Publication date: 15th December 2025
Next-generation III–V quantum dots (QDs) with enhanced optoelectronic properties and stability are expected to emerge from a complete understanding and control of their surface characteristics—specifically their three-dimensional (3D) atomic configuration, chemical composition, and bonding interactions. Although electron tomography is among the most powerful techniques for reconstructing the 3D atomic structure of nanomaterials, its application to III–V QDs remains challenging, as these systems are highly susceptible to structural degradation under prolonged electron-beam exposure. As a result, current 3D structural models of III–V QDs are typically inferred from 2D low-dose annular dark-field scanning transmission electron microscopy (ADF-STEM) projections through visual comparison and iterative manual adjustments.
To move beyond this qualitative and time-consuming approach, we developed a highly efficient AI-driven workflow that integrates quantitative transmission electron microscopy with molecular modeling. In this talk, we show how this combined experimental–theoretical protocol enables the generation of highly accurate and experimentally validated 3D models of novel InP QDs, including reliable reconstructions of their surface and interface atomic configurations.
J.Z. acknowledges FWO grant 12AA526N
