One well-known way to increase the efficiency of a solar cell is to decrease the reflectivity of the air/semiconductor interface; because semiconductors typically have high refractive indices and thus high reflectivities, this approach can substantially improve efficiency. The simpler the coatings, of course, the lower the cost of the resulting photovoltaic cells. Researchers at the Beijing University of Technology (Beijing, China) are optimizing three-layer broadband optical antireflection (AR) coatings for silicon using what is called an ant colony algorithm (ACA). Over the wavelength range from 400 to 1100 nm, the resulting coating has an average reflectance of 2.98% for incidence angles from 0° to 80° and 6.56% for incidence angles from 0° to 90°.
The ACA approach was originally conceived as a way to solve the classic "traveling salesman" problem, and is based on the capability of ants for finding the shortest path between a food source and their nest. In particular, ACA is insensitive to an optimization run's starting values. Each layer in the multilayer structure to be optimized is homogeneous and is characterized by its thickness and refractive index. The bottom absorption layer is assumed to be bulk silicon. The optimization algorithm is modeled on the behavior of actual ants, even including a simulation of the effects of ant pheromone. Each ant updates the amount of pheromone on its path, while any previously deposited pheromone is assumed to be volatile and thus gradually disappears. Contact Xia Guo at[email protected].