Low-cost prediction of molecular and transition state partition functions via machine learning

We have generated an open-source dataset of over 30 000 organic chemistry gas phase partition functions. With this data, a machine learning deep neural network estimator was trained to predict partition functions of unknown organic chemistry gas phase transition states. This estimator only relies on reactant and product geometries and partition functions. A second machine learning deep neural network was trained to predict partition functions of chemical species from their geometry. Our models accurately predict the logarithm of test set partition functions with a maximum mean absolute error of 2.7%. Thus, this approach provides a means to reduce the cost of computing reaction rate constants ab initio. The models were also used to compute transition state theory reaction rate constant prefactors and the results were in quantitative agreement with the corresponding ab initio calculations with an accuracy of 98.3% on the log scale.

Deep neural networks accurately predict transition state partition functions at the low cost of reactant and product input features for organic chemistry gas phase reactions.     

Advancing global chemical education through interactive teaching tools

This perspective highlights our recent efforts to develop interactive resources in chemical education for worldwide usage. First, we highlight online tutorials that connect organic chemistry to medicine and popular culture, along with game-like resources for active learning. Next, we describe efforts to aid students in learning to visualize chemical structures in three dimensions. Finally, we present recent approaches toward engaging children and the general population through organic chemistry coloring and activity books. Collectively, these tools have benefited hundreds of thousands of users worldwide. We hope this perspective promotes a spirit of innovation in chemical education and spurs the development of additional free, interactive, and widely accessible chemical education resources.

This perspective highlights the development of interactive chemical education resources for worldwide usage. We hope to promote a spirit of innovation in chemical education and spur the development of new chemical education resources.     

Data-driven discovery of cardiolipin-selective small molecules by computational active learning

Subtle variations in the lipid composition of mitochondrial membranes can have a profound impact on mitochondrial function. The inner mitochondrial membrane contains the phospholipid cardiolipin, which has been demonstrated to act as a biomarker for a number of diverse pathologies. Small molecule dyes capable of selectively partitioning into cardiolipin membranes enable visualization and quantification of the cardiolipin content. Here we present a data-driven approach that combines a deep learning-enabled active learning workflow with coarse-grained molecular dynamics simulations and alchemical free energy calculations to discover small organic compounds able to selectively permeate cardiolipin-containing membranes. By employing transferable coarse-grained models we efficiently navigate the all-atom design space corresponding to small organic molecules with molecular weight less than ≈500 Da. After direct simulation of only 0.42% of our coarse-grained search space we identify molecules with considerably increased levels of cardiolipin selectivity compared to a widely used cardiolipin probe 10-N-nonyl acridine orange. Our accumulated simulation data enables us to derive interpretable design rules linking coarse-grained structure to cardiolipin selectivity. The findings are corroborated by fluorescence anisotropy measurements of two compounds conforming to our defined design rules. Our findings highlight the potential of coarse-grained representations and multiscale modelling for materials discovery and design.

We present a data-driven approach combining deep learning-enabled active learning with coarse-grained simulations and alchemical free energy calculations to discover small molecules to selectively permeate cardiolipin membranes.     

A visible-light mediated ring opening reaction of alkylidenecyclopropanes for the generation of homopropargyl radicals

Classical cyclopropylcarbinyl radical clock reactions have been widely applied to conduct mechanistic studies for probing radical processes for a long time; however, alkylidenecyclopropanes, which have a similar molecular structure to methylcyclopropanes, surprisingly have not yet attracted researcher''s attention for similar ring opening radical clock processes. In recent years, photocatalytic NHPI ester activation chemistry has witnessed significant blooming developments and provided new synthetic routes for cross-coupling reactions. Herein, we wish to report a non-classical ring opening radical clock reaction using innovative NHPI esters bearing alkylidenecyclopropanes upon photoredox catalysis, providing a brand-new synthetic approach for the direct preparation of a variety of alkynyl derivatives. The potential synthetic utility of this protocol is demonstrated in the diverse transformations and facile synthesis of bioactive molecules or their derivatives and medicinal substances.

A non-classical ring opening radical clock reaction using the innovative NHPI esters bearing alkylidenecyclopropanes upon photoredox catalysis has been demonstrated, providing a brand-new synthetic approach to access a variety of alkynyl derivatives.     

