How OPC Mitigates Stochastic Effects
Optical Proximity Correction (OPC) was originally developed to correct
deterministic optical distortions.
For advanced nodes, OPC has evolved to address stochastic variability
by improving robustness rather than attempting perfect correction.
Design Philosophy
OPC does not eliminate randomness.
Instead, it reduces sensitivity to randomness and minimizes the probability
of failure in the tails of the distribution.
Targeting the Center of a Safe Distribution
Rather than targeting the nominal mean critical dimension (CD),
OPC may intentionally bias the target toward the
center of the process-safe distribution.
This reduces the likelihood that stochastic variations push features
beyond manufacturing limits.
Mitigating Photon Shot Noise
- Reducing CD sensitivity to dose variations
- Optimizing assist features to improve exposure uniformity
- Stabilizing edge placement under local dose fluctuations
Mitigating Resist Stochastic Variability
- Smoothing edge fragmentation to avoid amplifying local randomness
- Favoring robust feature shapes over aggressive correction
- Optimizing for CD distribution rather than single-value accuracy
Handling Failure-Prone Patterns
Certain layout patterns are statistically more likely to fail due to
stochastic effects.
These failure-prone patterns may not fail deterministically,
but exhibit higher probabilities of defects.
- Applying conservative bias to known weak patterns
- Reducing necking and marginal geometries
- Focusing optimization on high-risk hotspots
Summary:
Modern OPC is a statistical optimization process.
Success is measured by reduced variability, improved robustness,
and lower defect probability rather than perfect pattern fidelity.
↩ Go to EDA Portfolio
↩ Go to Home – Tech Notes Park