Search NASA⌕ Search

SEARCH · Search NASA

Results for “Convolution”

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 1,675 records · Page 93

Predicting Small Molecule Transfer Free Energies by Combining Molecular Dynamics Simulations and Deep Learning

Accurately predicting small molecule partitioning and hydrophobicity is critical in the drug discovery process. There are many heterogeneous chemical environments within a cell and entire human body. For example, drugs must be able to cross the hydrophobic cellular membrane to reach their intracellular targets, and hydrophobicity is an important driving force for drug–protein binding. Atomistic molecular dynamics (MD) simulations are routinely used to calculate free energies of small molecules binding to proteins, crossing lipid membranes, and solvation but are computationally expensive. Machine learning (ML) and empirical methods are also used throughout drug discovery but rely on experimental data, limiting the domain of applicability. We present atomistic MD simulations calculating 15,000 small molecule free energies of transfer from water to cyclohexane. This large data set is used to train ML models that predict the free energies of transfer. We show that a spatial graph neural network model achieves the highest accuracy, followed closely by a 3D-convolutional neural network, and shallow learning based on the chemical fingerprint is significantly less accurate. A mean absolute error of ~4 kJ/mol compared to the MD calculations was achieved for our best ML model. We also show that including data from the MD simulation improves the predictions, tests the transferability of each model to a diverse set of molecules, and show multitask learning improves the predictions. This work provides insight into the hydrophobicity of small molecules and ML cheminformatics modeling, and our data set will be useful for designing and testing future ML cheminformatics methods.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

Dynamic Profiling of β-Coronavirus 3CL Mpro Protease Ligand-Binding Sites

β-coronavirus (CoVs) alone has been responsible for three major global outbreaks in the 21st century. The current crisis has led to an urgent requirement to develop therapeutics. Even though a number of vaccines are available, alternative strategies targeting essential viral components are required as a backup against the emergence of lethal viral variants. One such target is the main protease (Mpro) that plays an indispensable role in viral replication. The availability of over 270 Mpro X-ray structures in complex with inhibitors provides unique insights into ligand–protein interactions. Herein, we provide a comprehensive comparison of all nonredundant ligand-binding sites available for SARS-CoV2, SARS-CoV, and MERS-CoV Mpro. Extensive adaptive sampling has been used to investigate structural conservation of ligand-binding sites using Markov state models (MSMs) and compare conformational dynamics employing convolutional variational auto-encoder-based deep learning. Our results indicate that not all ligand-binding sites are dynamically conserved despite high sequence and structural conservation across β-CoV homologs. This highlights the complexity in targeting all three Mpro enzymes with a single pan inhibitor.

59 BASIC BIOLOGICAL SCIENCES↗

A Multiple Filter Based Neural Network Approach to the Extrapolation of Adsorption Energies on Metal Surfaces for Catalysis Applications

Computational catalyst discovery involves the development of microkinetic reactor models based on estimated parameters determined from density functional theory (DFT). For complex surface chemistries, the number of reaction intermediates can be very large, and the cost of calculating the adsorption energies by DFT for all surface intermediates even for one active site model can become prohibitive. In this paper, we have identified appropriate descriptors and machine learning models that can be used to predict a significant part of these adsorption energies given data on the rest of them. Furthermore, our investigations also included the case when the species data used to train the predictive model are of different size relative to the species the model tries to predict—this is an extrapolation in the data space which is typically difficult with regular machine learning models. Due to the relative size of the available data sets, we have attempted to extrapolate from the larger species to the smaller ones in the current work. Here, we have developed a neural network based predictive model that combines an established additive atomic contribution based model with the concepts of a convolutional neural network that, when extrapolating, achieves a statistically significant improvement over the previous models.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

Modeling the Conformer-Dependent Electronic Absorption Spectra and Photolysis Rates of Methyl Vinyl Ketone Oxide and Methacrolein Oxide

Criegee intermediates are important atmospheric oxidants, formed via the reaction of ozone with volatile alkenes emitted into the troposphere. Small Criegee intermediates (e.g., CH 2 OO and CH 3 CHOO) are highly reactive, and their removal via unimolecular decay or bimolecular chemistry dominates their atmospheric lifetimes. As the molecular complexity of Criegee intermediates increases, their electronic absorption spectra show a bathochromic shift within the solar spectrum relevant to the troposphere. In such cases, solar photolysis may become a competitive contributor to their atmospheric removal. In this article, we report the conformer-dependent simulated electronic absorption spectra of two four-carbon-centered Criegee intermediates, methyl vinyl ketone oxide (MVK-oxide) and methacrolein oxide (MACR-oxide). Both MVK-oxide and MACR-oxide contain four low-energy conformers, which are convoluted in the experimentally measured spectra. Here, we deconvolute each conformer and estimate contributions from each of the four conformers to the experimentally measured spectra. We also estimate the photolysis rates and predict that solar photolysis should be a more competitive removal process for MVK-oxide and MACR-oxide (cf. CH 2 OO and CH 3 CHOO).

