Navigating the effect of tungsten oxide nano-powder on ethylene glycol surface tension by artificial neural network and response surface methodology

Abstract The quiddity of ethylene glycol (EG) surface tension is attributed to the cohesive forces between EG molecules. The presence of tungsten oxide (WO3) nanoparticles certainly affects the cohesive forces and the surface tension will be affected. In this study, SITA dynotester was utilized for measuring surface tension at 20–50°C and 0.005–5 wt%. Loading nanoparticles intensified the adhesive force between EG and WO3 molecules and simultaneously attenuated cohesive forces between EG molecules. According to the results, the positive effect of the adhesive force was overcome by the negative effect of reducing the adhesion force and therefore the surface tension was weakened. At 20 ° C the EG surface tension was attenuated up to 15.57% by adding WO3 nanoparticles (5 wt%). Performing response surface methodology and artificial neural network for navigating the surface tension was successful. The acceptable R-squared for RSM (0.982) and ANN (0.99) led to maximum error of 1.43% and 0.95%.