Bryan R. Goldsmith

A library of 186 million drug-like compounds was constructed and screened using a combined computational and experimental approach to identify three scaffold structures previously unreported as organic redox flow battery (ORFB) negolytes. Candidate scaffolds were identified through computational screening of reduction potentials, triaged via literature search, and characterized electrochemically. Three scaffolds—pyrido[1,2-a]pyrimidine, saccharin, and thiopyran—showed measurable redox activity. Initial cyclic voltammetry of synthesized derivatives revealed the key challenges for each that must be addressed for development into viable aqueous ORFB negolytes.

The electrochemical reduction of nitrate offers a pathway to convert industrial nitrate waste into valuable ammonia. However, developing dilute alloy catalysts (often referred to as “single-atom alloys”) that achieve both high selectivity and activity for the electrochemical nitrate reduction reaction (NO3RR) is limited by an incomplete mechanistic understanding of composition−activity relationships. Here, we prepared a series of shape-controlled, phase-pure, surfactant-free CuPd dilute alloys with systematically varied Pd dopant loadings at trace levels (0.05–0.9 wt%) into the Cu surface. Electrochemical studies in acidic phosphate buffer reveal a strong dependence of NO3RR performance on Pd content. Specifically, CuPd with 0.15 wt% Pd showed a marked enhancement in ammonium yields and Faradaic efficiency—achieving a maximum yield of 258 μmol mgcat−1 h−1 with over 80% Faradaic efficiency for NH4+ production at −0.8 V vs RHE. This rate of ammonium formation is greater than that observed on Cu2O, Cu, and other CuPd dilute alloys with lower (0.05 wt%) and higher (0.45 and 0.9 wt%) Pd loadings under identical conditions. Grand-canonical density functional theory calculations (GC DFT) reveal a Pd ensemble-size-dependent shift in selectivity between nitrate reduction to ammonium and the competing hydrogen evolution reaction. Post-electrolysis analysis demonstrates that the CuPd catalyst undergoes no substantial change in the morphology or composition after the NO3RR. These results underscore the advantages of trace Pd incorporation and ensemble control in tuning activity and selectivity, providing design principles for efficient NO3RR electrocatalysts.

Selective separation of like-charged ions is a central challenge in applications such as critical mineral recovery. Electrochemical membrane–based separations offer promising pathways to address this need, but a limited fundamental understanding of ion transport in charged polymer membranes hampers the development of highly selective materials. This study elucidates the role of specific ion effects (SIEs) in ion transport through charged polymer membranes by integrating experimental measurements of ion mobility and in situ ion/ion and ion/water interactions in model charged polymer membranes with results from molecular dynamics simulations on analogous systems. We demonstrate that solvent-mediated ion interactions drive pronounced SIEs within the studied membranes, with the ion softness, i.e., the malleability of ion hydration shells, from hard/soft acid/base (HSAB) theory emerging as a key predictor of transport properties. HSAB theory also explains the observed mechanism of solvent-mediated ion interactions. Our findings offer a mechanistic framework for designing membranes with tailored ion selectivity, potentially enabling efficient separations of chemically similar ions.

Nitrogen fixation heavily relies on the energy-intensive Haber-Bosch process, necessitating renewable alternatives. Here, we introduce a non-equilibrium spatially distributed electric field (SD-EF) strategy for nitrogen fixation in ambient air plasma. The optimized SD-EF strategy gives a NO_x⁻ yield of 9.8 mmol/h, tripling that of a uniform electric field and the N₂ conversion is three times higher than most discharge configurations at similar or lower energy consumptions. This high NO_x⁻ yield is achieved through simultaneously activating two beneficial kinetic networks via SD-EF by having both high and low electric fields present: O₃ and vibrational excitation of N₂ (N₂(v)) sub-mechanisms, which are revealed by developing a photonic crystal fiber diagnostic for in-situ quantification of molecules and ions (NO, NO₂, N₂O, O₃, NO₃⁻, and NO₂⁻) in gas-liquid plasma. The establishment of the SD-EF strategy, coupled with in-situ gas-liquid diagnostics, is broadly applicable to plasma-assisted nitrogen fixation and holds promise for other plasma-assisted chemical conversion processes.

Active data acquisition is central to many learning and optimization tasks in deep neural networks, yet remains challenging because most approaches rely on predictive uncertainty estimates that are difficult to obtain reliably. To this end, we propose Goal-Oriented Influence- Maximizing Data Acquisition (GOIMDA), an active acquisition algorithm that avoids explicit posterior inference while remaining uncertainty-aware through inverse curvature. GOIMDA selects inputs by maximizing their expected influence on a user-specified goal functional, such as test loss, predictive entropy, or the value of an optimizer-recommended design. Leveraging first-order influence functions, we derive a tractable acquisition rule that combines the goal gradient, training-loss curvature, and candidate sensitivity to model parameters. We show theoretically that, for generalized linear models, GOIMDA approximates predictive-entropy minimization up to a correction term accounting for goal alignment and prediction bias, thereby, yielding uncertainty-aware behavior without maintaining a Bayesian posterior. Empirically, across learning tasks (including image and text classification) and optimization tasks (including noisy global optimization benchmarks and neural-network hyperparameter tuning), GOIMDA consistently reaches target performance with substantially fewer labeled samples or function evaluations than uncertainty-based active learning and Gaussian-process Bayesian optimization baselines.

First-principles multiscale modeling shows that synergistic reactions of N˙ and H˙ radicals, rather than solely vibrationally excited N ₂ , control plasma-assisted ammonia synthesis on Fe and Ni catalysts, explaining their similar ammonia yields.

Bayesian optimization (BO) was used to accelerate the synthesis of hierarchical ZSM-5 with a balanced micro–mesoporous structure. Using mesoporosity and microporosity as dual objectives, a Gaussian process regression surrogate guided three successive BO iterations informed by 15 initial experiments. The optimized sample (HZ-R_0.55T₅₀t₂) exhibited the highest hierarchy factor (0.17), featuring similar mesoporosity (0.44) but higher microporosity (0.38) than the benchmark, indicating reduced diffusional resistance with preserved framework. Characterizations revealed that NaOH induced framework dissolution, whereas TPAOH moderated desilication; their mixture synergistically created uniform mesopores while maintaining crystallinity. ²⁷Al and ²⁹Si NMR confirmed realumination and silanol generation, and acidity analysis showed redistributed acid sites. Sensitivity analysis identified the TPAOH fraction and temperature as the dominant factors influencing the hierarchy factor. This study establishes a data-driven workflow for zeolite design, demonstrating that BO effectively accelerates hierarchical structure optimization while minimizing the experimental effort.