Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “constrained density functional theory”

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 145 records · Page 8
All sourcesNASA NTRSDOE OSTI
DOE OSTI2022

Comparison of Doppler back-scattering and charge exchange measurements of E × B plasma rotation in the DIII-D tokamak under varying torque conditions

Measurements of the E × B toroidal angular velocity, $\omega_{E \times B}\,=\,E_\mathrm{r}/RB_\theta$ ($E_\mathrm{r}$ is the radial electric field, B θ is the poloidal magnetic field), are made using the Doppler back-scattering (DBS) and charge-exchange recombination (CER) spectroscopy diagnostics. DBS uses the Doppler shift of wavenumber-resolved density fluctuations while CER uses the Doppler shift of impurity emission lines to independently measure plasma parameters for calculating the local radial electric field. DBS and CER profiles of $\omega_{E \times B}$ as a function of normalized toroidal flux (ρ) are compared at various levels of neutral beam applied torque on the plasma. Under standard neoclassical theory $\omega_{E\times B}$ is a flux surface quantity, making it appropriate to compare across diagnostics. DBS and CER generally show good agreement when comparing $\omega_{E \times B}$ profiles at different levels of neutral beam injection-applied torque. Furthermore, the DBS values have close to the same precision as CER values when averaged over a similar time-scale and effects, such as prompt-torque are considered. DBS is able to observe the rapid ($\lt$10 ms) modification of the $E_\mathrm{r}$ profile by the diagnostic neutral beam 'blips'. This modification is most pronounced when the blip applies a large relative change in torque on the plasma. Overall, these results could have implications on transport analysis and suggests using DBS and CER in conjunction to constrain values of the E × B-shear (sometimes called $\gamma_{E \times B}$).

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗
DOE OSTI2021

Bayesian uncertainty quantification for nuclear matter incompressibility

Within a Bayesian statistical framework using the standard Skyrme-Hartree-Fock model, the maximum a posteriori (MAP) values and uncertainties of nuclear matter incompressibility and isovector interaction parameters are inferred from the experimental data of giant resonances and neutron-skin thicknesses of typical heavy nuclei. Furthermore, with the uncertainties of the isovector interaction parameters constrained by the data of the isovector giant dipole resonance and the neutron-skin thickness, we have obtained K 0 = 223$^{+7}_{–8}$ MeV at 68% confidence level using the data of the isoscalar giant monopole resonance in 208 Pb measured at the Research Center for Nuclear Physics (RCNP), Japan, and at the Texas A&M University, USA. Although the corresponding 120 Sn data gives a MAP value for K 0 about 5 MeV smaller than the 208 Pb data, there are significant overlaps in their posterior probability distribution functions.

190 ≤ A ≤ 219↗
DOE OSTI2025

Highly dense atomic Fe–Ni dual metal sites for efficient CO 2 to CO electrolyzers at industrial current densities

Carbon-supported, atomically dispersed, nitrogen-coordinated metal sites (e.g., Fe and Ni) are arguably the most promising catalysts for the electrochemical reduction of CO 2 to CO due to their unique catalytic properties and the use of earth-abundant elements. However, conventional single metal sites are constrained by their structural simplicity, causing either too weak or too strong absorption/desorption of multiple critical intermediates (e.g., *COOH and *CO). Current catalysts also suffer from ultra-low loadings (<1.0 wt%) of atomic metal active sites in catalysts, leading to inadequate performance for CO 2 -to-CO conversion. Here, we develop dual Ni/Fe metal site catalysts with significantly increased atomically dispersed metal loadings (up to 4.8 wt%). A gas-phase chemical vapor deposition (CVD) approach to introducing single Ni sites was integrated with Fe 2 O 3 /ZIF-8 precursors, followed by an optimal thermal activation. The optimized CVD-Ni/Fe–N–C catalyst exhibited remarkable electrocatalytic performance for the CO 2 reduction to CO in a continuous membrane-electrode-assembly electrolyzer, achieving a maximum CO faradaic efficiency (FE CO ) of 96% at a current density of 700 mA cm −2 in a near-neutral electrolyte. Furthermore, a desirable but challenging acidic flow-cell electrolyzer was designed using this dual metal site catalyst to improve CO 2 utilization, accomplishing a FE CO of up to 95% at a CO partial current density close to 600 mA cm −2 . Density functional theory (DFT) calculations suggest a synergetic effect between Fe–Ni pairs facilitating *COOH intermediate formation and *CO desorption simultaneously during CO 2 to CO conversion. This is key to breaking the linear scaling relationship of conventional single-metal site catalysts during the CO 2 reduction reaction.

