Ag Surface Segregation in Sub-10-nm Bimetallic AuAg Nanoparticles Quantified by STEM-EDS and Machine Learning: Implications for Fine-Tuning Physicochemical Properties for Plasmonics and Catalysis Applications

Abstract

Mono-and multimetallic nanoparticles have been extensively studied in various fields due to their tunable physicochemical properties and potential for replacing expensive metals with more abundant and affordable ones. The chemical structure, i.e., the spatial distribution of elements inside nanoparticles, plays a crucial role in defining their properties, particularly in catalytic processes. However, accurately determining the spatial chemical distribution within sub-10-nm bimetallic nanoparticles remains a challenge. In this study, we have used scanning transmission electron microscopy associated with energydispersive spectroscopy to acquire hyperspectral images of gold− silver alloy nanoparticles in the 3−10-nm size range. We have quantified the chemical composition as a function of radial position; Ag enrichment toward the nanoparticle surface is robustly confirmed by statistical analysis, error bars, and nonoverlapping 3-sigma uncertainty intervals at the nanoparticle center and surface. Two complementary machine learning analyses (principal component analysis and non-negative matrix factorization) reveal that our experiments contain latent information on subtle composition variations inside the particles. The proposed data analysis procedures have also been validated by simulated data sets. These findings pave the way for more precise structural and chemical investigations of alloys on the nanoscale.

A 30-Year Review on Nanocomposites: Comprehensive Bibliometric Insights into Microstructural, Electrical, and Mechanical Properties Assisted by Artificial Intelligence

“…Machine learning (ML) is becoming a critical tool in various fields of nanomaterials research, such as nanoindentation [ 317 , 318 , 319 , 320 , 321 , 322 , 323 , 324 , 325 , 326 , 327 , 328 ], nanorobotics [ 329 , 330 , 331 , 332 , 333 , 334 ], and nanosensor [ 335 , 336 , 337 , 338 , 339 , 340 ] development. Its ability to analyze and interpret complex patterns from large datasets is particularly beneficial in advancing areas like nanostructured materials analysis, nanoscale manufacturing processes, and the development of nanotechnology applications in medicine and environmental monitoring [ 327 , 341 , 342 , 343 , 344 , 345 , 346 , 347 , 348 , 349 , 350 , 351 , 352 , 353 , 354 , 355 , 356 , 357 , 358 , 359 , 360 , 361 , 362 , 363 , 364 , …”