Filters for UV photography

Digital photography in reflected UV light (or, more properly, reflected UV radiation, since by definition UV is outside the range of visible light) requires a combination of unusual equipment, consisting mainly of a UV or multispectral light source, a UV pass filter, a lens that transmits sufficient amounts of UV radiation and a UV-sensitive camera. None of this equipment is trivial to obtain.

In the past, I discussed a few UV-pass filters on my site. However, the contents of these earlier pages may not be completely up to date. These pages on UV photography and the required equipment include, for instance, the following:

I recently obtained interesting results with a few filter types that I had not used in the past. These findings may be of interest to others, especially since a few of them may provide a cheaper pathway into the traditionally expensive field of reflected UV photography. There aren't too many manufacturers of UV-pass filters usable in UV photography and imaging, and therefore several of the filters discussed on this page have been mentioned or tested on other web sites and bulletin boards, including fotozones and ultravioletphotography. However, on this page I discuss my findings only, which may be partly different from the experiences of other photographers with the same filter types.

On this page, I discuss a few types of UV filters that I found useful in reflected UV photography with digital cameras. In the past 10-15 years, the subject of UV photography and filters for reflected UV photography with digital cameras has been discussed on several Internet sources. Some of the earlier sources are at present mainly useful to understand how ideas and techniques have developed among UV photographers, while more recent sources tend to contain updated information. In approximate chronological order, see for instance:

An additional reason why I added the publication year (when available) of the information at the above links is that at least some of these filters have been upgraded with respect to their IR rejection, and the published information may therefore not reflect the performance of currently produced specimens. The above list of links is not meant to be complete, and is only provided as a collection of examples.

I am aware of a couple of Internet sources who are using "fantasy names" for a few types of UV-pass filters and lenses, perhaps as an attempt to conceal their actual commercial sources and present themselves as the only distributors of these items. On this page I am only discussing filters that are available from well established (and often multiple) commercial sources, to make it possible for readers to shop around and find the best prices.

The scientific literature contains large numbers of papers dealing with UV imaging and its applications, starting from the early 20th century. A discussion and summary of the main references in this and related fields is available, for instance, in my book Digital photography for science.

Unless otherwise stated, test images on this page were recorded with a Bowens 1500Pro studio flash modified for UV photography, Optomax 35 mm f/3.5 lens at f/8 and multispectral Panasonic G3 at base ISO. The color balance was kept unchanged in all test images. Exposure was adjusted to obtain roughly comparable images with different filters by changing the flash power alone.

Reading transmission diagrams

The transmission diagrams of filters, and especially linear diagrams graduated in % of transmission, must be taken with a grain of salt. An 80% transmission peak, for instance, looks great on a diagram, and much better than a 60% transmission peak. However, this advantage can be completely offset by a narrower transmission band. A filter with a lower transmission peak and a wider transmission band can often transmit more UV than a filter with a higher peak and a narrower band.

The aperture scale almost universally used today to indicate the lens aperture is not linear, but geometric. The reason is that our perception of luminosity is non-linear. Halving the illumination intensity (which is equivalent, for instance, to decreasing the transmission of the whole system from 80% to 40%) is visually perceived as a small change in illumination intensity. This decrease is also equivalent to a difference of one stop in lens aperture. A one-stop difference in exposure is perceivable when images are compared side-by-side, but not dramatic. With most subjects, it is a simple matter to correct an over- or underexposure of one stop in post-processing. One can state that a one-stop difference in the transmission of a UV-pass filter is not a big deal. The difference between an 80% and a 60% filter transmission is less than half a stop. This is hardly noticeable in recorded images, and easily corrected in post-processing with no visible losses in image quality, except possibly in extremely over- or underexposed areas.