36 MATERIALS SCIENCE↗
DOE OSTI2024

Learning thermodynamically constrained equations of state with uncertainty

Numerical simulations of high energy-density experiments require equation of state (EOS) models that relate a material’s thermodynamic state variables—specifically pressure, volume/density, energy, and temperature. EOS models are typically constructed using a semi-empirical parametric methodology, which assumes a physics-informed functional form with many tunable parameters calibrated using experimental/simulation data. Since there are inherent uncertainties in the calibration data (parametric uncertainty) and the assumed functional EOS form (model uncertainty), it is essential to perform uncertainty quantification (UQ) to improve confidence in EOS predictions. Model uncertainty is challenging for UQ studies since it requires exploring the space of all possible physically consistent functional forms. Thus, it is often neglected in favor of parametric uncertainty, which is easier to quantify without violating thermodynamic laws. This work presents a data-driven machine learning approach to constructing EOS models that naturally captures model uncertainty while satisfying the necessary thermodynamic consistency and stability constraints. We propose a novel framework based on physics-informed Gaussian process regression (GPR) that automatically captures total uncertainty in the EOS and can be jointly trained on both simulation and experimental data sources. A GPR model for the shock Hugoniot is derived, and its uncertainties are quantified using the proposed framework. We apply the proposed model to learn the EOS for the diamond solid state of carbon using both density functional theory data and experimental shock Hugoniot data to train the model and show that the prediction uncertainty is reduced by considering thermodynamic constraints.

Sharma, Himanshu (ORCID:000900050235718X)↗
DOE OSTI2020

Correlations between triaxial shapes and formation history of dark matter haloes

ABSTRACT The shape of dark matter (DM) haloes plays a critical role in constraining cosmology with upcoming large-scale structure surveys. In this paper, we study the correlations between the triaxial shapes and formation histories in DM haloes in the MultiDark Planck 2 N-body cosmological simulation. We find that halo ellipticity is strongly correlated with halo properties that serve as proxies of halo formation history, such as halo concentration and the normalized peak-centroid offset. These correlations are nearly independent of the halo density peak height. We present a simple model for the correlation between halo ellipticity and concentration using conditional abundance matching, and provide fitting formulae for the multidimensional distributions of triaxial halo shape as a function of halo peak height. We apply our halo shape model to gauge the effects of halo ellipticity and orientation bias on the excess surface mass density profiles in cluster-sized haloes. Our model should be useful for exploring the impact of triaxial halo shape on cosmological constraints in upcoming weak lensing surveys of galaxy clusters.

79 ASTRONOMY AND ASTROPHYSICS↗
DOE OSTI2021

Exploring and enhancing the accuracy of interior-scaled Perdew–Zunger self-interaction correction

