35 mm lenses for UV photography

Photography in the near UV requires modified cameras, special filters and special light sources, as well as unusual lenses that transmit radiation between approximately 320 and 400 nm and form images of good quality in this wavelength range. UV photography has been done for more than a century, and its first use is generally credited to R.W. Wood. It has been for decades a standard tool in several fields of biology, medicine and forensics, and several hundreds of scientific papers deal with UV photography and its applications. However, it has remained a specialty technique, usually carried out with expensive purpose-made equipment. Nonetheless, ordinary equipment can be assembled and modified to access the portion of the UV spectrum closest to the visible range.

Several modern digital camera models can be modified for UV and multispectral photography at the cost of a few hundred $, and in some cases at no cost at all if you perform the modification yourself. About four or five modern types of UV-pass filters are available for this purpose, but almost invariably they are expensive. Lenses for near UV photography, on the other hand, may be the most critical and difficult items to acquire. With the exception of two current, high-priced models, i.e., the Tochigi Nikon UV Nikkor 105 mm and CoastalOpt 60 mm, one is forced to rely on legacy lenses that, accidentally or by design, happen to perform well in this type of photography.

"Accidental" UV lenses were not specifically designed for imaging in the UV range. In fact, this capability was at the time regarded as undesirable. In particular, in landscape photography with film, UV causes clear atmospheric conditions to be imaged as hazy. High-quality lenses were intentionally designed to be opaque to UV, and photographers were routinely told that it is always a good idea to use UV-blocking "haze" filters.

Legacy lenses intentionally designed to transmit varying amounts of near UV radiation and suitable for UV photography with digital cameras include several (albeit not all) EL Nikkor enlarger lenses. The usefulness of EL Nikkor lenses in UV photography has been recognized at least since 1999 (probably, initially by Rolland Elliott, and shortly thereafter by Bjørn Rørslett). The most useful focal lengths for UV photography in this series are 80mm and 105mm, and the most useful models are the earlier ones in metal barrels. For a few years, the El Nikkor 63 mm f/3.5 was regarded as especially good for UV photography (and as a consequence was sold at extremely high prices), but simple testing has shown the other El Nikkor models to be approximately as good. The Rodenstock UV Rodagon 60 mm f/5.6, designed for UV microfilming, is expensive and rare, but can still be found now and then. With few exceptions, lenses accidentally useful in UV photography have relatively few lens elements, and often no cemented element groups (because most types of lens cement absorb UV). Most EL Nikkor lenses do have two cemented groups, but it should be remembered that a moderately good transparency to UV is one of their intentional design parameters. Therefore, Nikon used a UV-transmitting cement in these lenses.

While only a lens that uses optical elements built with special materials (usually, fluorite and quartz or fused silica) can transmit UV wavelengths at 300 nm and below, the solid-state sensors of current digital cameras, at most, can record useful images at wavelengths down to about 320-340 nm. Therefore, accidental UV lenses are often sufficient to exploit the capabilities of these cameras, and using expensive "proper" UV lenses that transmit far into the UV would provide few advantages. One of the advantages afforded by special UV lenses is achromatic or apochromatic correction into the UV, which means an absence of UV "focus shift".

Note that the wavelengths of near-UV radiation are shorter than those of visible light. This makes it possible to close the lens aperture by at least one additional stop in the UV, compared to the point at which diffraction becomes a problem in visible light. Therefore, while f/16-f/22 is often too narrow an aperture for maximum resolution in macrophotography in visible light, it is usually acceptable in near UV macrophotography.

Shorter focal lengths than 50-60 mm are especially useful on APS-C and Micro 4/3 cameras. Until recently, the Novoflex Noflexar 35 mm f/3.5 was virtually the only accidental UV lens in this range of focal lengths known to work well for this purpose. It is a good performer at wavelengths down to 360 nm, but has become scarce and expensive in the last few years.

In 2012, information started to emerge about other accidental UV lenses of the same focal length. They have a UV performance comparable to, or better than, the Noflexar 35 mm. The following discussion deals with a few models of these recently discovered accidental UV lenses I own, as well as related models and brands.

Figure 1. The first group of tested 35 mm lenses.

Figure 1 shows a group picture of the 35 mm lenses I initially tested. These lenses and their mechanical characteristics are shown and discussed in detail at the links in Figure 2. Their UV performance is tested and discussed in the rest of this page.

My expectations before this test were mixed. Some of the tested lenses (in particular, the Kyoei type 2) were already known to be singularly good for UV photography. Others obviously use the same optical scheme, but were produced in a later period and the color of their coatings is obviously different. Still others use different optical schemes, and were included in this test mostly in the off-chance that they shared some of the characteristics that made the other lenses useful in UV photography. As a whole, I expected only two or three of the tested lenses to perform well in UV photography, and perhaps another couple to be usable to a lesser degree. Kyoei lenses are getting scarce and expensive, so I did not expect this test to reveal any equally excellent but simultaneously cheap UV lens.

