Versions Compared

Old Version 39

changes.mady.by.user Wirth, Justin C

Saved on

compared with

New Version Current

changes.mady.by.user Wirth, Justin C

Saved on

Key

  • This line was added.
  • This line was removed.
  • Formatting was changed.


Table of Content Zone
locationtop
styledisc

Table of Contents
outlinetrue
indent25px
stylenone


Overview

Birck supplies a selection of photoresists for both optical and electron beam lithography. Resists that are not supplied may be purchased for use by individual research groups. The chemical mechanism for each resist may be slightly different, so be sure to verify the appropriate process for each resist.

Optical Lithography:

BNC Supplied:

  • AZ1518: Positive photoresist, thicknesses from 1-4 μm
  • AZ9260: Positive photoresist, thicknesses from 5-20 μm

Used at facility (not supplied):

  • SU-8: Negative epoxy resist

Electron Beam Lithography:

BNC Supplied:

  • 950 PMMA A4: Positive resist, BNC supplies:
    • 950: A2, A4, A6, A8, A10

    • 495: A2, A4

  • XR1541 (HSQ): Negative resist

Used at facility (not supplied):

  • ZEP520A: Positive resist.


Optical Photoresists

Composition

The AZ positive resists consist of a resin (Novolak), a photoactive compound or PAC (a DNQ compound), and a solvent (PGMEA).

General Properties of AZ/Ti Photoresists

Via MicroChemicals:


AZ1518

BNC Supplied Photoresist
Tip
title

BNC supplied photoresist


Expand

Properties:

AZ1518 is optimized for adhesion and stability in wet etching. It has broad photosensitivity from 310-440 nm, a minimum resolution of 0.8-1.2 µm, and is optimized for thicknesses of 1.5-3 µm (films thicker than 3 µm may begin to show "bubbles", a result of the nitrogen released during exposure that cannot escape the top of the film).  It has a softening point at 110 °C, and will begin to formed rounded structures at and beyond this temperature. Beyond 120 °C, the resist will react with oxygen, become brownish, and cracks will begin to form. At 170-180 °C, the resist will crosslink. It is stable in HF/BOE and other acids, though strong acids (Sulfuric, Nitric) and bases will dissolve it. Dry etching performance will be limited by the thermal stability and rounding temperature.

Process:

Starting from a clean, hydrophobic substrate surface, the resist should be processed:

  1. Spin Coating: AZ 1518 will reach a thickness of ~1.8 µm when spun at 4000 RPM.
    1. Resist dispense: Dispense resist in center of wafer/sample in puddle.

    2. Spin recipe: It has been found (for 4" wafers) that coverage is best with a direct spin to 4000 RPM

. Via AZ,

    1. without a dispersal step. The spin recipe would then look like:

      StepRamp (s)RPMDwell (s)
      0000
      12.0400040
      22.000

      For other thicknesses, consult the spin speed chart, via AZ (for a static dispense on 6" wafers):
      Image Modified

  2. Hotplate Prebake: 100 °C for
60
  1. 110 s, or 110 °C for
45 s.Expose:18 s on the MJB3, 15 s on MA6
  1. 55 s. (Manufacturer recommends 100 °C for 60 s, but the cleanroom hotplates seem to run somewhat cold.)
  2. Expose: For a silicon substrate (other materials will have a different reflectivity and require different dosages),
    1. MJB3: 18 s.
    2. MA6: 15 s.
  3. Development: For bath development, mild agitation is recommended. Either AZ 340 or MF26A are recommended, with development in AZ Developer or AZ 400K also possible.
    1. NaOH based: 1:5 solution of AZ 340:H2O for 50-60 s.
    2. KOH based: AZ 400K for ?.
    3. TMAH based: Undiluted MF26A for
20-
    1. 30 s, ±10 s depending on substrate reflectivity and resist thickness.
  2. Postbake: No postbake required for most processes, especially if used for metallization or liftoff.
    1. Postbake will increase the chemical and thermal stability of the resist for wet or dry etching. In those cases:
      Hotplate Postbake: 115 °C for 60 s.
  3. Removal: Acetone, Remover PG (or other NMP based stripper), or PRS 2000. Postbaked (above 120 °C resist will require substantially more time to fully remove, and may require heated NMP or Nanostrip for removal.


PGMEA (as AZ EBR Solvent or another brand name) can be used for edge bead removal or dilution.

References:


AZ9260

title Supplied
Tip

BNC

AZ9260 is supplied by the facilitysupplied photoresist


Expand

For thick (>5 µm) films of AZ9260, the required exposure time will depend on the photoresist thickness.

