An evolutionary approach for semiconductor nanodevices optimization

Abstract

One of the main reasons for the large computational development seen in recent decades has been the continued miniaturization of transistors in size. However, we are close to the physical limit of miniaturization of the electronic components. Therefore, in order to maintain the advancement of the performance of processors, we need new alternative technologies and materials to be investigated. Nanocomputing aims to study the nanostructure and nanodevices for the development of a new generation of computers, with innovative and efficient architectures. Among the possible solutions we can highlight the semiconductor nanodevices. In this paper we investigate the optimization of two types of different semiconductor nanodevices: Photonic Crystals and Microcavities, that can enable the development of a future generation of computers that use light as a state instead of the electric charge of conventional processors variable.Typically, the process of optimizing these physical structures is empirical and slow, demanding knowledge and intuition of experts. The most robust and efficient optimization requires the existence of mathematical models and simulators able to represent the structural behavior of the devices. These models can take high computational complexity, making systematic evaluation of these structures a challenge. Moreover , the search for efficient solutions coming up in large search spaces and nonlinear behavior. Thus, in this work we apply evolutionary algorithms in the search for optimized satisfying the requirements for the development of future applications solutions.Photonic crystals are systems whose dielectric function is periodic in space. Generally, these structures are implemented in a semiconductor crystal, for example, Gallium Arsenide (GaAs), and have patterns of holes filled by air. By changing geometrical parameters of the cavities of the crystal, that is, controlled generation of defects in the structure, the light can be controled, a fact that allows the development of applications such as optical logic gates, high resolution sensors, quantum processing information, among others. In this project we focus on maximizing the quality known as L3 structure factor, varying the geometric positions and radii of the holes around the defect. The quality factor can be defined by the energy lost per cycle versus the energy stored in the defect. The results obtained in our case studies outnumber those previously presented in the literature. Furthermore, it is important to note that the simulation of each structure generated is computationally very expensive, which led to the development of a distributed and robust algorithm that could take advantage of the largest possible number of computers .Have the semiconductor microcavities can be considered as a dimension photonic crystals and are bases for a variety of optoelectronic devices such as lasers and light emitting diodes (LEDs), and optical transistors. Typically, these materials contain many different layers, whose growth conditions should be highly stable, with precise control over the composition and thickness of each individual layer. However, uncertainties in the physical process of growth beyond the control of experts and can significantly compromise the efficiency of the devices. Thus, in this project we aim to not only find the optimal parameters that lead to optimal solutions, but also to ensure the growth of robust and efficient devices. This is the first project we propose to structure optimization of the wells, mainly focusing on its robustness. Different types of wells were optimized and the strategy proposed in this study proved effective, leading to structures with higher quality factor than previously described in the literature guarantee robustness and high growth.