Machine learning–accelerated design and synthesis of polyelemental heterostructures

Abstract

Machine learning accelerates materials discovery by suggesting targets, yielding exceptionally complex biphasic nanoparticles.

Data‐Driven Materials Innovation and Applications

“…Researchers performed the data-driven design of materials through theoretical computing methods, which integrate the databases, ML algorithm, and high-throughput DFT calculations. Unexplored materials still need to be discovered and designed through data-driven strategies, such as alloys, [574] heterostructures, [575] 2D materials, [576] and even doped materials. [10] Data-driven studies of thermodynamic stability have important implications for the identification of novel materials, either as detailed complex features or simpler scalar features.…”

New challenges in oxygen reduction catalysis: a consortium retrospective to inform future research

“…After being trained on known data from SEM nanoparticle micrographs, a Gaussian-process-based agent predicted 18 new compositions that were confirmed to have a single interface of 19 that were experimentally tested, including the most complex bi-phasic nanoparticle known that contained seven elements. 118…”

Site-selective modification of metallic nanoparticles

“…81 It is possible to engineer various positions of the phases based on the composition and annealing time into multimetallic particles with various metallic junctions and predict which metals will incorporate into one another, providing a seemingly endless library of possible combinations which can be predicted by machine learning. 82…”

“…81 It is possible to engineer various positions of the phases based on the composition and annealing time into multimetallic particles with various metallic junctions and predict which metals will incorporate into one another, providing a seemingly endless library of possible combinations which can be predicted by machine learning. 82 Colloidal methods for incorporating several metal ions into a single particle can be achieved via cation exchange methods for metal oxide materials. Schaak and co-workers have devised schemes to incorporate up to 11 different cations into copper sulfide nanoparticles, all while preserving their colloidal stability and shape structures (Fig.…”