E-beam Lithography Resist Processing

Please note that there are new ebeam resist processes (below) and new ebeam resist handling procedures in place. See Resist List for complete list of resists and developers used in the lab.

This machine is capable of routinely producing 3:1 aspect ratios in a high resolution, high contrast resist, with a minimum feature size of roughly 50 nm. For example, if you want 50 nm features, plan on using resist no thicker than 150 nm; alternatively, if your process requires 0.30 µm resist, do not expect feature sizes below 100 nm. For greater process latitude, try to use lower aspect ratios. For sparse patterns with PMMA resist, it may be possible to make aspect ratios as high as 5:1.

In order to get features to print at their design size, a sizing may be specified in ELSA. Most positive resists require a sizing in the range of -0.04 to -0.08 depending on resist type, contrast, thickness, and development. Note that features equal to or smaller than the sizing get shrunk to nothing and will be deleted from your pattern. Although specifying a sizing will help linewidths in most cases, it will also result in an (apparent) increase in the proximity effect on some patterns. Negative resists generally require little or no sizing. Sizing may also be used to compensate for other effects such as etch bias. The resist thickness varies as the square root of the spin speed. However, the resist thickness is highly sensitive to the solvent concentration; ie diluting the resist by a factor of two will reduce the thickness by an amount much greater than a factor of two.

PMMA

PMMA is an ultra-high resolution, high current positive resist used for nanolithography. It has poor sensitivity and poor dry etch resistance. It sticks well to almost any surface, lasts forever, is not white light sensitive, and gives very reproducible results. There are several flavors of PMMA recipes that have been characterized at SNF:

Single Layer 2% 950K MW PMMA - Anisole <PDF>
Single Layer 5% 495K MW PMMA - Anisole <PDF>
Single Layer 5% 950K MW PMMA - Anisole <PDF>
Bilayer PMMA Resist Process for Liftoff
Single Layer 2% 950K MW PMMA - Chlorobenzene <PDF>
Single Layer 5% 495K MW PMMA - Chlorobenzene <PDF>

ZEP 520-12

ZEP is a very high resolution positive resist that like PMMA is simple to use and gives reproducible results. Compared to PMMA, it has an advantage of being 3 times faster and has good dry etch resistance. It has the disadvantages of poor adhesion (requires HMDS) and undercut profiles. This may be good for lift-off, however.

Although we post process notes for chlorobenzene-based ZEP520, please understand that we now discourage use of chlorobenzene-based resists in the lab. Instead, we now recommend the much safer, anisole-based ZEP520. Here are the spin-speed curves for anisole-based ZEP520.

APEX-E

Apex-E is a high performance, chemically amplified DUV positive resist that works well as an E-beam resist as well. It is extremely fast, has excellent dry etch resistance, and moderate resolution (0.15 µm). It's drawbacks are its high speed, which requires a large pixel spacing of .04 µm to keep the writing speed under 100Mhz, and sensitivity to airborne contamination. In order to get reproducible results, the resist should be spun, written, and developed all within a few hours. APEX-E is also very sensitive to contamination from underlying films on the wafer which may make it unusable in certain situations.

HMDS prime

Spin

PAB: 90°C 1 minute

Expose at 3 - 6 µC pattern dose

large area clearing dose is 2.5µC

PEB: 85°C 1 minute

Develop in MF319 1 minute (standard SVG track develop program)

Rinse: DI water

SAL-601

SAL-601 is a high sensitivity chemically amplified negative resist. It has very good resolution (sub 0.1 µm), high contrast, and moderate dry etch selectivity. The main drawbacks are scumming and bridging between features, particularly in dense patterns; poor adhesion (requires HMDS); a very short shelf life (hours on the wafer, days in a bottle at room temperature, months when refrigerated). Reproducibility is not great due to aging effects and sensitivity to bake time and temperature as well as minute levels of airborne contaminants. Because of the scum problem, this resist is best suited for isolated features. Since this is a negative resist, the wedge pattern does not work well. Try looking for color changes using wedge5, which requires a 0.04 µm pixel spacing to keep under 100 Mhz. SAL-601, like APEX-E, is also sensitive to the chemistry of the underlying film. In particular, nitride may reduce its sensitivity by 30%.

HMDS prime

spin

PAB: 90°C for 10 min.

expose at 5-15 µC/cm² (doses below 12µC may require 0.04 µm pixel spacing)

PEB: 115°C for 1 min on hotplate

Develop for 2 - 5 min in MF322 (try a develop time 2X the dose to clear)

May require a light RIE descum.

SNR-200

This resist has found to not have the scumming problems that SAL has and people are resolving smaller features. The following is a SNR recipe:

Singe 150°C for 30 minutes

HMDS prime (not critical)

Spin at 7.5KRPM for .5µm layer

Pre-bake at hotplate 120°C for 2 minutes

Expose. Do an exposure matrix here or start with 6.5µC.

Post exposure bake at 110°C for 2 minutes. (this is a critical step - do not vary time or temperature)

Develop for 20 sec in MF CD-14.

UVN2

UVN2 is high sensitivity chemically amplified negative photoresist with a wide process window overlap and PED stability > 2hours.

Singe 150°C for 30 minutes

HMDS prime

Spin at 7.5KRPM for 30 seconds = 4200A thickness

Pre-bake at hotplate 110°C for 1 minutes

Expose. 4.5 to 15 uC

Post exposure bake at 95°C for 1 minutes. (this is a critical step - do not vary time or temperature)

Develop for 30 sec in MF CD-26 developer.

UVN30

UVN30 is high sensitivity chemically amplified negative photoresist with a wide process window overlap and PED stability > 2hours. This resist is better at fine line-space resolution without scumming.

Singe 150°C for 30 minutes

HMDS prime

Spin at 7.5KRPM for 30 seconds =3800A thickness (or 4K =5300A)

Pre-bake at hotplate 140°C for 90 seconds

Expose. 4.5 to 15 uC (simular to UVN2)

Post exposure bake at 130°C for 40 secnds. (this is a critical step - do not vary time or temperature)

Develop for 30 sec in MF 702 developer, Normality 0.21(recommended by Shipley).

or Develop for 45 sec in MF 322 developer, Normality close to 0.21.

or Develop for (?30 sec) in MF CD-26 developer, Normality 0.26 (used by Cornell).

UV5

This resist has good plasma etch resistance (better than APEX-E). Resolution performance: For .5um thick resist: 0.2 um lines and spaces, 0.1um contact holes and 0.1um isolated lines.

Singe 150°C for 30 minutes

HMDS prime

Spin at 5KRPM for 30 seconds = 5000A thickness

Pre-bake at hotplate 130°C for 1 minutes

Expose. 10 to 20 uC

Post exposure bake at 115°C for 90 seconds. (The PEB delay stability has been shown to be greater than 90 minutes up to 2 hours.)

Develop for 45 sec in LDD26W developer

Hard Bake at 145C for 3 minutes

For more information contact James Conway

_______________________________________________________________
Stanford Nanofabrication Facility
Last Modified 02/29/2008