LiDAR and In-Cabin Sensing Push NIR Filter Specs Past the Catalogue

Automotive ADAS and LiDAR · 2026-10-12 · 3 min read

Automotive optical sensing has quietly become the most demanding volume market for interference filters. The reason is not the wavelength — 905 nm and 1550 nm are ordinary — but the combination of constraints that no catalogue part was designed for: wide angular range, wide temperature range, fifteen-year lifetime, and a cost target set by a car programme rather than by a laboratory.

LiDAR and In-Cabin Sensing Push NIR Filter Specs Past the Catalogue

Angle is the first casualty

A bandpass or longpass filter specified at normal incidence behaves differently at 30°. The passband shifts toward shorter wavelengths, the edge softens, and out-of-band blocking degrades. In a LiDAR receiver the returning light arrives across a range of angles determined by the scan architecture and the receiver aperture, so the filter is effectively asked to hold its specification across that whole range.

The conventional fix is to design the filter at a shifted centre wavelength so that normal-incidence performance is deliberately off-peak and the working angles land inside the passband. That is a design decision, not a purchasing one, and it has to be made with the actual angular distribution in hand. A supplier who is not told the angular range cannot make it.

Temperature is the second

Automotive grade parts see −40 °C to +85 °C, and sometimes more inside a roofline module in direct sun. Thin-film filters shift with temperature because the refractive index and the physical thickness of every layer change. For a narrow band this shows up as a few nanometres of drift, which is enough to lose signal at the edge of the passband if the design was centred exactly on the laser line with no margin.

The engineering response is to specify the passband with margin on both sides and to require the supplier to report the shift over temperature, not just at 25 °C. Parts built by electron-beam evaporation with poor packing density also drift with humidity, which is why dense processes — sputtering, ion-assisted deposition, IBS — dominate automotive programmes.

Blocking is the third, and it is under-specified

The filter has to reject sunlight, headlamps, and the pump light of a 1550 nm fibre laser system, which can be orders of magnitude stronger than the return signal. Blocking depth is usually quoted as an optical density over a range, but the range matters: blocking from 200 nm to 1200 nm is a different part from blocking 400–1100 nm, and the difference is layer count and cost.

What automotive buyers should put on the drawing

  • Centre wavelength or edge wavelength, defined at a stated angle of incidence.
  • The angular range the part must hold specification over.
  • Passband width and the minimum transmission inside it, measured.
  • Blocking range and optical density, with the measurement method.
  • Temperature shift, reported as a measured value across the operating range.
  • Environmental tests: damp heat, thermal cycling, abrasion, adhesion — with the standard named.
  • Substrate, thickness and any wedge or parallelism requirement.

The uncomfortable part

Most filter failures in automotive programmes are not coating failures. They are specification failures: an angle range never stated, a blocking band assumed rather than written, a temperature shift discovered during validation. The coating house can only design to the numbers it is given. Getting those numbers right at RFQ stage is cheaper than a requalification cycle in the middle of a build.

← All 400 articles

Need a Quote or Engineering Review?

Send your drawings, specifications or coating requirements — we reply within 24 hours.

Request a Quote

Related Articles

Request a Quote