UV photography

Quite a few web sites on UV photography seem to link to this page. They are welcome to do so, but this page contains only partly updated materials since I created it. Other pages of my site contain more recent information, as well as information on special topics in UV photography not mentioned on this page. If you want to make sure that you did not miss any subsequent additions on UV photography I made to this site, always check the main index of the photography section, or use the search function on the home page of this site.

What is UV

In addition to visible light (i.e., wavelengths ranging from 400 to 750 nm) and infrared, sunlight at ground level contains UV-A ( 400-320 nm), UV-B (320-280 nm) and small amounts of UV-C (280-200 nm) light. Above the atmosphere, UV-D, or vacuum UV, is an important part of the solar spectrum. Virtually no UV-D passes through the atmosphere. At even shorter wavelengths, the UV range grades into the soft-X-ray range.

UV-B is absorbed by most materials normally used to manufacture lenses, and ordinary digital camera sensors also possess low sensitivities in these ranges. UV-C is rapidly absorbed by atmospheric oxygen, and at ground level solar UV-C is irrelevant to photography. UV-D is very rapidly absorbed by atmospheric gases, including nitrogen.

How I arrived to UV photography

My early trials with near-UV photography, almost a decade ago, were not much successful, because of very large technical problems that stack up to make the odds almost impossible. Few photographers dealt with this - at the time - very exotic branch of photography, and reliable information was much more difficult to come by than now.

NIR photography, for initial experimentation, may require only a cheap, easy-to-find filter and a long exposure time (albeit modern digital cameras are much better than older ones at reducing the amount of NIR passing through their built-in filters to uselessly low levels). Getting even entry-level equipment for UV photography, instead, is both expensive and time-consuming. Essentially, cameras, digital sensors, lenses and ordinary filters are all designed to cut out UV radiation, and you have to swim upstream in order to obtain odd bits and pieces that, by accident or design, work in the opposite way of what is normally desired.

It is now much easier to start out in UV photography than a few years ago, thanks to current bulletin boards like ultravioletphotography.com and, before the latter, www.fotozones.com (link removed because no longer freely accessible) that freely distribute and discuss technical information. Nonetheless, UV photography requires a considerable determination, time and money to obtain the necessary equipment, and additional determination and time to learn how to properly use this equipment. At the risk of repeating myself:

UV photography is not an occasional pursuit that you can start practicing in an afternoon. Nor can you do so in a week. You might be ready to start after a few weeks or (more likely) months of preparations, if you have enough financial resources, luck, and the right sources of information. Initial mistakes in purchasing the right equipment can be very costly.
UV photography cannot be done with simple, cheap and easily procured equipment. The financial cost of good equipment for UV photography is significantly higher than for general-purpose photographic equipment.
The equipment for UV photography is largely not dual-purpose equipment. It can be used mostly, or only, for a single purpose: UV photography (or sometimes, also for IR photography and multispectral photography). Most items for UV photography are of no practical use, or too expensive, for more common types of photography.

The reward is that we can document and study a world otherwise invisible to us. In addition to its visual appeal, UV photography has numerous uses in science, technology and forensics. However, not too many scientists have used UV photography to document macroscopic objects (aside for flowers and birds plumage, which have been the subject of hundreds, or possibly thousands of papers). As an example, when I decided to look at the shells of land snails in UV (and IR) and did not quite understand what I was looking at, I found out that apparently no other scientist had thought of doing the same thing, and I was able to put to good use what I learned about UV photography in the following paper:

Savazzi, E. & Sasaki, T. 2013: Observations on land snail shells in near-ultraviolet, visible and near-infrared radiation. Journal of Molluscan Studies 79: 95-111. doi: 10.1093/mollus/eys039

A realistic time- and action plan goes as follows:

