-facing components than conventional fuzz.

[1] T.J. Petty, M.J. Baldwin, M.I. Hasan, et al., Nucl. Fusion 55, 093033 (2015)

[2] S. Kajita, S. Kawaguchi, N. Ohno, et al., Scientific Reports 8, 56 (2018)

[3] P. McCarthy, D. Hwangbo, M. Bilton, et al., Nucl. Fusion 60, 026012 (2020)

[4] K. Schmid, K. Krieger, S.W. Lisgo, et al., Nucl. Fusion 55, 053015 (2015)

[5] S. Kajita, T. Morgan, H. Tanaka, et al. J. Nucl. Mater. 548, 152844 (2021)

Work supported by US-DOE cooperative agreement, DE-SC0022528.

*Corresponding author: tel.: +1 858 534 8222, e-mail: m2patino@ucsd.edu (M.I. Patino)

Tungsten (W) is the prime candidate for the plasma facing material in present and future fusion devices. When exposed to a plasma containing He ions, W can exhibit creation of sub-surface nano-bubbles leading to formation of nano-tendrils called fuzz. Formation of W fuzz was confirmed by many laboratory experiments including He loading in linear plasma devices. There is, however, limited experience related to the parameter space for fuzz formation and re-erosion in a tokamak environment. For this reason a dedicated experiment was carried out in ASDEX Upgrade (AUG) during its 2022 He campaign to study the evolution of fuzz-like structures.

Twelve tungsten samples were mechanically polished to a mirror-like surface finish. Six of them were subsequently exposed to a He plasma in the PSI-2 linear device in order to establish a fuzz layer with a thickness of 600 – 800 nm. Before exposure in AUG, all samples were pre-characterized by Focused Ion Beam (FIB) cross-sectioning. In total 48 FIB cross-sections with line markings for quantification of local erosion, deposition and fuzz formation were prepared – 4 on each sample. The samples were placed in two parallel poloidal rows spanning a range of 20 cm around the outer strike line position (OSP). They were subsequently exposed in AUG to a series of 14 consecutive discharges, 8 in H-mode and 6 in L-mode.

Detailed analysis by means of electron microscopy revealed on the samples regions of erosion, deposition and fuzz formation. Below the H-mode OSP, homogeneous co-deposits containing W and O, with a thickness up to 400 nm, were found. In the close vicinity of the H-mode OSP, significant erosion of pre-exposure PSI-2 fuzz was observed. The erosion reached up to 100 – 250 nm, depending on the location. In addition, the initially polished samples did not show any newly formed fuzz in that zone. Above the H-mode OSP, new fuzz was formed with a thickness of up to 800 nm. Pre-exposure PSI-2 fuzz was either removed or modified. Below and near the L-mode OSP erosion below 80 nm was observed. Above the L-mode OSP co-deposition of W and O, with thickness reaching 200 nm, was identified. The co-deposition covered the pre-exposure PSI-2 fuzz without visible damage to its structure.

Additionally, arc traces, mostly at fuzzy surfaces, were found. In the arc track, the fuzz layer was completely removed, however without any visible damage of the underlying material.

Solid plasma facing components (PFCs) in divertor high heat flux areas are vulnerable to damage, particularly during off-normal events, which could ultimately lead to component failure. Liquid metals such as tin confined in a capillary porous structure (CPS) could provide a solution, as lost material can interact with and cool the plasma while also being replenished by capillary transport. However, due to the high Z number of Sn (50), only a limited fraction is acceptable in the plasma core. It is therefore vital to test this concept in a tokamak environment, particularly in a diverted plasma, which has not previously been done.

To that end, a test module was additively manufactured using Laser Powder Bed Fusion where a porous 1.5 mm W layer was directly attached to a solid W base with a plasma facing area of 16×40mm². The porous layer was then wetted and filled with 1.54 g of Sn. Prior to exposure in ASDEX Upgrade (AUG), it was exposed in the high heat flux facility GLADIS to determine its thermomechanical properties and to confirm its survival under the heat load expected during plasma exposure. The module was flush mounted in a dedicated target tile on the AUG divertor manipulator [1] and electrically pre-heated above the melting point of tin before each discharge. The diverted plasma was established with the outer strike point (OSP) above the module. During plasma flat-top the OSP was shifted down onto the module and kept there for a time period of 2 to 3.4s. After each discharge, pictures of the tile surface were taken, which revealed Sn leakage from the module after H-mode exposure, mainly from the downstream poloidal edge. Sn I line emission in the plasma close to the target was observed as soon as the OSP reached the module surface. The spectroscopic measurements indicated an acceptable tin erosion flux near the module, in good agreement with HeatLMD [2], a code for liquid metal surface and plasma interaction, which predicts that sputtering dominates over evaporation of tin. In contrast, the 1.5D-impurity transport STRAHL [3], verified against foil bolometers, which revealed a Sn concentration in the plasma core of up to 1.4×10^-4. This would be an unacceptably high concentration of Sn for a burning fusion plasma, especially since only 0.17% of the strike line was covered by the CPS in AUG. The high Sn fraction in the plasma core despite comparatively low tin erosion flux hinted at ejection of Sn droplets from the surface as the main tin source. This was confirmed by post-exposure analysis, revealing a substantial amount of Sn droplets on neighbouring tiles. Droplet ejection was observed previously in laboratory studies [4,5]. To enable CPS applications in future fusion reactors, further research and development is required in order to prevent macroscopic leakage and droplet ejection of liquid tin.

[1] A. Herrmann, et al., Fusion Engineering and Design 98–99 (2015)

[2] J. Horacek, et al., Nuclear Materials and Energy 25 (2020)

[3] R. Dux 2006 STRAHL User Manual Tech. Rep. 10/30 IPP

[4] A. Manhard, et al., Nucl. Fusion 60 (2020) 106007

[5] W. Ou. et al,. Nucl. Fusion 61 (2021) 014025

The Monte-Carlo code ERO2.0 traces impurity particles in a hydrogenic plasma background throughout the volume of fusion devices. It provides the local erosion and deposition fluxes at plasma-facing components or recessed objects, delivering important information about material sputtering or layer growth on those components. In recessed areas, as for example mirror assemblies in the diagnostic first wall (DFW) of ITER, the code is approaching its limits. The necessary resolution of information on mirrors far inside the assemblies cannot be achieved with standard simulations, as only a tiny fraction of impurity test particles and a large fraction of charge exchange hydrogenic neutrals reaches this volume located more than 50 cm away from the last closed flux surface.

A novel multi-staged approach of ERO2.0 simulations is employed to overcome this challenge. Impurity particles from a first global ERO2.0 simulation with its boundary close to the DFW are collected and subsequently injected into second and third stage local simulations covering only the mirror assembly volume. Importantly, the number of test particles representing the physical atom fluxes entering this region is scaled up at the interfaces between the simulation stages, to achieve much better resolution in the relevant distant regions. By using this approach, the computational power is focused on the test particles that are relevant for the impurity particle fluxes on the mirrors far inside the assemblies.

The full multi-stage approach of simulations has been conducted with three plasma backgrounds, namely a low-density L-mode case from the ITER pre-fusion operation phase with H fuel, and high- and low-density H-mode cases from the operation phase with D-T fuel. The results show that the sputtering of the mirror surfaces in each case is largely dominated by high energy hydrogenic charge exchange neutrals, while the patterns of the fluxes seem to be strongly influenced by the geometry of the assembly. In both the upper port and the equatorial port, overall negligible deposition mainly of housing material of the mirrors can be found on the first mirror considering the expected full operational time of ITER.

Beryllium is an essential material for a wide variety of nuclear application such as material testing fission reactors, fusion energy experimental and future commercial reactors, and target component materials of particle accelerator sources. In comparison to other nuclear materials, beryllium experiences extremely high transmutant helium accumulation rate. Both helium accumulation and displacement damage effects in beryllium are known to be highly irradiation-temperature dependent, and often lead to drastic changes of beryllium properties. Therefore, radiation damage and transmutant accumulation responses of beryllium and beryllium-based alloys are under investigation by fusion energy and other nuclear facility communities. Current work summarises results of a set of micromechanical and microstructural studies, carried out on beryllium samples irradiated by neutrons, high energy protons and on beryllium samples implanted in an ion accelerator with He ions.

Low temperature irradiation at 50°C to a dose of 0.04 dpa (160 appm of He) led to fracture mode change from transgranular cleavage to grain-boundary cracking. After irradiation, microcantilevers fracture tests showed significant increase of fracture load with almost complete loss of ductility. Nanoindentation tests demonstrated that irradiation induced hardening after 50°C implantation is about 2 times higher than after 200°C (0.1dpa, 2000appm of He). Analysis demonstrated, that observed point defect clusters (so-called “back dots”) created a low temperature irradiation (≤200°C) could only lead to about half of the measured hardness increase, while the rest of the hardening should originate from helium bubbles with the size below the TEM resolution (<2nm).

Microstructural changes in higher temperature (up to 650°C) neutron irradiation (up to 37 dpa, 6000 appm of He) are mainly manifested in the form of the high number density of He bubbles [1]. Crystallographic analysis demonstrated that high-angle grain boundaries favour creation of significantly larger bubbles, that leads to increased local microscopic and eventually total macroscopic swelling. For example, in the same sample irradiated at 600°C, local swelling of about 5% was measure inside grains (nanometric size bubbles) and up to 40% in the areas of high density of high-angle grain boundaries (micrometric size bubbles). Atom Probe Tomography investigations of the neutron irradiated samples showed presence of new isotopes with the mass-to-charge ratio appropriate for tritium. This transmutation product was highly heterogeneously distributed and has strong tendency to segregate to impurity precipitates and helium bubbles. This observation contributes to understanding of mechanism of the high tritium retention in beryllium at temperatures below 600°C.

[1] Klimenkov M, Vladimirov P, Jäntsch U et al. Sci Rep. 2020 May 15;10(1):8042.

*Corresponding author: tel.: +447474240269, e-mail: slava.kuksenko@ukaea.uk (V. Kuksenko)

Future generation tokamaks, including ITER, will strive for better confinement, which means a higher fusion plasma current to achieve a good Fusion Efficiency Factor value greater than 10 [1]. One of the major threats with this scenario is a potential for increased damage due to unmitigated plasma disruptions. Two possible consequences of disruptions are the high thermal loads with fast melting and electromagnetic forces acting on the molten layers as previously seen in JET [2] and TEXTOR [3] and secondly the high energy Runaway Electrons (RE) as reported in JET [4]. Unique data on the plasma-facing components, particularly the inner-wall beryllium tiles of the JET device following the impact with the REs is presented. The magnitude of the impact was firstly noted during in vessel inspection surveys periodically carried out in JET and later by high-resolution images taken during a shutdown period, when it was possible to deploy a camera into the vessel using a remote handling robotic arm. Further investigation of one RE damaged tile retrieved from JET during the 2017 shutdown campaign is presented here. The impact of RE damage was measured using 3D profiling to generate a surface map, revealing beryllium material dispersion and accumulation arising during melting. Initial assessment of the profiling data revealed a change in material height between 1 to 2 mm; material which was either dispersed generating “valleys”, or accumulated generating “hills”. Using the DINA-SMITER-MEMOS-U [5] together with GEANT4 toolkit, modelling of this damage was performed, revealing a very good agreement with the post-mortem measurements [6]. Optical microscopy to assess the damage depth was performed after tile cutting, revealing molten layers up to 1-2 mm thick and evidence of material bridging gaps between adjacent castellations. No accumulation of molten material in between the gaps was observed. In addition to the optical microscopy data, further analysis to evaluate morphological and structural changes and fuel retention in the damaged regions will be presented to provide a detailed assessment of RE damage on beryllium components. The work will inform assessments of the impact or RE damage in ITER and provide benchmark data for modelling.

[1] R. Pitts et al, Physics World 19: 20–26 (2006)

[2] I. Jepu et al, Nucl. Fusion 59 086009 (2019)

[3] G. Sergienko et al., Phys. Scr. T128 81–86 (2007)

[4] C. Reux et al Nucl. Fusion 55 093013 (2015)

[5] J. Coburn et al. Nuclear Materials and Energy, 28, 101016, (2021)

[6] L. Chen et al, 5th Asia-Pacific Conference on Plasma Physics, 26.09-1.10.2021

*Corresponding author: tel.: +44 1235 466208, e-mail: ionut.jepu@ukaea.uk (I. Jepu)

Recent experiments on the DIII-D tokamak evaluated the efficacy of various impurity control techniques (pellet pacing, gas puffing, resonant magnetic perturbations (RMPs), powder injection, and impurity seeding) to reduce W sourcing from the W-coated Small Angle Slot (SAS-VW) divertor into the main plasma. The most significant immediate findings from these studies were as follows:

(a) Tungsten sources from this slot-like divertor were systematically lower relative to analogous DIII-D experiments [1] conducted with a tungsten-clad "open" divertor;

(b) RMPs represented the most successful technique for W impurity control due to the complete suppression of edge-localized modes (ELMs) in certain scenarios- the first such observation with tungsten plasma-facing components (PFCs) in DIII-D.

Demonstration of a scenario with reduced high-Z leakage and without damaging edge transients and power exhausted into a slot-like, tungsten divertor represents notable progress towards an integrated core-edge solution for ITER and other future devices with high-Z PFCs. Specifically, the application of n=3 RMPs resulted in a decreased W source relative to the no-RMP cases at fixed pedestal conditions, regardless of ELM suppression or mitigation condition. ELMs apparently dominated W sourcing in SAS-VW because decreasing ELM frequency at constant ELM size resulted in substantial decreases to the W source. Linear increases in W source with ELM frequency observed during pellet injection were also consistent with this picture [3]. During injection of low-Z solid and gaseous impurities, W sourcing and leakage were highly sensitive to impurity injection rate, species, and location due to differences in the plasma response [4]. For example, when the Ne injection rate increased far from the target, the amount of core W decreased due to the decreasing target electron temperature, reducing sputter sources. When Ne was injected at similar rates within the slot itself, a similar decrease in tungsten erosion was not observed due to inefficient target cooling. During both B and BN powder dropping, the W source level decreased due to an increase in radiative dissipation and a corresponding decrease in the overall particle flux to the target.

Work supported by US DOE under DE-FC02-04ER54698, DE-FG02-07ER54917, DE-AC05-00OR22725, DE-SC0019256, DE-SC0023378, and DE-AC02-09CH11466.

[1] E.A. Unterberg et al., Nucl. Fusion 60 (2020) 016028

[2] T. Abrams et al., Phys. Scr. 96 (2021) 124073

[3] A. Cacheris et al., this conference.

[4] M.S. Parsons et al., this conference.

The EIRENE neutral transport solver is an established tool for edge and divertor fusion-relevant plasma modelling, a critical part of SOLPS-ITER (B2.5-EIRENE), EMC3-EIRENE and other comprehensive SOL/divertor codes. The focus of this work is on the validation of reaction rates in the EIRENE-related collisional radiative models (CRMs), in view of the prominent role of atomic and molecular processes during detachment [1] and of the necessary massive data extension to provide resolution by rovibrationally states. The dramatic increase of data amount and complexity requires improved tools for visualization, processing and quality/consistency control. The toolkit HYDKIN is being restructured to make it a standard pre-processing tool for EIRENE. A CRM for the decay of hydrogenic molecules is applied to JET and EU-DEMO benchmark cases and compared with EIRENE simulations. The equilibrium concentrations of neutrals are computed with HYDKIN and the established YACORA CRM flexible solver [2] in view of their different capabilities. The relevance of H₂ and H₂⁺ dissociation processes is strongly dependent on plasma temperature. Heavy particle collisions are the main H₂ dissociation channel at low temperature T_e<3eV, resulting in MAD and MAR chains, and ladder-like transitions provide the main populating/depopulating mechanisms for vibrational excitations of molecular electronic ground state (X), resulting in a Boltzmann-like distribution. Resolving the vibrational states was demonstrated to affect even total dissociation rate of ≃ 40%. A model for electronically excited states is defined, including vibrationally resolved excitation rates, derived from MCCC database [3], and transition probabilities for spontaneous decay [4]. The vibrationally resolved population coefficients for the excited state d³Π_u are shown and their dependence from X vibrational temperature (T_vibr) is outlined. Since they determine the intensity of the Fulcher band (d³Π_u àa³Σ⁺_g), most used band for diagnostic purposes, they can reveal prominent spectroscopic features characterising plasma regime. In this respect, it is foreseen the application to a parameter scan of plasma conditions with respect to the degree of detachment that is critical for real-time control of EU-DEMO operation.

[1] K. Verhaegh, B. Lipschultz, J.R. Harrison, et al., Nucl. Fusion 63, 016014 (2023)

[2] D. Wünderlich, M. Giacomin, R. Ritz, et al., J. Quant. Spectrosc. Radiat. Transfer 240, 106695 (2020)

[3] mccc-db.org

[4] D. Wünderlich, U. Fantz, Atomic Data and Nuclear Data Tables 92(6), 853 (2006)

*Corresponding author: tel.: +49 2461 618661, e-mail: f.cianfrani@fz-juelich.de (F. Cianfrani)

Many linear divertor simulators are widely used as basic research devices because of their ease of measurement and control. Most divertor simulators, however, cannot reproduce the actual detached plasma in the divertor due to the fact that the ion temperature of the simulated plasma is lower (a few eV) than that of the divertor plasma in a fusion device (over 10 eV). The main objective of this study is to investigate the effect of ICR heating on the production of detached plasmas in the expanding magnetic field region using a linear divertor simulator (TPDsheet-U) because a magnetized sheet plasma can be efficiently heated by ion cyclotron resonance (ICR) method [1,2], RF power in the ion cyclotron frequency band (1.0-1.3 MHz) was applied to a high-density hydrogen sheet plasma (ne~10¹⁹ m-3, Te~10 eV), and the energy of the heated plasma was measured using a magnetic-loop-coil for ICR heating effect. The detached plasma formation, such as the electron density, the electron temperature, and Hγ/Hα have been analyzed, using the Langmuir probe and the visible spectroscopy, respectively. As the RF power of ICR heating increases up to 500 W, the ion temperature is increased from 2.5 eV to 6.4 eV at the magnetic field of 0.08 T. It was observed for the first time that the electron temperature of detached plasma increased and the Balmer series emission intensity ratio became smaller when the ion temperature was increased by applying a resonance frequency ω_RF, which is about 1.2 times the ion cyclotron frequency ω_ci of hydrogen ions. This suggests that increasing the ion temperature causes a transition from a detached plasma to an attached plasma.

[1] Y.Ohara, et al., J. Plasma Fusion Res. SERIES, 8 (2009)pp888-892.

[2] T. Takimoto et al., Nucl. Mater. and Energy 19 (2019) pp352-357.

Diagnostics and control of edge/divertor plasmas is an important issue in fusion reactors. In DEMO, where access to the device will be much more limited than in ITER, optical emission spectroscopy (OES) will be a powerful tool for plasma diagnostics. A helium (He) line intensity ratio method has been widely used to measure the electron density, n_e, and the temperature, T_e, from a comparison with the collisional-radiative model (CRM). However, significant errors can occur because of some ambiguities in modeling the population distribution due to the effects of radiation trapping, transport of metastable states, etc. [1]. The recent application of machine learning to OES data in PISCES-A [2] (0.28 x10¹⁸ < n_e (m^-3) <3.8x10¹⁸) and Magnum-PSI [3] (2x10¹⁸<n_e (m^-3) <8x10²⁰) is promising. In [3], the use of neural network (NN) was found to reduce the measurement error to ~10%. In addition, in-Magnum-PSI, machine learning has been used to predict n_e and T_e from key machine settings [4]. In this study, we applied machine learning to OES and machine setting data from the linear plasma device NAGDIS-II to explore the potential application of machine learning for predicting n_e and T_e.

Pure He plasmas are used to build a data set, because of their simplicity as a first application of machine learning in NAGDIS-II. A spectrometer (HR2000, Ocean Insight) was used to measure the plasma, and an electrostatic probe was used to measure n_e and T_e. Measurements were made at five to seven different radial positions (0-18 mm from the center) for both the spectroscopy and the probe. By varying the discharge current (20, 40, 60, and 80 A), magnetic field strength (0.05, 0.1, 0.15, and 0.2 T), and He flow rate (200, 300, 400, and 500 sccm), a total of ~420 data points were prepared. Eleven identified lines at 388.9, 402.6, 438.8, 447.1, 471.3, 492.2, 501.6, 504.7, 667.8, 706.5, and 728.1 nm are used; a strong emission at 587.5 nm was eliminated due to the detector saturation. Covered ranges of n_e and T_e are 0.3-7.1 eV and 3.6x10¹⁷-2.4x10¹⁹ m^-3, respectively. A three hidden layer neural network (NN) is introduced to model the relationship between n_e/T_e and the combination of line intensities, radial position, and discharge data. The residual errors of the predicted values from the NN were 19% for n_e and 21% for T_e when using the emission data and radial position. By additionally including the discharge current, magnetic field strength, and He flow rate, the errors decreased by several % both for n_e and T_e. It is likely that the prediction error of ~20% is almost consistent with the measurement error in the probe.

[1] S Kajita, G Akkermans, K Fujii, et al., AIP Advances 10 (2020), 025225.

[2] D Nishijima, S Kajita, GR Tynan, Rev. Sci. Instrum. 92 (2021), 023505

[3] S Kajita, S Iwai, H Tanaka, et al. Nucl. Mat. Energy 33 (2022), 101281.

[4] H. J. N. van Eck, G. R. A. Akkermans, S. Alonso van der Westen et al., Fusion

Eng. Des. 142, 26 (2019).

Langmuir probe measurements in neutral-beam heated, low-confinement mode
plasmas in JET ITER-like Wall (JET-ILW) show that recycling in helium (He) is
significantly lower than in otherwise identical deuterium (D) plasmas; in high-recycling
conditions, the ion currents were a factor of five lower in helium than in deuterium.
This finding is further corroborated by fivefold lower atomic He than molecular D
pressures in the JET sub-divertor. These results were consistently obtained in two
divertor plasma configurations with the high-field side (HFS) strike line connected
to the HFS vertical plate and the low-field side (LFS) divertor strike line
connected to either the LFS horizontal (VH) or vertical plates (VV).

For the VH divertor plasma configuration, the experiments show that the core plasma
density required to reduce the ion current to the divertor target plates, which in D
plasmas is associated with detachment, was marginally (approx. 10%) higher in He
than in D plasmas, for both the HFS and LFS target plates. For the VV configuration,
the ion currents to the LFS target plate also peaked at approximately 10% higher
core plasma density in He than in D plasmas, but at the HFS target, 20% higher core
plasma density was required in He than in D plasmas. These observations are
consistent with those measured in ohmic confinement for hydrogen and deuterium [1]
and are indicative of atoms and molecules more efficiently connecting to the
cryogenic pump via the LFS pump duct in the VV configuration.

When raising the total input power from 2 MW to 7 MW, the He plasmas remained
in low-confinement mode and did not enter high-confinement mode. The core
plasma density required for saturation of the ion currents increased by 30%.
Furthermore, for auxiliary power levels below 5 MW, the reduction of the ion current
was observed as nearly bifurcated and accompanied by a move of He radiation from
the divertor into the core region. At 5 MW, the reduction of the ion currents is more
gradual, thus like previous observations in D plasmas. Bolometry measurements
indicate that the peak radiation moves from the HFS divertor to the LFS divertor and
X-point region as the core plasma density is increased.

