BPC5

BPC5Edit

mardi 05/01

last wafer from batch 2602 [100] >10 kΩ.cm

desox HF:ODI, 1:20, 1'
rinse ODI 1'

-> put right away in evaporator

mercredi 06/01

finally we decide that we will make an extra step to open windows in the ground plane of the resonators to evaporate Cu/Au for good contacts of the CPB to ground + QP traps.
So for now forget about complexity

• evap BPC5 1st step

pump Ti 20nm @ 0.2nm/s
evap Al @ 1nm/s -> P_ch increases above 1e-6mb !!
stop it
degas Al @ 2nm/s + scan through the surface of the ingot
pump Ti 20nm @ 0.2nm/s
evap Al 100nm @ 1nm/s, spin 16°/s, P_ch ~8e-7mb (before opening sutter, was 5e-7)

• Dose tests on Si wafer standard

Test dose for S1813 on Si wafer. (this was not the best test for the bobine which is the most delicate part of the pattern). Litho made using hard-contact mode.

With 11mW calibrated on MJB4
- tested dose 14sec (154mJ) (develop in MF319 40sec)

- and 16sec (176mJ) (develop in MF319 35sec)

• Litho BPC5 1st step

spin S1813 at 4000rpm for 30sec
bake S1813 at 110deg for 1min
dose 16sec (176mJ)(bad idea: Al reflects a lot more)
dev 40sec MF319

surdosé!
-> Ne pas faire les tests de dose sur substrat Si
S1813 pas adaptée pour faire la bobine (?)

recycle waffer using acetone + ipa in ultrasound for 10' each;
etch residual resine in plasma oxigen (100W 0.2mbar 41cc) for 5+5+10 min

I thought that there were still residues but the pattern left corresponds to the etching of Al by the developer.

jeudi 07/01

Decided to change to S1805 since the width is 500nm and should be more appropriate for the bobine.

• litho BPC5 (2nd time)

spin S1805 at 4000rpm for 30sec
bake S1805 at 110deg for 1min
align over etched pattern
dose 9 sec (99mJ, nominal 100mJ) and dev 35 sec on MF319

 Alignment worked perfect. Even the dose is the nominal, the bobine is interrupted at the corners
This might be due to bad spacing of the wafer / mask (hard contact mode)

• litho BPC5 (3rd time)

recycle wafer using acetone + ipa in ultrasound for 10' each;
etch residual resine in plasma oxigen (100W 0.2mbar 41cc) for 5 min

spin S1805 at 4000rpm for 30sec
bake S1805 at 110deg for 1min
align over etched pattern
dose 8 sec (88mJ, nominal 100mJ) --- Change to Vacuum mode for the lithography
dev 30 sec on MF319

 This time the dose is right but more development is necessary

(Afterthought 11/01: no in fact, the Al layer is already attacked on the image above)

I tried plasma oxigen 3 min to check if there were just residual resine. The result showed that this was not the case.

developed 5 more seconds in MF319 (35sec in total)

Development seems to be inhomogeneous. Some parts, the Al layer is already attacked, some other parts there is resist left... Should use a developper which does not etch Al. (like AZ remover)

vendredi 08/01

• litho BPC5 1st step (4th time)

recycle wafer using acetone + ipa ;
spin S1805 at 4000rpm for 60sec
bake S1805 at 110deg for 1min
align over etched pattern
dose 8 sec (88mJ, nominal 100mJ) in vacuum mode 
dev 65 sec on MF319

Below, this wafer developed resp. 50", 65", and 75".
On the inside part of the corners, the Al layer is progressively etched.
At 50", we still see resist residues, while above 65" we see no residue, even at large scales.

• remove resist residues

clean 30" in ozone plasma 100W, 0.2mb

• Etching of Al layer

Etch in HF:ODI 1:20 30" (from SID1, I determined that the etch rate is around 3-4")
Around 25", I see the resist layer stripping off

In HF etch web documentation, it is known that F ions can diffuse through the resist,
which strips the resist layer away, even when it is well sticked.
To avoid that we can only increase resist thickness

-> decide to switch to 'Al etch'
-> have to reprocess the wafer anyways. (the bobines are gone, and the edges of the remaining Al layer are lightly etched)

• Wafer cleaning / reprocessing

30" HF:ODI 1:20 (all patterns disappeared, but not shorter)

• Evap 1st step

pump 2h ->  P_ch / P_sas = 9e-8 / 1.5e-7mb
degas Al -> 5e-7 / 7e-7mb
pump Ti nm @ 0.2nm/s -> P_ch / P_sas = 3 / 7.8e-8mb (hot)
                        P_ch / P_sas = 2.7 / 7.5e-8mb (cold)
evap Al 100nm @ 1nm/s, P_ch / P_sas = 3 / 4e-7mb (hot)

mardi 12/01

• Litho S1813 mask 2015-11 BPCv2

bake wafer 10' @ 120°C (solvent evaporation)
spin S1813 60" @ 4krpm
bake 2' @ 110°C
Expo 150mJ/cm2 @ 365nm, Vacuum contact -> 14.5" @ 11mW
Dev MF319 50" @ 19.0°C, rinse ODI 60" beaker + 30" water tap

OK, patterns are perfect

• Al-Etch (BPCD9 only). Peremption 2002 (!!)

in tall beaker with magnetic agitator 
+ manual movement (slow, repetitive, from back to front) 

after 2' the whole wafer seems to be etched
leave 4' to insure etching of small patterns, rinse ODI in beaker + 1' water tap

bobins seem to be a little overetched, but patterns are globally nominal

lunedi 18/01

• BCB step

I developed this step through several tests and observations from (11/01 to 18/01). I put some BCB in the syringe as soon as possible to let all the possible bubbles go.   

bake wafer 5' @ 110°C 
spin AP3000 - 3" @ 1000rpm acc 500rpm/s; 60" @ 6000rpm ; bake 1' @ 110°C
spin BCB4024-40 - 3" @ 1000rpm acc 500rpm/s; 60"@5000rpm ; bake 2'30 @90°C
align layer 2 from mask with layer one 
Expo 71mJ/cm2 @ 365nm, Vacuum contact -> 6.5" @ 11mW
pre-bake wafer 30" @ 65°C before inmediate development
Dev DS3000 at 31-32°C (very important) for 3min 30sec 
(previous check with test waffers), rinse ODI 60" beaker + 30" water tap

Result: FAILED! (I suck!!) Since I aligned with small aligment marks and not with the big marks to center the waffer, the result is well aligned but rigidly displaced!!

maredi 19/01

• removing BCB 

I looked in the wiki and used a recipe involving AZ400 developer. Aluminium was eaten and probably also Si. Waffer is lost forever.