Chapter 6
Glass Bases

This Chapter discusses Archertypes, lantern slides, crystoleums, sphereotypes, and types of flat glass used in early photography, plus a description of "weeping glass".

Negatives on glass
Talbot's negative-positive calotype paper process was clearly a conceptual improvement over the Daguerreotype because it permitted multiple reproductions, but the texture of the paper fibers limited the sharpness of the finished picture. LeGray's wax impregnation of the negative helped reduce this texture, but still the paper was translucent, where complete transparency was wanted. Satisfactory transparent flexible films were not made until late in the 19th century, but glass was available much earlier in virtually any desired size. In 1858 John Kibble in Scotland made plates 36x44 inches in size. A camera named "The Mammoth" was built in Chicago in 1900 that used plates 4 1/2 by 8 feet; the loaded plate holder weighed 500 pounds according to Gernsheim.

The trouble was that glass could not simply be coated with a water solution of silver nitrate: it rubbed off when dry. Paper, on the other hand, retained silver nitrate when it was soaked in a solution, and the nitrate could then be converted to the more sensitive and water-insoluble chloride. The resulting image had an embedded appearance that today helps to identify the process.

A multitude of inventors experimented with coatings and binders on glass. A good coating had to be sufficiently durable to stick without peeling while going through various chemical baths; it had to be permanently transparent; and it had to be chemically compatible with the light sensitive ingredients. The most successful coatings turned out to be gelatin, collodion, and albumen (egg white). The first use of glass in quantity for photography was for the wet plate collodion process invented by Archer in 1854.

Rempel [124] discusses tests for identifying various coatings, and his work should be consulted for details. The tests are destructive but can be performed on very small regions under a microscope. Essentially they depend on the fact that water swells gelatin but not collodion, while ethyl alcohol dissolves collodion but not gelatin. Albumen is unaffected by either solvent. Infrared spectrophotometry is a non-destructive but more expensive analytical process that is quite reliable.

After the apparent solution of the adherence problem it soon became apparent that these coatings had more subtle shortcomings: low and erratic sensitivity to light continued to be a persistent difficulty. The Archertype, or wet plate collodion process, was far more sensitive than early dry albumen or gelatin emulsions, but it was clumsy. The sensitivity was so fleeting that the plates had to be exposed and processed within no more than ten minutes after coating, literally wet. One theory was that dried collodion prevented diffusion of processing chemicals to the silver. However, tintypes used dry collodion emulsion with no processing difficulty, so the problem was complex. The wet plate process survived for more than two decades because it took that long for a dry plate to be invented that approached or surpassed the sensitivity of wet collodion.

Besides George Eastman, other inventors were at work on the dry plate problem. Eder (48) describes a number of these experiments. Dry plates began to be marketed by various inventors in the 1870's; Eastman's plates appeared about 1880. An interesting sidelight on this work is that twin brothers in Maine, Frelan and Francis Stanley, manufactured successful dry plates until Eastman bought them out. They used the money to start an automobile company, making the Stanley Steamer.

A great amount of trial and error was expended to find a preservative that would slow the drying and prolong the sensitivity of collodion negatives. Some of the experimental preservatives that were concocted were more ingenuous than ingenious, as Gernsheim has recounted (61, 324): he called it "the culinary period of photography." Preservatives included caramel, camphor, coffee, gin and water, ginger wine, glycerine, honey, Iceland moss, lager beer, laudanum, liquorice, malt, magnesium nitrate, milk, morphine, morphine nitrate, nux vomica, raisin syrup, raspberry syrup, salicine, sherry, sugar, tannin, tea, tobacco (several brands), treacle, vinegar, whey, wormwood, and zinc nitrate. Whiskey was not listed in any of the four references that were consulted, an unexpected and mystifying absence. Perhaps it went into the photographer instead of the coating mixture.

Serendipity had its place, too. It is now known that some of these organic mixtures have the property of promoting the formation of organometallic complexes and colloids, with results that conceivably did benefit the photographic process. It is worth reflecting that a century from now some of our own efforts might fare no better in history's judgment.

In recent years some workers have reported on their use of modern analytical methods to investigate the composition of historic pictures for dating purposes (see the Bibliography of modern scientific studies.) Infrared and ultraviolet spectrophotometry and x-ray fluorescence are useful non-destructive analytical techniques, but interpretation of results can encounter formidable problems when the above list of "preservatives" is considered.