This contribution will further compare target profiles of the ion saturation
currents, electron temperature and density for otherwise identical D and He
plasmas.

[1] V. Solokha et al., Nuclear Materials and Energy 25 (2020) 100836.

In the facing wall of high heat flow plasmas such as divertors, hydrogen recycling releases vibrationally excited hydrogen molecule from the plasma facing material. Therefore, it is important to measure the vibrational temperature of hydrogen molecule Tvib in the electronic ground state with respect to changes in the temperature and bias voltage of the facing wall to indicate elementary atomic and molecular processes. The purpose of this study is to observe the vibrational temperature Tvib of hydrogen molecule in response to thermal electron emission and bias voltage changes of an electron emitter installed in a hydrogen plasma by means of vacuum ultraviolet (VUV) emission spectroscopy. In the experiment, an electron emitter was placed in a hydrogen plasma generated by a linear divertor simulator (TPDsheet-U) to measure the VUV spectra (80 nm to 150 nm) of hydrogen molecule versus thermal electron emission and bias voltage change. The vibrational temperature of hydrogen molecule was determined by fitting the relative intensities of the experimental spectra from the VUV spectrum measurements to the synthetic theoretical Lyman and Werner band spectra [1]. At a discharge current of 50 A and a gas pressure of 0.3 Pa, we observed that Tvib gradually increases from 3000 K to 6000 K as the electron emitter is heated and the bias voltage is further varied from the floating potential to -30 V. This indicates that the temperature and bias voltage of the facing wall affect Tvib in the plasma.

[1] A.Nakanowatari, at al., J. Nucl. Mater., 390–391 (2009) 311–314.

Analysis of the magnetic sheath is necessary to understand the energy and direction of motion of ions that cause plasma surface interaction in the presence of a magnetic field. Under the conditions that gyro-motion can be impeded by collisions, in particular, nuclear fusion reactor diverter under the detachment condition or a magnetron, the effect of collision changes the incidenct angle of the ions. In this study, the angle of incidence of ions incident on the surface is measured using a graphite mateiral probe under oblique magnetic field conditions, and the effect of collision on the angle of incidence of ions in the magnetization sheath is analyzed. The experiments were carried out by varying the magnetic field angle (ψ), with respect to the surface normal, from 0° to ~ 90° in a weakly collisional hydrogen plasma. The hydrogen ions actively reacted with carbon, leading to the formation of conical tips with axes directed along with the incident ion flux direction. The ion contact angle was obtained via the angle analyses of the etched graphite target images taken by scanning electron microscope. The measured angles were compared to those calculated using fluid magnetic sheath model[1]. In this model, the ion motion is assumed in the presheath region with a thermal ion flow along the magnetic field direction. The corresponding results showed decrease of ion incident angle by collision and the collisionality in magnetic sheath defined by the ratio between gyro radius and collision mean free path. Accordingly, in weakly collisional conditions, collisions increase E×B drift, but reduce the confinement effect in the magnetic field direction, and consequently change the drift velocity of ions in a direction close to the E-field direction (surface normal direction). This study shows that the collisional property of the ions is crucial to guide the ion motion in magnetic (pre)sheath and to determine the ion incidence angle at the material surface. (collision mean free path/ion source gyroradius).

Commonwealth Fusion Systems (CFS) is currently constructing the SPARC tokamak while in parallel designing the first ARC fusion power plant. Two of the key aspects of the CFS strategy for fusion energy are the use of our custom built high temperature superconducting magnets and a liquid immersion blanket design. Both of these have a large impact on the requirements and design of the plasma-facing components (PFCs) for our power producing ARC devices [1]. For SPARC, which will not have a blanket, the PFCs are fabricated from unalloyed tungsten in the highest heat flux regions and a tungsten heavy alloy in lower heat flux regions of the tokamak. In both cases, the tungsten-based components that directly face the plasma are fastened to a Nitronic-50 steel structure. With over 17,000 individual, inertially-cooled PFM components, coupled with high heat loads, thermal management is critical to the SPARC lifetime. A combination of optical and solid component diagnostics will monitor the PFM condition and inform SPARC operations. Additionally, high electromagnetic loads drive the size, shape, and fastening mechanisms of the PFC assembly. SPARC will explore ARC-relevant dissipative divertor conditions that will retire risk for ARC operations and inform the design of the ARC divertor.

ARC will have several new systems that will not be present on SPARC, including the FLiBe liquid immersion blanket, balance of plant power generation, and tritium separation and recycling systems. Because ARC must demonstrate power production and tritium breeding, the requirements for the PFCs are significantly different from those on SPARC. For example, ARC PFCs will have active FLiBe cooling (versus inertially cooled PFCs in SPARC) and thus will operate above 459°C, the melting temperature of FLiBe. At this stage in the ARC design, we have a shortlist of candidate plasma-facing and structural divertor and first wall materials. To ensure good enough thermal contact, and a thin enough structure for proper tritium breeding, the plasma-facing material will have to be monolithically joined to the underlying structural material. Research and development is currently ongoing for joining the different materials that will make up the ARC PFCs. The CFS PFC development roadmap includes ion irradiation, high heat flux testing, select neutron irradiation of candidate materials, and joining development. The design of the SPARC PFCs, strategy for ARC PFCs, and materials development roadmap will be presented.

[1] A.Q. Kuang, N.M. Cao, A.J. Creely, et al., Fusion Eng. Des. 137, 221–242 (2018)

*Corresponding author: tel.: +1 781-583-3100, e-mail: lgarrison@cfs.energy (L. M. Garrison)

Wall conditioning, which includes various conditioning methods [1], is essential for the operation of tokamaks and stellarators, and thus for the advancement of fusion research. For removing impurities accumulated at the plasma-facing surfaces and controlling the recycling of hydrogenic fuel fluxes at fusion devices, ion cyclotron wall
conditioning (ICWC) is used [1-3]. The present work compares the ICWC discharge in hydrogen (H) and helium (He) in JET. One monopole phased antenna was used to produce the uniform conditioning plasmas. The generated radio frequency (RF) power was up to 0.45 MW at a frequency of 29 MHz. With the toroidal magnetic field at 1.9 T, the ion cyclotron resonance (ICR) zone for the hydrogen fundamental harmonic is located at the torus axis, while it is outside the plasma boundaries for He²⁺ and He⁺. In H-ICWC discharges, a lower RF coupling efficiency was observed than in He-ICWC discharges, respectively up to 0.4 and 0.58. Accordingly, the
coupled RF power in the H discharge was lower than in the He discharge, up to up to 0.16 MW and 0.24 MW respectively. Measurements of the line-integrated electron density along the vertical chord in the H-ICWC discharge showed that the density is below the accuracy of the interferometer of 3x10¹⁷ m^-2. In the case of the ICWC discharge in He, the density was higher than 3x10¹⁷ m^-2. Video diagnostics showed significant differences for ICWC discharges in H and He. Shortly after the discharge initiation in He, the presence of horizontal filaments near the low field side was observed, suggesting that initially, the plasma flows from a localised plasma production zone along the barrel shaped magnetic field configuration. Also, a more intense glow in the ICR zone throughout the discharge in He is observed. Both features are not observed in H. The results are analysed and discussed using spectroscopic measurements and RF wave physics.
[1] T. Wauters, D. Borodin, R. Brakel, et al., Plasma Phys. Control. Fusion 62, 034002 (2020)
[2] E de la Cal and E Gauthier, Plasma Phys. Control. Fusion 47, 197 (2005)
[3] D. Douai, A. Lyssoivan, V. Philipps et al., J. Nucl. Mater. 415, S1021 (2011)

Corresponding author: tel.: +38 097 0137785, e-mail: kovtuny41@gmail.com (Y. Kovtun)

WEST is an actively cooled, long-pulse tokamak designed to study plasma operation with tungsten (W) plasma facing components (PFCs) in the divertor, in preparation for long-pulse operations on W divertor devices like ITER [1]. For long-pulse operation, the W impurity production and transport to the core plasma are critical concerns that require further measurement and interpretation in order to improve plasma performance and PFC lifetime. In low confinement mode (L-Mode) operation, the sputtering of tungsten is not expected to be driven by the plasma itself (D), but by the main light impurities (B, C, N, O). A refined light impurity content estimate at the WEST lower divertor was performed in L-Mode discharges, with an electron temperature range going from 1 eV to 30 eV (at the strike point), and in the high recycling regime (~ 10¹⁹ m^-3). The estimated W sputtering at the lower divertor including only O and C contribution could not match the visible spectroscopy measurements of neutral W production [2]. A better agreement is found with B, C, N, O impurities, for this specific discharge. Results show that B and C are the main contributors with 63% and 26% of the total W sputtering respectively. This work proposes the use of OEDGE-DIVIMP plasma background constrained with (n_e, T_e) from flush-mounted Langmuir probes located at the lower divertor and a reciprocating probe at the top of the torus [3]. The impurities launched are based on visible spectroscopy to model the transport in the divertor region. The radiative power is then calculated with ADAS and a synthetic diagnostic (SYNDI) is used to create synthetic bolometry measurements. These synthetic bolometry measurements are compared to experimental measurements from the WEST horizontal bolometer, which is used to resolve radiative power in the divertor region. This workflow is designed to obtain agreement between synthetic data and bolometry measurements, which can provide verification opportunities for the theory and modelling workflow for impurity transport in the edge plasma and improve the refined light impurity content estimates used in the divertor region.

*Work supported by US DOE under DE-SC0020414.

[1] J. Bucalossi et al., Fusion Eng. Des., vol. 89, no. 7, Art. no. 7, Oct. 2014, doi: 10.1016/j.fusengdes.2014.01.062.

[2] A. Grosjean et al., Nucl. Mater. Energy, 2023 (PSI issue)

[3] J. B. Maeker et al., Nucl. Mater. Energy, vol. 33, p. 101309, Oct. 2022, doi: 10.1016/j.nme.2022.101309.

*Corresponding author: tel.: +33 658136463, e-mail: agrosjea@utk.edu (A. Grosjean)

The final KSTAR target plasma performance with the planned maximum additional heating capability (~24 MW) will produce stationary divertor target heat fluxes far beyond those tolerable (~4.3 MW/m²) by the current graphite targets. The present lower divertor structure is thus being completely replaced by an actively cooled, tungsten (W) monoblock, cassette based system which bears some similarities to that now being procured for ITER.

The new W divertor system is designed to accommodate peak stationary target heat fluxes of £10 MW/m². It consists of 64 stainless steel (SS) cassette modules carrying plasma-facing units comprising W monoblocks bonded to CuCrZr cooling pipes via a Cu interlayer, as in the ITER design. The thermal stability of the entire divertor module at the peak heat flux of 10 MW/m² has been thoroughly examined by the CFD analysis [1].

Simple estimates, supported by modelling with the SOLPS plasma boundary code, show that the 10 MW/m² peak stationary power handling limit can easily be exceeded in 1 MA, H-mode KSTAR discharges with 24 MW of input power, so that divertor detachment will be mandatory under such conditions. The code simulations suggest that the required detachment can be achieved, for example, with nitrogen impurity seeding, but this will require the implementation of reliable detachment control schemes. For this reason, significant effort has been invested in the area of detachment control in the campaigns preceding the current machine shutdown in which the divertor is being replaced. Real-time control of divertor radiation in ELMing H-mode plasmas has been successfully demonstrated over timescales > 20 s by controlling the nitrogen injection rate using Langmuir probes as a monitor of the detachment state [2].

To enable detachment control and provide key measurements for divertor performance assessment, the W divertor is being accompanied by a new and comprehensive set of diagnostics, together with an improved gas puffing system. The former includes Langmuir probe arrays, neutral pressure measurements, IR thermography, divertor visible spectroscopy, divertor Thomson system and ITER-like cassette body mounted shunts for the monitoring of thermocurrents. The paper will provide an overview and status of the new divertor and associated systems.

The investigation of neutron-irradiated, Tritium- or Beryllium-containing plasma-facing-materials (PFMs) is of great interest for the investigation and prediction of PFM performance for ITER and future fusion reactors. Currently, only a few plasma facilities can handle such materials on a smaller scale outside of a fusion reactor. JULE-PSI at Forschungszentrum Jülich is a new linear plasma device, capable of investigating the plasma surface interactions of such PFM samples [1].

After the design and construction phase, JULE-PSI is currently setup on a test-stand outside the hot cell environment, with the exposure chamber temporarily replaced with a segmented plasma dump [2]. The implementation of the diagnostic systems is a critical step towards the use of JULE-PSI for plasma-wall-interaction studies: Especially the optical diagnostics are directly required for the in-situ analysis of plasma-wall-interaction processes such as erosion and desorption. Secondly, the plasma needs to be monitored, both for the commissioning phase and later during the operational phase.

The two main plasma diagnostics available at this stage are an optical emission spectrometer (OES, wavelength range 300-890 nm) and a single tip Langmuir probe. Both diagnostics have already been extensively used at the linear plasma device PSI-2, with the design and parameters of both systems installed at JULE-PSI being largely identical to the respective systems at PSI-2.

In this work, we will present the Langmuir probe and OES diagnostics used for the initial characterisation and optimization of the JULE-PSI plasma parameters. First results of electron temperature, electron density, ion flux density, plasma potential and plasma composition (impurities) measurements will be shown and discussed, also in comparison to existing devices such as PSI-2. In addition, we will compare both diagnostic methods directly and investigate if the OES spectroscopy can substitute the Langmuir probe for measuring certain plasma parameters. Since the OES is a passive diagnostic with few moving parts, this would be beneficial for the future operation in a hot-cell environment. The interpretation of the results will be supported by additional measurements from the segmented dump plate and a quadrupole mass spectrometer.

[1] B. Unterberg et al., Fusion Eng. Des. 86 (2011)

[2] R. Rayaprolu et al., 19^th PFMC Conference (2023)

[3] A. Kreter et al., Fusion Sci. Technol. 68 (2015)

*Corresponding author: tel.: +49 2461 615638, e-mail: m.reinhart@fz-juelich.de (M. Reinhart)

Next generation fusion devices will operate with long-pulse discharges in which tritium retention in the first wall and erosion of the first wall will be limiting factors [1]. During these discharges, plasma-facing-materials (PFM) will be exposed to a combination of high-energy neutrons and plasma particles. The energetic neutrons induce transmutation leading to chemical compositional changes and lattice damages in the PFMs. The damage is further compounded by plasma exposure, which saturates the lattice damage leading to an increase in retention, as shown by plasma exposure, post-ion-irradiation [2].

JULE-PSI is a new linear plasma device built for a hot-cell environment at Forschungszentrum Jülich. The device is designed to investigate plasma surface interactions on tritiated samples and neutron-irradiated PFMs [3]. The radioactive samples are introduced using a sample exchange chamber and exposed to deuterium, hydrogen or helium plasma. The device operation, sample exposure control, plasma characterisation and large portion of the post-exposure analysis will be done remotely.

Presently, JULE-PSI is setup on a test-stand outside the radiation area, with a heated disk LaB₆ cathode and a hollow anode in a linear vacuum vessel. The exposition chamber placed at the other end of the vacuum vessel is currently replaced with a segmented plasma dump, consisting of 59 sectorial segments in order to characterise the plasma. The plasma dump can be biased and used to measure the plasma current and temperature. Furthermore, a Langmuir probe and optical spectroscopy are installed on JULE-PSI to corroborate the plasma characteristics.

In this contribution, an update of the set-up of the nuclear-linear plasma device JULE-PSI along with the plasma characteristics will be presented. Furthermore, the hot cell implementation of the diagnostics which already installed on JULE-PSI and to be available in the future will be described. Lastly, measurements made using the segmented plasma dump will be examined and checked against expected values.

[1] Ch. Linsmeier et al 2017 Nuclear. Fusion 57 092007

[2] S. Möller et al 2020, Nuclear Materials and Energy 23 100742

[3] B. Unterberg et al 2011 Fusion Engg. And Design 86

[4] L. Scheibl et al 2015 Fusion Engg. And Design 98-99

Simulations comparing graphite to silicon carbide (SiC) walls in DIII-D indicate that SiC walls could lower Z_eff and thus improve core plasma performance in DIII-D. Additionally, SiC walls may “self-condition” and provide wall conditions similar to that of siliconization. SiC sputters ~5x less carbon than graphite, and even less silicon; Y_Si^SiC / Y_C^SiC ~ 0.1-0.01. Silicon ionizes rapidly in SOL plasmas, thus increasing the probability of prompt redeposition via rapid parallel transport back to the walls. New DIVIMP modelling has used a recently benchmarked SiC mixed-material model [1] to simulate sputtering from SiC walls. The modelling shows for a characteristic L-Mode shot in DIII-D that, even when neglecting prompt redeposition, Z_eff at the separatrix is reduced by about 5-20% when changing from graphite to SiC walls. These results qualitatively agree with a separate set of SOLPS-ITER/DIVIMP simulations on DIII-D [2]. Although beyond the scope of DIVIMP, regions of net deposition may be expected to consist of an amorphous mixed Si/C surface. These surfaces are similar to siliconized walls in a graphite-walled device, and therefore SiC may be expected to “self-condition” and to affect the core plasma similarly to siliconization. Studies on ASDEX-U, TEXTOR and other devices have shown siliconization to be quite effective at decreasing P_rad and Z_eff, primarily via effective oxygen gettering by the silicon on the wall. Results show similar/larger decreases in Z_eff and P_rad compared to boronization. Very little silicon is observed to penetrate the confined plasma in these experiments, supporting the modelling observation of rapid silicon ionization in the SOL. Siliconization traditionally uses highly flammable/toxic silane gas to condition the walls. SiC is not flammable or toxic, and is found to produce little/no silane under deuterium bombardment. The effect SiC walls have on core performance may be most reliably approximated by DIVIMP modelling, where the SiC surfaces remain exposed. Likewise, the contribution from regions of net deposition may be best described by the cumulative experimental evidence from siliconization. It is fortunate that both wall regimes are expected to lower Z_eff and P_rad. Finally, SiC is particularly resistant to neutron damage and it retains little tritium at elevated temperatures [3], making it a potentially attractive low/mid-Z reactor-relevant wall solution. Developing a core-edge integrated wall-solution with a neutron-resilient material in DIII-D has the potential to accelerate fusion power development.

[1] T. Abrams, et al., Nucl. Fusion 61, 6 (2021)

[2] G. Sinclair, T. Abrams and L. Holland, Fusion Sci. and Tech., 1-14 (2022)

[3] M.T. Koller et al., Nucl. Mat. Energy 20, 100704 (2019)

The lifetime of plasma facing components (PFCs) in magnetic confinement fusion devices is a critical issue for high performance steady-state operations [1]. In the EAST, the previous results show the regular damage, i.e. melting, distribution along toroidal direction on W divertor, which may be related to the heat load distribution. Thus, it is necessary to understand the the toroidal distribution characteristics of heat load on divertor target in the EAST. The plasma facing components flux (PFCFLUX) code basing on the fast and accurate field line tracking is employed to simulate the heat load on PFC surface [2]. By establishing an accurate model and setting reasonable boundary conditions, the heat load deposition patterns along toroidal direction on cassette modules (CMs) at lower divertor was successfully simulated and analyzed. The simulation results show that the heat load deposition pattern on a single CM presents obviously nonuniform distribution in toroidal direction, and the heat load on chamfer positions at gaps is increased suddenly. And the heat load deposition pattern for each CM is generally consistent and shows periodic change, which also coincides with the temperature distribution measure by infrared camera monitoring systems. It is found that one of two edge parts of each CM often be shadowed by neighbouring CM. This means that the divertor structure design on EAST is capable of alleviating leading edge effect, if no misalignment appear during installation. The heat flux on the surface of inner target (IT) and outer vertical target (OVT) is monotonically increasing or decreasing along the toroidal direction, while the heat flux of the outer horizontal target (OHT) is uniformly distributed except for the chamfer position, which is directly related to the geometric structure of the target geometry. Through specific analysis, it is proved that the change of the inclined angle of magnetic filed lines is one of the main reasons for the nonuniform distribution of the heat flux in the toroidal direction. Because of the symmetric structures of each CM, the changing of toroidal magnetic field direction from normal to reverse does not change its distribution trend. Based on the toroidal distribution simulation, the peak heat load during different plasma current operation platforms is analyzed. It is found that the peak heat load on the IT, OVT and OHT is generally higher in case of high plasma current operation. In addition, the influence of chamfer sizes on toroidal distribution of heat load is also discussed. Such study of the toroidal distribution characteristics of heat load on the divertor surface can not only verify the reliability of the component design, but also provide an important reference for the operation of the device.

[1] A. Loarte, B. Lipschultz, A. Kukushkin, et al., Nucl. Fusion 47 (2007) S203–S263.

[2] M. Firdaouss, V. Riccardo, V. Martin, et al., J. Nucl. Mater 438 (2013) S536–S539.

Experiments have recently been carried out in DIII-D with a new slot-like, tungsten-coated divertor to assess tungsten erosion and leakage from a divertor with this unique closed geometry. Reduced W leakage was demonstrated with neon impurity seeding using graphite Collector Probes (CPs) exposed to the far Scrape-off-Layer (SOL) at the outer midplane. The discharges with CPs used the same reference ELMy H-Mode scenario with the outer strike point on the tungsten-coated tiles.

In one series of discharges, neon gas was injected from four distinct poloidal locations. Spectroscopic measurements of tungsten emission from the core indicate that the core tungsten density decreased during neon injection. Similarly, the neon injection led to a clear suppression of tungsten deposition on the Outer-Target-Facing (OTF) side of the CPs. On the Inner-Target-Facing (ITF) side, the suppression of tungsten deposition is more modest although there is a clear change in the deposition pattern across the probe with neon injection. These changes to the tungsten deposition pattern on the CPs strongly imply changes in the SOL conditions, which would have a direct impact on the tungsten SOL transport itself.

Proxies for the tungsten divertor leakage can be calculated by taking one of the upstream tungsten measurements and normalizing it by a measurement of the tungsten erosion rate from the divertor. Here the tungsten erosion rate is determined by spectroscopic measurement of tungsten emission in front of the tungsten target. Three different upstream measurements are used: the relative tungsten core density, the mean tungsten deposition on the OTF side of the CP where the profile shows a radial decay, and the mean tungsten deposition on the ITF side of the CP over the same region. The three divertor leakage proxies all indicate a decrease in the tungsten divertor leakage as a result of neon injection, and all three proxies fall within the measurement error of each other.

This material is based upon work supported by the U.S. Department of Energy under Award Numbers DE-SC0020093, DE-SC0019256, FC02-04ER54698, DE-AC02-09CH11466, DE-AC52-07NA27344, DE-FG02-07ER54917, DE-AC05-00OR22725.

*Corresponding author’s e-mail: matthew.parsons@psu.edu (M.S. Parsons)

Bombardment by high energy hydrogen particles (deuterium and tritium) of the tungsten divertor surfaces will lead to a build-up of hydrogen inventory, which can induce material embrittlement and so reduce the lifetime of plasma-facing components. Because tritium can be retained in the material and permeate to the coolant, it also represents radioactive issue.

