Thanks all for your replies.
Gerry, I may try to torch it as you said, if a screwdriver can't get it out
Marylin, yes we are using the graphite funnel-tube
Wolfram, Stan, so this may be more normal than I thought. I didn't have this issue on the previous TCEA I used, but now that I think of it, I used to empty the crucible and cool down after each run, since I was dropping in large messy samples. The idea that silver slowly evaporates while the reactor is hot is interesting. In a way this would mean that you can run 300 samples without clogging and needing to empty the crucible, at the cost of risking to clogg the glassy tube below... it's almost like not using the crucible at all. We will play around the heating/cooling schedule to try and sort this out.
Harvard University, EPS, Johnston Lab