An Approach to Photographing Fluorescent Specimens with a Digital Camera

If you want to skip the technical details, click here to jump to the photo section.

Issues

While digital cameras are not overly sensitive to UV, they can certainly pick up some of it, and the results make it difficult to photograph fluorescent specimens with them. When I tried to take a photograph of the green fluorescence of a Calcium Tungstate crystal glaze that I had activated with Erbium Oxide, instead of getting a green spot where the UV from my ultraviolet LED was hitting the tile, I got a bright pink spot.

I was not surprised to find that the blue pixels in my camera were picking up a little of the near-UV, but it came as something of a shock to discover that the red pixels were sensitive around 355 or 360 nm: when I took the same photo through a prism, following a suggestion made by Nancy Lebovitz, I got a green spot, and also blue and red spots that overlapped —

(Notice that the red is actually on the short-wavelength side of the blue. My guess is that either the red filters in the sensor are fluorescent, or that the blue and green filters cut off these particular UV wavelengths but the red filter, for whatever reason, passes them, and the sensor then picks them up.)

Moreover, even filtered Mercury sources, which are extremely common in both shortwave and longwave lamps, emit copiously at the violet edge of the visible spectrum, and cast a “purple haze” (apologies to the late J. Hendrix) across the image. It is necessary to eliminate as much of this as possible, either from the the output of the lamp, or from the input to the camera. (The responsible Hg lines are at 404.66 and 407.78 nm, and possibly the very strong line at 435.8 nm. Hg has other lines in the visible, but they are at longer wavelengths, and are removed by the filter on the UV lamp.)

Solution

I tried various colored glass filters, all of which failed more or less miserably, and I eventually decided to use a sharp-cutoff longpass filter. I was able to find one with its cutoff centered at 420 nm and with 3% slope, which we purchased from Thin Film Imaging Technologies. I chose the TFITech filter because of its particularly steep slope: most other filters I checked had 6% slope, which would have eliminated more of the indigo and violet than I was comfortable with.

The filter comes mounted in a black-anodized aluminum ring (unless you specify otherwise), and I simply attached it to the front of a regular commercial UV filter (which I use to protect the lens of the camera) with three small dabs of silicone rubber aquarium adhesive. There was some chance that the 48-mm clear aperture wouldn’t be quite wide enough, but I haven’t noticed any vignetting problems.

Here is the measured performance curve of the filter:

Here are two more photos, showing the same test tile illuminated by a UVL-56 handheld lamp, also from UV Products, which almost certainly emits a spectrum that is slightly different from that of the BL 100. It is also a lower-power device, so these are not as bright, and they required a significantly longer exposure. Again, the photo on the left was taken with the filter in place; the one on the right was taken without it.

Under short-wave illumination only a little of the green shows up, and the color is primarily due to the fluorescence of the Scheelite crystals in the glaze. To show you what that looks like, here is a plain Scheelite glaze, without any additions, under short-wave UV:

Here is a different CaWO4 glaze, this one with 10% of Ruby Dust added to it. As you can see, the fluorescence is considerably different with different excitation wavelengths. (I am not entirely sure why the glaze is gray in roomlight, but it appears to be a feature of this particular batch of Ruby: it’s the same color when I add it to my regular clear glaze.) As usual, from left to right we have the test tile under roomlight, black light, and short-wave UV.

Here’s an overview in ordinary light, and a view of the fluorescence under black light. Note that the tiles in the top two rows were fired in reduction, and the ones in the bottom row were fired in oxidation.

Next photo: Samarium (fired in oxidation) on the left, Dysprosium (fired in reduction) on the right. Second photo: Terbium (fired in reduction) on the left, “Pink Snow” (recipe in the article) on the right. While the Erbium fluorescence in “Pink Snow” is not as bright as the typical Terbium fluorescence, it is a very nice clear green.

