Overcoming systematic photocurrent calculation errors in ray tracing simulations

The calculation of photo-generation current densities j(Ph) by means of Monte-Carlo ray tracing combined with Fresnel's equations for thin films is a common method in optical simulation of solar cells. We identified a reproducible and inherent error in ray tracing simulation software leading to systematic errors in j(Ph) calculation in the order of 0.1 to 0.4 mA/cm(2) when using defined textures like upright, inverted and in some cases random front side pyramids in combination with pyramidal or planar rear sides with the use of Fresnel's equations applying the transfer matrix formalism as physical interface model. It is argued that this error emerges from coupling effects of the front and the rear side interfaces, leading to regular and "dragon-back" shaped oscillations in the calculated j(Ph) when plotted as a function of the solar cell thickness. The amplitudes and frequencies of these oscillations correspond to the width of the pyramids used for the simulation. We show that reducing the base width of the pyramids is a simple way to solve this problem by reducing the amplitude to values where it meets the size of simulation-inherent statistic noise caused by the Monte-Carlo approach, without affecting other simulation parameters. (C) 2015 lite Authors. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).