Hydrogen transport in ITER monoblocks has already been modelled numerically with 1D and 2D simulations for large sets of irradiation conditions, assuming there is no effect of the monoblock axial thickness (poloidal direction) due to its large size defined for ITER design (12 mm). Since the conceptual design for DEMO monoblock can still change, the aim for this study is to explore the impact of the monoblock axial thickness on the retention and permeation during plasma operation. Desorption from both, toroidal and poloidal gaps, is also studied during the baking phase. A 3D FESTIM [1] model is first built and transient simulations up to 10⁶ s of continuous exposure are run with or without instantaneous recombination on poloidal side surfaces.

In the case of instantaneous recombination, the poloidal gaps act as a strong sink for hydrogen leading to a decrease of the monoblock inventory. The total desorption flux on poloidal surfaces is greater than on toroidal surfaces but remains orders of magnitude lower than the retro-desorbed flux at the plasma-facing surface. For a monoblock thickness of 4 mm, the relative difference in the hydrogen inventory per unit thickness between the two cases (with and without recombination on poloidal sides) is ~200%. As the thickness of the monoblock increases, this difference decreases (~30% at 14 mm). The monoblock’s response to baking is then studied at different baking temperatures. For example, at 600 K almost all the hydrogen content in the monoblock is removed after 15 days of baking (mostly outgassing ~70% from poloidal side surfaces). Finally, it is shown that assuming a non-instantaneous recombination on the tungsten surfaces according to the literature data [2] would not have a major impact for baking temperatures above 600 K.

[1] R. Delaporte-Mathurin et al., Nuclear Materials and Energy 21, p.100709 (2019)
[2] D. F. Anderl et al., Fusion Technology 21(2P2), pp. 745–752 (1992)

Long-term D outgassing from plasma-exposed surfaces in present day tokamaks is usually well described by an empirical power law decay of flux with time after exposure: q ~ t^‑n. The power law exponent n = 0.7 ± 0.1 in the decay of pressure and mass 4 (D₂) signals has been reported multiple times, being rather independent of wall materials and exposure conditions, e.g. in JET with C and ITER-Like wall [1, 2]. The recent second DT campaign at JET allowed monitoring of long-term outgassing after pure T and mixed D-T plasmas. Interpretation of results is complicated by the formation of heterogeneous molecules (in particular DT), but the trends for all related mass signals indicate faster decay of outgassing for T than for D. In particular, mass 6 signal corresponding to T₂ is following a power law decay with n ~ 0.9. In this contribution the analysis of the long-term outgassing after D, T, DT, and also He plasmas in JET‑ILW will be presented. The long-term outgassing over a period of 10 and more hours will be compared to short-term post-discharge / inter-discharge outgassing over a period of about 1000 seconds.

[1] V. Philipps and J. Ehrenberg, J. Vac. Sci. Technol. A 11 (1993) 437
[2] V. Philipps et al., J. Nucl. Mater. 438 (2013) S1067

*Corresponding author:
tel.: +49 2461 615625, e-mail: d.matveev@fz-juelich.de (D. Matveev)

In its Pre-Fusion Power Operation phase, ITER currently plans hydrogen (H) and helium (He) plasma operations, with the latter an option to improve access to H-mode. Whilst operation at lower toroidal field (1.8 T) is foreseen for these H-mode studies, L-mode plasmas in H at higher plasma current and half (2.65 T) and full field (5.3 T) is also planned. Input is needed from present devices with ITER-relevant plasma-facing components (PFC) on the best methods to achieve the required plasma composition in these experiments, namely when changing plasma operations between H and He, or between pure plasmas and mixed H-He discharges that will provide for efficient ion cyclotron (IC) heating schemes.

In WEST with full tungsten PFCs, a changeover between deuterium (D) and He plasma operation was conducted by repetitive lower hybrid heated L-mode plasmas on the passively cooled lower divertor at 70°C. The He concentration in the recycling flux, estimated from HeI and DI line intensities, increases above 75% within the first 10 seconds at the outer divertor strike point position, while further away in the outer divertor, at the inner divertor and at the upper divertor, more than 25 seconds are required to achieve similar concentrations. The He content in the neutral gas released after the discharge agrees with the slower changeover rate of the remote PFCs rather than with the high flux area at the strike point.

In JET with the ITER-like beryllium main chamber wall and tungsten divertor, a back and forth transition between He and H plasma operation was performed using IC wall conditioning (ICWC) followed by limiter and IC heated diverted plasmas. Optical penning and spectroscopy measurements indicate that the changeover from H to He plasmas is faster than from He to H, partially because of the low pumping efficiency of the He ions. By about 15 ICWC pulses, the gas composition in tokamak pulses is reduced to 5-20% He in H and 2-10% H in He. In both He and H pulses, a higher H content is observed during the limiter phase compared to the subsequent diverted phase. Finally, after 5 weeks of He plasma operation, almost 3 weeks are needed to reduce the He down to 1% in D plasma. A similar observation for the reverse transition (D to He) is reported earlier in the full tungsten ASDEX Upgrade [1].

In both JET and WEST, the changeover is mainly driven by wall exchange during the discharges followed by outgassing and pumping after the pulse. Outgassing of H after He discharges is characterized by a strongly delayed partial pressure peak.

[1] A. Hakola et al 2017 Nucl. Fusion 57 066015

*Corresponding author: tom.wauters@iter.org

In this study, we use two forms of low energy ion beam analysis, low-energy ion scattering (LEIS) and direct recoil spectroscopy (DRS), to probe the atomic-scale behaviour of hydrogen on surfaces. These measurements involve dosing the surface with atomic or molecular hydrogen while probing the surface with low energy ions (typically 500 eV - 3 keV He⁺, Li⁺, or Ne⁺). The recoiled H intensity varies with angle of incidence and azimuth enabling us to determine the chemisorbed hydrogen surface binding configuration. In our recent work, we have pursued the use of multi-angle scattering maps, which involve acquiring scattered and recoiled particle fluxes over a large angular sector. When coupled with binary collision or molecular dynamics simulations, the surface configuration can be determined with high precision < 0.02 nm [1]. We have applied these techniques to a variety of materials, ranging from model systems to technical surfaces [2, 3].

Like most surface analysis techniques, both LEIS and DRS are generally used at low pressures typical of ultra-high purity vacuum systems, usually < 10^-4 Pa. Here we have developed modifications to these techniques to allow them to be extended to higher pressures and more complex exposure environments. For these measurements, we use a time-of-flight (TOF) spectrometer that includes a differentially-pumped ion source and flight tubes. This equipment was attached to an ultra-high vacuum chamber where the scattered and recoiled particles are sampled at angles of 65° and 160° relative to the incident ion beam, enabling detection of both forward and back-scattered particles simultaneously. Initial characterization of 10 keV Li⁺ scattered from Au, W, Pd, Ni, and Al surfaces at pressures of 5×10^-2 Pa and above revealed minimal attenuation of the scattered signal and good mass resolution. Preliminary results from experiments to study adsorption and implantation of hydrogen onto / into different tungsten materials will be presented. These measurements will be calibrated against surface-sensitive temperature programmed desorption (TPD) as well as x-ray photoelectron spectroscopy (XPS) measurements.

Sandia is a multi-program laboratory managed and operated by Sandia Corporation, a wholly-owned subsidiary of the Honeywell Corporation, for the United States Department of Energy's National Nuclear Security Administration under contract DE-NA0003525.

[1] C. -S. Wong, R. D. Kolasinski, and J. A. Whaley, Surf. Sci. 729 (2023) 122229.

[2] C. -S. Wong, J. A. Whaley, Z. J. Bergstrom, Brian D. Wirth, and Robert D. Kolasinski, Phys. Rev. B 100 (2019) 245405.

[3] R. D. Kolasinski, J. A. Whaley, and D. K. Ward, Surf. Sci. 677 (2018) 176.

*Corresponding author: tel.: +1 925 294 2972, e-mail: rkolasi@sandia.gov (R. D. Kolasinski)

Fuel retention in plasma-facing materials (PFM) is one of the critical issues for fusion devices. Tungsten has been a promising candidate to PFM due to its high melting point, high thermal conductivity and low sputtering yields. In this work, rolled and recrystallized tungsten were studied. Damaged rolled and damaged recrystallized tungsten were prepared by iron ion irradiations at energy of 3.5 MeV and dose of 1.1×10¹⁹ ions/m². The distribution of vacancy defects on the near surface of pristine and damaged tungsten was analyzed by positron annihilation- Doppler broadening spectroscopy (PA-DBS). Two series of plasma exposure with different durations (90 mins and 130 mins) were carried out for the target samples. With using thermal desorption spectrum (TDS), deuterium retention of the target samples was compared after the low-temperature (380 K) and low-flux (~10²¹ D/m²s) plasma exposure. Surface observation indicates that surface blistering is barely formed on the samples except the pristine rolled tungsten where small-size blisters present with a low density.

PA-DBS results show that the S-parameter, which represents total volume of vacancy-type defects, is highest in the damaged rolled tungsten, followed by the damaged recrystallized one and lowest in the recrystallized one. It’s suggested that the number of defects is the highest in the damaged rolled tungsten, and the lowest in the recrystallized tungsten. The TDS results show that deuterium retention is highest in the damaged rolled tungsten, followed by the damaged recrystallized one and lowest in the recrystallized one. The trend of deuterium retention is consistent with the trend of PA-DBS result. It is likely that the amount of deuterium retention was dominated by the total number of defects in the material. For all samples, increasing exposure duration results in a slight increase in deuterium retention. Comparison uncovers that the amount of incremental deuterium induced by the longer exposure duration varies considerably between the samples. To analyze the effect of defect content on deuterium retention, ratio of deuterium retention to exposure fluence was calculated and compared. It is found that the ratio is highest in the damaged rolled tungsten and lowest in the pristine recrystallized tungsten, which is consistent with the volume of defects. In addition, the longer duration of exposure leads to a decreased ratio of deuterium retention in the tungsten samples. The decrease of the ratio becomes more significant with a greater number of defects. Results of this work provides a new insight of understanding the behavior of deuterium retention in tungsten.

The Atomic and Molecular Data (AMD) Unit of the International Atomic Energy Agency (IAEA) [1,2] evaluates, recommends and makes available atomic and molecular (AM) data as well as plasma-surface and plasma-material interaction (PSI and PMI) data for nuclear fusion energy research. The AMD Unit’s core activities are based on Coordinated Research Activities (CRPs) [3], maintaining and generating various searchable and freely available databases [4], and organizing various technical meetings and workshops [5].

A PSI-related CRP “Hydrogen Permeation in Fusion-relevant Materials” is currently being run by the AMD Unit. This CRP has participants from fusion research institutes representing 15 countries with the aim of studying parameters affecting plasma fuel permeation (deuterium and tritium) in fusion reactor components which influences the assessment of in-vessel and ex-vessel source terms for fuel retention. In addition to permeation, diffusion, trapping, retention, etc. studies, the Hydrogen Permeation CRP includes two experimental subtasks, i) neutron irradiation of selected fusion materials and a round-robin exercise to perform coordinated irradiation damage characterizations on these samples and corresponding hydrogen isotopes retention and permeation research, and ii) a round-robin activity on gas-driven permeation of deuterium in reduced-activation ferritic materials. Details and a progress report on the CRP activities will be given in this presentation.

The AMD Unit maintains various open databases including the ALADDIN database for numerical AM and PSI data, AMBDAS for bibliographical data, and Clerval database for institutions and events. Recent database activities include the deployment of a new database CollisionDB [6] for AM processes as well as the recent databases for computational PMI methodologies i) CascadesDB for neutron-induced cascades simulations in various plasma-facing component and structural materials [7] using classical Molecular Dynamics, and ii) DefectDB, which provides structural data on point defects and primary damage calculated with first-principles computational methodologies, such as electron Density Functional Theories (DFT) [8]. An update on these database developments will be given.

[1] https://www.iaea.org

[2] https://amdis.iaea.org

[3] https://amdis.iaea.org/CRP

[4] https://amdis.iaea.org/databases

[5] https://amdis.iaea.org/meetings and https://amdis.iaea.org/workshops

[6] interim URL https://db-amdis.org/collisiondb

[7] https://cascadesdb.iaea.org

[8] interim URL https://db-amdis.org/defectdb

Recently, we have shown that thin (33 to 55 nm), thermally grown oxide films on tungsten (W) prevent deuterium (D) uptake from a low-temperature plasma (5 eV/D; 370 K; 1.4×10²⁴ D/m²) into metallic W [1]. Here, we continue this investigation with higher D fluence up to 2.3×10²⁵ D/m² and, additionally, with higher D energy (up to 38 eV/D) or higher sample temperature of 500 K, in order to investigate the reduction mechanism of the oxide and to determine under which conditions the permeation barrier effect for D breaks down.

We show that during plasma exposure at 5 eV/D and 370 K the W-enriched zone, that forms on the surface of the oxide, saturates its growth with increasing fluence and protects the oxide beneath it against further reduction. For a fluence of 1.6×10²⁵ D/m², 99 % of the D uptake into metallic W is prevented by an originally 55 nm thick oxide film. D enters the metal solely by small cracks in the oxide that form during reduction by the plasma. For higher D energies or higher temperature, the reduction is stronger and laterally more inhomogeneous. The D uptake into the metal increases proportional to the area of the oxide/metal interface on which the oxide is reduced.

Additionally, we examined the D uptake through amorphous, electrochemically grown W oxide from atomic D or D plasma. We found that also the electrochemically grown oxide strongly reduces D uptake. However, the D retention in the oxide itself of (4 to 6 at.%) is significantly higher than for thermally grown oxide (1.3 at.%) and depends strongly on the reduction state after D exposure.

The new experiments confirm our previous assumption that D uptake into metallic W is blocked due to the difference in the enthalpy of solution of D in the oxide and the metal and that D can only enter the metal once the oxide at the interface is sufficiently reduced. Based on the obtained reduction rates of the oxide during D plasma exposure, a first tentative estimate on the relevance of natural oxide films for D uptake into the plasma-facing components of a fusion reactor was made. It was found that the natural oxide will have only a very minor effect in this case.

[1] K. Kremer et al., Nucl. Mater. Energy 27 (2021) 100991, DOI: 10.1016/j.nme.2021.100991

Development of structural materials, including first wall, is essential for the next generation of fusion devices. At ITER tokamak, the current plasma facing materials are beryllium and tungsten, which both have high melting points and thermal conductivities. Tungsten has additionally small erosion yields promoting prolonged sustainability of plasma, as a drawback the varying temperatures and the irradiation effects by plasma particles may lead to structural change lowering the ductile-to-brittle temperature. Also, embrittlement and material modification due to hydrogen and helium retention, respectively, may take place. Replacing tungsten with an alternative material could come under consideration in future tokamak designs.

Refractory high entropy alloys are a novel class of materials characterized by random equiatomic mixtures of five or more refractory elements [1]. The high configurational entropy gives rise to multiple lattice strengthening effects (i.e. lattice distortion, secondary phases), providing exceptional stability under elevated temperatures. In addition to advanced thermal properties, several alloys have been demonstrated to have good corrosion resistance, exceptional mechanical durability and high tolerance of irradiation [2][3]. All these features make them an interesting class of materials for extreme environments and promising candidates to withstand the conditions inside of the reactor vessel.

In the previous research [4] we investigated the suitability of refractory high entropy alloy, WMoNbTaV, as a plasma facing material from the perspective of tritium recycling, retention and trapping. In the study, tritium is modeled by non-radioactive isotope of hydrogen, deuterium. The results show a completely different deuterium trapping mechanism in WMoTaNbV alloy as compared to pure W. In W the deuterium is retained within small vacancy type irradiation damage close to the surface, while in WMoTaNbV the irradiation damage had no effect on deuterium trapping as it was more evenly distributed throughout the sample. To investigate whether the trapping mechanism is determined by other components of the alloy we have introduced deuterium into pure Mo, Ta, Nb and V and measured its retention with ion beam analysis tools and thermal desorption spectrometry.

The results indicate similar deuterium trapping behavior in Ta, Nb and V that was found in the earlier study in WMoTaNbV. Further research is necessary for deeper understanding of the nature of trapping mechanics behind deuterium retention.

[1] B. Cantor, I. Chang, P. Knight, et al., Mater. Sci. Eng. A. 375377 213 (2004)

[2] O. N. Senkov, G. B. Wilks, D. B. Miracle, et. al, Intermetallics 18 (2010)

[3] S. K. Bachani, C.-J. Wang, B.-S. Lou, et. al, Surf. Coat. Tech. 403 (2020)

[4] A. Liski, T Vuoriheimo, P. Jalkanen, et al., Materials, 15, 20 (2022)

Wall-conditioning is applied in current fusion machines, e.g. at JET and Wendelstein 7-X and are planned for future fusion reactors, i.e. ITER [1] to ensure reproducible conditions and plasma performance. Different techniques used and considered are baking, deposition of low-Z films, glow discharge cleaning (GDC), electron or ion cyclotron wall conditioning (ECWC and ICWC). Comparison of those techniques require a flexible machine such as TOMAS (TOroidal Magnetized System) [2], located at the Institut für Energie- und Klimaforschung - Plasmaphysik in Jülich, Germany. It is equipped with a range of plasma diagnostics, a sample load system and antennae for GDC, ICWC and ECWC.

To study the erosion of tokamak co-deposits, samples made out of TEXTOR (Tokamak EXperiment for Technology Oriented Research) tiles are exposed. These graphite samples with tokamak co-deposits contain mainly hydrogen (H), deuterium (D), carbon (C), oxygen (O), and boron (B). The exposures in TOMAS aim at the assessment of removal efficiency of co-deposited boronization layers. The samples were exposed to GDC in helium (He) and to ICWC in He and H, respectively. The plasma parameters were changed systematically to determine the best exposure conditions. The power level in ICWC plasmas was in the range from 1 to 4 kW. Additionally, the electrical connection of the sample holder was varied, i.e. samples were at floating potential, grounded to the vessel or biased with a negative voltage. Pre- and post-exposure analyses were carried out with heavy ion elastic recoil detection analysis (HI-ERDA) and nuclear reaction analysis (NRA). The main results are:

a) The consistent decrease of D throughout the full co-deposited layers could be observed only on samples exposed to ICWC plasmas powered above 4 kW.

b) H is removed from the surfaces more effectively than D.

c) O is eroded efficiently by GDC and ICWC in He and H: a decrease of O reaching up to 18 at. % in the uppermost sample layer of around 250 nm.

d) Biasing of the sample holder with negative voltages during ICWC exposures in He, reduced the removal efficiency of D from the samples.

The results are associated with plasma parameters obtained by Langmuir probe data. A clear relation between the power level and removal efficiency has been shown. For H plasmas neutral particle data are considered in the analysis.

[1] T. Wauters et al, Plasma Phys. Control. Fusion 62 (2020) 034002

[2] A. Goriaev et al, Rev. Sci. Instrum. 92 (2021) 023506

Tungsten (W) is the most promising candidate to be used as plasma-facing materials in future fusion reactors, owing to its high melting point, low sputtering yield, high heat conductivity and good mechanical properties [1-2]. When serving, W will subject to hydrogen (H) isotope environment including deuterium (D) and tritium (T) plasma implantation [2], H produced by high rate (n, p) transmutation induced by 14 MeV neutron irradiation, and also the permeation of gas state of hydrogen isotopes [1-2]. And the hydrogen isotope permeation and retention is of vital importance for fuel cycling, plasma confinement, and reactor operating safety. However, the hydrogen isotope permeation and retention behaviors in W and W alloys have not been fully studied.

In this work, a recent-developed carbide dispersion strengthened tungsten-base material (referred to ZrC-W) [3-4], fabricated by the Institute of Solid State Physics, Chinese Academy of Science, was selected to study. The D permeation properties of ZrC-W has been investigated by using the gas-driven permeation (GDP) apparatus located at USTC, covering the temperature range from 923 K to 1123 K, with the gas loading pressure from 5.0X10⁴ Pa to 1.0X10⁵ Pa. And the permeation signals were monitored by a quodrapole mass spectrometer (QMS) calibrated by a standard leak. The deuterium permeability, effective diffusivity, and solubility have been obtained. Moreover, deuterium retention characteristics was studied through thermal desorption spectroscopy (TDS) following 10-hour and 1.0X10⁵ Pa static thermal deuterium charging at 973 K. The microstructures (eg. grain boundaries, ZrC nanoparticles) was characterized by scanning electron microscope (SEM) and transmission electron microscope (TEM). And the relationship between TDS peaks and microstructures has been discussed. Finally, the D permeation and retention properties of ZrC-W have been compared with AT&M polycrystalline W and other W-base materials in literatures.

Further work will focus on the effect of H/He mixed plasma irradiation on the hydrogen isotope permeation and retention properties in ZrC-W.

[1] K. D. Hammond, Mater. Res. Express 4, 104002 (2017).

[2] M.Y. Ye, et al., J. Nucl. Mater. 241-243, 1243-1247 (1997).

[4] Z.M. Xie, et al., J. Nucl. Mater. 469, 209-216 (2016).

[5] Z.M. Xie, et al., J. Nucl. Mater. 496, 41-53 (2017).

*Corresponding author: tel.: +86 0551 63603224, e-mail: yemy@ustc.edu.cn (M.Y. Ye)

Laser Induced Breakdown Spectroscopy (LIBS) method uses short but intense laser pulses to ablate, evaporate and ionize the wall material into a light emitting transient plasma plume. The corresponding emission spectra is then used to determine the composition of the investigated material [1]. LIBS is considered for the assessment of hydrogen isotope retention in the ITER plasma facing components during the maintenance breaks when the reactor is filled with near atmospheric pressure nitrogen or dry air [1]. At these conditions the broadening of hydrogen isotope lines and the reduction of line intensities complicates the distinguishing of isotopes [2,3]. The use of argon or helium ambient atmosphere instead of air during LIBS measurements has shown to enhance the LIBS signal of hydrogen isotopes at atmospheric pressures [2,3]. The aim of present study was to investigate the effect of atmospheric pressure nitrogen, argon or helium ambient gas to the intensity and width of H_α line by LIBS. These measurements were supplemented by the study of spatial and temporal development of the plasma plume along the target’s axial direction.

Nd:YAG laser with 8 ns pulse width was used to ablate the molybdenum (Mo) target with hydrogen impurity and the development of the formed plasma plume was investigated by time and space-resolved emission spectra in the 20 nm range around the 656.28 nm H_α line [3]. For all gases used in the experiments, the intensity and width of H_α line decreased with the delay time between the laser pulse and the spectral registration. The slowest decrease of the H_α line intensity was observed in argon while the line-widths decreased fastest in helium. At the same line-width values, the helium atmosphere allowed to obtain highest intensity while lowest intensity was obtained in N₂. According to spatially resolved spectral measurements, H_α line was most intense near the Mo target while the Mo I lines peaked farther away from the target. The spatial development of the plasma plume was comparable in both argon and nitrogen atmospheres. The Mo intensities peaked between 2-2.5 mm (from the target) and within the delay time range between 1 to 6 μs, the Mo I line intensities remained practically same while H_α line decreased with the delay time. In the case of helium atmosphere, the Mo intensity was observed at longer distance from the target and both Mo and H_α lines decreased with the delay time. According to these results, helium is the most beneficial ambient gas for hydrogen isotope detection by atmospheric pressure LIBS. The use of argon ambient gas may be required when LIBS is used for the simultaneous determination of fuel and He retention in the wall material.

