Arctangent normalization and principal-component analyses merge method to classify characteristics utilizing...

Area of Science:

• Materials Science
• Analytical Chemistry
• Spectroscopy

Background:

• Characterizing time-dependent paint properties is crucial for understanding material evolution.
• Traditional methods struggle with diverse data dimensions from multiple analytical techniques.
• Accurate classification requires methods that integrate static and dynamic material information.

Purpose of the Study:

• To develop a new multivariate analysis technique for classifying paints with time-dependent properties.
• To introduce a "PCA-merge" method that integrates principal component analyses across different time components.
• To enable simultaneous analysis of static and dynamic paint characteristics.

Main Methods:

• Comprehensive characterization of paints during drying (1-48 h) using FTIR, ICP-MS, and HS-GC/MS.
• Data normalization using an angle parameter (θ) via arctangent transformation to standardize intensity and time variables.
• Development and application of the "PCA-merge" method for analyzing merged principal component analysis data groups.

Main Results:

• Arctangent normalization effectively reduced the influence of varying data intensities and analytical instrument differences.
• Multivariate analysis of normalized data enabled successful classification of paint samples into categories.
• The PCA-merge method successfully distinguished samples by utilizing shifts in the barycenter of the PCA score plot, integrating time-dependent data.

Conclusions:

• The proposed "PCA-merge" method offers a robust approach for classifying time-dependent materials like paints.
• This technique allows for the simultaneous analysis of static and dynamic properties, providing deeper material insights.
• The method enhances paint characterization by leveraging barycenter shifts in PCA scores, minimizing the need for detailed peak identification.