Recently, a couple of us were talking about white LEDs.  White LEDs are a somewhat recent commodity.  Red LEDs have been around forever, yellow LEDs for the limit approaching forever, and green LEDs for simply a really long time.  Blue LEDs are more recent, but they have been around long enough to have been installed in nearly every electronic device and some varieties of toilet paper.  White LEDs are considerably more recent than any of these.  Actual white light is composed of a wide spectrum of light, ranging from violet to red.  LEDs are typically constrained to a small color bands, so how do white LEDs work?

White light can be simulated by combining red, green, and blue light.  This is the method used by TVs, LCD screens, and compact florescent lights.  Could white LEDs be doing something similar?  Could they just be red, green, and blue diodes combined into one unit?  The simplest way to answer this question would be to google it, but let’s wait on that.  This gives us the perfect opportunity to do some empirical experimentation.  Let’s look at the light spectrum produced by a white LED.  To do this we need a spectroscope…or some cardboard and a CD.  There are multiple sites out there with instructions on building your own.  Here is a photo of the one I built.  You can find instructions to build a similar one here.  BTW: You can also use these spectroscopes to determine what gasses are used in your florescent lights – a topic for another time.

If the white LEDs were combining multiple colors to simulate white light, it would look something like the image below.  This image, by the way, is the spectrum of a typical compact florescent light.  You can see “bumps” in the image that represent individual wavelengths of light that compose the simulated white light.

However, this is not what you see when you look at a white LED.  Below is the image of the spectrum I captured from a common LED flashlight.  This is interesting!  ”Why?”, you may ask.  It is surprisingly continuous.  It seems to be composed mostly of a wide, smooth blue/green band and a wide, smooth red band.  There is practically no yellow though.  This seems to indicate that there are probably not three separate diodes in there, but the wideness and smoothness of the spectrum is a bit strange.

Another interesting aspect of this spectrum is that the red and blue ends look like they have a similar shape.  I’ve sketched this out in the plot below.  Now, I’ve hit a point where I need to google for answers.  After a quick search, I found that many white LEDs are actually blue LEDs that have been doped with phosphors that fluoresce red when exposed to blue light.  This explains the observation that the light was mostly composed of two similarly shaped blue and red spectra.  So there, a quick little experiment that taught us something about how LEDs work.