EUV InsightWhat interface roughness changes
Explore the effect of interface roughness on modeled Mo/Si reflectivity, with fixed layer geometry and a clearly defined roughness approximation.
Ideal interfaces are a useful reference, but they are an assumption. This example changes interface roughness while holding the nominal geometry and reference data fixed.
A controlled comparison
All comparisons use 40 Mo/Si bilayers on silicon, a 6.9 nm period, and a molybdenum thickness fraction Γ of 0.4. The incident side starts with Si. Mo and Si densities are 10.22 and 2.33 g/cm³. The sweep covers 12.4–14.6 nm in 0.01 nm steps with unpolarized illumination. Angles are measured from the surface normal.
Incidence is normal. The listed σ is applied to the top interface of every expanded film and the substrate boundary. The calculation uses the solver’s Névot–Croce roughness treatment.
| Interface σ (nm) | Peak reflectivity (%) | Peak wavelength (nm) | R at 13.5 nm (%) |
|---|---|---|---|
| 0.0 | 73.01 | 13.48 | 72.94 |
| 0.2 | 72.54 | 13.48 | 72.44 |
| 0.4 | 71.07 | 13.47 | 70.85 |
| 0.6 | 68.46 | 13.45 | 67.94 |
What the roughness model represents
The table describes changes in modeled specular reflectivity. It is not a calculation of a full off-specular scattering pattern. Roughness and chemically intermixed layers are different physical descriptions; do not assume a single σ reproduces both.
Reproduce the inputs
Open the default mirror in Pro and apply the listed σ to both group materials and the substrate. Keep thicknesses, density, wavelength range and angle unchanged. Changing only one boundary will describe a different sample.
Use measurements to constrain assumptions
Multiple combinations of roughness, density and thickness can influence a spectrum. A matching curve does not uniquely establish an interface structure. Compare independent metrology and document the chosen approximation when reporting the result.
Calculated examples use the bundled optical reference tables, not measured samples. Zr/Si/Mo example-data fingerprint: 92954053bf1c26ab. Assumptions and data sources.