In August 2003, Professor Ron Bailey, acting Chemistry chair at RPI was approached by Joyce Zucker, Painting Conservator, New York State Office of Parks, Recreation and Historic Preservation, Bureau of Historic Sites, regarding a fragile mural painting at Coe Hall at the Planting Fields Arboretum State Historic Park located on Oyster Bay, Long Island. This painting called the “Buffalo Room” (see Figure 2) was painted by Robert Chanler . According to Joyce Zucker, Chanler was known as a “quirky” artist with respect to the binders and media that he used. This mural painting began to rapidly deteriorate after completion and has undergone numerous coating treatments in an effort to conserve it. The difficulty with conserving this painting is that it is not clear what type of binder was used for this wall painting. Furthermore, since it was directly painted onto a wall, the layers are different than typical canvas paintings.
Figure 2. The “Buffalo Room” painting in the Breakfast Room at Coe Hall. Additional pictures are available at the web site denoted in Reference 35.
In September 2003, a meeting with Joyce Zucker, Ron Bailey and the PI, occurred to discuss the chemical analysis needs for the Buffalo Room painting. Many chemical analysis issues came up including the needs to identify the chemical components of the top coating layers as well as the binding layer with the wall. We decided to focus on the binding layer in the Instrumental Analysis course where the PI taught the lecture and Professor Bailey supervised the laboratory in the spring 2004. The primary issue was how to best integrate this project into the existing lecture and laboratory course. It was decided to cut two of the required experiments towards the end of the term so that students would have time to work on the project at the end of term as well as have time to devise a plan of attack after being exposed to many of the instrumental methods needed to solve the complex analysis problem.
At the beginning of the course, Ms. Zucker came to lecture on the 2nd day of class (Jan 15, 2004) to introduce the chemical analysis problem from an art conservator’s perspective to the students. There was active student participation during this hour. We overviewed the different challenges associated with the chemical analysis of these samples. In particular, the complexities of sampling with an inhomogeneous painting with layered structure were mentioned. Furthermore, we discussed the extreme limitations with sample size and quantity. Ms. Zucker was generous in the very large samples (many were approximately 5 x 5 mm or smaller), in her opinion, that were given to the class for analysis. Many of these samples were removed from the wall behind the radiator to allow for larger sample size. (The students readily recognized this as an additional complication.)
Several class lectures were used for discussion of analytical methods that could be used to determine the chemical source of the deterioration. The students hypothesized that chemical information about the binder was needed. This decision was based on knowing the sample deterioration history. LC and GC analyses for the amino acid contents were chosen. The binder content is important since binders can be classified by either their composition (proteinaceous or carbohydrate) or by the solubility of the binder (soluble or insoluble in water) . Protein-containing binders include casein (milk), egg yolk, egg white, and collagen (animal glue). Each group of approximately six to seven students had a team captain. The GC analysis was based on a paper found by the students in that group . The GC analysis was performed using the department’s Shimadzu QP5050 GC-MS instrument in the shared instrumentation facility and is shown in Figure 3. The LC analysis was performed using UV amino-acid derivatization chemistry that has also been previously described . The LC analysis was performed in Professor Stenken’s research laboratory with her Shimadzu gradient LC-UV instrument with autosampler and computer control. The current HPLC system in the undergraduate lab is anisocratic system with a diode array detector and manual injection with a strip chart recorder for data collection and is not suitable for the large data set collected.
Figure 3. GC-MS total ion current (TIC) chromatogram of amino acid standards.
Each sample set required acid dissolution to solubilize the paint and binder. The students were agreeable to performing 5 to 6-hour dissolution steps despite only being scheduled for four hours. Some initial GC-MS data was obtained, but was not conclusive with respect to content and concentrations of the amino acids in the sample. The LC analyses did not provide useful data. The students were only graded on participation for this special project. After this initial "failure", the course was officially over. More than half the class was completely dissatisfied with their results and truly felt as if they had let Joyce Zucker down because they had not produced anything. In other words, this group of students began to feel a sense of ownership in a short time period for their project. A couple of students actually worked past the end of the term through the reading days (2 days at RPI that are class-free prior to finals) and into finals week to retry the GC-MS analysis. The results improved, but most likely the complex samples as well as the inexperience of the students working with such complex chemistry may have caused additional problems.
One student (Mr. Dan Vissani) had a very strong desire to bring the project to completion and inquired if he could do undergraduate research on this project. He quickly produced in less than a week the two chromatograms shown in Figures 3 and 4. Figure 3 shows the GC-TIC (total ion current) chromatogram for the amino acid standards. Figure 4 shows the GC-TIC for a paint sample from the Buffalo Room. Furthermore, he wanted to see this project extended to students that would take Instrumental Analysis in subsequent years. Mr. Vissani had changed majors and thus was finishing in 4.5 years. He regretted finishing in the December term since he had an interest in being an undergraduate student TA for the course in the spring so as to continue this project.
Figure 4. GCMS analysis of paint sample. The red arrows denote peaks that have been identified using the GC-MS database software (1998 NIST database). Additional peaks are most likely organic fatty acids since this derivatization chemistry will react with these as well.