Fast predictions of liquid-phase acid-catalyzed reaction rates using molecular dynamics simulations and convolutional neural networks

The rates of liquid-phase, acid-catalyzed reactions relevant to the upgrading of biomass into high-value chemicals are highly sensitive to solvent composition and identifying suitable solvent mixtures is theoretically and experimentally challenging. We show that the complex atomistic configurations of reactant–solvent environments generated by classical molecular dynamics simulations can be exploited by 3D convolutional neural networks to enable accurate predictions of Brønsted acid-catalyzed reaction rates for model biomass compounds. We develop a 3D convolutional neural network, which we call SolventNet, and train it to predict acid-catalyzed reaction rates using experimental reaction data and corresponding molecular dynamics simulation data for seven biomass-derived oxygenates in water–cosolvent mixtures. We show that SolventNet can predict reaction rates for additional reactants and solvent systems an order of magnitude faster than prior simulation methods. This combination of machine learning with molecular dynamics enables the rapid, high-throughput screening of solvent systems and identification of improved biomass conversion conditions.

Solvent-mediated, acid-catalyzed reaction rates relevant to the upgrading of biomass into high-value chemicals are accurately predicted using a combination of molecular dynamics simulations and 3D convolutional neural networks.     

Controlling ultralong room temperature phosphorescence in organic compounds with sulfur oxidation state

Sulfur oxidation state is used to tune organic room temperature phosphorescence (RTP) of symmetric sulfur-bridged carbazole dimers. The sulfide-bridged compound exhibits a factor of 3 enhancement of the phosphorescence efficiency, compared to the sulfoxide and sulfone-bridged analogs, despite sulfone bridges being commonly used in RTP materials. In order to investigate the origin of this enhancement, temperature dependent spectroscopy measurements and theoretical calculations are used. The RTP lifetimes are similar due to similar crystal packing modes. Computational studies reveal that the lone pairs on the sulfur atom have a profound impact on enhancing intersystem crossing rate through orbital mixing and screening, which we hypothesize is the dominant factor responsible for increasing the phosphorescence efficiency. The ability to tune the electronic state without altering crystal packing modes allows the isolation of these effects. This work provides a new perspective on the design principles of organic phosphorescent materials, going beyond the rules established for conjugated ketone/sulfone-based organic molecules.

Sulfur lone pairs in bridged dimers enhance intersystem crossing and phosphorescence through orbital mixing and electrostatic screening.     

Studying the reactivity of alkyl substituted BODIPYs: first enantioselective addition of BODIPY to MBH carbonates

The first enantioselective addition of alkyl BODIPYs to Morita–Baylis–Hillman (MBH) carbonates is reported. This is the first reported enantioselective methodology using the methylene position of BODIPYs as a nucleophile. The reaction is efficiently catalyzed by cinchona alkaloids, achieving high enantioselectivities and total diastereoselectivity. The use of cinchona alkaloid pseudo enantiomers (chinine/cinchonine) allows us to obtain both pairs of enantiomers in similar yields and enantioselectivities, a common issue in this type of reaction. The photophysical study of these dyes (absorption and fluorescence) has been performed in order to determine their parameters and explore future possible application in bioimaging. In addition, electronic circular dichroism (ECD) studies supported by time-dependent density functional theory (TD-DFT) calculations were also performed.

The first enantioselective addition of alkyl BODIPYs to Morita–Baylis–Hillman (MBH) carbonates is reported.     

Graph-based machine learning interprets and predicts diagnostic isomer-selective ion–molecule reactions in tandem mass spectrometry

Diagnostic ion–molecule reactions employed in tandem mass spectrometry experiments can frequently be used to differentiate between isomeric compounds unlike the popular collision-activated dissociation methodology. Selected neutral reagents, such as 2-methoxypropene (MOP), are introduced into an ion trap mass spectrometer where they react with protonated analytes to yield product ions that are diagnostic for the functional groups present in the analytes. However, the understanding and interpretation of the mass spectra obtained can be challenging and time-consuming. Here, we introduce the first bootstrapped decision tree model trained on 36 known ion–molecule reactions with MOP. It uses the graph-based connectivity of analytes'' functional groups as input to predict whether the protonated analyte will undergo a diagnostic reaction with MOP. A Cohen kappa statistic of 0.70 was achieved with a blind test set, suggesting substantial inter-model reliability on limited training data. Prospective diagnostic product predictions were experimentally tested for 13 previously unpublished analytes. We introduce chemical reactivity flowcharts to facilitate chemical interpretation of the decisions made by the machine learning method that will be useful to understand and interpret the mass spectra for chemical reactivity.

