John A. Keith

Big‐data approaches to chemistry falter when experimental data remain scarce, expensive, and noisy. Here we demonstrate that physics‐informed descriptors overcome this limitation and enable accurate predictions from severely limited datasets. Taking capped diisocyanate deblocking as a test case, we introduce descriptors derived from the bond energies from bond orders and populations (BEBOP) method that capture electronic structure and bonding at low computational cost. Trained on only 19 experimental compounds, BEBOP‐informed LASSO regression predicts deblocking temperatures within an RMSE of 11 °C under fully nested leave‐one‐out cross‐validation over a 227–323 °C range, while conventional molecular descriptors fail to produce any meaningful correlation. We validate the resulting physically interpretable model on two new compounds, one within the training range and one outside of it. Embedding physical insight into descriptors rather than relying on algorithmic complexity reduces data requirements by an order of magnitude and offers a generalizable strategy for molecular design problems wherever large datasets are impractical.

We present PAMOS (Parameterized Approximate Molecular Orbital method with Slater orbitals), a semiempirical electronic structure method that keeps a minimal Slater-type basis set yet captures the response of atomic orbitals to their molecular environment. Rather than fit many coupled parameters to a training set, we build the model from the most elementary molecular species and introduce each parameter to emulate one feature of ab initio theory that a minimal basis cannot reach. Two devices distinguish the method. Different classes of integrals receive different effective Slater exponents, ζ _{e N} for electron–nuclear attraction and ζ _S for overlap, so a single basis function emulates the radial contraction and polarization that would otherwise require an extended basis. The off-diagonal Hamiltonian matrix elements satisfy a conservation-of-charge constraint that cancels the interatomic electrostatic interactions at large internuclear separation. Together with a bonding parameter λ and a heteronuclear weighting exponent p, these devices reproduce benchmark potential energy curves for ten one-electron diatomic ions, from H2++ to BeLi ⁶⁺ , to within ∼1 kcal/mol ( Z ² -scaled). A core-correction strategy extends the model to one-valence-electron ions from Li2+ to BC ⁶⁺ . We close with a preview of a two-electron treatment of static correlation and a route to dynamic correlation.

Big-data approaches to chemistry falter when experimental data remain scarce, expensive, and noisy. Here we demonstrate that physics-informed descriptors overcome this limitation and enable accurate predictions from severely limited datasets. Taking capped diisocyanate deblocking as a test case, we introduce descriptors derived from the Bond Energies from Bond Orders and Populations (BEBOP) method that capture electronic structure and bonding at low computational cost. Trained on only 19 experimental compounds, BEBOP-informed LASSO regression predicts deblocking temperatures within an RMSE of ~11 °C over a 227–323 °C range, while conventional molecular descriptors fail to produce any meaningful correlation. We validate the resulting physically interpretable model on two new compounds, one within the training range and one outside of it. Embedding physical insight into descriptors rather than relying on algorithmic complexity reduces data requirements by an order of magnitude and offers a generalizable strategy for molecular design problems wherever large datasets are impractical.

Per- and polyfluoroalkyl substances (PFAS) remain challenging for quantum mechanical (QM) modeling due to their atypical electronic structures, large number of compounds, and limited experimental data for validation. To address this, we constructed a comprehensive benchmarking database that evaluates the performance of a broad range of QM methods. The database benchmarks thermal properties (BDEs) and electronic properties (dipole moments, electron affinities, ionization potentials), validated against experimental data where available and against high-level theoretical references. All data are released through the open and extensible PFAS_Database on GitHub, which is designed to grow over time, both in number of PFAS and inclusion of newly identified top-performing methods, providing a foundation for more reliable modeling and remediation strategies.