[1] H.J. van der Meiden, S. Almaviva, J. Butikova et al. Nucl. Fusion 61, 125001 (2021)

[2] N. Idris, H. Kurniawan, T.J. Lie et al. Jpn. J. Appl. Phys. 43, 4221 (2004)

[3] I. Jõgi, J. Ristkok, J. Raud et al. Fusion Eng. Des. 179, 113131 (2022)

Currently, tungsten is considered the material of choice for the large plasma-facing
surfaces of future fusion reactors. For this reason oxidation-resistant self-passivating
tungsten alloys have been developed, which prevent mobilisation of radioactive
tungsten in case of an accident. In this contribution these are the two materials
W-10Cr-0.5Y and W-10Cr-0.5Y-0.5Zr produced by CEIT, San Sebastian, by
mechanical alloying, hot isostatic pressing, and optionally a subsequent heat
treatment.
To obtain information about the tritium inventory in such materials, we investigate the
retention of deuterium after introducing displacement damage in the crystal lattice.
10×10 mm² samples are irradiation-damaged by 20 MeV tungsten ions to a dose of
0.23 dpa. This is performed at temperatures from room temperature to 400°C. The
procedure introduces irradiation damage in the material down to a depth of about
2 µm. Subsequently, the samples are exposed to a deuterium plasma with a bias
voltage of –100 V. Here the exposure temperature is varied from 100°C to 250°C. Ion
fluences up to 10²⁵ m^-2 are applied. After the plasma exposure, the deuterium
distribution in the samples is investigated by nuclear reaction analysis using a ³He
ion beam at primary energies up to 6 MeV. This allows depth profiling up to 10 µm,
which includes the irradiation-damaged region as well as the region beyond. Finally,
we determine the total deuterium content of the samples by thermal desorption.
In general, the total deuterium content increases much faster with increasing fluence
in the alloy as compared to pure tungsten resulting in an order of magnitude higher
total retention. This is attributed to enhanced diffusion due to the ultrafine-grained
microstructure. In the displacement-damaged region, however, the deuterium
concentration diminishes significantly with increasing plasma exposure temperature
as compared to pure tungsten. This is attributed to a reduced amount of vacancies at
elevated temperatures. The latter could be explained by vacancy annihilation at grain
boundaries – enhanced due to the ultrafine-grained microstructure once mobile.
Another explanation could be the presence of chromium, which is known to suppress
vacancy formation during irradiation damage.

Plasma-facing materials in fusion devices experience extreme conditions as high fluxes of energetic particles hit the wall materials. Some of these particles are radioactive isotopes of hydrogen, tritium (T), which can be trapped into the wall materials, causing loss of fuel and making the walls radioactive. This hinders both the operation and wall maintenance in fusion devices. To choose proper materials and hydrogen removal methods for them, the details of hydrogen trapping must be known.

Hydrogen trapping and removal is still a known problem with the future fusion devices. We investigate this problem by measuring hydrogen trapping in tungsten (W), the divertor material for ITER. Previous studies on hydrogen isotope trapping on irradiated W have largely focused on saturating the materials’ trap sites with deuterium (D) [1–3]. Saturated, full trap sites only show the combination of all the trap filling levels. The different levels of trapping have different trapping and release energies on which our study focuses.

In our study to mimic the irradiation conditions of a fusion reactor, we first irradiated W samples using 4 MeV W ions to create vacancies and other defects which function as trap sites in the material. The samples were then annealed in D atmosphere at 200 °C using four different annealing times ranging from 4 hours to 7 days to fill the trap sites with D, hydrogen isotope chosen to work as a substitute to T for safety reasons. The annealing times were chosen to fill trap sites to various levels: depending on the annealing time, some samples had the trap sites saturated and some samples had only slightly filled traps. These samples were measured by Positron Annihilation Spectroscopy to investigate empty trap sites and by Thermal Desorption Spectrometry to investigate total D retention and D release temperatures from the trap sites. Two desorption temperatures were found with TDS. Ratio of the desorption peaks depended on the irradiation damage and annealing time in D atmosphere. Overall, this gives us a good overview of trapping mechanics of hydrogen in tungsten from partially filled trap sites. This information can be used to understand hydrogen trapping and release in the fusion devices.

[1] B. Wielunska, M. Mayer, T. Schwarz-Selinger, et al., Nucl. Fusion. 60, 096002 (2020)

[2] G.M. Wright, M. Mayer, K. Ertl, et al., Nucl. Fusion. 50, 075006 (2010)

[3] V.K. Alimov, Y. Hatano, B. Tyburska-Püschel, et al., J. Nucl. Mater. 441, 280–285 (2013)

Tungsten (W) is the divertor material in JET, ITER and WEST tokamaks. During tokamak operations, W will interact with the hydrogen isotopes (H) fuel leading to retention and outgassing of H. This can be detrimental for tokamak operations due to plasma control issues (spatial variation in fuel recycling) and nuclear safety concerns. Thus, the understanding of H retention properties in W materials are important to quantify the impact of plasma-wall interactions on plasma operations.

In a recent experimental study, Dunand et al [1] exposed a W(110) surface with oxygen intrinsic impurity (O) coverage from 0 to 0.75 monolayer (ML) to D₂ gas and to D₂⁺ ions. They observed that the clean W(110) retained more D when exposed to D₂ gas only instead of simultanenous D₂ and D₂⁺ ions. The suggested mechanism for explaining such behavior is a sputtering process of adsorbed D by D₂⁺ ions during the simultaneous exposure.

In order to test this interpretation, the code MHIMS (Migration of Hydrogen Isotopes in MaterialS) [2] is used to simulate these experiments. A sputtering process of adsorbed D by incident D ions is included. First, the desorption energy of D from W(110) surface with various O coverages are obtained by simulating the D desorption spectra after exposure to D₂ gas only. They are determined using a parametric optimization starting from first guesses calculated using density functional calculations (DFT) [3]. While the final desorption energies deviates from the DFT data by 0.1-0.2 eV, the trend is the same with lower and lower H desorption energy as the O coverage increases. Then, the bulk trapping parameters as well as the sputtering yields are obtained for the various surfaces. The simulations suggests that the addition of the sputtering mechanisms is mandatory to simulate the experimental results. This additional recycling mechanism may affect the overall fuel recycling as not only thermally released D₂ molecules but also D atoms adsorbed on the surface can be released from the materials upon D ion implanted.

[1] A. Dunand et al, Nucl. Fusion 62 (2022) 054002

[2] E. A. Hodille et al, Nucl. Fusion 57 (2017) 056002

[3] Y. Ferro et al, Submitted to Nuclear Fusion (2022)

Dynamic Measurements of Deuterium Retention in Tungsten
During Exposure to Plasma

W. Garcia^a,* G. Burns^a, and W. Wampler^a

^aRadiation-Solid Interactions, Sandia National Laboratories, Albuquerque, NM, 87123, USA

Tungsten (W) will be used for the divertor in ITER and is the leading plasma facing material for use in future fusion reactors. When in operation the ITER W divertor will face extreme environments with steady-state plasma-induced surface temperatures reaching 1000 °C and even higher transient temperatures. Helium (He) and Deuterium (D) particle fluxes of and high heat fluxes ranging are expected [1]. To investigate materials, under fusion relevant environments, the Ion Beam Laboratory (IBL) at Sandia National Laboratories at Albuquerque, NM has developed a new endstation in which materials can be exposed to a high flux (~1e20 /m²s) of low-energy (~100eV) D ions from an eH400 Kaufman & Robinson ion source, and MeV ions from an NEC Pelletron ion accelerator. This enables dynamic measurements of near-surface D content in materials during exposure to plasma using D(He3,p)He4 nuclear reaction analysis (NRA) with a detection limit around 10¹⁸ D/m² or a bulk concentration around 4x10^-5 D/W. Energetic protons are counted using a silicon particle detector external to the vacuum chamber. The objective is to examine the various physical mechanisms contributing to D retention, such as implantation and diffusion, trapping at lattice defects and gas-phase precipitation, all of which strongly depend on temperature [2]. D retention also depends on near-surface composition and microstructure of the material, for example from helium implantation [2]. The MeV ion beam also provides the capability to produce displacement damage as a surrogate for damage by 14 MeV DT neutrons in a fusion reactor. The endstation includes heating to measure and control sample temperature up to 1000 °C, and a Faraday probe to measure the flux of low-energy ions onto the sample. Preliminary tests have been done to characterize operation of the ion source using Argon, He and D. Measurements with a quartz sample of NRA yield versus time as the D plasma was turned on and off confirmed that NRA measures D content in the sample during D plasma exposure with little background from D in the plasma as predicted, and showed that the D content in the sample decreased when the plasma was turned off. Results from dynamic measurements of D retention in W versus sample temperature and D ion flux will be reported.

Sandia National Laboratories is a multimission laboratory managed and operated by National Technology & Engineering Solutions of Sandia, LLC, a wholly owned subsidiary of Honeywell International Inc., for the U.S. Department of Energy’s National Nuclear Security Administration under contract DE-NA0003525.

[1] D. Guilhem, J. Gaspar, C. Pocheau, and Y. Corre., IEEE Trans Plasma Sci, vol. 48, no. 7, (2020)

[2] W.R. Wampler, Phys. Scripta T171, page 014012, (2020)

*Corresponding author: tel.: +1(575)496-0764, e-mail: wagarci@sandia.gov (W. Garcia)

In ITER, the divertor exhaust made of tungsten (W) is subjected to high flux of hydrogen isotopes and He. It has been recently demonstrated that hydrogen isotopes retention increases locally where He is present in W [1, 2]. Therefore, understanding the fundamental mechanisms leading to He retention in W is essential to better estimate hydrogen isotopes retention. Most of models for He retention in W consider that He is a mobile species that agglomerates on preexisting bulk defects, such as grain boundaries, W vacancy and vacancy clusters. However, another mechanism is possible in theory, He self-trapping, when for sufficiently high implantation flux there is formation of mobile He clusters in W and an He cluster is immobilized on a “self-created” vacancy thanks to the emission of a W interstitial in its vicinity. Such self-trapping mechanism has been considered in several modeling works [3, 4] but there is currently no experimental proof that He can self-trap in W.

In the present contribution, we demonstrate experimentally the role of the He flux in He retention and interpret this result as the “self-creation” of vacancy-interstitial defects (Frenkel pair) through He self-trapping. We used a W(110) single crystal sample of high-purity (99,999%) that has been prepared until low energy electron diffraction (LEED) shows a 1×1 reconstruction typical of clean W(110) and auger electron spectroscopy (AES) confirms the absence of surface impurities. The clean W(110) sample was implanted in ultra-high vacuum (base pressure 2×10^-10 mbar) and at room temperature (300 K) with He ions having a kinetic energy of 130 eV (i.e. below the He displacement damage threshold) for a total fluence of 2×10²¹ He.m^-2, with five different flux in the 10¹⁶ – 10¹⁷ He.m^-2.s^-1 range. Temperature programmed desorption (TPD) from 300 K up to 2200 K was realized thanks to a continuous wave infrared laser delivering up to 1000 W coupled with two optical pyrometers and a quadrupole mass spectrometer. The presented results show that at the lowest investigated flux the He retention is below the detection limit. Then, He retention is observed well above the detection limit once the He flux crosses a threshold between 0.3 10¹⁷ and 0.7 10¹⁷ He⁺ m^-2 s^-1. Above this threshold, the He retention is increasing linearly with the He flux. At the lowest flux above the self-trapping threshold, two desorption peaks are observed at 950 K and 1700 K, consistently with the DFT literature for the dissociation energy of a He_n cluster from a single vacancy. Increasing further the flux, three desorption peaks are measured at 950 K, 1650 and 1800 K, corresponding to the DFT dissociation energy of He_n clusters immobilized on vacancy clusters V₂ and V₃, presumably resulting from trap-mutated He-vacancy clusters [5].

[1] S Markelj, T Schwarz-Selinger and A Založnik, Nucl. Fusion 57 (2017) 064002
[2] M Ialovega et al., Phys. Scr. T171 (2020) 014066
[3] T. Faney et al Nucl. Fusion 55 (2015) 013014
[4] R Delaporte-Mathurin et al., Sci. Rep. 11 (2021) 1468
[5] A Dunand et al., Nucl. Mater. Energy, accepted

Plasma-facing materials (PFMs) in ITER and future fusion reactors will be subjected to severe particle and heat loads that can damage the materials and degrade their properties. Tungsten was chosen as PFM for the ITER divertor due to its favourable properties such as high melting temperature, low erosion rate and low tritium retention. Tritium retention is of particular importance as, due to safety considerations, a tritium inventory of less than 700 g is to be kept in the reactor. Tritium fuel is also rare and should be recovered as much as possible. However, the low tritium retention property of tungsten can be deteriorated by macro- and microstructural changes in the material after withstanding thermal shocks from edge-localized modes (ELMs) and from neutrons and plasma loads. It is, therefore, important to study how tritium retention changes as the material structure could change along its lifetime in the reactor.

ITER-grade tungsten samples were exposed to a stationary D/He(6%) plasma and ELM-like thermal shocks in the linear plasma device PSI-2. Deuterium was used as a proxy for tritium, as tritium experiments are not possible due to its rarity and radioactivity. Deuterium retention was then measured in-situ with laser-induced breakdown spectroscopy (LIBS). It was found that the damage and microstructural changes caused by the particle and heat loads can strongly affect the deuterium retention behaviour of tungsten and this effect should be studied further in order to fully understand its possible effects in the ITER divertor and future fusion reactors.

Hydrogen isotope (HI) retention in plasma-facing components (PFC) of a fusion device remains one of the burning issues of self-sustainable magnetically confined fusion. Great experimental and theoretical efforts have been dedicated to studying hydrogen interaction with fusion-relevant materials and improving our understanding of the underlying phenomena. In addition to experimental work, models describing HI interaction with materials have been developed, which can be used to predict fusion fuel retention in PFC of a fusion device. Accurate models are needed to understand the conditions inside the device for better planning of its operation and maintenance.

Models have been significantly improved in recent years by including displacement damage creation, HI interaction with lattice defects and bubbles, etc. However, many phenomena have yet to be understood and included in the models, such as synergistic effects of HI irradiation and displacement damage creation. Another example where the shortcomings of current models have been exposed is deuterium plasma irradiation of tungsten under high ion flux (> 10²⁰ D/m²s). In [1], modeling consistently failed to accurately estimate the amount of retained D in W, more so as the D ion flux increased. The authors in [1] tackled this issue by significantly reducing the implantation flux in their model. In contrast, the author in [2] argued that the correct approach is to use a different value for the surface recombination coefficient. Such approaches do lead to modeling results agreeing with the experimental data, however, the altered modeling parameters can differ from their measured values by many orders of magnitude and such changes generally remain unjustified.

In an effort to find the most physically accurate modeling approach capable of describing a variety of implantation conditions, a series of D plasma irradiation experiments were performed using the PISCES-RF linear plasma device. Tungsten samples were divided into 3 groups, each group exposed to a different D ion flux (7x10²⁰ D/m²s, 4x10²¹ D/m²s, or 4x10²² D/m²s) at 400 K and ion energy of 60 eV. Within each group, samples were exposed for various lengths of times, fluences varying from 3.4x10²⁴ D/m² to 3.4x10²⁵ D/m². Deuterium content in the samples was investigated by thermal desorption spectroscopy and ³He nuclear reaction analysis.

Preliminary results show a deviation from modeling predictions. Samples exposed to higher flux consistently show higher total retained D amounts, even though modeling predicts total amounts almost independent of flux or with only minor changes as the flux varies and fluence remains the same. In order to understand this inconsistency, additional analyses will be performed, e.g., looking for bubble layers or possible D super-saturated layer that may change the properties of the material and the D diffusion and retention kinematics. The results will be presented at the conference.

In ITER, tungsten (W) will constitute divertor exhaust targets. An understanding of the interaction of W with fusion fuel (deuterium (D) and tritium (T)) is needed, because T is scarce and radioactive. Recent experiments showed that the (near-)surface of W could play a role in the D/T retention. ‘t Hoen et al. found that the bulk penetration flux of D in W was much smaller for 5 eV implantation than for 40 eV implantation [1]. They argued that the high flux (10²⁴ D m^-2 s^-1) should remove oxygen (O) impurities from W and thus a chemisorbed D surface layer would act as a bulk penetration barrier for 5 eV incident D. However, plasma removal of surface impurities was not backed up by in situ measurements. Recently, we showed by combining Ion Beam implantation, Temperature Programmed Desorption (TPD), Nuclear Reaction Analysis experiments and Macroscopic Rate Equations modelling [2] that the native oxide of the W surface could have a significant retention contribution in laboratory experiments. This prediction has just been confirmed experimentally in our group [3].

We performed a direct comparison of D retention in tungsten covered with a few nm thick native oxide versus the atomically clean W surface with increasing sub-monolayers coverage of O atoms. TPD evaluation of D retention after 250 eV/D ions implantation at 300 K shows that D surface retention is important on clean W but D bulk retention becomes predominant when the O surface coverage increases above 50% of a monolayer. Thus, a dense O surface layer could act as a re-surfacing barrier for D as suggested by recent DFT calculations [4]. This observation may explain the highest D retention observed for W with its native oxide.

To assess the effect of bulk O on D retention, we implanted 250 eV/D ions at 300 K in a ~200 nm thick WO₃ grown at 1073 K. The evolution of D retention as a function of ion fluence and storage time in vacuum was evaluated by TPD. For cumulated D fluence in the 10¹⁸ – 10¹⁹ D m^-2 range, WO₃ retained less D than the native oxide and D retention vanishes after several days in vacuum, in stark contrast to the native oxide [2]. For cumulated D fluence in the 10²⁰-10²¹ D m^-2 range, D retention in WO₃ is much higher than in the native oxide and retention is almost not dynamic. X-ray Photoelectron Spectroscopy (XPS) shows that an oxygen depleted WO_3-x layer is formed upon extended D implantation, consistent with Scanning Electron Microscopy of Focused Ion Beam (SEM-FIB) lamellae observations. Thus, defective WO_3-x acts as an effective near-surface retention layer.

[1] ’t Hoen et al., Phys. Rev. Lett. 111, 225001 (2013)

[2] Hodille et al., Nucl. Fusion 57, 076019 (2017)

[3] Dunand et al., Nucl. Fusion 62, 054002 (2022)

[4] Ferro et al., Nucl. Fusion, accepted (2023)

ITER will use beryllium as first wall material and tungsten as divertor material. Alloys can form due to erosion of beryllium and tungsten particles and their redeposition elsewhere. In the course of the plasma-wall interactions, tritium from the plasma can be deposited in the plasma-facing material. For safety, tritium breeding and economy considerations, the understanding of hydrogen retention in Be-W alloys is of central importance for the successful operation of ITER.

Fundamental experiments are required to understand the processes involved in hydrogen retention in Be-W alloys and in-situ ion beam experiments are performed to that end. A suitable system for this is ARTOSS, an ultra-high vacuum device, in which Be-W alloys can be produced, loaded with deuterium and examined using analysis techniques like thermal desorption spectroscopy (TDS) and ion beam analysis (IBA).

Recent studies have shown a low temperature deuterium desorption peak at around 400 K for beryllium, which shows a splitting into a fine structure from a threshold fluence of 1 ⋅ 10²¹ m^-2 executed with a sufficiently high resolution by a slow temperature ramp. There are several possible explanations for the formation of the sharp low-temperature peak and its splitting, such as the formation of beryllium hydrides [1] or the bursting of gas-filled bubbles [2, 3]. In order to decide exactly which mechanism is the reason for it, ramp-and-hold TDS experiments are used.

In this contribution, I will show results from previous studies and first own results from ramp-and-hold experiments with beryllium. In addition, the desorption spectra of the in situ loaded samples are to be compared to those of beryllium samples from the first wall of JET.

[1] Anže Založnik, Matthew J. Baldwin, Michael J. Simmonds, Nucl. Fusion 59, 126027 (2019)

[2] D. Matveev, M. Zlobinski, G. De Temmerman, B. Unterberg, Ch. Linsmeier, Physica Scripta T171, 014053 (2020)

[3] D. Matveev, M. Wensing, L. Ferry, F. Virot, M. Barrachin, Y. Ferro, Ch. Linsmeier, Nucl. Instrum. Methods Phys. Res. B 430, 23 (2018)

The divertor experiences hydrogen isotopes-helium (He) admixure plasma exposure. One of the most famous He plasma effects is the formation of fuzz structures on tungsten (W) [1], which alters retention properties of hydrogen as well as that of helium itself. Previous research has shown a presence of deuterium (D) and He desorption peak at ~350 K on nanostructured fuzz W samples [2, 3], while a mechanism for such a low temperature desorption is still unclear. In this study, we focus on revealing the causes of this low temperature desorption peak.

Experiments are conducted by using a compact RF linear plasma device APSEDAS (Advanced PWI Simulation Experimental Device and Analysis System). Tungsten samples of are first exposed to a He plasma with the ion fluence of , then to a D plasma with the fluence of . Note that the D plasma exposure is limited to the area of Φ8 mm on the sample surface. When using He plasmas, surface temperatures were set to make different types of surface structures: fuzz(~1200 K) and He-bubble (~1000 K) with the incident ion energies of 60 and 20 eV, respectively, while those parameters during the D plasma exposure were fixed to ~600 K and 20 eV. After the exposures, the samples are moved and heated with elevating temperature (~1 K/s) to perform the thermal desorption spectroscopy (TDS).

On the samples exposed to the He plasma, a desorption peak of He at ~350 K did not appear on the bubble sample but on the fuzz sample. The D desorption was also observed on the fuzzy surface exposed to the D plasma, whereas the bubble surface shows no clear desorption at this temperature. Considering a sample exposed to D plasmas only presented desorption peaks at ~400 and ~550 K, these results imply that the low temperature desorption peak is related to the presence of the fuzzy layer.

Here we speculate that the causes of the desorption peak are (1) the surface adsorption of He on a fuzz layer, which has larger surface area by a factor of several tens compared to a smoothed surface [4], and (2) the trapping in low binding energy trapping sites which are formed by displacements due to growth of nano-fibers.

For further investigation, a fuzzy W sample was prepared and experienced a series of treatments: ‘A‘-TDS (up to 450 K), ‘B‘-Leave until room temperature then TDS again, ‘C‘- Leave for 1 day in vacuum then TDS, ‘D‘-Leave until room temperature then TDS, ‘E‘-Expose to He gas for 1 hour in vacuum then TDS. As a result, the intensity of desorption peaks (~350 K) were in the order of ‘A‘>‘E‘>‘C‘>‘B‘>=‘D‘. The peak of ‘E‘ was larger than that of ‘C‘, indicating that He surface adsorption contributes to the formation of low temperature peak.

We estimate D trapping energy of low temperature desorption peak with Kissinger-Akahira-Sunose (KAS) model-free-kinetics method, by using TDS spectra with different rate of temperature increase.

[1] S. Takamura, et al., Plasma Fusion. Res. 1, 051 (2006)

[2] M. Yajima, et al., J. Nucl. Mater 449, 9-14 (2014)

[3] T. Sakai, et al., Plasma Fusion. Res. 17, 2405062 (2022)

[4] M. Yajima, et al., J. Nucl. Mater 438, s1142-s1145 (2013)

This work was supported in part by JSPS KAKENHI (20K22322, 21H01059, 21KK0048) and NIFS Collaboration Research program (NIFS20KUGM152).

