Material Analysis: Introduction
The material characterization of an artwork proceeds from material analysis. The material analysis of the Thomas Walther Collection photographs elicited data about the format and fiber content of the pictures’ paper supports, the thickness of the photographs, and the elements present in their paper supports, baryta layers, and emulsions. The material analysis techniques used in this project were selected from a series of established methods commonly used on works of art, though many are relatively new to photography. They include procedures as simple as the tape-measurement of vertical and horizontal dimensions, indicating the length and width of the object, and as complex as X-ray fluorescence (XRF) analysis, which yields data indicating the work’s internal elemental makeup. The information acquired through the material analysis of the photographs deeply informs the research and analysis throughout the Object:Photo website. It is presented on the individual artwork pages in the Photograph Index and incorporated into the site’s powerful visualization tools. In addition, comprehensive XRF and paper thickness data is presented in this section, which also describes the material-analysis techniques used in the project, including detailed procedural protocols.
Height and Width Measurement for Paper Format
Commercial photographic paper is sold in square or rectangular sheets (or rolls) of various precut sizes. Between 1909 and 1949, two standards of measurement were in common use: the metric system in France, Germany, Italy, Austria, Russia, Hungary, Czechoslovakia, and Poland, and the imperial system in Great Britain and, adapted slightly, in the United States. Photographic papers were sold in formats based on both standards (in centimeters and inches, respectively), largely segregated by region. However, it was possible for artists in New York, for example, to buy photographic papers in metric units. Some American manufacturers, such as Kodak, sold their products domestically and overseas, and so produced papers in both units.
For this study, prints in the Thomas Walther Collection were classified into one of three format categories: “imperial,” “metric,” or “unidentified.” Kelly Cannon, Cataloguer in the Department of Photography at The Museum of Modern Art, measured the length of the four sides of each print in July 2013. Measurements were taken with a flexible metric tape measure made of fiberglass (to reduce the chance of stretching over time); the positioning of the mark on the tape measure (at, just inside, or just outside the paper edge) was determined at the outset and consistency was maintained throughout the project. Each print was measured once, including any border around the image, and the values were entered into the Museum’s cataloguing database, which automatically calculated the imperial equivalents. Where the length of opposite sides differed by more than 2 millimeters (1/16 inch), the notation “irreg.” was added to the measurements.
The next step was to assess how closely each print’s dimensions matched a standard metric or imperial format, as published in manufacturers’ marketing material from the period. The edges of all the prints were examined under magnification for signs of trimming. Where the paper was intact, the closest paper size, whether metric or imperial, was considered a match, and the print was classified accordingly. Where there was evidence of trimming, classification proceeded differently for contact prints and enlargements. Manufacturers produced papers designed to exactly match common negative sizes, and when making contact prints, photographers overwhelmingly selected the paper designed for the negative size they were printing; thus, even when a contact print (identified by a number of factors, including evidence of a negative holder in the image) showed signs of trimming, the closest paper size, whether metric or imperial, was usually considered a match, and the print was classified accordingly. Enlargements that showed signs of trimming were more often designated as unidentified. Where a print’s dimensions were very close to both a metric and an imperial format (7.5 by 10 centimeters and 3 by 4 inches, for example), the life history of the artist was considered as well as format trends in his or her work, if available. Where uncertainty remained, the format was designated as unidentified.
Micrometry with Caliper for Paper Thickness
Between 1909 and 1949, commercial photographic papers were sold in a variety of thicknesses (designated by weight or volume). Papers of different thicknesses were designed for different uses, and photographers chose among them for reasons both aesthetic and practical. Heavy papers might be stiff enough not to require mounting, and they were often selected by professional photographers seeking a luxury product for their clients; they were also chosen for their durability—by press photographers, for example, whose prints circulated among press agencies and editors. Thinner papers might curl after drying, but they were cheaper, they processed more quickly, and they took up less space, making them ideal for certain industrial and military uses, such as document and map copying. For this study, thickness measurements were taken of prints in the Thomas Walther Collection, and four thickness categories were created to classify them. The category specifications were informed by thickness categories listed in manufacturer’s marketing material from the period and consultation of reference collections of historic photographic papers.
Hanako Murata, Assistant Conservator of Photographs at The Museum of Modern Art, devised the protocol and measured the thickness of each print, using an L. S. Starrett Company electronic micrometer, model number 733FL-1, between 2010 and 2013. The thickness, measured in inches to five decimal points, was displayed on the micrometer’s screen then recorded manually in a spreadsheet.
