Carbon nanotube membranes to predict skin permeability of compounds.

In the present study, carbon nanotube (CNT) membranes were prepared to predict skin penetration properties of compounds. A series of penetration experiments using Franz diffusion cells were performed with 16 different membrane compositions for model chemicals. Similar experiments were also carried out with same model molecules using five different commercially available synthetic membranes and human skins for the comparison. Model chemicals were selected as diclofenac, dexketoprofen and salicylic acid. Their permeability coefficients and flux values were calculated. Correlations between permeability values of model compounds for human skins and developed model membranes were investigated. Good correlations were obtained for CNT membrane, isopropyl myristate-treated CNT membrane (IM-CNT membrane) and bovine serum albumin-cholesterol, dipalmitoyl phosphatidyl choline-treated membrane (BSA-Cholesterol-DPPC-IM-CNT membrane). An artificial neural network (ANN) model was developed using some molecular properties and penetration coefficients from pristine CNT membranes to predict skin permeability values and quite good predictions were made.