*Corresponding author: tel.: +070 1005 3747, e-mail: saito_kota@prc.tsukuba.ac.jp

Hydrogen outgassing from tungsten materials seems to be rather logical given the fact that tungsten exhibits rather low H solubility and retention. However, so far, the significant outgassing of hydrogen isotopes (HI) from ion/plasma-exposed tungsten (W) during room temperature (RT) storage after irradiation at temperature below 450 K [1] is not well understood. In contrast, for tungsten-deuterium(W-D) co-deposited layers with rather high concentration of retained D [2], little D outgassing occurs. Both puzzles motivate the present work.

In the present work, we adopt the well-established fill-level-dependent trapping and de-trapping model [3] to interpret the aforesaid hydrogen degassing behaviour based on the occupation of multiple HI atoms in W monovacancy. For HI exposed W bulk, during the HI-loading phase, all traps will be gradually filled with HI atoms to the maximum fill level in equilibrium with the ambient interstitial HI concentration. Once the HI incident flux is switched off, the concentration of interstitial HI decreases rapidly and the equilibrium will break: HI that trapped in the defects with low binding energy will start immediately to de-trap until new equilibrium is established. Whereas for W-D co-deposited layers, the vast majority of HI is trapped in the W matrix via a quasi-equilibrium state during the co-deposition process. Therefore, how hydrogen is loaded, i.e., under equilibrium or non-equilibrium state, is proposed to determine the different hydrogen outgassing phenomena in D ion/plasma-exposed W bulk and W-D co-deposited layers. Experiments on W-D co-deposited layers are on-going.

It is further pointed out that the current simulation parameters for the fill-level-dependent trapping and de-trapping model derived from static DFT calculations [4] requires further adjusting to describe the dynamic HI outgassing from W at RT, Based on the electron gas model of effective-medium theory [5], we extended the embedding energy-electron density curve for multiple HI occupation in the W monovacancy and thus obtain the dynamic changes of the trapping and de-trapping energies of HI during trapping and de-trapping processes. In addition, we considered the thermal expansion of the W matrix at RT relative to 0 K. The results show that the HI de-trapping energies calculated by DFT calculation may be higher than the actual value, resulting in a potentially lower amount of HI degassing from W at RT storage than the experimental one. As for the exact value of the difference between them, it remains to be explored.

[1] K.A. Moshkunov, K. Schmid, M. Mayer, et al., J. Nucl. Mater. 404, 174 (2010)

[2] G. D. Temmerman, R. P. Doerner, J. Nucl. Mater. 389, 479 (2009).

[3] K. Schmid, U. V. Toussaint, and T. Schwarz-Selinger, J. APPL. PHYS. 116(13), 1475 (2014)

[4] N. Fernandez, Y. Ferro, and D. Katob, Acta Mater. 94, 307 (2015)

[5] F. Besenbacher, B.B. Nielsen, J.K. Nørskov, et al., J. Fusion Energy. 9, 257 (1990)

*Corresponding author: tel.: +49 2461 613113, e-mail: li.gao@fz-juelich.de (L. Gao)

The ITER (“The Way” in Latin) fusion device is currently built in Cadarache, France, with the aim to demonstrate the feasibility of producing energy from the fusion of hydrogen isotopes (HI) nuclei. ITER hosts a divertor at the bottom of the vacuum vessel, whose main goal is to extract heat and impurities from the plasma. The divertor is built with tungsten (W) monoblocks cooled by water flowing in copper alloy (CuCrZr) tubes, while the interlayer between W and CuCrZr alloy is made of pure copper (Cu). HIs are expected to diffuse across the monoblocks and possibly accumulate at the W/Cu interface, leading to defects formation and stabilization. Part of HIs will also permeate through W/Cu up to the water coolant. Both segregation and diffusion phenomenon will affect the mechanical properties of the monoblocks and can lead to safety issues. Therefore, it is fundamental to determine the amount of fuel segregated at the interface and lost through the divertor.

If the solubility and diffusivity of hydrogen are already known in pure W and Cu, they are to be established at the interface. To accomplish this objective, we built a theoretical model of the W/Cu interface that is well suited for Density Functional Theory (DFT) calculations. Two different interfaces were considered, the W(001)/Cu^fcc(001)R45^o and W(001)/Cu^bcc(001) orientations which minimize the mismatch between the Cu and W networks. Both interfaces converge to the same final geometry after volume and geometry relaxation is performed. In addition, the Cu slab experiment a significant geometry reconstruction close to the interface, which propagates deep into the bulk. We determined the cohesion energy as a function of the number of Cu and W layers to optimize the size of the system and ended up with 8 layers of Cu and W. The final model is a 2x2x8 interface unit cell, which is more than enough to converge the cohesion energy below 1 meV/Ų.

We subsequently determined the most stable interstitial positions for hydrogen. Starting from 53 different interstitial positions located at octahedral (OH) and tetrahedral (TH) sites, we ended up with 33 stable sites after geometry optimization. In Cu bulk, far away from the interface, the most stable interstitial site is of OH symmetry. In W bulk it is of TH symmetry. At the interface, HIs predominantly occupy the OH sites in Cu, but some TH positions become stable due to the distortion of the Cu network. Reversely in W, HIs predominantly occupy the TH sites but some of the OH sites become stable. The solution energy of H at the interface is 0.114eV, while it is around 0.371eV and 0.676eV in Cu and W, respectively. Therefore, we conclude that this interface behaves as a sink for HIs, where they will accumulate and segregate favoring defect formation.

For a future commercial fusion reactor, the joining of tungsten and steel will be of vital importance, covering the main part of the plasma facing area. However, the large difference, of more than a factor of 2, in the coefficient of thermal expansion (CTE) of W and steel results in high thermal stresses at their interface. The cyclic nature of the operation can cause fatigue effects and could result in a premature failure of the joint.

One possible solution is the insertion of a functionally graded material (FGM), with varying the CTE, as an interlayer between tungsten and steel, which could reduce these stresses. In this study two processes, atmospheric plasma spraying (APS) and spark plasma sintering (SPS), are utilized to manufacture such FGMs. The gradation was accomplished by using two or three layers with a thickness of 0.5 mm each.

Another principle is the insertion of a ductile metal interlayer which reduces the stress by plastic deformation. Vanadium foils of varying thickness were chosen, as V has a CTE in between W and steel and forms a solid solution with W and Fe. These and a direct W-steel joint as reference were made by diffusion bonding. For each variation, three identical samples were tested. All samples consist of 3 mm thick W and steel tiles allowing an easy and direct comparison of the different technologies.

A high heat flux benchmark test was performed to investigate and compare the potential of the different joining technologies. For this, the complete stacks were soldered on actively cooled copper cooling modules and tested with high stationary heat loads of up to 4 MW/m² with 200 cycles at each level in the JUDITH 2 facility. By monitoring the surface temperature using an IR camera, the cooling capabilities of each sample and any local overheating as indication of bond failure can be determined. Clear and distinct performance differences of the technologies were noticed. Detailed thermal analysis including comparison with prediction based on FEM simulation are accompanied with post mortem structural analysis of the stacks in order to find the cause and location of the failures. This study allows to focus the further development of W-steel joining technologies.

Removal of low-energy neutral hydrogen atoms reduces the number of charge exchange events and thus, the net energy losses in the plasma, resulting in a significant improvement of the performance of fusion devices. Effective control of hydrogen isotopes (HIs) residual pressure in the plasma edge can be achieved by utilizing a specially designed absorbing first wall interface capable of withstanding the damaging impact of the fusion environment. Tantalum (Ta) is a promising candidate material for the gettering interface for advanced fusion concepts due to its high melting point, good thermal conductivity, resistance to sputtering, low neutron activation, and unique corrosion resistance.

This work reports on the systematic study carried out to investigate the behavior of Ta coatings produced by the cold spray deposition technology which were thermally treated at 1100 K for 15 min. Pure Ta powder (99.95%) with a mean diameter of particles of ~10 µm was cold-sprayed on 316L stainless steel substrates. Five spray parameter sets were tested to optimize deposition parameters to achieve dense and well-adhered, thick and non-oxidized coatings with a uniform surface topography. Surface morphology of the as-deposited coatings was analyzed using confocal laser scanning microscopy. Observations revealed minor spallation on the surface of some coatings which was attributed to nonuniform powder flow rate during the deposition process. The local surface roughness of the coatings was found to be around 5 ± 2 µm. Mechanical durability of the coatings measured using a standard micro-hardness tester revealed the hardness of the coatings to be ~300 HV_0.050 confirming the severe plastic deformation and associated work hardening upon particle impact.

Prior to the annealing tests, the optimized coatings were mechanically polished up to a mirror-like surface finishing. Cross-sectional morphology observations carried out using scanning electron microscopy (SEM) before the thermal treatment revealed the coatings to be dense and well-adhered to the substrate with a negligible porosity. No morphology changes in the coatings were observed with the SEM after the thermal treatment at 1100 K for 15 min in an UHV setup. The phase stability of the coatings was confirmed with the X-ray diffraction spectroscopy technique. Finally, in-situ X-ray photoelectron analysis revealed a significant reduction of the native oxide layer on tantalum surface during the heat treatment at 1100 K for 20 min, however, some residual carbon and oxygen remained even after several annealing cycles at the same temperature.

*Corresponding author: tel.: +33650792662, e-mail: ialovega@wisc.edu (M. Ialovega)

In plasma facing material (PFM) analysis area, to know the components of the PFM in different depth is of great importance, they contain the information about how plasma and material interacts with each other. Therefore, by analysing the component of PFM, we can use the results to guide the designing of better PFM and to understand the operation principle of fusion plasma.

Laser induced breakdown spectroscopy (LIBS) has been applied to PFM analysis area successfully, especially in post mortem analysis for several fusion device [1, 2]. Quantitative and qualitative results of erosion and retention information in different PFM were obtained [3, 4]. However, the high roughness value of ablation crater, the ablated material in the crater edge (not belongs to the layer to be analysed), the thermal effect of laser and the difficulty for detecting elements with low concentration are still challenging [5]. To solve these problems or decrease the influence of these problems, a scanning LIBS system is developed.

In this system, scan mirror and ultrafast fiber laser with very high frequency (50 kHz) are implemented. With scan mirror and high frequency laser, flexible ablation rate (depth resolution), ablation area (crater edge effect) and accumulated laser number (sensitivity) can be modulated by changing the scan speed and scan length. Ultrafast laser can minimize the thermal effect and is significant for gas retention analysis. Furthermore, fiber laser improves the flexibility of LIBS and gives potential for in-situ LIBS system in the future.

Plasma-material interaction (PWI) processes are of paramount importance in the context of nuclear fusion devices, as they can have a severe negative impact on the performance of plasma-facing materials over extended periods of operation. In order to gain a more comprehensive understanding of these PWI processes, it is necessary to accurately identify the impurities and the manner in which they are distributed on the first wall. As such, the laser-induced breakdown spectroscopy (LIBS) technique is a powerful spectroscopic analytical tool that offers a range of advantages compared to other analytical methods. These advantages include its in-situ portability, its multi-elemental capability, and its fast analysis speed. These features make LIBS a highly attractive method for the composition and depth profile analysis of a nuclear fusion device, allowing for a more reliable and efficient assessment of the PWI processes that may be encountered[1,2].

In this paper, a novel and compact handheld LIBS system, including a diode-pumped Q-switched laser, a main unit with spectrometers, power supply, the control and processing software, was developed and tested. A preliminary measurement was carried out under atmospheric conditions after the 2021 EAST summer campaign. The calibration-free laser-induced breakdown spectroscopy (CF-LIBS) method was employed to calculate the concentration of the impurity depositions. Results showed the distribution of impurities, such as tungsten (W), molybdenum (Mo), iron (Fe), and lithium (Li) on the limiter and indicated that impurities deposited at the center of the limiter may have been transported by the plasma flux during the transportation process. The results of depositions also indicated that the high-Z impurities of W and Mo were derived from the tungsten components and the Mo first wall, respectively. In addition, trace carbon impurities were observed on the surface of both the upper and lower divertor, which is likely due to the transportation from graphite limiter and protect tiles. It was also noted that, due to the strongly broadening effect of the Hα line at 656.28 nm under atmospheric pressure, the Dα line of 656.1 nm could not be clearly observed. All of this evidence suggests that the handheld LIBS system mounted on a robot arm is a promising way to measure impurities and fuel retention in situ in a nuclear fusion device, which can enable the continuous monitoring of the impurities and the fuel retention in real-time, thereby facilitating the operation of the nuclear fusion device.

[1] S. Almaviva, L. Caneve, F. Colao, et al., Fusion Eng. Des. 157, (2020)

[2] S. Almaviva, L. Caneve, F. Colao, et al., Fusion Eng. Des. 169, (2021)

Plasma-facing components (PFCs) in nuclear fusion reactors are exposed to harsh conditions during operation. The combination of thermal loads, plasma exposure as well as neutron induced damage and activation limit the materials suitable for this application. Due to its unique properties, tungsten is foreseen as plasma-facing material (PFM) for the future DEMOnstration power plant. It is considered as such due to its exceptionally high melting point, excellent thermal conductivity, low tritium retention and high erosion resistance during plasma exposure. But even PFCs with tungsten armor have a limited lifetime due to, among many other factors, surface erosion and the resulting thickness reduction of the armor material.

In-situ local deposition of tungsten by means of the wire-based laser metal deposition (LMD-W) process could counteract surface erosion and thus increase the service life span of PFCs. First trials were carried out to check if it is possible to reliably deposit tungsten onto tungsten substrate using the LMD-W process. In these first attempts, single beads were generated, and in later experiments, entire layers were created from several beads which are arranged next to each other. To ensure reproducibility of the results, the substrate temperature was kept constant. Further experiments were aiming at the elimination or minimization of problems such as oxidation, arbitrary occurrence of balling defects, porosity, cracking, surface waviness and failed connection to the substrate. In order to increase the bead quality, the input parameters like laser power, tool speed, wire feed rate, inert gas flow, as well as the wire position were optimized. Furthermore, stacking of several layers, as well as the remelting of an already created layer, was carried out and investigated. This study represents the first steps in testing the feasibility of a tungsten surface regeneration concept for PFCs.

One of the challenges of any major engineering, first-of-a-kind undertaking such as the ITER project is that as component designs evolve in any given system, due to issues encountered during the design, there is an impact on others, which must adjust accordingly. One such case, are the ITER inner wall First Wall Panels (FWP), which are being subject to new constraints in several ways. For example, Remote Handling (RH) space tolerances required for FWP manipulation have had to be revised, forcing an increase in the horizontal and vertical gaps between the inner wall FWPs. These same inboard panels will also be the most heavily loaded during plasma current ramp-up and it has become clear only recently that longwave misalignments, coupled with other FWP manufacturing tolerances, may result in power handling issues in the early phase of discharges [1].

In view of the need to freeze the FWP models ready for series production, a very significant sensitivity analysis has been undertaken within the ITER Internal Components Division to map out the expected front surface heat loads in the critical inboard current ramp-up region. The assessment, reported in this paper, has been performed using the ITER Organization’s magnetic field line tracing environment, SMITER [2]. Comprising more than 100 individual 3D traces, the simulation database includes variation of all parameters playing a significant role in determining power fluxes. On the plasma side these are the ranges of heat flux channel widths, λq,near and λq,main defining the radial profile of parallel heat flux in the limiter plasma scrape-off layer [1] and the plasma currents, toroidal magnetic fields and magnetic equilibria, expected during the current ramp phase from circular to full bore elongated limiter plasmas. Regarding FWP engineering parameters, the assessment scans both short and longwave radial misalignments, ΔSW and ΔLW, respectively between toroidally neighbouring panels and the toroidal magnetic field structure. It also tests the impact of FWP tilting and examines possible modifications to front surface geometries to mitigate localized heat flux peaking.

Results demonstrate how power fluxes on the inboard FWP were found to be strongly sensitive to λq,near and ΔLW misalignment variation in accordance with [1] and result in an update of penalty factors previously defined in [3]. The study reveals that even when the ΔLW misalignment is reduced to zero and the scrape-off-layer parameters are in the ideal expected range, the combination of other misalignments with some plasma parameters could also provoke heat flux loads on the inboard FWPs closer to the design limit. Those factors are fundamental in order to know which parameters will be critical in the assembly of the FWPs as well as in ITER operation.

[1] Pitts, R.A. et al, Nucl. Fusion 62 096022 (2022)

[2] Kos, L. et al, Fus. Eng. Des. 146, 1796 (2019)

[3] Mitteau, R. et al, Fusion Eng. Des. 85, 2049 (2010)

*Corresponding author: e-mail: francisco.fernandezmarina@iter.org (F. Fernández-Marina)

In a tokamak, Radio Frequency (RF) antennas are subject to heat loads, coming from different sources. On average, the plasma radiation is the main contributor to the total power absorbed, but energy carried by accelerated ions and electrons may be responsible for larger heat fluxes locally.

To protect the RF antennas, poloidal limiters are installed on their sides. In WEST, ten antenna protection limiters are present since 1998 to protect the five antennas. They are made of Carbon Fiber Composite (CFC) brazed onto CuCrZr, and actively cooled by water. To prepare for WEST operations, the CFC was coated with tungsten in 2016.

Since then, limiters coatings have been damaged on some areas, allowing carbon to be eroded by plasma. Although coating reparation is technically possible, it would be time consuming, with no available spare in case of failure.

Therefore, design of new antenna limiters has been launched at CEA/IRFM. The key idea is to get rid of the carbon on the new limiters, while enhancing the maintainability, by easing the replacement of a damaged part. The main constrain is that the new limiter must have the same mechanical interface and performance than the existing one.

This paper presents the design of antenna limiters, composed of ten individual modules. Each module is made of CuCrZr assembled on 316L stainless steel by explosion bonding, with a cooling channel drilled from the backside. Coating can then be applied on the CuCrZr to improve the plasma compatibility. It will be either tungsten, or boron (recently developed process).

Results from different simulations are shown, starting with electro-magnetic loads during disruptions, and effects on the attachments. Thermal simulations are then presented, with local loads up to 10MW/m². Finally, water flow is analyzed, to ensure that the fluid velocity is relatively homogeneous, while pressure drop remains acceptable.

In future nuclear fusion reactors such as ITER, the tungsten (W) divertor will withstand hydrogen isotopes and helium (He) ion flux of the order of 10²⁴ m^-2 s^-1. The interaction of charged particles with W can induce modifications in the material, such as surface blisters (in the case of deuterium, D) and near-surface bubbles (for He) [1-3]. Such modifications can be responsible, for example, of an increased fuel inventory in reactor walls [4]. Moreover, D and He implantation can affect the optical properties of W, due to both a change of surface roughness and a change in the electronic properties of implanted W. A cursory knowledge of the evolution of the divertor’s optical properties during interaction with plasma may lead to inaccurate thermography measurements of plasma-facing components [5], which are realized through optical diagnostics whose response are based on the emissivity and reflectivity of the wall materials. For this reason, it is fundamental to improve the understanding of the variation of W optical properties during plasma exposure.

In this work, we investigate the evolution of W optical properties during the exposure to D and He plasma. All experiments were performed at Aix-Marseille University using the RF plasma chamber of the CAMITER setup, where in situ ellipsometric measurements (400-1000 nm) were carried out on polycrystalline W exposed to D and He plasma, exploring different ion flux values in the 10¹⁹-10²⁰ m^-2 s^-1 range. Such optical technique allows a real-time monitoring of the sample modification (in the first 100 nm) during plasma exposure, and through adequate modelling one is able to retrieve the plasma-exposed material optical constants n and k. Upon plasma exposure in CAMITER, the sample temperature increases up to about 300°C. Thus, we additionally studied the effect of sample temperature onto ellipsometric measurements and a proper modelling of the observed change [6] was used to subtract the temperature effect from the ellipsometric measurements. Atomic Force Microscopy and Confocal Microscopy was used to measure the change of roughness due to plasma exposure. Taking into account the variation of W temperature and roughness, we observed a small change in the reflectivity of W (~1%) due to D ion implantation. Preliminary measurements on He implantation will be presented.

In metallic fusion devices, parasitic light coming from multiple reflections on the wall is a major problem for interpretation of optical diagnostics. This is especially a critical issue for infrared (IR) thermography systems, ensuring the protection of plasma-facing components (PFCs). Indeed, the collected signals by cameras will include an important part of reflected flux, causing a high error in the surface temperature measurement [1]. This may lead to detecting false hotspots, which compromises the first function of IR diagnostics for monitoring heat loads on the first wall and the divertor of future devices ITER.

One possibility to cope with this reflected light is to use photonic simulation, which can accurately predict the behavior of light within complex 3D geometry. A prerequisite is to get a good description of the reflection model, represented by the Bidirectional Reflectance Distribution Function (BRDF), based on optical measurements of in-vessel materials. To avoid complicated measurements using goniophotometer to get the BRDF, one possibility is to link surface optical properties and topography characteristics, like roughness measurements, for example, using the classical Bennett's formula.

In this paper, we carried out BRDF measurements using two experimental goniophotometers to fully characterize the BRDF of tungsten samples with different roughness [2, 3]. Surface topography was measured using a three-dimensional laser scanning confocal microscope. From these measurements were extracted several parameters like the arithmetic average roughness (Ra), the root mean square roughness (Rq), the Surface Inclination Angle Distribution and furthermore its mean value d_m and the power spectral density (PSD). The correlations of BRDF model parameters deduced from the measurements are compared to the previous topographic parameters. The first results on several tungsten samples show that Ra, which is the usual measure of surface roughness, is not the most suitable metric to link with the reflection behavior of the surface. In contrast, the PSD and the surface inclination angle are interesting metrics for describing the reflected light.

[1] M-H. Aumeunier, et al., Nuclear Materials and Energy 26 (2021) 100879

[2] M. Ben Yaala, et al., Rev. Sci. Instrum. 92.9 (2021) 093501

[3] H. Natsume, et al., Plasma Fusion Res., 16 (2021) 2405019

The casting has become the widely used manufacturing process for bonding copper(Cu) as an interlayer between tungsten(W) block and CuCrZr cooling tube of divertor [1-2]. In order to reduce the pores inside Cu, a method of pressurizing gas through a porous tube during the casting cooling stage has been devised, and as a result, the mechanical strength of Cu was also improved. In the specimens manufactured by the gas pressure casting method, the amount of pores inside Cu was reduced and the interface between W and Cu was free of defects. SEM and EDS analysis showed Cu was diffused into tungsten to a depth of 2~3 μm and the hardness of Cu was increased by about 15% [3]. To evaluate the performance of the casted W/Cu monoblock, several PFC mock-ups for the high heat flux(HHF) test was produced. The W/Cu monoblock and CuCrZr heat sink was joined by hot radial pressing(HRP) method. The mock-up’s robustness to cyclic heat loading was tested by HHF at 10MW/m² up to 5,000 cycles. No abnormalities occurred during the HHF test in the mock-up specimen manufactured by the HRP method.

High heat flux effect on the tungsten monoblocks were observed by measuring microstructure and Vickers hardness on the cross section of the HHF tested monoblock specimens [4]. Though different performance was observed of the mock-ups under the HHF condition depending on the joining method; HRP or Brazing between the Cu interlayer and CuCrZr heat sink tube, the casted Cu interlayer itself remained intact. The material properties of tungsten armor showed degradation effect up to 500 mm manifested by the microstructure and hardness analysis from the surface into the depth. The degradation was influenced by the HHF test history and resultant heat removal capability of the PFC mock-ups.

