Search NASA⌕ Search

SEARCH · Search NASA

Results for “molecular generation”

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
All sourcesNASA NTRSDOE OSTI
DOE OSTI2025

Enhancing generative molecular design via uncertainty-guided fine-tuning of variational autoencoders

In recent years, deep generative models have been successfully applied to various molecular design tasks, particularly in the life and materials sciences. One critical challenge for pre-trained generative molecular design (GMD) models is to fine-tune them to be better suited for downstream design tasks that aim at optimizing specific molecular properties. However, redesigning and training an existing effective generative model from scratch for each new design task are impractical. Furthermore, the black-box nature of typical downstream tasks that involve property prediction makes it nontrivial to optimize the generative model in a task-specific manner. In this work, we propose an uncertainty-guided fine-tuning strategy that can effectively enhance a pre-trained variational autoencoder (VAE) for GMD through performance feedback in an active learning setting. The strategy begins by quantifying the model uncertainty of the generative model using an efficient active subspace-based UQ (uncertainty quantification) scheme. Next, the decoder diversity within the characterized model uncertainty class is explored to expand the viable space of molecular generation. The low-dimensionality of the active subspace makes this exploration tractable using a black-box optimization scheme, which in turn enables us to identify and leverage a diverse set of high-performing models to generate enhanced molecules. Empirical results across six target molecular properties using multiple VAE-based generative models demonstrate that our uncertainty-guided fine-tuning strategy consistently leads to improved models that outperform the original pre-trained models.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2024

CoarsenConf: Equivariant Coarsening with Aggregated Attention for Molecular Conformer Generation

Molecular conformer generation (MCG) is an important task in cheminformatics and drug discovery. The ability to efficiently generate low-energy 3D structures can avoid expensive quantum mechanical simulations, leading to accelerated virtual screenings and enhanced structural exploration. Several generative models have been developed for MCG, but many struggle to consistently produce high-quality conformers for meaningful downstream applications. To address these issues, we introduce CoarsenConf, which coarse-grains molecular graphs based on torsional angles and integrates them into an SE(3)-equivariant hierarchical variational autoencoder. Through equivariant coarse-graining, we aggregate the fine-grained atomic coordinates of subgraphs connected via rotatable bonds, creating a variable-length coarse-grained latent representation. Our model uses a novel aggregated attention mechanism to restore fine-grained coordinates from the coarse-grained latent representation, enabling efficient generation of accurate conformers. Furthermore, we evaluate the chemical and biochemical quality of our generated conformers on multiple downstream applications, including property prediction and large-scale oracle-based protein docking. Overall, CoarsenConf generates more accurate conformer ensembles compared to prior generative models.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗
DOE OSTI2025

A comparison of probabilistic generative frameworks for molecular simulations

Generative artificial intelligence is now a widely used tool in molecular science. Despite the popularity of probabilistic generative models, numerical experiments benchmarking their performance on molecular data are lacking. Here, in this work, we introduce and explain several classes of generative models, broadly sorted into two categories: flow-based models and diffusion models. We select three representative models: neural spline flows, conditional flow matching, and denoising diffusion probabilistic models, and examine their accuracy, computational cost, and generation speed across datasets with tunable dimensionality, complexity, and modal asymmetry. Our findings are varied, with no one framework being the best for all purposes. In a nutshell, (i) neural spline flows do best at capturing mode asymmetry present in low-dimensional data, (ii) conditional flow matching outperforms other models for high-dimensional data with low complexity, and (iii) denoising diffusion probabilistic models appear the best for low-dimensional data with high complexity. Our datasets include a Gaussian mixture model and the dihedral torsion angle distribution of the Aib9 peptide, generated via a molecular dynamics simulation. We hope our taxonomy of probabilistic generative frameworks and numerical results may guide model selection for a wide range of molecular tasks.

