(661i) An Autonomous Fluidic Lab for Accelerated Synthesis Science Studies of Eco-Friendly Semiconductor Nanocrystals

Copper (Cu)-based metal halide perovskite (MHP) nanocrystals (NCs) have emerged as environmentally benign alternatives to traditional lead (Pb)-based counterparts, offering promising features such as wide bandgaps, large Stokes shifts, and high emission stability. Despite their potential, achieving high photoluminescence quantum yield (PLQY) in Cu-based MHP NCs remains a major bottleneck, limiting their applicability in advanced optoelectronic and energy-related technologies. Recent efforts have shown that incorporating metal halide additives into precursor formulations of Cu-based MHP NCs can enhance their optical performance;^1–3 however, this strategy significantly expands the chemical parameter space, rendering conventional batch-based, trial-and-error approaches inefficient, labor-intensive, and resource-demanding.

To address this challenge, we developed an autonomous fluidic lab (AFL) that autonomously explores and optimizes the synthesis of Cu-based MHP NCs with metal halide additive-assisted strategies. The AFL integrates a modular droplet-based microfluidic reactor, in-situ real-time optical characterization, and a machine learning (ML)-guided optimization loop. By employing ensemble neural network-driven Bayesian optimization, the developed AFL efficiently navigates complex synthesis conditions, significantly accelerating the discovery process while minimizing material consumption and experimental waste.

We applied the AFL to three distinct precursor chemistries for the synthesis of Cs₃Cu₂I₅ NCs using zinc iodide (ZnI₂) as the metal halide additive. The real-time data stream enabled the construction of predictive digital twin models that offered mechanistic insights into the role of additive chemistry during NC formation. Iterative closed-loop optimization led to the identification of synthesis conditions that yielded Cs₃Cu₂I₅ NCs with a post-purification PLQY of ~61%, representing a more than 2.5-fold improvement over previously reported values.⁴

This study demonstrates the power of autonomous, microfluidics-enabled platforms in accelerating the discovery and optimization of eco-friendly semiconductor NCs. The integration of ML, real-time feedback, and miniaturized flow chemistry establishes a generalizable strategy for advancing next-generation photonic and energy materials.

References:

(1) Lian, L.; Zheng, M.; Zhang, W.; Yin, L.; Du, X.; Zhang, P.; Zhang, X.; Gao, J.; Zhang, D.; Gao, L.; Niu, G.; Song, H.; Chen, R.; Lan, X.; Tang, J.; Zhang, J. Efficient and Reabsorption‐Free Radioluminescence in Cs₃Cu₂I₅ Nanocrystals with Self‐Trapped Excitons. Adv. Sci. 2020, 7 (11), 2000195. https://doi.org/10.1002/advs.202000195.

(2) Li, C.-X.; Cho, S.-B.; Kim, D.-H.; Park, I.-K. Monodisperse Lead-Free Perovskite Cs₃Cu₂I₅ Nanocrystals: Role of the Metal Halide Additive. Chem. Mater. 2022, 34 (15), 6921–6932. https://doi.org/10.1021/acs.chemmater.2c01318.

(3) Qu, K.; Lu, Y.; Ran, P.; Wang, K.; Zhang, N.; Xia, K.; Zhang, H.; Pi, X.; Hu, H.; Yang, Y. (Michael); He, Q.; Yin, J.; Pan, J. Zn (II)‐Doped Cesium Copper Halide Nanocrystals with High Quantum Yield and Colloidal Stability for High‐Resolution X‑Ray Imaging. Adv. Opt. Mater. 2023, 11 (7), 2202883. https://doi.org/10.1002/adom.202202883.

(4) Sadeghi, S.; Bateni, F.; Kim, T.; Son, D. Y.; Bennett, J. A.; Orouji, N.; Punati, V. S.; Stark, C.; Cerra, T. D.; Awad, R.; Delgado-Licona, F.; Xu, J.; Mukhin, N.; Dickerson, H.; Reyes, K. G.; Abolhasani, M. Autonomous Nanomanufacturing of Lead-Free Metal Halide Perovskite Nanocrystals Using a Self-Driving Fluidic Lab. Nanoscale 2024, 16 (2), 580–591. https://doi.org/10.1039/D3NR05034C.