Do edge filters work differently at 0° vs 45° incidence?

Do Edge Filters Work Differently at 0° vs 45° Incidence?

When working with optical edge filters, understanding the impact of angle of incidence (AOI) is critical for achieving the desired filtering. AOI refers to the angle at which light strikes the filter relative to its surface normal. Changing this angle — especially using the filter at 45° instead of 0° — introduces several important physical effects:

1. Blue Shift of the Edge Wavelength
When you tilt an edge filter from 0° to 45°, the filter's edge wavelength shifts to a shorter ('bluer') wavelength. The shift typically amounts to about 5–10% of the center wavelength, depending on the coating design. This happens because the effective optical thickness of the filter stack gets shorter as AOI increases, so constructive interference favors shorter wavelengths.
For example, a longpass filter specified as '750 nm @ 0°' might perform as a 700 nm longpass when used at 45°.

2. Polarization Splitting (s vs p)
At higher AOI (like 45°), the edge transition is no longer sharp and uniform. The cutoff for p-polarized light is blue-shifted more than for s-polarized light. If your light is unpolarized, this appears as a broader, less steep edge with possible color fringing. Precision applications especially need to account for this effect.

3. Gentler Edge Slope and Weaker Blocking
The edge transition (how quickly the filter switches from blocking to transmitting) gets less steep at 45°. Blocking close to the edge may also become weaker, especially with fast optics (large cone angles), which increases the range of incidence angles and further broadens the transition region.

4. Angle (Cone) Sensitivity
Real beams usually fill a light cone, not just a single angle. The wider the cone (lower f/#), the more the cutoff wavelength spreads — leading to 'edge smearing' and less precise filtering.

5. Reflections and Mirror Use
At 45°, the filter can act as a turning mirror, and reflectivity depends on polarization. You are below the Brewster angle for glass, but polarization-dependent losses and performance changes are noticeable and need to be managed in optical designs that use dichroic beamsplitters.

You should use a filter designed and specified for 45° when you need it to both filter spectrally and act as a beamsplitter or folding mirror. Common examples:

• Fluorescence microscopy dichroics: Reflect short-wavelength excitation and transmit long-wavelength emission at a 45° angle.
• Machine-vision coaxial illumination: Inject illumination onto the lens axis with a 45° shortpass/longpass.
• Stage/architectural/projector optics: Use hot or cold mirrors at 45° for color management or to dump unwanted IR/UV energy.
• Compact systems: Save space by combining spectral separation and beam folding in a single optic.

• Order filters for the AOI you will use. If you need a 750 nm longpass at 45°, purchase one specified as such, not a 0° part.
• Declare polarization and cone angle to your supplier, especially for precision or imaging work.
• Avoid large tilts unless the filter is designed for it—edge sharpness and blocking decline fast beyond small angles.
• For s-polarization, the shift can be approximated by:
  λ_θ = λ₀ √[1 - (sin²θ) / n_eff²] With n_eff ≈ 2 and θ = 45°, this means around a 6–7% blue-shift.

• Edge filters at 45° incidence behave noticeably differently than at 0°. The cutoff blue-shifts, polarization effects show up, the transition edge is gentler, and blocking near the edge weakens.
• For applications that combine filtering and beam folding, always specify and buy filters rated for 45°. Otherwise, use 0° incidence for optimal performance.