Artificial intelligence↗
DOE OSTI2025

Impact of corrosion layer composition and thermal pretreatment on the radiolytic generation of molecular hydrogen from aluminum alloys

To fully evaluate the feasibility of extending the dry storage of aluminum-clad spent nuclear fuel (ASNF) in helium-backfilled cannisters, the amount of radiation-induced molecular hydrogen (H 2 ) generated and the impact of absorbed radiation dose on the ASNF corrosion layer composition must be accurately assessed. Here, in this study, we report a slowing in the rate of radiolytic H 2 generation for pre-corroded AA1100 and AA6061 aluminum alloy coupons irradiated to up to 53 MGy of absorbed cobalt-60 gamma dose. By exploring a variety of thermal pretreatment conditions for AA6061 coupons, we find that the “steady-state” yield of H 2 depends more on the aluminum alloy used than on the treatments applied prior to dry storage. Scanning electron microscopy and positron annihilation lifetime spectroscopy techniques also provided evidence for gamma radiation-induced defects in the corrosion layers of the investigated aluminum alloy coupons for high absorbed doses (~50 MGy), the consequences of which on cladding integrity and H 2 generation should be explored in future works.

38 - RADIATION CHEMISTRY, RADIOCHEMISTRY, AND NUCL↗
DOE OSTI2025

Leveraging Natural Language Processing and Generative Models in Molecular Chemistry: Property Prediction and Novel Compound Generation

The accurate prediction of molecular properties is important for the rational design and the advancement of green chemistry and sustainable materials research. However, the predictive power of traditional computational chemistry methods is limited due to computational restrictions. Here, in this study, we examine an alternative approach to the accurate prediction of properties of organic compounds: natural language processing (NLP)-based molecular embedding. Using viscosity, partition coefficient (log P), and enthalpy of vaporization as test properties through a survey of comprehensive datasets comprising 5695 data points for viscosity, 25 870 data points for log P, and 2296 data points for enthalpy of vaporization. These are important properties for the design of greener, safer, and sustainable chemical processes. Models were trained using NLP methods such as Mol2vec and fine-tuned ChemBERTa, and results were compared with traditional input featurization techniques such as Morgan fingerprints and quantum chemistry derived sigma profiles and DFT features. Among the various machine learning models, Mol2vec demonstrated superior predictive capabilities, achieving the highest correlation coefficient (R 2 = 0.945) and lowest RMSE (0.106 mPa s) for viscosity, as well as high accuracy for log P and enthalpy of vaporization predictions. These findings establish the Mol2vec featurization technique, graph-convolutional neural networks (GCNN), and fine-tuned ChemBERTa model as powerful tools for predictive modeling of organic compounds properties, offering a significant improvement over previously used featurization techniques and opening up strategies for very-high-throughput computational screening. Finally, we integrated ML models with hybrid language-model-based generative adversarial networks (LM-GAN) to generate novel molecular sequences with desirable properties for different research applications. The ability to computationally design solvents with lower viscosity, lower log P, and lower enthalpy of vaporization offers a data-driven route to accelerating the discovery of sustainable alternatives to traditionally toxic solvents.

ChemBERTa↗
DOE OSTI2025

Reorganization of Water at Aqueous Aluminum Chloride (AlCl 3 ) Interfaces: Vibrational Sum Frequency Generation and Molecular Dynamics Simulations

AlCl 3 hydration states and complexation are not well understood both in solutions and at the air–aqueous interface despite their potential significance in natural waters and their industrial and energy-related applications. Here, we investigated Al 3+ and Cl – ion behaviors in an AlCl 3 aqueous bulk solution and at the air–aqueous interface using interface-selective vibrational sum frequency generation (SFG), Raman and infrared spectroscopies, molecular dynamics (MD) simulation, as well as molecular-informed reduced modeling. Our reduced modeling reveals relatively long-range effects for Al 3+ as compared to monovalent ions such as Na + indicating that the interfacial depth of trivalent ions can be significantly larger than that of monovalent ions at the air–water interface. MD simulations reveal interfacial stratification and multiple layering of the ions. Compression of the Al 3+ and Cl – distributions with increasing concentrations from 0.5 to 2.5 m is also observed in the subsurface regions. Significant SSP- and PPP-polarized SFG OH spectral intensity increases are observed from 0.5 to 1.5 m and 0.5 to 2.5 m, respectively, indicative of interfacial depth increases and a change in average orientation above 1.5 m. Extensive evaluation of SFG spectra, Fresnel-corrected using several approaches, shows the same trends. The nonmonotonic trend points to a changing structure in surface and subsurface water orientation and hydrogen bonding environment generally consistent with the MD simulation of stratification and water orientation changes. Furthermore, solvent-shared ion pairing is implicated with MD simulation radial distribution analysis and consistent with infrared spectral identification of the hexaaqua aluminum ion in the solution phase. Spectral evidence of a strong Al 3+ hydration shell and the acidic behavior of the Al 3+ ions is obvious in the Raman and infrared spectra of the bulk solution. In conclusion, we show that the MD dipole potential is directly related to the MD second-order susceptibility of the interface, χ SFG–MD (2) , both of which correlate up to ∼35 Å with the spectral observations of increasing and then saturating intensities, suggesting that both ion stratification and interfacial depth determine the water orientations at an air–water interface of 1-3 electrolyte solutions.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗
DOE OSTI2025

