Message-passing neural networks for high-throughput polymer screening.

Machine learning methods have shown promise in predicting molecular properties, and given sufficient training data, machine learning approaches can enable rapid high-throughput virtual screening of large libraries of compounds. Graph-based neural network architectures have emerged in recent years as the most successful approach for predictions based on molecular structure and have consistently achieved the best performance on benchmark quantum chemical datasets. However, these models have typically required optimized 3D structural information for the molecule to achieve the highest accuracy. These 3D geometries are costly to compute for high levels of theory, limiting the applicability and practicality of machine learning methods in high-throughput screening applications. In this study, we present a new database of candidate molecules for organic photovoltaic applications, comprising approximately 91 000 unique chemical structures. Compared to existing datasets, this dataset contains substantially larger molecules (up to 200 atoms) as well as extrapolated properties for long polymer chains. We show that message-passing neural networks trained with and without 3D structural information for these molecules achieve similar accuracy, comparable to state-of-the-art methods on existing benchmark datasets. These results therefore emphasize that for larger molecules with practical applications, near-optimal prediction results can be obtained without using optimized 3D geometry as an input. We further show that learned molecular representations can be leveraged to reduce the training data required to transfer predictions to a new density functional theory functional.

Understanding the effects of electronic polarization and delocalization on charge-transport levels in oligoacene systems.

Electronic polarization and charge delocalization are important aspects that affect the charge-transport levels in organic materials. Here, using a quantum mechanical/embedded-charge (QM/EC) approach based on a combination of the long-range corrected ωB97X-D exchange-correlation functional (QM) and charge model 5 (CM5) point-charge model (EC), we evaluate the vertical detachment energies and polarization energies of various sizes of crystalline and amorphous anionic oligoacene clusters. Our results indicate that QM/EC calculations yield vertical detachment energies and polarization energies that compare well with the experimental values obtained from ultraviolet photoemission spectroscopy measurements. In order to understand the effect of charge delocalization on the transport levels, we considered crystalline naphthalene systems with QM regions including one or five-molecules. The results for these systems show that the delocalization and polarization effects are additive; therefore, allowing for electron delocalization by increasing the size of the QM region leads to the additional stabilization of the transport levels.

Close Packing of Nitroxide Radicals in Stable Organic Radical Polymeric Materials.

The relationship between the polymer network and electronic transport properties for stable radical polymeric materials has come under investigation owing to their potential application in electronic devices. For the radical polymer poly(2,2,6,6-tetramethylpiperidine-4-yl-1-oxyl methacrylate), it is unclear whether the radical packing is optimal for charge transport partially because the relationship between radical packing and molecular structure is not well-understood. Using the paramagnetic nitroxide radical as a probe of the polymer and synthetic techniques to control the radical concentration on the methyl methacrylate backbone, we investigate the dependence of radical concentration on molecular structure. The electron paramagnetic resonance data indicate that radicals in the PTMA assume a closest approach distance to each other when more than 60% of the backbone is populated with radical pendant groups. Below 60% coverage, the polymer rearranges to accommodate larger radical-radical spacing. These findings are consistent with theoretical calculations and help explain some experimentally determined electron-transport properties.

Quenching of the perylene fluorophore by stable nitroxide radical-containing macromolecules.

Stable nitroxide radical bearing organic polymer materials are attracting much attention for their application as next generation energy storage materials. A greater understanding of the inherent charge transfer mechanisms in such systems will ultimately be paramount to further advancements in the understanding of both intrafilm and interfacial ion- and electron-transfer reactions. This work is focused on advancing the fundamental understanding of these dynamic charge transfer properties by exploiting the fact that these species are efficient fluorescence quenchers. We systematically incorporated fluorescent perylene dyes into solutions containing the 2,2,6,6-tetramethylpiperidine-N-oxyl (TEMPO) radical and controlled their interaction by binding the TEMPO moiety into macromolecules with varying morphologies (e.g., chain length, density of radical pendant groups). In the case of the model compound, 4-oxo-TEMPO, quenching of the perylene excited state was found to be dominated by a dynamic (collisional) process, with a contribution from an apparent static process that is described by an ∼2 nm quenching sphere of action. When we incorporated the TEMPO unit into a macromolecule, the quenching behavior was altered significantly. The results can be described by using two models: (A) a collisional quenching process that becomes less efficient, presumably due to a reduction in the diffusion constant of the quenching entity, with a quenching sphere of action similar to 4-oxo-TEMPO or (B) a collisional quenching process that becomes more efficient as the radius of interaction grows larger with increasing oligomer length. This is the first study that definitively illustrates that fluorophore quenching by a polymer system cannot be explained using merely a classical Stern-Volmer approach but rather necessitates a more complex model.

Primary radiation defect production in polyethylene and cellulose.

Irradiation effects in polyethylene and cellulose were examined using molecular dynamics simulations. The governing reactions in both materials were chain scissioning and generation of small hydrocarbon and peroxy radicals. Recombination of chain fragments and cross-linking between polymer chains were found to occur less frequently. Crystalline cellulose was found to be more resistant to radiation damage than crystalline polyethylene. Statistics on radical formation are presented and the dynamics of the formation of radiation damage discussed.

Modeling reaction pathways of low energy particle deposition on polymer surfaces via first principle calculations.

The chemical processes that lead to polystyrene surface modification via low energy deposition of C(2)H(+), C(2)F(+), CH(2), CH(2)(+), and H(+) radicals and ions are examined using first principles calculations. Specifically, the reaction mechanisms responsible for products identified in classical molecular dynamics with reactive empirical bond-order potentials are examined using density functional theory. In addition, these calculations consider how the presence of charges on the incident particles changes the result for the CH(2) system through the comparison of barriers, transition states, and final products for CH(2) and CH(2)(+). The structures of the reaction species and energy barriers are determined using the B3LYP hybrid functional. Finally, CCSD/6-31G(d,p) single point energy calculations are carried out to obtain optimized energy barriers. The results indicate that the large variety of reactions occurring on the polystyrene surface are a consequence of complex interactions between the substrate and the deposited particles, which can easily be identified and characterized using advanced computational methodologies, such as first principle calculations.