Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “Chemical Bonds”

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 775 records · Page 43
All sourcesNASA NTRSDOE OSTI
DOE OSTI2022

Master equation study of hydrogen abstraction from HCHO by OH via a chemically activated intermediate

The abstraction reaction of hydrogen from formaldehyde by OH radical plays an important role in formaldehyde oxidation. The reaction involves a bimolecular association to form a chemically activated hydrogen-bonded reaction complex followed by a unimolecular reaction of the complex to generate the products. The reaction rate is usually considered to be pressure-independent by assuming equilibrium between the reactants and the complex. However, our nonequilibrium calculations based on the chemically significant eigenmode of the master equation, carried out with our recently developed TUMME program, indicate that the reaction complex makes the rate constant dependent on pressure at low temperatures (T < 200 K). The calculations include anharmonicity, variational effects, and multi-dimensional tunneling. We find that the reaction rate constant reaches a low-pressure limit at pressures below 10 Torr over the whole investigated temperature range (20–1800 K), which explains why the available low-temperature experiments, which are for pressures below 2 Torr, did not observe the pressure dependence. A new extension of the TUMME master-equation program is used to explore the time evolutions of the concentrations of the OH radical and the complex under pseudo-first-order conditions. The time-dependent evolution of the concentrations of the complex at a low temperature provide direct evidence for the stabilization of the reaction complex at high pressures, and it shows the negligible role of the stabilized reaction complex at low pressures. Finally, the picture that emerges is qualitatively consistent with our previous study of the reaction of methanol with OH in that the tunneling in the unimolecular step from the complex to the products affects the phenomenological reaction rate constants differently at high and low pressures and leads to a significant pressure effect.

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

Novel insight into the kinetics of amide bond glycolysis for nylon-6 depolymerization

Chemical recycling of nylon-6 to short-chain oligomers and monomer ε-caprolactam via catalytic glycolysis is a potential solution for plastic waste remediation. Here, in this work, the kinetics of amide bond glycolysis (with ethylene glycol) in nylon-6 and the model compound N-phenethyl-3-phenylpropanamide (M1) were evaluated at 473 K in the presence of the cyclic amidine catalyst 1,5,7-triazabicyclo[4.4.0]dec-5-ene. Rates of polymer glycolysis were determined by the time-dependent shift in molecular weight distribution, whereas rates of M1 glycolysis were determined using liquid chromatography. The similarity of the first-order rate constants for glycolysis of nylon-6 and M1 at 473 K with 0.03 M amidine catalyst (5 mol% relative to amide bonds), 1.22 × 10 −5 s −1 and 2.18 × 10 −5 s −1 , respectively, confirmed the suitability of M1 as a model compound for nylon-6 glycolysis. Similar rates of glycolysis in the presence of other cyclic amidine catalysts as well as sodium methoxide revealed little influence of base strength. Glycolysis rates were unexpectedly non-linear in catalyst loading and deactivation occurred with long reaction times, presumably by non-selective decomposition of products as detected by liquid chromatography.

Depolymerization rate and rate constant↗
DOE OSTI2020

Beyond Ammonia: Nitrogen–Element Bond Forming Reactions with Coordinated Dinitrogen

The functionalization of coordinated dinitrogen to form nitrogen-element bonds en route to nitrogen-containing molecules is a long-standing challenge in chemical synthesis. The strong triple bond and the non-polarity of the N 2 molecule pose thermodynamic and kinetic challenges for promoting reactivity. Although heterogeneous, homogeneous and biological catalysts are all known for catalytic nitrogen fixation to ammonia, the catalytic synthesis of more complicated nitrogen-containing organic molecules has far less precedent. Examples of silyl radical additions to coordinated nitrogen to form silylamines stands as the lone example of a catalytic reaction involving N 2 to form a product other than ammonia. Our review surveys the field of molecular transition metal complexes, as well as recent boron examples for the formation of nitrogen-element bonds. Emphasis is placed on the coordination and activation modes of N 2 in the various metal compounds from across the transition series and how these structures can rationally inform reactivity studies. Over the past few decades, the field has evolved from the addition of carbon electrophiles in a manner similar to protonation reactions to more organometallic-inspired reactivity including insertions, 1,2-additions and cycloadditions. Various N–C, N–Si and N–B bond-forming reactions have been discovered highlighting that the challenge for catalytic chemistry is not in the reactivity of coordinated dinitrogen but rather removal of the functionalized ligand from the coordination sphere of the metal.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗
DOE OSTI2021