Bifurcating reactions: distribution of products from energy distribution in a shared reactive mode

Bifurcating reactions yield two different products emerging from one single transition state and are therefore archetypal examples of reactions that cannot be described within the framework of the traditional Eyring''s transition state theory (TST). With the growing number and importance of these reactions in organic and biosynthetic chemistry, there is also an increasing demand for a theoretical tool that would allow for the accurate quantification of reaction outcome at low cost. Here, we introduce such an approach that fulfils these criteria, by evaluating bifurcation selectivity through the energy distribution within the reactive mode of the key transition state. The presented method yields an excellent agreement with experimentally reported product ratios and predicts the correct selectivity for 89% of nearly 50 various cases, covering pericyclic reactions, rearrangements, fragmentations and metal-catalyzed processes as well as a series of trifurcating reactions. With 71% of product ratios determined within the error of less than 20%, we also found that the methodology outperforms three other tested protocols introduced recently in the literature. Given its predictive power, the procedure makes reaction design feasible even in the presence of complex non-TST chemical steps.

Reactive Mode Composition Factor (RMCF) analysis is a powerful tool to forecast the product distribution of bifurcating reactions through analysis of the kinetic energy distribution within the first transition state traversed by the reacting system.     

Switchable photocatalysis for the chemodivergent benzylation of 4-cyanopyridines

We report a photocatalytic strategy for the chemodivergent radical benzylation of 4-cyanopyridines. The chemistry uses a single photoredox catalyst to generate benzyl radicals upon N–F bond activation of 2-alkyl N-fluorobenzamides. The judicious choice of different photocatalyst quenchers allowed us to select at will between mechanistically divergent processes. The two reaction manifolds, an ipso-substitution path proceeding via radical coupling and a Minisci-type addition, enabled selective access to regioisomeric C4 or C2 benzylated pyridines, respectively. Mechanistic investigations shed light on the origin of the chemoselectivity switch.

We report a photocatalytic strategy for the chemodivergent radical benzylation of 4-cyanopyridines. The chemistry uses a single photoredox catalyst to generate benzyl radicals upon N–F bond activation of 2-alkyl N-fluorobenzamides.     

Vessel effects in organic chemical reactions; a century-old,overlooked phenomenon

One of the most intriguing aspects of synthetic chemistry is the interplay of numerous dependent and independent variables en route to achieve a successful, high-yielding chemical transformation. The experienced synthetic chemist will probe many of these variables during reaction development and optimization, which will routinely involve investigation of reaction temperature, solvent, stoichiometry, concentration, time, choice of catalyst, addition sequence or quenching conditions just to name some commonly addressed variables. Remarkably, little attention is typically given to the choice of reaction vessel material as the surface of common laboratory borosilicate glassware is, incorrectly, assumed to be chemically inert. When reviewing the scientific literature, careful consideration of the vessel material is typically only given during the use of well-known glass-etching reagents such as HF, which is typically only handled in HF-resistant, polyfluorinated polymer vessels. However, there are examples of chemical transformations that do not involve such reagents but are still clearly influenced by the choice of reaction vessel material. In the following review, we wish to condense the most significant examples of vessel effects during chemical transformations as well as observations of container-dependent stability of certain molecules. While the primary focus is on synthetic organic chemistry, relevant examples from inorganic chemistry, polymerization reactions, atmospheric chemistry and prebiotic chemistry are also covered.

One of the most intriguing aspects of synthetic chemistry is the interplay of numerous dependent and independent variables en route to achieve a successful, high-yielding chemical transformation.     

A direct-to-biology high-throughput chemistry approach to reactive fragment screening

Methods for rapid identification of chemical tools are essential for the validation of emerging targets and to provide medicinal chemistry starting points for the development of new medicines. Here, we report a screening platform that combines ‘direct-to-biology’ high-throughput chemistry (D2B-HTC) with photoreactive fragments. The platform enabled the rapid synthesis of >1000 PhotoAffinity Bits (HTC-PhABits) in 384-well plates in 24 h and their subsequent screening as crude reaction products with a protein target without purification. Screening the HTC-PhABit library with carbonic anhydrase I (CAI) afforded 7 hits (0.7% hit rate), which were found to covalently crosslink in the Zn²⁺ binding pocket. A powerful advantage of the D2B-HTC screening platform is the ability to rapidly perform iterative design–make–test cycles, accelerating the development and optimisation of chemical tools and medicinal chemistry starting points with little investment of resource.

A photoreactive fragment screening platform employing direct-to-biology high-throughput chemistry (D2B-HTC) for the rapid iterative synthesis and screening of libraries of photoaffinity bits.     

