Sommaire
- 1. Fabrication of new wafer BPC8 High Z design
- 1.1. Litho BPC8 1st step
- 1.1.1. vendredi 12/02
- 1.1.2. lundi 15/02
- 1.1.3. lundi 7/03
- 1.2. Litho BPC8 2nd step
- 1.2.1. mardi 8/03
- 1.3. Litho BPC8 3rd step
- 1.3.1. mercredi 9/03
- 1.1. Litho BPC8 1st step
- 2. Fabrication of new wafer BPCT8
- 2.1. Litho BPC8 1st step
- 2.1.1. vendredi 12/02
- 2.1.2. lundi 15/02
- 2.1.3. mardi 16/02
- 2.1.4. jeudi 10/03
- 2.2. Litho 2nd step
- 2.2.1. lundi 14/03
- 2.3. Litho 3rd step
- 2.1. Litho BPC8 1st step
started 2016-02
Fabrication of new wafer BPC8
High Z design
lift-off process on 1st step
Batch 3481, Si <100>, intrinsic, rho > 20kohm.cm, 300µm
Litho BPC8 1st step
vendredi 12/02
- spin
Reprocess once because numerous tiny spots in the spin.
Second time is much better, even though there are still spots, mostly on the edges
bake wafer >10' @ 110°C (solvent evaporation)
spin AZ5214E (Aurelie bottle, per 12/2014) 60" @ 4krpm
bake 2' @ 110°C
Expo 30mJ/cm2 @ 365nm, Vacuum contact -> 3" @ 10mW
bake 2'30 @ 125°C
Flood expo 30"+20"
Dev MIF726 1'10 +20" +20" @ 19.4°C, rinse ODI 60" beaker + 30" water tap -> reprocess: aceton + US
lundi 15/02
Right after testing the procedure on BPCT8
bake wafer >10' @ 110°C (solvent evaporation) spin AZ5214E (main bottle, per 12/2014) 60" @ 4krpm bake 2' @ 110°C Expo 20mJ/cm2 @ 365nm, Vacuum contact -> 2" @ 10mW bake 3'30 @ 120°C (by mistake: I was aiming at 2'30) Flood expo 45" Dev MIF726 1'15 @ 19.4°C, rinse ODI 60" beaker + 30" water tap
-> OK
- Residue removal
2' ozone plasma @ 100W, 200µb
lundi 7/03
- multilayer evaporation (old plassys)
P_init = 1.4e-7mb, N2 trap -> 1.1e-7mb - Al 150nm @ 1nm/s, P_ev ~ 2.5e-7mb, 100nm @ +7° / 51nm @ 0° - Ti 10nm @ 0.5nm/s -> P_ev = 3.0e-7mb (hot) -> 3.5e-8mb (cold) - Pd 4nm @ 0.2nm/s -> P_ev = 6e-8mb /!\ Loadlock pressure always very high!!!
- Lift-off
remover PG 70°C, 30'
Litho BPC8 2nd step
Refill the BCB bottle: made a test on Si/SiO2, and the BCB bottle was not closed so well. Very viscous.
Measured 6µm, with 200nm variation from the center (thicker) to the edge (thinner)
I wait for the main bottle to thermalize, and as it was held vertical in the fridge, there is no bubble owing to transport.
I spin right away, and it's OK
Now thickness is 4640nm on center, 4600nm on the edge.
