Explore the vast range of titles available.

Three-dimensional (3D) printing has the potential to transform science and technology by creating bespoke, low-cost appliances that previously required dedicated facilities to make. An attractive, but unexplored, application is to use a 3D printer to initiate chemical reactions by printing the reagents directly into a 3D reactionware matrix, and so put reactionware design, construction and operation under digital control. Here, using a low-cost 3D printer and open-source design software we produced reactionware for organic and inorganic synthesis, which included printed-in catalysts and other architectures with printed-in components for electrochemical and spectroscopic analysis. This enabled reactions to be monitored in situ so that different reactionware architectures could be screened for their efficacy for a given process, with a digital feedback mechanism for device optimization. Furthermore, solely by modifying reactionware architecture, reaction outcomes can be altered. Taken together, this approach constitutes a relatively cheap, automated and reconfigurable chemical discovery platform that makes techniques from chemical engineering accessible to typical synthetic laboratories. A low-cost 3D printer is used to combine chemical reactions and the reactor to produce an active ‘reactionware’ system for organic and inorganic synthesis. Active elements such as catalysts can be incorporated into the walls of printed reactors, and other printed-in components that enable electrochemical and spectroscopic analysis can also be included.

Most chemical experiments are planned by human scientists and therefore are subject to a variety of human cognitive biases 1 , heuristics 2 and social influences 3 . These anthropogenic chemical reaction data are widely used to train machine-learning models 4 that are used to predict organic 5 and inorganic 6 , 7 syntheses. However, it is known that societal biases are encoded in datasets and are perpetuated in machine-learning models 8 . Here we identify as-yet-unacknowledged anthropogenic biases in both the reagent choices and reaction conditions of chemical reaction datasets using a combination of data mining and experiments. We find that the amine choices in the reported crystal structures of hydrothermal synthesis of amine-templated metal oxides 9 follow a power-law distribution in which 17% of amine reactants occur in 79% of reported compounds, consistent with distributions in social influence models 10 – 12 . An analysis of unpublished historical laboratory notebook records shows similarly biased distributions of reaction condition choices. By performing 548 randomly generated experiments, we demonstrate that the popularity of reactants or the choices of reaction conditions are uncorrelated to the success of the reaction. We show that randomly generated experiments better illustrate the range of parameter choices that are compatible with crystal formation. Machine-learning models that we train on a smaller randomized reaction dataset outperform models trained on larger human-selected reaction datasets, demonstrating the importance of identifying and addressing anthropogenic biases in scientific data. Human scientists make unrepresentative chemical reagent and reaction condition choices, and machine-learning algorithms trained on human-selected experiments are less capable of successfully predicting reaction outcomes than those trained on randomly generated experiments.

A new class of chemical instrumentation seeks to alleviate the tedium and complexity of organic syntheses. A new class of chemical instrumentation seeks to alleviate the tedium and complexity of organic syntheses. Close up view of dials and wires on Marty Burke’s synthesis machine which assembles complex molecules robotically.

Training algorithms to computationally plan multistep organic syntheses has been a challenge for more than 50 years 1 – 7 . However, the field has progressed greatly since the development of early programs such as LHASA 1 , 7 , for which reaction choices at each step were made by human operators. Multiple software platforms 6 , 8 – 14 are now capable of completely autonomous planning. But these programs ‘think’ only one step at a time and have so far been limited to relatively simple targets, the syntheses of which could arguably be designed by human chemists within minutes, without the help of a computer. Furthermore, no algorithm has yet been able to design plausible routes to complex natural products, for which much more far-sighted, multistep planning is necessary 15 , 16 and closely related literature precedents cannot be relied on. Here we demonstrate that such computational synthesis planning is possible, provided that the program’s knowledge of organic chemistry and data-based artificial intelligence routines are augmented with causal relationships 17 , 18 , allowing it to ‘strategize’ over multiple synthetic steps. Using a Turing-like test administered to synthesis experts, we show that the routes designed by such a program are largely indistinguishable from those designed by humans. We also successfully validated three computer-designed syntheses of natural products in the laboratory. Taken together, these results indicate that expert-level automated synthetic planning is feasible, pending continued improvements to the reaction knowledge base and further code optimization. A synthetic route-planning algorithm, augmented with causal relationships that allow it to strategize over multiple steps, can design complex natural-product syntheses that are indistinguishable from those designed by human experts.

Autonomous laboratories can accelerate discoveries in chemical synthesis, but this requires automated measurements coupled with reliable decision-making 1 , 2 . Most autonomous laboratories involve bespoke automated equipment 3 – 6 , and reaction outcomes are often assessed using a single, hard-wired characterization technique 7 . Any decision-making algorithms 8 must then operate using this narrow range of characterization data 9 , 10 . By contrast, manual experiments tend to draw on a wider range of instruments to characterize reaction products, and decisions are rarely taken based on one measurement alone. Here we show that a synthesis laboratory can be integrated into an autonomous laboratory by using mobile robots 11 – 13 that operate equipment and make decisions in a human-like way. Our modular workflow combines mobile robots, an automated synthesis platform, a liquid chromatography–mass spectrometer and a benchtop nuclear magnetic resonance spectrometer. This allows robots to share existing laboratory equipment with human researchers without monopolizing it or requiring extensive redesign. A heuristic decision-maker processes the orthogonal measurement data, selecting successful reactions to take forward and automatically checking the reproducibility of any screening hits. We exemplify this approach in the three areas of structural diversification chemistry, supramolecular host–guest chemistry and photochemical synthesis. This strategy is particularly suited to exploratory chemistry that can yield multiple potential products, as for supramolecular assemblies, where we also extend the method to an autonomous function assay by evaluating host–guest binding properties. A modular autonomous platform for general exploratory synthetic chemistry uses mobile robots to integrate an automated synthesis platform and two analysis platforms.