A novel chemical biology and computational approach to expedite the discovery of new-generation polymyxins against life-threatening Acinetobacter baumannii

Multidrug-resistant Gram-negative bacteria represent a major medical challenge worldwide. New antibiotics are desperately required with ‘old’ polymyxins often being the only available therapeutic option. Here, we systematically investigated the structure–activity relationship (SAR) of polymyxins using a quantitative lipidomics-informed outer membrane (OM) model of Acinetobacter baumannii and a series of chemically synthesized polymyxin analogs. By integrating chemical biology and all-atom molecular dynamics simulations, we deciphered how each residue of the polymyxin molecule modulated its conformational folding and specific interactions with the bacterial OM. Importantly, a novel designed polymyxin analog FADDI-287 with predicted stronger OM penetration showed improved in vitro antibacterial activity. Collectively, our study provides a novel chemical biology and computational strategy to expedite the discovery of new-generation polymyxins against life-threatening Gram-negative ‘superbugs’.

Multidrug-resistant Gram-negative bacteria have been an urgent threat to global public health. Novel antibiotics are desperately needed to combat these ''superbugs''.     

On-surface synthesis of organocopper metallacycles through activation of inner diacetylene moieties

The design of organometallic complexes is at the heart of modern organic chemistry and catalysis. Recently, on-surface synthesis has emerged as a disruptive paradigm to design previously precluded compounds and nanomaterials. Despite these advances, the field of organometallic chemistry on surfaces is still at its infancy. Here, we introduce a protocol to activate the inner diacetylene moieties of a molecular precursor by copper surface adatoms affording the formation of unprecedented organocopper metallacycles on Cu(111). The chemical structure of the resulting complexes is characterized by scanning probe microscopy and X-ray photoelectron spectroscopy, being complemented by density functional theory calculations and scanning probe microscopy simulations. Our results pave avenues to the engineering of organometallic compounds and steer the development of polyyne chemistry on surfaces.

The diacetylene skeletons of DNBD precursors are attacked on Cu(111) by copper adatoms resulting in the synthesis of organocopper metallacycles.     

Distinctive features and challenges in catenane chemistry

From being an aesthetic molecular object to a building block for the construction of molecular machines, catenanes and related mechanically interlocked molecules (MIMs) continue to attract immense interest in many research areas. Catenane chemistry is closely tied to that of rotaxanes and knots, and involves concepts like mechanical bonds, chemical topology and co-conformation that are unique to these molecules. Yet, because of their different topological structures and mechanical bond properties, there are some fundamental differences between the chemistry of catenanes and that of rotaxanes and knots although the boundary is sometimes blurred. Clearly distinguishing these differences, in aspects of bonding, structure, synthesis and properties, between catenanes and other MIMs is therefore of fundamental importance to understand their chemistry and explore the new opportunities from mechanical bonds.

Catenane chemistry is closely associated with that of rotaxane and knot, and this perspective highlights their similarities and differences in various aspects including synthesis, structure and properties.     

Leveraging autocatalytic reactions for chemical domain image classification

Autocatalysis is fundamental to many biological processes, and kinetic models of autocatalytic reactions have mathematical forms similar to activation functions used in artificial neural networks. Inspired by these similarities, we use an autocatalytic reaction, the copper-catalyzed azide–alkyne cycloaddition, to perform digital image recognition tasks. Images are encoded in the concentration of a catalyst across an array of liquid samples, and the classification is performed with a sequence of automated fluid transfers. The outputs of the operations are monitored using UV-vis spectroscopy. The growing interest in molecular information storage suggests that methods for computing in chemistry will become increasingly important for querying and manipulating molecular memory.

Kinetic models of autocatalytic reactions have mathematical forms similar to activation functions used in artificial neural networks. Inspired by these similarities, we use a copper-catalyzed reaction to perform digital image recognition tasks.     

Diaryliodonium salts facilitate metal-free mechanoredox free radical polymerizations

Mechanically-induced redox processes offer a promising alternative to more conventional thermal and photochemical synthetic methods. For macromolecule synthesis, current methods utilize sensitive transition metal additives and suffer from background reactivity. Alternative methodology will offer exquisite control over these stimuli-induced mechanoredox reactions to couple force with redox-driven chemical transformations. Herein, we present the iodonium-initiated free-radical polymerization of (meth)acrylate monomers under ultrasonic irradiation and ball-milling conditions. We explore the kinetic and structural consequences of these complementary mechanical inputs to access high molecular weight polymers. This methodology will undoubtedly find broad utility across stimuli-controlled polymerization reactions and adaptive material design.

Mechanically-induced redox processes offer a promising alternative to more conventional thermal and photochemical synthetic methods.     