The micrometer was secured on a tripod-base stand using a Castaloy adjustable-angle clamp and leveled using a bubble level. The measurement point of the micrometer was adjusted to the same height as the top surface of the weighted museum box that was placed beside it to support each print during the measurement process. This support was set 3/4 inch (1.9 centimeters) from the measurement point of the micrometer, to ensure that the edges of the prints could float freely during the measurement process and no pressure would be exerted on them by the built-up system, while the prints would not be susceptible to bending or creasing. A mirror set below the micrometer allowed for the observation of the verso of each print, so that locations with accretions, labels or hinges, or condition issues could be avoided during measurement, as these factors can affect the precision of the result.
The support was covered with blotter paper and topped with a 3 mil sheet of polyester. For measurement, each print was placed on this surface and covered with a second, identical polyester sheet. The polyester sheets, both of which were larger than the print, protected the delicate photographic surface from abrasion and ensured the safe and easy handling of the artwork. The 3 mil polyester was thick enough to protect the print surface from pressure by the metal disk of the micrometer. (Prior tests with thicker polyester sheets showed a wider range of standard deviation, and they were thus determined to be unsuitable for precision measurements.)
Three randomly selected locations on each print were each measured three times; the micrometer was zeroed before the first reading of each location. The nine values were entered into the spreadsheet, and standard deviation was calculated. If standard deviation was more than 0.0001 inches (0.0025 millimeters), the micrometer was zeroed and the print was measured again, following the above protocol. (For prints where a low standard deviation could not be achieved, see below.) The average of the nine readings was calculated, then the thickness of the two layers of 3 mil polyester was subtracted from the average. The resulting inch measurement was recorded to five decimal points in the spreadsheet, converted to millimeters, and, finally, rounded to three decimal points (with a margin of error of ± 0.003 millimeters).
Some prints showed a wide range in thickness across the three locations measured, and a standard deviation of less than 0.0001 inches could not be achieved. Such results were observed for thicker prints and for prints with cracks, cockling, and/or warping (and especially for thicker prints with these conditions). This suggests that thicker paper is more prone to distortion or is more likely, from point of manufacture, to have a wider range of thickness; if true, this could be considered a material characteristic itself, indicative of the original manufacturing process. In such cases, additional locations on the print were measured (the number of locations varying, case by case), following the protocol described above, in an attempt to find a more representative average print thickness. The result was recorded as the final measurement.
Not all the prints in the Walther Collection could be accurately measured. Those mounted to a secondary support or those whose entire recto was coated (with varnish or plastic laminate, for example) were not measured.
Fiber Analysis for Fiber Content and Pulping Method
The centuries-long history of paper making is marked by evolutions in fiber sources (from papyrus to linen to wood) and processing systems (from beating to cooking to chemical digesting, along with various bleaching methods) as well as the development of a plethora of specialty papers made with colorants, in special weights or formats, or with other physical or chemical modifications. Photographic papers are a small subset of this production, but their development over time mirrors industry-wide advancements. The analysis of a paper’s fiber content—determining the type and quantity of fibers present, as well as the manufacturing processes by which it was created—may provide information about its date and place of origin. A high degree of confidence can be achieved when fiber data is verified by comparison with data from reference collections of historical material samples of known date and origin.
For this study, a minute quantity of paper was extracted with microsurgical tools from the verso of prints in the Thomas Walther Collection. Not all prints in the collection could be tested; among those not sampled were prints mounted to secondary supports and prints whose entire versos were coated or covered (for example, with accretions, labels, or hinges) or subject to other condition issues that could contaminate the fiber sample. Lee Ann Daffner, Andrew W. Mellon Foundation Conservator of Photographs at The Museum of Modern Art, performed the sampling between March 2010 and March 2013. The entire procedure, including the location of the sampling on each work, was carefully documented, using the fiber worksheet developed for the project. Walther Rantenan of Integrated Paper Services in Appleton, Wisconsin, carried out fiber analysis following Technical Association of the Pulp and Paper Industry (TAPPI) Test Method T 401 om-03 for Fiber Analysis of Paper and Paperboard, identifying and counting the paper fibers in each sample and testing them to determine which pulping methods were used in its manufacture. The data collected in this study was compared with fiber data from two reference collections: the Paul Messier Historic Photographic Papers Collection, Boston, and the collection of historic European papers belonging to artist Alison Rossiter, New York.