Small-molecule synthesis usually relies on procedures that are highly customized for each target. A broadly applicable automated process could greatly increase the accessibility of this class of compounds to enable investigations of their practical potential. Here we report the synthesis of 14 distinct classes of small molecules using the same fully automated process. This was achieved by strategically expanding the scope of a building block–based synthesis platform to include even Csp3-rich polycyclic natural product frameworks and discovering a catch-and-release chromatographic purification protocol applicable to all of the corresponding intermediates. With thousands of compatible building blocks already commercially available, many small molecules are now accessible with this platform. More broadly, these findings illuminate an actionable roadmap to a more general and automated approach for small-molecule synthesis.

Automated synthesis platforms accelerate and simplify the preparation of molecules by removing the physical barriers to organic synthesis. This provides unrestricted access to biopolymers and small molecules via reproducible and directly comparable chemical processes. Current automated multistep syntheses rely on either iterative 1 – 4 or linear processes 5 – 9 , and require compromises in terms of versatility and the use of equipment. Here we report an approach towards the automated synthesis of small molecules, based on a series of continuous flow modules that are radially arranged around a central switching station. Using this approach, concise volumes can be exposed to any reaction conditions required for a desired transformation. Sequential, non-simultaneous reactions can be combined to perform multistep processes, enabling the use of variable flow rates, reuse of reactors under different conditions, and the storage of intermediates. This fully automated instrument is capable of both linear and convergent syntheses and does not require manual reconfiguration between different processes. The capabilities of this approach are demonstrated by performing optimizations and multistep syntheses of targets, varying concentrations via inline dilutions, exploring several strategies for the multistep synthesis of the anticonvulsant drug rufinamide 10 , synthesizing eighteen compounds of two derivative libraries that are prepared using different reaction pathways and chemistries, and using the same reagents to perform metallaphotoredox carbon–nitrogen cross-couplings 11 in a photochemical module—all without instrument reconfiguration. An automated synthesis instrument comprising a series of continuous flow modules that are radially arranged around a central switching station can achieve both linear and convergent syntheses.

Metal-organic frameworks (MOFs) are porous, crystalline materials with well-defined structures that can be used in many applications, from gas storage to catalysis and drug storage. This protocol is for the preparation of the MOF NU-1000. The synthesis of NU-1000, a highly robust mesoporous (containing pores >2 nm) metal-organic framework (MOF), can be conducted efficiently on a multigram scale from inexpensive starting materials. Tetrabromopyrene and (4-(ethoxycarbonyl)phenyl)boronic acid can easily be coupled to prepare the requisite organic strut with four metal-binding sites in the form of four carboxylic acids, while zirconyl chloride octahydrate is used as a precursor for the well-defined metal oxide clusters. NU-1000 has been reported as an excellent candidate for the separation of gases, and it is a versatile scaffold for heterogeneous catalysis. In particular, it is ideal for the catalytic deactivation of nerve agents, and it shows great promise as a new generic platform for a wide range of applications. Multiple post-synthetic modification protocols have been developed using NU-1000 as the parent material, making it a potentially useful scaffold for several catalytic applications. The procedure for the preparation of NU-1000 can be scaled up reliably, and it is suitable for the production of 50 g of the tetracarboxylic acid containing organic linker and 200 mg–2.5 g of NU-1000. The entire synthesis is performed without purification by column chromatography and can be completed within 10 d.

The infrastructure for chemical synthesis typically lies at either end of a spectrum: small-scale studies in ad hoc assemblies of glassware or large-scale production in capital-intensive custom reactors. Kitson et al. report a hybrid protocol that customizes a blueprint for synthesis of a target compound in a series of interconnected plastic modules, which can be assembled en masse by 3D printing (see the Perspective by Hornung). The approach, demonstrated for the commercial muscle relaxant baclofen, establishes a systematic workflow that is potentially amenable to automation: All that is necessary for synthesis and purification is the introduction of stock solutions and variation of temperature or pressure. Science , this issue p. 314 ; see also p. 273 A blueprint for chemical synthesis in plasticware offers an alternative to capital-intensive reactors for low-volume targets. Chemical manufacturing is often done at large facilities that require a sizable capital investment and then produce key compounds for a finite period. We present an approach to the manufacturing of fine chemicals and pharmaceuticals in a self-contained plastic reactionware device. The device was designed and constructed by using a chemical to computer-automated design (ChemCAD) approach that enables the translation of traditional bench-scale synthesis into a platform-independent digital code. This in turn guides production of a three-dimensional printed device that encloses the entire synthetic route internally via simple operations. We demonstrate the approach for the γ-aminobutyric acid receptor agonist, (±)-baclofen, establishing a concept that paves the way for the local manufacture of drugs outside of specialist facilities.