• Energy Research
• SI close

This work reports on the investigation of Eurofer-97 erosion behaviour when exposed to the deuterium plasma of the linear device GyM. The erosion dependence of Eurofer-97 on the deuterium ion fluence, Φ≤2.3×1025 m−2, and temperature of the samples, T = 600 K and 990 K, was addressed. A bias voltage of −200 V was applied to GyM sample holder during the experiments. Samples were deeply characterised by: profilometry, scanning electron microscopy, atomic force microscopy, energy-dispersive X-ray spectroscopy, time-of-flight secondary ion mass spectrometry, Rutherford backscattering spectroscopy and particle-induced X-ray emission.The behaviour of Eurofer-97 erosion rate with the ion fluence strictly depends upon temperature. At 600 K, it was ∼0.14 nm/s after 4.7 × 1024 m−2, then decreased, reaching a steady state value of ∼0.01 nm/s from 8.0 × 1024 m−2. At 990 K instead, the erosion rate was roughly constant around 0.019 nm/s for Φ≤1.24×1025 m−2. The value at 2.35 × 1025 m−2 was slightly lower. The erosion rate at 990 K was greater than that at 600 K for every ion fluence.Microscopy and surface analysis techniques showed that Eurofer-97 erosion rate dependence on Φ at 600 K was primarily determined by the preferential sputtering of iron and other mid-Z elements of the alloy, leading to a surface rich in W and Ta difficult to be sputtered. The erosion behaviour at 990 K was dominated by the morphology dynamics, instead. The different properties of the morphology developed at the two temperatures can explain the higher erosion rate at 990 K for all the ion fluences.

Net erosion of plasma-facing materials was investigated at the low-field-side (outer) strike-point area of the ASDEX Upgrade (AUG) divertor during H-mode discharges with small and frequent ELMs. To this end, Au andMo marker samples with different surface morphologies and geometries were exposed to plasmas using the DIMIIdivertor manipulator. The results were compared to existing erosion and deposition patterns from various LandH-mode experiments, in the latter case the main difference was the size and frequency of the ELMs.We noticed that increasing surface roughness reduces net erosion but less than what is the case in L-mode. Onthe other hand, net-erosion rates in H-mode are generally 2–5 times higher than the corresponding L-modevalues, in addition to which exposure in H-mode conditions results in strong local variations in the poloidal andtoroidal erosion/deposition profiles. The latter observation we associate with the large migration length, on theorder of several cm, of the eroded material, resulting in strong competition between erosion and re-depositionprocesses especially at poloidal distances > 50 mm from the strike point. Considerable net erosion wasmeasured throughout the analysed poloidal region unlike in L-mode where the main erosion peak occurs in thevicinity of the strike point. We attribute this qualitative difference to the slow decay lengths of the plasma fluxand electron temperature in the applied H-mode scenario.Both erosion and deposition require detailed analyses at the microscopic scale and the deposition patterns maybe drastically different for heavy and light impurities. Generally, the rougher the surface the more material willaccumulate on locally shadowed regions behind protruding surface features. However, rough surfaces alsoexhibit more non-uniformities in the quality or even integrity of marker coatings produced on them, thuscomplicating the analyses of the experimental data.We conclude that local plasma parameters have a huge impact on the PFC erosion rates and, besides incidentplasma flux, surface morphology and its temporal evolution have to be taken into account for quantitative estimatesof erosion rates and PFC lifetime under reactor-relevant conditions.