Programmable Light-Driven Color Tuning of Perovskite Quantum Dots.

Precise, sustainable, and scalable bandgap tuning of metal halide perovskite (MHP) nanocrystals (NCs) is critical for their integration into advanced optoelectronic and photocatalytic systems. Photoinduced anion exchange reactions (PIAERs) enable uniform halide delivery with spatiotemporal control, yet their complex parameter space has limited mechanistic understanding and rational optimization. Here, we introduce a material-efficient fluidic self-driving laboratory (FSDL) that integrates a single-droplet microfluidic photoreactor, multimodal in situ spectroscopy, and a multiobjective Bayesian optimization framework to navigate the ∼106-dimensional design space of PIAERs autonomously. Through machine learning-guided exploration of this coupled parameter landscape, the FSDL rapidly identifies synthesis conditions that simultaneously maximize photoluminescence quantum yield and minimize emission line width for any target emission wavelength across the UV-visible spectrum. Mechanistic trends derived from surrogate modeling revealed distinct kinetic regimes for Br-→Cl- and Br-→I- exchanges, governed respectively by reaction time and photon flux, enabling reaction-specific tuning strategies. Critically, synthesis protocols discovered at the droplet scale (∼10 μL) were directly translated to continuous-flow operation (∼50-250 mL·day-1) without reoptimization, maintaining optical performance and establishing knowledge scalability across 4 orders of magnitude in throughput, with low energy demand. This study demonstrates a reproducible, mechanistically informed, and industrially relevant route for programmable light-directed bandgap tuning in MHP NCs.