Department of Chemical and Biomolecular Engineering

Abstract: Amorphous aluminosilicates are essential components of fluid catalytic cracking (FCC) catalysts, where they provide structural support, hierarchical porosity, and acid functionality within the mesoporous matrix. A molecular-level description of the active acid site ensemble remains challenging because these materials are compositionally heterogeneous and structurally disordered. Most experimental and theoretical studies have focused on Si-doped γ–Al2O3 that serves as a convenient model system. Here, we move beyond this simplified representation toward more realistic and chemically complex aluminosilicate structures. We combine density functional theory (DFT), equivariant machine-learning interatomic potentials (MLIPs), and grand-canonical basin hopping (GCBH) to map structure-acidity relationships for aluminosilicate compositions that arise under FCC-relevant hydrothermal conditions (T = 550 °C and PH2O = 1.2 bar), representative of the reactor inlet. We study a family of mullite surfaces across Al2O3:SiO2 ratios (denoted n:m) and associated oxygen-vacancy contents, and we explicitly model additional SiO2 deposition to emulate silica redistribution during phase evolution from kaolin through spinel intermediates to mullite. Machine-learning potentials trained on structurally related sillimanite surfaces accurately describe Al–Si–O–H chemistry and transfer to vacancy-rich and silica-modified mullite phases. Using NH3 binding as a descriptor of Brønsted acidity, we find that mullite can host acid sites with strengths reaching ΔENH3 ≈ –140 kJ mol−1 in 3:2 mullite, which is comparable to external zeolitic acid sites and substantially stronger than those reported on Si-doped γ–Al2O3. Silica grafting onto otherwise weakly acidic mullite terminations further generates an ensemble of strong bridging and pseudo-bridging Brønsted acid sites. We attribute this enhanced acidity to the local emergence of tetrahedrally coordinated AlIV environments that strongly polarize Al–OH–Si linkages. These results identify defect-rich as well as silica-decorated mullite surfaces as a realistic and potentially dominant source of strong matrix acidity in FCC catalysts under operating conditions.

Abstract Supported Au nanoclusters are highly effective catalysts but are susceptible to deactivation via sintering. While Ostwald ripening (OR) is a primary sintering mechanism, its atomic-scale pathways on the anatase TiO2 surface remain poorly understood. Using neural network potential, we investigate the OR process of Au nanoclusters supported on TiO2(101) under oxidizing conditions, identifying Au and AuO as the primary mobile species. A systematic monomer detachment mapping (MDMAP) workflow is developed to efficiently sample the detachment landscape. Subsequent kinetic analysis reveals that low-energy metastable ensembles (LEMEs) serve as essential kinetic intermediates. By evaluating clusters with varying sizes and oxygen coverages, we identify interfacial oxygen as the key regulator of OR kinetics, where moderate oxidation promotes OR through charge stabilization and low Au–Au coordination, while excessive oxidation inhibits monomer detachment. This study provides a comprehensive atomistic perspective on Au nanocluster evolution, offering key insights into the rational design of stable supported catalysts.

Hydrogen dissociation is an essential step in hydrogenation catalysis and critical to catalyst performance. Recently, Au–Pd core-shell nanoparticle catalysts have emerged as a highly promising class of materials for hydrogenation catalysis. Yet, fundamental understanding on the origin of bimetallic synergy in these core-shell catalysts and their interaction with hydrogen is largely unknown. Here, we investigate the kinetics of hydrogen dissociation and recombination over Au–Pd core-shell catalysts using the hydrogen-deuterium (HD) exchange reaction. We employed colloidally prepared core-shell nanorod catalysts and used 4D scanning transmission electron microscopy to assess the degree and directionality of the strain in the Pd-shell. This revealed that even relatively thick Pd-shells are highly strained in the direction parallel to the surface. Catalytic testing showed that the HD exchange rate increases with increasing Pd-shell thickness and drastically decreases upon alloying. Interestingly, the core-shell nanorods showed temperature-dependent apparent activation energies and pre-exponential factors, indicating the presence of hydrides in the Pd-shell at the lower temperatures and gradual hydrogen release upon heating. Density functional theory calculations indicate that subsurface hydrogen plays a major role in accelerating the hydrogen-deuterium exchange reaction by providing a lower energy pathway for recombinative desorption of hydrogen and deuterium, which is the rate limiting step. Altogether, our combined experiment-theory approach revealed that lattice strain and hydride formation can significantly influence the hydrogenation behavior of Au–Pd core-shell nanoparticles and is relevant to further advance the design of core-shell materials for hydrogenation catalysis.