The Perdew–Zunger self-interaction correction (PZ-SIC) improves the performance of density functional approximations for the properties that involve significant self-interaction error (SIE), as in stretched bond situations, but overcorrects for equilibrium properties where SIE is insignificant. This overcorrection is often reduced by local scaling self-interaction correction (LSIC) of the PZ-SIC to the local spin density approximation (LSDA). Here, we propose a new scaling factor to use in an LSIC-like approach that satisfies an additional important constraint: the correct coefficient of the atomic number Z in the asymptotic expansion of the exchange–correlation (xc) energy for atoms. LSIC and LSIC+ are scaled by functions of the iso-orbital indicator z σ , which distinguishes one-electron regions from many-electron regions. LSIC+ applied to the LSDA works better for many equilibrium properties than LSDA-LSIC and the Perdew, Burke, and Ernzerhof generalized gradient approximation (GGA), and almost close to the strongly constrained and appropriately normed (SCAN) meta-GGA. LSDA-LSIC and LSDA-LSIC+, however, fail to predict interaction energies involving weaker bonds, in sharp contrast to their earlier successes. It is found that more than one set of localized SIC orbitals can yield a nearly degenerate energetic description of the same multiple covalent bond, suggesting that a consistent chemical interpretation of the localized orbitals requires a new way to choose their Fermi orbital descriptors. To make a locally scaled down SIC to functionals beyond the LSDA requires a gauge transformation of the functional’s energy density. Here, the resulting SCAN-sdSIC, evaluated on SCAN-SIC total and localized orbital densities, leads to an acceptable description of many equilibrium properties including the dissociation energies of weak bonds.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗
DOE OSTI2023

Nuclear and hybrid equations of state in light of the low-mass compact star in HESS J1731-347

Here, we sample over 500 relativistic mean-field theories constrained by chiral effective-field theory and properties of isospin-symmetric nuclear matter and test them against known stellar structure constraints. This includes a recent mass and radius measurement of a compact object in supernova remnant HESS J1731-347, with an unusually low mass of M = 0.77$^{+0.20}_{–0.17}$M ⊙ and a compact radius of R = 10.4$^{+0.86}_{–0.78}$ km. We show that none of the sampled nuclear models meet all constraints at the 68% credibility level, but that hybrid equations of state with a quark-matter inner core and nuclear outer core easily can. This indicates a tension between astrophysical constraints and low-energy nuclear theory.

73 NUCLEAR PHYSICS AND RADIATION PHYSICS↗
DOE OSTI2023

Synergistic double-atom catalysts of metal-boron anchored on g-C 2 N for electrochemical nitrogen reduction: Mechanistic insight and catalyst screening

The rational design of a novel catalytic center with a sound basis remains both challenging and rewarding for the electrochemical reduction of N 2 (eNRR), which has provided a feasible route for achieving clean and sustainable NH 3 production under ambient conditions. Herein, using density functional theory calculations, we demonstrate that hybrid metal (M)-boron (B) double-atom catalysts (DACs) embedded in g-C 2 N substrate (M-B@C 2 N, M = 3d, 4d and 5d transition metals) can achieve both high catalytic activity and high selectivity in eNRR. The proposed M-B@C 2 N DACs have exhibited impressive feasibility and stability thanks to the resilient and robust C 2 N substrate with abundant pyridinic N atoms distributed among right-sized pore structures. Our results reveal that like the metal center, the embedded B atom can actively involve in Ntriple bondN bond activation via π*-backdonation mechanism concomitant with the substantial charge transfer to adsorbed *N 2 , leading to sizable Nsingle bondN bond elongation. Accordingly, both adsorption energy and Nsingle bondN bond length of *N 2 can be employed as catalytic descriptors for predicting eNRR activity in terms of the limiting potentials (UL). Using high-throughput screening method, we found that six M-B@C 2 N candidates have stood out as the outstanding electrocatalysts for driving eNRR, namely, M = Ti (U L = 0 V), Mo (U L = 0 V), Nb (U L = -0.04 V), W (U L = -0.23 V), Zr (U L = -0.26 V), V (U L = -0.28 V). The underlying origin is attributed to the balanced and constrained N-affinity of M-B dual site working in synergy, which can thus be used as one important guide of catalyst design.

