2016-02

2016-02Edit

02/02

• Layer 3: LORPMGIAZ MASK

heat waffer in hotplate at 110°C for 10min
spin TI primer 60sec at 4000rpm
bake Ti primer at 110°C for 2min

spin LOR30B (per //) 60" @ 3krpm
bake LOR30B 5min at 170°C

---- Since the waffer is cut, LOR makes a real ugly thing; I decide to remove the risist.

Remove risist Remover PG @60°C for 30min + clean water 

mercredi 10/02

• cleaning BCB step

plasma clean: SF6:O2 (10:90 using 5cc and 45cc at P=250ubar -plasma off- and V=70V) 
for 30 sec (rate is 5-6nm/s)

heat waffer in hotplate at 110°C for 10min
spin TI primer 60sec at 4000rpm
bake Ti primer at 110°C for 2min

spin LOR30B (per //) 60" @ 3krpm
bake LOR30B 5min at 200°C
spin AZ5214E (per 16/12/2014) 60" @ 4krpm
bake 1'30 @ 110°C

Align+Expo 62mJ/cm2 (5.6" @ 11mW) , Vacuum contact + Alignment check

lunedi 15/02

bake 3' @ 120°C (=setpoint). 
Flood expo 45" 
Dev MIF726 1min @ 19.0°C, rinse ODI 60" beaker + 30" water tap 

It seems that 1 min is enough to get the little motivs developed.

I decided to clean the residues in a plasma oxigen

Plasma O2, 0.2mbar de O2, 40cc, P=100W, time=1min

Then I check at the microscope and change my mind. Decide to verify the effect of more development to be sure.

Dev MIF726 +10sec @ 19.0°C, rinse ODI 60" beaker + 30" water tap --> Obs
Dev MIF726 +30sec @ 19.0°C, rinse ODI 60" beaker + 30" water tap --> Obs

Result

Post-bake @ 150°C for 10min

Result: It can be observed that the motivs become larger

Decide to reprocess

Remove risist Remover PG @60°C for 40min + clean water 

-------------------------------------------------------------------END---------------------------------------------------------------------------

mercredi 10/02 

heat waffer in hotplate at 110°C for 10min
spin TI primer 60sec at 4000rpm
bake Ti primer at 120°C for 2min

spin LOR30B (per //) 60" @ 3krpm
bake LOR30B 5min at 180°C
spin AZ5214E (per 16/12/2014) 60" @ 4krpm
bake 1'30 @ 110°C

Expo 62mJ/cm2 (5.6" @ 11mW) mask IEF, Vacuum contact

lunedi 15/02 

bake 3' @ 120°C (=setpoint).
Flood expo 45"
Dev MIF726 40sec+10+20+60 @ 19.0°C, rinse ODI 60" beaker + 30" water tap 

Since the motiv corresponds to IEF mask, it is easy to know when all the resist is gone. The resist is not gone after 1min 10sec. I use another 60sec where I observe that the interference fringes start to collapse. The right time is in the middle.

Dev MIF726 +10sec @ 19.0°C, rinse ODI 60" beaker + 30" water tap --> Obs
Dev MIF726 +20sec @ 19.0°C, rinse ODI 60" beaker + 30" water tap --> Obs
Dev MIF726 +20sec @ 19.0°C, rinse ODI 60" beaker + 30" water tap --> Obs

Result

Post-bake @ 150°C for 10min

Result: It can be observed that the motivs become larger

Decide to reprocess

Remove risist Remover PG @60°C for 40min + clean water 

maredi 16/02

heat waffer in hotplate at 110°C for 10min
spin TI primer 60sec at 4000rpm
bake Ti primer at 120°C for 2min

spin LOR30B (per //2006) 60" @ 3krpm

Something happened to the LOR30B since the resist form very qnoying filaments in the borders of the waffer which makes it very difficult to continue.