Start slowly and read a lot in the initial phase, before you buy anything. Ask questions on bulletin boards. Do make your best to understand the technical aspects. Published opinions on equipment and methods in UV photography vary, and in a relatively short time (weeks or months) you will be able to distinguish knowledgeable sources from unreliable ones and/or sources with vested commercial interests.
Do not become dependent on a single source of information, and especially on a source that may have a commercial interest in "locking you in" with specific equipment they sell. Some web sites on UV photography are openly commercial, and only market a small subset of the equipment usable for this type of photography. Other web sites appear at first sight to freely communicate information on UV photography, but are less than forthcoming in clearly identifying the filters, lenses and techniques used to produce the published images, thus making it impossible for other photographers to use the same methods or building on them. Good criteria to apply in this evaluation are:
- How open is a specific information source in providing actual, usable information (including information on competitor's products or other sellers)?
- Are they trying to sell you a "black box" filter/lens/camera with no information on which company actually made it? For example, use of "fantasy names" for UV-pass filters or UV lenses is an early giveaway. Google the names of anything they propose to sell you. If you find nothing, or just the same single web site mentioning the specific item, ask them why. If they are not forthcoming with real, verifiable information, let them go and turn somewhere else. Reputable companies like Tochigi Nikon, CoastalOpt, Thorlabs, Edmund Optics, Asahi Spectra etc. can be trusted even when they are the exclusive maker or distributor of a specific product, and typically provide extremely detailed technical information on their products. A one-person company, on the other hand, is certainly not manufacturing UV-pass filters and lenses in a basement, but only marketing products made by larger companies and potentially available directly from these companies to any interested buyer. In this case, you might be able to avoid a price mark-up if only you manage to identify the actual source, which is therefore worth investigating.
- Does a specific information source recommend equipment that is well documented, tested and reviewed by multiple web sources but only available for purchase from a single source? This could be a warning flag, although not necessarily a sign to avoid them. Many of the filters most useful in UV photography are only made by one company. However, these filters are often available from multiple distributors, in addition to the maker itself. Other companies may be making slightly different but largely equivalent filters. Among the few currently produced lenses designed for UV and multispectral photography, the CoastalOpt 60 mm Apo Macro and Tochigi Nikon 105 mm UV Nikkor may be available through a few large distributors of photographic equipment, but individual buyers can directly order these items from the respective factories and usually save a substantial amount of money in the process. Window-shop before you take out your credit card.
- Does the information from a commercial source agree with information available from open sources like bulletin boards and web sites that do not sell equipment? If the latter sources are verifiably independent from the commercial one, this is an encouraging sign. If they are only parroting the information given out by the commercial source, be cautious, In this case, in reality you only have one information source, no matter how often the information has been copied and re-published.
Continue to keep yourself updated. The more you learn, the more you will become capable of separating wheat from chaff. A lot has changed in UV photography during the last five years, and a lot more can change in the next few years. Much of the generally accepted wisdom of a few years ago is still known to be true, but some of it is no longer regarded as valid. I recommend to start at ultravioletphotography.com, but do search for other sources of information - there are many, of greatly variable usability and helpfulness.
Build up your initial kit gradually in six months to one year, and spread your purchases over this time. "Deals of a lifetime" sometimes do occur, but often they turn out to be a good deal only for the seller. Good deals are sometimes available on eBay, but a deal that is too good to be true is an obvious giveaway. An eBay ad for a UV Nikkor 105 mm at 2,000 US $ stinks like July's prawn leftovers in August, especially if coming from a seller who registered on eBay only last week, has a user feedback of two, or until now has been selling old postcards and used kitchenware from the same eBay account (hijacked eBay accounts are commonly used by fraudsters).
Do not skimp on the camera. Start with a good UV camera because you will likely use it for many years. If the camera is not suitable for UV photography, you will never know what else, if anything, you are doing wrong.
Do start with relatively cheap UV lenses, but only one or two of them. There are enough pretty good ones to choose from. Do not start with very expensive ones, unless you already have them available for other reasons. Do not buy old lenses at random, hoping to chance on one that is good for UV photography but has remained undiscovered until now. The large majority of lenses, old and new alike, are no good for UV photography, and you might need to buy a hundred before you find a usable one. For your initial lens purchases, follow instead the advice of more experienced UV photographers.
Do not skimp on the filter. Start with just a single proven one, the Baader U. It is not cheap, but it will work (or rather: If you cannot get good results, the fault is somewhere else, not in the filter). Only after some time and practice (and enough investigation) try more exotic ones. Buy a filter that you can use on your UV lens(es) and camera. In many cases, with the right adapters on lenses of medium-long focal lengths, you can use a 25mm/28mm/1.25" filter without causing any vignetting. You can use a 52mm/48mm/2" filter, but in most cases this large diameter is not really necessary, especially on small sensors, and more expensive. You can test a given filter size for vignetting in the visible range (e.g. with a circular hole cut in a black cardboard sheet mounted in your filter holder) on a given camera and lens, before ordering a UV filter. Find out which filter size gives no vignetting with the lens aperture fully open, and buy this or a larger size.
Begin your trials of UV photography on subjects in full sunlight and use a tripod. Once you have debugged the initial problems, get a not-too-expensive, suitable electronic flash with uncoated tube (see www.ultravioletphotography.com, or some of the links below), or if you already have a studio flash and are interested in indoors subjects, consider investing in an uncoated tube and dome. You will eventually need an artificial UV source, but it is best to start with direct sunlight, unless you already have a studio strobe with easily-replaced tube and/or live in a place where sunlight is a rare or unreliable commodity (like I do).

Cameras and camera sensors

Camera sensors are moderately sensitive to UV-A, albeit to a lesser degree than to the visible range (400-750 nm). At shorter wavelengths than UV-A, the sensitivity of ordinary camera sensors quickly decreases with decreasing wavelength. This is due to a combination of factors, of which the most important is probably absorption by the Bayer color filters and microlenses built on top of the sensor chip. Depending on its thickness and composition, the glass window covering the chip and sealing it within its ceramic package can be an obstacle to transmission or UV-B and UV-C, but usually it is fairly well to very transparent to UV-A. Anti-reflection coatings of this window are typically designed to be most effective in the visible range, and some of them may reflect UV to varying degrees.

Videocamera sensors without microlenses and Bayer filters can be used in the UV-A and UV-B ranges. The window protecting the sensor is sometimes devoid of coatings, or treated with coatings optimized for UV transmission. The same type of videocamera sensor is sometimes completely devoid of a protective window, and completely exposed to whatever radiation passes through the lens. In this case, the sensor is extremely vulnerable to damage (e.g., atmospheric humidity may corrode the vacuum-deposited aluminium electrical pathways, thus requiring the use of gold for this purpose, and any dust particles are impossible to remove without damaging the sensor), and the lens should never be removed from the camera. This equipment is expensive and has specific technical uses, but it is simply the wrong type of equipment to start out in UV photography.

Fujifilm used to make DSLR models for UV (officially, only down to 380 nm) and multispectral photography. They are no longer available, and it is much more practical to modify a general-purpose DSLR or mirrorless camera that to try and get one of these special cameras.

UV and multispectral conversion

Virtually all sensors in general-purpose digital cameras are covered with a separate UV- and IR-blocking filter. In the large majority of cases, and especially in modern cameras, this filter is very effective for its intended purpose, and must be removed to allow imaging in the UV (and IR) range. Often, this filter also carries two thin layers of birefringent material, which provide anti-aliasing. Increasingly often, this anti-aliasing filter is absent, and aliasing and the related color moiré artifacts are corrected by the camera software.

In DSLRs, this filter is usually replaced with a window of optical glass transparent to the wavelengths of interest. This is necessary because simply removing the filter changes the optical distance between lens mount and sensor, while the optical distance between lens mount and the ground glass of the optical viewfinder is unchanged. Removing the filter without replacing it results in two separate problems:

The camera can no longer focus at infinity with ordinary lenses.
An image correctly focused in the viewfinder is out of focus on the sensor, and vice versa. It is still possible to use a DSLR modified in this way, if one uses only live view to focus.

