composite yield

**Composite Yield** is a **yield model that partitions die yield into systematic (fixed) and random (defect density-driven) components** — $Y_{composite} = Y_{systematic} imes Y_{random}$, allowing separate optimization strategies for each component. **Composite Yield Model** - **Systematic Yield**: $Y_{sys}$ — yield loss from design-process interactions, edge effects, and pattern-dependent failures that affect the SAME die every time. - **Random Yield**: $Y_{random} = e^{-D_0 A}$ (Poisson) or similar — yield loss from random defects (particles, contaminants) distributed across the wafer. - **Negative Binomial**: $Y_{random} = (1 + D_0 A / alpha)^{-alpha}$ — accounts for defect clustering ($alpha$ = cluster parameter). - **Separation**: Separate systematic and random yields by analyzing die failure patterns — systematic failures are spatially correlated. **Why It Matters** - **Targeted Improvement**: Systematic yield requires design or process changes; random yield requires defectivity reduction — different solutions. - **Mature vs. New**: New processes are dominated by systematic yield loss; mature processes by random defects. - **Prediction**: Composite models predict yield more accurately than single-component models. **Composite Yield** is **dividing blame between design and defects** — separating systematic from random yield loss for targeted improvement strategies.