From inert gas to fertilizer, fuel and fine chemicals: N 2 reduction and fixation

The 100th anniversary of a leading nitrogen fixation technology developer like CASALE SA is a reason to reflect over the 20th century successful solution of the problem of world food supply, and to look out for solutions for sustainable developments with respect to ammonia production. Here, we review the role of nitrogen as essential chemical constituent in photosynthesis and biology, and component of ammonia as it is used as fertilizer for primary production by photosynthesis for farming and food supply and its future role as energy carrier. While novel synthesis methods and very advanced synchrotron based x-ray analytical techniques are being developed, we feel it is important to refer to the historical and economical context of nitrogen. The breaking of the N≡N triple bond remains a fundamental chemical and energetic problem in this context. We review the electrochemical ammonia synthesis as an energetically and environmentally benign method. Two relatively novel X-ray spectroscopy methods, which are relevant for the molecular understanding of the catalysts and biocatalysts, i.e. soft X-ray absorption spectroscopy and nuclear resonant vibration spectroscopy are presented. We illustrate the perceived reality in fertilizer usage on the field, and fertilizer production in the factory complex with photos and thus provide a contrast to the academic view of the molecular process of ammonia function and production.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗
DOE OSTI2024

Impacts of Hydrogen Bond Donor Structures in Phenolic Aldehyde Deep Eutectic Solvents on Pretreatment Efficiency

As a green solvent for biomass processing, deep eutectic solvents (DESs) have shown effectiveness in biomass processing. Here, in this study, phenolic aldehydes with different numbers of methoxy groups, including 4-hydroxybenzaldehyde (HBA, no methoxy), vanillin (VA, monomethoxy), and syringaldehyde (SA, dimethoxy) were employed to synthesize DESs with choline chloride (ChCl). The presence of methoxy groups in the hydrogen bond donor structure affected DES properties, as well as biomass pretreatment performance. The high thermal stability of phenolic aldehyde DESs was shown with over 225 °C onset temperature. The hydrogen bond donor with one aldehyde and one hydroxyl group at the para position without a methoxy group (ChCl-HBA) showed the highest xylan removal and delignification, reaching 59.3 and 88.0%, respectively, leading to the highest enzymatic hydrolysis yield. Sonication after pretreatment further enhanced the hydrolysis yields, achieving 83.3% glucan conversion and 50.1% xylan conversion. In the lignin-rich fraction, the recovered lignin showed a low weight–average molecular weight under 2100 g/mol with a relatively uniform molecular weight dispersity below 1.5. This study provides insights into how the chemical structure of hydrogen bond donors in DESs affects biomass processing and paves the way for designing effective lignin-derived DES in future biorefinery processes.

09 BIOMASS FUELS↗
DOE OSTI2023

Harnessing Neural Networks for Elucidating X-ray Absorption Structure–Spectrum Relationships in Amorphous Carbon

Improved understanding of structural and chemical properties through local experimental probes, such as X-ray absorption near-edge structure (XANES) spectroscopy, is crucial for the understanding and design of functional materials. In recent years, significant advancements have been made in the development of data science approaches for the automated interpretation of XANES structure–spectrum relationships. However, existing studies have primarily focused on crystalline solids and small molecules, while fewer efforts have been devoted to disordered systems. Thus, in this work, we demonstrate the development of neural network models for predicting and interpreting XANES spectra of amorphous carbon (a-C) from local structural descriptors. Comparison between different structural descriptors expectedly shows that the inclusion of both bond length and bond angle information is necessary for an accurate prediction of the spectra. Among the descriptors considered in this work, we find that the local many-body tensor representation yields the highest accuracy and greatest interpretability so that it can be leveraged to understand the importance of structural motifs in determining XANES spectra. Furthermore, we also discuss performance of neural network models for predicting both local structure features, such as bond lengths and bond angles, and global chemical composition, such as the sp:sp 2 :sp 3 ratio.

