Homemade mask making
- Prepare film:
- Prepare file in LEDIT (
- Print at 4040 dpi at Pretech Graphics (6.5micron laser spot) (or 3386dpi at Laserlab), normal density. Or send file to photoplot store.
NOTE: for Pretech films, normal density exposure, dark field regions, light areas (e.g. spaces between dark lines) are smaller than design by about 15 microns (35+35u turns out 50+20).
- After receipt, film should stabilize in lab for at least 3 hours before use to ensure full dimensional tolerance.
- Cut film to size (4" square).
- Verify emulsion (printed) side by looking under microscope. Focus on features and then move microscope up; if you see bubbles in a plane above the features (features will then be out of focus), then the emulsion is down. Other indications: The emulsion side is somewhat less shiney. Also, with the present printer, the curl is such that the edges curl up when the emulsion is down.
- Blow off film with nitrogen gas. If necessary, it can also be cleaned with
DI water or isopropanol.
- Expose 8 seconds at 15mW/cm^2 (see note 1 below):
- Place new photomask plate on wafer tray so chrome and photoresist are up.
- Place film on top of photoplate, emulsion (less shiney) side down, so it is in contact with plate emulsion.
- Place glass plate on mask holder.
- Use 225 microns+filmthickness (*K) of pressure (1.5 turns up)
soft contact mode.
- Set timer for 8s (see note 1 below), and press EXPOSE.
- 30s puddle develop, 60s rinse, 60s spin dry.
- Place plate face up on spinner and roughly center.
- Turn on vacuum and rotate by hand, taking note of where each corner
pases over a fixed spot. Turn off vacuum and move mask toward center.
Repeat until it is centered to within about 1mm.
- Set spinner timer and speed to 60s @ 300 RPM, 60s @ 4000RPM for rinse and dry;
set portable timer tor 30s for develop.
- Dispense puddle of MF319 developer to cover photoplate, and start portable timer.
- When 30s is up, turn on spinner and dispense DI water for 60s.
- Allow to spin dry for 60s. Dispense nitrogen gas at center of plate
for a few seconds near beginning of spin period.
- 2s blanket exposure (see note 1 below)
- Remove blank glass plate from mask aligner mask holder.
- Center photoplate on wafer tray in mask aligner, and raise
- Expose 2s (see note 1 below)
- 15s puddle develop, 60s rinse, 60s spin dry.
- Etch chrome in CR7
- No post develop bake required for these large features.
- Use only beaker marked "CR7".
- Support mask in cutoff blue wafer holder.
- Agitate while etch to clear+25%
- strip photoresist in ACT (marked "for masks only")
film from laserlab had smaller space than line (about 45 micron line to 25 micron space). Wanted equal. 16 exposure gives close to equal, though some lines are wider than others (laserlab problem
At normal exposure, film from PreTech gives medium sized light features on a dark field which are about 15 microns smaller than design.
10 micron design lines still show up, but are a bit dark, and may not print
continuous. They are about 4 microns wide.
Minimum dark feature is about 18 microns. Minimum light feature is about 7
Note 1: 1/98 a light field mask quite overexposed using 8s+2s. Tried 6s+min
(at black edge) good with success, though still perhaps a bit overexposed.
In this case, a feature designed as 20 micron dot came out 30 microns on
the film, and 23 microns on the final mask.
I don't know if this overexposure is because the PR on this batch of
wafers is different or if there is some other change.
Light is still measuring 15mW/cm^2 on CI2.
Same 4/29/98--bad w/ 8s+2s, ok but maybe thin w/5.5s+1.5s.
1/5/00 26mW. neg photomask. 5s+2s very overexposed
Try 4s+1.5s. still overexposed.
1/6/00 Niyaz mask optimized at 3.5s + 1.0sblanket.
2/8/00 4s+1.5s(min on x10) is ok for IMPS mask
5/4/04 on new batch, good results with 6s + 1.75s (Igor Paprotny). note: 4.5 - 1.5s , very unexposed. 5 - 1.5 ok, but still some bubbles.