References:


AZ 5214

Note

 Not BNC supplied, must be purchased by individual research groups

LOR3B

Warning
titleIMPORTANT CONTAMINATION NOTE

Per Microchem: Ensure that acetone and resist waste are kept separate from LOR/PMGI waste streams. LOR/PMGI will precipitate in the presence of acetone, PGMEA, and ethyl lactate and may clog lines or form unwanted solids in the collection area.


Note

This resist is not supplied by BNC, it Not BNC supplied, must be purchased by individual research groups


Expand

LOR3B is a mixture of the solvents Cyclopentanone and 1-Methoxy-2-propanol, with a 'Polyaliphatic imide copolymer' as the polymer. Edge beads can be removed with EBR PG, and it can be lifted off with Remover PG.

References:


Electron Beam Photoresists

950

PMMA

A4

title
Tip

BNC supplied photoresist


Supplied
Expand

BNC

supplies a variety of PMMA blends. The PMMA specification (e.g. 950 PMMA A4 is supplied by the facility

Properties

Composition: 4% polymethylmethacrylate of

) indicates that the average molecular weight of the dissolved molecules (950 = 950,000

dissolved in anisole.

mw, 495 = 495,000 mw). The "A" indicates anisole is the primary solvent, which replaced older chlorobenzene blends. The number after the A indicates the weight percentage of polymethylmethacrylate dissolved. 

Currently stocked varieties:

  • 950: A2, A4, A6, A8, A10

  • 495: A2, A4


View file
namePMMA_Data_Sheet.pdf
height250

Properties

  • Positive tone.
  • Refractive Index: 1.49-1.52 at 632.8nm.
    Pros:
  • Good adhesion to most substrates
  • Very high resolution (down to 10 nm).
  • Long shelf life and spun film life
  • Insensitive to white light
  • Resistant to water, IPA/Methanol, TMAH based developers, and dilute acids for short periods
    Cons:
  • Poor etch resistance for dry etching
  • Low contrast
  • Attacked by acetone, HF, Piranha.

Manufacturer recommended process

(summarized from Microchem PMMA Data Sheet)

  1. Solvent clean of substrate.
  2. Dispense 5 - 8 mL for a 150 mm wafer.
  3. Ramp to 500 RPM for 5 s OR let sit without rotation for 10 s.
  4. Quick ramp to spin speed, holding for 45 s.
  5. Prebake on hot plate at 180 C for 60-90 s (Note: longer times will not negatively affect PMMA. Too cool of a bake causes burning/bubbling of resist in metallization chambers, and may gradually contaminate the chamber).
  6. Expose with dose between 50-500 μC/cm2 depending on equipment and polymer.
  7. Development: For high resolution, 1:3 MIBK to IPA for 60-120 s. Rinse in IPA or DI water immediately following develop to prevent scumming. Blow dry.
  8. (Optional) Postbake: 100 C hot plate for 60-90 s. Note that PMMA will reflow above 125 C.
  9. Removal: Will generally be removed by common positive PR strippers, including acetone. Thorough removal can be accomplished with Remover PG at 50-60 C.


Image Modified

 


Image Modified

Notes

Sensitivity of PMMA depends on the concentration, developer used, and accelerating voltage of the exposure (e.g. Rooks 2002) . At 100 kV for our PMMA, developed in 1:3 MIBK/IPA, 700 μC/cm2 may be a good dose for 2D features (Hoole 1997), with small isolated lines requiring a dose >3000 μC/cm2 .

PMMA References:

PMMA Processes/Common Issues

Burning/Bubbling in evaporators

PMMA burning/bubbling may be seen in metallization, particularly the E-beam evaporators. This is seen commonly as a result of insufficient pre-bake of the spun PMMA film before exposure. The recommended is 180 C for 60-90 s. If it's below that temperature, insufficient pre-bake temperature is likely the issue. If you're already baking at that temperature and time, it may be an issue with placement on the hotplate or poor contact.

If it's neither of those, it could be thermals in the machine, as PMMA is further crosslinked in the evaporators due to X-rays. However, this would likely be more of a cracking than burning.

If the sample was properly baked and thermal contact is good, it may be an issue with the evaporator. A crack in the crucible can cause thermal shorting.

Mechanisms

Cold Development


XR1541 (HSQ)

Tip
titleBNC Supplied

XR1541 is supplied by the facility

Properties

 

ZEP520A
Note

BNC Supplied photoresist, but HSQ is very different from other photoresists. Talk to Justin Wirth or Bill Rowe before using!


Expand

Hydrogen silsesquioxane is a spin on glass that crosslinks with exposure to e-beam. Unlike other resists, the spun layer is inorganic (contains Si, O, and H, no C) which makes it very useful as an etch mask or for high temperature processing. It functions as a negative resist that can be developed in 2.38% TMAH or 25% TMAH for higher contrast. HSQ is VERY expensive and much higher maintenance to work with than any other photoresist. HSQ spontaneously turns into glass, and this crosslinking is facilitated by higher temperatures, exposure to water vapor, exposure to glass, exposure to the e-beam, and with time.