With modern digital cameras, high ISO sensitivities are entirely usable, and UV-enabled electronic flash provides more than enough UV radiation for most uses. In fact, I routinely use without problems some UV bandpass filters with a transmission peak of only 25% and a narrow transmission band of 10 nm. A 25% transmission looks puny on a linear transmission diagram, but differs from an 80% transmission by only about a stop and a half. Bandwidth often has a proportionally higher effect on exposure than peak transmission and, unless you are interested in one specific and narrow band, with a broadband UV source like sunlight or electronic flash it does make sense to use a filter with a broad transmission band.

The UV contents of sunlight and, to a lesser degree, unfiltered electronic flash decrease rapidly toward shorter wavelengths. In addition, the sensitivity of the solid-state camera sensors used in general-purpose digital cameras also decreases rapidly toward shorter wavelengths in the UV range. The two effects combine together, and normally require an increased illumination intensity or exposure time at shorter wavelengths. Thus, filters with a similar height of their transmission peaks and a similar bandwidth, but with transmission peaks located at different wavelengths, frequently require a different exposure.

The bandwidth of UV-pass filters is generally specified as the width of the portion of transmission curve that lies at or above half the peak transmission. For instance, in a filter that exhibits a 50% transmission peak, the bandwidth is specified as the difference in nm between the longest and the shortest wavelengths that display a 25% transmission. This definition implies that wavelengths at either extremity of the bandwidth are attenuated by one stop with respect to the transmission peak. In practical photography, however, even wavelengths attenuated by two stops or more do significantly contribute to an image, especially if recorded by the camera as a different color than the peak wavelength. Therefore, in practice the usable bandwidth of a filter is a bit broader than the nominal one. It should not be regarded as surprising, for instance, that a filter with a transmission peak at 340 nm and a bandwidth of 10 nm (i.e., a nominal transmission band between 335 and 345 nm) still transmits detectable amounts of radiation at 360 nm.

An interesting discussion of the effects of interference coatings on filter transmission, and of the involved trade offs, is available here.

Traditional ionic filters

Traditionally, UV film photography has been carried out with UV-pass filters consisting of one of several types of ionic glass. This includes, for instance, the Hoya U325C, U330, U340, U350 and U360. The number following the U in the part name is the peak transmission wavelength in nm. Filters made with these glass types are available, for instance, from Edmund Optics. My experience with the U360 is described here. The example in Figure 1 was shot with the U360, and contains mostly near IR (NIR) information and just a little near UV (NUV). The little UV recorded in the image is sufficient to shift the prevalent recorded color from pink (typical or shorter NIR wavelengths) to orange-yellow. The U350 performed in a similar way in my tests. Schott markets a variety of roughly equivalent glass types, including UG1, UG5 and UG11. The B+W 403 is equivalent to the Schott UG1. Pentax marketed one of these filters as 363Mu, and Nikon as FF.

Figure 1. Example with U360.

It was well known by the early digital UV photographers that these traditional filters have massive transmission peaks in the NIR, which was not a problem in film photography but makes them useless with digital cameras. Digital cameras are so sensitive to NIR that the NIR radiation leaking through these filters overwhelms any UV information contained in the recorded images. In fact, early pioneers of UV digital photography like Bjørn Rørslett (see above links) used to stack an NIR-cut filter with these NUV filters to solve this very problem. Stacking two or more filters, however, lowers contrast and increases the likelihood of flare and internal reflections within the filter stack.

Today, these traditional filters are essentially useless for UV photography. There are much better alternatives, some of them discussed below. A few photographers use these UV filters for IR landscape photography, where the UV component contributes to a slightly different image character than a pure NIR-pass filter (see also here). The traditional UV-pass filters are obsolete also in microscopy, for similar reasons. The makers of these filters are well aware of this problem, and some of them offer coated alternatives with better NIR rejection. For instance, Schott currently provides a NIR-blocked version of its old UG11 filter, identified as DUG11. NIR blocking by this filter, however, is not as good as in filter types discussed below.