Electrocatalytic reactive carbon capture (RCC), or the direct electroreduction of captured CO2-sorbent adducts, could enable more efficient conversion of dilute CO2 streams to value-added products. Amines are among the best sorbents for industrial carbon capture, but achieving efficient RCC with amines typically leads to low Faradaic efficiency for carbon based products in favor of the hydrogen evolution reaction (HER). HER is prevalent on account of equivalent formation of protonated amines, which promote HER, along with formation of the captured CO2-amine adduct. We employed a joint theory and experimental approach to demonstrate Ni-N-C single atom catalysts as effective catalysts for RCC conversion to CO. Computational analysis revealed that RCC can proceed through direct reduction of the CO2-sorbent adduct or via an indirect C-N bond breaking mechanism with an adsorbed CO2 intermediate. The prevalence of the mechanisms is potential dependent, with the indirect mechanism being predicted to be dominant at lower overpotentials where we experimentally observe CO selectivity. Electrochemical testing showed that intermediate concentrations of diethanolamine sorbent paired with dilute (10-25%) streams of CO2 results in superior selectivity to bicarbonate solutions under the same dilute CO2 conditions. Through in situ X-ray absorption spectroscopy and computational stability analysis, we found that RCC conditions can trigger restructuring and instability of Ni-N-C catalysts at higher overpotentials, leading to deactivation. Taken together, we find that controlling the coordination environment of the catalyst and the solution speciation are key to achieving active and stable RCC. This work was conducted under the auspices of the U.S. Department of Energy by Lawrence Livermore National Laboratory (LLNL) under Contract DE-AC52-07NA27344. LLNL release number: LLNL-ABS-2014462.

In vivo cell engineering enables chimeric antigen receptor (CAR) production directly within the body. Most in vivo CAR approaches target broad T cell populations. However, antigen-presenting nanoparticles could be used to engineer specific T cell subsets, leveraging the phenotypic and functional properties of the cells for T cell therapy.

Inverse ZrO₂/Cu catalysts, where Zr oxide is deposited on Cu particles, show a high catalytic performance converting CO₂ to methanol. We employ density functional theory (DFT) calculations to investigate the CO₂ hydrogenation reaction mechanisms on a model of highly dispersed Zr oxide clusters on Cu (111). The exploration is not performed on a single active site configuration but across an ensemble of 83 formate configurations accessible under reaction conditions. Detailed reaction-pathway analysis reveals that structural sensitivity is pronounced, and only 10 of the catalyst configurations are significantly active across the full pathway. The turnover frequency of the studied inverse structures is largely determined by reaction steps after methoxy formation, rather than the formate hydrogenation steps, and the energy of the methoxy intermediate is a key reactivity descriptor. Two hypotheses are presented for the ensemble average activity: where the probabilities of site populations are determined at the formate intermediate, or at the methoxy resting state. The latter, compared to the former, drastically changes the site distribution, eliminating active structures and decreasing the average rate by a factor of 1000. Catalyst rigidity helps maintain activity by slowing down the structural evolution from the more active formate-bound states to the less active methoxy-bound states.

ABSTRACT Inverse ZrO 2 /Cu catalysts, where Zr oxide is deposited on Cu particles, show a high catalytic performance converting CO 2 to methanol. We employ density functional theory (DFT) calculations to investigate the CO 2 hydrogenation reaction mechanisms on a model of highly dispersed Zr oxide clusters on Cu (111). The exploration is not performed on a single active site configuration but across an ensemble of 83 formate configurations accessible under reaction conditions. Detailed reaction‐pathway analysis reveals that structural sensitivity is pronounced, and only 10 of the catalyst configurations are significantly active across the full pathway. The turnover frequency of the studied inverse structures is largely determined by reaction steps after methoxy formation, rather than the formate hydrogenation steps, and the energy of the methoxy intermediate is a key reactivity descriptor. Two hypotheses are presented for the ensemble average activity: where the probabilities of site populations are determined at the formate intermediate, or at the methoxy resting state. The latter, compared to the former, drastically changes the site distribution, eliminating active structures and decreasing the average rate by a factor of 1000. Catalyst rigidity helps maintain activity by slowing down the structural evolution from the more active formate‐bound states to the less active methoxy‐bound states.