Special handling procedures:

  • HSQ is ~$10/mL, with a 250 mL bottle costing ~$2500.
  • Store HSQ in the refrigerators at all times, except for when actively using a bottle of resist.
  • When removing a bottle from the refrigerator, it should sit out for ~20 minutes to warm some. Otherwise, water vapor from air will be more likely to condense inside the bottle.
  • Only use plastic pipettes with HSQ, glass pipettes with promote crosslinking and the formation of particles. We recommend 1 mL pipettes unless you have a very good reason to go larger.
  • Do not pipette out of the main bottle, get a small plastic bottle for your group. These can be obtained from the BRK 2251 Storeroom.
  • When done using the HSQ bottle, replace it in the refrigerator immediately.
  • You should always be mindful of surface conditions with HSQ as it is prone to adhesion issues if these are not considered. It spins best on hydrophobic surfaces (i.e. hydrogen terminated silicon).
  • It is critical to do a dehydration bake prior to spinning HSQ for good adhesion.
  • HSQ immediately begins to crosslink once you take it out of the refrigerator. This means that you should have a consistent process flow from removing PR → spinning → soft baking → exposure → development to get good results. Large swings in time (many hours) between steps will change feature sizes somewhat. Small changes (e.g. dozens of minutes to an hour or two) are not so critical.
  • Base dose will depend strongly on soft bake conditions and developer conditions. Because of this, you will find infinitely many HSQ recipes on the internet.
    • Justin recommends baking at 120 C for 120 s and developing in 2.38% TMAH (e.g. MF 26 A) followed by a 60 s rinse in DI H2O.
    • Higher temps or longer bakes will lower the base dose, resulting in more sensitivity and lower contrast.
    • 25% TMAH may be used for high contrast development (70 s soak, DI H2O rinse for 60 s), but this lowers sensitivity and is extremely dangerous.

Relevant Papers:

Improved time dependent performance of hydrogen silsesquioxane resist using a spin on top coat


ZEP520A

Note

Not BNC supplied, must be purchased by individual research groups


Expand

ZEP520 References:

Literature


AR-P 6200

Tip

BNC supplied photoresist


Expand

AR-P 6200 from AllResist, previously known as CSAR 62, is based on the same polymer as ZEP520A (a copolymer of a-chloromethacrylate and a-methylstyrene) with the addition of "0.5 % of a highly sensitive, thermo-stable, halogencontaining proprietary acid generator" (Chemical Semi-Amplified Positive E-Beam Resist CSAR 62 for Highest Resolution). The behavior of the two extremely similar, with AR-P6200/CSAR 62 exhibiting slightly increased sensitivity.

AR-P 6200 Product Info

CSAR 62 Info

https://arxiv.org/pdf/1611.07266.pdf

https://avs.scitation.org/doi/pdf/10.1116/1.4899239

Thinner: Anisole, sold as AR 600-02, ZEP-A

Adhesion Promoter: Diphenylsilanediol in PGMEA and Acetone, sold as AR 300-80

Some sources state that adhesion is typically excellent (e.g http://www.cnf.cornell.edu/cnf_process_ebl_resists.html#ZEP), and thus no promoter would be needed. 

Developer: Amyl acetate, sold as AR 600-546 or ZED-N50

Alternate developers may be explored, see below for internal results from AllResist on possible alternates:

Rinse: MIBK, sold as ZMD-D (e.g. http://research.engineering.ucdavis.edu/cnm2/wp-content/uploads/sites/11/2013/05/ZEP520ATechRepo.pdf) 

Stopper: Isopropanol, sold as AR 600-60

 Remover: NMP Soak at 85 C for liftoff, also dissovles in room temperature NMP and acetone (though these may leave a residue).

Heated NMP has been successfully used for liftoff, though the patterns still required sonication.

Other removers alternately suggested by manufacturer, but not used at BNC:

  1. AR 600-71, Mixture of >70% 1,3-Dioxolane and remainder 1-Methoxy-2- propanol AKA PGME
    1. AR 600-71 SDS
  2. AR 300-76, a mixture of Dibasic Esters (DBE): reaction mass of Dimethyl glutarate, Dimethyl adipate and Dimethyl succinate.
    1. AR 300-76 SDS




References

MicroChemicals Photoresist references:

General Properties of AZ/Ti Photoresists

Action of DNQ photoresists:

Diazonapthquinone-based Resists, Ralph Dammel, 1993.

"i-line Resists", Henderson Group, GA Tech