Autonomous Multi‐Objective Nanoscale Characterization of Combinatorial (Al, Sc, B)N Films Reveals Composition‐Dependent Ferroelectric Regimes

ABSTRACT Combinatorial materials libraries span multidimensional composition spaces efficiently, but high‐resolution functional characterization across them remains difficult. Here, we demonstrate an autonomous scanning probe microscopy platform for multi‐objective exploration of ferroelectric behavior in a ternary Al 1‐x‐y Sc x B y N thin‐film library. Automated piezoresponse force microscopy (PFM), guided by Bayesian experiment planning, maps surface morphology, piezoresponse amplitude, and local switching‐loop area across composition space. Rather than seeking a single optimum, the workflow reconstructs the Pareto landscape between structural stability and ferroelectric functionality. The autonomous decision loop operates on the PFM measurement alone. Wavelength‐dispersive spectroscopy and photoluminescence (PL) provide spatially registered correlative maps of the same library. X‐ray diffraction then tests the composition, defect, and property relations the exploration uncovers. The data resolve two composition‐dependent regimes. Below about 35% Sc on the cation sublattice, the films remain smooth and well textured. Switching descriptors then correlate with a defect‐sensitive PL signal. Above this composition, X‐ray diffraction shows a loss of crystalline texture which limits the switching response. The autonomous exploration located this boundary without human intervention, in six hours of instrument time against about a month on a grid. Combinatorial libraries thereby become tractable maps of coupled composition, defect, structure, and property relationships.