Evolutionary design algorithm for optimal light trapping in solar cells

Abstract

Light scattering at rough interfaces is a standard approach to enhance absorption of light in the absorber layers of solar cells due to light trapping. In this work, an automata optimization algorithm, based on the Constructal Theory, is proposed for the heuristic geometric design to enhance light trapping of silicon solar cell surface textures. Surface textures were investigated computationally for light absorption of a normally incident simulated solar spectrum, under constraints of transverse area of the texture. The photogenerated current density was calculated for textures to estimate light trapping. In the automata optimization, Elemental Constructals, consisting of dielectric material blocks, were assembled on top of the solar cell forming a texture. Three variations of the proposed automata algorithm were tested and their capabilities were determined. The best result for this optimization was a photocurrent J=17.48mA/cm2 for grid periodicity above the wavelength and J=18.31mA/cm2 for grid periodicity in the sub-wavelength range. The evolutionary algorithm was also applied to assemble a periodic-disordered texture from fixed shape gratings to better harness incident light. Different textures with fixed shapes of gratings had been investigated to determine optimal geometric parameters. Optimized triangular, trapezoidal, and rounded-tip textures achieved the best results in this phase and were implemented to assemble the periodic-disordered texture which reached J=19.75mA/cm2 for triangular gratings. Results found with this pristine evolutionary algorithm corroborate its efficiency in finding, practically, geometries that lead to greater light absorption with a minimum amount of dielectric materials.