I recently found that when I spin coat my patterned wafers, the photo
resist (AZ 3312) film thickness variation exists.

There are ~0.8um deep trenches on my wafer. The PR thickness at one
trench wall is different from at the opposite wall. (They appear to be
different color under optic microscope).

For the trench on the wafer right side, PR near the right wall is
thicker than the left wall.

For the trench on the wafer left, PR near the left wall is thicker than
the right wall.

For the trench on the wafer top side, PR near the top wall is thicker
than the bottom wall.

For the trench on the wafer bottom side, PR near the bottom wall is
thicker than the top wall.

>From the above facts, I think this radial uniformity issue comes from
the photo resist centrifuge movement typically from the spin coat
process.

My spin coat steps are:

1. Dispense PR to wafer with spin speed = 0
2. High speed spin, speed=4KRPM, time=25 sec, accelerate =5KRPM/Sec

Nominal PR thickness is ~1.1um under this condition.

I know that I could add a low speed spin step before the high speed
step. My question is if this will help in improving the uniformity issue
I mentioned above?


Thanks,

Wei Cheng