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

Comparison of Machine Learning Approaches for Prediction of the Equivalent Alkane Carbon Number for Microemulsions Based on Molecular Properties

The chemical properties of oils are vital in the design of microemulsion systems. The hydrophilic–lipophilic difference equation used to predict microemulsions’ phase behavior expresses the oils’ physiochemical properties as the equivalent alkane carbon number (EACN). The experimental determination of EACN requires knowledge of the temperature dependence of the microemulsion system and the effects of different surfactant concentrations. Thus, the experimental determination is time-intensive and tedious, requiring days to months for proper separations. Furthermore, the experiments require high purity of chemicals because microemulsions are sensitive to impurities. Our work focuses on the quick and reliable predictions of the EACN with machine learning (ML) models. Due to the immaturity of ML chemical predictions, we compare three graph neural networks (GNNs) and a gradient-boosted tree algorithm, known as XGBoost. The GNNs use the molecular structures represented as simplified molecular-input line-entry system (SMILES) codes for the initial input, which allows us to assess whether geometry optimization is necessary for reliable results. The XGBoost model also begins with the SMILES representations of the molecules but uses molecular descriptors instead of geometry optimizations. As a result, the best model tested (crystal graph convolutional neural network with Merck molecular force field-94) has an error of 1.15 EACN units of the true EACN for unknown data with the errors skewed toward zero and an R² score of 0.9

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

Self-Evolving Machine: A Continuously Improving Model for Molecular Thermochemistry

Because collecting precise and accurate chemistry data is often challenging, chemistry data sets usually only span a small region of chemical space, which limits the performance and the scope of applicability of data-driven models. To address this issue, we integrated an active learning machine with automatic ab initio calculations to form a self-evolving model that can continuously adapt to new species appointed by the users. In the present work, we demonstrate the self-evolving concept by modeling the formation enthalpies of stable closed-shell polycyclic species calculated at the B3LYP/6-31G(2df,p) level of theory. By combining a molecular graph convolutional neural network with a dropout training strategy, the model we developed can predict density functional theory (DFT) enthalpies for a broad range of polycyclic species and assess the quality of each predicted value. For the species which the current model is uncertain about, the automatic ab initio calculations provide additional training data to improve the performance of the model. For a test set composed of 2858 cyclic and polycyclic hydrocarbons and oxygenates, the enthalpies predicted by the model agree with the reference DFT values with a root-mean-square error of 2.62 kcal/mol. Finally, we found that a model originally trained on hydrocarbons and oxygenates can broaden its prediction coverage to nitrogen-containing species via an active learning process, suggesting that the continuous learning strategy is not only able to improve the model accuracy but is also capable of expanding the predictive capacity of a model to unseen species domains.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

Elucidating the Chemical Dynamics of the Elementary Reactions of the 1-Propynyl Radical (CH 3 CC; X 2 A 1 ) with Methylacetylene (H 3 CCCH; X 1 A 1 ) and Allene (H 2 CCCH 2 ; X 1 A 1 )