Well-known filters for digital UV photography

Astronomers and forensic technicians were probably the first to use digital cameras to record images in the NUV. A few filters were developed for astronomy, including the Schuler UV (actually Schüler in the original German version) and the currently popular Baader U. Both are discussed below. The clouds in the higher atmosphere of Venus have been known for a long time to display a clear banding structure in the NUV. For this reason, UV-pass filters originally developed for astronomy, like the Baader U, are often called Venus filters.

Foremost among the early types was the Schuler UV, made by Astrodon and now discontinued. It was more affordable than the Baader U described below, and provides a transmission peak around 370-380 nm and a relatively broad transmission band. Images taken with this filter (Figure 2) tend to have a prevalent purple/violet tone, which is characteristic of radiation prevalently concentrated around 380-390 nm as recorded by CMOS and CCD sensors of ordinary digital cameras after removing the built-in UV, IR and anti-aliasing camera filter. In my experience, the 2" Schuler UV has no problems caused by NIR leakage. It consists of multiple cemented layers of different types of glass (according to this Astrodon document, 1 mm UG1, 1 mm BG39, and a 2 mm WG295 substrate just to increase the thickness).

The Baader U seems to be the filter most broadly used today in UV photography. It is commonly available online through several amateur astronomy shops, and is occasionally sold on eBay. It provides a transmission peak around 350-360 nm (there is some minor disagreement about the peak wavelength among Internet sources) and attenuates the 380-390 nm band to a higher extent than the Schuler UV. According to some sources, it is based on a Schott UG11 substrate with added NIR-cutting and anti-reflection coatings. It is available in 1 1/4" and 2" sizes mounted in metal cells with threaded attachments. Early 1 1/4" specimens displayed a troublesome NIR leak, subsequently reduced by increasing the number of NIR-blocking interference layers. The Baader U can be purchased from dozens of online sources, for instance directly from Baader or from

The transmission peak of the Baader U is slightly higher than the Schuler UV (about 80% versus 65%). However, the rapidly decreasing sensitivity of solid-state sensors at shorter wavelengths and the rapidly decreasing amount of solar UV radiation at shorter wavelengths introduce a bias that often cancels out the perceived advantages of the Baader U and may result in a slightly shorter exposure with the Schuler UV than with the Baader U, especially with sunlit subjects. In addition, in practical photography and with all other things being the same, 65% and 80% transmissions only differ by about half a stop, which is hardly noticeable and easy to compensate in post-processing.

Figure 2. Example with Schuler UV.
Figure 3. Example with Baader U.

The Baader U produces grayish or reddish tones (Figure 3) characteristic of 370 nm, and with some subjects even yellow tones characteristic of 360 nm.

Figure 4. Old or mislabeled 1 1/4" Schuler UV.

A few years ago, I decided to get a smaller and handier version of the 2" Schuler UV, and purchased a 1 1/4" Schuler UV (Figure 4, shown mounted in a step-down filter adapter). This smaller filter displays a massive leak in the near IR and even in the deep-red portion of the visible range, and is completely unusable for digital UV photography. In fact, the results (Figure 5) are quite similar to those obtained with a NIR-pass filter (Figure 6).

This 1 1/4" filter is obviously made of different glass types than the 2" Schuler UV in my possession. For instance, the NIR-cut layer is colorless in my 1 1/4", but green/cyan in my 2" as seen from the filter shoulder. It is possible that this filter is a factory-mislabeled specimen, a filter of a different type that was inadvertently placed in the wrong box in the shop, or an earlier type of Schuler UV developed for film photography.

Figure 5. Example with (old or mislabeled) 1 1/4" Schuler UV.
Figure 6. Example with 780 nm NIR-pass interference filter.

