Numerical modelling of tungsten erosion under plasma irradiation PSI-2 facility
Interaction of hydrogen isotopes from the plasma with first wall materials of fusion reactors including tungsten (W) is one of the main problems for the controlled thermonuclear fusion. Plasma-surface interaction (PSI) determines the life time of the wall components and can lead to tritium retention, which must be avoided. Transport of sputtered impurities in plasma is responsible for severe energy losses during the operation. Tungsten (W) has been chosen as a main material for construction of ITER divertor due to its low sputtering at edge plasma temperatures, large melting temperature, and small uptake of tritium . However, due to its large mass, tungsten can cause strong radiative energy losses when moving in plasma. Therefore, experiments aimed at W transport in plasma as well as some theoretical calculations are necessary.
In many cases linear plasma devices such as PSI-2  have considerable advantages for conduction of the respective experiments e.g. continuous plasma operation, straightforward geometry, possibility to have more control on experimental conditions than in tokamaks. During well-defined and reproducible material exposures in small scale laboratory experiments it is possible to find answers to many particular questions of plasma-material interaction.
Despite experimental results obtained at linear devices are somewhat easier to analyze than that from tokamaks, numerical simulations are still needed to account for a variety of processes taken place in these experiments. 3D Monte Carlo code ERO has been applied many times for predictive modelling of erosion/deposition and the impurity transport in ITER . This code calculates transport of impurities in background plasma using Monte-Carlo approach. During the calculation a wide range of characteristic processes are considered.
In this work, we present a new version of 3D Monte Carlo code ERO, designed to calculate the plasma surface interactions and transport of impurities in linear plasma devices. Code modifications included modifications of the physical model (energy and angular distributions of sputtered particles, neutral metastable W tracking), magnetic field geometry, location of the diagnostics, 3D distribution of plasma density and temperature, position and shape of the target.
For testing of this new version of the code we used the results of the dedicated experiment on tungsten erosion recently carried out at PSI-2 installation. In this experiment a rectangular W target 80x100 mm was exposed to argon plasma with different parameters. Spectroscopic measurements, quartz micro-balance (QMB) for deposition at a distant location and weight loss measurements were conducted during these experiments. The weight loss also had a radial resolution and the WI 400.8nm line profiles were giving a 2D side view image.
Qualitatively the ERO simulations reproduce well all experimental dependencies: spectroscopy profiles, QMB measurements and weight loss. It was demonstrated that metastable tracking in ERO can explain deviations between modelled and observed axial WI intensity profiles, demonstrating their strong influence on the W spectroscopy. Characteristic lifetime of WI metastable states was fitted as t ≈ 1.4*10-4 s. The angular distribution of sputtered W atoms was determined by comparison of the modelled deposition on the QMB with the according experimental data and confirmed by additional molecular dynamics (MD) calculations. Erosion values were extracted from both experimental dependencies and ERO simulations. In general, our interpretation is in a good agreement with previous material sputtering simulations conducted by W. Eckstein (SDTrimSP code) .
 V. Philipps, Tungsten as material for plasma-facing components in fusion devices. Journal of Nuclear Materials. Vol. 415, Issue 1, Pages S2–S9 (2011)
 M. Reinhart et al., Transactions Fusion Sci. Technol. 63 (2013) 201.
 A. Kirschner et al. JNM, 363-365 (2007) 91
 J.P. Biersack, W. Eckstein, Appl.Phys. 34 (1984) 73