Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “Small molecules”

Search indexed NASA NTRS and DOE OSTI research on propulsion, heat transfer, battery materials and energy systems. Follow report and document links to the original sources.

Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.

At least 19 records

Deep learning to generate in silico chemical property libraries and candidate molecules for small molecule identification in complex samples

Comprehensive and unambiguous identification of small molecules in complex samples will revolutionize our understanding of the role of metabolites in biological systems. Existing and emerging technologies have enabled measurement of chemical properties of molecules in complex mixtures and, in concert, are sensitive enough to resolve even stereoisomers. Despite these experimental advances, small molecule identification is inhibited by (i) chemical reference libraries (e.g. mass spectra, collision cross section, and other measurable property libraries) representing <1% of known molecules, limiting the number of possible identifications, and (ii) the lack of a method to generate candidate matches directly from experimental features (i.e. without a library). To this end, we developed a variational autoencoder (VAE) to learn a continuous numerical, or latent, representation of molecular structure to expand reference libraries for small molecule identification. We extended the VAE to include a chemical property decoder, trained as a multitask network, in order to shape the latent representation such that it assembles according to desired chemical properties. The approach is unique in its application to metabolomics and small molecule identification, with its focus on properties that can be obtained from experimental measurements (m/z, CCS) paired with its training paradigm, which involved a cascade of transfer learning iterations. First, molecular representation is learned from a large dataset of structures with m/z labels. Next, in silico property values are used to continue training, as experimental property data is limited. Finally, the network is further refined by being trained with the experimental data. This allows the network to learn as much as possible at each stage, enabling success with progressively smaller datasets without overfitting. Once trained, the network can be used to predict chemical properties directly from structure, as well as generate candidate structures with desired chemical properties. Our approach is orders of magnitude faster than first-principles simulation for CCS property prediction. Additionally, the ability to generate novel molecules along manifolds, defined by chemical property analogues, positions DarkChem as highly useful in a number of application areas, including metabolomics and small molecule identification, drug discovery and design, chemical forensics, and beyond.

Colby, Sean M.↗

Deep learning to generate in silico chemical property libraries and candidate molecules for small molecule identification in complex samples

Comprehensive and unambiguous identification of small molecules in complex samples will revolutionize our understanding of the role of metabolites in biological systems. Existing and emerging technologies have enabled measurement of chemical properties of molecules in complex mixtures and, in concert, are sensitive enough to resolve even stereoisomers. Despite these experimental advances, small molecule identification is inhibited by (i) chemical reference libraries (e.g. mass spectra, collision cross section, and other measurable property libraries) representing <1% of known molecules, limiting the number of possible identifications, and (ii) the lack of a method to generate candidate matches directly from experimental features (i.e. without a library). To this end, we developed a variational autoencoder (VAE) to learn a continuous numerical, or latent, representation of molecular structure to expand reference libraries for small molecule identification. We extended the VAE to include a chemical property decoder, trained as a multitask network, in order to shape the latent representation such that it assembles according to desired chemical properties. The approach is unique in its application to metabolomics and small molecule identification, with its focus on properties that can be obtained from experimental measurements (m/z, CCS) paired with its training paradigm, which involved a cascade of transfer learning iterations. First, molecular representation is learned from a large dataset of structures with m/z labels. Next, in silico property values are used to continue training, as experimental property data is limited. Finally, the network is further refined by being trained with the experimental data. This allows the network to learn as much as possible at each stage, enabling success with progressively smaller datasets without overfitting. Once trained, the network can be used to predict chemical properties directly from structure, as well as generate candidate structures with desired chemical properties. Our approach is orders of magnitude faster than first-principles simulation for CCS property prediction. Additionally, the ability to generate novel molecules along manifolds, defined by chemical property analogues, positions DarkChem as highly useful in a number of application areas, including metabolomics and small molecule identification, drug discovery and design, chemical forensics, and beyond.

Colby, Sean M↗

Understanding Small-Molecule Interactions in Metal-Organic Frameworks: Coupling Experiment with Theory

Metal-organic frameworks (MOFs) have gained much attention as next generation porous media for various applications, especially gas separations/storage and catalysis. New MOFs are regularly reported; however, to develop better materials in a timely manner for specific applications, the interactions between guest molecules and the internal surface of the framework must first be understood. In this review, we present a combined experimental and theoretical approach that proves essential for the elucidation of small-molecule interactions in a model MOF system known as M 2 (dobdc) (dobdc 4– = 2,5-dioxido-1,4-benzenedicarboxylate; M = Mg, Mn, Fe, Co, Ni, Cu, or Zn), a material whose adsorption properties can be readily tuned via chemical substitution. Here, we additionally show that the study of extensive families like this one can provide a platform to test the efficacy and accuracy of developing computational methodologies in slightly varying chemical environments, a task that is necessary for their evolution into viable, robust tools for screening large numbers of materials.

