Covalent organic frameworks (COFs) have recently emerged as a new class of crystalline porous materials with many potential applications. The development of facile and effective synthetic methods of COFs is highly desirable for their large-scale applications. Herein, we demonstrate the room temperature batch synthesis of three classical two-dimensional (2D) COFs with various types of linkage, namely, COF-LZU1 (imine-linked), TpPa-1 (enamine-linked), and N 3 -COF (azinelinked). These obtained COFs exhibit good crystallinity and high porosity comparable to their counterparts synthesized solvothermally at higher temperatures. The facile formation of these COFs under such mild synthetic conditions can be attributed to (1) high solubility of monomers and (2) the strong π−π stacking interactions between monomers and π-systems of oligomers during the initial and the subsequent error-correction crystallization process. Based on this conclusion, two new iminelinked COFs named NUS-14 and NUS-15 were successfully synthesized with good crystallinity under ambient conditions. Moreover, continuous flow synthesis has been demonstrated in COF-LZU1 with a production rate of 41 mg h −1 at an extremely high space-time yield (STY) of 703 kg m −3 day −1 . This study represents the first example of synthesizing COFs by continuous processes, which sheds light on the scaled-up synthesis of these promising materials.

In materials science, the discovery of recipes that yield nanomaterials with defined optical properties is costly and time-consuming. In this study, we present a two-step framework for a machine learning-driven high-throughput microfluidic platform to rapidly produce silver nanoparticles with the desired absorbance spectrum. Combining a Gaussian process-based Bayesian optimization (BO) with a deep neural network (DNN), the algorithmic framework is able to converge towards the target spectrum after sampling 120 conditions. Once the dataset is large enough to train the DNN with sufficient accuracy in the region of the target spectrum, the DNN is used to predict the colour palette accessible with the reaction synthesis. While remaining interpretable by humans, the proposed framework efficiently optimizes the nanomaterial synthesis and can extract fundamental knowledge of the relationship between chemical composition and optical properties, such as the role of each reactant on the shape and amplitude of the absorbance spectrum.

Microfluidic tools and techniques for manipulating fluid droplets have become core to many scientific and technological fields. Despite the plethora of existing approaches to fluidic manipulation, non-Newtonian fluid phenomena are rarely taken advantage of. Here we introduce embedded droplet printing—a system and methods for the generation, trapping, and processing of fluid droplets within yield-stress fluids, materials that exhibit extreme shear thinning. This technique allows for the manipulation of droplets under conditions that are simply unattainable with conventional microfluidic methods, namely the elimination of exterior influences including convection and solid boundaries. Because of this, we believe embedded droplet printing approaches an ideal for the experimentation, processing, or observation of many samples in an “absolutely quiescent” state, while also removing some troublesome aspects of microfluidics including the use of surfactants and the complexity of device manufacturing. We characterize a model material system to understand the process of droplet generation inside yield-stress fluids and develop a nascent set of archetypal operations that can be performed with embedded droplet printing. With these principles and tools, we demonstrate the benefits and versatility of our method, applying it toward the diverse applications of pharmaceutical crystallization, microbatch chemical reactions, and biological assays.

Process intensification in a triphasic millireactor for nanoparticle-catalyzed gas–liquid reactions with facile catalyst recovery and recycle is demonstrated.

Robust, non-fouling, litres-per-day continuous synthesis of catalytically active palladium nanoparticles using triphasic segmented flow in a hybrid milli-meso flow reactor.

No abstract

<p>In materials science, the discovery of recipes that yield nanomaterials with defined optical properties is costly and time-consuming. In this study, we present a two-step framework for a machine learning driven high-throughput microfluidic platform to rapidly produce silver nanoparticles with a desired absorbance spectrum. Combining a Gaussian Process based Bayesian Optimization (BO) with a Deep Neural Network (DNN), the algorithmic framework is able to converge towards the target spectrum after sampling 120 conditions. Once the dataset is large enough to train the DNN with sufficient accuracy in the region of the target spectrum, the DNN is used to predict the colour palette accessible with the reaction synthesis. While remaining interpretable<i> </i>by humans, the proposed framework efficiently optimizes the nanomaterial synthesis, and can extract fundamental knowledge of the relationship between chemical composition and optical properties, such as the role of each reactant on the shape and amplitude of the absorbance spectrum.</p>

Combining high‐throughput experiments with machine learning accelerates materials and process optimization toward user‐specified target properties. In this study, a rapid machine learning‐driven automated flow mixing setup with a high‐throughput drop‐casting system is introduced for thin film preparation, followed by fast characterization of proxy optical and target electrical properties that completes one cycle of learning with 160 unique samples in a single day, a >10× improvement relative to quantified, manual‐controlled baseline. Regio‐regular poly‐3‐hexylthiophene is combined with various types of carbon nanotubes, to identify the optimum composition and synthesis conditions to realize electrical conductivities as high as state‐of‐the‐art 1000 S cm−1. The results are subsequently verified and explained using offline high‐fidelity experiments. Graph‐based model selection strategies with classical regression that optimize among multi‐fidelity noisy input‐output measurements are introduced. These strategies present a robust machine‐learning driven high‐throughput experimental scheme that can be effectively applied to understand, optimize, and design new materials and composites.