Multivariate statistical analysis of influencing factors in ecological multifunctional wool dyeing using Spartium...

Dyeing with natural dyes offers an eco-friendly and sustainable alternative to synthetic colorants, aligning with the increasing demand for environmentally responsible textile processing. This study investigates the aesthetic and functional enhancement of wool yarns dyed with Spartium junceum flower extract, employing a comprehensive mordanting strategy including single metal salts (Cr, Cu, Ni, Co, Zn), binary metal combinations, biomordants (oak hull, sumac fruit, eucalyptus leaves, gallnut, ascorbic, and gallic acids), and metal-biomordant hybrids, yielding 52 formulations in total. Colorimetric evaluation (K/S, L*, a*, b*, C*, h°) showed that Cr(III)- and Cu(II)-based systems markedly increased color strength and shade depth compared with unmordanted and biomordant-only samples, while hybrid systems such as Cr/GA, Cr/EU, Cu/GA, Cu/AA, and Zn/Cr produced the deepest, most saturated yellow-orange shades at reduced metal dosages. Fastness testing confirmed good-very good wash and rub fastness and substantially improved light fastness (6-7 and 7) for the best-performing hybrids, further supported by very low ΔE* values. Functional assessment revealed that selected metal-biomordant systems dramatically enhanced UV protection, increasing UPF from "Poor/Good" for raw and unmordanted yarns to "Excellent" levels, with several hybrids exhibiting UPF values well above 100. The same systems also showed strongly increased antioxidant activity in the DPPH assay relative to unmordanted controls. Principal Component Analysis (PCA) revealed that PC1 and PC2 together explained 73.78% of the total variance, confirming the multivariate interdependence between color depth, UV shielding, and radical scavenging. Box plots and radar charts further emphasized Cr-GA and Cu-AA as high-performance mordant systems with balanced functional and aesthetic profiles. A composite performance index was used to rank formulations, identifying 11 top-performing systems with scores exceeding 0.63. This study illustrates how bio-metal mordanting, coupled with multivariate analysis, can be leveraged to design natural textiles with high multifunctionality. Finally, a machine learning (i.e., random forest) model was used to predict colorimetric attributes, and the satisfactory performance of the model was noted, implying the potential of using machine learning in prediction of similar process parameters.