Searching the Chemical Space of Bicyclic Dienes for Molecular Solar Thermal Energy Storage Candidates

Photoswitches are molecular systems that are chemically transformed subsequent to interaction with light and they find potential application in many new technologies. The design and discovery of photoswitch candidates require intricate molecular engineering of a range of properties to optimize a candidate to a specific applications, a task which can be tackled efficiently using quantum chemical screening procedures. In this paper, we perform a large scale screening of approximately half a million bicyclic diene photoswitches in the context of molecular solar thermal energy storage using ab initio quantum chemical methods. We further device an efficient strategy for scoring the systems based on their predicted solar energy conversion efficiency and elucidate potential pitfalls of this approach. Our search through the chemical space of bicyclic dienes reveals systems with unprecedented solar energy conversion efficiencies and storage densities that show promising design guidelines for next generation molecular solar thermal energy storage systems.

Evaluating Scalable Supervised Learning for Synthesize-on-Demand Chemical Libraries

Traditional small-molecule drug discovery is a time-consuming and costly endeavor. High-throughput chemical screening can only assess a tiny fraction of drug-like chemical space. The strong predictive power of modern machine-learning methods for virtual chemical screening enables training models on known active and inactive compounds and extrapolating to much larger chemical libraries. However, there has been limited experimental validation of these methods in practical applications on large commercially available or synthesize-on-demand chemical libraries. Through a prospective evaluation with the bacterial protein-protein interaction PriA-SSB, we demonstrate that ligand-based virtual screening can identify many active compounds in large commercial libraries. We use cross-validation to compare different types of supervised learning models and select a random forest (RF) classifier as the best model for this target. When predicting the activity of more than 8 million compounds from Aldrich Market Select, the RF substantially outperforms a naïve baseline based on chemical structure similarity. 48% of the RF's 701 selected compounds are active. The RF model easily scales to score one billion compounds from the synthesize-on-demand Enamine REAL database. We tested 68 chemically diverse top predictions from Enamine REAL and observed 31 hits (46%), including one with an IC50 value of 1.3 μM.

Physics-inspired machine learning of localized intensive properties

Machine learning (ML) has been widely applied to chemical property prediction, most prominently for the energies and forces in molecules and materials. The strong interest in predicting energies in particular has led to a 'local energy'-based paradigm for modern atomistic ML models, which ensures size-extensivity and a linear scaling of computational cost with system size. However, many electronic properties (such as excitation energies or ionization energies) do not necessarily scale linearly with system size and may even be spatially localized. Using size-extensive models in these cases can lead to large errors. In this work, we explore different strategies for learning intensive and localized properties, using HOMO energies in organic molecules as a representative test case. In particular, we analyze the pooling functions that atomistic neural networks use to predict molecular properties, and suggest an orbital weighted average (OWA) approach that enables the accurate prediction of orbital energies and locations.

Evaluation of tight-binding DFT performance for the description of organic photochromes properties

Photochromic molecules are widely studied and developed for their many potential applications. To optimize the required properties through theoretical models, a considerable chemical space is to be explored, and their environment in devices is to be accounted for.. To this end, cheap and reliable computational methods can be powerful tools to steer synthetic developments. As ab initio methods remain costly for extensive studies (in terms of the size of the system and/or number of molecules), semiempirical methods such as density functional tight-binding (TB) could offer a good compromise between accuracy computational cost. However, these approaches necessitate benchmarking on the families of compounds of interest. Thus, the aim of the present study is to evaluate the accuracy of several key features calculated with TB methods (DFTB2, DFTB3, GFN2-xTB, and LC-DFTB2) for three sets of photochromic organic molecules: azobenzene (AZO), norbornadiene/quadricyclane (NBD/QC), and dithienylethene (DTE) derivatives. The features considered here are the optimized geometries, the difference in energy between the two isomers (ΔE), and of the energies of the first relevant excited states. All the TB results are compared to those obtained with DFT methods and state-of-the-art electronic structure calculation methods: DLPNO-CCSD(T) for ground states and DLPNO-STEOM-CCSD for excited states. Our results show that, overall, DFTB3 is the TB method leading to the best results for the geometries and the ΔE values and can be used alone for these purposes for NBD/QC and DTE derivatives. Single point calculations at the r2SCAN-3c level using TB geometries allow circumventing the deficiencies of the TB methods in the AZO series. For electronic transition calculations, the range separated LC-DFTB2 method is the most accurate TB method tested for AZO and NBD/QC derivatives, in close agreement with the reference.

Molecular design of redox carriers for electrochemical CO2 capture and concentration

Developing improved methods for CO₂ capture and concentration (CCC) is essential to mitigating the impact of our current emissions and can lead to carbon net negative technologies. Electrochemical approaches for CCC can achieve much higher theoretical efficiencies compared to the thermal methods that have been more commonly pursued. The use of redox carriers, or molecular species that can bind and release CO₂ depending on their oxidation state, is an increasingly popular approach as carrier properties can be tailored for different applications. The key requirements for stable and efficient redox carriers are discussed in the context of chemical scaling relationships and operational conditions. Computational and experimental approaches towards developing redox carriers with optimal properties are also described.

