When an historic albumen photograph is seen in cross-section, using an environmental electron scanning microscope (ESEM), the nature of material can be seen. The albumen layer is sitting on paper (lower). It exhibits behavior characteristic to denatured, filtered, egg albumen: it shrinks and curls. The lower part of the fragment is attached to the paper (which prevents some shrinkage), while the upper portion is free to move as the forces equalize into a curl. This process can be seen in a video clip where this albumen fragment is dried from wet.
All of the albumen photographs exhibited some degree of cracking. This is true of all images seen by the authors and most other conservators and curators. The degree of variation can be seen in Table 3 in Messier & Vitale (1994).
Early in the history of photography conservation, cracks were a tool to identify albumen photographs.
Based on the 20 vintage prints used in the study an average square inch of print (before treatment) will have 2,000 cracks, with over 1 million in an average 8 X 10 print. The average crack width was found to be 11.7 microns, before treatment.

Exposure to Moisture Causes Changes in Albumen Prints

In the Messier & Vitale (1994) study of 20 vintage images it was found that even minor surface cleaning made a 21% increase in crack of existing cracks and a 48% increase after water immersion for backing removal. Crack population also increases as the result of wetting. A 14% increase was seen with simple aqueous surface cleaning and an addition 27% after immersion for backing removal

Rapid Local Humidity Change Results in Cracking

A study was done on a loose chip new albumen in 1993. Initially the study was to demonstrate the differences in deformation behavior for fast and slow application of stress; see those results below. However, it was discovered that small humidity changes, from 50% RH to 75% RH, caused by the light used for photography resulted in extensive cracking of the chip due to the internal stress created by the humidity change. The sample was photographed at the start of the localized humidity change, then in 30 seconds intervals for about 2 minutes, and then after 5 minutes, with the following results. The 5-minute image is worse than the 90 second version but not by much.

Chip at 50% RH and 70F
Chip after 30 seconds
Chip after 1 minute
Chip after 90 seconds
Chip after 5 minutes

See an Albumen Chip in Cross-Section Go from Dry, to Wet to Dry

Wet (immersed in water)

Speed of Deformation Effects Behavior

The following experiment was done to show that there is a time dependency for the cracking (failure) behavior. Faster strain rates results in brittle behavior while slow intrusion results in plastic behavior. The experiment was to push a needle probe into the surface of a chip of modern albumen made using the Towler method. The first part was done, fast, as one would normally
probe a material to see how hard or soft it might be. The slow part was done over about 2 minutes. The result of the slow experiment could only be seen when the probe was removed, revealing a depression; as can be seen in the image on the far right. This parallels results found in Vitale & Messier (1994), where the ultrasonic modulus (made at very high rates of strain) is high in comparison to those made in using the slow "equilibrium stress-strainng" protocol used in all
the remaining mechanical measurements. This highly interesting and revealing experiment was suggested by Duane Chartier (Culver City, CA).

Chip before experiment
Start of fast intrusion
Result of fast intrusion
Start of slow intrusion
Result of slow intrusion, magnified

Why Albumen Behaves this Way

Humidity: Albumen swells and shrinks to a greater degree than many materials. What makes this unfortunate property problematic is that swelling caused by to a 20% change in humidity (around 50% RH) is enough to cause already cracked albumen to increase crack sizes and possible to form new cracks. For modern albumen which is crack free, a 35% change in humidity (around 50% RH) will cause the creation of new cracks.

Wetting: If an albumen layer is swelled in water (immersion, for example), it swells so much (17+%) that a massive number for new cracks form and existing cracks increase in width; see above and Messier & Vitale (1994).

This behavior is explained by the following two images. The plot above shows the dimensional of an albumen bar change due to swelling in air with increasing, and then decreasing, amounts of water vapor. Between 50% RH and 85% RH albumen swells 12.4%, and shrinks 11.8%. Note that the slight difference between swelling and deswelling is know as hystersis. In the second plot below, the result of a very slow straining process, shows that already cracked albumen (E-B Method) can only be strained 1.5% before it fails. Another sample of modern albumen made using a method which produced an albumen bar which was crack-free (Towler Plate Method) can be strained 3.4% before it fails.


Taking this data into consideration, the following chart shows the problem. If albumen goes from 50% RH to 20% RH, a strain of 1.7% will results from the humidity change. Our sample of cracked albumen failed when stretched (strained) 1.5%. Thus, cracks grow and "possibly" new ones form from a strain of 1.5% or greater. Note that the strain (inches per inch) is converted to percent by multiplying by 100; a strain of 0.01 is 1% strain (0.01 X 100) of the material.

Work was done at CAL, Smithsonian Institution

Conservation research was done at the Conservation Analytical Laboratory (CAL, now called SCMRE, Smithsonian Center for Materials Research and Education) by the authors between 1990-92. The goal was the personal need for understanding and to provide information to conservators, curators and collections managers so they could balance the obvious benefits of (1) aqueous treatment and (2) humidity changes with their drawbacks.