Measuring UV radiation from bulbs

It's known that CFL and LED bulb emit UV radiation. Which Arduino sensors can measure this radiation? Maybe if someone has built a UV sensor he could check if this measures anything from such bulbs?

UV is a whole range of the spectrum, it matters a lot where the energy is, UVA/UVB/UVC, so
the best answer is a spectrometer, or a UV probe with a range of optical filters. I don't think
it will be cheap.

Normal white LEDs are based on GaN heterojunction devices, which have an inherently limited
spectrum and no UV output as the output drops off before 400nm. Other LED types produce
UV below this wavelength, but these are normally explicitly used for UV generation (sun lamps
for instance), rather than general illumination.

Usually the manufacturers of quality LED lights love to show off the spectral response graph
in their literature, as cheap LEDs have poor colourer accuracy and more gaps in the spectrum.

Thanks for the detailed answer!
However if I don't want to rely on manufacturer's graphs (or he doesn't publish them) can I use Arduino sensor to tell the difference between low-UV home bubs and high-UV home bulbs (I read claims that also normal LED bulbs sometimes emit UV)?

One can use inexpensive UV LEDs as photodiodes to detect UV light.

A particular LED will detect a band of wavelengths, about as wide as the emission bandwidth, but shifted typically 50 nm to the blue. So, a 405 nm LED will detect (roughly) 335-365 nm UV. To cover the spectrum of interest you may need several different LEDs, or a different detection method.

My understanding is Ultra Violet (above violet) takes in a very wide range of the electromagnetic spectrum. Something like 400 nm down to 10 nm. I doubt you will find a one sensor fits all. They do make sensors with both an analog output as well as I2C. A simple Google of “UV Light sensor” or “UV Light sensor Arduino” will bring up a dozen hits. You need to decide what portion of the spectrum you are interested in. UV-A, UV-B, UV-C and shorter wavelengths. Since there is no one size fits all you need to decide what sensor you want for what range of UV.


The plastic of the LED housing may absorb in the UV, depending on what it is, which could confuse things.

If you want anything like a reasonably meaningful measurement you need something calibrated, preferrably
a spectrometer as the precise wavelength affects the permissable exposure level drastically, ie UVA is
much less damaging than UVB, and those are only rough categories that span a continuum of effects

There’s a very simple test for UV though, find something with fluorescent paint and a UV (high-pass) optical filter
that lets only UV < 400nm through, and then you can see any fluorescence easily - note that optical
filters like this are very expensive as they have to be made of exotic materials and sputter-coated with
multi-layer structure to provide the filtering. Normal glass blocks a lot of UV.

BTW if you’re avoiding UV light in winter you may need vitamin D supplements(!)