I remember a project in Scientific American many years ago which proposed using a distributed bunch of sensors using reverse-biased green LEDs to measure atmospheric pollution levels, but I can't immediately find it now.

Internet Search has become near useless of late, regardless of which Search Engine is used.

Hopefully you find it and let us know.

Forrest Mims was the first to propose using the LED as a sensor in the early 70s, not long after LEDs became common.

https://hackaday.com/2015/10/13/citizenscience/

Archive.org is the closest thing I have found for the old internet. If I'm interested in a topic the first I go to is archive.org.

probably you should donate and also see if you can save what you can: https://www.youtube.com/watch?v=a4KkiKDKw88

governments burn libraries

Probably something from Forrest Mims. I have his stuff going back to the 1970s and he was using LEDs like this for all sorts of communication and environmental stuff way back.

Here we go! https://web.archive.org/web/19980131092505/https://www.scien...

Found via https://www.researchgate.net/publication/234345754_An_Intern...

Very probably. This isn't it, but it's very closely related: https://www.researchgate.net/publication/243581659_LED_sun_p...

Yes. We both posted that he originated the idea at the same time.

> Which is good for ~30 Baud and distances of inches at best.

This feels like it would be part of the plot of a Neal Stephenson or Andy Weir book.

Our intrepid hero has to send super secret plans to save humanity through glass to his partner, so they can prevent Jupiter going supernova! But the Bad Guys are jamming wifi signals!

As a university project I created a Morse code transmitter and receiver circuit that used the same type of LED on both sides. It worked pretty well up to a few inches apart.

the distance and baud rate depend a lot on the specific led you're using and how you're measuring it