Microstrip on top of a Hall Sensor
The idea of our recipe is to have a dielectric layer between the 2DEG wafer and the stripline, and
to be able to remove it in case of bad news later, without damaging the Hall bar. The main
difficulty is to get some layer which can stand the UV lithography + lift off process, but in the
same time which can be easily removed in a soft chemical. There is was a photoresist who could
be lifted off in isopropanol, rather than in acetone.
Photoresists: LOR-3A, HSQ and AZ resist
Developer: AZ developer
Other Chemicals: Acetone, Isopropanol, ethanol, H2O2, H2SO4, DI-water, Hydrogen peroxide,
Metals for evaporation: Platinum and gold
Wafer: 2DEG GaAs/AlGaAs
Tools: glass slides, pipettes, diamond knife, lint-free paper, hot plate, ultrasound bath, spinner, mask
aligner and evaporation chamber.
1. Prepare AZ developer by mixing (1:4 with DI water).
2. Take AZ resist and HSQ out of the refrigerator to let them warm up, wait 30
minutes at least.
3. Pour only the HSQ amount that will be used into the white ceramic beaker and
use the plastic pipette to place on wafer. AZ can be poured into a glass vile, cover
with Al foil to prevent contamination, use a glass pipette to place on wafer.
4. Take hall sensor and clean by spraying acetone, IPA and ethanol followed by
blow dry in N2 gas. Use sonic only if needed, but only as last resort.
Acetone (5 min.) – Isopropanol (5 min.) – Ethanol (5 min.)
5. Blow dry with N2 gas.
6. Spin LOR photo resist on top of a hall sensor, using program H (40”@3000rpm).
This is the sacrifying layer which can be removed in EBR-PG in case something
wrong is happening later.
7. Bake for 5 minutes at 170 C
8. Spin HSQ resist using program H (40”@3000rpm). This is the dielectric layer, a
kind of glass, which is generally known to be resistant to solvent
9. Bake for 20 minutes at 190 C
10. Spin AZ resist using program (40”@4000rpm). This is the famous photoresist
which can be removed by isopropanol
11. Bake for 10 minutes at 100 C
12. Align the hall cross with the very center of either the microstrip or the polarizer
using the mask aligner dials.
13. Expose desired pattern in the UV lamp for 15 seconds.
14. Develop in the AZ developer for 70 seconds (very important that this time is
tracked as accurately as possible)
15. Immediately dip sensor into DI water to stop the development and wave it around
a bit in the water.
16. Blow dry with N2 gas
17. Deposit 3 nm of Pt (0.1 A/s, 150 -180 nm of Cu (0.4A/s) and 5-10 nm of Au
(0.1/s) following the standard chamber procedure, but evaporating slowly to keep
the resist from cracking.
18. Lift off the metal with warm (60oC) isopropanol. Use ultrasound at the end to
help, if necessary.
19. Rinse with DI water
20. Blow dry with N2 gas.
This is a single layer lift-off. In consequence, the edge of the stripline are usually not as
nice as with the LOR/S1813 process.
During the exposure/development process, part of the HSQ layer could be developed as
well. Students working with HSQ also noticed that. The recipe is anyway working
because there is still some HSQ on top of the LOR under the stripline. Don't be thus
surprised by that.
You can safely use ethanol to clean the integrated sensor at the end because it does not
remove any of the layers. Acetone would have a bad effect on the underlying LOR layer.
HSQ: The exact chemical we used is: FOX 12, by the Dow Corning Corporation Midland,
AZ P4110 photoresist. This one is from AZ electronic materials (70 Meister Avenue,
Somerville NJ 08876, tel: 800-515-4164).
AZ 400K developer.