Optimizing Pulsed Laser Deposition of LaVO₃ by Active Learning

The synthesis of emerging correlated quantum materials poses inherent challenges due to electron correlations where small changes to composition and microstructure drive large variations in function. Studies on the pulsed laser deposition (PLD) of LaVO₃, a correlated perovskite, have reported growth parameters that address these difficulties, but no consensus on optimal growth conditions exists. Several studies have also reported the dependence of optical, electronic, and structural properties on growth parameters, underscoring the need to link growth conditions to structural and functional properties. Moreover, synthesizing optimal LaVO₃ contends with the effects of compositional and microstructural variability by minimizing oxygen vacancies, cation non-stoichiometry, and competing phases among other limitations to film quality. Here, we show the integration of active learning with multimodal characterization facilitates the optimization of LaVO₃ PLD. Our approach also provides insights into the film quality landscape as a function of growth parameters, enabling inferences about growth mechanisms such as film-substrate interactions. Functional properties strongly depend on growth conditions with implications for LaVO₃-based applications, such as altermagnets where orbital orders and local symmetries dominate functionality. Adaptable to other material classes and growth techniques, our methodology represents an efficient and quantitative approach to thin film synthesis.