36 MATERIALS SCIENCE↗
DOE OSTI2021

Chemical durability and surface alteration of lanthanide zirconates (A 2 Zr 2 O 7 : A = La-Yb)

Chemical durability of lanthanide zirconates (A 2 Zr 2 O 7 ) (A = La-Yb) under near-field environments is important for evaluating their application as potential nuclear waste forms. In this work, A 2 Zr 2 O 7 (A = La-Yb) are synthesized by spark plasma sintering with controlled microstructure and their chemical durability are evaluated in a nitric acid solution (pH = 1). Scanning transmission electron microscopy analysis reveals an amorphous passivation film either enriched with Zr or lanthanide. The complex chemistry of the passivation films can be correlated with a transition in corrosion mechanisms from a preferential release of lanthanide in La 2 Zr 2 O 7 to a preferential release of Zr in Er 2 Zr 2 O 7 and Yb 2 Zr 2 O 7 . These results suggest a dominant mechanism of incongruent dissolution and surface reorganization for the formation of passivation films. Strong correlations are identified between the leaching rates and cation ionic size, ionic potential, electronegativity differences between A-site cation and Zr, and bonding valence sum of oxygen, suggesting important impacts of structural and bonding characteristics in controlling chemical durability of lanthanide zirconates.

12 MANAGEMENT OF RADIOACTIVE AND NON-RADIOACTIVE W↗
NASA NTRS1995

Interfacial Effects on the Thermal and Mechanical Properties of Graphite/Copper Composites

Graphite surfaces are not wet by pure copper. This lack of wetting has been responsible for a debonding phenomenon that has been found in continuous graphite fiber reinforced copper matrix composites subjected to elevated temperatures. By suitably alloying copper, its ability to wet graphite surfaces can be enhanced. Information obtained during sessile drop testing has led to the development of a copper-chromium alloy that suitably wets graphite. Unidirectionally reinforced graphite/copper composites have been fabricated using a pressure infiltration casting procedure. P100 pitch-based fibers have been used to reinforce copper and copper-chromium alloys. X-ray radiography and optical microscopy have been used to assess the fiber distribution in the cast composites. Scanning electron microscopy and Auger electron spectroscopy analyses were conducted to study the distribution and continuity of the chromium carbide reaction phase that forms at the fiber/matrix interface in the alloyed matrix composites. The effects of the chromium in the copper matrix on the mechanical and thermal properties of P100Gr/Cu composites have been evaluated through tensile testing, three-point bend testing, thermal cycling and thermal conductivity calculations. The addition of chromium has resulted in an increased shear modulus and essentially zero thermal expansion in the P100Gr/Cu-xCr composites through enhanced fiber/matrix bonding. The composites have longitudinal tensile strengths in excess of 700 MPa with elastic moduli of 393 GPa. After 100 hr at 760 deg C 84 percent of the as-cast strength is retained in the alloyed matrix composites. The elastic moduli are unchanged by the thermal exposure. It has been found that problems with spreading of the fiber tows strongly affect the long transverse tensile properties and the short transverse thermal conductivity of the P100Gr/Cu-xCr composites. The long transverse tensile strength is limited by rows of touching fibers which are paths of easy crack propagation under low tensile loads. The short transverse thermal conductivity is dictated by the fiber/matrix interface. Conduction across this interface has been estimated to be two orders of magnitude lower than that across the composite. This is due to the mechanical, and not chemical, nature of Gr/Cu bond.

Devincent, Sandra Marie↗
DOE OSTI2026

Boron Coordination in Multicomponent Glasses: Analytical Models and Machine Learning With Uncertainty

Borosilicate glasses are extensively used in a variety of applications from kitchenware to nuclear waste immobilization due to the strong network formed by the Si-O-B bond that makes it resistant to chemical corrosion and gives it a low thermal expansion. Boron, however, exists in both trigonal BO3 and tetrahedral BO4 bonds in glass systems, which impacts the chemical durability and thermal resistance of the glass, amongst other properties. Boron coordination (N4), or the ratio of the amount of BO4 to BO3 within a glass, may aid in predicting these properties but is difficult to derive without experimental data due to the complexity of impacts from varied glass compositions and processing factors. For this reason, compositional models have been developed to predict boron coordination, but the models typically include a limited number of glass components. To help fill this gap in the models, in this work, a diverse multicomponent glass dataset of 809 glasses is compiled from a literature search, and then a number of analytical and machine learning (ML) models are trained on the dataset. Previously developed modified Bernstein and modified Du Stebbins analytical models were fitted to update parameters with the new dataset. Then, partially Bayesian neural networks, Gaussian process regressor, and heteroskedastic deterministic neural networks were evaluated. The ML models examined all have different strategies to overcome the potential for overfitting as a result of a limited training dataset, and return results that account for model uncertainty, which can be valuable for understanding model reliability. For the first time, cooling rate is introduced as an input parameter for ML models, showing consistent improvements in performance and solidifying the importance of including parameters outside of composition alone for N4 prediction. The machine learning models examined here show promise in accurate predictions of boron coordination in borosilicate glasses, all achieving R2 values of 0.91.

