APEX-SZ Bolometer Array Fabrication Steps v2
Jared
Mehl
6-6-2005
- Obtain Si wafers, either 500 um thick or 150 um thick
(bonding to handle wafer still a research project in progress)
- Clean wafers in sink6 piranha bath and DI
water rinse, then spin dry in srd.
- Deposit ~ 1
um silicon nitride (SiN) onto wafers in tystar17. Use LSNSTDxxx recipe and calculate the
deposition based on current process monitoring data posted on Microlab
website. Time should be roughly 4
hours.
- Aluminum and titanium deposition in cpa. Load wafers and wait for system to pump
down to base pressure P < 2e-7 Torr.
Turn throttle on and adjust Ar gas flow obtain pressure of 6 mTorr. Deposit 400 Å Al using 1kW, 40.0
cm/min and 1 pass. Deposit 800 Å
Ti using 1.5 kW, 31.3 cm/min and 2 passes. Be sure to coat a few dummy wafers for alignment corrections
in later steps, focus/exposure tests, and T_c checks.
- If using 150 um wafers and bonding to 350 um handle
wafer, you would also coat backside of thin wafer with >1 um Al and do
the bonding step. Still a work
in progress!
- Ti lithography step:
primeoven setting “0” (1 min HMDS), svgcoatx “1” “1” (1.1 um I-line PR, 60 sec
bake). Shrink gcaws2
aperture to 37.5 + 5 small divisions in x direction and 37.5 + 15 small
divisions in y direction (7 mm x 6 mm aperture). Load APEX ALIGN mask and run ALIGN;1 program to print
alignment marks. Develop alignment
marks in svgdev “1” “7” (no PEB yet). Back to gcaws2, load ARRAY TITANIUM mask and run program
HEXPIX,1,2 to print array, then load blank mask and run HEXPIX;3 to clear
contact pads. Be sure to print at
least one dummy wafer here for later alignment corrections. Then svgdev “8” “1” for Post
Exposure Bake and svgdev “1” “7” to develop wafers.
- Etch titanium in ptherm. Run oxygen scourge on chamber and
spacer wafers (15 min, 100 sccm O2, and 200 W) and SF6 scourge (2-3 min,
60 sccm SF6, 125 W) to clean chamber.
Run wafers in ptherm, with base pressure < 15 mTorr on
left side gauge ( < 0.040 Torr on MKS gauge) and cooling water temp low
(< 20 C), for roughly 3:30 minutes at 60 sccm and 125 W. Then clean off PR with acetone bath,
isopropanol rinse, DI water rinse, and N2 gun dry.
- Al lithography step: same as Ti (#6) except do a 90s
bake on PR in svgcoatx, and only shrink gcaws2 aperture 10
small divisions per side. Load
ARRAY ALUMINUM mask, run a focus/exposure test, then run program
HEXPI2,5-8 with appropriate shiftx and shifty settings on your dummy to
correct for any alignment shifts.
Incorporate any alignment shifts into HEXPI2,1,2 program and print
your real wafers with this program.
Do not develop real wafers yet.
Then, contact print APEX ARRAY mask with ksaligner, using
vacuum contact, 9.0 s exposure time, and 30 um alignment gap. You have to align the wafer alignment
mark to the bottom of the mask alignment mark to account for gcaws2
offsets. Run your dummy wafer
first to check alignment and exposure time. Run PEB and develop in svgdev.
- Etch Al in lam3. Run 3 dummies first.
Then run standard recipe with following changes: Reactor #2 COMPL:TIME and WAIT:00:25,
Reactor #3 OVERETCH:0%, and AIRLOCK #1 WAIT:00:30. Clean off PR, either with acetone or ptherm
O2 plasma if acetone doesn’t work (technics-c has been giving us
trouble…).
- Gold web lithography step: same as Al (#8) but no ksaligner,
except on need 60s PR bake. Use
ARRAY GOLDWEB mask and run HEXPI2,1,2 (with appropriate alignment shifts).
- Liftoff gold web:
Clean off gold wire bundles and boats in acetone/isopropanol with
sonicator. Deposit 20 Å Ti on crystal monitor and 300 Å Au in v401. Deposit gold with a rate of ~10
Å/s. Leave in acetone bath
overnight (cover in aluminum foil) for liftoff to occur. Rinse wafer with isopropanol and DI
water afterwards.
- Gold lead lithography step: same as goldweb
(#10). Load mask ARRAY GOLDLEAD
and run HEXPI2,1,2 in gcaws2.
- Liftoff gold lead: same as goldweb (#11) except
deposit 20 Å Ti and ~400 Å Au (this still needs to be calibrated).
- Gold blob lithography step: same as goldweb (#10),
except use 2.8 um PR and bake for 120s.
Load mask ARRAY GOLDBLOB and run HEXPI2,1,2 in gcaws2.
- Liftoff gold blob: same as goldweb (#11) except now
deposit 20 Å Ti and ~1 um of
gold. Will need to do multiple
v401 runs to reload boats with gold.
Only depostion ~2000 Å at a time and wait ~ 10 minutes to prevent
wafer overheating.
- Nitride web lithography: same as goldweb (#10) except use 2.8 um PR and bake for
120s. Load mask ARRAY NITRIDE and
run HEXPI2,1,2 in gcaws2.
- Etch SiN in ptherm.
You have to be very careful here, if you overetch the nitride at
all you will weaken the spiderwebs to the point that they disintegrate on
release step. Do same scourge
step as #7. Etch with SF6 at 60
sccm and 125 W for roughly 14-16 minutes (???). The etch time seems to vary a bit. One should run a dummy nitride wafer first to calibrate
this.
- Dice wedge out in disco. Dice to the inside of aluminum cutting
lines except for line near contact pads, where you should dice to outside
to limit undercutting of contact pads.
- Deposit gold backshort in v401. You can cut a wedge-sized triangle in
aluminum foil and put kapton tape in the corners to hold the wedge onto
the top of the chimney. Use 20 Å
Ti sticking layer and ~ 2500 Å Au backshort.
- Release wedge in xetch. Use the program parameters 1.5 Torr
XeF2, 0.0 Tor N2, 20 seconds/cycle, and roughly 150 cycles. The number of cycles will vary and you
must verify total undercutting in microscope (you should be able to see
this on fake TES island in web opposite the real TES).
- Remove PR in ptherm with O2 plasma at 100 sccm, 200 W
for ~7 minutes. Check for total PR
removal.
*** Note we may soon be reversing
the order of Ti and Al lithography and etching steps. Stay tuned! ***