Search

Heterogeneous nanocatalysts play a crucial role in both the chemical and energy industries. Despite substantial advancements in theoretical, computational and experimental studies, identifying their active sites remains a major challenge. Here we utilize atomic electron tomography to determine the three-dimensional atomic structure of PtNi and Mo-doped PtNi nanocatalysts for the electrochemical oxygen reduction reaction. We then employ the experimental atomic structures as input to first-principles-trained machine learning to identify the active sites of the nanocatalysts. Through the analysis of the structure–activity relationships, we formulate an equation termed the local environment descriptor, which balances the strain and ligand effects to provide physical and chemical insights into active sites in the oxygen reduction reaction. The ability to determine the three-dimensional atomic structure and chemical composition of realistic nanoparticles, combined with machine learning, could transform our fundamental understanding of the active sites of catalysts and guide the rational design of optimal nanocatalysts. (Figure presented.)

Abstract: Heterogeneous catalysts play a key role in the chemical and energy industries1. To date, most industrial-scale heterogeneous catalytic reactions have relied on nanocatalysts2,3. However, despite significant progress from theoretical, experimental and computational studies4-18, identifying the active sites of alloy nanocatalysts remains a major challenge. This limitation is mainly due to an incomplete understanding of the three-dimensional (3D) atomic and chemical arrangement of different constituents and structural reconstructions driven by catalytic reactions19-22. Here, we use atomic electron tomography23 to determine the 3D local atomic structure, surface morphology and chemical composition of 11 Pt alloy nanocatalysts for the electrochemical oxygen reduction reaction (ORR). We reveal the facet, surface concaveness, structural and chemical order/disorder, coordination number, and bond length with unprecedented 3D atomic detail. The experimental 3D atomic coordinates are used by first-principles trained machine learning to identify the active sites of the nanocatalysts, which are corroborated by electrochemical measurements. A striking feature is the difference of the ORR activity of the surface Pt sites on the nanocatalysts by several orders of magnitude. Furthermore, by analyzing the structure-activity relationship, we formulate an equation named the local environment descriptor to balance the strain and ligand effects and gain quantitative insights into the ORR active sites of the Pt alloy nanocatalysts. The ability to determine the 3D atomic structure and chemical composition of realistic nanoparticles coupled with machine learning could transform our fundamental understanding of the catalytic active sites and provide a guidance for the rational design of optimal nanocatalysts.

The rationale of the catalytic activity observed in experiments is a crucial task in fundamental catalysis studies. Efficient catalyst design relies on an accurate understanding of the origin of the activity at the atomic level. Theoretical studies have been widely developed to reach such a fundamental atomic scale understanding of catalytic activity. Current theories ascribe the catalytic activity to the geometric and electronic structure of the active site, in which the geometrical and electronic structure effects are derived from the equilibrium geometry of active sites characterizing the static property of the catalyst; however catalysts, especially in the form of nanoclusters, may present fluxional and dynamic structure under reaction conditions, and the effect of this fluxional behavior is not yet widely recognized. Therefore, this Account will focus on the fluxionality of the active sites, which is driven by thermal fluctuations under finite temperature.Under reaction conditions, nanocluster catalysts can readily restructure, either being promoted to another metastable isomer (named as plastic fluxionality) or presenting ample deformations around their equilibrium geometry (named as elastic fluxionality). This Account summarizes our recent studies on the fluxionality of the nanoclusters and how plastic and elastic fluxionalities play roles in highly efficient reaction pathways. Our results show that the low energy metastable isomers formed by plastic fluxionality can manifest high reactivity despite their minor occurrence probability in the mixture of catalyst isomers. In the end, the highly active metastable isomer may dominate the total observed reactivity. In addition, the isomerization between the global minimum structure and the highly active metastable isomer can be a central step in catalytic transformations in order to circumvent some difficult reaction steps and may govern the overall mechanism. In addition, the thermal fluctuation driven elastic fluxionality is also found to play a key role, complementary to plastic fluxionality. The elastic fluxionality creates substantial structural deformations of the active site, and these deformed geometries enable low activation energies and high catalytic activity, which cannot be found from the static equilibrium geometry of the catalyst. A dedicated global activity search algorithm is proposed to search for the optimal reaction pathway on fluxional nanoclusters. In summary, our studies demonstrate that thermal-driven fluxionality provides a different paradigm for understanding the high activity of nanoclusters under reaction conditions beyond the static description of geometric and electronic structure. We first summarize our previous results and then provide a perspective for further studies on how to investigate and take the advantage of the fluxional geometry of nanoclusters. We will defend in this Account that the static picture for the active site is not complete and might miss critical reaction pathways that are highly efficient and only open after thermally induced restructuring of the active site.