The Material Plasma Exposure eXperiment (MPEX) is a linear plasma device being built at Oak Ridge National Laboratory. This device has been designed to be capable of reaching ion fluences up to 10³¹ m⁻². The MPEX device final design is nearly complete, and civil construction work has begun. This device will be used for plasma–material interaction (PMI) studies by exposing neutron-irradiated materials to divertor-relevant plasmas. These studies will help develop an understanding of plasma’s effects on materials and contribute to developing fusion reactor materials that can withstand high-heat fluxes and high ion fluences. MPEX will generate a bidirectional plasma using a high-power helicon source (200 kW, 13.56 MHz). The target is the last plasma-facing component on the downstream side of the MPEX device. The target’s primary function is to hold the material sample as it is exposed to fusion prototypic plasmas. The target assembly interfaces with the target exchange cart (TEC), which retrieves the sample from the PMI region and transports it to the surface analysis station (SAS) for analysis. The target assembly is designed to withstand 10 MW/m² heat fluxes from direct exposure to the plasma while remaining a removable component for post-PMI analysis. The target assembly design consists of a water-cooled Glidcop AL-15 body, a tungsten clamp to hold the target sample, and Grafoil. Grafoil is used as a thermal interface material for an effective transfer of heat from the target sample and clamp to the heat sink. Computational fluid dynamics and structural analysis were performed to optimize the design. A target assembly article was produced and tested for thermal performance and reliability using an electron beam up to 4 MW/m² heat flux. This paper discusses the details of the target design, the article’s high-heat flux test results, and lessons learned from high-heat flux testing that were incorporated into the final design.

Structured tungsten surfaces could be a promising option for plasma-facing materials in future nuclear fusion reactors. They have advantages over conventional tungsten such as better radiation resistance and, in particular, a strong retardation of tungsten fuzz growth under helium bombardment [1]. In addition, also the sputtering yield of nano-columnar tungsten is significantly reduced in comparison to flat surfaces as was already shown for seeding gas ions [2]. In a series of experiments, the role of surface topography on the sputtering behaviour of nanostructured tungsten surfaces was investigated [2, 3]. One clear result was that the surface inclination angle distribution is crucial for describing the sputtering of conventionally rough surfaces. This distribution can be determined from topographic surface data acquired via Atomic Force or Secondary Electron Microscopes. Understanding the origin of the sputtering yield reduction of conventionally rough surfaces, it was possible to optimize the surface by implementing nanoscale structures to thereby achieve very low sputtering yields with the benefit of a significantly improved radiation resistance.
A systematic comparison between flat and oriented nanocolumnar [2] tungsten surfaces will be presented, using results from experiments and simulations by Binary-Collision-Approximation-based or Molecular-Dynamics codes.

[1] W. Qin, et al., Acta Mater. 153 (2018) 147

[2] A. Lopez-Cazalilla, C. Cupak, et al., Phys. Rev. Mat. 6 (2022) 075402

[3] C. Cupak, et al., Appl. Surf. Sci. 570 (2021) 151204

As known to all, W and W-based alloys are the most promising candidate material for plasma-facing components of fusion reactors, while W has fatal weakness of brittleness, i.e., low temperature brittleness, high temperature brittleness, recrystallization brittleness and irradiation brittleness. Researchers from Institute of Nuclear Materials (INM), University of Science and Technology Beijing (USTB) have studied W-based alloys via experiments and theoretical. According to first-principles simulation, Y atoms at grain boundaries (GB) can effectively strengthen GB when Y concentration lower than 3.367%. Besides, it is found that incorporation of K atoms results in a slight increase of GB fracture energy. What’s more, microstructure and tensile performance detection of K doped W (KW) indicate that K bubbles can inhibit grain boundaries transverse movement, emit dislocation and barrier to dislocation movement, and improve the KW strength. Furthermore, K doping can significantly improve the high-temperature stability and mechanical properties without reducing its thermal conductivity, and effectively inhibit the formation and propagation of cracks during thermal shock. While Fe ion damage of KW can effectively inhibit surface blistering and reduce D retention, and KW shows better irradiation hardening resistance.

In addition to the plasma facing materials, the breeders of D-T fusion reaction also have been systematically studied, which are lithium-based ceramics Li₄SiO₄ and Li₂TiO₃. Li₂TiO₃ pebbles were prepared by hydrothermal and solid-state method combined with agar-based wet method. Crushing load and relative density of the obtained Li₂TiO₃ pebbles are 58.6N, 90.2% and 53.9N, 86.9%, respectively. Besides, lithography-based 3D printing was applied to prepare Li₂TiO₃pebbles, whose relative density, packing factor and compressive density are 92.3%TD, 80.99% and 84.4MPa. Using SiC as sintering aid via simple wet method Li₄SiO₄ pebbles were prepared, whose crushing load, relatively density and thermal conductivity are 65.3N, 90.3% and 1.97Wm^-1K^-1, respectively. What’s more, we proposed a facile rolling ball method to prepare Li₄SiO₄ sorbent pebbles, whose sphericity, crushing load, porosity, and CO₂ adsorption ability are 0.95, 65N, 6.28% and 0.304 g CO₂ /g sorbent, respectively.

In this study an efficient blanket module based on the HCPB structure with Li4SiO4 in the form of pebble bed breeder, Beryllium pebbles as neutron multiplier, and ODS ferritic steels as structural material simulated with the MCNPX code and COMSOL Multiphysics software. To reach a TBR of 1.15, the Li6 enrichment in the breeder beds should be increased 20% for the front of the module and up to 40% for the back of it. The coolant He at high pressure (8 MPa) and inlet temperature of 330 oC flows in the first wall and the breeding zone in small channels. The maximum He temperature is 540°C with respect to the upper design limit of 650°C for Be. In the figure below, a scheme of the blanket module is presented.

Based on the consideration of a full tungsten divertor, the ITER plans to employ tungsten divertor Langmuir probe (DLP)[1, 2]. In 2022, several ITER-like DLPs have been installed at the lower horizontal target in the EAST device, aiming to test the performance and function under long-pulse operations[3]. Due to the no actively cooling system, the thermal performance is unclear under the in-situ complex heat flux loading conditions. Thus, a transient thermal analysis employing ANSYS was performed on the DLP, in order to explore the melting threshold parameters as well as the cooling time according to the actual heat flux distribution features.

The FEM model considers detail actual conditions, such as the 3D structure and temperature-dependent materials’ properties, the engineering misalignment and leading edge effect, the magnetic configuration and plasma platforms, and so on. The parallel heat flux from 40-100 MWm^-2 which is current level of EAST was used. It is found that for lower current (<500kA) platform corresponding to an incident angle below 2.8^o, there is no plasma heat flux along magnetic field line that deposited on the DLP, and the radiation power from core plasma only induce a temperature rise about 277^oC during 1000s pulse, which is acceptable and safe. In case of higher current (>500kA) corresponding to an incident angle above 2.8^o, the plasma heat flux along magnetic field line will doubtly deposit on the DLPs, which induce the temperature increasing quickly with time extension. The threshold parameters, i.e. the special heat flux and its loading time, for the melting of W DLP, were obtained after a series of thermal simulations. Taking the 600kA current platform (corresponding to an incident angle below 2.8^o) and 100MWm^-2 parallel heat flux as an example for illustration, it only need 142s to reach to the melting point of W. In addition, the cooling time from the peak temperature to room temperature for different conditions were also obtained. It need at least 1.8 hour to return back to RT from melting point.

Such results on the thermal performance of W DLPs provide important data reference for the plasma operations of the EAST, and key experience for the application of W DLPs in the ITER.

[1] Y. Jin, W. Zhao, C. Watts, et al., Fusion Sci. Technol. 75, 120 (2019)

[2] W. Zhao, Y. Jin, L. Nie, et al., Fusion Eng. Des. 177, (2022)

[3] L. Meng, J. Xu, J. Liu, et al., Fusion Eng. Des. 175, 113011 (2022)

*Corresponding author: Dahuan Zhu, tel.: +86-0551-65593291, e-mail: dhzhu@ipp.ac.cn

Eurofer rectangular channels are foreseen to be adopted in both the DEMO WCLL-BB (Water Cooled Lithium Lead –Breeding Blanket) [1] and the ITER WCLL-TBM (Water Cooled Lithium Lead – Test Blanket Module) [2] for cooling the First Wall. The nominal cooling conditions considered for each channel are the PWR (Pressurized Water Reactor) conditions: 15.5 MPa with a target value of inlet/ outlet temperatures of 295°C/ 330°C [2]. This will result in boiling over a portion of the total length. Therefore, a two-phase thermal hydraulic analysis is required to predict correctly the cooling performance in the First Wall.

Structural mechanical analysis based on finite element method is usually used to solve complex structural engineering problems faster and more efficiently. This approach makes use of heat transfer correlations as boundary conditions. These semi-empirical correlations have been elaborated in the past (“Nukiyama correlations”) based on experimental data for one-sided heated ‘’circular’’ channels [3] and a limited range of hydraulic parameters as pressure, temperature and tube diameter. These correlations need thus to be adapted by corrective factors to cover the new one-sided heated “rectangular” geometry and the current PWR conditions expected in WCLL-BB and TBM. Therefore, it is important to establish another analysis approach that could evaluate the heat transfer in these conditions.

This work aims to evaluate the validity of the computational fluid dynamic (CFD) analysis (coupled approach) to perform two-phase thermal hydraulic analysis for heat transfer between the heatsink and the cooling water in case of circular and rectangular channels and for multiple regimes: single-phase (forced convection), two-phase (nucleated boiling and forced convection), and near critical heat flux (i.e. ebullition crisis). Comparison between steady-state thermal modelling using semi-empirical heat transfer Nukiyama correlations and CFD modelling (Fluent model) using Eulerian multiphase model (RPI boiling model) and SST k-omega turbulent flow is reported and discussed in this paper.

Additionally, the influence of main parameters (geometrical and thermal-hydraulic) is investigated to design relevant mock-ups dedicated to high heat flux tests (up to critical heat flux).

[1] P. Arena, A. Del Nevo et al, The DEMO Water-Cooled Lead–Lithium Breeding Blanket: Design Status at the End of the Pre-Conceptual Design Phase. Appl. Sci., 11 (2021)

[2] J. Aubert et al, Design and preliminary analyses of the new Water Cooled Lithium Lead TBM for ITER, Fusion Eng. and Des., 160 (2020)

[3] J. Boscary et al, Critical heat flux of water subcooled flow in one-side heated swirl tubes, Int. J. Heat Mass Transfer, 42 (1999)

KSTAR (Korea Superconducting Tokamak Advanced Research) which has been in operation since 2008 is a worldwide renowned research device using superconducting tokamak technology for fusion research. The final heating power target is set to about 24 MW by utilizing an additional heating system called NBI (Neutral Beam Injection) to enhance the plasma performance. The upgrade of the KSTAR divertor is essential to exhaust the increased heat load on the divertor. The upgraded divertor system must be able to handle high heat flux and cool down the excessive power generated in the scrape-off layer domain. The upgrade process for the KSTAR divertor began in 2019, and the new divertor is set to be installed and tested in 2023. The peak heat fluxes on the divertor target can reach up to 10 MW/m² and 20 MW/m² for steady-state and slow transition operations respectively.

The upgraded KSTAR divertor system basically applies the ITER-like divertor concept, which is a water-cooled tungsten monoblock. Tungsten is a highly durable material in high heat flux plasma conditions, and the combination of a copper alloy heat sink and pressurized water coolant is an effective cooling method due to its high heat transfer efficiency. The upgraded KSTAR divertor system has a single null configuration, and 64 cassette divertor modules will be placed at the bottom of the vacuum vessel. Each divertor module includes an inner target, central target, outer target, and a cassette body with supports to connect each part. In previous studies, CFD analysis was conducted to ensure the thermal stability of the entire divertor module [1]. Results showed that the design could operate within an allowable thermal range at 10 MW/m² heat flux. The temperature distribution from the CFD analysis is applied to one-way FSI (Fluid-Structure Interaction) for thermo-mechanical analysis. The structural analysis also considered the deadweight of the components, coolant pressure, pre-load at joints, and estimated electromagnetic forces. Finally, based on the ASME code, the upgraded KSTAR divertor design was evaluated for plastic collapse, ratcheting, fatigue and buckling, and it was found to be reliable from both thermal and mechanical perspectives. The high heat flux test was carried out in GLADIS to verify the thermomechanical integrity under the peak heat flux of 20 MW/m². The result shows no flaw like crack or delamination in the bonding surface.

[1] S. Kwon et al., Fusion Science and Technology 77 699-709 (2021).

Additively manufactured tungsten capillary porous structures (CPS) with liquid tin are a promising alternative for conventional solid divertor plasma-facing units to deal with high peak heat fluxes in a sustainable way. It gives a high level of flexibility in designing the CPS, significantly impacting its performance. Our goal is to design for maximum lifetime of the CPS. Some requirements are: the ability to remove high heat loads to keep the surface cool, reduce stresses and have sufficient wicking speed to replenish the tin in case of a disruption. Most proposals (e.g. [1]) use engineering insight, and although promising, we do not know if the selected design is optimal.

We apply an adjoint gradient-based numerical optimization technique, it can deal with multiple design variables and keep the computational cost low. This approach has been successfully applied for the design of functionally graded tungsten-copper monoblocks [2] and divertor shape optimization [3]. We aim to optimize the combined thermal and mechanical performance of the CPS by balancing the stress and temperature contributions in the objective function. We solve the optimization problem by applying a steepest descent method. The main design variable is the three-dimensional spatial porosity distribution of the structure, which is linked to a parameterization of the predefined unit cell topology. We propose a homogenization method to extract the effective material properties starting from the unit cell topology. By using both temperature and stress in the objective function, the optimization problem is non-convex. Hence different solutions are obtained depending on their relative importance. We visualize these solutions in a Pareto frontier diagram, and discuss how to go about selecting a design. Later, we will add more contributors to the model and use the Pareto frontier diagram introduced here to select the best design. Finally, we will print this design and confirm that the CPS meets the requirements for use in a fusion reactor.

KSTAR has been able to determine a suppression window of high energy edge localized modes (ELMs) in H-mode discharges by use of the n=1 mode of a resonant magnetic perturbation by fixing the top and bottom coil currents at 5 kA and varying the middle coil current and phasing as 𝜙_𝑀=110°−15𝐼_𝑀 [1]. The use of a resonant field minimizes ELMs but introduces asymmetries in the plasma edge, which manifest in lobes of concentrated heat flux along the divertor plates. A consequence of this is increased ion temperatures at the plasma edge which lead to large erosion fluxes of the target material and these impurities can lead to undesired effects in the plasma.

This work focuses on the use of the 3D kinetic Monte Carlo ERO2.0 code [2] to determine carbon erosion and impurity transport present in the double null configuration for the KSTAR ELM suppression window. EMC3-EIRENE plasma backgrounds are used for a range of RMP coil currents (I_M = 0, 1, 2, 3, 4, 5 kA) and a no RMP reference. The ERO2.0 code makes use of the full 3D wall geometry as the n = 1 mode is not periodic and travels along the full toroidal direction of the lower target. Erosion due to molecules is also taken into account as these rates are not negligible in H-Mode discharges.

Results for these simulations show strong carbon erosion of the lower outer divertor target plate, with toroidally averaged profiles showing decreasing net erosion with increasing coil current. As the coil current increases, the lobe footprint is widened which leads to decreased edge temperatures and subsequently, softer erosion profiles. This also leads to a decreased carbon density in the plasma with increasing coil current, with the baseline no-RMP scenario leading to insignificant erosion fluxes by comparison. When sampling data for the carbon charge states along the midplane, the higher states dominate in the core (Z=+4,+5,+6) with C+4 dominating in the scrape-off layer and extending all the way to the wall. Consequently, this can lead to increased erosion due to charge-exchange events as there is significant recombination at the far SOL. These results lead to the calculation of the Zeff in the range of 3-4 integrated across the full volume, decreasing with increasing coil current. When looking at the effect of the double null configuration in the results, the secondary SOL remains separate from the primary one, and insignificant carbon fluxes from the lower inner target are able to deposit along the upper divertor targets. By contrast, most of the carbon eroded in the outer lower divertor target (~80%) is deposited along the upper divertor target, and the remaining fraction is deposited along the first wall, which suggests a good location for impurity pumping.

[1] J.K. Park, Nature Physics, Vol 14, 1223-1228 (2018)

[2] J. Romazanov, Nuclear Materials and Energy, Vol 18, 331-338 (2019).

*Marcos Navarro: tel.: +1 608-598-9802, e-mail: navarrogonza@wisc.edu

An accurate prediction of material redeposition is critical to determining the lifetime and survivability of high-Z plasma facing components (PFCs) in magnetic confinement fusion. For clarity of the underlying physics, we categorize high-Z redeposition into three primary mechanisms: prompt (geometric-dominated), local (sheath-dominated), and far (scrape-off-layer-dominated) redeposition. This categorization allows the development of quantitative diagnostic methods to distinguish each mechanism and experimentally determine the dominate redeposition type by poloidal location. Towards this goal, we use an experimentally well-diagnosed plasma background via OEDGE [1] and the 3D, global Monte Carlo particle tracker GITR [2] to simulate total redeposition of an azimuthally-symmetric tungsten source in representative DIII-D divertor conditions. We discriminate each redeposition mechanism through charge state distribution and mapping particle trajectories. Moreover, we manipulate several key parameters in the plasma background that determine sheath thickness in order to predict the associated scaling for each redeposition mechanism.

Additionally, we explain with simple-as-possible models an observed radial difference in the maximal redeposition location between prompt, local, and far. By exploiting this radial difference, we may use radial deposition ranges as an ex situ technique using isotopic methods to experimentally discriminate the redeposition contribution of each mechanism. In particular, we propose a removable coupon with a surface containing concentric disks of disparate isotopes to maintain spectroscopic resolvability while allowing for ex situ surface discrimination (via deposition radius) within the spectroscopic field of view. This method proves useful for resolving existing discrepancies between current models of redeposition scaling. Furthermore, the proposed experimental workflow with a synthetic diagnostic framework helps to lay the foundation for comparing ex situ and in situ redeposition diagnostics, all towards improving predictive, high-Z net-erosion models for extending PFC lifetimes.

This work was supported in part by the U.S. DOE SCGSR 2020 Solicitation 2 award and DE-SC0019256.

[1] P.C. Stangeby, et al., J. Nucl. Mat., 313-316, 883-887 (2003)

[2] T.R. Younkin, et al., Computer Phys. Comm. 264, 107885 (2021)

In-depth understanding of near-surface modification of materials due to plasma exposure is decisive for sustainable operation of future fusion devices [1]. When studying properties of the outermost surface layers, exposure to ambient conditions can significantly alter materials properties, which severely affect the capability of advanced analytical tools to assess e.g. fuel retention or sputter yields from such surfaces.

In this contribution, we present recent upgrades to the Time-of-Flight Medium Energy Ion Scattering (ToF-MEIS) set-up at Uppsala University [2][3] which enable in-situ studies of near-surface material modification processes, highly relevant for studies of plasma-material interaction. Specifically, this system, which has earlier been successfully employed for studying temperature and irradiation-induced surface segregation of tungsten in EUROFER and FeW model films [4][5], has been equipped with an ultra-high vacuum sample preparation chamber capable of

a) (reactive) thin film deposition from multiple e^--beam evaporators.

b) (reactive) thin film deposition by DC/RF magnetron sputtering.

c) low-energy ion irradiation for sputtering, implantation or other applications using a keV ion gun with Wien-filter.

d) annealing of samples to temperatures of up to 1500 K using an e^--beam heater.

e) annealing of samples, including self-supporting targets using a retractable button heater to approx. 800 K.

f) residual gas analysis during experiments.

All as such prepared or modified sample systems can be analysed using ion beams of light and heavy species in backscattering, forward scattering and transmission geometries. Detection of scattered particles, photons [6], sputtered and desorbed species [7] and recoils [8] with two position sensitive detectors yields information on near-surface composition and crystal structure.

[1] S. Brezinsek et al., Nucl. Fusion 57 (2017) 116041

[2] M.K. Linnarsson et al., Rev.Sci. Instr. 83 (2012) 095107

[3] M. A. Sortica et al., Nucl. Instr. Meth. B 463 (2020) 16

[4] P. Ström et al., Nucl. Materials & Energy 12 (2017) 472

[5] P. Ström et al., J. Nucl. Materials 508 (2018) 139

[6] S. Lohmann et al., Nucl. Instr. Meth. B 417 (2018) 75

[7] S. Lohmann et al., Nucl. Instr. Meth B 423 (2018) 22

[8] R. Holeňák et al., Vaccuum 204 (2022) 111343

Physical sputtering is one of the important erosion processes in plasma surface interaction. For several decades, many research groups have contributed to the basic understanding of that issue [1], which resulted in well-established simulation techniques, e.g., the TRIM family of codes [2]. Mostly, the crystallinity of the material is not taken into account, even though it has long been known that the sputter yield can vary by more than one order of magnitude due to the crystal orientation [3]. Therefore, some aspects of the modelling still need some discussions, such as: Is the assumption of an amorphous/random sample for erosion prediction by simulation programs adequate? How is the crystallinity of the material treated in the simulation programs? Are the available experimental data sufficient to benchmark simulations including crystallinity? In order to deliver such experimental data, we developed a measuring and evaluation strategy to obtain the physical sputter yield for quasi all crystal orientations [4,5].

The development in the detector technology for electron backscattering diffraction (EBSD) performed in a scanning electron microscope (SEM) allows the analyses of sufficient large areas of polycrystalline samples in a reasonable time in respect to their crystal orientation. Combining the EBSD data of areas with their three-dimensional profile data obtained (e.g., by confocal laser scanning microscopy (CLSM)) after sputtering results in sputter yield determination for a huge number of crystal orientations. The results can be visualized in so-called inverse pole figures (IPF), a representation of all possible crystal orientations. For the combination of EBSD and CLSM data and their evaluation, a Python script was written [4]. First, physical sputter yield data obtained for tungsten bombarded by gallium ions were compared to simulations of crystalline material with remarkable agreement [5].

In this contribution, the results for sputtering by Ga ion with energies between 2 and 30 keV will be presented for four metals: two with body-centred cubic lattice structure (tungsten, molybdenum) and two with face-centred cubic structure (copper, platinum). These results will be discussed with respect to accepted models, but also in the light of recent simulations and their extrapolation potential, especially for predicting erosion in fusion devices.

[1] Review books: “Sputtering by Particle Bombardment I-IV”, ed. R. Behrisch, Topics in Applied Physics vols. 47 (1981), 52 (1983), 64 (1991) and 110 (2007).

[2] H. Hofsäss, K, Zhang, A. Mutzke, Applied Surface Science 310 (2014) 134–142.

[3] H.E. Roosendaal in “Sputtering by Particle Bombardment I”, ed. R. Behrisch, Topics in Applied Physics vols. 47 (1981).

[4] K. Schlueter, M. Balden, T. da Silva, Int. J. Refract. Met. Hard Mater. 79 (2019) 102–107.

[5] K. Schlueter, et al., Phys. Rev. Lett. 125 (2020) 225502.