X-ray Fluorescence Spectrometry for Elemental Identification and Location
X-ray fluorescence (XRF) spectrometry is one of the most useful and widespread analytical methods for the examination of photographs. In XRF spectrometry, a sample is bombarded with high-intensity X-radiation, which in turn stimulates the emission of low-intensity X-radiation (X-ray fluorescence) from the sample. Individual elements fluoresce at characteristic energy levels, and so an analysis of the energy spectrum emitted allows for the identification of the individual elements present in the sample. XRF spectrometry can be used to detect the presence of a wide range of elements, from low to high atomic number (from magnesium through uranium, with the instrument used in this study), in trace amounts or as major constituents. Each element is identified by a characteristic set of emission lines (peaks in the spectrum), and the intensity of these peaks is proportional to both the abundance of the element and the sensitivity of the technique for the particular element (sensitivity is reduced for lighter elements, which produce a weaker signal).
With a careful selection of instrument working conditions and areas on the photographs to examine, this non-invasive technique can simultaneously detect elements related to the image material (including silver, platinum, palladium, and inorganic compounds added to the gelatin emulsion, such as matting agents); elements indicative of the composition of the baryta layer (including barium sulfate, with varying amounts of strontium sulfate impurities or, in more recent papers, titanium oxide); and elements associated with the paper support (such as compounds added to the paper pulp for bulking, buffering, or homogenizing). An XRF spectrum acquired from the recto of a photographic print is better suited for the identification of image material and elements in the emulsion and baryta layers, while elements present in the paper support are best identified by acquiring a spectrum from the verso. While the majority of the elements typically encountered in photographic papers can be detected by obtaining a spectrum from the recto or the verso, a comparison of the relative intensity of peaks obtained from both readings can in some cases indicate whether an element is found in the emulsion, baryta layer, and/or paper support. On the other hand, comparison of spectra acquired from areas of maximum and minimum image density (Dmax and Dmin) on the recto may assist in distinguishing elements specifically related to the image material and toning agents. (For these elements, the intensity of the signal is directly related to the density of the image.) If a work is mounted and analysis from the verso is thus unviable, a spectrum acquired from the mount is required to help tease out the contribution of the mat to the measured signal. Likewise, any contribution from the instrument (rhodium, copper, and nickel, for example) can be identified by acquiring a background spectrum.
For this study, XRF analysis of the prints in the Thomas Walther Collection was conducted in the Department of Conservation at The Museum of Modern Art in 2011 and 2012, following a protocol developed by Ana Martins, Associate Research Scientist. Conservation Science Interns James Newhouse and Marc Vermeulen assisted Martins with the analysis. Analysis was conducted on the recto (Dmax and Dmin, no contact) and verso (paper or mount, usually in direct contact) of each print. The spectra were collected at 40 kiloelectron volts (keV) and 20 microamperes (µA) for 300 seconds using a Bruker Tracer III-SDD handheld XRF instrument with a rhodium excitation source and silicon drift detector, employing a helium purge to improve the detection of the lighter elements (magnesium, aluminum, and silicon). The XRF instrument was secured to an industrial-grade instrument tripod with rack-and-pinion X and Y axis adjustment.
Each print was placed on an acrylic sheet lined with Japanese paper (MMN-1 Tengucho 5g/m2 by Hiromi Paper, Inc.) and secured with one or two polyethylene straps (depending on print size). This mount was designed to position each photograph safely on the tabletop easel and in a vertical orientation for the analysis without introducing any interference in the spectra. The background spectra were also acquired with this support.
Ultraviolet Examination for the Presence of Optical Brightening Agents
The prints in the Thomas Walther Collection were examined under ultraviolet (UV) illumination to identify the presence of optical brightening agents in their photographic-paper supports. A characteristic blue-white fluorescence identifies the presence of these agents, which, in turn, suggests that the paper was fabricated after the early to mid-1950s. 
When possible, both the recto and verso of each print were examined; for prints mounted on secondary supports, only the recto surface was examined. The examination, in the Department of Conservation at The Museum of Modern Art, took place under long-wave (365 nanometer) UV illumination using a portable Q-22NF Spectroline lamp (115 volts, 60 hertz, 0.36 amperes). The lamp was held by hand approximately 8 inches (20 centimeters) from the surface, which was exposed to the UV light for 1.4 to 1.8 seconds. Researchers wore protective glasses during the examination.
A number of variations were observed in this survey, but the results were simplified to indicate whether fluorescence was observed (positive) or was not observed (negative), on the recto and verso of the prints, as applicable.
 Paul Messier, Valerie Baas, Diane Tafilowski, and Lauren Varga, “Optical Brightening Agents in Photographic Paper,” Journal of the American Institute for Conservation 44, no. 1 (2005): 1–12.