Tracing Long-Lived Atomic Coherences Generated via Molecular Conical Intersections

Accessing coherences is key to fully understand and control ultrafast dynamics of complex quantum systems like molecules. Most photochemical processes are mediated by conical intersections, which generate coherences between electronic states in molecules. We show with accurate calculations performed on gas-phase methyl iodide that electronic coherences of spin-orbit-split states persist in atomic iodine after dissociation. Our simulation predicts a maximum magnitude of vibronic coherence in the molecular regime of 0.75% of the initially photoexcited state population. Upon dissociation, one-third of this coherence magnitude is transferred to a long-lived atomic coherence where vibrational decoherence can no longer occur. To trace these dynamics, we propose a tabletop experimental approach—heterodyned attosecond four-wave-mixing spectroscopy. This technique can temporally resolve small electronic coherence magnitudes and reconstruct the full complex coherence function via phase cycling. Hence, heterodyned attosecond four-wave-mixing spectroscopy leads the way to a complete understanding and optimal control of spin-orbit-coupled electronic states in photochemistry.

Chemical Physics↗
DOE OSTI2025

Solar-driven selective conversion of millimolar dissolved carbon to fuels with molecular flux generation

Abstract The direct utilization of dissolved inorganic carbon in seawater for CO 2 conversion promises chemical production on-demand and with zero carbon footprint. Photoelectrochemical (PEC) CO 2 reduction (CO 2 R) devices promise the sustainable conversion of dissolved carbon in seawater to carbon products using sunlight as the only energy input. However, the diffusion-dominant transport mechanism and the near-zero concentration of CO 2 (aq) (CO 2 dissolved in aqueous solution) in static seawater has made it extremely challenging to achieve high solar-to-fuel (STF) efficiency and high carbon-product selectivity. Here, where CO 2 (aq) as a reactant generated in situ by acidification of HCO 3 - flows continuously from BiVO 4 photoanodes to Si photocathodes, enabling a single-step conversion of dissolved carbon into products. Our PEC device significantly increases the CO selectivity from 3% to 21%, which approaches the 30% theoretical limit according to multi-physics modeling. Meanwhile, the Si/BiVO 4 PEC CO 2 R device achieved a STF efficiency of 0.71%. Such flow engineering achieves flow-dependent selectivity, rate, and stability in simulated seawater, thus promising practical solar fuel production at scale.

Liu, Bin↗
DOE OSTI2026

Discovery of hydrogen storage molecules using large language models and machine learning

Accelerating the discovery of new molecules with targeted properties is a central challenge in molecular design. In this contribution, we present an AI-driven molecular discovery framework that integrates Large Language Models (LLMs) for generative molecular design with Machine Learning (ML)-based screening to identify novel Liquid Organic Hydrogen Carrier (LOHC) candidates. Using the developed framework, LOHC molecules were systematically generated, evaluated, and refined iteratively, combining LLM-guided molecular generation and ML-predicted hydrogenation enthalpies (Δ H ), under physicochemical property constraints such as optimal melting points (MP), desired hydrogen storage capacity (wt% H 2 ), and synthetic accessibility (SA) scores. This approach enabled the discovery of 42 new LOHC candidates in two distinct campaigns, one seeded with experimentally known and another with previously computationally identified LOHCs, respectively. Although we began with different numbers of starting molecules (31 vs . 7 seed molecules), both runs yielded a comparable number of viable candidates, suggesting an influence of chemically intuitive seed molecule selection for success. Selected LOHC molecules, such as 3-methyl pyridine, 1-ethylnapthalene, 1,1-diphenylethane, and benzofuran, were experimentally tested and compared with benchmark LOHCs (toluene and 9-ethylcarbazole) for hydrogenation using a series of commercial supported metal catalysts. The order of conversion into fully hydrogenated products at 200 °C was 3-methyl pyridine (100%) > 9-ethyl carbazole (86.4%) > 2,3-benzofuran (74%) > 1,1-diphenylethane (66.9%) > 1-ethylnapthalene (66.7%) > toluene (57%), further validating the AI-guided molecular design. This study demonstrates promise of LLM-driven molecular design in conjunction with ML-based screening for accelerated discovery and design of molecules.