The reactions of the 1-propynyl radical (CH 3 CC; X 2 A 1 ) with two C 3 H 4 isomers, methylacetylene (H 3 CCCH; X 1 A 1 ) and allene (H 2 CCCH 2 ; X 1 A 1 ), along with their (partially) deuterated counterparts were explored at collision energies of 37 kJ mol –1 , exploiting crossed molecular beams to unravel the chemical reaction dynamics to synthesize distinct C 6 H 6 isomers under single collision conditions. The forward convolution fitting of the laboratory data along with ab initio and statistical calculations revealed that both reactions have no entrance barrier, proceed via indirect (complex-forming) reaction dynamics involving C 6 H 7 intermediates with life times longer than their rotation period(s), and are initiated by the addition of the 1-propynyl radical with its radical center to the π-electron density of the unsaturated hydrocarbon at the terminal carbon atoms of methylacetylene (C1) and allene (C1/C3). In the methylacetylene system, the initial collision complexes undergo unimolecular decomposition via tight exit transition states by atomic hydrogen loss, forming dimethyldiacetylene (CH 3 CCCCCH 3 ) and 1-propynylallene (H 3 CCCHCCCH 2 ) in overall exoergic reactions (123 and 98 kJ mol –1 ) with a branching ratio of 9.4 ± 0.1; the methyl group of the 1-propynyl reactant acts solely as a spectator. On the other hand, in the allene system, our experimental data exhibit the formation of the fulvene (c-C 5 H 4 CH 2 ) isomer via a six-step reaction sequence with two higher energy isomers—hexa-1,2-dien-4-yne (H 2 CCCHCCCH 3 ) and hexa-1,4-diyne (HCCCH 2 CCCH 3 )—also predicted to be formed based on our statistical calculations. Here, the pathway to fulvene advocates that, in the allene–1-propynyl system, the methyl group of the 1-propynyl reactant is actively engaged in the reaction mechanism to form fulvene. Because both reactions are barrierless and exoergic and all transition states are located below the energy of the separated reactants, the hydrogen-deficient C 6 H 6 isomers identified in our investigation are predicted to be synthesized in low-temperature environments, such as in hydrocarbon-rich atmospheres of planets and their moons such as Titan along with cold molecular clouds such as Taurus Molecular Cloud-1.

74 ATOMIC AND MOLECULAR PHYSICS↗

Characterizing the Self-Assembly Properties of Monoolein Lipid Isosteres

Living cells feature lipid compartments which exhibit a variety of shapes and structures that assist essential cellular processes. Many natural cell compartments frequently adopt convoluted nonlamellar lipid architectures that facilitate specific biological reactions. Improved methods for controlling the structural organization of artificial model membranes would facilitate investigations into how membrane morphology affects biological functions. Monoolein (MO) is a single-chain amphiphile which forms nonlamellar lipid phases in aqueous solution and has wide applications in nanomaterial development, the food industry, drug delivery, and protein crystallization. However, even if MO has been extensively studied, simple isosteres of MO, while readily accessible, have seen limited characterization. An improved understanding of how relatively minor changes in lipid chemical structure affect self assembly and membrane topology could instruct the construction of artificial cells and organelles for modeling biological structures and facilitate nanomaterial-based applications. Here, we investigate the differences in self-assembly and large-scale organization between MO and two MO lipid isosteres. We show that replacing the ester linkage between the hydrophilic headgroup and hydrophobic hydrocarbon chain with a thioesther or amide functional group results in the assembly of lipid structures with different phases not resembling those formed by MO. Using light and cryo-electron microscopy, small-angle X-ray scattering, and infrared spectroscopy, we demonstrate differences in the molecular ordering and large-scale architectures of the self-assembled structures made from MO and its isosteric analogues. These results improve our understanding of the molecular underpinnings of lipid mesophase assembly and may facilitate the development of MO-based materials for biomedicine and as model lipid compartments.

60 APPLIED LIFE SCIENCES↗

Enhancing Nanoparticle Detection in Interferometric Scattering (iSCAT) Microscopy Using a Mask R-CNN

Interferometric scattering microscopy (iSCAT) is a label-free optical microscopy technique that enables imaging of individual nano-objects such as nanoparticles, viruses, and proteins. Essential to this technique is the suppression of background scattering and identification of signals from nano-objects. In the presence of substrates with high roughness, scattering heterogeneities in the background, when coupled with tiny stage movements, cause features in the background to be manifested in background-suppressed iSCAT images. Traditional computer vision algorithms detect these background features as particles, limiting the accuracy of object detection in iSCAT experiments. Here, in this paper, we present a pathway to improve particle detection in such situations using supervised machine learning via a mask region-based convolutional neural network (mask R-CNN). Using a model iSCAT experiment of 19.2 nm gold nanoparticles adsorbing to a rough layer-by-layer polyelectrolyte film, we develop a method to generate labeled datasets using experimental background images and simulated particle signals and train the mask R-CNN using limited computational resources via transfer learning. We then compare the performance of the mask R-CNN trained with and without inclusion of experimental backgrounds in the dataset against that of a traditional computer vision object detection algorithm, Haar-like feature detection, by analyzing data from the model experiment. Results demonstrate that including representative backgrounds in training datasets improved the mask R-CNN in differentiating between background and particle signals and elevated performance by markedly reducing false positives. The methodology for creating a labeled dataset with representative experimental backgrounds and simulated signals facilitates the application of machine learning in iSCAT experiments with strong background scattering and thus provides a useful workflow for future researchers to improve their image processing capabilities.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

