For decades, scientists developing high-performance semiconductors for solar cells, LEDs and quantum devices have been plagued by lattice mismatch. Mismatched atomic lattices produce structural defects that drastically reduce device efficiency. Researchers long hypothesized that a nanoscale organic “molecular cushion” could buffer interfacial strain, yet direct atomic-scale observation of this buffer remained unachieved. Our joint team from Westlake University and ShanghaiTech University constructed a robust van der Waals epitaxy platform to resolve this issue and discovered spiral chiral heterostructures with distinctive luminescent properties.
The research originated from an accidental experimental phenomenon: scattered 3D perovskite microcrystals emerged on 2D crystal templates. While other researchers regarded these grains as experimental noise, I decided to investigate this anomaly with Prof. Enzheng Shi’s approval. I spent nearly a large of times conducting continuous solution screening experiments to distinguish controllable epitaxial growth from disordered random crystallization. After hundreds of trials, we found that precursor supersaturation and cation concentration govern nucleation behavior. By tuning solvent composition, solute dosage and temperature, we converted disordered crystal islands into highly oriented 3D single crystals grown tightly on 2D substrates. This coherent van der Waals epitaxy enables precise control over crystal density and size, ranging from microscale domains to continuous millimeter films. The method is universally applicable to most lead and tin halide perovskite systems with diverse compositions.
We then collaborated with Prof. Yi Yu’s team to solve the characterization challenge of beam-sensitive tin perovskites. We established a complete inert-gas-protected low-dose 4D-STEM testing pipeline to avoid sample oxidation and damage. We finally captured atomic images of the interfacial molecular cushion, confirming merely 2 % lattice mismatch and periodic dislocation-free interfaces. We further fabricated rectifier devices with a record rectification ratio of 6000000 and ultralow leakage current, which remained stable after 500 switching cycles. We also scaled the synthesis to obtain large-area transferable single-crystal films, connecting fundamental crystal research with industrial optoelectronic manufacturing.
The most surprising outcome of our robust van der Waals epitaxy lies in the spontaneous formation of spiral heterostructures, which redefines how we view crystal defects and chiral optoelectronics. In conventional epitaxy research, screw dislocations are universally treated as fatal structural flaws. They warp atomic lattices, introduce charge traps and ruin the uniformity of functional thin films, so researchers spend enormous effort eliminating them entirely during synthesis. Our system, however, completely reverses this mindset: the soft strain-buffering effect of interfacial molecular cushions grants the material unprecedented defect tolerance, turning screw dislocations into natural, self-assembled growth templates.
When we fine-tuned solution supersaturation to a mild range, the edges of dislocation steps on 2D perovskite sheets became preferential nucleation sites for 3D MASnI₃ crystals. The 3D phase faithfully follows the twisted staircase contour of the 2D template, winding upwards to form integrated bilayer spiral architectures with consistent left-handed or right-handed geometric configurations. We successfully produced pure single-handed spirals and even mixed left–right spiral domains on one single 2D template, giving us full geometric control over material chirality without any chiral organic additives or expensive chiral dopants.
This geometry-derived chirality delivers extraordinary circularly polarized luminescence (CPL), a core property demanded by next-generation optoelectronic hardware. Most existing chiral perovskite materials rely on specially designed chiral cations embedded inside crystal lattices, which raises manufacturing costs and often weakens charge transport performance. Our spiral heterostructures generate strong chiroptical signals purely from their twisted atomic stacking. Characteristic luminescence dissymmetry factors reach up to 0.29 for epitaxial 3D spiral segments, far exceeding the performance of many dopant-based chiral perovskite systems.
These spiral architectures carry tangible industrial value across multiple frontier fields. For advanced 3D display panels, circularly polarized light from spiral perovskites eliminates glare and enhances display color accuracy. In quantum communication, the distinct left/right circular light signals act as stable optical carriers for quantum state transmission. For information encryption, the switchable handedness of spiral luminescence creates unique, unclonable optical security tags impossible to counterfeit with ordinary luminescent materials.
This entire project is a reminder that transformative scientific breakthroughs often hide in ignored experimental anomalies. What started as a batch of unremarkable stray microcrystals evolved into a universal, highly robust coherent van der Waals epitaxy platform. This soft buffer layer fundamentally differs from rigid chemical bonding in traditional epitaxy, offering a brand-new design logic for all layered semiconductor heterostructures. Moving forward, our team will prioritize optimizing spiral film scalability, aiming to fabricate low-cost circularly polarized light detectors and integrated spiral optoelectronic chips for commercial 3D display and quantum encryption applications.