Other potentially interesting (but untested) brands and models

This section contains brands or lens models I have only seen in pictures. As far as I know, none of these lenses have been mentioned as useful in UV photography by other sources. However, some of them seem to be rebranded clones of lenses known to be good, or at least usable, for this purpose. Others were made in the same period, possibly by the same factories, and might be worth testing.

  • Taika Terragon, as shown on Capt. Jack's site. This seems identical to the Kyoei type 2.
  • Hanimar. Some of the preset models might be optically identical to some of the Photax discussed below, but in a different barrel and perhaps with different lens coatings.

Multicoated lenses

UV radiation contained in daylight imparts a "hazy" appearance to landscapes in film photography, because photographic film is very sensitive to UV (even to extreme UV wavelengthts that cannot be recorded by digital cameras). Camera lenses that transmit significant amounts of UV were therefore regarded as of lower quality than UV-blocking ones. In the 1950s and 1960s, it was virtually obligatory to use a UV-cutting "haze" filter in outdoors photography. This habit developed into a general recommendation to use UV filters on all lenses at all times. This habit is still common today, in spite of the fact that digital cameras are equipped with internal filters that completely cut UV, and make a UV or haze filter useless. In fact, at least half of the cheap UV filters sold today are not UV-blocking filters at all (they freely transmit UV radiation between 300 and 400 nm), and the only reason left to use them is to provide a mechanical protection of the front lens element against accidental damage or contamination.

In the 1970s, lens makers learned to make multicoated lenses that provide clearer images than single-coated lenses in the visible range and block virtually all UV. Naturally, this renders multicoated lenses useless for UV photography. Some of the lenses made in this period, made with poorer and cheaper coatings designed to "look" multicoated, may still be usable in UV photography, but the large majority are unsuitable. From the 1980s onward, all lenses are multicoated and transmit practically no UV (with few exceptions, like the EL Nikkors, intentionally designed to transmit long-wave UV). In this period, increasingly large numbers of optical elements became the rule in lens designs, further cutting the amount of transmitted UV. As a result, since the 1980s virtually all general-purpose camera lenses totally block UV.

The Nikon Nikkor-E 35 mm f/2.5 from the early 1980s uses an optical scheme with 5 elements in 5 groups. Bjørn Rørslett reported a moderate success in UV photography with this lens after painstakingly removing the coating of the front element of this lens with a slurry of fine abrasives. This is a slow and high-risk operation that takes several days of work to carry out properly. Doing this in a hurry is very likely to turn the lens into a soft-focus, low-resolution lens useful, at most, for lomography. The simple optical design without cemented elements probably contributes to the relatively good UV performance.

The Nikkor-E series was meant as a series of optically good but relatively cheap lenses in all-plastic barrels except for the bayonet mount. It is quite possible that, to save on production costs, only the outer coating of the front element was designed to block UV, while the internal surfaces used cheaper single coatings. This may explain why removing the coating from the outermost element surface has a major effect on UV transmission. In lenses where all elements are multicoated, this likely would not produce a significant improvement.

UV tests

The following tests were made by photographing a set of filters in transmitted illumination by a broadband light source (a modified Bowens 1500 Pro studio flash). The camera is a Panasonic G3 modified for multispectral photography. From left to right in the images, the filters have peak transmission wavelengths of 325, 340, 370, 385 and 400 nm, respectively. 325 nm lies close to the practical UV sensitivity limit of CMOS sensors. 400 nm is already in the visible range and corresponds to indigo. Note how the different UV wavelengths are recorded as different colors by the camera sensor. This false-color response seems to be fairly constant among different makes and models of sensors. All test images were shot at the same flash power, with the lenses stopped down to f/8 and in manual exposure mode, and allow a qualitative comparison of the transmission levels of the tested lenses. Lighter filter images correspond to a higher lens transmission at the specific wavelengths.

Noflexar
 
Kyoei type 1
 
Kyoei type 2
 
Optomax
 
Soligor
 
Soligor (with very worn out coatings)
 
Hanimex
 
Photax type 1
 
Photax type 2
 
El Nikkor 63 mm f/3.5
 
UV Rodagon 60 mm f/5.6
Figure 2. Test results. Filters from left to right: 325 nm, 340 nm, 370 nm, 385 nm, 400 nm.

Figure 2 shows the filter strip imaged with the lenses discussed at the respective links, and in addition with the El Nikkor 63 mm f/3.5, which has a fame as a good UV lens (although most of the other EL Nikkor lenses are very similar in practical use) and with the Rodenstock UV Rodagon 60 mm f/5.6, which is designed for UV microfilming up to 365 nm. Some of the images are not perfectly focused, but this is irrelevant to the present test.

