LAUSR

Abstract Nitrogen seeding, necessary for divertor heat-load mitigation in ITER, has been shown to lead to ammonia formation which would be a severe operational and safety issue in ITER. Predictions of ammonia production in ITER are based on data from present day fusion devices. Ammonia is mainly detected by residual gas analysis (RGA). Detection of ammonia is impeded by the presence of water and methane which, in a mixed H-D system, leave signatures in the same range of the mass spectra. A statistical model is used to ascribe an average isotope ratio to each gaseous species. The model is tested with simulated RGA recordings with varying concentration of ammonia to evaluate the sensitivity to fitting parameter boundaries, noise in the recordings and mis-matching cracking patterns. The analysis shows that the fitting procedure may in some occasions substitute species among each other, resulting in faulty concentrations. Nevertheless, the right choice of parameter boundaries ensures correct fitting results. Finally, the fitting procedure is applied to experimental data from nitrogen-seeeded discharges at AUG and JET.