Enhanced Sampling of Crystal Nucleation with Graph Representation Learnt Variables

In this study, we present a graph neural network-based learning approach using an autoencoder setup to derive low-dimensional variables from features observed in experimental crystal structures. These variables are then biased in enhanced sampling to observe state-to-state transitions and reliable thermodynamic weights. Our approach uses simple convolution and pooling methods. To verify the effectiveness of our protocol, we examined the nucleation of various allotropes and polymorphs of iron and glycine from their molten states. Our graph latent variables when biased in well-tempered metadynamics consistently show transitions between states and achieve accurate free energy calculations in agreement with experiments, both of which are indicators of dependable sampling. This underscores the strength and promise of our graph neural net variables for improved sampling. Furthermore, the protocol shown here should be applicable for other systems and with other sampling methods.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

What is the best method to fit time-resolved data? A comparison of the residual minimization and the maximum likelihood techniques as applied to experimental time-correlated, single-photon counting data

The need for measuring fluorescence lifetimes of species in subdiffraction-limited volumes in, for example, stimulated emission depletion (STED) microscopy, entails the dual challenge of probing a small number of fluorophores and fitting the concomitant sparse data set to the appropriate excited-state decay function. This need has stimulated a further investigation into the relative merits of two fitting techniques commonly referred to as “residual minimization” (RM) and “maximum likelihood” (ML). Fluorescence decays of the well-characterized standard, rose bengal in methanol at room temperature (530 ± 10 ps), were acquired in a set of five experiments in which the total number of “photon counts” was approximately 20, 200, 1000, 3000, and 6000 and there were about 2–200 counts at the maxima of the respective decays. Each set of experiments was repeated 50 times to generate the appropriate statistics. Each of the 250 data sets was analyzed by ML and two different RM methods (differing in the weighting of residuals) using in-house routines and compared with a frequently used commercial RM routine. Convolution with a real instrument response function was always included in the fitting. While RM using Pearson’s weighting of residuals can recover the correct mean result with a total number of counts of 1000 or more, ML distinguishes itself by yielding, in all cases, the same mean lifetime within 2% of the accepted value. For 200 total counts and greater, ML always provides a standard deviation of <10% of the mean lifetime, and even at 20 total counts there is only 20% error in the mean lifetime. Here, the robustness of ML advocates its use for sparse data sets such as those acquired in some subdiffraction-limited microscopies, such as STED, and, more importantly, provides greater motivation for exploiting the time-resolved capacities of this technique to acquire and analyze fluorescence lifetime data.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

Quantifying Charge Carrier Localization in PBTTT Using Thermoelectric and Spectroscopic Techniques

Chemically doped poly[2,5-bis(3-alkylthiophen-2-yl)thieno[3,2- b ]thiophene] (PBTTT) shows promise for many organic electronic applications, but rationalizing its charge transport properties is challenging because conjugated polymers are inhomogeneous, with convoluted optical and solid-state transport properties. Herein, we use the semilocalized transport (SLoT) model to quantify how the charge transport properties of PBTTT change as a function of iron(III) chloride (FeCl 3 ) doping level. We use the SLoT model to calculate fundamental transport parameters, including the carrier density needed for metal-like electrical conductivities and the position of the Fermi energy level with respect to the transport edge. We then contextualize these parameters with other polymer-dopant systems and previous PBTTT reports. Additionally, we use grazing incidence wide-angle X-ray scattering and spectroscopic ellipsometry techniques to better characterize inhomogeneity in PBTTT. Our analyses indicate that PBTTT obtains high electrical conductivities due to its quickly rising reduced Fermi energy level, and this rise is afforded by its locally high carrier densities in highly ordered microdomains. Ultimately, this report sets a benchmark for comparing transport properties across polymer-dopant-processing systems.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

Structural Evolution of a Cyclooctatetraene Adlayer on Cu(111) during Isothermal Desorption

