Accurate Modelling of a GaAs-Based Laser Power Converter

High power laser transmission (HPLT) is a technology capable of transferring energy using a monochromatic laser onto a laser power converter (LPC). This avoids the main limitations of conventional wiring, such as sparks or electromagnetic interferences[1], which is desirable in applications as monitoring systems for refineries or mines[2]. The technology also allows to optical power devices such as antennas[3], drones[4] or satellites[5]. HPLT has been pointed out as a key on the development of the wireless power transfer field[6], which has already become a billionaire market[7]. Although LPCs with efficiencies above 65% at room temperature have been reported[8, 9], the overall efficiency of the whole system is less than 20%[6]. To develop the technology’s full potential, new paths towards high-efficiency LPCs are required. For instances, the Vertical Epitaxial Hetero-Structure Architecture (VEHSA) can reduce series resistance and increase the efficiency and input power density (Pin) due to the distribution of the current between the multi-junction subcells[10]. Also, the use of high-bandgap materials has been proposed[11, 12] as a new route to reduce intrinsic entropic losses and series resistance losses, greatly improving the performance of the LPC and allowing the transmission of higher power densities. As the experimental implementation of these or other new potential improvements requires long developing times, with generally high manufacturing costs associated, a preliminary TCAD modelling and optimization study is highly advised. In this context, the usage of trustworthy and reliable TCAD tools and procedures is needed. Therefore, the objective of this work is to analyze the reliability of TCAD studies for the characterization of real devices. For that reason, we present a meticulous TCAD modelling of a GaAs-based LPC, the current preferred base material in the state-of-the-art devices[13]. We chose the experimental LPC published by Shan et al.[14] due to the detailed information on design parameters provided by the authors. Fig. 1 shows the experimental structure and doping values of the LPC. We carried the simulation with Silvaco Atlas[15], a TCAD simulator able to provide realistic and trustable results when modeling a wide variety of devices, including photovoltaic cells[16, 17]. The Poisson and continuity equations are solved to obtain the characteristics of the device. For carrier mobility we use doping concentration-dependent tabulated data, available via Silvaco. The Shockley-Read-Hall (SRH) recombination also considers the doping concentration, following the experimental data published by Lush et al.[18]. We also include optical and Auger recombinations in the simulation framework. All simulations were performed at T=298 K. The absorption coefficient used depends on both the doping concentration and the wavelength, and fits the experimental curves published by Casey et al.[19]. Figs. 2 and 3 show, respectively, the J-V and P-V curves for the experimental and modelled cells at input power densities of 3 and 5 W/cm2. Note that the modelled curves fit remarkably well the experimental ones, properly matching the distinctive J-V points, such as the short-circuit current density (Jsc), open circuit voltage (Voc) and maximum power voltage (Vm). Indeed, all the modelled figures of merit (shown in Table 1) are within a 98% accuracy with respect to the experimental ones for the two input power densities. The energy bands, shown in Fig. 4, reflect the proper performance of the window and back surface field (BSF) layers, acting as barriers for the minority carriers, which is a relevant feature to avoid major recombinations in contacts. As a conclusion, the results of the modelled LPC show an excellent agreement with the experimental ones, supporting the reliability of TCAD modeling. This has relevance in the context of designing and optimizing novel photovoltaic cells in a timely manner, aiding the research in new generation ultra-high efficient LPCs.