Bleaching is a crucial step in oil refining that removes unwanted pigments and oxidative products, which degrade oil quality. This study evaluated the effects of temperature (75–115 °C), adsorbent concentration (1–3 wt%), and time (20–40 minutes) on sunflower oil bleaching using a Box-Behnken factorial design in Minitab 21. Responses were divided into two groups: (I) peroxide value and spectroradiometric-based chroma and hue-angle, and (II) peroxide value and spectrophotometric-based bleaching efficiency. Under optimal conditions (approximately 78 °C, 2.95 wt% adsorbent, and 20 minutes), peroxide value decreased from 1.8 meq/kg to 0.398 meq/kg, chroma reduced from 45.89 to 9.61, hue-angle increased from 92.69 to 105.55, and bleaching efficiency reached 71.86%. Composite desirability was higher in the first group (0.97 vs 0.89), primarily due to the more objective nature of spectroradiometric parameters (hue-angle and chroma).

This study presents a chemical and mineralogical analysis of stone samples from Jošanica quarry collected from three different locations—Field A, Field B, and Field C. Mineralogical analyses were conducted using XRD analysis. The analyses showed that calcite was the dominant mineral in most of the samples, while dolomite was significantly present in some of the samples. Chemical analysis revealed that calcium was dominant in samples 2 (Field B), 3a, and 3b (Field C), with only negligible amounts of magnesium. In contrast, samples 1a, 1b, and 1c (Field A) contained a significant amount of magnesium. Based on the MgCO₃ content, the amount of dolomite in the stone samples was calculated. The content of CaCO₃ in its bound form in dolomite was lower than that in the stone samples, indicating that CaCO₃ was present in another form. According to the dolomite content, samples 1a, 1b, and 1c (Field A) are classified as limestone–dolomite, while samples 2 (Field B), 3a, and 3b (Field C) are classified as limestone due to their high calcium carbonate content. The results of a mineralogical analysis confirmed the results of the chemical analysis.

Silica gel is a widely used desiccant with a high water sorption capacity, influenced by factors such as specific surface area, pore volume, and pore size. This study develops mathematical models using multiple linear regression analysis (MLRA) to predict the water sorption capacity of silica gel based on experimental data. The independent variables include the relative humidity (RH), specific surface area (Sp_BET), pore volume (Vp), and mean pore diameter (d). The Pearson correlation coefficient reveals strong positive correlations between water sorption capacity (w) and RH, and moderate correlations with Sp_BET, Vp, and d. Additionally, Sp_BET is strongly correlated with both Vp and d, while Vp and d also show a strong relationship. Two MLRA models are developed to predict sorption capacity: one based on RH and Sp_BET, and the other on RH and Vp. The R-squared values for these models are 0.886 and 0.902, respectively, indicating that they are strong predictors of water sorption capacity in silica gel.