Interfacial Behavior from the Atomic Blueprint: Machine Learning-Guided Design of Spatially Functionalized a-SiO2 Surfaces

Abstract: Precise control over surface chemistry is essential for tuning interfacial behavior in technologies ranging from catalysis and protective coatings to energy conversion systems. Although chemical functionalization of alpha-quartz (alpha-SiO2) with hydroxyl (OH) and methyl (CH3) groups has been extensively studied, the impact of their spatial distribution at the atomic scale remains largely uncharted. In this work, we integrate density functional theory (DFT), ab initio molecular dynamics (AIMD), and on-the-fly machine-learned force fields (MLFFs) to systematically investigate how different arrangements of OH/CH3 groups modulate surface properties. Our results reveal that spatial patterning governs the formation of hydrogen-bonding networks, alters vibrational signatures, and has a significant influence on the thermodynamic stability of the functionalized surfaces. The MLFF framework enables high-fidelity simulations at unprecedented scales, bridging the gap between quantum accuracy and statistical sampling. By uncovering structure-property relationships inaccessible to conventional approaches, this study establishes spatial arrangement of functionalized groups as a critical and tunable design axis, paving the way for the predictive engineering of silica-based materials with optimized interfacial performance.