Research on Quantitative Model of Corrosion Products of Iron Artefacts Based on Raman Spectroscopic Imaging

Cheng Xiao-xiang,Wu Na,Liu Wei,Wang Ke-qing,Li Chen-yuan,Chen Kun-long,Li Yan-xiang
DOI: https://doi.org/10.3964/j.issn.1000-0593(2023)07-2166-08
2023-01-01
Spectroscopy and spectral analysis
Abstract:Iron artefacts are important part of the cultural heritage in China. Due to the high activity of iron, iron artefacts are prone to corrosion and deterioration. Corrosion products greatly influence the stability of iron cultural relics. Therefore, determining the composition of iron corrosion products is significant for evaluating iron artefacts' stability. In this study, pure chemical reagents were used to simulate three types of corrosion products commonly found on iron artefacts, including hematite (alpha-Fe2O3) magnetite (Fe3O4) and akaganeite (beta-FeOOH). Raman spectroscopic imaging, combined with Principal Components Regression (PCR), Partial Least Squares (PLS) and multiple spectral pretreatment methods, were applied to establish quantitative models for two sets of a binary mixture of standard corrosion samples (alpha-Fe2O3 + Fe3O4, alpha-Fe2O3 + beta-FeOOH). The results indicate that, for alpha-Fe2O3 + Fe3O4 mixed standard samples, the model effects of PCR and PLS algorithms are not much different. The quantitative model results show that the best spectral processing method of PCR modeling is first derivative +Savitsky-Golay (S-G) smoothing (9). For alpha-Fe2O3 + beta-FeOOH mixed standard samples, the model constructed by the PLS method is superior to the PCR method. The best PLS modelling spectral processing method is MSC+S-G smoothing (5). The research results provide an effective method for quantitatively evaluating the chemical stability of corrosion products of iron artefacts.
What problem does this paper attempt to address?