The atomic-scale structure of a metal catalyst surface controls its catalytic performance. Through a combination of high-pressure scanning tunneling microscopy (HP-STM), ambient-pressure X-ray photoelectron spectroscopy (AP-XPS), and machine learning-accelerated computational studies, we uncovered that Pt nanoclusters, formed by restructuring of hex-Pt(100) under reaction conditions involving CO, experience major structural evolution when exposed to increasing CO pressure in the range of 2 × 10^-8-750 Torr. Atom-resolved images and simulations demonstrate that CO binds strongly on these nanoclusters, leading to a CO coverage of one molecule per Pt atom, while the lateral CO-CO repulsion is released by tilting CO molecules outward at the nanocluster edge. Metal nanoclusters break down along as CO pressure is increased from 2 × 10^-8 to 1 Torr with the average size decreasing from 3.2 ± 1.5 to 2.3 ± 1.0 nm, consistent with Pt 4f_7/2 photoemission feature evolution observed with AP-XPS. In contrast, in the pressure range of 1-750 Torr at 25 °C, HP-STM observed a decrease of nanocluster density by 4-5 times, consistent with the growth of the average nanocluster size from 2.3 ± 1.0 nm in 1 Torr CO to 4.6 ± 1.8 nm in 750 Torr. This uncovers a reactant pressure-driven coalescence of nanoclusters even at room temperature. Nanoclusters formed at 25 °C in 750 Torr CO require annealing to 100-130 °C to reach equilibrium size, indicating akinetic control of nanocluster growth at low preparation temperatures, such as room temperature. Our neural network potential (NNP) coupled with basin-hopping (BH) simulations determined the optimal CO coverage and configuration at various CO pressures and showed, in agreement with experiments, an optimum nanocluster size resulting from a competition between the size-dependent energy cost of nanocluster formation and the energy gain through CO adsorption. The formation of Pt nanoclusters, followed by their breakdown with increasing CO pressure from 2 × 10^-8 to 1 Torr and coalescence in CO pressure from 1 to 750 Torr highlights the significance of imaging catalyst nanoparticle surfaces in gas phase at a specific reactant pressure toward establishing a direct structure-catalytic performance correlation.

Accurate adsorption thermodynamics of reactive intermediates at electrified metal interfaces are central to predictive surface science, heterogeneous catalysis, and electrochemical modeling, yet remain challenging for standard density functional approximations and difficult to benchmark experimentally. Here we apply electrochemical-aware random phase approximation (RPA) methods that incorporate solvent dielectric screening and grand-canonical constant-potential control to establish a reference-quality description of hydroxyl (*OH) adsorption on Cu(100) in electrocatalytic conditions. At the potential of zero charge (pzc), we quantify the site dependence of *OH binding across terrace sites and representative defective motifs. Under applied potential, electrochemical-aware RPA reproduces the experimentally inferred *OH desorption fingerprint, predicting an *OH desorption potential of -0.54 V vs SHE, in excellent agreement with experimental value -0.56 V vs SHE at pH = 7, whereas widely used GGA functionals (PBE, RPBE) underestimate *OH stability and compress the stability window under reducing conditions. G₀W₀-RPA electronic structure analysis provides a mechanistic rationale for the discrepancy. Finally, we reassess the Cu(100) surface-state diagram with *OH and *CO coadsorption under electrochemical conditions, including the stability of metastable CuCO(OH)_n motifs, and show that many-body accuracy can qualitatively alter predicted interfacial speciation. Overall, our results establish electrochemical-aware RPA as a broadly applicable, systematically improvable many-body framework for quantitative adsorption thermodynamics at constant potential, enabling predictive surface-state maps for complex electrochemical interfaces.