Description
Oxygen impurities in fusion reactors impair plasma startup in limiter configuration, which restricts the transition to high-performance discharges. Effective oxygen gettering is thus critical, and boronization addresses this issue by improving impurity control [1]. This requires boron layers 50 to 100 nm thick to reduce the oxygen content. One method for depositing such layers is glow discharge boronization, creating B:D coatings. The bombardment of these layers with plasma particles limits their lifetime. The eroded boron can migrate into recessed areas, potentially prolonging the gettering capabilities. Codes, such as ERO2.0, simulate this plasma wall-interaction, but require validated data on physical and chemical erosion yields. The linear plasma device PSI-2 has been essential for validating ERO2.0, particularly for tungsten studies [2]. Thus, PSI-2 is an excellent device for investigating the interaction of the boron layer with deuterium plasmas.
This contribution presents the erosion behaviour of 115 nm thick B and B:D layers deposited onto tungsten under deuteron bombardment in PSI-2. Biasing the target enables scanning of the ion impact energy across the near-threshold erosion regime from 40 to 100 eV. Intact boron layers completely covered the samples when the deuteron fluence was limited to $3\cdot10^{23}$ m$^{-2}$. However, the thickness decreased by tens of nanometers, so post-mortem layer thickness measurements provided net erosion rates [3]. During plasma operation, optical emission spectroscopy of the B I 2p-3s (250 nm) transition and the BD Q-branch of the A$^1$Π-X$^1$Σ$^+$ system (432.8 nm) gives time-resolved gross erosion rates. Inverse photon emission efficiencies applied to D$_\gamma$ and BD A-X Q-branch suggest a constant chemical erosion yield of ~0.1% under PSI-2 conditions (ion flux density $4\cdot10^{21}$ m$^{-2}$s$^{-1}$; surface temperature <200°C). Additionally, the spatial emission profile of the 2p-3s emission, essentially the penetration depth, agrees well with the emission simulated by ERO2.0.
These physical and chemical erosion yields are critical for refining the input data into ERO2.0, addressing a key uncertainty in predicting boron erosion and migration in ITER.
[1] J. Winter et al. J. Nucl. Mater. 713 162-164 (1989)
[2] A. Eksaeva et al. NME 12 253-260 (2017)
[3] M. Sackers et al. NME 45 102003 (2025)