Converting carbonyl index values in microplastics studies

Abstract

The variability in carbonyl index (CI) calculation methods among studies has made direct comparison of CI values challenging. To address this, we developed a CI conversion methodology that enables the unification of CI values across studies. Using spectra data from environmental microplastic samples, a polynomial regression model was developed to convert predictor CI values into a target CI. The optimal model order was determined through five-fold cross-validation, which also demonstrated the model's ability to explain 36 %–84 % of the variance depending on the predictor CI method. The developed model was further applied to convert average CI values from different studies in coastal regions around Japan. Results revealed that microplastics in the southeast coastal regions of Japan are more oxidized than those in the west coastal regions, likely due to differences in environmental exposure and retention time. The developed methodology is also adaptable to CI calculation methods not explicitly covered in this study.