boron coordination↗
NASA NTRS1985

Microstructure of reaction-bonded silicon nitride consolidated by isostatic hot-pressing

A microstructural examination and chemical analysis are conducted for reaction-bonded Si3N4 specimens that were isostatically hot pressed at 1850, 1950, and 2050 C. The present observations indicate that the previously reported increase in the 1200 C bend strength of reaction-bonded Si3N4, after pressing at 2050 C exclusively, is due to the relative oxygen levels of the isostatically hot pressed specimens; calculations show that an SiO2 content of less than about 2.2 vol pct is needed to yield the 1200 C strength increase.

Mieskowski, D. M.↗
DOE OSTI2025

Structure and Sulfur: Tuning the Viscoelastic and Surface Properties of Natural Keratin Fibers

Natural keratin fibers, such as wool, possess a complex hierarchical structure that governs their mechanical properties and surface energy. However, the extent to which these characteristics are influenced by combined contributions of structural variations (e.g., fiber diameter, intermediate filament (IF) packing) and chemical composition (e.g., disulfide bond density) remains poorly understood. In this study, we investigate wool fibers from five sheep breeds (Merino, Polwarth, Cheviot, Eider, and Devon) to elucidate how these factors influence viscoelasticity and surface interactions. Using a multimodal approach integrating interfacial and bulk characterization methods, including inverse gas chromatography (IGC), atomic force microscopy-infrared spectroscopy (AFM-IR), X-ray photoelectron spectroscopy (XPS), uniaxial tensile testing, and synchrotron small-angle X-ray scattering (SAXS), we show that the nanometer-thick 18-methyleicosanoic acid (18-MEA) layer is consistently present across all wool types and plays a key role in governing hydrophobicity and surface heterogeneity. A controlled isothermal treatment at 200 °C, designed to cleave disulfide bonds, results in a nearly 40% reduction in specific surface area across all fiber types, accompanied by a significant decrease in tensile strength and 80% reduction in elongation at break for Merino and Devon wool, but limited influence on the mechanical properties of Eider fibers. Furthermore, rate-dependent tensile testing within the elastic regime reveals distinct viscoelastic responses among the fiber types, suggesting that the sulfur-rich protein matrix surrounding IFs and its structure contribute actively to stress partitioning. Altogether, when combined with conclusions from SAXS measurements of IF spacing, our work offers compelling insights into the role of the keratin-associated protein (KAP) matrix in shaping wool fiber mechanics. Differences in mechanical behavior among wool types, despite similar IF spacing or sulfur content, highlight the importance of matrix composition and cross-linking density, suggesting that the molecular architecture of the KAP network may be a dominant factor in determining fiber performance.

X-ray scattering↗
DOE OSTI2023

Permethylation as a strategy for high–molecular–weight polysaccharide structure analysis by nuclear magnetic resonance—Case study of Xylella fastidiosa extracellular polysaccharide

Current practices for structural analysis of extremely large-molecular-weight polysaccharides via solution-state nuclear magnetic resonance (NMR) spectroscopy incorporate partial depolymerization protocols that enable polysaccharide solubilization in suitable solvents. Non-specific depolymerization techniques utilized for glycosidic bond cleavage, such as chemical degradation or ultrasonication, potentially generate structural fragments that can complicate complete and accurate characterization of polysaccharide structures. Utilization of appropriate enzymes for polysaccharide degradation, on the other hand, requires prior structural knowledge and optimal enzyme activity conditions that are not available to an analyst working with novel or unknown compounds. Herein, we describe an application of a permethylation strategy that allows the complete dissolution of intact polysaccharides for NMR structural characterization. This approach is utilized for NMR analysis of Xylella fastidiosa extracellular polysaccharide (EPS), which is essential for the virulence of the plant pathogen that affects multiple commercial crops and is responsible for multibillion dollar losses each year.

