New Alt. Membrane Release Recipe (Nitride RIE)
I have lost many chips during my MgF2 or Au or Pt process due to yield problems with my RIE.
I have wondered if the problem was due to power spikes. I have observed that my most poorly yielding etch runs usually have uneven power or power spikes during the run. Increasing window size (active area) also has some association with having poor yield, but this is uncontrollable from the perspective of RIE. With this in mind, I lowered the etch power by 25% to see if I could prevent myself from hitting a critical breaking threshold. Lowering the power also means that I need to determine the etch rate, as it will not necessarily be 75% of the original.
Etch recipe
- 2.5 mTorr O2
- 33 mTorr CHF3
- 75 W operating power (0-1 reflected)
- 75 mTorr operating pressure
- 5e-5 Torr base pressure
- 50-60 nm/min Si3N4 etch rate
I ran a series of etch times (30, 45, 60, 75 90 seconds) on gold (~55nm) and platinum coated (~40 nm) NPN , and then measured the combined film stack. Due to charging and the film motion, I really wasn’t able to distinguish one film from the other. I measured a high estimate and a low estimate, and tried to control for the film on each chip where I was able to.
So the etch rate is somewhere between 35-60 nm/min. I tend to believe the higher estimate of 48-60 nm/min, because it appears the film thickness plateaus around 60s, and a 50 nm/min etch rate would match nicely with a 50 nm thick film. This is an average etch rate; no entrance effects are considered here.
75 seconds at 75 W should maintain a complete removal of the nitride substrate, releasing a free standing film.