Volume 2 Issue 6, June 2023

Automated experiments with integrated characterization techniques greatly accelerate materials synthesis and provide data to be used by machine learning algorithms. We reflect on the current use of data-driven automated experimentation in materials synthesis and consider the future of this approach.

Automation and real-time reaction monitoring have enabled data-rich experimentation, which is critically important in navigating the complexities of chemical synthesis. Linking real-time analysis with machine learning and artificial intelligence tools provides the opportunity to accelerate the identification of optimal reaction conditions and facilitate error-free autonomous synthesis. This Comment provides a viewpoint underscoring the growing significance of data-rich experiments and interdisciplinary approaches in driving future progress in synthetic chemistry.

Andrew Cooper, a professor at the University of Liverpool and Academic Director of the Materials Innovation Factory, talks to Nature Synthesis about the use of robotics and artificial intelligence for the synthesis and discovery of materials and chemicals.

King Kuok (Mimi) Hii, a professor at Imperial College London and director of the Centre for Rapid Online Analysis of Reactions (ROAR) and the Engineering and Physical Sciences Research Council (EPSRC) Centre of Doctoral Training in Next Generation Synthesis & Reaction Technology (rEaCt), talks to Nature Synthesis about reaction monitoring in automated workflows as well as the challenges to be overcome in automated synthesis.

Nature evolves proteins by iterating through an untold number of mutations over time. Now, a method is reported to prepare and optimize synthetic polypeptides in an automated high-throughput fashion driven by artificial intelligence.

Retrosynthesis has served as a playground for computer-aided design for many decades. Computer-aided methods are usually predicated on human-expert rules or learning algorithms that extract the rules from literature data. Now, an approach that bridges the gap between these computer-driven methods and the traditional, intuition-driven, ‘chalk board’ retrosynthetic methods is reported.

The elementary steps of transition-metal catalysis can be thwarted by high energy barriers. Now, by designing light-harvesting ligands on rhodium centres, these barriers are lowered or circumvented by accessing the excited state of the metal, enabling otherwise challenging reactivity at room temperature.

Bicyclo[1.1.1]pentyl (BCP) fragments are typically made from heat- and air-sensitive propellane, which decreases their potential impact in drug discovery. Now, easily accessible BCP sulfonium salts that are storable at room temperature enable the synthesis of a diverse set of functionalized BCP fragments.

Solar-driven photosynthesis offers a sustainable approach to directly producing hydrogen peroxide from oxygen and water but remains inefficient owing to the low photocatalytic efficiency of reported photocatalysts. Now, a Ga–N₅ atomic site on macroporous inverse-opal-type carbon nitride is introduced for the visible-light-driven photosynthesis of hydrogen peroxide with a solar-to-chemical conversion efficiency of 0.4%.

Self-driving labs (SDLs) combine machine learning with automated experimental platforms, enabling rapid exploration of the chemical space and accelerating the pace of materials and molecular discovery. In this Review, the application of SDLs, their limitations and future opportunities are discussed, and a roadmap is provided for their implementation by non-expert scientists.

Combinatorial synthesis has historically been the cornerstone of high-throughput experimentation. In this Review, we discuss the evolution of combinatorial synthesis and envision a future for accelerated materials science through its integration with artificial intelligence. We also evaluate the key aspects of combinatorial synthesis with respect to workflow design.

Trial-and-error synthesis and labour-intensive characterization procedures hinder the development of nanocrystals. Now, a data-driven robotic synthesis approach is used to prepare gold and double-perovskite nanocrystals. This approach combines data mining of synthesis parameters, robot-assisted synthesis and characterization, and machine-learning-facilitated inverse design of the nanocrystals.

High-throughput synthesis of polypeptides through ring-opening polymerization of N-carboxyanhydride is challenging. Now a diversification approach is developed based on the post-polymerization modification of a selenium-containing polypeptide. With the assistance of automation and model-guided optimization, this approach enables the discovery of functional polypeptides from chemical space with little previous knowledge.

Complex molecule synthesis involves speculative retrosynthetic planning and resource-intensive experimental evaluation. Now, a complementary strategy is reported that combines human-generated synthetic plans with computational prediction to accelerate this process. A machine learning model was trained to predict the yield of radical cyclization and guide the syntheses of clovane sesquiterpenoids.

Photoinduced catalytic systems typically consist of a transition metal catalyst and a photoredox catalyst. Now spiro-fluorene-indenoindenyl-Rh(I) complexes are reported as a single catalytic system that extends the scope of C–H borylation of arenes and [2+2+2] cycloaddition of alkynes to challenging substrates under irradiation with blue light.

1,3-Disubstituted bicyclo[1.1.1]pentanes are linear bioisosteres for para-substituted benzene rings; however, the lack of practical reagents for the introduction of bicyclopentane currently impedes their application, especially in drug development. Now, stable thianthrenium-based bicyclopentane reagents are reported and their use in O-, N- and C-alkylation reactions demonstrated.

Hydrogen peroxide is an important industrial feedstock but its synthesis is energy intensive. Now, a highly efficient Ga-N₅ atomic site is reported with a high solar-to-chemical conversion efficiency for direct photocatalysis of water into hydrogen peroxide.

BODIPYs possessing boron-stereogenic centres are rare and it is challenging to develop catalytic methodologies to enantioselectively prepare these molecules. Now, a palladium-catalysed desymmetric intramolecular C–H arylation reaction for the enantioselective synthesis of boron-stereogenic BODIPYs is reported, which gives access to various six- to nine-membered chiral boron heterocycles with good enantioselectivity.

Preparing enantioenriched aliphatic amines from readily available feedstocks is challenging to achieve. Now, direct enantioconvergent amination of racemic secondary alcohols using a variety of aliphatic primary amines is reported, catalysed by chiral iridium and phosphoric acid species. This atom-economical strategy streamlines the enantioselective synthesis of N-containing commercial drugs and analogues.