We combine mass spectrometry with machine learning that is predictive and explainable using chemical reactivity flowcharts for diagnostic ion–molecule reactions.     

Hexa-peri-benzocoronene with two extra K-regions in an ortho-configuration

There are three possible isomers for hexa-peri-hexabenzocoronene (HBC) with two extra K-regions, but only two of them have been reported, namely with the meta- and para-configurations. Herein, we describe the synthesis of HBC 4 with two extra K-regions in the ortho-configuration, forming a longer zigzag edge compared with the other two isomers. The structure of 4 was validated by laser desorption/ionization time-of-flight mass analysis and nuclear magnetic resonance spectra, as well as Raman and infrared spectroscopies supported by density functional theory calculations. The optical properties of 4 were investigated by UV/vis absorption and ultrafast transient absorption spectroscopy. Together with the analysis of aromaticity, the influence of the zigzag edge on the π-conjugation pathway and HOMO–LUMO gaps of the three isomers were investigated.

We reported the synthesis of hexa-peri-benzocoronene (HBC) with two extra K-regions adopting an ortho-configuration. The systematical study provides deep insights about the effect of zigzag edge on the π-conjugated pathway and molecular design.     

Accelerating high-throughput virtual screening through molecular pool-based active learning

Structure-based virtual screening is an important tool in early stage drug discovery that scores the interactions between a target protein and candidate ligands. As virtual libraries continue to grow (in excess of 10⁸ molecules), so too do the resources necessary to conduct exhaustive virtual screening campaigns on these libraries. However, Bayesian optimization techniques, previously employed in other scientific discovery problems, can aid in their exploration: a surrogate structure–property relationship model trained on the predicted affinities of a subset of the library can be applied to the remaining library members, allowing the least promising compounds to be excluded from evaluation. In this study, we explore the application of these techniques to computational docking datasets and assess the impact of surrogate model architecture, acquisition function, and acquisition batch size on optimization performance. We observe significant reductions in computational costs; for example, using a directed-message passing neural network we can identify 94.8% or 89.3% of the top-50 000 ligands in a 100M member library after testing only 2.4% of candidate ligands using an upper confidence bound or greedy acquisition strategy, respectively. Such model-guided searches mitigate the increasing computational costs of screening increasingly large virtual libraries and can accelerate high-throughput virtual screening campaigns with applications beyond docking.

Bayesian optimization can accelerate structure-based virtual screening campaigns by minimizing the total number of simulations performed while still identifying the vast majority of computational hits.     

Super-capacity information-carrying systems encoded with spontaneous Raman scattering

Optical multiplex barcode systems have been significantly boosting the throughput of scientific discovery. A high volume of barcodes can be made from combinations of distinct spectral bands and intensity levels. However, the practical capacity often reaches a ceiling due to the overlaps of signal frequencies or intensities when massive information is written on individual carriers. In this paper, we built super-capacity information-carrying systems by tuning vibrational signals into octal numeral intensities in multiple bands of Raman-silent regions. This novel approach experimentally yielded the largest capacity of distinct optical barcodes to date. The experiments of encoding ASCII and Unicode systems to write and read languages indicate that the Raman coding method provides a new strategy for super-capacity data storage. In addition, multiplex screening of a cell-binding ligand was implemented to demonstrate the feasibility of this technology for fast and in situ high-throughput bio-discovery. These information-carrying systems may open new scenarios for the development of high-throughput screening, diagnostics and data storage.

Super-capacity information-carrying systems are fabricated by tuning vibrational signals into octal numeral intensities in multiple bands of Raman-silent regions.     

Cobalt-catalyzed intramolecular decarbonylative coupling of acylindoles and diarylketones through the cleavage of C–C bonds

We report here cobalt–N-heterocyclic carbene catalytic systems for the intramolecular decarbonylative coupling through the chelation-assisted C–C bond cleavage of acylindoles and diarylketones. The reaction tolerates a wide range of functional groups such as alkyl, aryl, and heteroaryl groups, giving the decarbonylative products in moderate to excellent yields. This transformation involves the cleavage of two C–C bonds and formation of a new C–C bond without the use of noble metals, thus reinforcing the potential application of decarbonylation as an effective tool for C–C bond formation.