Controlling cyclization pathways in palladium(ii)-catalyzed intramolecular alkene hydro-functionalization via substrate directivity

We report a series of palladium(ii)-catalyzed, intramolecular alkene hydrofunctionalization reactions with carbon, nitrogen, and oxygen nucleophiles to form five- and six-membered carbo- and heterocycles. In these reactions, the presence of a proximal bidentate directing group controls the cyclization pathway, dictating the ring size that is generated, even in cases that are disfavored based on Baldwin''s rules and in cases where there is an inherent preference for an alternative pathway. DFT studies shed light on the origins of pathway selectivity in these processes.

We report a series of palladium(ii)-catalyzed, intramolecular alkene hydrofunctionalization reactions with carbon, nitrogen, and oxygen nucleophiles to form five- and six-membered carbo- and heterocycles.     

Manifold dynamic non-covalent interactions for steering molecular assembly and cyclization

Deciphering rich non-covalent interactions that govern many chemical and biological processes is crucial for the design of drugs and controlling molecular assemblies and their chemical transformations. However, real-space characterization of these weak interactions in complex molecular architectures at the single bond level has been a longstanding challenge. Here, we employed bond-resolved scanning probe microscopy combined with an exhaustive structural search algorithm and quantum chemistry calculations to elucidate multiple non-covalent interactions that control the cohesive molecular clustering of well-designed precursor molecules and their chemical reactions. The presence of two flexible bromo-triphenyl moieties in the precursor leads to the assembly of distinct non-planar dimer and trimer clusters by manifold non-covalent interactions, including hydrogen bonding, halogen bonding, C–H⋯π and lone pair⋯π interactions. The dynamic nature of weak interactions allows for transforming dimers into energetically more favourable trimers as molecular density increases. The formation of trimers also facilitates thermally-triggered intermolecular Ullmann coupling reactions, while the disassembly of dimers favours intramolecular cyclization, as evidenced by bond-resolved imaging of metalorganic intermediates and final products. The richness of manifold non-covalent interactions offers unprecedented opportunities for controlling the assembly of complex molecular architectures and steering on-surface synthesis of quantum nanostructures.

A real-space characterization of dynamic non-covalent interactions in molecular assemblies and chemical reactions at the atomic bond level.     

Paving the way to conformationally unravel complex glycopeptide antibiotics by means of Raman optical activity

It is crucial for fundamental physical chemistry techniques to find their application in tackling real-world challenges. Hitherto, Raman optical activity (ROA) spectroscopy is one of the examples where a promising future within the pharmaceutical sector is foreseen, but has not yet been established. Namely, the technique is believed to be able to contribute in investigating the conformational behaviour of drug candidates. We, herein, strive towards the alignment of the ROA analysis outcome and the pharmaceutical expectations by proposing a fresh strategy that ensures a more complete, reliable, and transferable ROA study. The strategy consists of the treatment of the conformational space by means of a principal component analysis (PCA) and a clustering algorithm, succeeded by a thorough ROA spectral analysis and a novel way of estimating the contributions of the different chemical fragments to the total ROA spectral intensities. Here, vancomycin, an antibiotic glycopeptide, has been treated; it is the first antibiotic glycopeptide studied by means of ROA and is a challenging compound in ROA terms. By applying our approach we discover that ROA is capable of independently identifying the correct conformation of vancomycin in aqueous solution. In addition, we have a clear idea of what ROA can and cannot tell us regarding glycopeptides. Finally, the glycopeptide class turns out to be a spectroscopically curious case, as its spectral responses are unlike the typical ROA spectral responses of peptides and carbohydrates. This preludes future ROA studies of this intriguing molecular class.

Raman optical activity tackles the complex conformational space of glycopeptide antibiotics.     

Structure-mechanics statistical learning uncovers mechanical relay in proteins

A protein''s adaptive response to its substrates is one of the key questions driving molecular physics and physical chemistry. This work employs the recently developed structure-mechanics statistical learning method to establish a mechanical perspective. Specifically, by mapping all-atom molecular dynamics simulations onto the spring parameters of a backbone-side-chain elastic network model, the chemical moiety specific force constants (or mechanical rigidity) are used to assemble the rigidity graph, which is the matrix of inter-residue coupling strength. Using the S1A protease and the PDZ3 signaling domain as examples, chains of spatially contiguous residues are found to exhibit prominent changes in their mechanical rigidity upon substrate binding or dissociation. Such a mechanical-relay picture thus provides a mechanistic underpinning for conformational changes, long-range communication, and inter-domain allostery in both proteins, where the responsive mechanical hotspots are mostly residues having important biological functions or significant mutation sensitivity.