In so-called multispectral conversions, a window transparent to most or all wavelengths that can be recorded by the sensor replaces the built-in filter. In UV conversions, a UV-pass filter is instead used as replacement. A variety of IR-pass filters are also used as replacements to allow IR photography. This modification is essentially permanent. Therefore, the most versatile conversion is the multispectral type, which requires a filter to be mounted in front (or at the rear) of the camera lens to isolate a subset of the light spectrum.

Most DSLRs allow the distance between ground glass and lens mount to be adjusted, and in some camera models the adjustment range may be large enough to allow viewfinder focus to be recalibrated after removal of the UV and IR filter, without needing a replacement window. The thickness of the original UV- and IR-cut filter is very variable among camera makes and models, and this also plays a role in the availability of this solution. This type of recalibration does not correct the problem of lenses no longer focusing at infinity. For this reason, DSLR conversions usually replace the built-in filter with a transparent window or filter. On the other hand, when using a mirrorless camera and lenses mounted on an adapter, it is possible to simply remove the built-in filter without replacing it, and to use a slightly shorter adapter to restore infinity focus with lenses that can be mounted on an adapter (native lenses designed for direct mounting on the specific mirrorless cameras, without an adapter, remain unable of focusing at infinity).

The infinity focusing position of many lenses can be recalibrated, and in some cases the adjustment range is large enough to allow focusing at infinity on cameras where the built-in filter has been removed but not replaced. This is the case, for instance, of the Novoflex Noflexar 35 mm.

Among Nikon DSLRs, the old D70 and D70s are frequently reported to be particularly suitable for UV photography. Many other models are also suitable, albeit not all. Several Micro 4/3 cameras also work well. I discuss here a Panasonic G3 that I modified for multispectral (including UV) photography with very good results. In the past, it was often stated that Canon DSLRs were unsuitable. It was also stated in the past that CCD sensors were more suitable that CMOS sensors, but this is not supported by the current knowledge. Very capable UV cameras like the Panasonic G3 and GH series have CMOS sensors. It is stated on some sites that some of these cameras are suitable for UV photography even without modification. I prefer to qualify this statement by saying that some of these unmodified cameras allow imaging in the (very) near UV range at the price of extremely long exposure times, even in the presence of abundant UV radiation, but certainly cannot be called optimal for this application.

A further factor to consider is that live view is very convenient for framing and focusing with a UV-pass filter mounted on the lens (provided that a sufficiently strong continuous UV illumination is available). Live view also allows a precise focusing (by electronically magnifying a portion of the live view image), although this requires a stronger UV source than necessary for viewing the whole frame. Older cameras like the Nikon D70 lack live view, and focusing and framing must be carried out without a UV-pass filter. UV focus shift (see below) and, at high magnification, the thickness of the filter can slightly alter the plane of focus, and with these older cameras this problem becomes evident only after examining the pictures.

Technical aspects of UV photography

UV images record radiation outside the human visible range. Therefore, there is no "correct" white balance to display these images, and if you only want to experiment with the visual impression of false-color images, you are free to try the most outrageous white balancing and color remapping techniques.

With cameras, lenses and UV-pass filters that allow the recording of a relatively wide UV spectrum, Bayer sensors most often provide a range of false colors, which include red, magenta, yellow, green and shades of blue and violet. They are called false colors because they have no relationship to the actual colors of the subject as displayed in the visible range. Images of flowers generally provide the widest gamut of UV false color. However, most flowers display just one or two UV false colors, and to experience the full gamut you need to experiment with a rather broad range of flowers.

If your lens only transmits the longest UV wavelengths (i.e., those closest to the VIS range), your UV images will be a monochromatic violet (if you use the built-in sunlight white balance of the camera), and regardless of the subject you will get no other false color. Also, if you use a largely monochromatic illumination source (e.g., a UV LED torch), you will get little or no false color, even if potentially present in the subject.

There seems to be a relationship between UV false colors and the approximate UV wavelengths reflected by the subject. However, this relationship may be somewhat different in different camera models, and is strongly affected by the color balance being used, as well as other factors that include the subject and the illumination source.

"Bee vision" and all that

Sometimes, UV false color is combined with VIS colors, and the color channels of the image are remapped and remixed in post-processing to give an impression of how insects, birds and other animals capable of vision in the UV spectrum see a given subject.

I strongly object to these images being called "bee vision", "butterfly vision", "bird vision" and the like. Animal (and human) vision involves large amounts of data post-processing in the retina as well as the brain. Recording the way this species visually perceives a given subject and translating it to a human perception experience is way outside our present capabilities, and may well be impossible. The visual perception neural signals of another species might very well be completely alien to our brain.

For example, some animals can perceive four or more component colors. Our brains are wired to process only three component colors, so they are incapable of processing a fourth color channel in the same way as done by these species, just as they are incapable of processing the perception of a four-dimensional solid (we can at best visualize the data by using a number of diagrams and formulas, but communicating the actual perception is a whole different thing). Therefore, these "bee vision" etc. images are, at best, only a curiosity useful in science fairs and exhibits for the general non-scientific public, and should always be labelled with the above warnings.

Good science can indeed be done on bee vision etc. It is done within an experimental scientific framework and with well-designed experiments, for example by observing bee behavior and responses to natural and artificially altered flowers and illumination spectrum under controlled settings, and by measuring neural activity related to the vision system while the bee is presented with different radiation wavelengths. See for example:

Horridge A. 2019 - The discovery of a visual system: The honeybee. 296 pp, CABI.
Horridge A. 2011 - What does the honeybee see? And how do we know? A critique of scientific reason. 360 pp, ANU E Press.

My point is that good science cannot be done by shooting images of flowers with UV and VIS filters, remapping their color channels in Photoshop and calling it "bee vision".

Links to cameras for UV photography on this site:

See the cameras section index.

Lenses

Special lenses can be designed for UV-A and UV-B photography. This includes, for instance, the Nikon UV Nikkor 105 mm (and the optically identical Tochigi Nikon UV Nikkor 105 mm), CoastalOpt Apo Macro 60 mm and UV Rodagon 60 mm. None of these lenses are cheap, and most of them are no longer manufactured. Several of these lenses also have relatively long focal lengths that limit their usefulness for landscape photography.

