Data-Driven Insights into Copper Nanoparticle Morphology via Image Analysis and Chemometric Modeling of Electrochemical Syntheses

In fifteen words or less, explain the significance of this contribution (Novel Aspect).: Establishing quantitative predictive descriptors of nanoparticles from scanning electron microscopy images for machine learning

Abstract Text:

Understanding and controlling nanoparticle morphology is critical for tuning catalytic performance, yet the relationship between synthetic conditions, particle structure, and catalytic behavior remains poorly defined. Copper is the only known monometallic catalyst that produces appreciable quantities of high-value multi-carbon products, and its product distribution is strongly influenced by nanoparticle morphology, size, composition, and facet expression. Despite this, no quantitative framework currently links these morphological features to synthesis parameters or catalytic outcomes. Here, we present a chemometric framework to explore the morphological design space of electrochemically synthesized copper nanoparticles. Using a two-level fractional factorial design of experiments (DoE), we systematically investigate the influence of five synthetic variables – copper salt concentration, deposition time, overpotential, pH, and temperature – on nanoparticle morphology in a nitrate-directed system. This approach yields diverse structures ranging from cubes to multipods and quatrefoils. Secondary electron scanning electron microscopy (SEM) images were segmented via image analysis pipelines to isolate individual particles. Particles were then parameterized yielding 56 features capturing size, shape, and texture characteristics. Principal component analysis (PCA) enabled exploratory data analysis, while predictive synthesis-structure relationships were modeled using partial least squares discriminant analysis (PLS2-DA) and support vector machines (SVM). This work establishes a scalable approach for navigating complex synthesis landscapes through chemical imaging and multivariate analysis, offering a foundation for rational nanoparticle design in catalysis and beyond.