3D Device Simulation and First Demonstration of BaSi2 Thin Film Solar Cells

Abstract

Barium disilicide (BaSi2) shows great promise as a new material for thin film solar cells (Fig. 1) [1]. It has a suitable bandgap of 1.3 eV, a large optical absorption coefficient of 3×104 cm−1 for a photon energy of 1.5 eV, and a large minority-carrier diffusion length of about 10 μm. Furthermore, it is composed of only earth abundant elements and highly stable. Therefore, BaSi2 can be used for future terawatt-class power generation. We have achieved the operation of BaSi2 homojunction solar cells (Fig. 2) [2]. In this structure, an open-circuit voltage V OC beyond 0.8 V and a conversion efficiency (η) beyond 25% are expected [3]. However, the achieved η was very small. In this study, three-dimensional (3D) optical simulation using a 3D pyramid texture that is effective in preventing reflection was performed on a BaSi2 solar cells in order to increase the photocurrent. Futhermore, we suggested a new device structure using Al-doped ZnO (AZO) electron transfer layer (ETL).In this study, we performed a 3D optical simulation using the GenPro4 model developed at TU Delft [4]. In this simulation, the absorptance of each layer is calculated as a function of wavelength λ. We introduced the 3D-texture structure on the front and back surfaces of the Si substrate to investigate the anti-reflection and light trapping effects (Fig. 3(a)). The thickness and complex refractive index n + ik of every layer are given as input (Fig. 3 (b,c)). In addition, to increase light absorption in BaSi2 light absorber layers, we investigated n+-AZO/p-BaSi2 heterojunction solar cells in which the ETL was changed from n+-BaSi2 to n+-AZO.The absorption spectra of the BaSi2 homojunction solar cells are shown Fig. 5. By changing the front surface structure from flat to texture, the reflectance in the λ range 700 – 1200 nm reduced and the J ph in p-BaSi2 increased by 1.2 mA/cm2. On the other hand, the light trapping effect by the texture structure on the back surface could not be observed well. This is probably because almost all the transmitted light was absorbed by the Si substrate before reaching the back surface. Photons with λ between 800 and 1200 nm are absorbed in the Si substrate. In addition, it was found that the largest factor inhibiting light absorption in the p-BaSi2 layer was parasitic absorption in the ETL n+-BaSi2. This parasitic absorption is unavoidable when BaSi2 is used for ETL, because BaSi2 has a very large light absorption coefficient.Other thin film solar cells, e.g. CIGS, Perovskite, are a kind of heterojunction solar cells which consist of p-type absorbers and n-type window layers. The wide bandgap of the window layer allows more photons to reach the absorber layer. Another advantage of a heterojunction is that the recombination in the wide bandgap window layer is quite low in comparison to that in homojunctions. Now based on these basics, we proposed n+-AZO/p-BaSi2 heterojunction solar cell in which the ETL was changed from n+-BaSi2 to n+-AZO (Fig. 6). Because the E g of AZO is very large (~3.3 eV), much light can be transfer to BaSi2 light absorber layer. The absorption spectra of the n+-AZO/p-BaSi2 heterojunction solar cells were shown in Fig. 7. The parasitic absorption in the ETL was significantly reduced. The J ph in the p-BaSi2 layer increased by more than 10 mA/cm2 compared with that of the homojunction solar cells and reached a maximum of 30.23 mA/cm2 using the front textured surface.Based on these simulated results, we tried to form n+-AZO/p-BaSi2 heterojunction solar cells experimentally. p-BaSi2 and p+-BaSi2 films were grown epitaxially on a Cz-p+-Si(111) substrate by molecular beam epitaxy, and ZnO and AZO were deposited by sputtering. The J-V characteristics under AM1.5 illumination and the internal quantum efficiency (IQE) spectrum for an n+-AZO/p-BaSi2 heterojunction solar cell are shown in Fig. 8. It shows η = 0.04%, a short-circuit current density J SC of 3.7 mA/cm2, and an V OC of 50 mV. IQE exceeded 30% at λ = 600 nm. This efficiency is almost the same value as that obtained for BaSi2-pn homojunction solar cells. Therefore, we can state that we succeeded the demonstration of n+-AZO/p-BaSi2 heterojunction solar cells for the first time.[1] T. Suemasu and N. Usami, J. Phys. D: Appl. Phys. 50, 023001 (2017).[2] K. Kodama et al., Appl. Phys. Express 12, 041005 (2019).[3] T. Suemasu, Jpn. J. Appl. Phys. 54, 07JA01 (2015).[4] R. Santbergen et al., IEEE J. Photovoltaics, 7, no. 3, 919–926 (2017). Figure 1