At least some of the general-purpose photographic lenses also transmit rather low amounts of these wavelengths (see links below).Therefore, it is possible to use modified general-purpose digital cameras and general-purpose lenses to take sunlit pictures in the near UV range, at least for static subjects that allow relatively long exposures.

Nikon EL-Nikkor enlarger lenses are often used for UV photography (see links below). Once regarded as the best general-purpose lenses for UV photography, they transmit less deeply into the UV than the lenses mentioned above. Nonetheless, they provide good alternatives for wavelengths down to approximately 365 -380 nm. The EL-Nikkor 63 mm f/3.5 was once regarded as singularly good for this purpose, but several other models (especially the older series in metal barrels) are just as suitable.

Lenses of short focal length, commonly called wideangles, suitable for UV imaging are a special problem. These lenses typically use a large number of optical elements, which makes them intrinsically poor for UV imaging. There are several cheap and easily available models of legacy 35 mm lenses suitable for UV imaging (their availability on eBay seems to have shrunk since the UV properties these lenses became known, but that's life). Lenses of shorter focal lengths usable in UV imaging are much scarcer. My current favorites are the Enna Lithagon 28 mm f/3.5 (and its Porst 28 mm f/3.5 "clone"), and the Nikon AI-S Nikkor 24 mm f/2.8. All these lenses are designed to cover a full-frame film/sensor, and therefore, to fully use their image circle, one should use them on a full-frame camera. They still can be used for UV imaging on APS-C or Micro 4/3 cameras, but on these cameras they behave as medium-focal length "normal" lenses. There are a few legacy lenses usable in UV that behave as "proper" wideangles even on Micro 4/3, for example the Pentax-110 18 mm f/2.8. Even some modern autofocus lenses, like the Sigma 19 mm f/2.8 and 30 mm f/2.8, can be used at the highest UV wavelengths between 380 and 400 nm.

The largest lists of legacy lenses suitable for UV imaging are those on ultravioletphotography.com, but these lists have been compiled from multiple sources that used different test criteria. Before purchasing anything, you should do your homework, Google for actual UV images taken with the lens model you are interested in, and compare them with UV images taken with other lenses of comparable focal lengths.

A problem with many general-purpose lenses used in UV photography is a so-called focus shift between the visible and UV ranges (strictly speaking, this is not focus shift, which is a different phenomenon, but a type of chromatic aberration). This problem is usually not serious when focusing with live view in UV-only imaging, but must still be taken into account in multispectral photography, when superposing images recorded at widely different wavelengths.

Field UV photography

Swapping filters is the most common method for recording images in UV and VIS, when it is necessary to compare images in the two bands. If a precise registration of the images in the different bands is desired, it is also necessary to work with a tripod. Many public places, however, place limitations on the use of tripods, monopods and selfie sticks. In several cases, I have heard that photographers in botanical gardens have been accosted, and even harassed, by overzealous staff even while using tripods in a safe and unobtrusive way. It seems their motivation is often not the well-being of visitors and plants, but rather the worry that the photographers are carrying out commercial work. Since universities and non-profit research institutions in most of the world are in a chronic state of economic strangulation by their respective governments, they appear to have instated requirements to extract even the smallest thinkable profits from commercial photographers, by demanding the payment of fees, licenses and copyright proceedings for any even remotely commercial photographic activities carried out on their premises and assets.

When a perfect registration is not essential, or unwelcome attention should be avoided, I find that the easiest, fastest method is to carry with me two cameras, one equipped for UV and the other for VIS. They do not need to be the same model or format, as long as they provide the same angle of view. For example, an Olympus E-M1 Mark II with 12-40 mm f/2.8 Pro zoom at 12 mm is an almost perfect match in VIS for my full-spectrum Sony 7R II with AI-S Nikkor 24 mm f/2.8 in UV.

The unusual appearance of dichroic-coated filters like the Bader U and Primalucelab U can prompt stares and questions from passers-by, in turn attracting the attention of staff who might interfere with your photography plans. Using a rear-mounted filter avoids this potential problem (although this is not the most important reason to use rear-mounted filters).

First I shoot the UV image, which takes a little time because it requires manual focusing and the use of image magnification, plus typically manually stopping down the lens. Then I let the UV camera hang down on my chest, lift up the VIS camera and snap a quick picture with AF and auto exposure in A mode, which takes a couple of seconds or so. As long as the neck straps of the two cameras are adjusted at different lengths, the risk of banging the two cameras together is low. A variation on this method can use a waist- or shoulder-mounted bag as a double camera holster.

VIS image — Finspång Castle, Sweden, top: VIS, bottom: UV.

Although UV images have an esthetic appeal on their own, one of the technical and forensic uses of UV images is that sometimes they reveal obvious differences in UV reflectivity where none are displayed in VIS. In the above example, the UV image reveals that extensive repair/repainting has been carried out on the left wing of the castle, while none of this is displayed in the VIS image.

Links to lenses for UV photography on this site:

See the UV lenses section index.

Filters

UV photography normally requires a filter that blocks longer wavelengths than UV, so that they do not contribute to pictures. When photographers talk about UV filters, usually they mean filters that are opaque to UV and transmit longer wavelengths. These filters are more properly called UV-cut filters. This is the opposite of what we need for UV photography. What we want is UV-pass filters. A special problem with digital cameras is that their sensors are very sensitive to near infrared (NIR), and any NIR that leaks through a UV-pass filter will usually spoil the recorded image.

Traditionally, UV-pass filters were made from different types of ionic glass. Virtually all these types of glass also transmit NIR. For this reason, traditional types of UV-pass filters are essentially useless for digital UV photography, unless combined with a NIR-cut filter. Multiple filters increase the risk of internal reflections and loss of contrast.