I find that the best option is to remove the edges with Remover PG. This has to be done very carefully.  

bake LOR30B 5min at 180°C
spin AZ5214E (per 16/12/2014) 60" @ 4krpm

I observe that there are some spots or little black dot after the spin. I had to reprocess many times.

bake 1'30 @ 110°C

mercoledi 17/02

Expo 62mJ/cm2 (5.6" @ 11mW) mask Layer3, Vacuum contact
bake 3' @ 120°C (=setpoint).
Flood expo 45"
Dev MIF726 1min 5sec @ 19.0°C, rinse ODI 60" beaker + 30" water tap

The motifs are still not develpeded. I do 10sec more then 20sec and most of the motifs are there except the bridge. Therefore I do another 10sec and then it is. So it is 1min45sec

I think we have to think in another method since there is still a lot of undercut.

-------------------------------------------------------------------------------------------------------------------------------------------------

I process a new wafer for etch process
wafer Si intrinsic [100] 5-10 kΩ.cm, lot PCQ 129133-02 from University Wafer
Started jeudi 11/02
See BPC7 page for details

vendredi 12/02
Al 100 / Ti 20 / Pd 5

On standard Si <100> 300µm / SiO2 500nm

• evap Old Plassys

Al 100nm @ 1nm/s, P_ev = 5e-7
Ti 2*10nm @ 0.5nm/s (error on evap, make a skip step at 10nm, have to relaunch)
Pd 5nm @ 0.2nm/s (wait about 10min, because I initially don't have rights to use Pd)

• observation

No apparent default on optical microscope

• Spin 1 st step, exp old mask CPBv1 with DC pattern

bake wafer > 10' @ 110°C (solvent evaporation)
spin S1813 60" @ 4krpm
bake 1' @ 110°C
Expo 150mJ/cm2 @ 365nm, Vacuum contact -> 15" @ 10mW (fluctuates 10.2-10.5)
/!\ I made a mistake, and set the time to 3" initially.
So exposed a second time, with realignement -> not easy
Dev MF319 45" @ 19.4°C, rinse ODI 10" beaker + 30" water tap
Hard bake 10' @ 150°C

The hard bake has the following effect: reflow of the resist

• Residue removal

3' ozone plasma @ 100W, 200µb

I tried many different recipes for etching, none was convenient. After 24h I resign. Better adapted for lift off.
TODO table etch solutions + rates:

The best amongst all trials was:
- dip 5" in aqua regia (10mL HNO3 70%, 60mL HCl 37%)
supposed to be 3 HCl : 1 HNO3 : 2 H2O but could not find appopriate products.
supposed to yield Pd etch rate 6.5nm/s, Al etch rate 8.5nm/s, Ti alsmost 0
- followed by a dip in a mixture of diluted HF and Al-etch.

For aqua regia I observe:
- when no resist, the layer goes away completely in 3"
- when resist, need at least 3.5" to ensure no residues after lift, but above 7" the resist is gone
- it seems to be exothermic reaction, which provokes lift of the resist
- the rates are not really reproducible

For HF mix I observe:
- Al-etch seems to stabilize reaction (the resist is no longer heavily attacked and lifted)
- the etch rate is a lot faster than Al-etch alone: no more than 30" to finish etching of Al100-Ti20 after 3" dip in aqua regia

jeudi 18/02

There is now another idea to test related to the third layer for the bridge. The requirements we have:

- We have to go over BCB, it means that the resist has to be thick, at least 3microns.

- In order to avoid interference, the resist has to planarize.

- Since we need to do a lift-off, it has to be suitable, with a good undercut.

- Since we also need to recover ohmic contact with the first Al layer which is oxidized, the mask has to hold a strong Ar ion milling.

- We have a mask suitable for a negative resist.

Taking this into account we have tried:

• LOR5B+AZ5214E

AZ5214E (1.4 microns, inversible). LOR5B (500nm, conformal, high undercut rate). Using this bilayer it was not possible to go over BCB.

• LOR30B+AZ5214E

AZ5214E (1.4 microns, inversible). LOR30B (3000nm, conformal, high undercut rate). Using this bilayer we could go over BCB, not so strong edge effects. The problem is that we have to etch to recover contact and the mask does not hold.

