APEX Fabrication Steps

12-16-2002

1) Check out test wafers (p-type; boron dopant) from Microlab office.

2) Clean wafers in sink6 (piranha bath; Eva didn’t bother with HF bath because the wafers were new). Wafers have to be cleaned at clean sink (sink6) before they can go into the furnace.

3) Deposition of low stress silicon nitride in tylan 18: 1 micron thick, BSLOWF recipe (standard low stress nitride recipe) (we end up with .75 - .8 micron thickness, later got 0.9-1.1 micron thickness with time of 3 hrs 20 mins).

4) Al / Ti bilayer deposition in CPA. 300 Å Aluminum => 1 kW, 53.3 cm/min, one pass (subject to change, this is as of 7-22-02 for T_c of 500 K, depends on film quality of CPA; used to be 500 Å w/ 32.0 cm/min); 800 Å Titanium => 1.5 kW, 31.3 cm/min, two passes.

5) Ti photolithography (Ti mask).

a) HMDS in primeoven “1”

b) 1 micron photoresist (I-line) in svgcoat1/2

c) gcaws, program 8X8 (as of 6-20-2002 we use 3.8 sec exposure for 3.0s baseline on side of GCAWS and 255 focus, but these are subject to change, check side of GCAWS or do focus/expo test)

d) develop resist svgdev; prebake first, then develop

6) Ti etch in ptherm: first, clean machine and aluminum support wafers with O2 plasma (200 W, 100 sccm) for 20 minutes, then add wafer and etch Ti for 2 minutes, 200 W, 100 sccm SF6

7) Strip photoresist with acetone, isopropanol, DI water, then air blow dry

8) Al photolithography (Al mask); see step 5, using program APEXTI.

9) Al etch in lam3, standard Al etch process except: in Reactor #2 COMPL: TIME, WAIT: 00:25 secs, in Reactor #3 (Overetch) : 00%, and Airlock #1 set to 00:30s. Titanium part of sensor bigger than aluminum part, excess gets etched away in lam3, which keeps the edges of the sensor straight and prevents noise.

10) Remove resist with acetone and isopropanol, but if this doesn’t work, you’ll have to strip photoresist in technics-c (standard recipe 7 minutes, 51.7 sccm O2, 300 W), had to ash in Technics-C as of 12-9-02.

11) Thin gold photolithography (gold web mask), see step 5, using program APEXAL, NOTE: new liftoff method:

a) HMDS primeoven “1”

b) SVGcoat 1 micron I-line

c) Developer Presoak SVGDEV “1” “7” 60sec I-line no bake

d) Expo GCAWS as usual

e) Develop in SVGDEV as usual (“8” “2” PEB then “1” “7”)

12) Gold deposition in v401 (veeco) for liftoff process. Always wear polyethene gloves (clear plastic) when handling gold, boats, and stuff inside bell jar of Veeco. Clean gold first (*). First deposit thin layer (~20 Å) of Ti to help gold stick, then 100 Å (probably not right, gives way too much DC resistance ) of Au (this can all be done in one vacuum step).

13) Liftoff 1: let sit in acetone until the resist with gold layer comes off as one sheet. Sits at least a few hours, better to leave it overnight if possible. Can stick in ultrasound bath too if you’re feeling saucy.

14) Thick gold photolithography (gold lead mask), see step 5, using program APEXAL

NOTE for liftoff

a) HMDS primeoven “1”

b) SVGcoat1 “7” 2.8 micron I-line, “2” 90 C 90 sec

c) Developer presoak SVGDEV “1” “7” 60s I-line no bake then “3” “7” 30s I-line no bake

d) Expo GCAWS as usual, except exposure time should be longer for ~3micron resist (I used +0.6 seconds for each micron of resist above 1micron as a general rule of thumb)

e) Develop in SVGDEV as usual (“8” “2” PEB then “1” “7”)

15) Thick gold deposition in v401 (veeco) again. (clean *). Thin (20 Å or more) of Ti for stickiness, 1000 Å of Au (still tentative 12-2-2002). Its not really relevant anymore, but you should stop evaporation every 2500 Å to let photoresist cool down so that it does not harden too much.

16) Liftoff 2: same as step 13

17) SiN photolithography (nitride web mask), see step 5, using program APEXAL, and 2-3 micron resist.

18) Nitride etch in ptherm, ~9 minutes works for roughly 9550 Å of nitride at 200 W and 100 sccm SF6 (we have gold on wafer so we can’t use technics-c in VLSI room). Etch rate is just slightly greater than 1000 Å / min.

19) Dice wafers in disco.

20) Use xetch to release the spiderweb. 45 cycles, 15 seconds, 1.5 torr Xenon Difluoride, 0.0 torr everything else

21) Use ptherm (O2 plasma) to strip off the resist. We’ve had some trouble with this when the resist is on there for a long time (months). O2 setpoint 100.0 sccm, Power 200 W, time 7:00 minutes has worked in past.

* for steps 12 and 15, check out gold wire from the microlab office; before deposition clean it (and boats) in acetone 3-5 minutes in ultrasonic bath, then isopropanol, blow dry.

jjm 12-16-02