(Demontration of Quenching.) First photo, daylight; second photo, black light. The tile on the left has 1% Eu2O3 added to it; the tile on the right has 3%, and is, somewhat surprisingly, not as bright. It also shows slightly more greenish color in daylight. (Please note: “.19c.” is the largest available size of the daylight photo.)

Europium and Cerium: color difference. (1% Eu on the left, 3% Ce on the right.)

Terbium and Willemite: color difference. (3% Tb in a clear glaze on the left, 2% Mn [as carbonate] in a crystal glaze on the right.) My apologies for the lint on the crystal tile.

Ruby Dust in a clear glaze on the left, 3% Eu2O3 on the right. Both tiles were fired in oxidation.

File this one under “Rare-Earth Fluorescences in Everyday Life” — this is a 500-ml jelly jar, under black light. In ordinary light it’s just a clear jar, but there’s some Cerium in it (probably intended as a decolorizer), so it fluoresces:

On the left in both photos is Europium, on the right Samarium. The photo on the left shows them under black light, and the photo on the right shows them under short-wave UV.

Note that some of the Eu seems to have dissociated, even in an oxidized firing, so that under black light the tile on the left is more or less magenta or “hot pink”, a mixture of red and blue emissions. Eu(II), which is responsible for the blue part of the color, doesn’t respond much to short-wave excitation, so only the orangy-red of Eu(III) shows up in the photo on the right. (Sm(III) doesn’t appear to be particularly responsive to short-wave UV either, and that tile is quite dim.)

Here is a china paint test, in which I added Thulium Oxide in approximately half-percent increments to a base that consists of equal parts Ferro Frit 3195, Gillespie Borate, Lithium Carbonate, and Silica. There is a stripe of the plain uncolored base down the left edge of the tile. It is clear that a percent or so of Tm is about all that is required for reasonable fluorescence, which is a good thing: Thulium Oxide retails for about $10 a gram in small quantities.

The peculiar shape in the bottom right corner is a self-supportingΔ 016, slightly overfired, being viewed from above — I glued the cone to the tile with a dab of china paint and fired the tile flat on its back, so that I’d have a record of the firing conditions.

Note the fluorescence of the pre-existing glaze on the commercial tile I used as a substrate. If it were important to eliminate it, I’d probably make up a nonfluorescent slip and fire it onto the tile before applying the china paint; but for my purpose here, an informal test of Tm, it was not an issue.

It does, however, illustrate the fact that fluorescence can be used in some cases as a diagnostic or analytical tool. Tiles made by different factories have slightly different fluorescences; if a tile is already mounted and the back cannot be examined, or if only a chip is available, the fluorescence could conceivably help identify the source.

Here is another china paint test, this one a demonstration of “secret writing”. (Clockwise from the top: Sm plus Eu; Tb; Tm. I used about 1% of each colorant, so there is roughly 2% colorant in the orange vertical shape and about 1% in each of the others. The paint is the same “1111” base as in the test tile shown above. This tile, like the other, was fired (several times) to Δ 016.) It isn’t quite right yet — the fluorescent regions are faintly visible in roomlight — but it will serve as a demonstration of the general principle.

(Note, added later)

I have also managed to use “Glow Powders” in china paint base, which results in phosphorescent or “Glow-in-the-Dark” paint. This is easier to deal with than the cone 08 glazes that were described in Ceramics Monthly a while ago, but those were intended as actual glazes, not as enamels or paints. Depending on the color, you may be able to use any of several types of paint base; I suspect that the less reactive the base is, the better the resulting brightness will be — you do not want to dissolve the phosphorescent material.

Here’s an odd one. As you can see in the first image, this is a rather ordinary Homer Laughlin saucer, in their “Virginia Rose” pattern. I do not yet know why it is fluorescent; the color resembles that of the Willemite fluorescence, but it is less pronounced under short-wave UV than I’d have expected, and I don’t actually observe any crystals — the glaze, though crazed, is entirely vitreous. It also resembles the fluorescence of glasses that contain Uranium Oxide, but the roomlight color is not as distinct a yellow as I’d expect. (When I get a chance I’ll take a counter to it and see whether it emits any alpha. If it is Uranium, there isn’t very much of it present here, and the activity will be minimal; still, it may be detectable.)

