Application of artificial neural network (ANN) in order to predict the surface free energy of powders using the capillary rise method

Abstract

Capillary rise method was used to determine the surface free energy of 15 different powders. This method is based on measuring the penetration time needed for a liquid to rise to a certain height. The normalized wetting rates as a function of surface tension of the test liquids for a given powder will show a maximum, which is the solid–vapor surface tension of that powder. The powders used covered a wide range of surface free energy (25.5–63.9 mJ/m 2). An artificial neural network (ANN) was used to predict the normalized wetting rates for the powders. The network's inputs were particle size, bulk density, and packing density for the powders and surface tension for the liquids. Using the designed and trained network, for each investigated powder, values of surface tension were made to vary in the range of 15.45–71.99 mJ/m 2 (i.e. surface tension range of the available liquids) in increments of 0.01 units and the normalized wetting rates were recorded. The surface tension equivalent to the maximum normalized wetting rate was reported as the solid–vapor surface tension for the powder being investigated. As a result, the individual surface free energy of these powders based on the capillary rise method, was determined without need to obtain the surface tension of each liquid experimentally.