This work reports on the properties of heterojunctions consisting of n-type Ga₂O₃ layers, deposited using ultrasonic spray pyrolysis at high temperature from water-based solution, combined with p-type NiO and Cu₂O counterparts, deposited by radio frequency and reactive, direct-current magnetron sputtering, respectively. After a comprehensive investigation of the properties of the single layers, the fabricated junctions on indium tin oxide (ITO)-coated glass showed high rectification, with an open circuit voltage of 940 mV for Ga₂O₃/Cu₂O and 220 mV for Ga₂O₃/NiO under simulated solar illumination. This demonstrates in praxis the favorable band alignment between the sprayed Ga₂O₃ and Cu₂O, with small conduction band offset, and the large offsets anticipated for both energy bands in the case of Ga₂O₃/NiO. Large differences in the ideality factors between the two types of heterojunctions were observed, suggestive of distinctive properties of the heterointerface. Further, it is shown that the interface between the high-temperature-deposited Ga₂O₃ and the ITO contact does not impede electron transport, opening new possibilities for the design of solar cell and optoelectronic device architectures.

The discovery of new materials with a well-defined set of properties is a work-intensive and time-consuming task, when relying on conventional experimental routines. The employment of high-throughput (HT) techniques speeds up the screening of material properties and facilitates the generation of material libraries for data-driven optimization. Ultrasonic spray pyrolysis (USP) is an up-scalable technique, well-suited for creating combinatorial thin films, enabling two-dimensional variation of the film composition and/or thickness which can be used in an HT approach. In this work, we upgraded a commercial USP tool with a custom-built, electronically controlled pump system that allows for a gradual composition change of the precursor solution during the deposition process. The capabilities of the realized equipment are demonstrated by depositing a 2D composition gradient of copper-gallium-iron oxides, with demonstrated process reliability and material stability under ambient conditions. This elemental system is relevant for applications in photovoltaic, photo-electrochemical or optoelectronic devices, where the materials can be used as transparent electrodes, charge carrier selective or absorber layers, depending on the obtained phases and composition. Spatially resolved elemental quantification of the 2D deposits is performed by HT-SEM/EDS, revealing a concentration distribution of the metal oxides with a stoichiometry range of Cu_14–49Ga_21–59Fe_14–44. Crystallographic information is gathered through HT-XRD point measurements which yield maps of identified structures, i.e., spinel, delafossite, CuO, and Cu₂O. The film thickness distribution in a range of 400–650 nm is obtained through Monte Carlo simulations of EDS measurements and verified using tactile profilometry. The optical properties of the thin films are determined by HT-FT transmission measurements, yielding maps of band gap energies ranging from 1.9 to 3.0 eV. The presented platform facilitates high-throughput screening of solution-based semiconductor films through combinatorial deposition and (semi-)automatized analysis, enabling a 10- to 100-fold speed-up with over 96 % compositional reproducibility and over 98 % reproducibility of the evaluated band gap energy.

The accelerated discovery and optimization of materials relies on the integration of advanced experimental techniques with data-driven methodologies. In this work, Bayesian optimization (BO) is applied to optimize the ultrasonic spray pyrolysis (USP) process for the deposition of copper oxides, targeting high-quality Ga₂O₃-Cu₂O heterojunctions for optoelectronic applications. By employing BO with an initial data set of 12 samples and conducting 4 USP parameter optimization cycles, significant improvements in device performance are achieved, with the open-circuit voltage increasing from 288 to 804 mV. During the optimization process, the performance of the model declines, necessitating the identification of a reliable subset of samples from the full data set. Through the application of BO, the cross-validation error of the model is minimized based on the sample selection, whereby accuracy is restored and generalizability is achieved. The subsequent model evaluation reveals two distinct deposition regimes, each characterized by unique process conditions, leading to specific material properties and device performances. These findings not only demonstrate the application of a data-driven experimental workflow in the context of thin film deposition but also highlight the importance of robust data validation and model evaluation.

Accelerating materials discovery and optimization is crucial for transitioning to sustainable energy conversion and storage. In this regard, materials acceleration platforms (MAPs) can significantly shorten the discovery process, cutting material and labor costs (Aspuru-Guzik & Persson, 2018). Combinatorial and high-throughput methods are instrumental in developing said MAPs, enabling autonomous operation and the generation of large datasets (Maier et al., 2007). Therefore, in a previous work, we developed combinatorial deposition and analysis techniques for the discovery of new semiconductor materials (Wolf et al., 2023). To drive further innovation in the field, COHESIVM was created, which facilitates combinatorial analysis of material and device properties through the following key features: