A Green Alternative in Rubber Technology: Properties and Applications of Epoxidized Soybean Oil

Although epoxidized and conventional vegetable oils are increasingly being considered as sustainable processing aids for natural rubber (NR) compounds, the relationships between their physicochemical characteristics, vulcanization behavior, and vulcanizate performance remain insufficiently understood. This study investigated the effects of epoxidized soybean oil and six conventional vegetable oils at oil loadings of 10, 20, and 30 phr. Physicochemical characteristics, mixing power consumption, rheometric parameters, and selected mechanical properties were evaluated. Principal component analysis (PCA) and Pearson correlation analysis were used to examine associations within the dataset, while artificial neural network models were developed for tensile strength (TS), stress at 300% elongation (M300), and maximum torque (Mmax). The first three principal components explained 89.74% of the total variance. Oil loading, rheometric torque parameters, viscosity, surface tension, and iodine number were strongly associated with variation in the investigated compound properties. Within the original random data split, the selected ANN models reproduced a substantial proportion of the variation in the modeled responses. Sensitivity analysis indicated that the calculated M300 and Mmax responses were most strongly dependent on oil loading, whereas the calculated TS response showed the greatest dependence on mixing power consumption and oil loading. However, leave-one-oil-out grouped validation yielded negative pooled R2 values for all three responses, indicating limited generalization to oil types excluded from model development. The PCA, correlation, and ANN results should therefore be interpreted as exploratory descriptions of relationships within the investigated oils and loading range rather than as evidence of causal mechanisms, externally validated predictive performance, or demonstrated applicability to formulation optimization.