Genetic Algorithm Optimization of Core-Shell Nanowire Betavoltaic Generators

Numerical optimization has been used to determine the optimum junction design for core-shell nanowires used in betavoltaic generators. A genetic algorithm has been used to calculate the relative thickness, height, and doping of each segment within silicon, gallium arsenide, and gallium phosphide nanowires. Using the simulated spectra and energy deposition of nickel-63, nickel citrate, tritium, and tritiated butyl, devices with power output and overall efficiency up to 8 µW.cm^-2 and 12%, respectively, have been predicted. Compared to previously investigated axial nanowires, the core-shell structures simulated here have realized drastic improvements by reducing surface recombination for longer nanowires. In addition, core-shell nanowires are shown to be capable of nearly matching the ideal performance predicted for this device structure. A new approach for calculating the practical upper limit of betavoltaic performance is presented and additional methods for improvement are discussed.