During the last decades, substantial modelling work has been performed in order to simulate material sputtering process under ion impact. This phenomenon is critical in numerous situations, especially in nuclear fusion devices [1]. It has been identified that ion sputtering could be a critical process in the context of magnetic confinement fusion (MCF) as it limits plasma facing materials lifetime and it induces plasma core contamination by sputtered material transport [2], [3]. This phenomenon could eventually jeopardize nuclear fusion process. Ion sputtering modelling is thus a key element in order to estimate plasma facing material lifetime and improve them. Several codes such as TRIM, SDTrimSP, TRIDYN and ACAT have been developed for sputtering modelling, while SPRAY, IM3D and ERO2.0 perform particles transport simulation in nuclear fusion devices. Another code, called Csipi, has been developed at ONERA for the simulation of material erosion under Xenon ion irradiation in space applications and based on the same physical processes.

Comparisons have been carried out between these codes at high incident ion kinetic energy (above 1 keV) [4]. However, it has been evaluated that plasma ions will hit the tokamaks walls with kinetic energies closer to 100 eV than to 1000 eV [5]. Besides, when these codes take into account the surface roughness, their conclusions can be contradictory [5]–[7]. This suggests a lack of understanding of ion sputtering phenomenon in presence of surface roughness.

In this work, the ONERA code CSIPI is presented and compared with the other codes in order to assess their pros and cons as far as MCF applications are concerned. Simulations are run for low incident ion energy (below 1keV) and by taking into account miscellaneous surface roughness. It is evaluated in which situations these codes are diverging in order to assess their representativeness and identify possible lacks in the modelling of the physical processes taken into account. This allows determining which cases can be considered critical and could be used in the future to improve ion sputtering models.

[1] J. Linke et al., Matter Radiat. Extrem., vol. 4, no. 5, p. 056201, Sep. 2019.

[2] I. S. Landman et al., J. Nucl. Mater., vol. 337–339, pp. 761–765, Mar. 2005.

[3] J. N. Brooks, in 2014 IEEE 41st International Conference on Plasma Sciences (ICOPS) held with 2014 IEEE International Conference on High-Power Particle Beams (BEAMS), May 2014, pp. 1–1

[4] H. Hofsäss et al., Appl. Surf. Sci., vol. 310, pp. 134–141, Aug. 2014.

[5] A. Eksaeva et al., Phys. Scr., vol. 2020, no. T171, 2020.

[6] A. Eksaeva et al., Nucl. Mater. Energy, vol. 19, pp. 13–18, May 2019

[7] A. Eksaeva et al., Nucl. Mater. Energy, vol. 27, p. 100987, Jun. 2021.

*Corresponding author: e-mail: luca.chiabo@onera.fr (L. Chiabò)

Dust-wall interactions in plasma facing material can have a significant impact on the performance and lifespan of the tokamak [1]. Factors such as temperature, dust velocity, and dust size all play a role in determining the extent of this interaction. High temperatures in the tokamak can cause dust particles to vaporize upon impact with the wall, leading to increased erosion and damage. Additionally, high dust velocities can also contribute to increased erosion due to the kinetic energy of the impacting particles [2]. The size of the dust particles also plays a role; however, the effect of particle size on the interaction is complex and depends on several factors, such as the size distribution of the particles. We conduct an atomistic study to investigate and gain new insights into the interactions between dust and walls in tokamaks and how these interactions can alter the mechanical properties of the target material. We use extensive molecular dynamics simulations to examine the atomic processes that occur when dust impacts the target wall at different velocities. We simulate large samples containing up to 200 million atoms. The study examines various aspects of the dust impact, such as the temperature, dust velocity, and dust size, all play a role in these interactions and provide an in-depth understanding of the physical mechanisms involved, which will help to minimize the damage caused by dust-wall interactions and improve performance of the device.

References:

[1] Ratynskaia, S., Bortolon, A. & Krasheninnikov, S.I. Rev. Mod. Plasma Phys. 6, 20 (2022).

[2] C. Castaldo, S. Ratynskaia et al, Nucl. Fusion 47, L5–9 (2007).

Erosion, transport and redeposition of plasma facing materials lead to mixed-material formation during operation of fusion devices; such as, tungsten beryllide (W-Be) formation on the W divertor surface in JET [1] (and, as is expected in ITER). Further, the resulting alloy formation and growth are very sensitive to implantation conditions (e.g., deposition rate and temperature) [1,2]. Therefore, a model capable of capturing mixed material formation and growth, and its dependence on reactor-relevant parameters is needed to accurately predict material performance in future fusion reactors. In this work we outline a staged approach to developing and benchmarking a W-Be mixing model against film-deposition and linear plasma experiments, and present the atomistic work used as basis and for parametrization of the model.

Molecular Dynamics (MD) simulations were conducted using the Spectral Neighbor Analysis Potential (SNAP) for W-Be to understand the thermodynamic and kinetic drivers for intermixing of Be in W. Previous static atomistic calculations indicated that Be forms well-ordered structures on defect-free W surfaces [3]. However, our simulations find that the presence of defects on the W surface, such as W vacancies, can trigger Be intermixing, with a progressive growth of an intermetallic layer that evolves towards WBe₂, thermodynamically the most stable phase per 0 K calculations. These and other [2] observations imply that the intermetallic layer growth is dependent on the transport of Be through the intermetallic phases. Hence, the diffusivity of Be interstitials was studied in each stable W_xBe_y convex hull.

Alongside this, we use atomistic simulations to parameterize Be loss mechanisms in the W-Be mixing model. MD simulation of 60 eV Be implantation in the three stable intermetallics, WBe₂, WBe₁₂, and WBe₂₂, at 1000 K, found that the depth profiles were similar for all three intermetallics but the sputtering rate was higher and Be retention lower for WBe₂. Static atomistic calculations [4] show that surface and adatom binding energies of Be to the surface of stable intermetallics is too large to expect evaporation in the experimental range (<1200 K). Consistent with these calculations, our MD simulations found no significant evaporation, with only a few atoms leaving the surface when an adatom island or an unfinished layer were present, while the material was heated from 300 K to 2000 K.

The diffusion-driven mixing model parametrized by these atomistic simulations agrees with W-Be mixed layer growth rates measured in Be deposition experiments [2]. Future model extensions, including Be loss mechanisms, will also be discussed.

[1] M. Rubel et al., Nuclear Fusion 57 (2017)

[2] M.J. Baldwin et al. Journal of Nuclear Materials 363 (2007)

[3] P. Hatton, et al. Acta Materialia 235 (2022)

[4] Chen, L., et al. Nuclear Materials and Energy 16 (2018)

The main concerns with respect to the presence of contaminants in tokamak plasma are related to the safety and performance of the fusion machine operation [1]. Apart from the air (a common contaminant for any vacuum process), another possible contaminant is water which is a coolant candidate to be used in ITER. Here, we addressed the issue of air and water leaks over the dust production of beryllium, using ball milling. The aim of these studies is to improve the prediction of Be dust generation from plasma facing components.

Beryllium dust was obtained using a ball milling procedure of a beryllium pebble in air and water environment and collected on different size range by sieving (i. >160 µm, ii. 90-160 µm, iii: <36 µm). The dust shape and size obtained by ball milling were similar with thise collected from the JET machine. Formation of oxides and nitrides as follows: i) BeO, H₂O, N₂, Be(OH)₂ content - higher for wet environment samples, ii) Be, O₂ and Be₃N₂ content - higher for dry condition samples were highlighted.

A ball milling device (Retsch type) having alumina bowl of 250 ml and 40 alumina balls of 10 mm diameter was used to produce beryllium powder. As raw material was used Be flake, (Good Fellow) of 10 mm maximum flake size, electrolytic. The milling conditions: i) Milling in air atmosphere: 5g Be flake, 40 alumina balls, 200 rpm (8÷30 hours), ii). Milling in water: 5g Be flake, 50 ml distilled water, 40 alumina balls, 200 rpm (8÷30 hours). Beryllium dust was collected on different size range by sieving ( i.>160 µm, ii. 90-160 µm, iii: <36 µm).

Rough surfaces were identified using SEM technique on dust particles prepared in dry and wet condition. The dust collected from JET and the one prepared by ball milling has shown similar shape and morphology as revealed by SEM analysis.

Crystalline phase composition analyzed using Rigaku ULTIMA IV X-ray diffractometer apparatus shows peaks of beryllium, beryllium oxide and Al₂O₃ and SiO­­­₂. The observed shifts to small angles suggest the modification of interplanar spacing of the Be crystalline lattice, due the presence of a uniform stress state responsible for the lattice strain. The BeO diffraction peaks are present for samples prepared in wet condition. Alumina (Al₂O₃) and agate (SiO2) used in the milling process appeared too: more obvious in “wet” conditions.

Using the Thermal Desorption System developed at NILPRP, the behavior of different dust powders prepared in wet and dry conditions during 10; 20; 35, 50 and 61 hours of milling process was studied. Were highlighted the followings: i) formation of oxides and nitrides: BeO, H₂O, N₂, Be(OH)₂ content - higher for wet environment samples, ii) Be, O₂ and Be₃N₂ content - higher for dry condition samples, suggesting that the water is present in Be dust as: a) pure H₂O non-reacted; b)oxygen to form BeO; c) (OH) to form Be(OH)₂ . The maximum desorption temperatures were found for H₂O~190^oC and BeO~ 540^oC.

[1] M. Shimada, R.A. Pitts, S. Ciattaglia, et al., In-vessel dust and tritium control strategy in ITER, J. Nucl. Mater., 438 (2013), p. S996

The stellarator W7-X operated with a carbon divertor, the so-called Test Divertor Unit (TDU) during the operational phase OP 1.2 [1, 2]. During the whole operational phase about 70±20 g carbon were eroded at the strike line of the TDU. Some redeposition was observed in poloidal direction left and right of the strike line, but this amount of redeposited carbon was by far smaller than the observed carbon erosion.

Carbon-containing deposits were observed on all first wall components. These were analysed using elastic backscattering spectrometry (EBS) with incident protons at 2.5 and 3.5 MeV incident energy, with scanning electron microscopy (SEM), and with focused ion beam cross-sectioning (FIB).

Thick deposits exceeding thicknesses of 1 μm were found on some baffle tiles, the thickest observed deposits exceeded thicknesses of 10 µm. Moderately thick deposits with inhomogeneous thickness profiles were found on toroidal closure tiles, while only thin deposits were observed on heat shield tiles and the outer wall panels. The surface of deposits was sometimes smooth, but sometimes showed characteristic, conical structures parallel to the surface. Deposits consisted mostly of carbon with 20-40 at.% of hydrogen, 10-20 at.% oxygen, and <10 at.% of boron. Some deposits flaked off. The rupture line was sometimes between deposit and substrate, but sometimes also inside the layered structure of the deposit. The appearance of flaking was not directly connected to layer thickness or layer composition.

In total about 30 g of carbon were found in redeposited layers or were pumped out as volatile carbon-containing molecules such as hydrocarbons or as CO or CO₂ [3]. The amount of observed redeposited carbon is therefore within a factor of about two to the amount of eroded carbon. Keeping the highly inhomogeneous redeposition pattern in mind, this can be considered as reasonable agreement.

[1] M. Mayer et al., Phys. Scr. T171 (2020) 014035

[2] M. Mayer et al., Nucl. Fusion 62 (2022) 126049

[3] C.P. Dhard et al., Phys. Scr. 96 (2021) 124059

Microstructural analysis of JET deposition was carried out for a tungsten lamella from a shadowed region of divertor tile 5 retrieved from the first ITER-like wall campaign (ILW1). Tile 5 was subjected to ITER relevant deuterium plasma loading scenarios and are therefore extremely valuable for studying synergistic thermal and particle loading on tungsten. A C3 lamella exposed in JET during ILW1, when the outer strike point remained mostly on tile 5, has been studied. The lamella was located in a region shadowed from the plasma by the preceding tile 5 assembly and as a result over 20 tungsten droplets, originating from elsewhere in the divertor, remained on the surface. TEM liftouts were prepared from the droplet-lamella interface and EELS analysis showed that there was beryllium at this interface. The oxygen signal confirmed that most of the beryllium had formed BeO which was decorating void-like structures in most areas and had formed layers of up to 200 nm.

Beryllium was not observed in the absence of a droplet but could however be present on the surface in concentrations below the limit of detection of the techniques used here. Ion beam experiments, separate from this study, have been able to detect low levels of beryllium on the surface of lamellae in this region of tile 5. However the data presented suggests that beryllium is associated with the tungsten droplet and therefore arrives as a part of a tungsten melt event. Possible sources of tungsten melt events are Langmuir probes or flaking from tungsten coated carbon fibre composite tiles, both of which may contain some beryllium from deposits before melting.

Furthermore, underneath every droplet there were bubbles up to 100 nm in diameter which extended a few microns into the surface of the lamella. The correlation between the two suggests that the droplets may be involved with the formation the bubbles. EELS linescans of the bubbles followed by a multiple linear least squares method revealed the onset of a shoulder in the spectrum at 14 eV. Similar spectra have been seen in beryllium and zirconium which suggests that these bubbles were filled with a hydrogenic gas. Bubbles or voids of a similar size and distribution were also present on the surface of the as received material which likely occurred due to the evaporation of impurities during sintering and rolling.

Bubbles and void-like structures have widely been observed in tungsten as a result of helium coalescence or the evaporation of impurities. However, there was no helium injected into JET during ILW1, neither is it detected by EELS in the exposed lamella and was therefore not linked to void formation. The molten droplets are thought to be responsible for introducing impurities and sufficiently high temperatures to create the voids. These impurities then evaporate in a similar manner to those during manufacture and hydrogenic gas diffuses into the voids from the plasma.

The formation of void networks near the surface of the lamella increases the probability of dust creation. Some regions may become thermally isolated which causes recrystallization and grain boundary weakening. Furthermore, these regions are more likely to flake off in the presence of a W-BeO interface which introduces a discrepancy in thermal conductivity and coefficient of expansion.

Plasma-facing materials in tokamak nuclear reactors are subjected to intense heat and radiation and should be able to withstand high temperatures and yet hold a tight seal to resist the influence of the high-temperature plasma. One of the potential materials for the divertor part of the tokamak wall is liquid metals, which has several advantages over solid divertor wall materials. Hence, it is crucial to study the surface of the liquid metal divertor (LMD) to understand its material migration, re-deposition, and fuel retention on Plasma-Facing Components. Materials like lithium (Li), lead (Pb) and eventually tin (Sn), or the combination alloy of these elements, are preferred as the materials for the liquid metal divertor. The study of sputtering and material erosion can be done using LIBS considering its effective quantitative in-situ diagnostics of the first walls during maintenance breaks. Quantitative information including depth profile analysis could be obtained by Calibration-Free Laser Induced Breakdown Spectroscopy (CF-LIBS), where the elemental composition at each depth of a sample is determined based on the accurate knowledge of the electron temperature Te and the electron density ne. The high-power handling capacity of Li and LiSn-based divertor modules based on a Capillary Porous System (CPS) has been tested in COMPASS tokamak during two experimental campaigns. The samples after the CPS experiment have been analyzed by the CF-LIBS method and were also used for the depth profile analysis of the deposited layer on the Ni-Cr based screws which were situated at different places in the COMPASS tokamak vessel during the first week of Li LMD campaign measurements. The LIBS experiment was carried out using a Nd:YAG laser operating at 1064 nm and producing a pulse of 5 ns. The laser ablation was performed at atmospheric pressure [1]. Optimized experimental conditions were configured for the detection and quantification of Pb in traces and bulk using CF-LIBS [2]. In this work, we present the CF-LIBS depth profile analysis of the deposited layer on the Ni-Cr based screws which were situated at different places in the COMPASS tokamak vessel during the second LiSn LMD campaign measurements. A new approach of differential LIBS spectra analysis was applied, and the simulated Ni and Cr LIBS spectra were subtracted from the measured one, for better analysis of weak Sn lines and trace elements quantification of B and C. The simulation of LIBS spectra requires the knowledge of experimental parameters like electron temperature (Te) and electron density (ne) beforehand. The parameters were evaluated from non-subtracted LIBS spectra using mainly Ni and Cr lines. The simulated spectra are from the NIST database [3], and the results are compared with those obtained from conventional CFLIBS quantification.

Erosion and deposition resulting from plasma surface interactions within magnetic confinement fusion reactors may lead to restrictions on the selection of plasma-facing materials (PFMs). Specifically, the formation of codeposits, which trap the hydrogenic fuel (crucially tritium), results in both safety and economic concerns.

Ceramic materials have recently garnered attention as potential PFMs due to their good thermal properties and high operating temperatures. Additionally, low-Z ceramic materials have the advantage of high plasma contamination tolerance. Partly for these reasons, silicon carbide (SiC) has been proposed as a first wall material in the ARIES-AT reactor design [1]. More recently, SiC-coated tiles have been tested in the DIII-D tokamak [2] and the recent General Atomic’s demonstration reactor design involves a W/SiC composite first wall structure [3]. While carbon-based materials are known to have high affinity for hydrogen, there is currently little information available pertaining to the hydrogen retention properties of Si-containing deposits. In a previous study we investigated the deuterium (D) retention in sputter-deposited C-Si codeposits which showed similar retention characteristics to carbon deposits [4]. To address the hydrogen retention issue, thermo-oxidation (heating the deposits in an O₂ atmosphere) has successfully been used to remove trapped hydrogen from carbon-based tokamak codeposits [5]. The process can be highly effective, but is less so where some non-carbon elements (eg. B) make up a significant fraction of the deposits [5].

In this study, C-Si codeposits are produced via the sputtering of a SiC target with a D ion beam and collecting the sputtered C and Si atoms, along with reflected D ions, on a tungsten substrate. The collection surfaces can be heated during deposition to simulate conditions which might occur in a reactor environment. The deposits range in thickness from 100 to 1000nm, and the D content is measured by laser desorption spectroscopy (LDS). D reclamation is evaluated through thermo-oxidation treatments under various temperatures (300-400°C) and exposure times (0.5 to 6 hrs).

Our results generally show reduced D removal from the C-Si deposits as compared to most tokamak deposits [5]. C removal was also reduced. To ensure an accurate comparison, we have also investigated D removal from pure carbon specimens prepared by the same sputter deposition process. The reduced effectiveness of thermo-oxidation is attributed to the enrichment of SiO₂ in the films, which appear to stabilize the films against further oxidation. While SiC PFMs have a number of advantages over carbon components, they may result in greater effort required to remove the codeposited layers.

[1] F. Najmabadi et al., Fus Eng and Design 80 (2006) 3

[2] G. Sinclair et al., Nucl Mater and Energy 26 (2021) 100939

[3] M. S. Tillack et al., Fus Eng and Design 180 (2022) 113155

[4] J. A. Lantaigne, J. W. Davis, Nucl Mater and Energy 33 (2022) 101288

[5] J. W. Davis, AA Haasz, J Nucl Mater 390-391 (2009) 532

*Corresponding author: e-mail: james.davis@utoronto.ca

Tungsten (W) dust particles generated by plasma-wall interactions have been
observed many times during tokamak operation with full W plasma facing
components using fast camera diagnostics [1]. In our case, several distinct dust
species have been identified on twelve post-mortem inner-wall tungsten samples
extracted from their individual holders. Sample holders are positioned on toroidal
sectors Q3 and Q6 at poloidal locations of +50°, 0°, -50°, and -70° with respect to the
midplane of the WEST tokamak after its third and fourth operational campaigns [2].
Various thin film deposits have been identified on these samples, with composition
dependant on tokamak operation. Deposited on the surface of these thin films are
newly found stethoscope like dust particles, many of which have been identified on
samples nearest to the lower divertor. In addition to these, many spheroid dust
particles have been identified. These species indicate a deposition in liquid state with
important information on their direction of deposition. Large dust flakes and chunks
are also observed on all samples.

Dust features were measured using confocal, scanning electron and atomic
force microscopy (AFM), and energy dispersive x-ray spectroscopy (EDS). Size
distribution functions and aspect ratios have thus far been determined for
stethoscopes and spheroid dust on samples from the four poloidal locations across
Q3 and one sample from Q6 at -70°, closest to the lower divertor. Size distributions fit
well to a lognormal distribution with a most likely size for spheroid dust and
stethoscopes of 70–100 nm and 400–500 nm, respectively. The dominant range for
spheroid dust and stethoscopes is 50–200 nm and 0.2–1 μm, respectively. Both
stethoscopes and spheroid dust have a most likely aspect ratio (AR) of about 1.3–
1.4. However, stethoscopes may have AR above 3.5, while spheroid dust tends to
have a maximum AR of approximately 2.5. The deposited stethoscope dust particles
are found to exhibit preferential alignment on the sample surface at specific angles
(depending on the sample) between -50 and +50 degrees, with respect to the vertical
direction. Ongoing analysis of the variation of dust distribution functions along the
WEST inner-wall suggests that spheroid dust number density decreases from 3 x 106
mm-2 at the -70° poloidal location to 104 mm-2 further from the divertor at -50°.
Spheroid dust number density is found to vary toroidally by a factor of 2, while
stethoscope dust varies from 5 x 103 mm-2 to 2 x 104 mm-2. Primary composition of
both stethoscopes and spheroid dust is determined by EDS to be W with varying
degrees of oxidation.

[1] F. Brochard, A. Shalpegin, S. Bardin. Nucl. Fusion 57, 3 (2016)
[2] M. Diez, Y. Corre, E. Delmas, Nucl. Fusion 61, 10 (2021)

The effect of D₂ pellet injection and D₂ gas puffing on W erosion and peak intra-ELM heat flux has been studied in the Metal Rings Campaign (MRC) open and Small Angle Slot (SAS) closed divertor geometries in the DIII-D tokamak, and results are compared between the two geometries. Ongoing SAS analysis is being performed to filter background light due to visible bremsstrahlung (VB) and/or contaminant emission lines to reduce visible light (400.9 nm) W spectroscopy uncertainties. Likewise, SAS geometry peak intra-ELM heat flux is under analysis through the use of surface eroding thermal couples (SETCs) and Langmuir probes. Inter- and Intra-ELM W divertor erosion and intra-ELM peak divertor heat flux during D₂ pellet injection has been fully assessed from the DIII-D MRC, where toroidally symmetric W-coated tiles were installed in the lower open divertor [1]. Fast D₂ pellet mass injection ranging from 34-41 in arbitrary units (A.U.) and no D₂ injection resulted in a similar total W erosion rate during ELMs (intra-ELM). On average, results show a 29% increase in the total gross W erosion rate with intermediate mass injection rates (~13-23 A.U.) compared to the no pellets and the fastest injection rate cases. On average, the fast D₂ mass injection rate cases had 15% less erosion between ELMs (inter-ELM) than cases with no pellets. An inverse correlation between inter-ELM density and W erosion during D₂ pellet injection is also observed. Simulations by the ‘free-streaming plus recycling model’ (FSRM) under and over-predict the average W erosion per ELM by approximately 25% when including and excluding a C/W mixed material model, respectively. The peak intra-ELM heat flux during the MRC decreased steadily with increasing D₂ pellet mass injection rates. Heat flux values were approxiamtely, on average, 45% lower for fast D₂ injection rate cases compared to no D₂ injection cases. Inter- and Intra-ELM erosion from the SAS campaign during D₂ pellet injection and D₂ gas puffing into the tightly baffled W-coated closed upper divertor is currently under investigation to distinguish the effect that adding cold neutrals and triggering smaller, more frequent ELMs has on W erosion. D₂ pellets and gas were injected and puffed, respectively, at various rates to determine trends in W erosion compared to the total D₂ influx. Moreover, collector probes were exposed to study W transport and help investigate net W erosion. The impact on W sputtering and transport and peak intra-ELM divertor heat flux from D₂ pellet and gas insertion will be discussed for the SAS experiment and further details on the impact on W sputtering and peak intra-ELM heat flux from various D₂ pellet injection rates in the Metal Rings and SAS campaigns will be compared.