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

Baryon fraction from the BAO amplitude: a consistent approach to parameterizing perturbation growth

Galaxy clustering constrains the baryon fraction Omega_b/Omega_m through the amplitude of baryon acoustic oscillations and the suppression of perturbations entering the horizon before recombination. This produces a different pre-recombination distribution of baryons and dark matter. After recombination, the gravitational potential responds to both components in proportion to their mass, allowing robust measurement of the baryon fraction. This is independent of new-physics scenarios altering the recombination background (e.g. Early Dark Energy). The accuracy of such measurements does, however, depend on how baryons and CDM are modeled in the power spectrum. Previous template-based splitting relied on approximate transfer functions that neglected part of information. We present a new method that embeds an extra parameter controlling the balance between baryons and dark matter in the growth terms of the perturbation equations in the CAMB Boltzmann solver. This approach captures the baryonic suppression of CDM prior to recombination, avoids inconsistencies, and yields a clean parametrization of the baryon fraction in the linear power spectrum, separating out the simple physics of growth due to the combined matter potential. We implement this framework in an analysis pipeline using Effective Field Theory of Large-Scale Structure with HOD-informed priors and validate it against noiseless LCDM and EDE cosmologies with DESI-like errors. The new scheme achieves comparable precision to previous splitting while reducing systematic biases, providing a more robust way to baryon-fraction measurements. In combination with BBN constraints on the baryon density and Alcock-Paczynski estimates of the matter density, these results strengthen the use of baryon fraction measurements to derive a Hubble constant from energy densities, with future DESI and Euclid data expected to deliver competitive constraints.

Crespi, Andrea [U. Waterloo (main); Waterloo U., I↗
DOE OSTI2022

Platform for probing radiation transport properties of hydrogen at conditions found in the deep interiors of red dwarfs

We describe an experimental concept at the National Ignition Facility for specifically tailored spherical implosions to compress hydrogen to extreme densities (up to ∼800× solid density, electron number density ne∼4×1025 cm−3) at moderate temperatures (T∼200 eV), i.e., to conditions, which are relevant to the interiors of red dwarf stars. The dense plasma will be probed by laser-generated x-ray radiation of different photon energy to determine the plasma opacity due to collisional (free–free) absorption and Thomson scattering. The obtained results will benchmark radiation transport models, which in the case for free–free absorption show strong deviations at conditions relevant to red dwarfs. This very first experimental test of free–free opacity models at these extreme states will help to constrain where inside those celestial objects energy transport is dominated by radiation or convection. Moreover, our study will inform models for other important processes in dense plasmas, which are based on electron–ion collisions, e.g., stopping of swift ions or electron–ion temperature relaxation.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗
DOE OSTI2022

Construction of meta-GGA functionals through restoration of exact constraint adherence to regularized SCAN functionals

The strongly constrained and appropriately normed (SCAN) meta-GGA exchange–correlation functional is constructed as a chemical environment-determined interpolation between two separate energy densities: one describes single-orbital electron densities accurately and another describes slowly varying densities accurately. To conserve constraints known for the exact exchange–correlation functional, the derivatives of this interpolation vanish in the slowly varying limit. While theoretically convenient, this choice introduces numerical challenges that degrade the functional’s efficiency. We have recently reported a modification to the SCAN meta-GGA, termed restored-regularized-SCAN (r 2 SCAN) , that introduces two regularizations into SCAN, which improve its numerical performance at the expense of not recovering the fourth order term of the slowly varying density gradient expansion for exchange. Here, we show the derivation of a progression of density functional approximations [regularized SCAN (rSCAN), r++SCAN, r 2 SCAN, and r 4 SCAN] with increasing adherence to exact conditions while maintaining a smooth interpolation. The greater smoothness of r 2 SCAN seems to lead to better general accuracy than the additional exact constraint of SCAN or r 4 SCAN does.

, correlation energy↗
DOE OSTI2021

r 2 SCAN-D4: Dispersion corrected meta-generalized gradient approximation for general chemical applications

We combine a regularized variant of the strongly constrained and appropriately normed semilocal density functional with the latest generation semi-classical London dispersion correction. The resulting density functional approximation r 2 SCAN-D4 has the speed of generalized gradient approximations while approaching the accuracy of hybrid functionals for general chemical applications. We demonstrate its numerical robustness in real-life settings and benchmark molecular geometries, general main group and organo-metallic thermochemistry, and non-covalent interactions in supramolecular complexes and molecular crystals. Main group and transition metal bond lengths have errors of just 0.8%, which is competitive with hybrid functionals for main group molecules and outperforms them for transition metal complexes. The weighted mean absolute deviation (WTMAD2) on the large GMTKN55 database of chemical properties is exceptionally small at 7.5 kcal/mol. This also holds for metal organic reactions with an MAD of 3.3 kcal/mol. The versatile applicability to organic and metal–organic systems transfers to condensed systems, where lattice energies of molecular crystals are within the chemical accuracy (errors <1 kcal/mol).