- spin
bake wafer 10' @ 110°C spin AP3000 - 5" @ 500rpm acc 200rpm/s; 60" @ 6000rpm ; bake 1' @ 110°C spin BCB4024-40 - 5" @ 500rpm acc 200rpm/s; 60" @ 5000rpm acc 1500rpm ; bake 3'30 @90°C
- expose
Expo 70mJ/cm2 @ 365nm, Vacuum contact -> 7" @ 10mW/cm2
- develop
pre-dev bake 30" @ 65°C dev DS3000 at 31-32°C (very important) for 5min rinse DS3000 1' @ 20°C rinse ODI 1' water tap
NB: made a unexposed test wafer for the development time: 2'30 - 3' when interferences finish
added a factor 2 for the dev time of BPC8
mardi 8/03
- Hardening bake
Hot plate 1h @200ºC
Obs: no residues
- BCB RIE cleaning
SF6:O2 5:45, 250µb, 25W -> 70V, 35"
-
Litho BPC8 3rd step
- spin PMGI SF15
bake wafer 5' @ 110°C (solvent evaporation) spin PMGI SF15 (Leandro bottle, per 12/2014) 60" @ 4krpm bake 5' @ 200°C
-> 4.5µm thick
pattern height on top of BCB is still 2.2µm above transmission line patterns, and 1.4µm above coil bridge
- spin AZ5214
NB: there are still a lot of tiny spots on the spun resist!!! I try to clean the pipette the best I can.
spin AZ5214E (Aurelie bottle, per 12/2014) 60" @ 4krpm bake 1' @ 110°C Expo 60mJ/cm2 @ 365nm, Vacuum contact -> 6" @ 10mW bake 5' @ 125°C Flood expo 50" Dev MIF726 1'30 + rinse ODI 60" beaker + 30" water tap try in steps of 45" up to 5'15 where I see the Al starting to be etched
2 problems
- I get always the same problems with the AZ5124 resist, on small patterns
6" was ~OK for small patterns (here 3µm) which were well pushed on the mask,
but patterns where the mask was spaced (bad shape due to interferences) show that the dose was probably too high
-> ? revert to the initial AZ dose??? 30mJ/cm2
- Note that even at 5'15, the bottom of some patterns is not cleared from PMGI
So the other problem is that we get a too different etch rate for the 2 resists. We have to go to more than 5' to clear the bottom of PMGI, whil AZ is already largely undercut... Bad association.
I should have done
-> probably lower exp time to 4"
-> lower bake temperature to 120°C
-> lower bake time to 3'
-> press harder the wafer onto the mask
-> DON'T USE PMGI, but LOR instead, or bake PMGI at lower temp
-> Lift-off layer baked such that development takes 1'30 to clear the bottom of all patterns
For this step without etch, some evolution of the following recipe could do
(see 2016-01 test LORAZ-30B, BCB test2):
spin LOR30B (per //) 60" @ 3krpm bake LOR30B 5min at 170°C spin AZ5214E (per 16/12/2014) 60" @ 4krpm bake 1'30 @ 110°C Expo 62mJ/cm2 (5.6" @ 11mW) , Vacuum contact bake 3' @ 120°C (=setpoint). Flood expo 45" Dev MIF726 1min @ 19.0°C, rinse ODI 60" beaker + 30" water tap
mercredi 9/03
- clean
Ozone plasma, 40cc, 200µb, 100W, 1'
- etch + bridge evaporation (new plassys)
P_init = 1.4e-7mb - etch 500V, 65mA, 10", 16°/s planetary - Ti 5nm @ 0.2nm/s - Al 250nm @ 1nm/s -> P_ev ~ 4e-7mb
- Lift-off
remover PG 70°C, 1h
finish with US (Doesn't go after 25')
Almost all chips were etched by the developer, as I suspected. Too long to remove PMGI, cap layer not protective enough (also because on the initial evap, the mask closes, so the edges of Al are not protected.
Probably none will be useful, but we can still try to measure BCB losses.
RIP!!!
- PMGI/PMMA Bilayer spin
bake 120°C 10' spin TI prime @ 4000rpm 60" bake 120°C 1' spin PMGI SF8 (batch 12060407, exp. 7/1/2013) @ 3000rpm 60" bake hot plate setpoint 180°C, 5', no beaker (measured: 164.4°C) spin PMMA A6 (batch 14020130, exp. 3/1/2015) @ 6000rpm 60" bake hot plate setpoint 180°C, 15', under beaker spin UV III (viven bottle) @ 4000rpm 60" bake hot plate 140°C, 90"
Fabrication of new wafer BPCT8
In parallel with BPC8, to do the exact same steps in advance
Take wafer recycled from an old design BPC5. Lot 2451. It is DSP, so I take the back side to make the new litho (otherwise, litho would be annoyed by ghosts of previous patterns)
Litho BPC8 1st step
vendredi 12/02
- spin AZ
It's awful, I have to recycle the wafer 3 times, there are always tiny spots in the spin, I don't understand where they come from. Hopefully not from the resist.
