Hurdles and Early Benchmarking Results for 157 nm Resists

IMEC (Leuven, Belgium) is conducting lithography research three generations ahead of state-of-the-art manufacturing, for the 45 nm device generation and beyond. The research organization has concurrent projects aimed at making 157 nm lithography a viable technology for the 45 or 32 nm nodes (Figure ). And, though there is much debate surrounding whether 193 nm immersion lithography might prove successful and extend through the 45 nm node, activity in 157 nm research continues, of which photoresist development is of fundamental importance. Though resist platforms have been identified, resist suppliers are still experimenting with many different approaches to maximizing the transparency of 157 nm resists, while meeting other necessary requirements for 157 nm patterning.

Significant progress must be made in several areas of 157 nm lithography development simultaneously. (Source: IMEC)

The use of vacuum UV (157 nm) lithography will likely require thin resists (~150 nm) because of the need to pattern <100 nm CDs and the high absorption of materials at this wavelength. Indeed, air, water and most organic compounds are opaque at 157 nm. Resists must also be thinner for higher-NA tools, better depth of focus and the prevention of post-develop pattern collapse. This thinning makes the challenge of building in etch resistance to a single-layer resist substantially more challenging than in previous generations, even as etch resistance remains a concern for the more mature, single-layer 193 nm resists (see "Fine Tuning Today's Photoresists ").

The thickness of the underlying antireflective coating is at issue as well. Since the coating does not scale as drastically, either a faster-etching antireflective coating or one with higher etch selectivity to resist is needed to minimize resist erosion, leaving enough for the substrate etch.

The enormously difficult absorption characteristics narrow down the possible resist chemistries to two basic categories: fluoropolymers and polysiloxanes, most likely a polysilsesquioxane (polySSQ). In addition to the right optical properties, the resist must demonstrate good solubility in a standard alkaline solution, low levels of outgassing, high sensitivity and good adhesion to the substrate. In a single-layer resist, good etch resistance is necessary as well. Fluoropolymers are a single-layer system, whereas polySSQ is a bilayer resist approach with the thin silicon layer acting as imaging layer and a thicker underlying organic resist providing etch resistance.

Intel compared the performance of several proprietary 157 nm resist platforms as well as multiple variations of the hexafluoroisopropanol-norbornene systems, and demonstrated the feasibility of some fluoropolymer-based resists. The researchers reported on the study at last year's SPIE Microlithography conference in a paper titled, "Intel Benchmarking and Process Integration of 157 nm Resists." They used an Exitech exposure system with 0.60 and 0.85 NA at International SEMATECH (Austin), benchmarking the lithographic performance by patterning lines and contact holes on a variety of substrates.

Using the 0.60 NA optics and 110 nm resist films, reasonable top-down CD-SEM images were produced with most of the resists. However, at 0.85 NA, the researchers reported clear differences. Only those resists with absorbance <1/µm provided acceptable imaging in thicker resists (180-200 nm). Intel went on to further test the top-performing resists to determine compatibility with bottom antireflectants, etch profiles, and resist erosion and dry ashing behavior of these fluoropolymer-based resists. The oxidizing etch chemistry for the bottom antireflective coating layer severely degraded the resist. However, patterning using the fluoropolymer resist directly on the substrates resulted in reasonable selectivity (as high as 4.5:1) and etch profiles. Standard resist ashing after etch did not degrade underlying features.

Outgassing is an extremely serious problem in 157 nm lithography because it leads to contamination of the scanner optic. Researchers from Selete (Ibaraki, Japan) and Osaka Prefecture University (Osaka, Japan) recently quantified resist outgassing during VUV irradiation using an in situ quartz crystal microbalance method. They determined that the outgassing rate strongly depends on the structure of the polymer backbone and blocking units on the resist. Commonly used t-butoxycarboyl (t-BOC) units attached to phenolic hydroxide and hexafluoroisopropanol groups readily decomposed upon VUV (146 nm) exposure.

BOC and methoxymethyl (MOM) units attached to hexafluoroisopropanol groups decomposed even faster than those attached to phenolic OH. Ethoxyethyl units decomposed faster than MOM units in the polymers. FTIR study showed the outgassing was primarily the result of photogeneration of acid labile protecting groups in the resist films. The researchers summarized their findings in a paper titled, "F₂ Resist Outgassing Studied by In Situ QCM Technique," also presented at Microlithography.