Experimental Screening and Statistical Modeling of Mass Loss as a Function of Solvent Type and Process Parameters in the Hydrothermal and Solvothermal Conversion of Carp (Cyprinus carpio) Scales Toward Carbon Quantum Dot Production

Abstract

This study systematically investigates the hydrothermal and solvothermal parameters influencing the production of carbon quantum dots (CQDs) from the collagen-hydroxyapatite composite structure of carp (Cyprinus carpio) scales. The effects of different solvent systems, raw material-to-solvent ratios, total raw material loading, 1% NaOH alkali pretreatment, acidic/neutral additives, and reaction medium conditions on biomass conversion behavior were screened at a constant temperature of 200 °C for 4 hours. Mass loss was utilized as the primary process indicator to compare reaction efficiency and mass transfer tendencies. To optimize process parameters within this framework, a 40 run experimental design was implemented, evaluating the influence of solvent type (distilled water, ethanol, and acidified aqueous/alcoholic solutions), sample amount (1g and 3g), and solvent volume (5mL to 60 mL). Due to the binary-level constraints of the sample amount factor, a first-order linear interaction model was successfully fitted using ordinary least squares regression to avoid matrix singularities. Analysis of Variance (ANOVA) revealed that all main factors and their combined interactions exerted a statistically significant effect on the conversion process (p < 0.01). The model exhibited exceptional predictive accuracy, achieving a high coefficient of determination (R² = 0.9913 and 0.9612), an adjusted coefficient of determination (R²adj = 0.9882 and 0.9413), and a low root mean square error (RMSE = 0.0732g and 0.0382g). Comprehensive residual diagnostics confirmed the statistical validity, normality, and homoscedasticity of the model. Furthermore, 2D contour and 3D response surface profiles demonstrated that sample amount, solvent volume, and their synergistic interaction (SxL) are critical determinants governing the recovery efficiency, providing a robust mathematical foundation for scale-up and subsequent analytical characterizations. The findings are not interpreted directly as CQD product yield; rather, they provide a preliminary evaluation framework establishing a foundation for selecting optimal experimental conditions for subsequent photoluminescence (PL), UV-Vis, FT-IR, and advanced microscopic characterizations.