Automating van der Waals Assembly: A Machine-Vision Guided Setup for High-Yield 2D Material Heterostructures

The fabrication of van der Waals heterostructures based on two-dimensional (2D) materials has become increasingly critical for advancing quantum studies and next-generation device applications. However, the conventional manual assembly process is notoriously time-consuming, suffers from high failure rates, and is significantly complicated by the necessity of performing delicate deterministic transfers inside the constrained environment of a glovebox. To overcome these critical bottlenecks, we present a novel automated transfer setup designed to streamline the entire heterostructure fabrication pipeline. Our system autonomously scans substrates to identify target 2D flakes and systematically extracts multi-modal metrological data, including precise spatial coordinates, flake size, photoluminescence (PL) spectra, and dark-field imaging. This comprehensive material data is autonomously registered into a centralized database. Researchers can then interact with a Graphic User Interface (GUI) to intuitively browse through the stored crystals and select the ideal components for their desired heterostructure stack. Once the selection is finalized, the setup utilizes advanced machine vision to autonomously pick and place the selected flakes, building the complex heterostructure with high precision and minimal human intervention. By eliminating the manual constraints of optical searching and alignment, this automated approach significantly reduces fabrication time, maximizes yield, and provides a scalable, reliable pathway to produce complex 2D quantum material platforms.