The most frequently used modern filters for UV photography are of three types:

Sandwich-type filters, made by cementing together a stack of different types of ionic glass to allow the transmission of UV but not of IR. The optical glue used to cement the layers must transmit UV (Canada balsam is not usable). Multi-layer filters of this type often display a lower contrast and image resolution than other filter types.
Combination dielectric and ionic filters, which combine a single layer, or multiple stacked layers, of ionic glass with dielectric coatings on one or more surfaces of the filter. The most useful filter types for UV photography have a single layer of ionic glass and different types of coatings on either side. The dielectric coatings may consist of up to one hundred layers.
Dielectric, or interference, filters, which use dielectric coatings on a substrate of fused silica that transmits UV, visible and NIR radiation.

Dielectric coatings have variable spectral transmission properties, depending on the angle of incidence of the radiation. For this reason, they should not be used with wideangle lenses. Anti-reflection coatings are similar to dielectric coatings, but typically consist of fewer layers. Some of the traditional UV-pass ionic filters are available in anti-reflection coated versions (e.g. from Thorlabs) and perform better than their uncoated counterparts in terms of image contrast. However, anti-reflection coatings do not solve the NIR-leak problem.

Ionic UV-pass filter, degraded by exposure to air.

Detail of degradation of optical surface.

Many traditional types of ionic UV-pass optical glass deteriorate on exposure to air and/or humidity. This deterioration makes the surface of the filter "matte" and strongly degrades its performance. Some of these types of glass, depending on the climate and humidity, may visibly deteriorate in as little as a few months, typically developing visible "flowering", "icing", or moss-like matte areas, on their optical surfaces (see above example). This degradation may look like molds, but is not of biological origin, and cleaning the filter only makes it worse. In other types of UV-pass ionic glass, deterioration is not immediately visible, but cleaning one of these filters causes microscopic glass particles to detach from its optical surfaces and makes the filter unusable for UV photography.

It has been reported that appropriate cleaning of air- and humidity-sensitive UV-pass ionic glass filters is feasible while the damage is still in its initial, barely visible state. Dielectric coatings usually protect the filter surface from this type of damage, but require extra care when cleaning. In general, unless I need to clean an accidental fingerprint, I only use an air blower to clean coated filters. It takes a lot of dirt and dust on a filter before image quality is affected, especially if the filter is protected by a lens shade or rear-mounted between lens and camera. Molds, on the other hand, should be cleaned out before they have a chance of etching themselves into filter coatings or optical surfaces.

Unless the optical surfaces of these ionic glass types are protected by sufficiently thick dielectric coatings or by sandwiching with more resistant types of glass and air- and humidity-proof optical glue, these filters are only usable for photometry (although they may require a periodic re-calibration of the equipment), but not for imaging and photography. The filter in the above images has a dielectric coating on the opposite surface, which is still pristine. Any scratch in dielectric coatings, however, opens the way to more significant, and steadily growing, damage. In extreme cases, degradation that starts at the edges of a sandwich- or combination-type filter, where the ionic glass is exposed to air, can eat its way through the glass toward more central regions of the filter. In these cases, the filter edges should be air-proof sealed with epoxy or silicone within a metal cell.

Other UV-pass filter types exist, but are typically very expensive. For instance, so-called solar blind filters (which transmit only UV wavelengths that are selectively absorbed by the atmosphere or not emitted by the sun) use a nano-scale grid of metal wires on a fused quartz substrate to reflect higher wavelengths. A partially similar technology, which uses parallel nanowires, allows the construction of polarizing filters effective in the UV. These are very expensive specialty filters with military and industrial applications, and rarely seen on the second-hand market.

Strong and/or lengthy exposure to UV radiation degrades many types of UV-pass filters and results in a lower UV transmission. This is sometimes called solarization (and is unrelated to the darkroom printing process known with the same name). Typically, this type of degradation does not affect image quality, but does require a longer exposure or stronger UV source. This is especially likely to happen to filters placed on a UV source, rather than on a camera lens. For some uses, manufacturers recommend replacing the UV-pass filters after six months or one year of use, although this recommendation may be dictated by the necessity of avoiding a combination of solarization and air degradation.

For maximum durability in UV photography use, one should choose only filters with optical surfaces coated with dielectric layers, if possible use them on the camera lens rather than light source, and protect them from excessive humidity and strong light when not in use. The use of lens shades left permanently attached to UV-pass filters, in addition to improving image contrast and reducing the risk of internal flare (see here), reduces the risk of accidentally touching the filter surfaces.

A lens cap mounted on a filter also makes it easier and less risky to handle the filter. Remove the lens cap after screwing the filter into the filter mount of the lens, and reattach the cap before unscrewing the filter.

Wipe optical surfaces only with a new (never before used) optical-grade paper towel, and only with the towel moist (not dripping) with alcohol, lens cleaner or water. Throw away a lens-cleaning paper towel immediately after using it, to make sure you will not re-use it by mistake. The portion of surface of the lens towel that will wipe an optical surface must not touch or wipe anything else (including your fingers) before being used. Never wipe an optical surface with a dry lens cleaning towel, or with anything else than a lens towel (I do not trust reusable micro-pore towels on optical surfaces, although they are fine for other surfaces). Practice cleaning a piece of glass or a cheap UV-cut filter with a new cleaning fluid or lens cleaning towel type, before using them on expensive equipment.

Two streaks of different types of sun lotion, in visible light (top) and UV (bottom).

Sun lotion, cosmetics and all sorts of spray cans should be kept well away from photographic equipment. They can be very difficult to remove from optical surfaces and some of them block UV radiation extremely well.

Links to filters for UV photography on this site:

See the UV and IR filter section index.

Sources of UV radiation

Warning. UV damages the eyes and skin (especially UV-B and UV-C, but UV-A is by no means safe), so you should not stare into UV light sources and avoid skin exposure. Even if the sources are not powerful, exposure to UV is compounded by the fact that the eye's iris does not contract in the presence of UV, thus increasing the effective irradiation of the eye. It is also a good idea to switch off the UV sources when not in use and to wear protecting goggles that cut UV radiation (e.g., made by UVEX and specifically classified for UV protection, not just ski or fashion goggles), to prevent long-term exposure to indirect UV during photography sessions. Some prescription glasses do not cut UV, and many cheap sunglasses are transparent to UV.