Collodion-based sensitive layers were used in three applications:

1) Glass negatives, described above as Archertypes.

2) Collodion-coated paper, late in the 19th century.

3) Tintypes.

The sensitivity-stability problem existed mainly in connection with glass negatives. Collodion-coated paper could easily be given whatever exposure was needed in the darkroom. Tintypes were less affected than Archertypes for reasons that are discussed in Chapter 7, basically having to do with the superior speed of short focal length lenses.

Generally speaking, photographic plates and papers were coated on only one side, with the exception of very early salt prints. Coating both sides by dipping was easy, but it not only doubled material costs, it also produced out-of-register ghost images from the back side. Coating machines were put into production in the latter part of the 19th century, and manufactured dry plates (mostly gelatin silver bromide) can be recognized by their uniformity in thickness compared with the hand-coated product. Collodion plates were hand-coated by the user at the time of use, and film thickness often varied at the edges because of uneven drainage, and the fact that collodion would not adhere to as-cut edges (scored and broken). The edges of collodion plates were therefore usually roughened or polished, which also reduced handling injuries. They were often salvaged and reused several times to save cost. Plate thickness was not standardized, but they were considerably thicker than the dry plates introduced in the 1880's, which usually had as-cut edges.

Hand coated plates often contained blisters and occluded dirt particles; at the factory such defective plates were (usually) discarded. Sometimes the glass showed faint markings caused by the factory practice of marking lot numbers with soap; the alkaline soap slightly etched the glass, preventing collodion adherence, and the marks could only be removed by abrasive polishing. There were probably more flaws in the collodion coating on average than in the glass.

Visual Appearance of Emulsions
As Gill [67] and Rempel [124] have described, observation of reflected and transmitted light from images on glass can often differentiate between the emulsion types. Collodion is creamy or milky by reflection and a neutral black by transmission.
Gelatin-silver images are neutral black by both transmitted and reflected light. Woodburytypes are usually brown in transmitted light and dark by reflection. Carbon transfer prints were pigmented with many colors, which show by transmission.
Hand-tinted colors can cause confusion, but some areas were fortunately left clear, so the basic appearance of the medium can be observed.

Albumen on glass was tried as early as 1847 but because of low sensitivity it was seldom used commercially for negatives in spite of its popularity for paper prints. It was used on glass as positives in several forms, described below. It has a creamy appearance by reflected light, black and white by transmitted light.

Positives on glass
Glass was a natural base for lantern slides, which had already found some vogue with hand-painted images. The Langenheim brothers of Philadelphia patented photographic albumen glass lantern slides in 1850 under the name Hyalotype. They are brown by transmitted light, milky by reflected light, and survivors are somewhat rare. Woodburytype, carbon, and collodion transparencies were also made for lantern slides. They are difficult to distinguish visually from each other. Woodburytypes and carbon positives, like Hyalotypes, are usually brown, but they have a dark reflection rather than milky.

Collodion negatives on glass were the basis for ambrotypes, as discussed in Chapter 7. Collodion positives were sometimes printed on opal glass, also known as milk glass by some collectors and dealers. Opal glass contains colloidal crystallites, usually sodium or lithium fluorides, that scatter light and produce a pleasing translucent white color. Opal glass superficially resembles ivory, but collodion portraits were not made by the same process as ivorytypes or Eburneums (see Chapter 9). Collodion portraits on opal glass were often vignetted, framed, and tinted. It bears repeating that collodion prints on opal glass are not "opal ambrotypes", as we have seen at least one specimen mislabeled. They are positive collodion prints on a white glass, whereas ambrotypes are negative collodion prints on clear glass against a black backing.

Crystoleum
The crystoleum was representative of several types of decorative pictures on glass. An albumen print was glued to the inner side of a slightly curved glass, and the paper was removed by soaking, leaving the transparent albumen image on the glass. The image was tinted with oil colors and sealed with wax. A second curved glass was tinted with broad expanses of color and mounted behind the image; the two glasses were bound together with a separator to give a three dimensional effect. Details of the process are given in Cassell's [84, 154-5].

The sphereotype, patented by Albert Bisbee in 1856, was made somewhat similarly on the bottom of curved paperweights. The spherical glass acted as a magnifier. Other similar processes were the diaphanotype, the ectograph, and the opalotype (see Chapter 14, Section 3, for references). Some were transfer processes, others direct printing, and their classification is somewhat arbitrary. See also Chapter 7 for further information on variations of ambrotypes, and Chapter 9 for information on transferotypes.

