Rapid and robust quality evaluation of fragrances using near-infrared spectroscopy compensated with multi-color parameters.

Authors

Jiaxiao Cai, Yuanlin Tan, Yuqi Luo, Bo Kong, Ting Deng, Ziyu Zhao, Jianna Yu, Suxing Tuo, Wen Liu

Published in

Analytical methods : advancing methods and applications. Oct 02, 2026. Epub Oct 02, 2026.

Abstract

This study investigated the influence of sample color variations on near-infrared (NIR) spectra and spectral modeling for fragrance quality evaluation. The color characteristics of 73 fragrance samples of different types and batches were first analyzed. Hierarchical clustering of RGB and L*a*b* color parameters was performed to objectively assign color labels to fragrance samples, and then violin plots using different color labels revealed distinct color differences across sample types and batches. Principal component analysis (PCA) of the NIR spectra revealed a clear separation trend of samples of different colors in PCA score plots and indicated that the primary spectral variations were largely associated with color differences. Correlation analysis further demonstrated interrelations between color parameters and spectral absorption, particularly in regions corresponding to C-H, O-H, and N-H functional groups. These findings showed that color parameters can assist in revealing additional chemical information embedded in the spectral data. Thus, quantitative prediction models for the refractive index of fragrances were developed by NIR spectra compensated with color parameters, combined with standard normal variate (SNV) or multiplicative scatter correction (MSC) preprocessing. Additionally, the impacts of full-wave, long-wave, and short-wave on the constructed model performance were also compared. Modelling results indicated that the models combined with RGB or L*a*b* parameters both exhibited obviously enhanced performance, as evidenced by increased R_c² and R_p² and a decrease in RMSEP, compared to the original model without color parameter integration. The partial least squares (PLS) model integrated with RGB values pretreated with the SNV method achieved the best performance, with a predictive coefficient of determination (R_p²) of 0.9866, an RMSEP of 0.0021, and an RPD of 8.6387, indicating excellent accuracy and stability. These results confirm that the proposed color compensation strategy effectively considered the effects of color differences on NIR spectral signals, thereby enabling reliable and robust quantitative analysis for diverse fragrance products in practical production settings.

PMID:
42825420
Bibliographic data and abstract were imported from PubMed on 02 Oct 2026.

Read full publication at:
Please sign in to see all details.

Stats

1-terrible, 9-excellent. How would you rate this publication? Sign in in to submit your rating.

Post a comment

You need to be signed in to post comments. You can sign in here.

Comments

There are no comments yet.