Inverse molecular design of alkoxides and phenoxides for aqueous direct air capture of CO2

Aqueous direct air capture (DAC) is a key technology toward a carbon negative infrastructure. Developing sorbent molecules with water and oxygen tolerance and high CO₂ binding capacity is therefore highly desired. We analyze the CO₂ absorption chemistries on amines, alkoxides, and phenoxides with density functional theory calculations, and perform inverse molecular design of the optimal sorbent. The alkoxides and phenoxides are found to be more suitable for aqueous DAC than amines thanks to their water tolerance (lower pK_a prevents protonation by water) and capture stoichiometry of 1:1 (2:1 for amines). All three molecular systems are found to generally obey the same linear scaling relationship (LSR) between [Formula: see text] and [Formula: see text], since both CO₂ and proton are bonded to the nucleophilic (alkoxy or amine) binding site through a majorly [Formula: see text] bonding orbital. Several high-performance alkoxides are proposed from the computational screening. Phenoxides have comparatively poorer correlation between [Formula: see text] and [Formula: see text], showing promise for optimization. We apply a genetic algorithm to search the chemical space of substituted phenoxides for the optimal sorbent. Several promising off-LSR candidates are discovered. The most promising one features bulky ortho substituents forcing the CO₂ adduct into a perpendicular configuration with respect to the aromatic ring. In this configuration, the phenoxide binds CO₂ and a proton using different molecular orbitals, thereby decoupling the [Formula: see text] and [Formula: see text]. The [Formula: see text] trend and off-LSR behaviors are then confirmed by experiments, validating the inverse molecular design framework. This work not only extensively studies the chemistry of the aqueous DAC, but also presents a transferrable computational workflow for understanding and optimization of other functional molecules.

A Neural Network Approach for Property Determination of Molecular Solar Cell Candidates

The dihydroazulene/vinylheptafulvene (DHA/VHF) photocouple is a promising candidate for molecular solar heat batteries, storing and releasing energy in a closed cycle. Much work has been done on improving the energy storage capacity and the half-life of the high-energy isomer via substituent functionalization, but similarly important is keeping these improved properties in common polar solvents, along with being soluble in these, which is tied to the dipole properties. However, the number of possible derivatives makes an overview of this combinatorial space impossible both for experimental work and traditional computational chemistry. Due to the time-consuming nature of running many thousands of computations, we look to machine learning, which bears the advantage that once a model has been trained, it can be used to rapidly estimate approximate values for the given system. Applying a convolutional neural network, we show that it is possible to reach good agreement with traditional computations on a scale that allows us to rapidly screen tens of thousands of the DHA/VHF photocouple, eliminating bad candidates and allowing computational resources to be directed toward meaningful compounds.

Data-driven modeling of S0 → S1 excitation energy in the BODIPY chemical space: High-throughput computation, quantum machine learning, and inverse design

Derivatives of BODIPY are popular fluorophores due to their synthetic feasibility, structural rigidity, high quantum yield, and tunable spectroscopic properties. While the characteristic absorption maximum of BODIPY is at 2.5 eV, combinations of functional groups and substitution sites can shift the peak position by ±1 eV. Time-dependent long-range corrected hybrid density functional methods can model the lowest excitation energies offering a semi-quantitative precision of ±0.3 eV. Alas, the chemical space of BODIPYs stemming from combinatorial introduction of-even a few dozen-substituents is too large for brute-force high-throughput modeling. To navigate this vast space, we select 77 412 molecules and train a kernel-based quantum machine learning model providing <2% hold-out error. Further reuse of the results presented here to navigate the entire BODIPY universe comprising over 253 giga (253 × 10⁹) molecules is demonstrated by inverse-designing candidates with desired target excitation energies.

Virtual screening of norbornadiene-based molecular solar thermal energy storage systems using a genetic algorithm

We present a computational methodology for the screening of a chemical space of 10²⁵ substituted norbornadiene molecules for promising kinetically stable molecular solar thermal (MOST) energy storage systems with high energy densities that absorb in the visible part of the solar spectrum. We use semiempirical tight-binding methods to construct a dataset of nearly 34 000 molecules and train graph convolutional networks to predict energy densities, kinetic stability, and absorption spectra and then use the models together with a genetic algorithm to search the chemical space for promising MOST energy storage systems. We identify 15 kinetically stable molecules, five of which have energy densities greater than 0.45 MJ/kg, and the main conclusion of this study is that the largest energy density that can be obtained for a single norbornadiene moiety with the substituents considered here, while maintaining a long half-life and absorption in the visible spectrum, is around 0.55 MJ/kg.

Organic reactivity from mechanism to machine learning

As more data are introduced in the building of models of chemical reactivity, the mechanistic component can be reduced until 'big data' applications are reached. These methods no longer depend on underlying mechanistic hypotheses, potentially learning them implicitly through extensive data training. Reactivity models often focus on reaction barriers, but can also be trained to directly predict lab-relevant properties, such as yields or conditions. Calculations with a quantum-mechanical component are still preferred for quantitative predictions of reactivity. Although big data applications tend to be more qualitative, they have the advantage to be broadly applied to different kinds of reactions. There is a continuum of methods in between these extremes, such as methods that use quantum-derived data or descriptors in machine learning models. Here, we present an overview of the recent machine learning applications in the field of chemical reactivity from a mechanistic perspective. Starting with a summary of how reactivity questions are addressed by quantum-mechanical methods, we discuss methods that augment or replace quantum-based modelling with faster alternatives relying on machine learning.