Astrodon is currently marketing an alternative to the Schuler UV, called Astrodon UVenus, in mounted 1.25" and unmounted round 50 mm sizes. A 50 mm unmounted filter can usually be mounted in the empty frame of a 52 mm camera filter (especially the B&W filter frames, which possess a wider internal flange that better covers the edge of slightly too small filters). The UVenus specifications guarantee a minimum 90% transmission between roughly 320 and 380 nm, very sharp edges of the transmission band and a complete NIR blockage. It is specifically designed for CCD sensors and uses interference coatings on a transparent substrate. Based on these specifications, the Astrodon UVenus appears to be better than both the Schuler UV and the Baader U. In some shops, it may be cheaper than the Baader U. It is frequently used by astronomers, and their posted opinions on this filter are positive.

Figure 7. Example with 1 1/4" Astrodon UVenus.
Figure 8. Example with Astrodon UVenus and B+W 486.

Images with the Astrodon UVenus are slightly more blue than those obtained with the Baader U, but not very different (Figure 7). With some subjects, yellow tones are stronger than with the Baader U. Most of the image information is clearly recorded in the NUV, but some subjects look more faithful to their usual rendering in visible light. This is explained by the fact that the Astrodon UVenus transmits a small but easily detectable amount of the shortest wavelengths of visible light, albeit no NIR. Placed on one eye, it clearly shows strong light sources in transparency in a grayish-violet color. With the addition of a B+W 486 filter, which cuts both UV and NIR (and a bit of the extreme visible spectrum at both its ends), images clearly show a moderate amount of "royal blue" light (Figure 8, taken with same exposure as Figure 7).

Baader U.
Schuler UV.
Astrodon UVenus.
Visible light.
Figure 9. Comparison of the three UV filters discussed in this section and visible light.

The UVenus may find specialized uses, especially when a modest amount of added visible light makes UV subjects appear more "natural" to the untrained eye, but requires an adequate understanding of its properties (see Figure 9 for a different subject imaged with these three filters). The higher transmission of the UVenus filter makes hand-held UV landscape photography slightly easier, and sometimes produces results uniquely different from the other UV-pass filters. The UVenus is probably less suitable than the Baader U for recording separate information channels in the visible and UV that are meant to be reassembled into a false-color composite.

Less well known current filters

UV photography is not the main use of UV-pass filters. These filters are routinely used in a variety of scientific, technological and military devices, including microscopes, spectroscopes and remote sensing imagers and analyzers. Since most solid-state UV sensors are similar in construction to the photo-sites of CMOS and CCD camera sensors and are likewise sensitive to NIR, it is not surprising that all these devices need UV-pass filters with minimal NIR transmission. Therefore, it pays off to browse the catalogs of optical filters for scientific and technological uses. It turns out that the Schuler UV, Baader U and Astrodon UVenus filters are by no means unique. Instead, numerous sources provide similar filters, usually consisting of a combination of ionic glass and interference coatings.

While the coatings of the Baader U filter are external and unprotected against damage, some of the currently available filters protect the interference coatings by sandwiching them among cemented layers of optical glass. The design choice between these two alternatives yields a reduction in production costs and a higher resistance to ghosting and flare for the single-layer design, versus a higher durability and the possibility of combining the properties of multiple glass types for the multi-layer design.

In many cases, an excellent NIR rejection is obtained by combining dielectric coatings, a traditional type of UV-pass ionic glass and one or more layers of different types of ionic NIR-cut glass. Thin outermost claddings of sapphire glass, fused silica glass or other UV-transparent materials are sometimes used to protect the softer filter glass from mechanical damage and, in some cases, to protect special glass types from oxidation and corrosion by air or humidity.

During my search, three classes of filters turned out to be of special interest:

The transmission peak is 30% for the 390 nm model and 25% for all others. All are blocked to 2,000 nm, which is more than enough to completely prevent NIR leakage with CMOS and CCD sensors. There is also an FL355-10 which is blocked to 1,150 nm, which may or may not be enough to avoid problems with NIR contamination. These are multilayer filters in metal cells with a 25 mm outer diameter and a clear (optically usable) diameter of about 23 mm. I discuss below whether, and how, this diameter is sufficient for UV photography. The narrow bandwidth naturally requires a higher illumination intensity when used with a broadband source, especially near the short end of the wavelength range of these filters. The difference may be as much as 4-5 stops, compared to the Baader U and Schuler UV filters. Nonetheless, it is entirely possible to carry out close-up and macrophotography with the radiation source used for these tests.

