Facile Synthesis of Ultrasmooth Au Nanospheres into Macroscopic 3D Supercrystals for Machine-Learning-Driven Analysis of Thiram in Soil

Self-assembly of nanocrystals into a macroscopic 3D superlattice has emerged as a promising nanostructure due to its collective optical and electrical properties. In this work, a macroscale supercrystal was constructed under constant temperature and humidity with an ultrasmooth Au nanosphere (NS) as a building block synthesized by a facile route. In particular, a reproducible and facile strategy was developed to fabricate highly spherical Au NSs with nanoscale size tenability. Importantly, this method does not require a syringe pump and breaks the limitation of instrumentation for the synthesis of Au NS, significantly contributing to the promotion and popularization of Au NSs. Meanwhile, an operable and cost-effective droplet slow evaporation self-assembly approach was applied to form a homogeneous 3D superlattice array against coffee rings with an enhancement factor as high as 1.37 × 10⁸ verified by FDTD simulation results. Furthermore, integrating a convolutional neural network (CNN) model with the ultrasensitive 3D superlattice SERS platform, the precise and wide-range concentration prediction of the pesticide thiram in soil was successfully demonstrated, breaking the limitation of high-concentration saturation in the conventional concentration-intensity fitting curve. Consequently, this innovative Au NS 3D supercrystal facilitates the fabrication of novel nanoassemblies with ideal plasmonic functions for extensive applications in the fields of food safety, environment, medicine, and biology.