In fact the tubes are Sapphire - which is ok up to 2000C+ (2332K) , and operation is intermittent, so the plasma temperature might be too high for comfort, but the tube and the electrode temperatures fall within the bounds of possibility. Intermittence is even alluded to in the paper.
Alan, if the tubes are sapphire, why would they ascribe an emissivity of 0.9 and make their calculations based on blackbody emission from the tube surface? Surely it would have been recognized immediately that the tube was at a different temperature than the plasma if it was obviously clear like a sapphire tube. It makes it seem even more of a rookie mistake.