Date Thesis Awarded
5-2022
Access Type
Honors Thesis -- Access Restricted On-Campus Only
Degree Name
Bachelors of Science (BS)
Department
Chemistry
Advisor
Kristin Wustholz
Committee Members
Elizabeth J. Harbron
William R. McNamara
Nicole M. Santiago
Abstract
Surface-enhanced Raman Scattering (SERS) is a powerful tool to detect fading organic colorants in historical oil paintings. However, SERS analysis of red paint samples without chemical pretreatments reveals difficulties in identifying madder lake, one of the most important red lake pigments used by artists since antiquity. The pretreatment of red paint samples proposed by other research groups involved using hydrofluoric acids, which is hazardous and corrosive. In this thesis, a safer and effective pretreatment strategy upon red paint samples is developed to obtain better SERS signals of madder lake. The pretreatment method is then used on samples of two historical portraits from the Colonial Williamsburg Foundation. The appropriateness of using SERS to detect multiple red dyes in a single paint sample is also assessed. Finally, a neural network-based program is developed to automatically classify different red dyes and pigments based on their unique SERS spectra. The proficiency of the program is tested by having the model predict SERS spectra acquired from artworks.
Recommended Citation
Zheng, Zhaoyun, "SERS Analysis of Pretreated Red Lake Colorants in Historical Oil Paintings & Development of a SERS Spectral Analysis Program with Deep Learning" (2022). Undergraduate Honors Theses. William & Mary. Paper 1775.
https://scholarworks.wm.edu/honorstheses/1775
