Are dirty lenses faster?
OK, "dirty" is a bit strong. Back over 100 years ago, words like "hazy" and "tarnished" were used. However, my point is this: It may sound paradoxical but old uncoated lenses actually became faster as they got older and more weathered, about one full f-stop faster. How can this be? The answer led to the Nobel Prize of 1904 for Lord Rayleigh, a patent for H. Dennis Taylor for the first lens coating that same year, inappropriate credit going to Zeiss for having invented lens coatings in 1936, and eventually to the world of coated lenses on your new DSLR.
Comments on Lens Coatings and History
Lord Rayleigh:
John William Strutt, known to history as Lord Rayleigh, was a physicist in the 19th century who eventually received the Nobel Prize in 1904. Lord Rayleigh made a fascinating observation the relates to photography and lenses. He found that old noticeably foggy or "tarnished" camera lenses performed better than new clean ones in terms of their speed. The older and more hazy the lens, the more efficient it was in gathering light, as much as one full f-stop faster. This basic observation is easily explainable now but it was quite paradoxical then. The clean glass of new lenses have significantly higher reflectance than old lenses since the old lenses are coated with oxides and "dirt" which act as anti reflectance coatings in much the same was as modern thin film coatings. When light hit the surface of old lenses, it was significantly more likely to hit the thin film of oxides first rather than the actual glass surface. The result is that the light will encounter the interface of the coating and glass where the reflectance is significantly lower than at an air-glass interface. The consequence is an increase in light transmission which is an increase in lens speed. Much of the later work on thin film coatings of optical glass was based upon Lord Rayleigh's simple original observation.
H. Dennis Taylor:
The earliest lens coating appears to be that first described by H. Dennis Taylor some time around 1896. As Rudolf Kingslake comments (see: Kingslake, R. A History of the Photographic Lens. Academic Press, 1989. pp. 16,17). Taylor described how old "tarnished" lenses performed better than newly polished ones and he suggested that this was due an altered glass surface that had a lower refractive index. In 1904 Taylor received a British Patent on the use of chemicals to create a "tarnish" on new lenses to improve performance (see British Patent 29561, 1904). Although the process of chemically tarnishing optical glass worked well in principle, in practice it was not sufficiently uniform and reproducible for mass production of lenses. It is not clear to me how Taylor's observation was related to that attributed to Strutt in 1886, though they appear to be concerned with the same phenomenon.
Alexander Smakula:
For almost 40 years it had been known that a thin non-reflective coating on lenses could markedly improved their performance. However, Taylor's first patent in 1904 described a process of lens coating that proved to be impractical. In 1935 while working for Zeiss, a Ukrainian physicist named Alexander Smakula invented the first commercially successful method for coating optical lenses. He is occasionally credited with having invented optical coatings which is simply not true. However, he did invent the first practical method for creating non-reflecting optical coatings on lenses for the consumer market.
*****************************
First, all glass surfaces reflect some light; lens coatings can effectively reduce stray light from reaching the film by reducing the reflectance of glass surfaces and thus increase the contrast of images. The reflectance of light is a characteristic of the type of glass itself and is described by both the color of the light (i.e. wavelength) and the angle at which it strikes the glass (i.e. incidence angle). It is also a function of the type of materials at each interface, that is, air-glass, glass-glass, or coating-glass, and each type can produce different degrees of reflection. Reflectance can reduce the amount of light that passes through to the film and appears to reduce contrast in the image. However, this is not due to reflected light that goes back into the environment. We are talking about light that reflects off the multiple surfaces of the elements within a lens housing. Anti-reflectance coatings first developed during WWII reduce reflectance by introducing a very thin layer of material that causes a portion of the reflected light to self-destruct, literally destructive interference. However, as first described by Rayleigh well over 100 years ago, the interface between the coating and the glass reflects less light than the air-glass interface. Inside of the lens housing, less light bounces off of multiple surfaces to re-enter the glass as stray light. However, less light reflects from the coating-glass interface and more light actually enters the glass as a consequence. The net effect is to increase the total amount of light that reaches the film as well as to reduce the amount of stray light that reduces contrast.
In general, there is no such thing as "color coating" a lens per se. Color in the transmitted light and color accuracy of the image is more a function of the lens design and less the nature of the thin film coating within the limits of our eyes to detect such color accuracy, though one cannot completely separate one from the other. Although the efficiency of anti-reflectance coatings is effected by the wavelength of light, such coatings are normally selected to increase contrast in the visible range as recorded by B&W or color films. This is related to the fact that coated lenses appear to be purplish, indicating that the rest of the visible spectrum of reflected light has been disproportionately reduced by destructive interference within the thin film of coating. At the same time, the coated lenses do not have a tinted appearance since the transmitted spectrum of light hasn't changed very much. Where did the light go that produced the purple tinge to the reflected light? The answer is that it was destroyed as soon as it reflected back from the coated surface.