A method for cobalt–N-heterocyclic carbene catalytic systems for the intramolecular decarbonylative coupling of ketones was achieved.     

Electronic and vibrational surface-enhanced Raman scattering: from atomically defined Au(111) and (100) to roughened Au

In surface-enhanced Raman spectra, vibrational peaks are superimposed on a background continuum, which is known as one major experimental anomaly. This is problematic in assessing vibrational information especially in the low Raman-shift region below 200 cm⁻¹, where the background signals dominate. Herein, we present a rigorous comparison of normal Raman and surface-enhanced Raman spectra for atomically defined surfaces of Au(111) or Au(100) with and without molecular adsorbates. It is clearly shown that the origin of the background continuum is well explained by a local field enhancement of electronic Raman scattering in the conduction band of Au. In the low Raman-shift region, electronic Raman scattering gains additional intensity, probably due to a relaxation in the conservation of momentum rule through momentum transfer from surface roughness. Based on the mechanism for generation of the spectral background, we also present a practical method to extract electronic and vibrational information at the metal/dielectric interface from the measured raw spectra by reducing the thermal factor, the scattering efficiency factor and the Purcell factor over wide ranges in both the Stokes and the anti-Stokes branches. This method enables us not only to analyse concealed vibrational features in the low Raman-shift region but also to estimate more reliable local temperatures from surface-enhanced Raman spectra.

Both electronic and vibrational information at the metal/dielectric interface were explicitly extracted from surface-enhanced Raman spectra.     

Vibrational exciton delocalization precludes the use of infrared intensities as proxies for surfactant accumulation on aqueous surfaces

Surface-sensitive vibrational spectroscopy is a common tool for measuring molecular organization and intermolecular interactions at interfaces. Peak intensity ratios are typically used to extract molecular information from one-dimensional spectra but vibrational coupling between surfactant molecules can manifest as signal depletion in one-dimensional spectra. Through a combination of experiment and theory, we demonstrate the emergence of vibrational exciton delocalization in infrared reflection–absorption spectra of soluble and insoluble surfactants at the air/water interface. Vibrational coupling causes a significant decrease in peak intensities corresponding to C–F vibrational modes of perfluorooctanoic acid molecules. Vibrational excitons also form between arachidic acid surfactants within a compressed monolayer, manifesting as signal reduction of C–H stretching modes. Ionic composition of the aqueous phase impacts surfactant intermolecular distance, thereby modulating vibrational coupling strength between surfactants. Our results serve as a cautionary tale against employing alkyl and fluoroalkyl vibrational peak intensities as proxies for concentration, although such analysis is ubiquitous in interface science.

Coupling between surfactant molecules at the air/water interface bleeds intensity into a diffuse background, such that single-wavelength vibrational intensity is effectively depleted at high surface coverage.     

Discrete Helmholtz model: a single layer of correlated counter-ions. Metal oxides and silica interfaces,ion-exchange and biological membranes

The mechanism by which interfaces in solution can be polarised depends on the nature of the charge carriers. In the case of a conductor, the charge carriers are electrons and the polarisation is homogeneous in the plane of the electrode. In the case of an insulator covered by ionic moieties, the polarisation is inhomogeneous and discrete in the plane of the interface. Despite these fundamental differences, these systems are usually treated in the same theoretical framework that relies on the Poisson–Boltzmann equation for the solution side. In this perspective, we show that interfaces polarised by discrete charge distributions are rather ubiquitous and that their associated potential drop significantly differs from those of conductor–electrolyte interfaces. We show that these configurations, spanning liquid–liquid interfaces, charged silica–water interfaces, metal oxide interfaces, supercapacitors, ion-exchange membranes and even biological membranes can be uniformly treated under a common “Discrete Helmholtz” model where the discrete charges are compensated by a single layer of correlated counter-ions, thereby generating a sharp potential drop at the interface.

Electrolytes in solution are strongly correlated with discrete charges at insulating interfaces inducing a behavior significantly different from that of conducting interfaces.     

Directional asymmetry over multiple length scales in reticular porous materials

In nature and synthetic materials, asymmetry is a useful tool to create complex and functional systems constructed from a limited number of building blocks. Reticular chemistry has allowed the synthesis of a wide range of discrete and extended structures, from which modularity permits the controlled assembly of their constituents to generate asymmetric configurations of pores or architectures. In this perspective, we present the different strategies to impart directional asymmetry over nano/meso/macroscopic length scales in porous materials and the resulting novel properties and applications.

