I’ve done some work with near infrared spectroscopy on a similar problem to the tricorder “molecular scan.”. There are two-three main problems as I see it.
A typical lab spectrometer might collect on 3,000 different frequencies to cover the spectrum. Meanwhile the sensor that is cheap enough to put in a tricorder has around 10 channels.
The lab instrument probably has expensive and fragile optics. You can’t do the same thing on the tricorder because the optics will break when you drop it.
Lab procedures rely on carefully controlling the illumination so it’s the same every time. Hard to do in the field, even in relatively benign field environments. This even comes down to using sample cuvettes that are precision machined to have two sides extremely parallel. You can’t make a tricorder that dispenses precision cuvettes for sample collection. If you can’t control the illumination, you have to measure it and calibrate.
I’ve done some work with near infrared spectroscopy on a similar problem to the tricorder “molecular scan.”. There are two-three main problems as I see it.