I finally wash the wafer in the washing machine, and it seems to improve a bit.
note 2016-03-09:
Problem also appears on different resists (S18...), on the 2 spinners.
It may come from the pipettes. Indeed, when I wash one, it improves, when I wait a long time before spinning it improves as well -> Should try the siringes instead.
bake wafer >10' @ 110°C (solvent evaporation) spin AZ5214E (Aurelie bottle, per 12/2014) 60" @ 4krpm bake 2' @ 110°C Expo 30mJ/cm2 @ 365nm, Vacuum contact -> 3" @ 10mW bake 3' @ 125°C Flood expo 30" Dev MIF726 50" +15" +15" +20" +15" @ 19.4°C, rinse ODI 60" beaker + 30" water tap +15" flood +15" dev
50" 50+15 50+15+15+20
2µ patterns not open 1µm patterns not open 1µm patterns still not open
What is happening is that as patterns are quite small here, in places where the optical mask was not well pressed on the resist, they are exposed at the initial exp, and hardened. So the AZ mask is hardly open, and is blocked on the top while the bottom is developped. This explains why the dev time is so much longer.
-> reprocess: aceton + US
lundi 15/02
bake wafer 5' @ 110°C (solvent evaporation) spin AZ5214E (Aurelie bottle, per 12/2014) 60" @ 4krpm bake 1' @ 110°C Expo 20mJ/cm2 @ 365nm, Vacuum contact -> 2" @ 10mW bake 3' @ 120°C Flood expo 40" Dev MIF726 50"+20+20 rinse ODI 60" beaker + 30" water tap
There is still the same issue, though now I analyze it better. All the previous analysis was completed only now
(see below for a series of pictures describing well the issues)
I should have done (adjust for BPC8)
-> lower exp time to 2"
-> lower bake temperature to 120°C
-> lower bake time to 2'30
-> press harder the wafer onto the mask
mardi 16/02
- Residue removal
2' ozone plasma @ 100W, 200µb
Difficult to say if it is really cleaner
Below series of pictures taken along the AZ step on 2 resonators,
1 good from the beginning (left), one bad (right)
dev 50"
dev 1'10
dev 1'30
+ ashing 2'
- Al /Pd evaporation (old plassys)
- Al 150nm @ 1nm/s, P_ev ~ e-7mb - Pd 5nm @ 0.2nm/s
- Lift-off
remover PG @ 60°C, 2h
- spin protective S18
- Dice few chips
- Etching tests
The layer is etched in both MIF726 and MF319, at a rate much slower than aluminium (~15' in total to remove the layer), but still higher than Al/Ti/Pd (nothing visible in 15')
- Low temperature Measurements
Tc ~935mK
jeudi 10/03
With the rest of the wafer (remains 2/3)
Litho 2nd step
BCB bottle refilled on 07/03
thickness was 4640nm on center, 4600nm on the edge.
Now is 4590nm
- spin BCB
bake wafer 10' @ 110°C spin AP3000 - 5" @ 500rpm acc 200rpm/s; 60" @ 6000rpm ; bake 1' @ 110°C spin BCB4024-40 - 5" @ 500rpm acc 200rpm/s; 60" @ 5000rpm acc 1500rpm ; bake 3'30 @90°C
- expose
Expo 70mJ/cm2 @ 365nm, Vacuum contact -> 7" @ 10mW/cm2
- develop
pre-dev bake 30" @ 65°C
dev DS3000 at 31-32°C (very important) for 7 min
rinse DS3000 1' @ 20°C
rinse ODI 1' water tap
NB: made a unexposed test wafer for the development time: 4' when interferences finish
OBS: The patterns are much wider than they should, we see residues at the bottom, and yet it is clearly overdevelopped!!! (because patterns at 7min or BPCT8 are better shaped than those of BPCT9 that had 8min, both wafers made at the same time)
WE SHOULD NOT KEEP ON USING BCB THAT WAS NOT HELD IN THE FRIDGE ... Goddam
I try cleaning the bottom of the BCB before curing it, to check if I reprocess it or not
- BCB RIE cleaning
SF6:O2 5:45, 250µb, 25W -> 70V, 35"
EXTREMELY BAD IDEA!!!