Protein residues exhibit specific routes of mechanical relay as the adaptive responses to substrate binding or dissociation. On such physically contiguous connections, residues experience prominent changes in their coupling strengths.     

Mechanical compression in cofacial porphyrin cyclophane pincers

Intra- and intermolecular interactions are dominating chemical processes, and their concerted interplay enables complex nonequilibrium states like life. While the responsible basic forces are typically investigated spectroscopically, a conductance measurement to probe and control these interactions in a single molecule far out of equilibrium is reported here. Specifically, by separating macroscopic metal electrodes, two π-conjugated, bridge-connected porphyrin decks are peeled off on one side, but compressed on the other side due to the covalent mechanical fixation. We observe that the conductance response shows an exceptional exponential rise by two orders of magnitude in individual breaking events during the stretching. Theoretical studies atomistically explain the measured conductance behavior by a mechanically activated increase in through-bond transport and a simultaneous strengthening of through-space coupling. Our results not only reveal the various interacting intramolecular transport channels in a molecular set of levers, but also the molecules'' potential to serve as molecular electro-mechanical sensors and switches.

A two-order conductance increase upon stretching in porphyrin cyclophane pincer junctions is measured. Atomistic studies explain experimental observations by characteristic intramolecular changes in through-space and through-bond transport.     

Accurate vertical ionization energy and work function determinations of liquid water and aqueous solutions

The absolute-scale electronic energetics of liquid water and aqueous solutions, both in the bulk and at associated interfaces, are the central determiners of water-based chemistry. However, such information is generally experimentally inaccessible. Here we demonstrate that a refined implementation of the liquid microjet photoelectron spectroscopy (PES) technique can be adopted to address this. Implementing concepts from condensed matter physics, we establish novel all-liquid-phase vacuum and equilibrated solution–metal-electrode Fermi level referencing procedures. This enables the precise and accurate determination of previously elusive water solvent and solute vertical ionization energies, VIEs. Notably, this includes quantification of solute-induced perturbations of water''s electronic energetics and VIE definition on an absolute and universal chemical potential scale. Defining and applying these procedures over a broad range of ionization energies, we accurately and respectively determine the VIE and oxidative stability of liquid water as 11.33 ± 0.03 eV and 6.60 ± 0.08 eV with respect to its liquid-vacuum-interface potential and Fermi level. Combining our referencing schemes, we accurately determine the work function of liquid water as 4.73 ± 0.09 eV. Further, applying our novel approach to a pair of exemplary aqueous solutions, we extract absolute VIEs of aqueous iodide anions, reaffirm the robustness of liquid water''s electronic structure to high bulk salt concentrations (2 M sodium iodide), and quantify reference-level dependent reductions of water''s VIE and a 0.48 ± 0.13 eV contraction of the solution''s work function upon partial hydration of a known surfactant (25 mM tetrabutylammonium iodide). Our combined experimental accomplishments mark a major advance in our ability to quantify electronic–structure interactions and chemical reactivity in liquid water, which now explicitly extends to the measurement of absolute-scale bulk and interfacial solution energetics, including those of relevance to aqueous electrochemical processes.

A generalised liquid-phase photoelectron spectroscopy approach is reported, allowing accurate, absolute energy scale ionisation energies of liquid water and aqueous solutions, as well as liquid water''s work function to be reported.     

Physically inspired deep learning of molecular excitations and photoemission spectra

Modern functional materials consist of large molecular building blocks with significant chemical complexity which limits spectroscopic property prediction with accurate first-principles methods. Consequently, a targeted design of materials with tailored optoelectronic properties by high-throughput screening is bound to fail without efficient methods to predict molecular excited-state properties across chemical space. In this work, we present a deep neural network that predicts charged quasiparticle excitations for large and complex organic molecules with a rich elemental diversity and a size well out of reach of accurate many body perturbation theory calculations. The model exploits the fundamental underlying physics of molecular resonances as eigenvalues of a latent Hamiltonian matrix and is thus able to accurately describe multiple resonances simultaneously. The performance of this model is demonstrated for a range of organic molecules across chemical composition space and configuration space. We further showcase the model capabilities by predicting photoemission spectra at the level of the GW approximation for previously unseen conjugated molecules.

A physically-inspired machine learning model for orbital energies is developed that can be augmented with delta learning to obtain photoemission spectra, ionization potentials, and electron affinities with experimental accuracy.