Optimization of Gold Tailings Filtration with Chemical Aids and Filter Cloths to Produce Dry Stackable Tailings: Important Factors and Potential Pitfalls

Abstract Dry stacking of tailings is a safer and more sustainable alternative to conventional tailings dams; however, its implementation is often constrained by the fine particle size of the material. This study evaluates strategies to enhance gold tailings filtration through the combined use of chemical filtration aids and optimized filter cloth selection, while also identifying potential methodological and operational pitfalls relevant to filtration optimization. Bench-scale vacuum filtration tests (Leaf Tests) were performed using process water, varying reagent type and dosage, filter media, and solids content, with the target cake moisture defined based on compaction tests. Tailings characterization revealed a high fines content, with a Fraunhofer mean diameter of 37.41 μm (d 50 = 29.57 μm; d 90 79.65 μm), and quartz identified as the dominant crystalline phase, together with muscovite, kaolinite, illite, and goethite. Statistical design of experiments identified solids content and reagent dosage as the primary factors influencing filtration rate, whereas filter cloth type significantly affected cake moisture and filtrate turbidity. The optimal condition, consisting of 15 g/t Floticor FA 18,899 (anionic polyacrylamide), AS–619.CA polyamide cloth, and 60 wt% solids, achieved cake moisture within the target range and increased filtration productivity by approximately 15%, corresponding to an absolute increase of 40 kg·h⁻¹·m⁻², relative to the reference filter paper test. However, the practical significance of laboratory-derived improvements, the limited benefit of increasing reagent dosage, and the use of dried and reconstituted tailings represent potential pitfalls that should be considered when interpreting the optimized conditions. Validation using as-collected slurry under pilot-scale or industrially representative conditions is therefore required to assess their applicability before full-scale implementation.