All tested lenses give a good UV image resolution on Micro 4/3 cameras. This is not surprising, since these lenses were designed to cover 36 by 24 mm film. On Micro 4/3, only the central, best corrected part of their image circle is used. The Photax type 1 is surprisingly good in close-up UV photography, and better than the Noflexar in this respect. This is remarkable, given that the Photax was sold as a general-purpose, low-cost lens targeted to the mass market, while the mechanics of the Noflexar were specifically designed for close-up and macrophotography. On the other hand, if it is correct that the Noflexar 35 mm inherited the optical formula of the Staeble Lineogon 35 mm f/3.5, then the optics of this lens were not originally designed for macrophotography, but only as a general-purpose 35 mm that was later repurposed for Novoflex in a barrel suitable for macrophotography.

The lenses discussed on this page were not specifically tested for focus shift between UV and visible ranges (which is actually a type of axial chromatic aberration). In general, focus shift is low among these 35 mm lenses. A low amount of focus shift, or ideally none, is a desirable property in multispectral photography. On the other hand, focus shift is unimportant in UV-only photography with cameras capable of live view and when a continuous UV source (e.g. LEDs) is available for live-view focusing.

The tests were carried out with a high-contrast subject, and the black areas of the test strip clearly show which lenses have a lower contrast and a tendency to flare or "fog". However, with ordinary subjects, the contrast differences among these lenses are much more limited.

Test results and discussion

The results can be qualitatively summarized as follows:

  • All tested lenses perform well down to 360-370 nm. There is no particular reason to prefer a specific model over the others in this respect.
  • Noflexar 35 mm f/3.5 - Usable down to 340 nm, hardly any transmission at 325 nm.
  • Kyoei types 1 and 2 - Both usable down to 325 nm. The type 2 has better contrast than the type 1.
  • Soligor, Optomax and Hanimex- Essentially undistinguishable from the Kyoei types 1 and 2. Worn out and damaged lens coatings result in lower contrast, but there is no particular reason to prefer the Kyoei to these brands in spite of the obviously different color of the lens coatings.
  • Photax types 1 and 2 - The type 1 trasmits only low amounts at 340 nm, and the type 2 no visible amounts at this wavelength. They should not be chosen if wavelengths below 360-370 nm are important.
  • EL Nikkor 63 mm f/3.5 - In spite of its fame as an exceptionally good UV lens, it transmits only faint amounts of radiation at 340 nm. It should only be used if wavelengths below 360-370 nm are unimportant. This lens often sells for 400-800 $, but a Photax 35 mm f/3.5 type 1 gives roughly equivalent results (except for the different focal length) and can be bought for 10-20 $. A 20-30 $ Soligor is clearly superior to the EL Nikkor 63 mm f/3.5.
  • UV Rodagon 60 mm f/5.6 - At 325 and 340 nm, this lens clearly transmits the largest proportion of radiation of any of the tested lenses. However, good luck finding a specimen of this lens, regardless of price.

After testing the above lenses, I found that also a Prinz Galaxy 35 mm f/3.5 , a Galaxy 35 mm f/3.5 and a Soligor 35 mm f/3.5 in a different barrel than those tested above use the same optics as the Kyoei, Soligor, Optomax and Hanimex discussed above, and give the same results in UV photography.

An additional result is that the known-good Kyoei, Soligor, Optomax, Prinz Galaxy, Galaxy and Hanimex discussed above are not only usable down to 325 nm, but are even capable of recording, with an Asahi Spectra XRR0340 filter and a strong UV source, useful amounts of wavelengths as low as 305-315 nm.

Recommendations

It should be remembered that the lens specimens most likely to work well in UV photography are the earliest produced ones. Thus, my summary of recommendations to improve your chances of obtaining a good 35 mm UV lens is:

