Measuring the difference between gross and net erosion

A series of systematic experimental investigations are reported, using a beryllium-containing oven to seed a controlled and adjustable amount of impurities into a steady-state plasma. Beryllium impurity seeding of the plasma is used to simulate redeposition of eroded material to investigate the relationship between gross and net erosion. The measured net erosion rates from beryllium targets do not change when the influx of beryllium ions is increased unless the Be-seeding rate exceeds, by an order of magnitude, the beryllium surface loss rate, which was determined without beryllium seeding of the plasma. Measurements of the penetration distance of eroding beryllium atoms axially into the plasma column reveal that the reflection coefficient of the incident beryllium ions is unchanged, but the sputtering yield of the plasma-deposited surface beryllium is increased with respect to the bulk beryllium material by a factor of similar to 10.