Capacitively coupled RF discharge in a magnetized environment is widely used for material processing applications and first mirror (FM) cleaning operations in tokamaks thanks to its versatility [1, 2]. In an asymmetric discharge, ions are accelerated in the DC self-bias voltage developed at the active electrode surface. The magnitude of this DC voltage is controllable by varying the operational parameters such as gas pressure, the strength of the magnetic field, RF power and the angle between the magnetic field and the RF electrode [3]. When the electric field on the surface of the electrode has a component perpendicular to the magnetic field, the uniformity in particle flux to the active electrode is modified due to the ExB drift. While performing experiments with a one-faced tungsten electrode, a shift in the DC self-bias voltage from negative to positive values was observed at low pressure (1 Pa) and 0.1 T magnetic field at grazing angles [4]. In addition, the density gradient due to the stochastic heating in the direction perpendicular to B creates a diamagnetic drift above the RF electrode. Under the light of this complex physics above the active electrode, a series of experiments has been performed in the ALINE plasma device [5] to study the evolution of these drift effects at different grazing angles. The potential and density profiles inside the magnetic flux tube have been investigated using the RF-compensated Langmuir probes. The effect of drift and contribution to particle flux to the electrode is investigated using the Langmuir probe data. The motion of the drift structures around the electrode is recorded using a fast camera. The drift velocity evaluated from the video data is compared with the values obtained analytically from the probe results.

[1] P. Chabert, Physics of Radio-Frequency Plasmas, Vol. 64 pp. 197–223. (2011)

[2] K. Soni et al 2022 Nucl. Fusion 62 126009 (2022)

[3] M. Oberberg, et al.,Plasma Sources Science and Technology, 27(10) (2018)

[4] A. Chreukulappurath Mana, et al., Submitted to Nucl. Fusion (2022)

[5] E. Faudot, S. Devaux, J. Moritz, et al., Rev. Sci. Instrum. 063502 (2015)

Surface roughness is known to have a huge impact on sputtering yields, especially at oblique and grazing incidence [1]. However, this property is generally not taken into account when a sample’s sputtering yield is measured. For relatively flat metallic samples, this is not a problem, but after prolonged plasma exposure, most surfaces tend to roughen [2], so their sputtering yield may vary during the experiment. Moreover, most ceramic materials, even pristine samples, usually exhibit a rather rough surface due to the sintering process. Therefore, ceramics sputtering yield values are largely scattered, as observed on boron nitride samples.

A sputtering simulation Monte-Carlo program, called CSiPI [3], has been developed at ONERA in order to model collision cascades in matter and compute sputtering yields. In this conference, we will present some recent developments we made to implement in this software a roughness module. In this module, the code takes a surface profile as input, searches for potential shadowed areas and runs collision cascades all along the profile, with specific (local) incidence angle at each location (related to the local curvature of the surface). The program also tracks sputtered particles, and only those which do not intercept the surface after leaving the target are counted in the sputtering yield. We will present our results regarding sputtering yield with different roughness levels, but also atomic re-deposition fraction on the surface and emitted particles profiles.

In our laboratory, we also run an erosion-dedicated facility, equipped with a Kaufman source, in which we use Quartz Crystal Microbalances (QCMs) to collect sputtered particles. Eroded particles collection measurements are then used to compute both the total sputtering yield of the target and emission profiles. In order to validate our simulation model, we will perform some experiments on samples with different roughness, previously measured by optical profilometry. Surface profiles obtained with this instrument will serve as inputs for our simulations. The goal is also to study the influence of surface topography on ejected particles directions, which may be of primary interest regarding erosion induced contamination of tokamaks walls and plasma core [4]

[1] M. Küstner, W. Eckstein, E. Hechtl, and J. Roth, Journal of Nuclear Materials, vol. 265, no. 1–2, pp. 22–27, Feb. 1999.

[2] W. L. Chan and E. Chason, Journal of Applied Physics, vol. 101, no. 12, p. 121301, Jun. 2007.

[3] T. Tondu, V. Inguimbert, and F. Darnon,European Space Agency, (Special Publication) ESA SP, Sep. 2006.

[4] J. Linke et al., Matter and Radiation at Extremes, vol. 4, no. 5, p. 056201, Sep. 2019.

To assess the long-term stability of deuterium retention in co-deposits two tiles from JET with the ITER-Like Wall (JET-ILW) were examined: a tungsten-coated divertor tile exposed during the first ILW campaign in 2011–2012 (ILW-1); a beryllium dump plate tile exposed during the second campaign in 2013–2014 (ILW-2). Studies were performed with ion beam analysis techniques: Rutherford Backscattering Spectrometry (RBS), Proton Induced X-ray Emission (PIXE, for metal impurities) and Nuclear Reaction Analysis (NRA).

The D concentration was measured twice in the two studied tiles: shortly after the end of respective campaigns and then a few years later (in 2020). In between, the tiles were stored at the Beryllium Handling Facility (BeHF) at JET. The second measurement was done to assess the aging effect on the D inventory after long-term storage. This work has been motivated by the need to recognise the stability of retention in view of future D-T operations in JET-ILW and in ITER. The main results are summarised in a few points.

• A belt of Be-rich co-deposit was formed in the region shadowed of divertor from the direct plasma impact (s = 712–762 mm). The layer thickness reached ~1.5 μm with D content of up to 1 ×10 ¹⁹ D/cm ², as determined in the first measurement done three months after ILW-1.
• The second measurement done 7.5 years later revealed a very similar D content (up to 1 ×10 ¹⁹ D/cm ²) thus indicating a negligible impact of the long-term storage on the fuel concentration.
• The results obtained for the D concentration on the dump plate tile were also similar in two measurements performed two years after ILW-2 and then four years later (up to 3 ×10 ¹⁷ D/cm ² and 4 ×10 ¹⁷ D/cm ² respectively)

The results prove the long-term stability of the deposits and the retention of the D in the tiles when stored at room temperature and ambient atmosphere. For future operations with tritium, the stability of the fuel in deposits implies that over this period there is no significant change in activity in terms of waste inventory categorisation.

On the safe operation for a magnetic confinement fusion device, tritium accumulation is one of the main concerns. Simultaneous deposition of eroded plasma facing materials (PFM) and plasma particles, so-called co-deposition can be a source of hydrogen fuel retention in tokamaks [1]. On the other hand, it is widely understood that helium (He)-induced surface changes strongly disturb the retention of deuterium (D) into W bulk area [2]. A previous study has shown that accumulation of He with high binding energies is mitigated for co-deposits [3], while D retention strongly increased with the presence of He-W co-deposition layers [4]. In this study, He-W co-deposition layers are created with simultaneous operation of He plasmas and W sputtering, then exposed to pure D plasmas. Total D and He desorption peaks are measured by thermal desorption spectroscopy (TDS).

In this study, He and D plasmas were generated by using a compact RF plasma device APSEDAS, which can control the plasma operations between inductively coupled mode and helicon modes. W deposits are produced by introducing a negative voltage of -200~-400 V to spiral-shaped W wires placed ~10 mm above a W sample. Resultant co-deposition layers had various surface morphologies and layered structures as the biasing voltage, the sample temperature and the plasma modes changed. When these conditions were kept constant, a series of plasma exposure on several W samples with increasing He fluences presented consistent growth of co-deposits with similar internal structures in the layers.

At the sample temperature above ~1400 K, co-deposits were recrystallized, forming larger grain combined to W bulk area, while pressurized bubbles up to ~100 nm were distributed in depth of sub-mm and many pinholes existed on the surfaces. At the temperature under 1270 K, blisters were formed at the boundary of W bulk surface. SEM observations revealed that the layers internal composition of the layers differed with the ratio of W and He particles, which were speculated from emission spectra of W I and He I lines, and the bias voltage which alters sputtering yield of W.

Succeeding D plasmas were exposed to the He-W co-deposition layers at the surface temperature of < 700 K. On a pristine W surface with D plasma exposure only, broad desorption peaks appeared at ~400 K and ~550 K. On the other hand, new D desorption peaks were detected apparently at ~350-360 K, while the peak at ~550 K significantly lowered or disappeared, which is similar to the previous study [3]. Total D retention amounts decreased by 50-90% compared to the pristine W sample. This result hints that He-induced co-deposits also play as a D diffusion barrier, which hinders D to diffuse into a W bulk area.

Acknowledgments

This work was supported in part by JSPS KAKENHI (20K22322, 21H01059, 21KK0048) and NIFS Collaboration Research program (NIFS22HDAF009, NIFS20KUGM152).

References

[1] A. Widdowson, J.P. Coad, E. Alves et al., Nucl. Fusion 57, 086045 (2017)

[2] K.D. Hammond, Mater. Res. Express 4, 104002 (2017)

[3] S. Krat, E. Fefelova, A. Pryshvitsin et al., Nucl. Mater. Energy 34, 101336 (2023)

[4] S. Kajita, K. Asai, N. Ohno et al., J. Nucl. Mater. 540, 152350 (2020)

*Corresponding author: tel.: +29 853 7474, e-mail: hwangbo@prc.tsukuba.ac.jp (D. Hwangbo)

A fuzzy layer formed on a tungsten (W) surface can lead to significant line-of-sight redeposition of sputtered W atoms, lowering its sputtering yield and changing its angular distribution. While it is generally agreed that fuzz reduces the sputtering yield, the changes in the angular distribution seem inconclusive. When sputtering W by argon (Ar) plasmas, vertical profiles of W I line emission observed at a single axial location in the PISCES-B linear plasma device suggested that the angular distribution from a fuzzy surface was similar to that from a smooth surface [1], while quartz crystal microbalance measurements showed that the “backward” sputtering was more preferred from a fuzzy surface [2].

In our previous work [3], hyperspectral imaging (HSI) was proven a powerful technique to investigate angular distributions of sputtered impurities, since the 2-D profiles of electron density and temperature of a helium (He) plasma can also be obtained from HSI images, facilitating impurity transport simulations in the plasma. Although the low sputtering yield of W in He plasmas makes HSI measurements rather difficult, similar fuzzy structures also form on a molybdenum (Mo) surface, making the measurements much easier due to its higher sputtering yield in He plasmas. Since the line-of-sight redeposition is a purely geometrical effect, investigations using Mo fuzz should give insight into the physical sputtering of W fuzz.

In the present work, we use a Mo sample with a 10 mm diameter, since the 22 mm sample diameter used in Ref. [1] might have been too large for the plasma column with a diameter of ~50 mm FWHM (full width at half maximum), so that the geometrical details of the emission profiles were smeared out. A fuzzy layer is first grown on the Mo sample by He plasma exposure with sample temperature, T_s, ~1023 K, He ion energy, E_i, ~85 eV and He ion fluence ~10²⁶ m^-2 in PISCES-A. Next, the fuzzy surface is sputtered by a He plasma at T_s < 423 K and E_i ~132 eV, while the 2-D profile of Mo I line emission at 550.6, 553.3 and 557.0 nm is recorded by an HSI camera. The emission profile is then compared with that from a smooth Mo surface. Our results show that after the formation of fuzz, the emission intensity is reduced by a factor of 3 due to the sputtering yield reduction, and the emission profile is narrowed along the vertical direction, suggesting that the angular distribution from a fuzzy surface prefers backward sputtering. Simulations, considering fuzz structures, will be performed to quantitatively verify if a fuzzy layer is indeed the cause of these changes.

[1] D. Nishijima, et al. J. Nucl. Mater. 415, S96-S99 (2011)

[2] R. Stadlmayr, et al. J. Nucl. Mater. 532, 152019 (2020)

[3] F.J. Chang, et al. Nucl. Mater. Energy 29, 101077 (2021)

Work supported by US-DOE cooperative agreement, DE-SC0022528.

*Corresponding author: tel.: +1 858 531 4337, e-mail: fchang@eng.ucsd.edu (F. Chang)

Investigation of the He content in W layers, deposited by Ar-He magnetron discharges, by LIDS, LIBS and TDS

L. Laguardia ^a, *, M. Iafrati^b, M. Alonzo^b, S. Almaviva^b, M. Pedroni^a, E. Vassallo^a,

M. De Angeli^a, F. Ghezzi^a, G. Gervasini^a, A. Uccello^a, A. Cremona^a, V. Mellera^a,

A. Rufoloni ^b

^a Istituto per la Scienza e Tecnologia dei Plasmi-CNR, Via R. Cozzi 53, I-20125 Milan, Italy

^b ENEA, Department of Fusion and Nuclear Safety Technology, 00044 Frascati, Rome, Italy

Fuel accumulation in plasma-facing and structural materials used in fusion devices is highly important both for radiation safety and for the assessment of the impact of gas recycling on plasma operation. Co-deposition, that is simultaneous deposition of previously eroded plasma facing materials and plasma particles, is considered one of the primary sources of hydrogen fuel accumulation in tokamaks [1]. Helium (He) quantity in future reactors will be about the same as that of fuel particles due to deuterium-tritium (D-T) fusion reactions. It is well known that radiation/exposure causes much more material damage than hydrogen atoms such as He bubbles and nano-structures, known as “fuzz” [2]. As laboratory experiments have demonstrated [3], He is trapped in tungsten (W) samples by ion and plasma implantation. Therefore, it is important to investigate how He is accumulated in co-deposited or re-deposited layers observed in fusion devices. As reported in the literature [4] W-He co-deposits, which simulate the re-deposited layers, can be produced in the laboratory by exploiting magnetron sputtering technology.

In this contribution, the laboratory production of reference coatings mimicking tokamak W-He deposits and the characterization of their morphology and He content are presented. The effect of the different process parameters on the properties of the coatings is also addressed.

W-He films were deposited by magnetron sputtering with variations of pressure ranging from 1-5 Pa. The morphology of the coatings was investigated by Scanning Electron Microscopy (SEM) whereas the He content within the layer was measured by Laser Induced Desorption Spectroscopy (LIDS), Laser-Induced Breakdown Spectroscopy (LIBS) and Thermal Desorption Spectroscopy (TDS).

He TDS spectra exhibit a broad desorption peak in the 500-600 K range and another one, significantly increased, at ~960 K. To quantify He content in the samples, a calibration procedure that takes into account the conductance and pumping speed of the device has been performed allowing the determination of the sensitivities of the mass spectrometers used as detectors for TDS and LIDS.

[1] A. Widdowson et al. 2017 Nucl. Fusion 57 086045

[2] D. Nishijima, et al. 2011 J. Nucl. Mater, 415

[3] Y. Gasparyan et al. 2020 Phys. Scr. T171

[4] S. Krat et al. 2020 J. Phys. Conf. Ser. 1686

*Corresponding author: tel.: +39 0266173208, e-mail: Laura.Laguardia@istp.cnr.it (Ms. Laguardia)

Advancements in ultraviolet spectroscopic measurements and tungsten (W) atomic data have made time-resolved measurements of erosion and re-deposition of W from plasma facing components (PFC) possible. Erosion of W, especially during Edge Localized Modes (ELMs), can potentially degrade plasma confinement through core contamination and radiative loses. Spectroscopic measurements in combination with atomic physics coefficients known as S/XB ratios can be used to infer high-Z PFC erosion and re-deposition rates.

Experiments have been conducted in the DIII-D tokamak to examine the dynamics of tungsten erosion and re-deposition using a recently commissioned high-resolution spectrometer optimized for ultraviolet wavelengths. Measured erosion rates from ion beam analysis of removable tungsten coated graphite samples and spectroscopic emission from multiple spectral lines in the UV range are compared over a range of electron temperatures (10 – 25 eV) and densities (1.5 – 6 × 1019 m-3) during L-mode and H-mode plasmas. Neutral (W I) and singly-ionized tungsten (W II) lines have both been observed for determination of W re-deposition rates. Tungsten erosion and re-deposition rates during ELM plasma intervals are being investigated.

Spectroscopic techniques providing time-resolved erosion, re-deposition, and transport require accurate atomic data, with the near-neutral charge states being the most critical. Further, the impact of metastable level populations on W spectra require multiple emission lines to be monitored simultaneously [1] to accurately infer erosion rates. New atomic calculations of W+ [2] and W2+ excitation have been benchmarked against tungsten emission in the Compact Toroidal Hybrid (CTH) experiment and will be utilized for accurate inference of tungsten re-deposition fractions. Supporting measurements in the CTH device have been accomplished utilizing high-resolution spectroscopy (~4 pm wavelength resolution) to evaluate potential impurity blending with tungsten emission.

Work supported by USDOE grants DE-SC0015877, DE-FC02-04ER54698 & DE- FG02-00ER54610.

[1] Johnson et al., Plasma Phys. Control. Fusion 62, 125017 (2020)

[2] Dunleavy et al., J. Phys. B: At. Mol. Opt. Phys. 55, 175002 (2022)

*Corresponding author: tel.: +1 (334) 844-4366, e-mail: ennis@auburn.edu (D. A. Ennis)

Over the first phase of WEST exploitation, the full tungsten environment consisted in a mixture of W coated and actively cooled ITER-like plasma facing units (PFU). With an X-point divertor configuration during the WEST C3 and C4 campaigns, stable discharges, long pulse duration, a dedicated helium campaign and steady state heat loads up to 6 MW/m² on the lower divertor were achieved [1]. First post-mortem analyses were performed on the entire tiles [2] or on collected deposits on PFUs [3], they allowed to obtain an overview of the erosion/redeposition pattern, thickness and content of co-deposited layers from the lower divertor [4]. Except for the two erosion-dominated areas corresponding to the inner and outer strike point region (ISP and OSP) respectively, the divertor is covered by layered deposition mixing mainly (W, C, O, B with D and He). Thicker deposits were observed on the inner side of the ISP zone elsewhere most of the divertor is covered with thin coloured deposits (<1 µm) leading in some area to rainbow iridescence which correlates to a sharp emissivity evolution questioning on the origin of such property changes of deposits [5].

This contribution focuses on the composition (W, C, O and B) and the roughness of thin (1 < µm) coloured deposits observed on PFU for surfaces of both top and poloidal gap side of W-monoblocks (MBs). Electron and optical confocal microscopies, Xray and Raman spectroscopies were performed along the radial direction with spatial sampling of about few mm allowing to follow the sharp variation of property changes for selected MBs of the PFU#13 (MB02-03 on inner side, MB17-18 in the private flux area and MB32-35 on outer side). Coupling Raman and Xray results allowed to identify the relative content and also the material phases of the deposit (aromatic carbon, tungsten oxide, boron oxide etc…). The higher content of C and B was observed on the inner side whereas only tungsten oxide film of WO₃ type was evidenced on the outer side for the MBs which were shadowed from the plasma by the baffle component. The results will be put into perspective with the other coloured thin films found in gaps and first wall. The goal is to improve the understanding of the role played by light impurities in the plasma-wall interactions. For example, the formation of carbon rich and oxide deposits, thins but covering large area, could contribute for example to the D retention.

[1] E. Tsitrone et al. 2021 Nucl. Fusion 61

[2] M. Balden et al. 2021 Phys. Scr. 96 124020

[3] M. Martin et al 2021 Phys. Scr. 96 124035

[4] M. Diez et al 2022 contribution PSI to be published

[5] J. Gaspar et al 2022 Nucl. Fusion 62 096023

The material balance, global erosion and deposition patterns of the JET ITER-Like Wall from the perspective of ex-situ analysis of plasma facing components (PFC) is discussed to give an overview spanning three operating periods: 2011-12, 2013-14, 2015-16. Mass change data is summarised bringing newly presented data for the net erosion of beryllium limiters and deposition divertor tiles based on weighing and surface profiling analysis of PFCs providing the following highlights:-

• Net deposition in the divertor is of the order 60 g after 19 hours (one operating period) of divertor plasma operations and the mass of dust collected is of the order 2 g, giving an upper dust conversion factor ~2-4% which is an order of magnitude lower than estimated for the JET carbon wall. This value assumes that dust is a product of the disintegration of deposits. However dust sources in JET are various, including from tungsten coatings on divertor tiles (specific to JET), and beryllium droplets and therefore in practice the conversion factor is lower.
• The net erosion and deposition distribution from weighing follows the generally understood pattern of net deposition at the inner divertor tiles and outer divertor horizontal corner tile 6 and net erosion at the vertical outer divertor tiles. New confirmation of net mass loss at tiles B and C further outboard is supported by images showing tungsten coating removed from the surfaces of B and C tiles.
• Beryllium limiters are known to contribute to the source of eroded beryllium. In this work the quantified erosion masses from weighing and surface profiling are presented. The sectioning of the JET beryllium limiter tiles enables central and end sections of the tiles to be investigated independently. Mass change data shows that net erosion from the outer and inner mid-plane limiter tiles, central section only, are the largest source or eroded material arising mainly from limiter plasma contact on start-up and ramp-down of plasmas. When considering net erosion and deposition across whole tiles and interpolating along all limiter beams (twelve outer vs ten inner beryllium limiter beams) the outer limiter remains the largest global erosion source, due in part to higher surface area, accounting for > 50 g net erosion per operating period compared with < 15 g at the inner limiter.

The global material balance will be discussed taking into consideration the differing pathways for erosion, material migration and deposition in limiter plasma and divertor plasma, along with supporting plasma operations related data on beryllium spectroscopy and ion flux data. By way of conclusion for future fusion PFC analysis programmes the lessons learned from the ex-situ analysis techniques for assessing net erosion and deposition for global material balance will be discussed.

The JET ITER like Wall (ILW) divertor mostly consists of CFC tiles coated with a
traditional 20 microns thick tungsten (W) layer over a thin molybdenum (Mo)
interlayer. Some poloidal sets of tiles are coated with ~3 microns of Mo and ~4
microns of W on top of ~12 microns of W (and the thin Mo interl ayer) to act as
markers to measure the erosion/deposition during ILW campaigns. A few of the
marker tiles only have the Mo coating to facilitate transport studies of Mo and W
within the divertor. Transport to and from one of these Mo only markers (tile 14ING3B
a tile 3 from the lower part of the inner divertor wall) to tile 14BNG4D (an innermost
base tile4 with a standard Mo + W marker) has been extensively studied using SIMS
and IBA techniques including RBS and HIERDA.

Metallic elements are principally eroded by incident plasma ions as atoms or ions and
most probably return to the surface a short distance further in the plasma field line
direction where they may be eroded again: migration may thus occur by a series of
“hops ”. There was measurable Mo erosion from the top of 14ING3B after the first ILW
campaign (ILW1) as for the majority of discharges the inner strike point was at the top
of tile 3/bottom of tile 1 (plasma configuration V5) whereas for ILW 2+3 the inner
strike point was u sually nearer th e bottom cor ner of the divertor (configuration VT).
As the incident plasma ions have both toroidal and poloidal components the eroded
Mo may migrate toroidally to an adjacent tile 3 or poloidally down onto the nearby tile 4.
After both the IL W1 and the ILW2+3 campaigns significant amounts of Mo had been
transferred to 14BNG4D and were located towards the top of the tile (s-coordinates
~840mm).