13C↗
DOE OSTI2020

Linking thermoset ink rheology to the stability of 3D-printed structures

Thermoset polymer composites show promise for additive manufacturing (AM) applications to address some of the limitations of the more widely used thermoplastic feedstock materials. Thermosets offer attractive mechanical properties while providing excellent interlayer bonding, high thermal and chemical stability, and reduced energy consumption as a result of deposition at room temperature. However, since thermoset resins rely on a crosslinking reaction to solidify, rather than quickly cooling like thermoplastics, viscoelastic properties must be relied upon to maintain deposited shape after deposition until crosslinking can occur. This fact has not impeded development and characterization of new thermoset feedstocks on the small scale, but recent efforts to increase scale of thermoset printing have highlighted issues with structural stability under self-weight. This study addresses issues of self-weight by investigating the mechanisms that cause collapse of tall, thin printed walls. Using nanoclay- and fumed silica-filled epoxy feedstocks, this work compares the collapse height for printed walls to stability models based on yielding and buckling mechanics. Inputs for these models – shear yield stress and storage modulus – were taken directly from parallel plate rheometry measurements. Model predictions were found to be in good agreement with experimental results, where both yielding and buckling behavior were observed, provided the rheological properties after a shear excursion were used as inputs. This work establishes a direct link between basic rheological properties of the feedstock, geometry of the printed object, and achievable height. Overall, the results presented highlight the importance of understanding recovery behavior in thermoset feedstocks and provide valuable guidance on the development of more effective direct-ink writing feedstock materials.

36 MATERIALS SCIENCE↗
DOE OSTI2021

Multiple N–H and C–H Hydrogen Atom Abstractions Through Coordination-Induced Bond Weakening at Fe-Amine Complexes

We report the use of the reported Fe-phthalocyanine complex, PcFe (1; Pc = 1,4,8,11,15,18,22,25-octaethoxy-phthalocyanine), to generate PcFe–amine complexes 1-(NH 3 ) 2 , 1-(MeNH 2 ) 2 , and 1-(Me 2 NH) 2 . Treatment of 1 or 1-(NH 3 ) 2 to an excess of the stable aryloxide radical, 2,4,6-tritert-butylphenoxyl radical ( t Bu ArO•), under NH 3 resulted in catalytic H atom abstraction (HAA) and C–N coupling to generate the product 4-amino-2,4,6-tritert-butylcyclohexa-2,5-dien-1-one (2) and t Bu ArOH. Exposing 1-(NH 3 ) 2 to an excess of the trityl (CPh 3 ) variant, 2,6-di-tert-butyl-4-tritylphenoxyl radical ( Tr ArO • ), under NH 3 did not lead to catalytic ammonia oxidation as previously reported in a related Ru-porphyrin complex. However, pronounced coordination-induced bond weakening of both α N–H and β C–H in the alkylamine congeners, 1-(MeNH 2 ) 2 and 1-(Me 2 NH) 2 , led to multiple HAA events yielding the unsaturated cyanide complex, 1-(MeNH 2 )(CN), and imine complex, 1-(MeN=CH 2 ) 2 , respectively. Subsequent C–N bond formation was also observed in the latter upon addition of a coordinating ligand. Detailed computational studies support an alternating mechanism involving sequential N–H and C–H HAA to generate these unsaturated products.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗
DOE OSTI2024

A DFT Comparison of C–C Reductive Coupling from Terminal Cyanido and Cyaphido Complexes of Nickel

The density functional theory study of the thermal C–C reductive coupling from terminal cyanido and hypothetical cyaphido complexes of [Ni(dmpe)] (dmpe = 1,2-bis(dimethylphosphino)ethane) revealed the key reaction intermediate in the reductive C–CP coupling being a σ-CC complex unlike an η 2 -aryl complex in the Ni C–CN system, as already observed in our previous studies. The reaction in THF is endothermic by 4.9 kcal/mol for cyanido with a 32.0 kcal/mol activation barrier and exothermic by 28.5 kcal/mol for cyaphido with an 11.3 kcal/mol activation barrier. To compare our results with the existing experimental data, we chose mesityl as the aryl group and also studied the CP reaction with [Pt(dmpe)] and [Pt(dmpm)] (dmpe = 1,2-bis(dimethylphosphino)methane) fragments. Our findings are consistent with the thermodynamically uphill photolytic C–CP bond activation in phosphaalkynes with Pt and a faster thermal back-reaction with [Pt(dmpe)] compared to that of [Pt(dmpm)]. Furthermore, based on the natural population analysis, when the polarity of the C–C bond is inverted, the sign of ΔG° is also inverted.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗
DOE OSTI2020

Coupled Cluster Study of the Interactions of AnO 2 , AnO 2 + , and AnO 2 2+ (An = U, Np) with N 2 and CO