  • First of all, look only for 35 mm f/3.5 models. The more modern 35 mm f/2.8 and faster models are likely unsuitable.
  • Look for any of the known-good models discussed above. If this fails, expand your search, but keep in mind the following points.Several of the early 35 mm f/3.5 lenses seem to come from the same factory as the Kyoei/Kuribayashi.
  • Look for models with manual aperture or manual preset aperture. Lenses with manual aperture have an aperture ring and no aperture transmission from the camera body. Manual preset apertures are characterized by the presence of two aperture rings, or an aperture ring and an additional open/close ring, and likewise no aperture actuator controlled by the camera body. These rings are usually (but not always) mounted near the front of the lens. Lenses with manual or preset apertures are older and therefore far more likely to be suitable than models with automatic apertures.
  • Golden lens coatings and bluish lens coatings seem to be equivalent, in spite of my expectations. Multicoated lenses that display reflections of multiple colors from their front and internal surfaces are the least suitable in this context. Non-coated lenses might be good, but none of the 35 mm lenses I inspected is completely devoid of coatings.
  • Don't bother trying lenses from the Topcor UV series. It is specifically designed to block UV, and therefore is useless for UV photography.
  • Look for models with a 23 mm diameter of the front element. This is the diameter of the glass as visible at the front of the lens, not the diameter of the lens barrel or filter mount. Models with 31.6 mm or larger front elements are different and more recent lens designs, which in the tested lenses display a more limited UV transmission.
  • Look for models equipped with a 46 mm filter mount. All the known-good models seem to use this filter mount. More recent, less suitable models like the Photax tend to use 49 mm or wider filter mounts. Models with 46 mm filter mounts also have shorter barrels than the more modern versions, and a shorter distance between front and rear optical surfaces.
  • Models in M42 mounts, or old bayonet mounts like Exakta, are more likely to be suitable than more modern mounts like Olympus OM or Pentax K. Remember that, ideally, you want a lens manufactured in the 1950s or 1960s. However, watch for damage to the rear element or its coating in lenses with M42 mounts, because the rear element projects far at the back of the M42 lens mount. Placing the lens on a table without a suitable rear cap touches the table surface with the rear lens element. I have also seen damage to the rear element caused by using too shallow a rear lens cap. Exakta and other mounts with a lower registration distance do not have this problem.
  • Models in M39 or L39 mounts, or other rangefinder mounts, may also be interesting to test on mirrorless cameras. Kyoei made at least two different models of these that use the same optical scheme of the lenses tested above.
  • Models equipped with a built-in T2 lens mount and an adapter friction-mounted with three set screws (usually T2 on the lens side and M42 or Exakta on the camera side) may be more likely to be suitable than models with non-interchangeable lens mounts. However, some of the highly desirable early Kyoei etc. models may have a non-removable mount, or a removable mount attached with a non-standard thread.
  • The large majority of brands and models of 35 mm f/3.5 lenses are, most likely, unsuitable for UV photography. Until proven otherwise, I recommend to avoid all other unknown/undocumented brands and models, unless you already own them or can borrow them for testing without committing to a purchase. According to Klaus Schmitt, who tested over 100 legacy, general-purpose lens models for UV transmission, only 3 of his tested lenses turned out to be usable for this purpose, so the odds are against your finding another usable model by blindly purchasing 35 mm f/3.5 lenses to test.
  • In addition to the above recommendations, follow all the general advice and common sense for buying second-hand lenses:
    • Select specimens without visible or described faults that will affect the intended use of the lens, and ask the vendor specific information if in doubt, especially when buying online.
    • Do not hesitate to ask about molds, scratches and/or damage to the lens coatings, stuck or sticky aperture blades, irregularly shaped aperture openings, rusty aperture blades or hard-to-turn focusing rings if the vendor does not say anything specific about the state of these lens parts.
    • If possible, select a vendor with a good and extensive feedback history of selling second-hand camera lenses.
    • As usual, a small amount of internal dust in a second-hand lens is normal, and causes no image degradation.
    • A couple of small scratches or cleaning marks are likewise unimportant to image quality, but usually warrant a lower lens price.
    • Lenses with optical surfaces that have been polished by hand to remove or hide small scratches are instead completely useless (just test one of these lenses and you will believe me). They must be returned and a complete refund obtained. A vendor who is knowingly trying to sell you this type of item is clearly intent on committing a fraud, and should be treated accordingly.
    • The quality of the description of an item in an online ad and the vendor's replies to your questions will also give you an idea of whether the seller is knowledgeable or knows nothing of/cares nothing for what he/she is selling.

If choosing a lens suitable for UV photography based on the above criteria sounds difficult and risky, to a certain extent I agree with you. I can only try to give my best advice, but I cannot give any assurances or warranties that the above recommendations will lead to a foolproof purchase. The high number of existing models and their varying degrees of suitability for UV photography imply that some risk will always be involved. Ultimately, do not purchase anything that you cannot afford to throw away or resell at a lower price if it should turn out to be unusable for UV photography.

Summary

Several models of 35 mm f/3.5 lenses are usable in UV photography down to 325 nm if losses of up to 3 stops are acceptable, or 340 nm if an approximately linear response is required. Thus, this lens type currently seems to be the most promising, in terms of availability and low prices, among all lenses for UV photography. This focal length is very suitable for use with Micro 4/3 mirrorless camera. The UV Rodagon 60 mm confirms in this test its excellent performance down to 325 nm. The EL Nikkor 63 mm f/3.5, instead, ranked as one of the worst among the tested lenses for UV photography.