[NOTE, added 2008 June 25: I just put this saucer in front of an alpha/beta counter, and although it does emit, it is only about 10X the background level that I see here; this strongly suggests Uranium, which is not very active.]

Again, this points toward the use of fluorescence as an analytical technique: without the use of UV it is not at all obvious that there is anything unusual about the glaze on this saucer, and there is no reason to think about taking a counter to it.

As usual, roomlight, long-wave, short-wave. I had to use a 4-second exposure for the short-wave image, also fairly typical: that lamp is of relatively low power.

Finally, here’s an ordinary Δ 9-10 crackle clear, first under roomlight and then under short-wave UV. Many (if not most) ceramics have at least some fluorescence, and the color is often different under different wavelengths. This particular glaze is more interesting under short-wave than under black light.

Next, a difficult mineral: some regions of this specimen fluoresce in the deep blue, and it would probably be better if I could remove all of the visible emission from the lamp instead of filtering the camera, because the filter on the camera pushes the color slightly to longer wavelengths. Zircon and Feldspar, from Wisconsin; roomlight on the left, long-wave UV on the right.

Here’s a different face of the same specimen, first under long-wave illumination, then short-wave. Please ignore the background, which is visible in both, but is particularly evident in the shortwave image.

This is a specimen from Franklin, NJ. It appears to contain Willemite and Calcite; there may be small amounts of other fluorescent minerals as well. Photos, from left to right: roomlight, long-wave, short-wave. (The short-wave image was taken earlier than the other two, and shows the piece in a different orientation.)

I believe that this is Calcite. Photos, from left to right: roomlight, long-wave. (I saw very little fluorescence with the short-wave lamp, and it looked essentially the same as the long-wave fluorescence.)

Here is a conglomerate (I think that’s the correct term) with little fossils in it. My apologies for the dust, which is mostly visible in the black light photo.

This is a “Mini Orbit Ball”, purchased at GrandRabbit’s Toy Shoppe, which is in Boulder, Colorado. As usual, roomlight on the left and long-wave on the right. (I didn’t see much under short-wave.)

Conclusion

Even with the one caveat about violet and deep blue fluorescences this is vastly better than I have ever been able to do before, with film or digital camera, and I am extremely pleased. The use of the TFITech 420LP filter is clearly a viable method for photographing fluorescent specimens under either short- or long-wave excitation, provided that the source is adequately filtered to reduce its visible output to a minimum. This filter costs $165 at 50 mm diameter, and although that appears to be somewhat expensive, the performance cannot be denied. Moreover, it will be difficult to duplicate with any other filter that I’ve been able to find.

I did look into the Baader UV/IR blocking filter, which is also available in 50 mm diameter; but I note (see the upper curve on the Baader page) that it passes some UV near 355 nm, exactly where I expect to have trouble with the red pixels in my camera. This filter is less expensive than the TFITech filter, but the fact that it requires another filter to compensate for the UV passband increases both the complexity and the expense. Nonetheless, it may be a viable alternative if you can find the other filter at a reasonable price, or if your camera is not sensitive to 355-nm UV.

If your funding is unlimited, you may want to look into the possibility of using a large shortpass filter with cutoff at 380 or 385 nm on your lamp. This would probably allow you to use a longpass filter with cutoff at 395 or 400 nm on the camera, and thus permit you to record fluorescences essentially all the way to the edge of the visible range. Unfortunately, such a shortpass filter would be a special-order item, would probably need to be quite large (depending on the size of your lamp), and is certain to command a very high price.

Contact Information:

My email address is a@b.com, where a is jon and b=joss.

Phone: +1 240 604 4495.

Last modified: Thu Jun 13 23:54:46 EDT 2013