Dissolution of Clinker Minerals C3S, C2S, C3A, and C4AF: Application of a Machine Learning Model

The study focuses on the quantitative analysis of the dissolution kinetics of clinker minerals (C3S, C2S, C3A, C4AF) during the isothermal hardening of Portland cement within the temperature range of40-70 °C. Based on quantitative phase analysis data (Rietveld method) and stoichiometric ratios of hydration reactions, a hybrid approach integrating a dual-exponential kinetic model and a machine learning method (Random Forest Regressor) was developed. The model is grounded in experimental results on the mass fraction of Portland cement hydration products at temperatures of 40, 50, and 70 °C. The quantitative content of the products was determined by the Rietveld method as a function of hardening time. It was established that the temperature dependence of the dissolution rate is most pronounced for C3S. At 70 °C, the residual mass fraction of alite decreases to 10% within 51 hours. The model separates the contributions of surface chemical dissolution and diffusion mass transfer through the layer of hydration products. The parameters of the dual-exponential model were approximated with high accuracy (convergence criterion Rwp = 7.0-9.4%). The kinetic part of the model considers the dissolution rate constants (ki, кг), mass change, ion concentrations (Ca2+, SiO44-, Al3+, Fe3+), and the thickness of the diffusion layer (5) of the clinkers. It was found that increasing temperature leads to a significant acceleration of dissolution, a reduction in diffusion layer thickness, and an increase in ion concentration. Moreover, the most pronounced effect is observed for alite (C3S) and aluminate (C3A). The activation energy values (45-55 kJ/mol) indicate a mixed mechanism controlling the kinetics. The hybrid model demonstrates good accuracy (R2 > 0.92) and enables the prediction of the phase composition kinetics of cement systems, accounting for the formation of metastable hydration products. The obtained results are significant for selecting optimal heat treatment regimes for cement compositions with specified service properties. Contribution of the authors: the authors contributed equally to this article. The authors declare no conflicts of interests.

Keywords

Authors

References