Speaker
Description
As part of ITER’s 2023 re-baselining, the first-wall material was changed from beryllium to tungsten. The absence of the beneficial ability of Be to getter low-Z impurities like oxygen will be compensated by covering the armour tiles with a boron layer of 10-100 nm (boronisation). Broadening the understanding of physics behind B erosion and re/co-deposition is a top priority for ITER to effectively control plasma-wall interactions.
This work investigates erosion of pulsed-laser deposited B coatings: 100 nm compact films (35% porosity), representing pristine B layers, on mirror-polished W substrates, and 1-8 $\mu$m porous films (50-95% porosity), mimicking re-deposits, on silicon wafers. B model systems were exposed to deuterium plasma in the GyM linear device under ITER first-wall-relevant conditions (3-4e20 m$^{-2}$s$^{-1}$ ion flux).
Within the EUROfusion work-programme, thin B films were exposed to E$_{ion}$=76 eV for comparison with experiments in PSI-2, on RF-magnetron-sputtered coatings [1], and in a D$^+$ source, on plasma-sprayed coatings [2]. Here, a low ion fluence $\Phi$=4e23 m$^{-2}$ was selected to avoid complete layer erosion. After exposure, samples were characterised using mass loss, AFM, SEM, ToF-ERDA, ToF-MEIS, SIMS, Raman spectroscopy, and XPS.
Regarding the pristine B layer, SIMS data interpretation was hindered by partial oxidation and is being disentangled through comparison with pure B samples and related compounds. ToF-ERDA shows $\sim$13.3 at.% of oxygen. Combining areal density ($\sigma$) and film thickness, a calculated mass density of 2.39 g/cm$^3$ is obtained. Post-exposure, $\sigma$ corresponds to 6.8 nm. High-resolution ToF-MEIS analysis indicates W at the surface, suggesting its incorporation into the B-O layer. The effective sputtering yield (Y$_{eff}$) agrees within a few percent with [1–2], despite differences in layer structure and a two-orders-of-magnitude variation in ion flux.
Porous B films were exposed to E$_{ion}$=43-223 eV and $\Phi$=7.0e23-2.8e24 m$^{-2}$. Y$_{eff}$ of all layers, notwithstanding differing morphologies, is generally within 20% of SDTrimSP results for E$_{ion}$>43 eV, while at 43 eV it reaches up to four times higher, possibly due to ion-assisted chemical erosion at 550 K. The limited sensitivity of Y$_{eff}$ to film morphology may, for instance, have important implications for plasma-cleaning of B-based re/co-deposits on ITER first-wall optical diagnostic mirrors. Although the structure and porosity of re/co-deposits may vary depending on mirror location in the vessel, their erosion rates are expected to be similar, potentially enabling a single plasma-cleaning strategy for all optical diagnostics.
[1] M.Sackers, et al., Nucl. Mater. Energy 45(2025)102003
[2] E.Hechtl, et al., J. Nucl. Mater. 196-198(1992)713-6