Layer-by-layer decoding of contemporary and historic painting composition using MALDI mass spectrometry imaging and machine learning

TL;DR

This paper introduces a new method using MALDI mass spectrometry imaging and machine learning to analyze the composition of modern and historic paintings layer by layer.

Contribution

The novel contribution is a high-chemical-specificity method for automated composition assignment of individual layers in multilayered paintings.

Findings

The method provides unprecedented molecular detail using MALDI mass spectrometry imaging.

Machine learning enables automated identification and mapping of organic and inorganic compounds in historic painting cross sections.

Abstract

Layer-by-layer characterization of modern and historic paintings is essential for understanding how an artwork was created and how it has changed over time. This information can reveal historical or societal insights, inform attribution and classification, and support preservation efforts. The determination of the identity, structure, and composition of art material remains challenging due to the complex, multilayered nature of paintings. These structures often contain mixtures of organic and inorganic components, with unknown in situ chemical interactions, that create cross-linked and chemically modified molecular networks and assemblies. In this work, we have developed a method with high chemical specificity to identify and map organic and inorganic components in modern and historic multilayered paint systems. Our approach achieves an unprecedented level of molecular detail using a single technique. In addition, we have designed the method for automated composition assignment of individual layer through high-resolution matrix-assisted laser desorption/ionization (MALDI) mass spectrometry imaging (MSI) and a dedicated composition prediction model.

MALDI-MSI with machine learning identifies and maps organic and inorganic compounds in historic painting cross sections.

Figures