Methods and assumptions

Review EUV Insight’s multilayer solver, optical reference data, numerical checks and model limitations before using a result.

The optical calculation

EUV Insight models planar thin-film stacks using a Parratt multilayer recursion. Inputs include layer thickness, composition, density, interface roughness, substrate, wavelength, angle and polarization. Transmission, reflection and absorption describe the specified model.

Angles are measured from the surface normal. Layer order is displayed from the incident side. Repeated groups expand into physical layers; the layer count includes films, not the surrounding media.

Reference data and attribution

Bundled atomic scattering-factor tables are sourced from the Center for X-Ray Optics (CXRO) at Lawrence Berkeley National Laboratory. The reference data draws on the work of B. L. Henke, Eric M. Gullikson, J. C. Davis, and subsequent contributors.

B. L. Henke, E. M. Gullikson and J. C. Davis (1993). Atomic Data and Nuclear Data Tables 54(2), 181-342.

CXRO reference and data Element-specific updates

EUV Insight is an independent application. Source attribution does not imply endorsement by the named researchers or institutions.

Bundled elemental scattering factors are combined with formula composition and density to calculate optical response. Interpolation is linear in log energy. The underlying data may have energy-range restrictions and missing real scattering factors at low energies; check application warnings before interpreting a result.

The bundled silicon table differs from a current upstream snapshot. The installed bundle is versioned; do not assume every table is identical to the latest upstream release. Imported optical data is a separate input and must be documented with the result.

Numerical checks and their limits

Ideal-stack energy-balance checks compare reflection, transmission and summed absorbed power.
A separate characteristic-matrix implementation provides numerical regression comparisons with the production solver.
Fixed Mo/Si examples check peak and band metrics for unexpected changes.
The public example tables are regenerated from the production solver and bundled data during the build.

These checks assess numerical consistency and detect regressions. They are not independent measurements of a manufactured optic and do not establish experimental accuracy for every material or wavelength.

Reproducing a public example

Each guide specifies geometry, wavelength, angle, polarization, density and interface assumptions. Default spectra and tables use the same solver as the app. Values are rounded for display; comparisons should use unrounded outputs and a stated numerical tolerance.

Model boundaries

A planar optical model does not establish mechanical strength, thermal response, contamination lifetime, or fabrication feasibility. Supporting structures, oxidation and interlayers require explicit treatment where applicable. Interface roughness is represented by a Névot–Croce approximation; it does not provide a full diffuse-scattering model.

Statistical manufacturing outputs depend on the chosen error distribution, correlations, sample count and acceptance conditions. They describe those assumptions, not a guaranteed factory yield.

Questions about an input or result?

Contact us with the public example, settings and discrepancy you want to discuss. Share only information you are comfortable providing.