My original goal was to design an instrument to advise me of the levels of ultraviolet light (UV) being radiated down by the sun. Being the easy-to-burn type, that's a big deal. Many commonly available photodiodes, being primarily made of silicon, are more sensitive to infrared (IR) than UV. So, I went looking for a UV-enhanced photodiode. My first place to go was Hamamatsu, who make a wide variety of optoelectronics (And I also found a photodiode designed to detect X-rays- I will check that one out later!). But a UV photodiode will probably require a filter to block visible light. Then I remembered an idea that I had seen years ago- using an LED as a narrowband photodiode (Forrest Mims III wrote the book on this approach. See Applied Optics Vol. 31, No. 33, 6965-6967. Also see a project that he is involved in to measure atmospheric haze.). With the plethora of LEDs now available- including UV LEDs- it was worth a try.
I built a basic amplifier using the LM324 op-amp (operational amplifier), and began trying LEDs. My initial attempt was with a super-bright red LED (630 nm), which was quite sensitive. I later tried both green (525 nm) and blue (465 nm) LEDs, with sensitivity decreasing with wavelength. With demonstated success, next, I will get for experimentation two IR LEDs (940 and 880 nm), a yellow LED (595 nm?), an aqua LED (502 nm) and a UV LED (395nm). So far, my initial tests show that I can reliably build a solar photometer (LEDs as photodetectors are not as sensitive as conventional photodiodes, so what levels of light they can respond to is limited. But LEDs are cheap and easy to obtain.) that can respond to various wavelengths.
I just finished my testing, and all but one (the 880 nm LED) worked well. Even illuminated by direct sun, the blue, aqua and UV LEDs will need two op amps to boost the signal to usable levels. Also, an LED as a detector operates at a shorter wavelength than it does as an emitter (perhaps by 30 to 50 nm). This seemed more pronounced at shorter wavelengths (525 nm green on down). IR (940 nm), red (630 nm) and yellow (595 nm) responded strongly or very strongly to illumination by another like LED, while green (525 nm), for example, worked as a better detector of blue (465 nm- strong response) than of green (very weak response).
So, now with seven usable LEDs, I have the makings of both a broadband photometer and a limited absorption spectrometer. I plan to amplify the signal of each and route with a CMOS analog switch (4066) to an A/D converter. The controller will read out the data and send it to an LCD.
My original goal was to design an instrument to advise me of the levels of ultraviolet light (UV) being radiated down by the sun. Being the easy-to-burn type, that's a big deal. Many commonly available photodiodes, being primarily made of silicon, are more sensitive to infrared (IR) than UV. So, I went looking for a UV-enhanced photodiode. My first place to go was Hamamatsu, who make a wide variety of optoelectronics (And I also found a photodiode designed to detect X-rays- I will check that one out later!). But a UV photodiode will probably require a filter to block visible light. Then I remembered an idea that I had seen years ago- using an LED as a narrowband photodiode (Forrest Mims III wrote the book on this approach. See Applied Optics Vol. 31, No. 33, 6965-6967. Also see a project that he is involved in to measure atmospheric haze.). With the plethora of LEDs now available- including UV LEDs- it was worth a try.
