Building a Tubidity meter

OK Gustav, you will know that low energy (infra-red 800 - 1000nm) will penetrate MUCH better than high energy (blu-violet 380 - 480nm)

unfortunately silicon (& germanium) photodevices are MUCH more sensitive to infra-red (peak typically 900nm) and have little sensitivity to blue light.

So ideally you'ld like a detector that is uv/blue sensitive. You could use a VEML6070 and pair it with a UV LED.

As a compromise I'd recommend you use a high intensity blue LED (around 460nm) - or even white - and a silicon phototransistor that DOES NOT HAVE a black coating.

Ebay - loads for little money (a few pence). EG

An interesting alternative would be to use an LED as a detector (yes you can)
https://wiki.analog.com/university/courses/electronics/electronics-lab-led-sensor?rev=1551786227