Addition of a thin film anti-reflection coating on the air surfaces of lens elements will increase the apparent contrast, perceived brightness, and the spatial resolution of the resultant image at the film film plane (please note that I am not referring to the color sensitivity of the film or the spatial resolution that the film is capable of producing). However, this apparent increase in color brightness and contrast is primarily a function of the multi-element lens design in terms of selection of the type of glass, the curvature of the lens elements, and how these multiple elements may be assembled in groups, cemented or left with air gaps between elements. Coating optimize but do not appreciably change the performance of the lens design.
That being said, optical coatings to reduce reflectance were certainly not equally effective across the color spectrum. In fact, the principle on which anti-reflecting coatings are based is described by destructive interference of reflected light from the glass surface. Lens coatings were originally composed of a single layer thin film such as calcium or magnesium fluorides which were selected because they are effective within the visible spectrum. However, at either extreme, that is red or violet, the reflectance may appear to be increased giving the lens surface a purplish hue.
So, are hazy lenses comparable to coated lenses? Not exactly. Hazy lenses may be faster than perfectly clean lenses, but they produce an image that has reduced contrast. Since the purpose of photographic lenses is to produce more or less sharp images, and not to simply gather as much light as possible, producing a hazy lens by coating with very fine particles is not a practical technique for improving lens design. However, the phenomenon did lead directly to the study of thin film coatings of lenses and soon to the development of the modern coated lens used today in all camera lenses.
The New Coated Lenses and Marketing
Color photography was most actively developed in the 1930's primarily for professional and amateur consumer products. At the same time, lens designs had advanced to a point where multiple elements began to introduce serious limitations on lens performance due to unavoidable internal reflections. This made it impossible to derive the full benefit of the lens designs which were aimed at increasing the resolution and speed of the lenses. As the design of the glass optical elements made it possible to correct for color and minimize aberrations, the design engineers had to contend with the behavior of light at the interface of air and glass and the fact that some of the light is reflected from the glass surface (i.e. about 3-4%). Also note that when light strikes glass at very high angles, it is totally reflected (called the "grazing angle") and may then bounce about striking and scattering off lens mounts and the ground surfaces at the circumference of the lens. All of this uncontrolled light may eventually hit the film and effectively reduce the image quality. However, it took more than a decade of research in the 1930's and 40's to find a thin coating material that functioned as an anti reflective surface but also was sufficiently hard to resist scratching and being simply rubbed off the glass during routine cleaning.
For instance, Kodak's early magnesium fluoride coating on the Ektar lenses from the late 1930's and through the 1940's was sufficiently hard to resist damage but it was used primarily on the outer glass surfaces. The inner surfaces were coated with the older calcium fluoride coating which is soft and easily damaged is a lens is dismantled for cleaning today. However, Kodaks "Lumenized" coating was a significant technological step forward and paved the way for further development of more advanced multi-element lens designs during the 1950's (see Kodak Ektra 1941).
Since lens coatings reduce reflectance at the air-glass interface, they effectively increase the amount of light that passes through the lens to reach the film. When lenses were designed, that is when the type of glass and the precise curvatures of the elements were calculated, a lens was assigned a certain light gathering efficiency which is normally called the "lens speed" and is recognized by the minimum f stop. The speed of a camera lens was determined by several factors, most importantly, the maximum aperture that produced minimal aberrations. Historically, the revered Rapid Rectilinear lens often found on early cameras was designed to flatten the image but it did so at the expense of introducing considerable astigmatism. Since most lenses perform best toward the center and away from the lens periphery, you will notice that old RR lenses are rather large yet still have limited speed and the RR lenses were limited to about f/8 (i.e. the ratio of the focal length of the lens to the maximum aperture of the diaphragm). In this instance, f/8 was the maximum useful aperture of the RR lens and some photographers in the early days would remove or widen the Waterhouse Stop or aperture in order to increase the speed of the lens, all at the expense of the image quality.
This gave rise to the question of what is the "true" speed of a lens and during the early days this was not at all clear. However, curiously Clarus Camera Mfg. in their manual published in 1947 touted their Clarar coating as effectively increasing the lens speed (for more information about Clarus cameras see Clarus Camera Mfg. Co. 1946-52). Clarus obtained their lenses from Wollensak and trademarked the coatings put on the lenses for a time. They point out in their manual for their MS-35 model (their only model) that the optical coating increased the speed of their lenses as shown here (right). Thus, an f/2 lens is claimed to have a true speed of f/ 1.75 which translates into an increase in light gathering capacity of over 30% due to coating alone. No other lens manufacturer of the time or since has pointed out that fact but is it true? What these marketing people of 1946 may have meant was their coated f/2 were equivalent to f/1.75 lenses that were not coated. Based on the Rayleigh effect described above, coating can theoretically increase the total amount of light gathered by a lens. The question is whether and to what degree do such coatings did so in practice?