Here, the use of helium diffraction patterns to study desorption processes is explored as a novel extension to traditional methods based on helium specular reflection. The sample, cyclooctatetraene, adsorbed on Cu(111) provides a rich but complex structure. The modulation of cyclooctatetraene by Cu(111) is manifested as a convolution in the diffraction pattern, displaying an averaged super cell symmetry of (7$\sqrt{3}$ x 7$\sqrt{3}$)R30°. The adlayer expands during isothermal desorption, and the change in lattice constant provides a direct measure of the coverage as a function of time. We find a desorption energy of 0.96 +/- 0.01 at saturation of the first layer and an upper limit of 1.62 +/- 0.07 for isolated molecules. These values and details of the assigned structure indicate chemisorbed molecules with a planar conformation.

36 MATERIALS SCIENCE↗

Physical Origins of the Transient Absorption Spectra and Dynamics in Thin-Film Semiconductors: The Case of BiVO 4

Transient absorption (TA) spectroscopy is uniquely suited for understanding kinetic processes initiated by light over vast ranges of time. In combination with white light probes, the recorded differential absorption spectra can contain spectroscopic signatures characteristic of specific charge carrier population densities. However, disentangling the often-complex and convoluted spectra is made challenging without robust analysis methods relating the underlying physical mechanisms to the spectral components. In this work, we address the origin of the transient spectra of a model system of emerging solar energy harvesting materials using a monoclinic BiVO 4 thin film. Using ground-state optical properties of the semiconductor, we find the main derivative-like spectral response to be related to shifting and broadening of oscillators, rather than specific carrier-related transitions. However, by using the Drude optical model of free carriers, we also identify the transient response related to free-hole density. Importantly, sample heating from the optical pumping, which begins at ~10 ps and plateaus by ~200 ps, dominates the overall spectral response at longer times. On the basis of a physical model of the spectral response, a kinetic model is developed that describes the pump power dependence of the free-hole density, as well as the temporal evolution of the spectral changes associated with shifting and broadening of oscillators. First-principles density functional theory calculations are used to rationalize experimental measurements. Here, this comprehensive approach to analyzing and modeling the TA spectra offers a generalizable basis for understanding the complex pump-probe data, reveals thermal heating artifacts that are frequently erroneously assigned to long-lived photocarriers, and offers a path to eliminating ambiguity in analysis of photocarrier dynamics in solid-state systems.

36 MATERIALS SCIENCE↗

Measuring Activation and Luminescence Time Scales of Upconverting NaYF 4 :Yb,Er Nanocrystals

Accurate determination of upconversion and luminescence lifetimes requires kinetic modeling of the complete time-resolved response for upconversion luminescence of NaYF4:Yb,Er nanocrystals. Prior investigations typically perform exponential fitting of the tail in the time profile or employ complex systems of differential equations to extract lifetimes. To simplify analysis while fitting the entire time-resolved response, this work introduces a set of simplified models that model the response as a convolution of upconversion and fluorescence processes. Models for two- and three-photon upconversion processes are developed and tested for NaYF4:Yb,Er nanocrystals excited by a 980 nm laser. The results are presented for the transitions 2 H 11/2 , 4 S 3/2 → 4 I 15/2 (530 nm, green emission), 4 F 9/2 → 4 I 15/2 (650 nm, red emission), and 2 H 9/2 → 4 I 15/2 (400 nm, blue emission). Even with the same number of fitting parameters, the two- and three-photon models resulted in better fitting than the simpler models of single and two convolved exponentials for 30 nm, 400 nm, and silica-coated 30 nm nanoparticles. We provide evidence that the longest time scales (0.5–4 ms) are due to the luminescence of the final state, that the energy transfer waiting times leading to upconversion are shorter (<0.1 ms), and that more than two waiting times are required to explain the entire time response. The long-time response deviates from single-exponential behavior more strongly in uncoated and silica-coated 30 nm particles than for 400 nm particles. This deviation may result from heterogeneity of the local environment of ions within the 30 nm particles.