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗
DOE OSTI2022

Structural and dynamic properties of solvated hydroxide and hydronium ions in water from ab initio modeling

Predicting the asymmetric structure and dynamics of solvated hydroxide and hydronium in water from ab initio molecular dynamics (AIMD) has been a challenging task. The difficulty mainly comes from a lack of accurate and efficient exchange–correlation functional in elucidating the amphiphilic nature and the ubiquitous proton transfer behaviors of the two ions. By adopting the strongly constrained and appropriately normed (SCAN) meta-generalized gradient approximation functional in AIMD simulations, we systematically examine the amphiphilic properties, the solvation structures, the electronic structures, and the dynamic properties of the two water ions. In particular, we compare these results to those predicted by the PBE0-TS functional, which is an accurate yet computationally more expensive exchange–correlation functional. We demonstrate that the general-purpose SCAN functional provides a reliable choice for describing the two water ions. Specifically, in the SCAN picture of water ions, the appearance of the fourth and fifth hydrogen bonds near hydroxide stabilizes the pot-like shape solvation structure and suppresses the structural diffusion, while the hydronium stably donates three hydrogen bonds to its neighbors. We apply a detailed analysis of the proton transfer mechanism of the two ions and find the two ions exhibit substantially different proton transfer patterns. In conclusion, our AIMD simulations indicate that hydroxide diffuses more slowly than hydronium in water, which is consistent with the experimental results.

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

Compositional phase stability of correlated electron materials within DFT+DMFT

Predicting the compositional phase stability of strongly correlated electron materials is an outstanding challenge in condensed matter physics, requiring precise computations of total energies. In this work, we employ the density functional theory plus dynamical mean-field theory (DFT+DMFT) formalism to address local correlations due to transition metal $\textit{d}$ electrons on compositional phase stability in the prototype rechargeable battery cathode material Li x CoO 2 , and detailed comparisons are made with the simpler DFT+$\textit{U}$ approach (i.e., the Hartree-Fock solution of the DMFT impurity problem). Local interactions are found to strongly impact the energetics of the band insulator LiCoO 2 , most significantly via the $E_g$ orbitals, which are partially occupied via hybridization with $\textit{O p}$ states. We find CoO 2 and Li 1/2 CoO 2 to be moderately correlated Fermi liquids with quasiparticle weights of 0.6–0.8 for the $T_{2g}$ states, which are most impacted by the interactions. Further, as compared to DFT+$\textit{U}$, DFT+DMFT considerably dampens the increase in total energy as $\textit{U}$ is increased, which indicates that dynamical correlations are important to describe this class of materials despite the relatively modest quasiparticle weights. Unlike DFT+$\textit{U}$, which can incorrectly drive Li x CoO 2 toward spurious phase separating or charge-ordered states, DFT+DMFT correctly captures the system's phase stability and does not exhibit a strong charge-ordering tendency. Most importantly, the error within DFT+$\textit{U}$ varies strongly as the composition changes, challenging the common practice of artificially tuning $\textit{U}$ within DFT+$\textit{U}$ to compensate the errors of Hartree-Fock. DFT+DMFT predicts the average intercalation voltage decreases relative to DFT, opposite to the result of DFT+$\textit{U}$, which would yield favorable agreement with experiment in conjunction with the overprediction of the voltage by the strongly constrained and appropriately normed DFT functional.

