Ultrasol S10 polish rate for poly and oxide
cmcg at stanford.edu
Fri Mar 21 01:09:56 PDT 2008
I went through another round of polishing wafers using the new parameters
Gary sent out earlier, and the new yellow pad. Qualitatively, the slower
rpm's and lower pressure seems to give a more uniform polish.
New settings, both heads down: 30sec --> 7.55 A/sec, 60sec -->
12.24A/sec, 90sec -->
Old settings, both heads down: 30sec --> 17.45 A/sec, 60sec --> 4.70 A/sec
(nanospec gave suspicious measurement for this wafer), 90sec --> 20.81 A/sec
Old setting, head 2 up: 60sec --> 11.30 A/sec, 90sec --> 22.69 A/sec
New settings, both heads down: 30sec --> 8.9 A/sec
Old setting, both heads down: 30sec --> 20.44 A/sec
This seems to suggest the polish rate for poly using S10 slurry is not
linear with time. Maybe there is some ramp up time before the polishing
actually kicks in.... I settled with 13A/sec, and assumed this would polish
entirely through the poly layer, and slightly into the oxide base layer,
which seems to have been the case.
As for selectivity of polishing rates b/w poly and oxide, the new settings
give roughly 13:9, while the old settings give 1:1. My prior run showed
faster polishing rates for poly (between 30-60A/sec) and similar polishing
rates for oxide, giving selectivity ratio somewhere between 3:2 to 3:1. The
difference may be due to the new pad, and might suggest the old pad is
preferable for improved selectivity. More data for oxide polishing rates
would be useful.
Should I post this information on the website somewhere?
-------------- next part --------------
An HTML attachment was scrubbed...
More information about the cmp