Extraction and multivariate spectral analysis of bioactive compounds in Peganum harmala L. seeds using hyphenated SFE-FIA-DAD system

This doctoral study presents a pioneering example in analytical chemistry for the monitoring and evaluation of multicomponent natural products through the integration of green extraction, online analysis, and chemometric modeling, based on Peganum harmala L. seeds. The developed system is envisioned as a powerful tool for future industrial-scale applications such as real-time production monitoring, process validation, and product quality control. Real-time monitoring and quantitative determination of bioactive compounds during plant extraction processes still represent a significant challenge in analytical chemistry. In this study, a novel integrated platform SFE-FIA-DAD was developed as a multi-purpose analytical process combining Supercritical Fluid Extraction (SFE) with Flow Injection Analysis (FIA) and Diode Array Detection (DAD). The system allows simultaneous execution of extraction, monitoring, and determination processes, and was specifically applied to the seeds of Peganum harmala L., known for their richness in bioactive compounds. Extraction conditions were optimized using a three factor, five level Central Composite Design (CCD) involving extraction pressure, temperature, and CO₂ flow rate. A total of 20 extracts were obtained and analyzed using standard chromatographic and spectroscopic techniques (HPLC-DAD/FLD, GC-FID, GC-MS). These same extracts were then analyzed using the developed hybrid-mode SFE-FIA-DAD system, aligned with Process Validation Approach principles. Spectral data collected from the system were evaluated through PCA, LDA, k-NN, SIMCA, PLS-DA, and multi-output PLS regression models. Comparison of the spectral modeling results with reference chromatographic and spectroscopic data revealed high correlation values (R² > 0.93), indicating strong analytical agreement. In total, 11 target compounds were included in the modeling process. Among them, harmaline, chlorogenic acid, apigenin-7-glucoside, gamma-tocopherol, and delta-tocopherol were identified with high reliability, showing prediction accuracy between 85% and 110% and AUC values of 1.00, reflecting perfect classification performance. During the extraction process, the system was operated in a loop-based discontinuous mode, where spectral data were collected via DAD at specific time intervals. This enabled time-resolved spectral monitoring, and the elution profiles of bioactive compounds over time were successfully recorded. The results indicated that alkaloids were predominantly extracted within the first 120 minutes, followed by a higher extraction of phenolic compounds in later stages. The developed system not only contributes to the green analytical approach by significantly reducing solvent consumption but also merges separation, qualitative identification, and quantitative determination into a single integrated framework. This integration renders the platform a highly applicable real-time process monitoring tool. Owing to these features, the system can be readily adapted for use in pharmaceutical, food, and herbal product industries for process tracking, quality control, and standardization applications.