AI summary by TwinMind
EUV Photoresists: The Next Generation?
Asianometry
22m 22s
Traditional Chemically Amplified Resists (CARs) Background
- Dominated semiconductor industry for nearly 40 years
- Critical to photolithography process using high-powered UV light exposure
- Applied via spin coating, then post-exposure bake and development
- Extremely delicate formulations, often custom mixed
- TSMC contamination incident: killed 10,000-30,000 wafers, cost $500 million
CARs Technical Components and Process
- Special chemical compounds with multiple components
- Photo acid generator (PAG) activated by UV light
- Generates acid that catalyzes dissolution inhibitor removal
- Single PAG molecule can catalyze hundreds of reactions
- Pioneered at IBM in early 1980s for KrF 248nm lithography
- Market dominated by Japanese companies:
- Tokyo Oka: ~30% market share
- JSR: ~30% market share
- Shin-Etsu: significant player
RLS Trade-off Performance Metrics
- Resolution: pitch resolution (smallest printable distance between features) and feature resolution (finest details)
- Line Edge/Width Roughness: measures displacement from ideal line, "shakiness" between edges
- Sensitivity: dose needed to change resist behavior
- Three metrics are coupled - improving one worsens others
- RLS equation: resolution³ × line edge roughness² × dose = constant
- More concerning as devices shrunk below 100nm
EUV Challenges for CARs
- Resist materials absorb DUV at molecular level vs EUV at atomic level
- Traditional CARs made from oxygen, carbon, hydrogen (low Z-numbers)
- Poor at "stopping" EUV photons due to elemental composition
- First generation EUV (2017-18, 7nm node) used modified 193nm CARs
- Photon shot noise issues: ~30mJ/cm², 2,024 photons per 10nm² plot, 2.2% random variation
- Solution: raised EUV exposure doses rather than increase sensitivity
CARs Evolution for EUV
- Acid diffusion control strategies by JSR
- Quenchers to restrain acid catalyst
- New fluorinated PAGs more sensitive to EUV wavelengths
- Metal-based sensitizers to boost existing PAGs
- Photosensitized Chemically Amplified Resist (PSCAR) by Tokyo Electron and JSR:
- Extra photosensitizer precursor (PP)
- Additional post-exposure step with 365nm UV flood
- Creates more acid at EUV photon impact sites
High-NA EUV Challenges
- ASML's next-generation EUV machines with higher numerical aperture
- Higher NA enables smaller features but creates problems:
- Depth of focus inversely proportional to NA²
- Thinner required resist layers
- Concerns about proper image resolution and etch resistance
- Expected to regularly print below 20nm pitch
- Molecular concerns: CARs are 3-5nm polymer chains, leaving rough edges
- Mechanical concerns: pattern collapse from liquid surface tension during development
Metal Oxide Resists Introduction
- Direct change when hit by EUV photons (no chemical amplification)
- Core materials: zinc or tin oxide attached to organic molecules
- Higher Z-number metals: 4-5x more sensitive to EUV than CARs organics
- Molecular rather than polymeric structure (smaller size)
- Better resolution and line roughness performance
- No unpredictable acid catalyst diffusion
- Robust enough to skip hard mask deposition step
Key Players and Patent Disputes
- Dr. Robert Brainard (University at Albany): pioneering work on metal-based EUV resists
- Inpria Corporation (spun out of Oregon State University 2007):
- Based on Professor Doug Keszler's work
- Promoting tin metal oxide resists for ~decade
- Funding rounds:
- 2014: $7.3M Series A (Samsung Ventures lead, Intel, Applied Materials)
- 2017: $23M Series B (Samsung lead, JSR joined)
- 2020: $31M Series C (JSR lead, added SK Hynix, TSMC)
- 2021: JSR acquired remaining 79% for >$500M
- Patent lawsuits:
- Oct 2022: Inpria sued Lam Research for patent infringement
- Jan 2024: SUNY sued Inpria/JSR claiming original invention by Brainard in 2011
Lam Research's Dry Resist Technology
- "Ether" dry EUV photo resist
- Metal oxide resist applied via atomic layer deposition (ALD)
- Dry development using plasma/non-plasma chemicals
- Eliminates liquid development process
- Reduces pattern collapse risk for small features
- Challenges: requires high selectivity, risk of corner rounding
- Ion beam post-exposure treatment to smooth patterns and reduce EUV dose
Implementation Challenges
- Substantial retooling required for coater/developer systems
- Fabs have 50-60 EUV machines requiring modification
- Metal contamination concerns:
- Surface anomalies on silicon wafers
- Transistor gate oxide integrity degradation
- Electron trapping in silicon
- IMEC testing showed low contamination levels but concerns remain
- Cost savings must justify substantial investments
Industry Outlook
- Metal oxide resists offer benefits but require significant fab investment
- Potential savings: reduced defects, lower EUV exposure costs, consolidated patterning passes
- CARs industry (40-year dominance) developing intense countermeasures for high-NA EUV era
- Competition between established CARs and emerging metal oxide technologies ongoing