Notes on the History of Flat Glass

19th century photography was one of a growing number of new industries that demanded better raw materials. Photography soon exerted sufficient commercial leverage to bring about improvements in paper making (see Chapter 1). Better and cheaper glass was also needed with the advent of Archer's wet collodion process in 1854.

Good quality flat glass was difficult to make in the nineteenth century. "Good quality" means flat parallel surfaces, uniform thickness, smooth grainless surfaces, neutral coloration, freedom from pits, stones, bubbles, and striae; and all at the lowest cost, naturally.

In the nineteenth century there were two principal glass compositions: lead and lime glass. Both were used in photography; lead glass was heavier and more expensive, but because of its early availability as plate glass, it was used for wet plates. Later in the century it was phased out in favor of lime glass, which had been made as early as 1864. The quality gradually improved so that it could be used for window glass without grinding.

The commonest chemical impurity in the glass was iron, producing a green color that did not bother negative processes but caused unpleasant effects in glass positives. Only 500 parts per million of iron will give window glass a green color that can be seen through the edges (optical glass is permitted only 10 PPM or less). Lead glass is dark in edge viewing, while lime glass commonly shows a green tint.

A booklet from the Corning Museum Of Glass, reference I-12, states "At the beginning of the 20th century, there was no way to mass-produce flat glass". Several methods of making flat glass were in use in the 19th century, each with its own peculiarities:

1. Cast glass:
One of the oldest ways of making flat glass was to cast molten glass and then roll to thickness on flat iron tables (molten glass does not stick to iron or carbon except at red heat, so these materials are used for tools.) The bottom surface was always optically spoiled by contact with the casting surface, and ripples and striae were common. Grinding and polishing the contact surface made a good product ("plate" glass) for windows and mirrors, but it was expensive and reserved for those who could afford it. There is a diamond-polished mirror on display in President James Monroe's home from early in the 19th century.

Steam power was used for grinding plate glass as early as 1789. The surfaces were seldom as brilliant (grainless) as fire polished surfaces because of the particle size of polishing media, and body flaws were common, especially in larger sizes. Grinding the two sides was done separately until 1937 when a twin grinder was developed in England that ground both sides simultaneously.

2. Blown cylinder glass:
This method consisted of blowing as long a cylinder of molten glass as possible; after cooling, the ends were cut off and the cylinder was scribed lengthwise. When reheated in a furnace, the cylinder opened and sagged flat on a table. Only the inner surface remained fire polished; the outer surface was somewhat deteriorated by contact with the table, but not as seriously as table-cast glass. A specimen of an uncut cylinder, about eight feet long, is on display at the Corning Museum Of Glass, Corning, New York.

This process was not mechanized until early in the 20th century. The product was wavy but tolerably good for windows: the Crystal Palace, built in England in 1851, used 300,000 panes of cylinder-blown glass four feet long. At first it did not make a very good negative photographic base without grinding, but selected pieces were occasionally used because it was cheaper than ground and polished glass. Apparently the quality improved in the 1870's, in time for the gelatin dry plate.

3. Disc glass:
This was an old process consisting of blowing a glob of glass into a sphere, opening the end, and spinning rapidly while molten. Centrifugal force could form a disc as large as a meter in diameter, which even today is a tricky manual operation. It was not suitable for photography without grinding because cut pieces did not have parallel surfaces, and the surfaces were marked with concentric ridges. It was used for small windows such as leaded diamond panes, where the defects were less apparent and even attractive.

The thick center from which the disc formed during rotation was called the "crown". Crown glass was scrap and was used for lens-making when big enough pieces could be found that were not too bad in quality. Crown and flint glasses were used together in compound lenses to correct some lens aberrations. Flint glass contained lead oxide, while crown glass did not, and the refractive indices and optical dispersion of the two glasses were substantially different. Lens formulas increasingly made use of these properties in compound lenses to meet the demand for better photographic sharpness.

According to Archer [2], crown glass was sometimes flattened by melting to produce sheet glass, actually a form of casting with its characteristic defects on one side.

Two other kinds of glass have had some photographic use as light diffusers. One is commonly called "ground glass", used as viewing screens in view cameras. It is usually made by sand blasting or fluoride etching. The other is "opal" glass, and is used as a light diffuser in enlargers and other light sources. It consists of a thin layer of opal glass fused to a base of clear glass. The opal layer is thinner than a solid piece of opal glass and therefore has less light loss, while the clear glass provides strength for the thin layer. Neither of these two types appear to have been used as photographic image bases. However, solid opal glass was sometimes used as a substrate, as mentioned above. The composition of opal glasses is discussed in Scholes (A - 326).

