Closed-loop, autonomous experimentation enables accelerated and material-efficient exploration of large reaction spaces without the need for user intervention. However, autonomous exploration of advanced materials with complex, multi-step processes and data sparse environments remains a challenge. In this work, we present AlphaFlow, a self-driven fluidic lab capable of autonomous discovery of complex multi-step chemistries. AlphaFlow uses reinforcement learning integrated with a modular microdroplet reactor capable of performing reaction steps with variable sequence, phase separation, washing, and continuous in-situ spectral monitoring. To demonstrate the power of reinforcement learning toward high dimensionality multi-step chemistries, we use AlphaFlow to discover and optimize synthetic routes for shell-growth of core-shell semiconductor nanoparticles, inspired by colloidal atomic layer deposition (cALD). Without prior knowledge of conventional cALD parameters, AlphaFlow successfully identified and optimized a novel multi-step reaction route, with up to 40 parameters, that outperformed conventional sequences. Through this work, we demonstrate the capabilities of closed-loop, reinforcement learning-guided systems in exploring and solving challenges in multi-step nanoparticle syntheses, while relying solely on in-house generated data from a miniaturized microfluidic platform. Further application of AlphaFlow in multi-step chemistries beyond cALD can lead to accelerated fundamental knowledge generation as well as synthetic route discoveries and optimization.

Colloidal organic/inorganic metal-halide perovskite nanocrystals have recently emerged as a potential low-cost replacement for the semiconductor materials in commercial photovoltaics and light emitting diodes. However, unlike III–V and IV–VI semiconductor nanocrystals, studies of colloidal perovskite nanocrystals have yet to develop a fundamental and comprehensive understanding of nucleation and growth kinetics. Here, we introduce a modular and automated microfluidic platform for the systematic studies of room-temperature synthesized cesium–lead halide perovskite nanocrystals. With abundant data collection across the entirety of four orders of magnitude reaction time span, we comprehensively characterize nanocrystal growth within a modular microfluidic reactor. The developed high-throughput screening platform features a custom-designed three-port flow cell with translational capability for in situ spectral characterization of the in-flow synthesized perovskite nanocrystals along a tubular microreactor with an adjustable length, ranging from 3 cm to 196 cm. The translational flow cell allows for sampling of twenty unique residence times at a single equilibrated flow rate. The developed technique requires an average total liquid consumption of 20 μL per spectra and as little as 2 μL at the time of sampling. It may continuously sample up to 30 000 unique spectra per day in both single and multi-phase flow formats. Using the developed plug-and-play microfluidic platform, we study the growth of cesium lead trihalide perovskite nanocrystals through in situ monitoring of their absorption and emission band-gaps at residence times ranging from 100 ms to 17 min. The automated microfluidic platform enables a systematic study of the effect of mixing enhancement on the quality of the synthesized nanocrystals through a direct comparison between single- and multi-phase flow systems at similar reaction time scales. The improved mixing characteristics of the multi-phase flow format results in high-quality perovskite nanocrystals with kinetically tunable emission wavelength, ranging as much as 25 nm at equivalent residence times. Further application of this unique platform would allow rapid parameter optimization in the colloidal synthesis of a wide range of nanomaterials (e.g., metal or semiconductor), that is directly transferable to continuous manufacturing in a numbered-up platform with a similar characteristic length scale.

With the rise of self-driving labs (SDLs) and automated experimentation across chemical and materials sciences, there is a considerable challenge in designing the best autonomous lab for a given problem based on published studies alone. Determining what digital and physical features are germane to a specific study is a critical aspect of SDL design that needs to be approached quantitatively. Even when controlling for features such as dimensionality, every experimental space has unique requirements and challenges that influence the design of the optimal physical platform and algorithm. Metrics such as optimization rate are therefore not necessarily indicative of the capabilities of an SDL across different studies. In this perspective, we highlight some of the critical metrics for quantifying performance in SDLs to better guide researchers in implementing the most suitable strategies. We then provide a brief review of the existing literature under the lens of quantified performance as well as heuristic recommendations for platform and experimental space pairings.