Boron-doped Sucrose Carbons for Supercapacitor Electrode: Artificial Neural Network-Based Modelling Approach

Here, a simple yet efficient and economic strategy was demonstrated for the production of multiporous boric acid-doped sucrose carbon (Bx–pC) for supercapacitor application. The electrochemical performance was established through cyclic voltammetry and galvanostatic charge/discharge tests. Bx–pC samples were characterized by X-ray diffraction, scanning electron microscope, Raman spectroscopy and nitrogen adsorption/desorption at − 196 °C. The results reveal that the optimum boron dopant is 2 wt.%; and B2–pC containing 2 wt.% boron exhibited honeycomb-like porous structure (2.88 nm) and a high specific surface area of 1298.9 m2g –1 . The B2–pC based symmetric supercapacitor delivered a remarkable energy density of ~56 Wh kg−1 , a high power density of 1300 W kg−1 and superior capacitance of 239 F g−1 at 1 A g−1 in 1 M H2SO4 electrolyte. To establish the complex relationships between the electrode structure, active operating conditions and electrochemical performance of the supercapacitor, an artificial neural network (ANN) methodology was utilized herein. After several random runs, the ANN maintained satisfactory predictive performance with an average error rate of ~1.06% and desirability function of 0.93 which is closer to 1.0. Keywords: supercapacitor performance; sucrose-based porous carbons; artificial neural network optimization; electrochemical analysis.