Search

We present a robotic chemical discovery system capable of navigating a chemical space based on a learned general association between molecular structures and reactivity, while incorporating a neural network model that can process data from online analytics and assess reactivity without knowing the identity of the reagents. Working in conjunction with this learned knowledge, our robotic platform is able to autonomously explore a large number of potential reactions and assess the reactivity of mixtures, including unknown chemical spaces, regardless of the identity of the starting materials. Through the system, we identified a range of chemical reactions and products, some of which were well-known, some new but predictable from known pathways, and some unpredictable reactions that yielded new molecules. The validation of the system was done within a budget of 15 inputs combined in 1018 reactions, further analysis of which allowed us to discover not only a new photochemical reaction but also a new reactivity mode for a well-known reagent (p-toluenesulfonylmethyl isocyanide, TosMIC). This involved the reaction of 6 equiv of TosMIC in a “multistep, single-substrate” cascade reaction yielding a trimeric product in high yield (47% unoptimized) with the formation of five new C–C bonds involving sp–sp2 and sp–sp3 carbon centers. An analysis reveals that this transformation is intrinsically unpredictable, demonstrating the possibility of a reactivity-first robotic discovery of unknown reaction methodologies without requiring human input., A robotic discovery system driven by machine learning is able to navigate a chemical space and successfully discover two unpredictable reactions.

Interpreting the outcome of chemistry experiments consistently is slow and frequently introduces unwanted hidden bias. This difficulty limits the scale of collectable data and often leads to exclusion of negative results, which severely limits progress in the field. What is needed is a way to standardize the discovery process and accelerate the interpretation of high-dimensional data aided by the expert chemist’s intuition. We demonstrate a digital Oracle that interprets chemical reactivity using probability. By carrying out >500 reactions covering a large space and retaining both the positive and negative results, the Oracle was able to rediscover eight historically important reactions including the aldol condensation, Buchwald–Hartwig amination, Heck, Mannich, Sonogashira, Suzuki, Wittig, and Wittig–Horner reactions. This paradigm for decoding reactivity validates and formalizes the expert chemist’s experience and intuition, providing a quantitative criterion of discovery scalable to all available experimental data.

We present a robotic chemical discovery system capable of learning the generalized notion of reactivity using a neural network model that can abstract the reactivity from the identity of the reagents. The system is controlled using an algorithm that works in conjunction with this learned knowledge, the robot was able to autonomously explore a large number of potential reactions and assess the reactivity of mixtures, including unknown datasets, regardless the identity of the starting materials. The system identified a range of chemical reactions and products, some of which were well-known, some new but predictable from known pathways, but also some unpredictable reactions that yielded new molecules. The search was done within a budget of 15 inputs combined in 1018 reactions, which allowed us not only to discover a new photochemical reaction, but also a new reactivity mode for a well-known reagent (p-toluenesulfonylmethyl isocyanide, TosMIC). This involved the reaction of six equivalents of TosMIC in a ‘multi-step, single-substrate’ cascade reaction yielding a trimeric product in high yield (47% unoptimized) with formation of five new C-C bonds involving sp-sp2 and sp-sp3 carbon centres. Analysis reveals that this transformation is intrinsically unpredictable, demonstrating the possibility of reactivity-first robotic discovery of unknown reaction methodologies without requiring human input.

We present an artificial intelligence, built to autonomously explore chemical reactions in the laboratory using deep learning. The reactions are performed automatically, analysed online, and the data is processed using a convolutional neural network (CNN) trained on a small reaction dataset to assess the reactivity of reaction mixtures. The network can be used to predict the reactivity of an unknown dataset, meaning that the system is able to abstract the reactivity assignment regardless the identity of the starting materials. The system was set up with 15 inputs that were combined in 1018 reactions, the analysis of which lead to the discovery of a ‘multi-step, single-substrate’ cascade reaction and a new mode of reactivity for methylene isocyanides. p-Toluenesulfonylmethyl isocyanide (TosMIC) in presence of an activator reacts consuming six equivalents of itself to yield a trimeric product in high (unoptimized) yield (47%) with formation of five new C-C bonds involving sp-sp2 and sp-sp3 carbon centres. A cheminformatics analysis reveals that this transformation is both highly unpredictable and able to generate an increase in complexity like a one-pot multicomponent reaction.

The sorting of objects into groups is a fundamental operation, critical in the preparation and purification of populations of cells, crystals, beads, or droplets, necessary for research and applications in biology, chemistry, and materials science. Most of the efforts exploring such purification have focused on two areas: the degree of separation and the measurement precision required for effective separation. Conventionally, achieving good separation ultimately requires that the objects are considered one by one (which can be both slow and expensive), and the ability to measure the sorted objects by increasing sensitivity as well as reducing sorting errors. Here we present an approach to sorting that addresses both critical limitations with a scheme that allows us to approach the theoretical limit for the accuracy of sorting decisions. Rather than sorting individual objects, we sort the objects in ensembles, via a set of registers which are then in turn sorted themselves into a second symmetric set of registers in a lossless manner. By repeating this process, we can arrive at high sorting purity with a low set of constraints. We demonstrate both the theory behind this idea and identify the critical parameters (ensemble population and sorting time), and show the utility and robustness of our method with simulations and experimental systems spanning several orders of scale, sorting populations of macroscopic beads and microfluidic droplets. Our method is general in nature and simplifies the sorting process, and thus stands to enhance many different areas of science, such as purification, enrichment of rare objects, and separation of dynamic populations.

Herein we present a chemistry capable robot built with a standard set of hardware and software protocols that can be networked to coordinate many chemical experiments in real time, such that the different chemical reactions can be distributed over many sites simultaneously. We demonstrate how multiple chemical processes can be done with two internet connected robots collaboratively, exploring a set of azo-coupling reactions in a fraction of time needed for a single robot, as well as encoding and decoding information into a network of oscillating BZ reactions transferring a message between two different locations using chemical reactions. The system can also be used to assess the reproducibility of chemical reactions and discover new reaction outcomes using game playing to explore a list of reaction conditions not accessible when the robots instead take it in turn to each a pre-define reaction from a list.

Cinchona alkaloid derivatives featuring a guanidinium group in diverse positions efficiently catalyze the cleavage of the RNA model compound 2-hydroxypropyl p-nitrophenyl phosphate (HPNP). Their high catalytic efficiency as phosphodiesterases and the potentiometric and kinetic investigations indicate the existence of a high degree of cooperation between the guanidinium group and the quinuclidine moiety with the operation of a general acid/general base mechanism. The performance of these compounds was investigated and compared in the kinetic resolution of HPNP. These data were also compared with the results of DFT calculations on the transition states of the transesterification reaction that, in part, predict and rationalize the experimental data.