Harb, Hassan [Argonne National Laboratory (ANL), A↗
DOE OSTI2024

Molecular catalyst coordinatively bonded to organic semiconductors for selective light-driven CO 2 reduction in water

The selective photoreduction of CO 2 in aqueous media based on earth-abundant elements only, is today a challenging topic. Here we present the anchoring of discrete molecular catalysts on organic polymeric semiconductors via covalent bonding, generating molecular hybrid materials with well-defined active sites for CO 2 photoreduction, exclusively to CO in purely aqueous media. The molecular catalysts are based on aryl substituted Co phthalocyanines that can be coordinated by dangling pyridyl attached to a polymeric covalent triazine framework that acts as a light absorber. This generates a molecular hybrid material that efficiently and selectively achieves the photoreduction of CO 2 to CO in KHCO 3 aqueous buffer, giving high yields in the range of 22 mmol g -1 (458 μmol g -1 h -1 ) and turnover numbers above 550 in 48 h, with no deactivation and no detectable H 2 . The electron transfer mechanism for the activation of the catalyst is proposed based on the combined results from time-resolved fluorescence spectroscopy, in situ spectroscopies and quantum chemical calculations.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗
DOE OSTI2026

EvoDiffMol: evolutionary diffusion framework for 3D molecular design with optimized properties

Designing molecules with specific target properties remains a fundamental challenge in computational chemistry. While existing approaches show promise, most rely on simplified representations like SMILES strings or 2D graphs that lack essential three-dimensional geometric information. We present EvoDiffMol, a computational framework that integrates evolutionary algorithms with three-dimensional diffusion models for property-driven molecular generation. The method operates through adaptive evolutionary optimization, where population-based selection guides the generation process toward desired property landscapes. EvoDiffMol supports both unconstrained molecular design and scaffold-constrained generation that preserves fixed substructures while optimizing complementary regions. Comprehensive evaluation demonstrates exceptional performance, achieving the highest drug-likeness score (0.94) among all compared state-of-the-art methods while maintaining excellent validity, uniqueness, and novelty. Beyond single property optimization, the framework demonstrates flexible multi-property optimization capabilities, simultaneously controlling multiple molecular descriptors including synthetic accessibility, lipophilicity, topological polar surface area, and clinically relevant ADMET properties such as cardiotoxicity (hERG) and intestinal permeability (Caco-2). This adaptability spans from simple descriptors to practical pharmaceutical endpoints without requiring complete model retraining. The framework achieves precise control over target property values, generating molecules with properties closely matching specified targets for both single and multiple descriptors. Scaffold-constrained experiments preserve fixed molecular cores while maintaining effective property optimization. The three-dimensional representation offers advantages in maintaining structural validity during iterative optimization, with potential for geometry-aware applications in materials science and drug discovery.

3D molecular generation↗
DOE OSTI2025

RLMolLM: Reinforcement Learning-Enhanced Language Model Framework for Inverse Molecular Design

Inverse molecular design faces significant challenges due to vast chemical space and complex property requirements. While language models show promise for molecular generation, they struggle with validity, multi-property optimization, and structural constraints. This work presents RLMolLM, a reinforcement learning framework combining Proximal Policy Optimization (PPO) with genetic algorithms to address these limitations. Our approach optimizes multiple user-specified properties including quantitative estimates of drug-likeness (QED), synthetic accessibility (SA), and ADMET (absorption, distribution, metabolism, excretion, and toxicity) endpoints without requiring complete model retraining, while maintaining capability for scaffold-constrained generation where specific substructures must be preserved. We outperform state-of-the-art methods for molecular optimization, achieving best QED scores across GDB13, Moses, and Zinc datasets with up to 31% improvement over previous methods while maintaining excellent validity, uniqueness, and novelty metrics. For simultaneous multi-property optimization, our framework achieves substantial improvements in ADMET properties including 4.5-fold reduction in hERG toxicity and enhanced Caco-2 permeability compared to Moses dataset. Under structural constraints, the framework significantly improves molecular validity while preserving scaffolds and effectively optimizing properties. In conclusion, this versatile solution advances pharmaceutical and materials molecular design through effective integration of reinforcement learning and genetic algorithms with multi-property optimization and scaffold preservation.

