Lamp spectra - first try

kw: analysis, spectroscopy, lighting

In the past few years we have tried several lower-wattage "bug lights" as an alternative to the yellow 40-watt incandescent bulbs we've used before in our porch light fixture. When the one we had 4 years ago burnt out we got a 13 watt, yellow compact fluorescent spiral lamp by Sylvania. Though it was not marketed as a bug light, it worked pretty well, though some insects came to it. The next year I saw a 6 watt LED bug light, marketed as such by Feit, so we got that. It worked about equally well. Then I went looking for something that might be a little bit better, and got a 3 watt amber bug light, also by Feit. It doesn't draw insects, but it is pretty dim.

I decided to find out whether a little blue light is getting out of these lamps, so I made a crude spectroscope from a piece of diffraction grating and a short length of PVC pipe plus some odds and ends. In this photo it is on a tripod aimed at a test lamp. I aim a camera with a telephoto lens at the black aperture at the left, where the spectrum emerges.

I cut the end of the PVC for the grating at an angle so the spectrum would exit at right angles to the grating. It has the added benefit that, for visual use, looking the "back way" yields a spectrum about twice as wide. But the focal plane is strongly tilted, making it a poor choice for photography (though I tried!). The instrument has a number of shortcomings, but I think I know how to produce a better next version. For one thing, I'll use a different exit angle, so the diffraction grating doesn't reflect the camera and photographer! (see below)

I photographed the spectrum of nine lamps, the three test bug lights and several others either for spectrum reference or to see the spectral coverage of both incandescent and non-incandescent lamps. Eight of the lamps are shown here, and their spectra are tagged in the next image, followed by some explanation.

These are in the order listed in the spectra image.

The first three spectra are for reference. The 4000K (cool white) CFL shows a combination of spectral lines for mercury (Hg) and for the phosphors used to "whiten" the harsh blue-green light of raw Hg lamps. Mercury has a strong green spectral line at 546 nm, as seen in both this lamp and the 13W yellow CFL (A nearby strong green line is from a phosphor) and a strong blue-violet line at 405 nm, which excites some of the fluorescence, but a stronger near-UV line at 365 nm does most of that. The strong red-orange line at or near 615 nm is from a phosphor, as are the yellow-orange-green and green-blue-violet bands. The 40W incandescent lamp shows the smooth spectrum characteristic of a thermal source. The near-lack of yellow in this spectrum is because a camera's sensor sees colors differently from our eyes, but this is only evident when photographing spectra! The 60W "Reveal" lamp has a filter that cuts out most of the yellow and yellow-orange, making the light appear bluer and closer to daylight.

The next three spectra are for the bug lights. The 13W yellow CFL has the same spectrum as the white CFL from green through red, but with extra yellow and orange, and the green-blue-violet phosphor is left out. Also, a filter removes the blue and violet lines of Hg. The two LED's have nearly identical spectra. The blue-violet light from the fluorescence-exciting blue LED is filtered out, leaving only light from the broad band phosphors. The 3W lamp has a little more red-orange than the 6W lamp, and this is visible when they are lit side-by-side; the 3W lamp's color is amber. In the photo of the lamps above, the filter is inside the 3W lamp's envelope, which is white. For these three spectra, the brownish features seen below the green band are reflections of either me or the camera off the diffraction grating film.

The 8.5W LED is the kind of "warm white" bulb we have begun to use around the house. It has a spectrum very similar to incandescent; it just has a dip in the mid-blue range, and a bright band in the blue-violet range, which is from the fluorescence-exciting LED. The UV CFL is a "black light", very similar to old black light fluorescent tubes used at parties, but in spiral form. Most of the visible light is filtered out. The green and violet lines at 546 and 405 nm are a little visible anyway, and the camera is barely able to record the 365 nm line that does all the work of making fluorescent things glow. I am puzzled by the line in between, at about 385 nm. I don't know what it could be from. However, I know that these lamps use a phosphor that responds to a strong Hg line at 254 nm and converts it to longer-wave UV, to get more "black light". Perhaps it is the source of the 385 nm line and other faint features in that space, but I think it mainly adds more 365 nm light.

Finally, the 40W fluorescent tube is of the kind that has been in use for nearly my whole life (7 decades), now mostly supplanted by CFL's and LED's. The two lines of Hg in blue-violet and green come through, but broad-band phosphors fill out the light making these pretty good for most uses. They actually have better color rendering values than CFL's, at the cost of using nearly twice the power: a 40W "tube" and a 23W CFL both emit about 1,600 lumens, but strongly colored items may look a little odd with the CFL.

As crude as it is, this simple spectroscope helped me understand these lamps better. I think the reason that some insects still come to the three non-incandescent bug lights is that they can see the green light. I don't have an incandescent bug light, but I suspect it to have less green light than the CFL or the LED's. This has been an instructive exercise.