We report thermochemical and spectroscopic properties for actinyl complexes involving UO 2 2+/1+/0 and NpO 2 2+/1+/0 with N 2 and CO, together with the UO 2 -O 2 , UO 2 + -O 2 , and UO 2 + -NO complexes, have been studied for the first time using an accurate composite coupled cluster approach. Two general bonding motifs were investigated, end-on (η 1 ) and side-on (η 2 ) relative to the metal center of the actinyls. For end-on CO complexes, both C-coordinated (An-C) and O-coordinated (An-O) structures were examined, with the former always being lower in energy. All of the η 1 complexes were calculated to be stable, with dissociation energies ranging from 2 to 36 kcal/mol, except for that of UO 2 + -O 2 (the η 1 orientation for UO 2 + -NO was not amenable to single reference coupled cluster). In agreement with a previous study, the η 2 structure for UO 2 + -O 2 was calculated to be relatively strongly bound, by 22.3 kcal/mol in this work. The closely related NO complex, however, had a calculated dissociation energy of just 4.0 kcal/mol. The binding energy of O 2 to neutral UO 2 in a η 2 orientation was calculated to be very strong, 75.4 kcal/mol, and strongly resembled a UO 2 + (O 2 – ) complex at equilibrium. The N–N and C–O bonds were found to be somewhat activated for all the side-on (η 2 ) neutral An(IV) complexes, with stretching frequencies of N 2 or CO being red-shifted by as much as 480 cm –1 with a 0.06 Å bond length elongation. Dissociation energies for the η 1 complexes are strongly correlated with the extent of electron transfer from ligand to actinyl. The nature of bonding in the actinyl complexes is examined using natural resonance theory (NRT). The correlation between bonding motif and small molecule activation is in agreement with experiments in condensed phases.

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

HEPOM: Using Graph Neural Networks for the Accelerated Predictions of Hydrolysis Free Energies in Different pH Conditions

Hydrolysis is a fundamental family of chemical reactions where water facilitates the cleavage of bonds. The process is ubiquitous in biological and chemical systems, owing to water’s remarkable versatility as a solvent. However, accurately predicting the feasibility of hydrolysis through computational techniques is a difficult task, as subtle changes in reactant structure like heteroatom substitutions or neighboring functional groups can influence the reaction outcome. Furthermore, hydrolysis is sensitive to the pH of the aqueous medium, and the same reaction can have different reaction properties at different pH conditions. In this work, we have combined reaction templates and high-throughput ab initio calculations to construct a diverse data set of hydrolysis free energies. The developed framework automatically identifies reaction centers, generates hydrolysis products, and utilizes a trained graph neural network (GNN) model to predict ΔG values for all potential hydrolysis reactions in a given molecule. The long-term goal of the work is to develop a data-driven, computational tool for high-throughput screening of pH-specific hydrolytic stability and the rapid prediction of reaction products, which can then be applied in a wide array of applications including chemical recycling of polymers and ion-conducting membranes for clean energy generation and storage.

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

Including Physics-Informed Atomization Constraints in Neural Networks for Reactive Chemistry

Machine learning interatomic potentials (MLIPs) have emerged as powerful tools for investigating atomistic systems with high accuracy and a relatively low computational cost. However, a common and unaddressed challenge with many current neural network (NN) MLIP models is their limited ability to accurately predict the relative energies of systems containing isolated or nearly isolated atoms, which appear in various reactive processes. To address this limitation, we present a mathematical technique for modifying any existing atom-centered NN architecture to account for the energies of isolated atoms. The result produces a consistent prediction of the atomization energy (AE) of a system using minimal constraints on the model. Using this technique, we build a model architecture that we call hierarchically interacting particle neural network (HIP-NN)-AE, an AE-constrained version of the HIP-NN, as well as ANI-AE, the AE-constrained version of the accurate NN engine for molecular energies (ANI). Our results demonstrate AE consistency of AE-constrained models, which drastically improves the AE predictions for the models. We compare the AE-constrained approach to unconstrained models as well as models from the literature in other scenarios, such as bond dissociation energies, bond dissociation pathways, and extensibility tests. These results show that the constraints improve the model performance in some of these tasks and do not negatively affect the performance on any tasks. The AE constraint approach thus offers a robust solution to the challenges posed by isolated atoms in energy prediction tasks.

74 ATOMIC AND MOLECULAR PHYSICS↗