Genetic algorithms↗
DOE OSTI2024

Readout of Oriented Triplet Excitons in Linear Acenes via Room-Temperature Electrically Detected Magnetic Resonance

In this study, optically generated molecular spin centers offer an attractive platform for room-temperature spintronic and quantum applications. The linear acene family of molecules are especially good candidates due to their efficient generation of highly polarized triplet excitons via singlet fission. However, the direct detection and manipulation of these spin centers in thin films via the electrical means desirable for ultimate microelectronic devices has proven challenging. In particular, highly oriented triplet features have previously been detected in crystalline anthracene but longer acenes reveal only doublet features in Electrically-Detected Magnetic Resonance (EDMR). In this work we present EDMR spectra of highly oriented triplet excitons in pentacene for the first time, using a host-guest style device made of tetracene and pentacene. The guest acts as an energetic trap site, permitting the isolation and detection of molecular triplets at room temperature. Modeling of these results shows that the observed resonance features correspond to triplet sublevel transitions on isolated pentacene guest molecules. Rotation of the applied field confirms the tendency of the linear acenes to self-orient with the longest molecular axis perpendicular to the device substrate. Lastly, we find the disappearance of resonant triplet features in the neat acenes is not primarily due to the effects of exciton delocalization, but a broader mechanism of spin relaxation primarily influenced by exciton diffusivity.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗
DOE OSTI2026

Expanding the Domain of Applicability of Machine Learning Models with Limited Data for Drug Property Prediction

Accurate machine learning models for predicting small molecule interactions with biological targets are essential for therapeutic discovery, biothreat response, and computational drug design, but their performance is often limited for understudied targets with sparse experimental data. To address this challenge, we developed and evaluated methods to improve molecular property prediction under low-data conditions, using the NimA-related kinase (NEK) family as a proof-of-concept. This work focused on two complementary goals within the ATOM Modeling PipeLine (AMPL) and the Generative Molecular Design (GMD) loop: expanding model applicability through transfer learning, representation learning, feature scaling, sampling strategies, and active-learning-inspired compound selection; and enabling efficient virtual screening to prioritize compounds that balance predicted activity, design objectives, and synthetic accessibility.

organic↗
DOE OSTI2024

Revealing the Reaction Path of UVC Bond Rupture in Cyclic Disulfides with Ultrafast X-ray Scattering

Disulfide bonds are ubiquitous molecular motifs that influence the tertiary structure and biological functions of many proteins. Yet it is well known that the disulfide bond is photolabile when exposed to UVC radiation. The deep-UV induced S-S bond fragmentation kinetics on very fast time scales are especially pivotal to fully understand the photostability and photodamage repair mechanisms in proteins. In 1,2-dithiane, the smallest saturated cyclic molecule that mimics biologically active species with S-S bonds, we investigate the photochemistry upon 200 nm excitation by femtosecond time-resolved X-ray scattering in the gas phase using an X-ray free electron laser. In the femtosecond time domain, we discover a very fast reaction that generates molecular fragments with one and two sulfur atoms. On picosecond and nanosecond timescales, a complex network of reactions unfolds that, ultimately, completes the sulfur dissociation from the parent molecule.

74 ATOMIC AND MOLECULAR PHYSICS↗
DOE OSTI2025

Electrocatalytic oxidation of hydrothermal liquefaction-derived aqueous phase for on-site wastewater treatment and H 2 production

Electrocatalytic oxidation (ECO) is a promising method for generating molecular hydrogen (H 2 ) while simultaneously treating the aqueous phase (AP) from hydrothermal liquefaction (HTL) of biomass-derived feedstocks such as algae, food waste, sludge, and wood. This study highlights the impact of HTL-AP composition on ECO performance, mainly activity, stability, and efficiency in a batch and flow electrolyzer. We demonstrated current efficiency (CE) for chemical oxygen demand (COD) removal ranges from 14% to 85% in the flow electrolyzer with electrode stability increasing from 20 to ≥2,800 h. Decreasing the applied potential enhances the CE. High ammonium content can accelerate deactivation, yet chloride ions appear to aid oxidation of organic compounds and enhanced anode stability. A preliminary energy and H 2 balance for an HTL sewage sludge plant shows that the HTL-AP contains sufficient COD to produce all the H 2 needs for bio-oil hydrotreating, with a 26% surplus available for other uses.

Electrolysis↗