As UV-only sources, initially I tried both "black light" fluorescent tubes (of the type that can be screwed into an ordinary socket for a light bulb) and UV LED banks (which can be mounted in 12 V sockets for halogen lights). The latter seem to produce more blue and indigo light and less UV than fluorescent tubes (at least in the models I tried). LED banks can be battery-powered or fed with a 12V DC power supply, but they draw so little current that the 12 V solid-state transformers used for halogen lights sometimes are fooled into thinking that no lamp is present, and refuse to provide a current. Ordinary incandescent light bulbs (especially those of higher power and higher Kelvin temperatures) do emit potentially usable amounts of UV, but they also emit a lot more IR. I tested additional UV sources here and here.

Specialist continuous light sources are capable of producing relatively large amounts of UV (e.g., for adhesive curing and fluorescence microscopy). The tube of electronic flashes does produce a lot of UV, which normally is filtered out by a yellowish coating on the outer surface of the tube (sometimes the filter is colored glass used to make the tube, and therefore is not removable). The plastic exit window of flashes also cuts UV, and sometimes has a distinct yellowish color that indicates the presence of UV-absorbing materials.

There are a few specialist electronic flashes designed to provide unfiltered UV, but they are scarce, and probably mostly discontinued (this includes the Nikon SB-140). In principle, one might attempt to convert a general-purpose electronic flash with uncoated tube by removing the plastic exit window. Among old (and usually cheap) battery-operated flash models known to have been modified with success are the Vivitar 285 (and its more recent and more expensive revival, the 285HV) and the Metz Mecablitz 45 CT-1.

Removing the flash window of the Vivitar 285 is difficult and messy (it is actually better to just cut a rectangular hole in the window), while doing it with the Metz 45 is easy. The Metz 45 CT-1 has been suggested as a UV source for film UV photography (even without removing the front plastic window, which means it would only be usable down to 360-370 nm) as early as the 1990s. It is frequent and cheap on the second-hand market, and recommended by a few UV photographers. Once the window is removed, I found it usable at least down to 325 nm, and possibly 300 nm. It produces a significantly higher amount of UV than the Vivitar 285.

The original rechargeable Ni-Cd battery packs of the Metz 45 CT-1 are almost always unusable. A replacement battery holder for 6 AA alkaline batteries, made in China, is cheap and commonly available. It is often mentioned on web sites and bulletin boards that the rechargeable and alkaline battery packs are different, and use different electrical contacts (which is true). However, I and other photographers have been successful in loading the unmodified alkaline battery holder with six Eneloop nickel hydride batteries, and using the flash in this way without any obvious problems. With these batteries, voltage is about 1.5 V lower than with alkalines, and the electronics seem to be able to cope with this difference.

It is not necessary to add a capacitor in the alkaline battery pack for this purpose (the capacitor in the original Ni-Cd pack is only used to smoothen out the recharging voltage, and early original rechargeable packs don't have it).

The trigger voltage of the Metz 45 CT-1 can be quite high, and it is recommended to fire it through an optical or radio slave instead of connecting it directly to a digital camera.

It is stated on some bulletin boards that the plastic front window of the CT-1 is easily popped out without opening the flash casing. This is not quite true, as doing so is likely to break the window and/or casing. The head part of the casing is easily opened by removing four screws (attention must be paid not to touch any of the electronics, see below). The window then slides out, and can be put back if desired.

Some of the subsequent Metz models may have UV-coated tubes.

Studio strobes can be modified by replacing a UV-coated tube with an uncoated one. Uncoated replacement tubes for studio strobes are cheaper than the ones coated with a UV-cut layer. Cheap, amateur-level studio strobes made in China sometimes use uncoated tubes. These cheap strobes are neither durable, nor reliable, nor especially powerful, but they may be an acceptable compromise for studio UV photography. See here and here for an updated discussion of studio flash units for UV photography.

A word of caution when modifying electronic flashes: their capacitors store a potentially deadly high-voltage charge even when the power is turned off and power cords and batteries are removed. The capacitor may remain charged for hours or days, especially in battery-operated units and cheap studio units that do not actively discharge the capacitor when power is switched off.

There may be a dangerous charge left in the capacitor even after manually triggering the flash. This applies also to battery powered units. Make sure you know what you are doing and check for high voltages before replacing flash tubes.

In my personal experience, flash units with uncoated tubes are by far the most practical sources of radiation for UV photography (above examples). With powerful ones, I can record radiation down to about 300 nm (imaged as blue in the picture at the right; see details here). However, a continuous UV source is also very handy, because it allows focusing and framing the subject in live view, without removing the UV-pass filter from the lens. UV LED torches are useful for this purpose down to about 365 nm, but LEDS that emit shorter wavelengths are weak and expensive. LEDs nominally specified as emitting 365 nm radiation often emit 370, 375 or even 380 nm instead. This is especially common in cheap Chinese UV flashlights sold on eBay (some of them are even known to emit only at 405 nm, which is in the visible range).

Certain types of specialized fluorescent tubes emit a more broadband UV radiation in the UV-A, UV-B and UV-C ranges. Different types of tubes must be used for different wavelength bands. As a rule, their emission is too week for photography, and their usefulness mainly in framing and focusing.

Links to electronic flash for UV photography on this site:

Bowens Gemini 500R
Bowens Gemini 1500Pro

Links to continuous UV sources for UV photography on this site:

Continuous sources of near-UV
Building a UV LED lamp

White balance in UV photography

The title of this section sounds like a conundrum, since UV photography records a part of the electromagnetic spectrum invisible to human eyes. However, the white balance (WB) set in a UV camera has a large impact on the false color recorded by the camera in UV photography. An inconsistent WB makes the false colors of images recorded in the UV impossible to compare across cameras. Unlike WB in the visible range, in UV photography there is no single "correct" WB, and the main purpose of WB in UV photography is to obtain comparable results with different cameras.