Never do RIE cleaning before baking!!!
I had an enormous quantity of tiny spots of BCB redeposition, and it seems also that the Al/Pd layer was etched...???
- BCB hardening bake
Hot plate 1h @200ºC
OBS: broadening of the patterns
Try to clean the tiny spots
In the meantime I clean the RIE with O2:SF6, 10:1, 250µb, 60V for 30min
- BCB RIE cleaning
SF6:O2 5:45, 250µb, 25W -> 70V, 35"
No effect (except increasing the etch of BCB bridges)
But nothing on the tiny spots, neither increase of the quantity, size, no change of shape..., nor removal.
lundi 14/03
Litho 3rd step
- spin PMGI SF15 +AZ5214
bake the wafer 30' @ 110°C for water removal tape the wafer on blue clean room tape, cut to a ~3" square (half the width of blue tape) dispense PMGI SF15 1-1.5mL (no more!) slowly from a pipette well centered on wafer 2" @ 300rpm acc 250rpm/s 3" @ 800rpm acc 500rpm/s 60" @ 1500rpm acc 500rpm/s bake SF15 5min at 150°C spin AZ5214E (Aurelie bottle, per 12/2014) 60" @ 4krpm bake 1'30 @ 110°C Expo 50mJ/cm2 @ 365nm, Vacuum contact -> 5" @ 9.5mW bake 3'30 @ 120°C Flood expo 45" Dev MIF726 2' + rinse ODI 60" beaker + 30" water tap
Patterns are OK, but the PMGI unsticks from noble metal parts
Pd is not sufficiently protective and Al is etched.
We clearly see that the problem is only on metal parts, not on Si
Confirmation of the troubles we had with BPCv1
RIP!!!
I will perform a new test on this wafer to try improving PMGI SF15 adhesion, before complete abandon and revert to LOR20B
- Was this page helpful?
- Mots clés (Modifier les mots clés)
- Liens sur cette page
Fichiers 6
| Fichier | Taille | Date | Attaché par | |||
|---|---|---|---|---|---|---|
| BPCT8_1-10GHz_15mK_-105.5dBm.JPG Aucune description | 44.76 Ko | 17:45, 15 Mar 2016 | Helene_Le_Sueur | Actions | |
| BPCT8_2.6368GHz_15mK_-105.5dBm.JPG Aucune description | 55.56 Ko | 17:45, 15 Mar 2016 | Helene_Le_Sueur | Actions | |
 | BPCT8_3rd-step_SF15-150°C_exp47.5mJ_bake120°C-3'30_dev-MFCD26-2'_a.JPG Aucune description | 132.12 Ko | 12:28, 15 Mar 2016 | Helene_Le_Sueur | Actions | |
 | BPCT8_3rd-step_SF15-150°C_exp47.5mJ_bake120°C-3'30_dev-MFCD26-2'_b.JPG Aucune description | 85.23 Ko | 12:28, 15 Mar 2016 | Helene_Le_Sueur | Actions | |
 | BPCT8_3rd-step_SF15-150°C_exp47.5mJ_bake120°C-3'30_dev-MFCD26-2'_x5.JPG Aucune description | 138.61 Ko | 12:28, 15 Mar 2016 | Helene_Le_Sueur | Actions | |
 | BPCT8_3rd-step_SF15-150°C_exp47.5mJ_bake120°C-3'30_dev-MFCD26-2'_x5_2.JPG Aucune description | 118.37 Ko | 12:28, 15 Mar 2016 | Helene_Le_Sueur | Actions | |
| Images 6 | ||
|---|---|---|
 |  |  |
 |  |  |