Layer Sensitivity
Layer Sensitivity shows which layers are the most critical to make accurately. It nudges each layer’s thickness by a small amount, measures how much the optical performance changes, and ranks the layers by that change. The layers at the top of the ranking are the ones that need the tightest monitoring and process control.
Settings
Section titled “Settings”Probe size: how big a thickness change to test on each layer. Choose Relative (a percentage of the layer’s own thickness, 1 % by default, and the usual choice since real errors scale with thickness) or Absolute (a fixed amount in nanometres). The sign does not matter; the tool always tests the same amount above and below nominal.
Scale: Normalized sets the most sensitive layer to 100 % and scales the rest against it (the easiest way to compare layers), or Absolute shows the raw change in merit on a logarithmic axis (use this when one layer dominates and you want its true size).
The ranking is evaluated for the surface mode set in the Design Editor (front, back, both, or symmetric), shown as a badge on the window.
How to read it
Section titled “How to read it”The window is the ranked table: one row per layer, most critical first, with its material and the change in merit its perturbation produced. Read from the top down, and the layers that need the tightest monitoring are the first few rows. Export writes the ranking to CSV.
Open Chart below the table for the same numbers as bars. A tall bar means a small thickness error there produces a large change in performance. A short bar means the layer is forgiving. A design whose bars are all about the same height is robust, with no single layer that can ruin it.
Layers are numbered as in the Design Editor (layer 1 touches the substrate), and locked layers are excluded. Sensitivity is measured against the optical performance, so the ranking reflects the effect on the spectrum.
References
Section titled “References”- H. A. Macleod, Thin-Film Optical Filters, 5th ed., §13.7.