The illumination source is also important in this context. WB in the visible range is typically set for a subject in direct sunlight or electronic flash. Given that these seem to be the most frequently used illumination sources in UV photography, it does make sense to use one of these sources for WB.

My current understanding of this problem is that two types of WB can be used in UV photography for the purpose of making images comparable across cameras:

Setting the WB while using no filters except for a UV-cut and NIR-cut filter (like the B+W 486 or the Baader UVIR cut) that approximately restores the original spectral performance of a camera before its built-in UV- and IR-cut filter was removed. Usually, this is moderately, but not very, different from the "standard" sunlight or electronic flash WB of the camera.
Setting the WB on a reference substrate while using the same UV-pass filter used to record images. The reference substrate can be one of the following:

Spectralon white target. Spectralon is chemically similar to PTFE (also known as Teflon), albeit Spectralon seems to be a highly reflecting sintered powder containing plenty of microscopic air spaces, while Teflon is usually a compact, less reflecting solid. For this reason, Spectralon is sensitive to contamination by oily substances, which permeate its pores, and even by airborne traffic pollution. Spectralon is often found on the second-hand market in the form of integrating spheres recovered from optical equipment, as well as flat, round targets (either calibrated or uncalibrated.
PTFE white target. As mentioned above, solid PTFE is less reflective than Spectralon. However, its spectral properties remain similar. A problem with PTFE is its relatively high translucency, which means a thickness of at least 5-10 mm should be used.
A matte aluminium panel. The reflectivity of clean, non-anodized aluminium is practically linear well into the far UV. This highly desirable property is counterbalanced by the difficulty of obtaining a non-directional reflectivity by brushing an aluminium surface. Sand-blasting may be better than brushing.
A combination of a thin (about 2 mm) PTFE sheet backed with a slightly matte aluminium foil or sheet. This combines the best properties of the two materials. No glue should be used to keep the two materials together.

The first method uses the same WB, regardless of the UV-pass filter being used. This is the type of white balance I use for my pictures when I need to quickly alternate between VIS and UV imaging. The second method can be used with one specific UV-pass filter to set a "reference" WB, or can be set individually for each filter type. This means that the second method is more dependent on the used equipment. Both methods have advantages and disadvantages and, as mentioned above, there is no objectively right method.

Ordinary white and gray cards used for WB in the visible range have unpredictable UV and IR reflectance, and should not be used for this purpose. However, sintered PTFE, unless contaminated by other chemicals, performs very well as a VIS and UV reflection target. Even a clean, ordinary PTFE sheet with a thickness of several mm can be satisfactory for relatively non-demanding uses.

The problem of white balance in UV imaging is discussed above. Without an appropriate white balance, even if the image contains some false color, with the wrong white balance this false color may pass unnoticed and the image will look monochromatic.

If you wish to publish UV images that need to be compared with the UV images recorded by other photographers, you need to use a white balance that produces a consistent visual impression. If you need to compare images shot with the same camera and filter but with different lenses, you need to use exactly the same white balance and the same subject and illumination source for the comparison shots. For reliable comparisons, this rules out sunlight, which changes with weather, location, time of day and season. Electronic flash, and especially studio flash units with non-coated tubes, are a more consistent source for this type of tests.

See here for more on white balance in UV imaging.

My current kit for UV photography

In the end, my current, satisfactorily working solution for UV photography involves a few unusual pieces of equipment (not counting various adapters needed to make this equipment work together):

A Most often, Panasonic G3 modified for UV + visible + IR photography. Earlier, I used with success a D70s modified for UV + visible + IR photography, but the G3 performs better.
A set of several UV-pass filters with different transmission characteristics. This is the type of equipment I update and expand most often, since they are the components most likely to affect the results.
A Jenoptik CoastalOpt 60 mm f/4 Apo when I need the best results (i.e., most often), one of the several "accidental" 35 mm UV lenses when I need a shorter focal length, and an AI-S Nikkor 24 mm f/2.8 (on full frame) when I need a "real" wideangle for landscape UV imaging. I use a UV Rodagon 60mm lens to save space and weight in travel photography. This is a good lens, but otherwise not particularly useful since I also own a CoastalOpt 60 mm. I used in the past Nikon El Nikkor enlarger lenses, which are a good and cheap starter solution if one does not need to go below 370 nm (The El Nikkor 80 mm and 105 mm are most usefult in this context).
Bowens 1500Pro studio strobe with uncoated tube and dome, and Bowens 500R. The 1500 Pro is my workhorse for studio UV photography, while the 500R is more appropriate for studio photography in the visible range.
MTE U301 or 365 nm Convoy S2+ LED torch (the latter is smaller and cheaper), especially useful for Live View framing and focusingwithout removing the UV-pass filter. Note that there are several variants of Convoy S2+. The "right" one uses a Nichia 365nm UV LED.
These torches are equipped with a 3W Nichia 365 nm LED. They are mostly useful for focusing and framing with a Baader U or other filters that transmit well at this wavelength. They are occasionally useful as an alternative to electronic flash for heat-sensitive subjects, but only for subjects that do not require a more broadband UV source. This LED is quite powerful (650 mW radiated power) but the beam of these torches is narrow and not focusable. It needs to be diffused either with an external aluminium reflector (cheap, inefficient, enlarges the source area) or a concave fused silica lens mounted directly in front of the torch (expensive, efficient, but source area remains small).
Three types of small fluorescent tubes.When swapping different filters, fluorescent UV tubes are more versatile than LEDs for framing and focusing, because they emit a more broadband UV radiation.