Asahi Spectra markets comparable 25 mm bandpass filters for similar and shorter wavelengths. For digital UV photography, the 300 nm, 310 nm (part no. XBPA310), 320 nm and 330 nm types with close to 60% peak transmission and bandwidths of 10 nm could be interesting. Asahi Spectra also markets more expensive filter types with a higher peak transmission, for instance a 310 nm (part no. XHQA310) model with close to 80% peak transmission. However, the peak transmission of 80% in these filters is obtained at the expense of a relatively high leakage in the NIR, which makes them very likely unsuitable for NUV photography without an additional NIR-cut filter.

Additionally, Asahi Spectra offers bandpass filters for much shorter wavelengths, which most likely cannot be recorded by ordinary camera sensors and therefore are not usable in digital UV photography.

So far, I did not test any filters with peak wavelengths shorter than 320 nm. Eventually I may do so, gradually moving toward shorter wavelengths one filter at a time, and at some point (probably between 300 and 315 nm) I should meet the limit beyond which digital UV photography with current equipment becomes impossible or impractical. I know from other digital UV photographers that filters with peak wavelengths as low as 310 nm are still usable when plenty of UV radiation is available. Of course, an important question is whether imaging at these wavelengths will show any interesting information, which is not necessarily the case because solar radiation at these wavelengths is much weaker than at longer wavelengths, and may lie outside the vision range of most organisms.

Figure 10. Example with Asahi Spectra XRR0340.
Figure 11. Another example with Asahi Spectra XRR0340.

This filter is based on a Hoya U340 substrate with multiple visible- and NIR-rejection dielectric coatings on one side, apparently anti-reflection coatings on the opposite side, and is blocked to 2,000 nm. An example taken with this filter is shown in Figure 10. The prevalently yellow color is typical of the rendering of radiation around 360 nm. It must be remembered that the sensitivity of the camera sensor decreases at shorter wavelengths, so the 340 nm and shorter wavelengths, even if present, contribute less than higher wavelengths to the recorded image.

It is available in a circular 25 mm unmounted size (part no. XRR0340) with a 23 mm usable diameter or a square 2" unmounted size (part no. ZRR0340).

25 mm unmounted filters are easily mounted in the frame of a cheap 28 mm UV or protector filter, where they fit perfectly. This holder from Edmund Optics can be used to mount 2" square filter on the threaded filter attachment of a camera lens.

The Asahi Spectra XRR0340 produces remarkable high-contrast and high-resolution images. Yellow tones are usually prevalent, but with some subjects a dark, lifeless blue color is produced. The left half of Figure 11 shows an example (visible mostly in the central areas of the petals). In the right half of Figure 11, the blue channel has been removed in post-processing to show the difference.

This blue color is different from the violet-blue typical of 370-390 nm (which normally is not transmitted by the present filter), and seems to be the typical way in which wavelengths around 300-305 nm are imaged by solid-state Bayer sensors.

Unlike the higher wavelengths at the upper end of the NUV range, this color is only recorded in the blue channel (rather than both blue and red). It is therefore possible to isolate this band by separating the blue channel from the rest of the image. Photography at these UV-B wavelengths is further discussed on this page.

This filter also allows the separation of a green channel, which records information prevalently in the 320-340 nm band. Thus, with the XRR0340 it is feasible to separate two adjacent NUV bands from a single color image, a feat that is not practical with other NUV-pass filters that transmit higher wavelengths (which are recorded in both blue and red channels).

Ocean Thin Films' 335nm FWHM 80nm, among other types, is roughly comparable in specifications to the Asahi Spectra XRR0340, but its transmission band is more sharply delimited than the XRR0340. However, its NIR leak is very high, and may be located within the sensitivity range of some camera sensors. I did not test any of these Ocean Thin Film filters.

