Home › Guide › Detectors
Silicon APD vs InGaAs APD: why the wavelength decides the detector
905 nm and 1535 nm are not two settings on the same instrument. They need different detectors, and the detector is where the cost and the supply risk live.
Silicon stops at about 1100 nm
Silicon's bandgap makes it essentially blind past ~1100 nm, which is why every 905 nm and 940 nm part in the catalogue can use a cheap, fast, well-supported silicon APD or SiPM. High repetition rates (kilohertz), small apertures, single-digit-gram modules — all downstream of that one material property.
1535 nm needs InGaAs
Indium gallium arsenide responds across 800–1700 nm, so a 1.54 µm return can be detected at all. The trade: InGaAs APDs cost more, need a temperature-compensated bias (the multiplication factor moves with temperature and bias), and have higher dark current. Expect the datasheet to specify bias stability, bandwidth and multiplication range rather than just "sensitivity".
What that means for your design
- Bias supply: InGaAs needs a stable high-voltage bias, often with a temperature compensation loop. Budget board area for it.
- Noise: dark current rises with temperature, so hot-environment performance is a bias-design problem as much as an optics problem.
- Cost curve: at volume, silicon parts fall to single-digit dollars while InGaAs stays an order of magnitude above. That gap, not the laser, is why 1535 nm modules cost more.
- Supply: InGaAs APD supply is concentrated. Qualify a second source or accept single-source risk — and ask for the APD part number before you commit.
The filter matters more than people expect
A narrowband filter at the operating wavelength buys background rejection, which converts directly into range in daylight. A 1535 nm module with a 1535 ±5 nm filter is doing far more work than a 905 nm part with a 40 nm band — and that is part of why the long-range family performs in haze. The pages in our catalogue list the filter where the manufacturer publishes it.