Optimization of Gypsum Filtration in Phosphoric Acid Production

Abstract

Filtration is one of the main operational bottlenecks in the wet-process phosphoric acid industry, as it directly affects plant productivity, operating costs, and product quality. During the desulphation stage, the filtration behaviour of the calcium sulphate precipitate is influenced by multiple interacting factors, including phosphoric acid composition, phosphate rock characteristics, and impurity content. However, the combined effect of these variables remains insufficiently understood under industrial operating conditions. The objective of this work was to evaluate the influence of phosphoric acid composition and phosphate rock type on filtration performance during wet-process phosphoric acid desulphation. A mixture Design of Experiments (DOE) was employed to investigate eight industrially relevant acid compositions prepared from Benguerir, Engis, and Foskor phosphoric acids using two phosphate rocks, Moroccan K09 and Algerian. Sixteen desulphation experiments were carried out under industrially representative conditions. Filtration time was selected as the primary response, while filtered-acid composition, cake properties, filtration parameters, particle-size distribution, and qualitative microscopic observations were used to support the interpretation of the process. Response-surface methodology, Principal Component Analysis (PCA), and Partial Least Squares Regression (PLS) were applied to investigate the relationships between impurity composition and filtration performance. The results demonstrated that phosphate rock type exerted the strongest influence on filtration behaviour. Under the studied conditions, K09 consistently produced filtration times approximately four times shorter than Algerian rock and generated filter cakes with significantly lower specific cake resistance. Acid composition also affected filtration performance, although its influence depended on the phosphate rock employed and resulted from interactions among the three acid streams rather than from the concentration of any individual acid. Particle-size distribution alone was insufficient to explain the observed differences, indicating that cake morphology and structure also VIII Influence of Acid Composition and Phosphate Rock Type on Filtration Performance During Wet-Process Phosphoric Acid Desulphation played an important role. The multivariate analysis further showed that impurity composition contained sufficient information to accurately predict filtration time within the experimental domain, with the final PLS model achieving an R² of approximately 0.95 and a cross-validated Q² of approximately 0.88. Overall, this work demonstrates that combining mixture design, filtration theory, and multivariate statistical analysis provides an effective framework for understanding and optimizing phosphoric acid desulphation. The methodology developed in this study establishes a foundation for future predictive tools that may assist industrial raw material selection and acid-blending strategies while improving filtration performance.