36 MATERIALS SCIENCE↗

Correction to “Frank–Kasper Phases of Diblock Copolymer Melts Studied with the DPD Model: SCF Results”

In a recent study, (1) we performed the self-consistent field (SCF) calculations of a dissipative particle dynamics (DPD) model of conformationally asymmetric diblock copolymer A-B melts to study the stability of various Frank–Kasper (FK) phases formed by such A-B melts. While all the equations in ref (1) are correct, we did not calculate the dimensionless (mean-field) Helmholtz free energies per chain βf c = βu c,χ + βu c,κ – s c /k B and the dimensionless stresses dβf c /dθ of ordered phases accurately. For example, the dimensionless (nonbonded) internal energy per chain due to the system compressibility βu c,κ = (N/2κ)∑ q βû 0 (|q|)(ϕ̂ A (q) + ϕ̂ B (q) – δ q,0 )(ϕ̂ A (−q) + ϕ̂ B (−q) – δ– q,0 ) should be calculated in the reciprocal space as given (note that βû 0 (|q|) is obtained analytically, and that since ϕ A (r) is real 𝜙̂ A (−𝐪)= $\overline{𝜙̂ A}$(𝐪), i.e., the complex conjugate of ϕ̂ A (q)); this corresponds to numerically evaluating βu c,κ = (N/2κ)∫dr(ϕ A (r) + ϕ B (r) – 1)ψ(r) with ψ(r) ≡ (∫dr′βu 0 (|r–r′|)(ϕ A (r′) + ϕ B (r′) – 1)/V in the real space via the (composite) trapezoidal rule. While the trapezoidal rule is less accurate than the Romberg integration (2) in general, the opposite occurs when the integration domain is periodic; this was pointed out by Matsen in ref (3) and explained in detail in refs (4and5). With our ignorance of this exceptional convergence property of the trapezoidal rule for periodic functions, we calculated all convolutions having the form of ψ(r) in the reciprocal space via the fast Fourier transforms (6) but those having the form of βu c,κ in the real space via Romberg integration; the latter leads to inaccurate results and unfortunately incorrect phase diagrams reported in ref (1), which are corrected here.

He, Juntong↗

Machine Learning Framework for Characterizing Processing–Structure Relationship in Block Copolymer Thin Films

The morphology of block copolymers (BCPs) critically influences material properties and applications. This work introduces a machine learning (ML)-enabled, high-throughput framework for analyzing grazing incidence small-angle X-ray scattering (GISAXS) data and atomic force microscopy (AFM) images to characterize BCP thin film morphology. A convolutional neural network was trained to classify AFM images by surface features, achieving 97% testing accuracy. Classified images were then analyzed to extract 2D grain size measurements from the samples in a high-throughput manner. ML models were trained to predict domain orientation based on processing parameters such as solvent ratio, additive type, and additive ratio. GISAXS-based properties were predicted with strong performances (R 2 > 0.75), while AFM-based property predictions were less accurate (R 2 < 0.60), likely due to the localized nature of AFM measurements compared to the bulk information captured by GISAXS. Beyond model performance, interpretability was addressed using SHapley Additive exPlanations (SHAP). SHAP analysis revealed that the additive ratio had the largest impact on morphological predictions, where additive provides the BCP chains with increased volume to rearrange into thermodynamically favorable morphologies. This interpretability helps validate model predictions and offers insight into parameter importance. Altogether, the presented framework combining high-throughput characterization and interpretable ML offers an approach to exploring and optimizing BCP thin film morphology across a broad processing landscape.

36 MATERIALS SCIENCE↗

Efficient Multiscale Optoelectronic Prediction for Conjugated Polymers

Conjugated polymers represent a high-potential material class due to the tunability of their optoelectronic properties via straightforward processing protocols. However, to in silico tailor these optoelectronic properties, one must compute the conformationally dependent electronic structure over mesoscopic length and time scales. This task represents a challenging multiscale computational problem in which quantum chemistry and atomistic or coarse-grained molecular dynamics must be integrated. Recently, we introduced a methodology, called electronic coarse graining (ECG), that utilizes artificial neural networks to compute ab initio quality electronic structures using only the reduced degrees of freedom of coarse-grained models. Here, we adapt ECG to variable-molecular weight conjugated oligomers by casting the problem in the framework of sequence detection. A machine-learning (ML) architecture employing one-dimensional convolutional neural networks and long short-term memory networks is utilized to compute ground-state orbital energies, charge density distributions, and optical spectra solely from the coarse-grained model's configurational degrees of freedom. Robust molecular weight transferability for ECG is established via a A-ML approach that leverages model electronic Hamiltonians for ground and excited-state property determination. The accuracy and transferability of the ECG methodology opens the door for scalable optoelectronic property prediction in conjugated polymers directly from coarse-grained degrees of freedom.

36 MATERIALS SCIENCE↗