We have tried baking harder the LOR30B (200°C) to get a smaller undercut rate but it is hard to get a smaller undercut in the patterns of the different sizes. We tried to bake the AZ to make it harder but the motifs get larger due to constraints.

Possibility 1: Add Germanium layer. Not guaranteed that it would work.

Possibility 2: Do not do the etch. The bad contact is a resistance in parallel with a capacitance. We have to check the rf impedance to know what is stronger

Possibility 2b: Add a new step with big squares to etch, and put a cap layer

• PMGI SF15 + AZ5214E

AZ5214E (1.4 microns, inversible). PMGI SF15 (3000nm, planarizing, moderate undercut rate). Possibility 3: do this. SF15 it would be optimal to planarize and has not so large undercut in comparison with LOR.

• S1828 

S1828 (3 microns, positive). Using this resist implies Possibility 4: transfer the mask LAYER 3 (first to mirror the design and second to inverse the mask and the design). The other Possibility 5: is to command a new inversed mask. This resist should have a small natural undercut which helps the lift. We should use less Al, 100nm for example. To test: does it goes on top of BCB, does it planarize?, does it hold the etching? To check all this.

venerdi 19/02

 I did a modification to the mask of the new design adding some squares to etch and then deposit a cap layer. I tried SPEC litho laser KLOE facility.

spin AZ5214 @ 4000rpm, bake @110°C 2min
expose 1micron beam, 13% modulation, speed...
develop in AZ351 B diluted 1:3 in H2O for 25sec;

Alignement is not optimal. It took 15min to write 10% of the mask. The squares qre not well defined. Needs to try more things and don't think will get a good result.

maredi 23/02

• S1828-TEST1 

I spinned S1828 in two waffers at 1000rpm and 2000rpm. The expecs are

I have measured in Filmmetrics

I keep the waffer spinned at 2000rpm which I will use to test the dose

spin S1828 @ 2000rpm, bake @110°C 1min

To test the dose I will writte different parts of the waffer so I can test at the same time the alignement. The recomended dose 300 mJ.

To take into account: S1813 160mJ mientras que S1828 300mJ. 300/160=1.875 and 2.8/1.3=2.15. Elodie says that for S1813 1micron beam 20% is good fro filling. So 38% should work. But since I want 4.5 microns and not 2.8microns there is also another factor which gives approximately 52%. I should do the dose test 30-40-50-60 in the four quarters of the waffer.

• TEST1 - optic lithography

S1828 spined @ 2000 rpm for 60sec, baked @110°C for 3min, exposed with 300-350-450-500mJ. After 4min10sec dev in MF319 only the litho with 500mJ dose is ready and after 5min10sec the litho with 450mJ is almost ready.

• TEST2 - optic lithography

S1828 spined @ 2000 rpm for 60sec, baked @110°C for 3min, exposed with 500-600-700-800mJ. After 2min dev in MF319 the three more exposed lithos are ready.

giovedi 25/02

• A comment on how to use your gds file in KLOE litho laser

EXPLAIN

• Taking PREVIOUS info into account, I will test with 50-60-70-80 percent for the litho laser

EXPLAIN

dose 50 close to focus point

dose 80 far away from focus point

As focus is lost, the result becomes terrible;

maredi 01/03/16

I try the new mask NEGATIVE LAYER 3 with S1828 resist. I use previous values of dose as a guide.

spin S1828 @ 2000rpm, bake @110°C 1min
expose 600mJ dose 
develop in MIF726 2min+30sec+30sec+1min;

Result:

Development times are longer than previous test. Need higher dose. Possible cause? Dunno. Maybe there is a remaining layer of Cr but this is rather unlikely.

remove resist in acetone @ 60deg (thermal bath)+IPA for 10min
spin S1828 @ 2000rpm, bake @110°C 1min (in paralell with a Si temoin waffer)
expose 800mJ dose 
develop temoin in MIF726 2min+2min;
develop true in MIF726 3min;

Result temoin:

2min -->

2min +2min -->

Result true:

Put the sample in the evaporator

mercoledi 02/03/16

Etch and evap

Etch Ar 500V 130mA for 1min40sec + 100nm Al evap at 1nm/s 

Result awful, terrible:

Two things can be observed:

- first: there was some resist remaining which did also reacted to the etch

-second: the resist was very damaged by the etch. Motives are lost in bubbles and explotions. 

