“…k p was found to depend weakly on power, typically k p $ P 0:2 inp or weaker [7,8]. As a consequence of k p being narrower than expected, installation inaccuracies ($0.5 mm), radial modulation of the field lines at the limiter position caused by field ripple ($1.5 mm) and up/down power asymmetries [2], the actual power handling capability of the limiter was insufficient for the achieved plasma parameters and substantial Be melting could be observed for global energy deposition levels much lower than those of the design. This was mostly due to local overheating of the tile edges which were exposed to an estimated local power density >100 MW/m 2 , although melting was also found on the flat surface of the tiles.…”