Study the role of roughness in the sputtering process of tungsten by GyM helium plasma: experiments and ERO2.0 modelling

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 (Ra?3, 300, 600, 900 nm), and graphite substrates with irregular surface and three different Ra (?7, 90, 280 nm), as well as reference polished bulk W samples (Ra?10 nm), have been exposed in GyM changing the incident He+ energy (EHe+) 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 EHe+>=250 eV. This allowed to evaluate the quasi-static effective sputtering yield (YW|GyM) from mass loss data, on the one hand, and run single time step ERO2.0 simulations, on the other hand. For EHe+<=200 eV, YW|GyM is negligible. For higher energies, YW|GyM decreases by increasing the mean value of the surface inclination angle distribution (?m), in agreement with ERO2.0 results. ?m is thus the key-parameter determining the erosion of the samples rather than Ra, as also pointed out in [3]. Moreover, YW|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 YW from the code is also higher than YW|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.