Some time ago, I used a number of small UV bandpass filters, probably made for microscopy and/or photometry, in the construction of a filter strip I made to test lenses for suitability in UV photography. These filters are quite similar in construction to the Thorlabs filters discussed above, except for their diameter. The ones I used are between 10 and 15 mm in diameter. Some of them are designed for photometric use and show uneven surfaces and a blotchy appearance, while others are proper optical filters. Only the latter can be used as UV-pass filters for digital imaging. Among these filters, the following type attracted my attention:

This filter is only 12.7 mm in diameter and has a clear diameter of barely 10 mm, but its specifications sounded very interesting, not the least in view of its 19 $ price (between one tenth and one twentieth of typical 2" filters mentioned above). In practice, it transmits between 300 and 380 nm, which is somewhat different from other filters and therefore potentially useful as an alternative. It is also well NIR-blocked in my tests, although the transmission diagram shows approximately 10% NIR leakage around 1,100-1,200 nm. The 330WB80 is marketed by Omega Filters in a 25 mm diameter as part n. XB04. The 330WB80 is also identified as XF1001 and marketed for fluorescence microscopy by Omega Filters and Horiba in a variety of sizes.

The interesting question, of course, is whether a very small diameter of 12.7 mm makes it possible to use this filter in UV photography with anything larger than a microscope objective (see below). The same eBay seller offers several other, sometimes larger types of UV-pass optical filters of potential interest for UV photography.

Figure 12. Example with 325BP10 and CoastalOpt 60 mm.
Figure 13. Same example, 1:1 crop. The dark globules are at the tips of microscopic hairs. Some of the visible detail is at the cell level.
  • Updated 325BP10, eBay item n. 310295977900, sold in 25.4 mm metal cells by eBay member bjomejag, with peak wavelength of 325 nm and a very narrow and rectangular transmission band of 10 nm. Peak transmission is relatively low at around 40-45%. Although the peak wavelength is similar to the 330BW80 discussed above, its results are very different. All UV wavelengths transmitted by this filter are imaged as "green" by Bayer sensors, and therefore the images are visually close to those produced by the Thorlabs FB340-10 discussed below. However, the 325BP10 gives a much higher contrast and resolution.

    The sensitivity of Bayer sensors is quite low at 325 nm, and this makes use of the 325BP10 difficult. I had to push sensitivity to ISO 1,600 or 3,200 and shoot with the Bowens 1500Pro at full power from just 20 cm away. None of the 35 mm lenses useful at longer UV wavelengths can realistically be used at 325 nm. Therefore, the above test images were recorded with the CoastalOpt 60 mm. The Asahi Spectra XRR0340 (see above) can help to record even shorter wavelengths with certain subjects, but requires substantial post-processing.

    A plus of this filter is that it has a dielectric coating on both surfaces, so it should be sealed against the atmospheric degradation common in UV-pass ionic glass types. However, it has a troublesome NIR leak that produces spurious red tones with some subjects.
Asahi Spectra XRR0340.
Figure 14. CoastalOpt 60 mm.
  • Updated 330WB60, eBay item n. 150805840293, sold in 25.4 mm unmounted glass by eBay member bjomejag, with center of a broad transmission region at 330 nm and a broad but rather shaply delimited transmission band of 60 nm. This filter gives results quite similar to, but not identical with, the Asahi Spectra XRR0340. It still renders a mostly yellow image, but a "colder" yellow than the Asahi Spectra filter.
  • Updated 330WB70, eBay item n. 310793801058, sold in 25.4 mm unmounted glass by eBay member bjomejag, For all practical purposes, almost identical to the preceding one, but with some subjects it does show a slightly higher amount of wavelengths near the high and low ends of its band.
  • It would have been interesting to test also the 330WB80 of the same size, which would be interesting. However, the seller shipped it to me without tracing information, and it was lost in the mail.
  • I also bought a thicker 330 nm filter similar to the 330WB70, with sputtered magnetron coatings on both sides. This remarkably expensive filter supposedly has a much better NIR rejection. Unfortunately, my specimen of the filter arrived with the coating on one face lifted off from the glass in a "bubble" covering a large portion of the filter surface, and image quality was extremely poor.
  • Updated PrimaLuceLab U is a UV-pass filter originally developed for astronomy. It is somewhat similar to the 330WB70, but not identical. It transmits a large amount of UV in a broad band that extends to both longer and shorter wavelengths than the 330WB70. This is an interesting, well-performing and reasonably priced filter, which I tested here.

