Spectra of Some Light Sources


An associate of mine built a spectrometer one weekend1. So I asked him to measure the spectrum of some common light sources. Here are the results.

It will be useful to have a legend which translates wavelength to color:

(1) Actually, he said, "This spectrometer is NOT completely scratch built. This one was constructed by extracting the optical bench (fiber coupling, grating, mirrors, and the very fragile linear CCD) as a complete unit from an older ISA bus spectrometer (ala eBay). I laid out a board and then built an Atmel controller and a 1Msps 12bit A/D to control the CCD and download the data over serial. In this new form it is similar to what you can buy currently. The cheapest one available is about $600-$650 (only just recently available) and they go up fast from there depending on various things."

Source: Wikipedia "visible light spectrum"
(units in nanometers)

Visible light

This first spectrum was made from a 5 watt tungsten-halogen bulb.


Notice that it is continuous, meaning the curve is relatively smooth. The quantization effects in the far left are due to dynamic range and the limits of amplitude resolution in his spectroscope's A/D converter. It has no sharp peaks or valleys. Our eyes perceive this light to be "white".

This next spectrum was made by pointing his fiber probe up through his glass skylight during daylight.


It illustrates a number of intresting features known as Fraunhofer lines— absorption of light at certain frequencies by elements on the surface of the sun or in our atmosphere.

Here is a similar graph from Wikipedia which identifies the various Fraunhofer lines with their absporptive constituents.


The next spectrum was made from the compact florescent (CCFD) bulb in his ceiling fixture. This is one of those heavily pushed "energy saver" bulbs that cost 5x normal bulbs, are supposed to live much longer but in fact die at about the same rate as normal bulbs, so that all we're really doing is paying five times as much to push the energy useage from our homes to the manufacturers.


They are also pretty lousy sources of light as you can see. Rather than containing a full spectrum they consist of a series of more or less distinct colors separated by dark regions. What this means is that if you have an object whose color is restricted to a region missed by the light — most notably reds and blues — it will appear dark. (Hey, Mark – can you show an example of this?) This light does best at illuminating greens, yellows and oranges.

This is a 405 nm (blue) laser.

spectrum of blue 405 nm laser.

This laser is in the blue-violet range, just above ultra-violet. It is a very pure color (narrow spectral peak). The sub-harmonic at 810 nm is not actually present in the light, but is an artifact of the grating used to split the beam in the spectrometer.

When 405 nm light shines on a ruby it fluoresces — absorbs the light and re-emits it at a different frequency. This effect is shown in next spectrum made from a natural ruby (corundum) that Mark bought at a rock shop in Northern California.


Fluorescence generally occurs at longer wavelengths (lower energy) than the excitation frequency — in this case, at 694 nm &mdash. Very cool!

That last one looks like he may have had the room lights on at the time. Here is another made in my presence where the room lights were definitely off.

spectrum of corundum (ruby) florescense.

Again, the 405 nm spike is the laser excitation wavelength, the 694 nm spike is the resulting fluorescense from the ruby, and the 810 spike is an artifact, not actually present.

Oh, by the way, you should never look into a laser or point one in someone's eye - the damage happens so fast it's over before you are even aware of it.

laser warning sign

spectrum of He-Ne laser.

The spectrum below was made from a blue CCFL bulb (cold cathode fluorescent lamp).

spectrum of blue CCFL bulb.

The blue CCFL bulb above appears a very similar color to the blue LED below:

spectrum of blue LED.

UV Light

Ultra-violet light is, technically speaking, invisible to human eyes, although generally there are enough other components in these light sources so you can see them quite clearly. Some wavelengths of UV are quite dangerous - both to the eyes and to skin and are even used to destroy bacteria, doing so by causing damage to their DNA.

This first UV source is an EPROM erasor. The strongest line at 250 nm is invisible, but the lines at around 435, 545 and 580 are easily seen.

spectrum of EPROM erasor

This next UV source is a mercury vapor in a quartz lamp (normal glass blocks UV). It has the same spectral characteristics as the EPROM erasor, leading me to believe the former is also a mercury vapor source. There is a higher ratio of visible light in this source however.

spectrum of intense UV quartz lamp.

This next UV source was a home-made rock light built in the fifties by my father, who ran a gold-mine with his brother back in the thirties (in the White Mts bordering Nevada and California) and maintained an avid life-long interest in geology. It was made of two dark purplish fluorescent bulbs mounted with a couple pieces of angle-aluminum and a transformer ballast with scary-looking wires hanging out. It has an interesting spectrum, predominantly between about 345-365 nm. Dried dog and cat urine on carpets fluoresces brightly under this light. The light itself appears purple due to the strong line at 405 nm.

spectrum of old long-wave UV rock light.

Here is a zoom of the previous spectrum (immediate above).

zoom of spectrum of long-wave UV rock light.

Usually when you buy a UV flashlight you get one that produces a spectrum with one peak at 390 nm. It is apparently cheap to manufacture these LED's. Unfortunately, dog and cat urine and most minerals do not fluorese at this frequency, or not sufficiently to be very useful. There are probably other uses I am not aware of.

spectrum of 390nm UV LED.

True 365 nm LED's are apparently quite expensive and due to the above mentioned uses they are in demand. Which leads the unscrupluous and/or ignorant to sell LED flashlights advertised as 365nm which are actually producing longer wavelengths. Here is one such flashlight which, as you can see, actually has a peak at 374 nm, not 365 as advertised (to make matters worse, he wouldn't take it back!). If you click on the spectra below you can see an enlargement.

spectrum of supposed 365nm UV flashlight.