Design strategies for the controlled assembly of discrete and extended reticular materials with asymmetric configurations of pores or architectures.     

Access to substituted cyclobutenes by tandem [3,3]-sigmatropic rearrangement/[2 + 2] cycloaddition of dipropargylphosphonates under Ag/Co relay catalysis

We present herein an unconventional tandem [3,3]-sigmatropic rearrangement/[2 + 2] cycloaddition of simple dipropargylphosphonates to deliver a range of bicyclic polysubstituted cyclobutenes and cyclobutanes under Ag/Co relay catalysis. An interesting switch from allene–allene to allene–alkyne cycloaddition was observed based on the substitution of the substrates, which further diversified the range of compounds accessible from this practical method. Significantly, preliminary biological screening of these new compounds identified promising candidates as suppressors of cellular proliferation.

In situ generation of allenes through [3,3]-sigmatropic rearrangement of propargylphosphonates. Divergent allene–allene or allene–alkyne cycloaddition by Ag/Co relay catalysis. Products as promising suppressors of cellular proliferation.     

Diversified strategy for the synthesis of DNA-encoded oxindole libraries

DNA-encoded library technology (DELT) employs DNA as a barcode to track the sequence of chemical reactions and enables the design and synthesis of libraries with billions of small molecules through combinatorial expansion. This powerful technology platform has been successfully demonstrated for hit identification and target validation for many types of diseases. As a highly integrated technology platform, DEL is capable of accelerating the translation of synthetic chemistry by using on-DNA compatible reactions or off-DNA scaffold synthesis. Herein, we report the development of a series of novel on-DNA transformations based on oxindole scaffolds for the design and synthesis of diversity-oriented DNA-encoded libraries for screening. Specifically, we have developed 1,3-dipolar cyclizations, cyclopropanations, ring-opening of reactions of aziridines and Claisen–Schmidt condensations to construct diverse oxindole derivatives. The majority of these transformations enable a diversity-oriented synthesis of DNA-encoded oxindole libraries which have been used in the successful hit identification for three protein targets. We have demonstrated that a diversified strategy for DEL synthesis could accelerate the application of synthetic chemistry for drug discovery.

Constructing DNA-encoded oxindole libraries by a diversified strategy.     

The electronic structure and deexcitation pathways of an isolated metalloporphyrin ion resolved by metal L-edge spectroscopy

The local electronic structure of the metal-active site and the deexcitation pathways of metalloporphyrins are crucial for numerous applications but difficult to access by commonly employed techniques. Here, we applied near-edge X-ray absorption mass spectrometry and quantum-mechanical restricted active space calculations to investigate the electronic structure of the metal-active site of the isolated cobalt(iii) protoporphyrin IX cation (CoPPIX⁺) and its deexcitation pathways upon resonant absorption at the cobalt L-edge. The experiments were carried out in the gas phase, thus allowing for control over the chemical state and molecular environment of the metalloporphyrin. The obtained mass spectra reveal that resonant excitations of CoPPIX⁺ at the cobalt L₃-edge lead predominantly to the formation of the intact radical dication and doubly charged fragments through losses of charged and neutral side chains from the macrocycle. The comparison between experiment and theory shows that CoPPIX⁺ is in a ³A_2g triplet ground state and that competing excitations to metal-centred non-bonding and antibonding σ* molecular orbitals lead to distinct deexcitation pathways.

Near-edge X-ray absorption mass spectrometry (NEXAMS) and restricted active space (RAS) quantum mechanical calculations at the metal L-edge reveal the electronic structure and orbital-specific deexcitation pathways of gas-phase metalloporphyrins.     

Retrosynthetic accessibility score (RAscore) – rapid machine learned synthesizability classification from AI driven retrosynthetic planning

Computer aided synthesis planning (CASP) is part of a suite of artificial intelligence (AI) based tools that are able to propose synthesis routes to a wide range of compounds. However, at present they are too slow to be used to screen the synthetic feasibility of millions of generated or enumerated compounds before identification of potential bioactivity by virtual screening (VS) workflows. Herein we report a machine learning (ML) based method capable of classifying whether a synthetic route can be identified for a particular compound or not by the CASP tool AiZynthFinder. The resulting ML models return a retrosynthetic accessibility score (RAscore) of any molecule of interest, and computes at least 4500 times faster than retrosynthetic analysis performed by the underlying CASP tool. The RAscore should be useful for pre-screening millions of virtual molecules from enumerated databases or generative models for synthetic accessibility and produce higher quality databases for virtual screening of biological activity.