Modern flat glass is made by several methods: grinding and polishing cast glass, continuous vertical drawing of sheets, and floating on molten tin; the latter currently dominates the window-glass industry. The quality is so good that grinding is not necessary for most applications.

There are also many specialized methods to meet modern requirements. One example is the very thin, optically perfect sheet glass used for screens in laptop computers such as the one on which I am typing these words. Reference I-22 describes the fusion draw process used for this type of sheet glass.

Wet Plates and Dry Plates

When Archer invented the wet plate collodion negative in 1851, the best available glass was polished plate glass. It was usually lead glass; later in the century, lime glass supplanted lead glass because it was cheaper and lighter. By this time lime glass was universally used for ordinary window glass.

I have not found reliable information on the sources of glass for gelatin silver dry plates, so the following remarks are speculative. In the context of the technology, the most likely source was soda lime cylinder glass, selected for uniform thickness within lots, and minimum waviness. It seems unlikely that it was ground and polished because of cost and industrial capacity; the fact that the plates had as-cut edges argues for cost constraints even in early days of factory production. Slight variations in thickness would probably have been tolerated at a time when attention was concentrated on the sensitivity question.

Weeping Glass

This is a term that has been given [Ref 152] to destructive deterioration of glass under certain storage conditions. It is irreversible and may completely ruin glass photographic plates, even in archival storage. The explanation is necessarily technical, but understanding may help save some valuable plates.

It manifests itself as a sticky wet coating on the glass surface (not the emulsion side) in an apparently dry room. The coating may remain wet in room environment. If the glass is washed in clean water and dried, the coating will be gone but the glass will appear frosted or etched. A photographic plate will be hazy, and a good clear print cannot be obtained from it, nor can the original clarity be restored by chemical treatment.

It can occur in archival storage if the environment undergoes a temporary excursion of high humidity, such as might happen if the air conditioning fails, or a sprinkler system nearby is energized, or the roof leaks. The restoration of normal conditions may not save the day if the damage has been done, and once started, it can continue to progress under benign storage. The combination of circumstances causing the condition are fortunately rather uncommon, but it can occur in climate-controlled archival storage that is usually considered safe. This writer has seen it happen.

The chemistry of the problem is well described in Scholes [A-408]. Werner [152] has a similar discussion. Glass can be attacked by water but most glasses are not water soluble. If a thin film of water is allowed to condense on glass and remain, hydrogen ions diffuse into the glass, displacing sodium ions. This sodium diffuses into the water, forming a solution of sodium hydroxide. If the body of water is small (such as a thin film of condensate), the sodium hydroxide may become quite concentrated with a high pH. Such an alkaline solution rapidly etches the glass, destroying the Si-O bonds, and does not readily evaporate to dryness at room temperature. It feels wet and "soapy" to the touch, and the etching is progressive and irreversible. A concentrated electrolyte of this kind has a reduced vapor pressure and low evaporation rate at room temperature, so its drying rate is much reduced.

The buildup of a concentrated solution of sodium hydroxide requires a thin undisturbed film of water. The time scale depends on temperature and film thickness, but damage can occur in a few hours. Etching is more likely to take place on the reverse side of photographic plates rather than the emulsion side, although water swells gelatin emulsion and affects its optical properties.

Soda lime glass is particularly susceptible, which was used for gelatin dry plate negatives rather than the heavier and more expensive lead glass. Of course, glass photographic plates can withstand darkroom chemical processing with no observable change. Glass is a durable and ubiquitous material, evidenced by long service in windows and other objects. But window glass may exhibit faint cloudiness after many years of weathering, and other glass objects stored in a damp environment can deteriorate. Antique glass vessels often show interior cloudiness; it is sometimes mistaken for calcium deposits. Acetic acid will remove calcium deposits but it has no effect on water-damaged glass.

The conditions conducive to the formation of a film of condensed water on archival photographic glass plates are fortunately uncommon. But this writer has seen storage racks in two modern museum archives draped with sheet plastic because of roof leaks, for periods of days or weeks. When a water problem is present, archivists may be more concerned about the threat to paper artifacts than to glass plates, because glass is considered to be "waterproof". Archival storage is usually thought to be safe and secure, but eternal vigilance is necessary to avoid false security.