Machine Learning-Guided Construction of Cobalt Hexacyanoferrate Space-Confined in Hollow Carbon with Yolk–Shell Structure for Advanced Aqueous Zinc-Ion Batteries

Carbon-modified materials have shown great promise as cathode candidates for aqueous zinc-ion batteries (AZIBs). However, the indistinctness of screening standards along with the time-consuming experimental procedures inevitably restricts their practical uses. With the increasing prevalence of artificial intelligence, machine learning (ML) offers an avenue to enhance the performance of energy storage devices. In this study, we have presented an ML-directed method to quicken the development pace of carbon-modified materials. Specifically, encapsulation or physical mixing, specific surface area after carbon modification, carbon mass fraction, specific surface area before carbon modification, and pore size are selected as screening parameters to realize the efficient screening of carbon-modified materials. Combined with experimental verification, a three-dimensional conductive yolk-shell structured composite, composed of hollow mesoporous carbon spheres as the shell and cobalt hexacyanoferrate (Co HCF) as the yolk, is designed. The composite exhibits excellent zinc-ion storage performance: a reversible capacity of 100 mAh g^-1 at 0.05 A g^-1 and a high-capacity retention of 75% after 1500 cycles at 2 A g^-1, significantly outperforming pure Co HCF. This work integrates ML technology with material development, providing a strategy for accelerating the screening of carbon-ba<b>s</b>ed cathode materials for AZIBs and paving a promising research path.