Other UV-pass filters

It was not my goal to test and compare all the currently available UV-pass filters, especially since many of them are expensive and I have already found several good choices. I cannot say much of the filters I did not test, except that most of them are probably suitable for the stated use (although of course I cannot give any guarantees). The differences one is most likely to notice in practical use are caused by the filters' different spectral responses (i.e., peak wavelength and bandwidth). As discussed above, differences in transmission that reach or exceed one stop are noticeable from a direct comparison test with constant illumination, ISO sensitivity and aperture, but are often irrelevant when the exposure can be changed accordingly. Some difference in contrast and tendency to flare may also exist, but these differences can be compensated for by using appropriate illumination and lens hoods.

Commercial sources for some of the UV-pass filters I never tested can be found at the following links. Tests of some these filters on other web sites are mentioned above.

What filter sizes can be used, and how?

Small filters are significantly cheaper than larger ones. Especially at high magnification, a small filter often makes it easier to position light sources around the lens barrel. A small filter can often be mounted in a recessed position at the front of the lens, where it is more protected against accidental touching and stray illumination, or even internally between the lens and the camera sensor. This section examines these alternative placements and the feasibility of using smaller filter diameters than the currently "standard" and expensive 2".

There are no strict standards for the diameter of UV-pass filters, but some sizes and shapes are especially common. Among astronomy filters, 1 1/4" and 2" round filters mounted in metal cells with a threaded attachment are common. Filters for technical use, on the other hand, usually come in unmounted 2" square, unmounted 1" round, or mounted in cylindrical unthreaded metal cells, most commonly with external diameters of 1" or 25 mm, 12.5 or 12.7 mm and smaller.

The most common way to use UV-pass filters is by mounting them in front of the camera lens. Filters can be screwed to the filter mount of the lens, via a step-up or step down filter adapter if necessary. Many China-based eBay sellers offer well-made adapters of this type at unbeatable prices. Alternatively, one may use a hinged adapter like the one shown here. This saves time when inserting or removing the filter from the optical path, but requires a considerable free space in front of and around the lens, which may not be available in macrophotography.

Ordinarily, a filter suitable for mounting at the front of a lens must provide a clear filter area somewhat larger than the front element of the lens. Wideangle lenses typically require filters substantially larger than their front element. This guarantees that the filter will cause no vignetting with the lens aperture wide open. However, depending on the lens design and on the distance between front lens element and filter, when the aperture is stopped down it may be possible to use a smaller filter without vignetting. In particular, and especially with lenses of short focal lengths, the filter should be placed as close as possible to the front lens element. Using a camera with a small sensor, like 4/3 or Micro 4/3, also allows smaller filters to be used than APS-C or full-frame cameras, all other factors remaining the same. Working in the close-up and macrophotography ranges also allows a smaller filter diameter than focusing at infinity.

Figure 15. Thorlabs FB340-10 on lens.

Exactly how small a filter can be used without causing unacceptable vignetting at a given aperture requires some experimentation. For instance, with lenses of focal length equal to 35 mm known to perform well in UV photography, like several Kyoei, Petri Kuribayashi, Soligor, Optomax, Hanimex and other brands (but actually all made by Kyoei), I have been able to use Thorlabs 25 mm filters at f/8 and larger f/values without detectable vignetting on Micro 4/3 cameras. I discuss on this page how to use small filters and lens shades on lenses equipped with larger filter mounts.

