Lift-Off with DUV Imaging + PMGI Underlayer

Developed by Demis D. John, ~2017-2019

This process is intended for Deep-UV Exposure on the ASML DUV Stepper. PMGI is used as the underlayer, which is exposed at the same time as the imaging resist.

Adjust spin speed or switch to a different PMGI formulation to tailor the underlayer thickness to your desired metal thickness.

Tips

Please see our Bi-Layer Lift-Off Tutorial to understand the limitations and requirements for good lift-off.

• Remember that gaps in between adjacent metal traces can lift-off during the develop!

• Need underlayer thickness approx. 2x the desired metal thickness.

• The underlayer will develop laterally from both sides which can lift-off the imaging resist, so:
  • Minimum gap between adjacent metals should then be: greater than 0.5 * underlayer thickness
  • Want at least 30-50nm of underlayer width left to support the imaging resist, so the PR doesn't fall over/collapse.

• Make sure to use a vertical evaporation line-of-sight such as EBeam4 or EBeam1. EBeam3 is not vertical unless special care is taken.

Process Limits

This process is written with PMGI SF-5 @ 4krpm, which spins to about 130nm

• Maximum metal thickness = 130nm/2 ≈ 65nm (can possibly increase up to 130nm* 2/3 ≈ 85nm)
• Undercuts laterally about 130-140nm, so minimum gap between metals = (2 x 140nm) + ~30nm ≈ 350nm

Adjust the PMGI layer thickness to adjust these process limits.

PMGI could be spun thicker, or double-spun to get thicker underlayers (develop time & undercut will increase accordingly).

PMGI underlayer could be replaced with LOL1000/LOL2000, or another developable underlayer.

Suggested Process for Liftoff

De-H2O Bake (eg. 220°C, 1m+) - or - Technics O2 ash (300mT, 100W), 15sec	Oxygen ash is faster and more effective
Spin PMGI SF-5 @ 4krpm (rcp 7)
POLOS underside clean: 2000rpm, ACE/ISO/N2
PMGI-Bake: 220°C, 3min (BrewerSci lift-pin hotplate).	Can just place wafer directly onto hotplate surface. Optionally can use lift-pins & Recipe "00 220deg, 3min Vac"; --> To prevent wafer sliding: Enable Vacuum with overhead valve, only for wafer landing and then turn off.
Spin UV-6-0.8 @ rcp 6 (3.5krpm)
POLOS underside clean: 2000rpm, ACE/ISO/N2
Soft-Bake = 135°C, 1min (builtin hotplate)
(check underside for particulates)
ASML Exposure
Default: Exp = 37.5mJ // foc = –0.10	Try this exposure dose but might need to do a FocArray (smaller Image Distribution eg. 5x5) to find proper exposure.
PEB = 135°C, 1.5min (built-in hotplate)
Dev (300MiF) = 50sec (CRITICAL time)	Use cassette & H2O rinse dish prepared --> not stirring, instead very slow/gentle lift/drop at various angles or gentle swish just to mix developer slightly.
DI rinse	Very Gently! Dunk into prepared DI dish, and dump/fill gently 2x. Don’t allow direct hard water to hit PR surface.
PEii Technics O2: 30sec (100W/300mT)	May increase feature size openings by ~50nm.
Metal Evaporation (insert steps)	EBeam#4 or EBeam#1 with vertical/4-inch holders (not planetary)
Lift-off in NMP	Facing down or vertical. Optional: going directly into continuous Ultrasonic can reduce residual particles.

Data

• For PMGI Baked at 220°C for 3min, the PMGI dissolves in AZ 300MIF at a rate of approximately ~600nm/min.

Images / Examples

SEM of holes in PR with above process. Top layer is UV6, 140nm underlayer is PMGI, substrate is Silicon. Lateral measurement of undercut is only approximate because circle may not have been cut through the exact diameter. (Demis D. John, 2017)