Try to lift in acetone and there is Al all over (in the holes left by the resist as can be guessed from the pictures) and there is also a part of the resist that is caramelized and does not leave.

---------------------------------------------------------END-------------------------------------------------------------------------------------- 

giovedi 25/02

•  First Transfer --> positive resist

I take a Cr mask with AZ1518 ready to transfer.

expose AZ1518 for 9.1sec (91mJ)
develop in AZ400K:H2O 1:4 during 1min +clean water

Result

etch Cr etch for 1min + 30sec 

result

The resist is damaged after the etch. The motif is fine but there are some Cr residues. I have to try to clean it.

remove resist in Remover 1165 for 30min + clean water 

After the resist is gone the residues are still there.

Result (I took the photo on monday 29/02/16)

I noticed days after that it is in fact the original mask which was dirty.

• Negative transfer

I took a 4" Cr mask with resist and remove it

remove resist in Remover 1165 @ 60deg for 30min + clean water
spin AZ5214E 60" @ 4krpm (no acceleration at the end)
bake 2' @ 110°C
Expo 62mJ/cm2 (6.2" @ 11mW), Hard contact (cannot do Vacuum contact)
bake 3' @ 120°C (=setpoint).
Flood expo 50" 
Dev MIF726 1min @ 19.0°C, rinse ODI 60" beaker + 30" water tap  

RESULT

The result is awful, very dirty, full of little dots.- I decide to remove resist and restart this step.-

Then I realized that the first inversion mask was wrong and that the original mask was dirty.

lunedi 29/02/16 --- día bisiesto

I have a lot of troubles to get a proper mask.

•  sur First Transfer mask

I try to clean this mask

1' ozone plasma @ 100W, 200µb ---> little dots still there
etch Cr etch for 5sec + clean water 

Result

• sur Negative transfer mask

I remove the resist of the last essay

remove resist in Remover PG @ 60deg for 30min + clean water 

!!The Mask broke when I tried to spin the AZ again...!! Took a new mask, remove resist

remove resist in Remover PG @ 60deg for 1h + clean water 

A lot of problems to spin properly. Cleaned again and put it aside to use 3" mask.

• Negative transfer mask 3"

remove resist in Remover PG @ 60deg for 1h + clean water 
spin AZ5214E 60" @ 4krpm (no acceleration at the end)
bake 3' @ 110°C
Expo First transfer mask 62mJ/cm2 (6.2" @ 11mW), Hard contact (cannot do Vacuum conta6ct)
bake 3' @ 120°C (=setpoint).
Flood expo 50" 
Dev MIF726 2min @ 19.0°C, rinse ODI 60" beaker + 30" water tap  

Result

but there is still some strange features

and some residues which I didn't understand

1' ozone plasma @ 100W, 200µb ---> little dots still there

Result

I check the original mask

and it is very dirty...

• original mask

remove resist in Remover PG @ 60deg for 30min + clean water
15' ozone plasma @ 100W, 200µb   

• First Transfer 3"

In a new mask 3"

expose original mask Layer 3 in AZ1518 for 9.1sec (91mJ)
develop in AZ400K:H2O 1:4 during 40sec +clean water

Result

better!

1' ozone plasma @ 100W, 200µb
Cr etch 1min+1min in new etchant + clean watter

Result

remove resist in Remover PG @ 60deg for 30min + clean water

Ended the day with a nice lithography

maredi 01/03/16

Did a lot of things to end up with a negative mask for layer 3