The example in Figure 17 was taken with a Thorlabs FB340-10. The green color is the typical rendering of radiation at 320-340 nm. Focusing with this filter can be difficult, because continuous sources of these wavelengths, necessary for focusing with the filter mounted on the lens, are not easy to obtain. None of the fluorescent tubes I tested emits usable amounts of radiation at this wavelength, and even 340 nm LEDs do not emit much radiation at this wavelength. Focusing with a 360 nm filter and then switching filters does help to reduce focus shift.

My 25 mm filters are mounted in empty 28 mm metal filter rings (aluminum-colored in Figure 12) for easy screwing into step-down adapters. Mounting the filters in a recessed position within the front of the lens by screwing them in at the rear of the step-down adapter, as shown in Figure 12, helps to reduce vignetting, improves contrast and reduces the risk of accidentally touching the filter surface. It is also possible to use multiple, nested step-down adapters to mount the filters even closer to the front element of the lens.

Figure 16. 330WB80 sitting at rear of lens.

The very small 330WB80 is a different problem. Even mounting it in contact with the much larger front element of one of these 35 mm lenses would not prevent vignetting at any reasonable aperture. However, it can be noted at the left how the opening in the metal ring surrounding the rear element of this lens is barely smaller than the filter. In fact, the filter sits quite stable in this position when the lens points straight downward, mounted on a Micro 4/3 camera on a macro stand. It is of course possible to add here a friction or threaded mount for holding filters with the lens in any orientation.

Testing (Figure 18) shows no detectable vignetting at f/8, and both contrast and resolution are higher than with some of the other filters mounted in front of the lens. The prevalently yellowish color indicates that the recorded UV radiation in this case is mainly a mixture of wavelengths in the 360 nm band.

Mounting and dismounting a filter at the rear of the lens is of course time-consuming and requires the lens itself to be dismounted. The filter could be mounted in a wider adapter that fits within the rear cavity of the lens mount, making the operation easier. It is even possible to build a sliding filter strip capable of holding two or more filters and operated from the outside, without dismounting the lens. Housed in a short extension ring, such a filter strip could be used with different lenses.

Figure 17. Example with Thorlabs FB340-10.
Figure 18. Example with 12.7 mm Omega Filters 330WB80.

A 25 mm filter is usable even in front of lenses of higher focal lengths and with larger front and rear elements, like the CoastalOpt 60 mm, without vignetting on a Micro 4/3 camera. Rather than discounting the possibility of using small UV-pass filters on a given lens and camera, it may be worth testing the lens for vignetting in visible light with a clear or color filter of the same size (or even with an empty filter frame or step-down adapter), before purchasing the actual UV-pass filter.

These tests show that, with some lenses, not only it is possible to use smaller filters than the typical 2", but doing so is actually advantageous in several respects. Even some of the very small UV-pass filters may be put to good use in UV photography with certain lenses. In particular, the price for a combination of this cheap (19 $) 12.7 mm filter and one of these cheap (20-40 $) 35 mm lenses is hard to beat.

Further examples of pictures taken with some of the filters discussed on this page can be seen here.


Several types of currently available UV-pass filters are suitable for reflected-UV photography. Unlike traditional UV-pass filters, most modern filters block near-IR radiation and prevent IR contamination of digital UV images, so this is no longer a problem. The well-known Baader U remains useful and a good generic alternative, but other filters with different peak transmission wavelengths and bandwidths can be used to record different and/or narrower UV bands and can reveal details and absorption patterns not easily visible when using a single broadband filter.

With some camera lenses, it is entirely possible to use filters with diameters of 25 mm, or even 12.5 mm. These small filters are significantly cheaper than 2" filters of the same types, especially on the surplus market, also give excellent results, and often are handier to use than the larger 2" filters.