Development of nano-colloidal system for fullerene by ultrasonic-assisted emulsification techniques based on artificial neural network

Abstract

Propagation of high intensity ultrasonic waves and shearing effect in formulating nanoemulsion loaded fullerene for drug delivery was investigated. Artificial neural network (ANN) was applied to optimize the emulsification process by varying the ultrasonic and homogenization parameters. Control of operating conditions, such as sonication amplitude (30–70%) and duration (60–120 s), as well as homogenization rate (4000–5000 rpm), was tested to determine the physical attributes of nanoemulsion. Ultrasonic cavitation showed far greater effects as compared to high shear homogenization in controlling the droplet size (sonication time) and viscosity (sonication amplitude) of the nanoemulsion system. Levenberg–Marquardt algorithm produced the optimum topology with network architecture of three inputs, four hidden nodes, and two outputs. Validation further confirmed the aptness of the proposed model with low root mean square error. In this study, ANN has superior predictive ability by yielding low percentage of residual standard error. An ultrasonic approach in formulating fullerene nanoemulsion system is a powerful technique in minimization of droplet size and acquisition of desirable viscosity. This serves as a platform to advance fullerene in nanomedicine field despite its hydrophobicity. Keywords: Nanoemulsion, Fullerene, Artificial neural network, Ultrasonic cavitation, High shear homogenization