75 CONDENSED MATTER PHYSICS, SUPERCONDUCTIVITY AND↗
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↗
DOE OSTI2021

encore : an O ( N g2) estimator for galaxy N -point correlation functions

ABSTRACT We present a new algorithm for efficiently computing the N-point correlation functions (NPCFs) of a 3D density field for arbitrary N. This can be applied both to a discrete spectroscopic galaxy survey and a continuous field. By expanding the statistics in a separable basis of isotropic functions built from spherical harmonics, the NPCFs can be estimated by counting pairs of particles in space, leading to an algorithm with complexity $\mathcal {O}(N_\mathrm{g}^2)$ for Ng particles, or $\mathcal {O}(N_\mathrm{FFT}\log N_\mathrm{FFT})$ when using a Fast Fourier Transform with NFFT grid-points. In practice, the rate-limiting step for N > 3 will often be the summation of the histogrammed spherical harmonic coefficients, particularly if the number of radial and angular bins is large. In this case, the algorithm scales linearly with Ng. The approach is implemented in the encore code, which can compute the 3PCF, 4PCF, 5PCF, and 6PCF of a BOSS-like galaxy survey in ${\sim}100$ CPU-hours, including the corrections necessary for non-uniform survey geometries. We discuss the implementation in depth, along with its GPU acceleration, and provide practical demonstration on realistic galaxy catalogues. Our approach can be straightforwardly applied to current and future data sets to unlock the potential of constraining cosmology from the higher point functions.

79 ASTRONOMY AND ASTROPHYSICS↗
DOE OSTI2021

Information content of higher order galaxy correlation functions

The shapes of galaxy N-point correlation functions can be used as standard rulers to constrain the distance–redshift relationship. The cosmological density fields traced by late-time galaxy formation are initially nearly Gaussian, and hence, all the cosmological information can be extracted from their two-point correlation function. Subsequent non-linear evolution under gravity, as well as halo and then galaxy formation, generates higher order correlation functions. Since the mapping of the initial to the final density field is, on large scales, invertible, it is often claimed that the information content of the initial field’s power spectrum is equal to that of all the higher order functions of the final, non-linear field. This claim implies that reconstruction of the initial density field from the non-linear field renders analysis of higher order correlation functions of the latter superfluous. We show that this claim is false when the N-point functions are used as standard rulers. Constraints available from joint analysis of the two and three-point correlation functions can, in some cases, exceed those offered by the initial power spectrum. We provide a mathematical justification for this claim and demonstrate it using a large suite of N-body simulations. In particular, we show that for the z = 0 real-space matter field in the limit of vanishing shot-noise, taking modes up to k max = 0.2 h Mpc −1 , using the bispectrum alone offers a factor of 2 reduction in the variance on the cosmic distance scale relative to that available from the linear power spectrum.

79 ASTRONOMY AND ASTROPHYSICS↗
DOE OSTI2025

Data-Driven Kinetic Reaction Networks for Separation Chemistry

Understanding complex, multistep chemical reactions at the molecular level is a major challenge whose solution would greatly benefit the design and optimization of numerous chemical processes. The separation of rare-earth (4f) and actinide (5f) elements is an example where improving our chemical understanding is important for designing and optimizing new chemistries, even with a limited number of observations. Here, in this work, we leverage data-driven artificial intelligence and machine-learning approaches to develop kinetic reaction networks that describe the liquid–liquid extraction mechanism of uranium using N,N-di-2-ethylhexyl-isobutyramide (DEHiBA). Specifically, we compare and contrast the properties of two classes of models: (1) purely data-driven models that are regularized using chemistry-agnostic, L1 regression and (2) chemistry-informed models that are regularized using relative reaction energies provided by quantum mechanical calculations. We observe that purely data-driven models are unbiased, simple, and accurate in their predictions of experimental measurements when provided with sufficient data but are difficult to fully constrain and interpret. In contrast, chemistry-informed models exhibit significantly improved chemical interpretability and consistency, providing a detailed description of the separation process while achieving high accuracy through ensemble averaging. Overall, the dominant species predicted to be extracted into the organic phase is UO 2 (NO 3 ) 2 (DEHiBA) 2 , agreeing with experimental slope analysis, thermodynamic modeling, EXAFS, and crystal structures. This work demonstrates that leveraging the fundamental structure of the problem can lead to efficient learning schemes that provide both accurate predictions and chemical insights at a low computational cost.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