The Lowly Meniscus Lens and some misconceptions:
Let me begin by saying that the term "meniscus lens" is not a cute name invented by manufacturers to disguise the fact that their camera has a cheap lens on it. The word meniscus comes from the Greek word meniskos which refers to the "lunar crescent". The term "meniscus lens" came into common in use in the early 19th century when it was found that a convergent convex-concave lens produced a relative distortion-free image compared to that produced by bi-concave lenses when used in a camera obscura. This was important since the camera obscura was used primarily for drawings, especially drawings of architecture where distorted images had been a serious problem.
However, the term "meniscus lens" appeared 200 years earlier in the writings of Stephen Gray in the 17th century (if you are really serious, try Philosophical Transactions, Volume 4, 1694-1702, pp. 97-101). Gray was a noted physicist who described the behavior of a convex-concave lens formed with a droplet of water placed and held in a very small metal ring. This water drop lens is necessarily a convergent lens, the type used to form images at a distant plane. A "meniscus" commonly refers to the curvature formed by liquids in contact with solids such as at the edge of a container, or the curved edges of water in a narrow tube, the degree and direction of the curvature depends upon the amount of molecular attraction of the liquid to the solid. Hydrophilic surfaces attract water and produce a concave curvature of the water. Whereas, hydrophobic surfaces repel water and produce the opposite curvature.
In Gray's experiments in the 17th century, a true meniscus lens was formed by a combination of factors: the surface tension of the water itself which held the droplet in the ring, the weight of the water droplet relative to the diameter of the ring, and the attraction of the water molecules to the metal surface of the ring (there are others but they make less important contributions to the phenomenon). The result was a lens made of water that could be used as a simple microscope since the two surfaces of the water droplet had different curvatures thus forming a convergent (i.e. positive) lens. When simple lenses were formed using solids such as resins, crystals or glass instead of a liquid, the curvature of the two opposing surfaces could be varied independently to produce either positive or negative lenses; the term meniscus lens has been commonly applied to all variations of this simple design becoming somewhat of a metaphor.
I have seen several rather over wrought interpretations of the term "meniscus lens" including the claim that a "true" meniscus lens cannot even form an image. It is always risky to speculate on the etymology of technological terms in use today; after all the word "lens" could be restricted to "bean-shaped" lenses since the word came from lens, lentis in Latin. So it is not correct to say that a "true" meniscus lens has the same curvature on both surfaces since that would be simply a glass dome (meaning "housetop" in ancient Greek) and a dome would not have been possible using the phenomenon of the liquid meniscus (i.e. put more simply: you can't make a water droplet lens with the same curvature on both surfaces). The only form of the meniscus lens that is relevant to photography is that of the positive or convergent lens since divergent lenses do not form images - nor is the name meniscus lens just a marketing ploy for hiding the fact that a camera has a cheap lens as I recently read on an Internet site populated by "experts".
Of course, it is a fact that the cheapest vintage cameras usually have simple meniscus lenses or what was called "doublet" lenses, but are they to be disregarded as cheap or do they in fact represent major historical advances in photographic lens design? Back in 1812 the meniscus lens was very significant advance in optics. It was known as the landscape lens because it was most effectively used for bright landscapes and not close portraiture. Since a meniscus lens produced so much spherical aberration, the maximum useful aperture was f/11 and the maximum useful field is 40 degrees off axis because of the terrible astigmatism. These shortcomings are less apparent in long daylight landscape photos.
Eventually, the use of the cemented doublet improved design of the basic meniscus lens reduced lateral chromatism, especially for the yellow and blue regions of the spectrum, useful when the photographic emulsions of the 19th and early 20th century had their peak sensitivity in this region. Since the meniscus lens is relatively free of distortion, and if the world were simply B&W, then you would probably have a meniscus lens on your new Leica. Right now, though, you're still likely have one in your cellphone camera.
One more note related to vintage cameras. You will notice that some old cameras such as the Kodak Vest Pocket cameras with meniscus lenses, have the diaphragm or shutter located in front so when the shutter is closed, the lens cannot be seen. When the shutter or aperture is placed in front of a meniscus lens, the result is barrel distortion of the image. Alternatively, when it is placed behind the lens as with some other cheap cameras, the result is pincushion distortion. Kodak engineers decided that barrel distortion was less offensive than pincushion and less persceptible when one looked at the photographic images. The Holga, however, has a plastic mensicus lens located in front of the aperture of which there are only two giving f/8 and f/11 (the typical limits of a meniscus lens), thus using only the center of the lens to form the image in order to minimize aberrations. The use of plastic in place of glass in the lens was not only cheaper to produce, but also somewhat improved the lens performance over glass.