The retrosynthetic accessibility score (RAscore) is based on AI driven retrosynthetic planning, and is useful for rapid scoring of synthetic feasability and pre-screening of large datasets of virtual/generated molecules.     

Enhanced voltammetric anion sensing at halogen and hydrogen bonding ferrocenyl SAMs

Halogen bonding mediated electrochemical anion sensing has very recently been established as a potent platform for the selective and sensitive detection of anions, although the principles that govern binding and subsequent signal transduction remain poorly understood. Herein we address this challenge by providing a comprehensive study of novel redox-active halogen bonding (XB) and hydrogen bonding (HB) ferrocene-isophthalamide-(iodo)triazole receptors in solution and at self-assembled monolayers (SAMs). Under diffusive conditions the sensory performance of the XB sensor was significantly superior. In molecular films the XB and HB binding motifs both display a notably enhanced, but similar, response to specific anions. Importantly, the enhanced response of these films is rationalised by a consideration of the (interfacial) dielectric microenvironment. These effects, and the resolved relationship between anion binding and signal transduction, underpin an improved fundamental understanding of anion sensing at redox-active interfaces which will benefit not just the development of more potent, real-life relevant, sensors but also new tools to study host–guest interactions at interfaces.

Surface enhancement effects in the sensing of anions at redox-active molecular films are investigated in detail and rationalised based on a consideration of the dielectric binding microenvironment.     

The central role of the metal ion for photoactivity: Zn– vs. Ni–Mabiq

Photoredox catalysts are integral components of artificial photosystems, and have recently emerged as powerful tools for catalysing numerous organic reactions. However, the development of inexpensive and efficient earth-abundant photoredox catalysts remains a challenge. We here present the photochemical and photophysical properties of a Ni–Mabiq catalyst ([Ni^II(Mabiq)]OTf (1); Mabiq = 2-4:6-8-bis(3,3,4,4-tetramethyldihydropyrrolo)-10-15-(2,2-biquinazolino)-[15]-1,3,5,8,10,14-hexaene1,3,7,9,11,14-N₆)—and of a Zn-containing analogue ([Zn^II(Mabiq)OTf] (2))—using steady state and time resolved optical spectroscopy, time-dependent density functional theory (TDDFT) calculations, and reactivity studies. The Ni and Zn complexes exhibit similar absorption spectra, but markedly different photochemical properties. These differences arise because the excited states of 2 are ligand-localized, whereas metal-centered states account for the photoactivity of 1. The distinct properties of the Ni and Zn complexes are manifest in their behavior in the photo-driven aza-Henry reaction and oxidative coupling of methoxybenzylamine.

The development of earth-abundant photoredox catalysts remains a challenge. Studies of Ni- and Zn-Mabiq complexes demonstrate how the coordinating metal ion influences the photochemistry, photodynamics and reactivity of photocatalysts.     

Design of BODIPY dyes as triplet photosensitizers: electronic properties tailored for solar energy conversion,photoredox catalysis and photodynamic therapy

BODIPYs are renowned fluorescent dyes with strong and tunable absorption in the visible region, high thermal and photo-stability and exceptional fluorescence quantum yields. Transition metal complexes are the most commonly used triplet photosensitisers, but, recently, the use of organic dyes has emerged as a viable and more sustainable alternative. By proper design, BODIPY dyes have been turned from highly fluorescent labels into efficient triplet photosensitizers with strong absorption in the visible region (from green to orange). In this perspective, we report three design strategies: (i) halogenation of the dye skeleton, (ii) donor–acceptor dyads and (iii) BODIPY dimers. We compare pros and cons of these approaches in terms of optical and electrochemical properties and synthetic viability. The potential applications of these systems span from energy conversion to medicine and key examples are presented.

BODIPYs offer a versatile platform to build organic triplet photosensitisers for PDT, TTA upconversion and photocatalysis. Tuning their properties provides the opportunity of replacing heavy-metal complexes and can lead to improved sustainability.