Rapid assessment of algal biomass and pigment contents using diffuse reflectance spectroscopy and chemometrics

Abstract

Rapid and cost-effective analysis of algal pigments and biomass with minimum processing would be an invaluable tool for industrial applications especially in pharmaceutical and food industries. In this study, diffuse reflectance spectroscopy (DRS) was used for quantifying biomass and pigments from Chlorella vulgaris, Nostoc muscorum and their mixed culture via partial least squares regression (PLSR) algorithm. Results indicated that all PLSR models were able to accurately predict algal biomass using visible to near infrared (VisNIR) spectrum, producing an identical R2 of 0.98. While using pigment spectra for modeling pigments for individual culture, DRS exhibited excellent model generalization capability with calibration R2 ranging from 0.92–0.99. Qualitative spectral analysis has identified culture specific spectral signatures. Moreover, while using the pooled data for predicting algal pigments, culture spectra produced similar model accuracies to those produced by using pigment spectra. For predicting total carotenoids, the PLSR model using culture spectra (R2=0.94) outperformed the model, which used pigment spectra (R2=0.35), indicating the potential for developing a new approach for estimating pigment concentrations in algae samples without consuming the sample.