Molecular dipole moments are required for empirical correlations used to estimate thermophysical properties of zinc dialkyldithiophosphates (ZDDPs), an important class of anti-wear additives. We calculated dipole moments for three ZDDPs and their precursor dialkyldithiophosphates using density functional theory and explicitly accounted for their dependence on molecular conformation. We unexpectedly found that dipole ∗Corresponding author. Email: karlj@pitt.edu moments are highly sensitive to conformational degrees of freedom. The dipole moments for each of these six molecules range from almost zero to almost 5 D across different conformers. Surprisingly, this variation appears to be dominated by the conformations of the alkyl chains rather than the polar inorganic core. Despite this wide range, the arithmetic mean and Boltzmann weighted average dipole moments fall within a narrow range of 2.2–2.4 D for all six molecules examined. This consistency arises because the dipole moment shows no correlation with the conformational energy. The robust average value significantly simplifies property estimation by eliminating the need for conformation-specific dipole calculations. We recommend that thermophysical property correlations use a constant dipole moment of 2.3 D for ZDDPs and their precursors. We predict that experimental measurements will yield values close to this average, since they inherently sample an ensemble of conformations.

Coupled cluster theory with a Kohn-Sham reference (KS-CC) can dramatically outperform its Hartree-Fock counterpart for strongly correlated systems, but the origin of these improvements has remained unclear. Here we demonstrate that these improvements arise from differences in the one-particle density matrix that are encoded into the non-canonical Fock matrix and not from the nature of the KS orbitals, as is commonly assumed. Equipped with this insight, KS-CCSD(T) can be leveraged to achieve near-chemical-accuracy for electronic and thermochemical properties of transition-metal dimers and main-group compounds. Most strikingly, KS-CCSD(T) qualitatively recovers the entire Cr$_2$ potential energy surface, a notorious failure case for HF-CCSD(T) and single-reference density functional theory. We further introduce a density difference diagnostic that identifies multireference character and guides practitioners toward rational selections of optimal references at mean-field cost. These results establish KS-CCSD(T) as a practical route to treat strong correlation within the "gold standard" framework, and this has immediate implications for machine learning potential development and materials research, areas that heavily rely on KS-DFT for model-parameter fitting.

Metal oxide corrosion severely limits anodic electrocatalysis, particularly at high potentials in acidic environments, where degradation pathways remain poorly defined. This study establishes explicit connections between corrosion and electrocatalysis on tin oxide during water oxidation by examining the roles of lattice defects, reactive oxygen species, interfacial pH variations, and speciation of corroded tin in acid. We first demonstrate the presence of structural defects such as oxygen vacancies and substoichiometric Sn­(II) species by integrating electron paramagnetic resonance spectroscopy, ultraviolet photoelectron spectroscopy, and Mott–Schottky analysis. Kohn–Sham density functional theory calculations reveal that explicit water structures thermodynamically stabilize reaction intermediates and lower reaction overpotentials. Moreover, we propose that water dissociation leads to hydrogen-bonding networks formed by H* and OH* intermediates, which may span the entire catalyst surface and decrease the interfacial pH to drive corrosion. In contrast, the electrochemical generation of reactive oxygen species is shown to play a minor role in catalyst corrosion during water oxidation using inductively coupled plasma mass spectrometry coupled with selective chemical scavengers. Square-wave voltammetry combined with rotating ring-disk electrodes is used to reveal that under open-circuit conditions, only Sn­(IV) cations chemically dissolve from tin oxide, while both Sn­(IV) and Sn­(II) species electrochemically corrode during water oxidation. Our results unveil a dynamic and complicated interplay between corrosive and catalytic pathways on metal oxide electrocatalysts: a decrease in interfacial pH due to water oxidation exacerbates Sn­(II)/Sn­(IV) corrosion. Subsequently, the electrochemical corrosion of Sn­(II)/Sn­(IV) facilitates product formation from lattice oxygen, while the redeposition of corroded Sn­(II) as Sn­(IV) can enable oxygen exchange with water. By elucidating the roles of defects and interfacial chemistry, this work provides a roadmap for engineering improved electrocatalysts that balance activity and stability, a critical step toward scalable and durable energy technologies.