36 MATERIALS SCIENCE↗

Hybrid Nanocrystals of Small Molecules and Chemically Disordered Polymers

Organic crystals formed by small molecules can be highly functional but are often brittle or insoluble structures with limited possibilities for use or processing from a liquid phase. A possible solution is the nanoscale integration of polymers into organic crystals without sacrificing long-range order and therefore function. This enables the organic crystals to benefit from the advantageous mechanical and chemical properties of the polymeric component. We report here on a strategy in which small molecules co-crystallize with side chains of chemically disordered polymers to create hybrid nanostructures containing a highly ordered lattice. Synchrotron X-ray scattering, absorption spectroscopy, and coarse-grained molecular dynamics simulations reveal that the polymer backbones form an “exo-crystalline” layer of disordered chains that wrap around the nanostructures becoming a handle for interesting properties. The morphology of this “hybrid bonding polymer” nanostructure is dictated by the competition between the polymers’ entropy and the enthalpy of the lattice allowing for control over the aspect ratio of the nanocrystal by changing the degree of polymer integration. Further, we observed that nanostructures with an exo-crystalline layer of polymer exhibit enhanced fracture strength, self-healing capacity, and dispersion in water, which benefits their use as light-harvesting assemblies in photocatalysis. Guided by computation, future work could also explore these hybrid nanostructures as building blocks for functional materials.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

MolDiscovery: learning mass spectrometry fragmentation of small molecules

Identification of small molecules is a critical task in various areas of life science. Recent advances in mass spectrometry have enabled the collection of tandem mass spectra of small molecules from hundreds of thousands of environments. To identify which molecules are present in a sample, one can search mass spectra collected from the sample against millions of molecular structures in small molecule databases. The existing approaches are based on chemistry domain knowledge, and they fail to explain many of the peaks in mass spectra of small molecules. Here, we present molDiscovery, a mass spectral database search method that improves both efficiency and accuracy of small molecule identification by learning a probabilistic model to match small molecules with their mass spectra. A search of over 8 million spectra from the Global Natural Product Social molecular networking infrastructure shows that molDiscovery correctly identify six times more unique small molecules than previous methods.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗

Mixing Behavior in Small Molecule: Fullerene Organic Photovoltaics [On the Mixing Behavior in Small Molecule: Fullerene Organic Photovoltaics]

Here, we report a novel method to determine the amount of pure, aggregated phase of donor and acceptor in organic photovoltaic (OPV) bulk heterojunctions. By determination of the diffraction intensity per unit volume for both donor and acceptor, the volume content of pure, aggregated donor and acceptor in the blend can be determined. We find that for the small molecule X2:[6,6]-phenyl-C61-butyric acid methyl ester (PCBM) system, in contrast to most polymer systems, all the PCBM is aggregated, indicating there is negligible miscibility of PCBM with X2. This provides an explanation why the performance of OPV devices of X2:PCBM are high over a large range of PCBM concentrations. This is in contrast to many other OPV blends, where PCBM forms a mixed phase with the donor and does not provide sufficient transport for electrons when the PCBM concentration is low. This study demonstrates that a mixed phase is not necessarily a requirement for good OPV device performance.

14 SOLAR ENERGY↗

A modular molecular framework for utility in small-molecule solution-processed organic photovoltaic devices

We report on the design, synthesis and characterization of light harvesting small molecules for use in solution-processed small molecule bulk heterojunction (SM-BHJ) solar cell devices. These molecular materials are based upon an acceptor/donor/acceptor (A/D/A) core with donor endcapping units. Utilization of a dithieno(3,2-b;2',3'-d)silole (DTS) donor and pyridal[2,1,3]thiadiazole (PT) acceptor leads to strong charge transfer characteristics, resulting in broad optical absorption spectra extending well beyond 700 nm. SM-BHJ solar cell devices fabricated with the specific example 5,5'-bis{7-(4-(5-hexylthiophen-2-yl)thiophen-2-yl)-[1,2,5]thiadiazolo[3,4-c]pyridine}-3,3'-di-2-ethylhexylsilylene-2,2'-bithiophene (6) as the donor and [6,6]-phenyl-C 71 -butyric acid methyl ester (PC 71 BM) as the acceptor component showed short circuit currents above -10 mA cm -2 and power conversion efficiencies (PCEs) over 3%. Thermal processing is a critical factor in obtaining favorable active layer morphologies and high PCE values. A combination of UV-visible spectroscopy, conductive and photo-conductive atomic force microscopies, dynamic secondary mass ion spectrometry (DSIMS), and grazing incident wide angle X-ray scattering (GIWAXS) experiments were carried out to characterize how thermal treatment influences the active layer structure and organization.

14 SOLAR ENERGY↗

First-principles Hubbard U approach for small molecule binding in metal-organic frameworks

We apply first-principles approaches with Hubbard U corrections for calculation of small molecule binding energetics to open-shell transition metal atoms in metal-organic frameworks (MOFs). Using density functional theory with van der Waals dispersion-corrected functionals, we determine Hubbard U values ab initio through an established linear response procedure for M-MOF-74, for a number of different metal centers (M = Ti, V, Cr, Mn, Fe, Co, Ni, and Cu). While our ab initio U values differ from those used in previous work, we show that they result in lattice parameters and electronic contributions to CO{sub 2}-MOF binding energies that lead to excellent agreement with experiments and previous results, yielding lattice parameters within 3%. In addition, U-dependent calculations for an example system, Co-MOF-74, suggest that the CO{sub 2} binding energy grows monotonically with the value of Hubbard U, with the binding energy shifting 4 kJ/mol (or 0.041 eV) over the range of U = 0-5.4 eV. These results provide insight into an approximate but computationally efficient means for calculation of small molecule binding energies to open-shell transition metal atoms in MOFs and suggest that the approach can be predictive with good accuracy, independent of the cations used and the availability of experimental data.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

Diazo coupling for surface attachment of small molecules to TiO 2 nanoparticles

Robust surface attachment of molecular species to metal oxide semiconductors is desirable for many applications. In this work, we report the interfacial diazo coupling of surface-bound amines with aromatics to bind them to the surface of TiO 2 nanoparticles via a robust siloxane anchor and a diazo link. The technique shows potential for the inexpensive, stable, modular and tunable attachment of molecules to metal oxide surfaces.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