Things that I attempted but did not work well

I attempted several of the commonly discussed "fixes" for near-ultraviolet photography, (i.e., attempts at using and/or modifying lenses not designed for this purpose), with scarce success. At the time not being able to afford a true UV lens, which would have removed most problems, I tried, for instance:

Removing the coating on the front lens element in a Nikkor E 35mm f/2.5 and a Nikkor E 100mm f/2.8. These lenses are described as the subject of a similar treatment on another web site. I did this by rubbing the lens surface with a damp cotton cloth dusted with cerium oxide polishing powder. This substance is very hard, and used for polishing gemstones. Also the lens coating is very hard, but rubbing by hand for a couple of hours did remove most of the coating. An unexpected problem was that the lens material is much softer than the coating, so the lens surface turned slightly bumpy on a microscopical scale. The polished lenses did transmit fair amounts of UV (see example here), but also turned into interesting examples of soft focus lenses (the effect is slight, but interesting for portraits, and subtly different from using a soft filter). However, a soft focus lens is not what I wanted. Removing the coating by chemical treatments has been attempted by other photographers, albeit with no reports of success.
Using some enlarger lenses that are transparent to the near-UV. The EL-Nikkor 63mm f/3.5 is often recommended for this purpose. Perhaps because of this fame, it seems to be hoarded by collectors, and is much scarcer and more expensive than other EL-Nikkor models and focal lengths. However, my tests show that eight other EL-Nikkor models are roughly as good for near-UV photography. (see also my comparison of 63 mm EL-Nikkor models in near-UV and photomacrography). Probably, all EL-Nikkors, especially those from older series in metal barrels, are approximately as good for this purpose (see also independent information here, or the whole blog here). This did work, but not very far into the UV. Cheaper legacy lenses are actually better suited for photography below 370 nm.
Using normal, modern lenses with a UV-pass filter and enormous exposure times. Eventually, some UV becomes detectable, but usually buried in noise. Modern lenses are less likely to work in UV photography than legacy ones (with the exception of lenses specifically designed for UV photography). However, even among legacy lenses, perhaps only 1% are of any use in UV photography. This is why the cumulative experience of dozens of UV photographers is so useful.

Fluorescence is not UV photography

This is photography of UV-induced fluorescence in the visible range.
*This is not UV photography*.

UV photography is very different from the photography of UV-excited fluorescence in the visible range. Fluorescence is the emission of longer wavelengths than those of incident light. Thus, illumination with UV radiation can cause the subject to fluoresce by emitting visible light. This is, for all practical purposes, photography in the visible range, and in fact it does not require special lenses and may even be carried out with a UV-cut filter on the lens. The above pictures are an example of UV-excited visible fluorescence. Some of the markings on a banknote, which are almost invisible in ordinary light (left), fluoresce strongly when illuminated with near-UV light (right). Visible light illumination was used in both pictures (in addition to UV illumination in the second picture).

For years, Google ranked the above picture high in searches for "UV photography". I had to fix both the picture metadata and figure caption to decrease the chance that this misleading search result comes up in the context of UV photography.

Some materials fluoresce in the near IR when illuminated with visible light. This fluorescence is typically weak, but has scientific applications. In rare cases, materials fluoresce at shorter wavelengths than the incident radiation (by absorbing two photons in quick succession before releasing their energy as a single photon). Also this type of fluorescence is typically very weak.

Fluorescent materials release photons a very short time after capturing incident photons. For practical purposes, absorption and fluorescent emission are simultaneous. Phosphorescence is partly similar to fluorescence, but phosphorescent materials capture photons and store their energy for a longer time (even minutes or hours) before releasing it. This characteristic allows the imaging of phosphorescence by starting the exposure after the stimulating radiation source has been switched off, and in this way requires no filter to cut out the incident wavelengths.

Phosphorescent materials also emit photons when exposed to other types of radiation than photons. A small amount of radioactive material is sometimes mixed with phosphorescent materials to make them weakly glow even in the absence of any ambient light. This use is now less common than in the past, because of health concerns about low-dosage radiation.

Quite a few web sites talk about "underwater UV photography". In all cases I am aware of, they do not deal with UV photography, but with UV-excited fluorescence in the visible range (or sometimes, blue LED-excited fluorescence in the visible range). Many of their images are very interesting, but I have yet to see true examples of underwater UV photography.

The above picture shows visible fluorescence of a white cardboard sheet illuminated by the UV fluorescent tube visible in the picture. The fluorescence in the picture is much brighter that the surface of the UV bulb. This picture was taken without additional light sources. This example shows that, even with a UV-only light source, often it is necessary to use a UV-pass, visible-cut filter on the camera, in order to eliminate visible fluorescence of the subject that may overwhelm its actual UV reflectance. Some UV fluorescent tubes do emit also large amounts of IR.

Some people report "seeing" UV light. Most of these people do not possess unusual capabilities. There are documented genetic traits in humans that produce minor differences in spectral response of the eye photo-pigments, and some people are indeed slightly tetra-chromatic, but they do not include UV vision. Instead, those who report seeing UV at improbably short wavelengths may be, more likely, experiencing UV-induced fluorescence of parts of their eyes, which causes a stimulation of the retina in the visible range. The most common result of fluorescence of the cornea, crystalline or vitreous humor is a perception of "haziness" in the presence of UV. If the retina itself is fluorescing, instead, the images of bright UV sources can be perceived as very sharp. This is especially likely in cataract patients who have had the crystalline surgically removed and are using prosthetic lenses transparent to UV. In other cases, reports of people "seeing" the UV output of a monochromator are simply seeing the equipment leaking radiation in the visible range.

The human eye crystalline acts as a strong UV-cut filter, and usually prevents UV damage to the retina. The crystalline itself, however, can be damaged by long-term UV exposure, often leading to cataracts. In people who have had the crystalline surgically removed (usually to restore vision in cataract patients), UV from sunlight reaches the retina almost unimpeded. As a result, the retina is easily damaged by UV (and perhaps even by blue and violet light), and these persons must protect their eyes from sunlight with UV-absorbing glasses or contact lenses. They may also need to wear UV protecting lenses when indoors in buildings with large windows. In my house, double-pane windows transmit about 50% of solar UV at 340 nm, and still a relatively high amount at 325 nm.

I installed so-called UV-blocking film (metallic gray) on some of my windows, but these windows became only slightly less transparent to UV than windows without this UV-blocking film. The difference is perhaps less than one stop, or 50% of transmission. I would not call this an effective UV-cut filter.