2014-02

2014-02Edit

lundi 03/02

We are going to developp only on the real wafer now, as the process seems irreproducible.
Undercut + Dose Test on a trash chip on wafer BPC1 and cleaving it. One part is to check undercut by sputtering Au on it and the other part is to check the overlap size after two angle evaporation of Au.

expo sap
dev sap
Obs
Below, dose 1.1 (left) dose 1.2 (right), and test dose for true pattern

Again we find this high inhomogeneity.

1st part, Evap: Au sputtering

SEM images  (BPC1 Test.zip):

-PMGI thickness (631+/- 27) nm, PMMA thickness (258 +/- 22) nm. In comparison to BPCD7-2 with PMGI (615+/- 15) nm  and PMMA (253 +/- 21) nm. So the thicknesses of the resists seems to be reproducible.
-For rel. doses x0.9-x1.1 undercut was visible but for x1.2-x1.4 the resist collapsed.
-comparing the undercuts with  BPCD7-2 (red numbers in the table), only structures with dose x1.0

undercuts are very similar but slightly larger for BPC1 Test.

mercredi 05/02

2nd part, Evap: Ar-ion milling and deposition at +/-22 degrees of 30 nm Au each.
lift: sap (issue with clean room flux down, have to do it in old clean room)

Obs of the lines:

- gap between the two gold lines: (538 +/- 22) nm
- the width difference between the first and the second Au line is in the range 22-57 nm which is below 60 nm (given by the deposited thickness of 30 nm which can close the PMMA opening on each side by 30 nm)
-Designed PMMA linewidth versus extracted PMMA linewidth (calculated from the width of the first Au line, angle of evaporation and PMMA thickness):

-The shown plot is in agreement with previous findings.
-Figures of the Au lines BPC1-TEST Au lines.zip
-summary of extracted values Summary 20140203.xlsx

Obs of the BPC:

- lots of dirt with size 200-500+ nm
- BPC in general look good and there are no significant issues, also they look similar (reproducible).
- Design changes to make:

note to Simon: no in fact, the pattern we chose was line 2 col 1 (with longer island, made from a polygon; this one is made from 3 paths), and the size of the wire was decreased both in length (decrease by 800nm, and add a path of 800nm dosed 0.5 on top of the end of the wire) and width from 80nm to 70nm

-images for the Au BPC BPC1-TEST Au JJ.zip

Meanwhile lifting the 2nd part of the test chip, we have bake the 3 diced chips from top of wafer BPC1 for another 25 min at 170°C as we observed the collapse of the PMGI on one part of BPC1-TEST.

-> Baked: 25 min on hot plate at 170 Celsius.

jeudi 06/02

BPC1_1
Exposure: sap. Design includes above (05/02) changes.
Dev: as usual with 1 min MIF726.
obs: the litho is quite clean, the undercut went slightly lower than on BPC1_test, but I think the small vertical wires are not here, cause I don't see them.
 

BPC1_2
Exposure: Increase the dose of the vertical wires to 1.5, and increment by 0.1 the other doses (even for the JJ and DC patterns)
dev: sap
obs: Again, the mask seems closed on the vertical wires
Now there are also residues in the openings of the mask -> we will try to remove them with ozone plasma

-> We shouldn't have baked prior to observation of BPC1_test
-> go cut some other chips from BPC1

BPC1_3 (a chip from the center of the wafer, close to the test chip)
Exposure: Go back to the pattern that went well on BPC1_test (ie pattern line 2 col 1, for dose 1 and above, with changes mentioned above)
dev: sap
obs: This time it's OK!

For comparison below, on the left how it looked like on BPC1_test, before evaporation (where all patterns went very well), how it looks like on BPC1_1, and how it looks like on BPC1_3

evap: sap
lift: sap
measurement probe station:
- bottom left JJ : open
- upper left JJ (47*43 nm):  kΩ ->  Ω.µm²
- bottom right squid (55x45nm and 55x42nm):  kΩ ->  Ω.µm²
- upper right squid (57*55 nm and 70*44 nm):  kΩ ->  Ω.µm²

Obs SEM:
Too bad: one gate is open (bad connection between large pads), and the other might be touching the island...

vendredi 07/02

BPC1_1 & BPC1_2: 
Test ozone plasma 100W 1'
obs: The mask has opened. See below, comparison before and after ozone

BPC1_4
exp same as previous, design arranged: gap increased from 70 to 80nm, and arms of the squid moved up by 50nm.
dev: sap
evap: sap together with BPC1_1 and BPC1_2
lift: sap

lundi 10/2

obs on ozone tests BPC1_1 and _2
Went bad!!! See below on BPC1_1 (left before ozone, right after ozone -see mask opening on the arms, bottom after double angle evap)

Measurements Probe station on BPC1_4:
- bottom left JJ (53*53 nm): 200 Ω -> 0.6 Ω.µm²
- upper left JJ (53*47 nm): 202.3 kΩ -> 503 Ω.µm²
- bottom right squid (66*50 nm with arm partly cut, and no left arm): 115.2 kΩ -> <380 Ω.µm²
- upper right squid (40*58 nm and 53*54 nm): 12.5 kΩ -> 65 Ω.µm²

Obs SEM:
patterns are OK on BPC1_4

NB: on DC part, we had a test of the symmetric design, which went very well:

mardi 11/02

first need to solder the SMP connectors onto the PCB, and to slightly widen the cutout for the chip, as it is cleaved and has a cleaving angle which makes it larger than the cutout (see top left on photo below)

bonding on TPT HB10 (power - time - force): 
first bond: 350 - 450 - 300
second bond: 380 - 500 - 320
All went very well, either from Au pad to Nb or from Nb to Nb