Single-atom catalysts (SACs) have become a prominent theme in heterogeneous catalysis, not least because of the potential fundamental insight into active sites. The desired level of understanding, however, is prohibited due to the inhomogeneity of most supported SACs and the lack of suitable tools for structure-activity correlation studies with atomic resolution. Herein, we describe the potency of electrospray ionization mass spectrometry (ESI-MS) to study molecularly defined SACs supported on polyoxometalates in catalytic reactions. We identified the exact composition of active sites and their evolution in the catalytic cycle during CO and alcohol oxidation reactions performed in the liquid phase. Critical information on metal-dependent reaction mechanisms, the key intermediates, the dynamics of active sites and even the stepwise activation barriers were obtained, which would be challenging to gather via prevailingly adopted techniques in SAC research. DFT calculations revealed intricate details of the reaction mechanisms, and strong synergies between ESI-MS defined SAC sites and electronic structure theory calculations become apparent.

Dehydrogenation reactions are central for the production of functional molecules. Steam plays a pivotal role in ZnFe2O4-catalyzed 1-butene oxidative dehydrogenation (ODH). However, the essential effect of steam on this reaction is still unclear. Herein, we describe the structure-performance relationships of ZnFe2O4 in the presence/absence of steam by combined density functional theory and experimental studies. The catalytic performances of ZnFe2O4 under different reaction conditions were investigated. The ZnFe2O4(110) surface properties under reaction conditions and molecular reaction pathways were modeled. Free-energy profiles were calculated. We found that an oxygen-excess ZnFe2O4(110)-O termination, with an extra O atom bridging two Fe cations, is preferred under oxygen-rich conditions. However, both experimental and theoretical approaches indicate that this surface bicoordinated O is not stable at relatively high temperatures in the absence of steam, resulting in the reduction of surface Fe3+ cations to Fe2+ which are inactive in the 1-butene ODH reaction. It was found that steam interacts strongly with the ZnFe2O4(110)-O surface. Steam stabilizes the catalyst surface Fe3+ ions by converting bicoordinated O to more thermally stable hydroxyl groups. Surface hydroxyls are active sites for C-H bond cleavage in the 1-butene ODH reaction in the presence of steam. We propose that the role of steam elucidated here represents a general mode of steam influence in ODH reactions over oxide surfaces.

The sluggish electrochemical kinetics of sulfur species has impeded the wide adoption of lithium-sulfur battery, which is one of the most promising candidates for next-generation energy storage system. Here, we present the electronic and geometric structures of all possible sulfur species and construct an electronic energy diagram to unveil their reaction pathways in batteries, as well as the molecular origin of their sluggish kinetics. By decoupling the contradictory requirements of accelerating charging and discharging processes, we select two pseudocapacitive oxides as electron-ion source and drain to enable the efficient transport of electron/Li+ to and from sulfur intermediates respectively. After incorporating dual oxides, the electrochemical kinetics of sulfur cathode is significantly accelerated. This strategy, which couples a fast-electrochemical reaction with a spontaneous chemical reaction to bypass a slow-electrochemical reaction pathway, offers a solution to accelerate an electrochemical reaction, providing new perspectives for the development of high-energy battery systems.

Alumina supported Pt nanoclusters under a hydrogen environment play a crucial role in many heterogeneous catalysis applications. We conducted grand canonical genetic algorithm simulations for supported Pt8 clusters in a hydrogen gas environment to study the intracluster, cluster-support, and cluster-adsorbate interactions. Two alumina surfaces, α-Al2O3(0001) and γ-Al2O3(100), and two conditions, T = 600 °C, pH2 = 0.1 bar and T = 25 °C, pH2 = 1.0 bar, were considered corresponding to low and high hydrogen chemical potential μH, respectively. The low free energy ensemble of Pt8 is decorated by a medium (2-12 H), respectively, high (20-30 H), number of hydrogen atoms under equilibrium at low μH, respectively, high μH, and undergoes different morphological transformations on the two surfaces. On α-Al2O3(0001), Pt8 is mostly 3D but very fluxional in structure at low μH and converts to open one-layer 2D structures with minimal fluxionality at high μH, whereas on γ-Al2O3(100), the exact opposite occurs: Pt8 clusters present one-layer 2D shapes at low μH and switch to compact 3D shapes under high μH, during which the Pt8 cluster preserves moderate fluxionality. Further analysis reveals a similar Pt-Pt bond length increase when switching from low μH to high μH on both surfaces although morphological transformations are different. Electronic structure analysis shows the existence of bonding interactions between Pt and Lewis acidic Al3+ sites along with the Pt-O interaction, which implies the necessity to include Al neighbors to discuss the electronic structure of small Pt clusters.

Data-driven machine learning force fields (MLFs) are more and more popular in atomistic simulations and exploit machine learning methods to predict energies and forces for unknown structures based on the knowledge learned from an existing reference database. The latter usually comes from density functional theory calculations. One main drawback of MLFs is that physical laws are not incorporated in the machine learning models, and instead, MLFs are designed to be very flexible to simulate complex quantum chemistry potential energy surface (PES). In general, MLFs have poor transferability, and hence, a very large trainset is required to span all the target feature space to get a reliable MLF. This procedure becomes more troublesome when the PES is complicated, with a large number of degrees of freedom, in which building a large database is inevitable and very expensive, especially when accurate but costly exchange-correlation functionals have to be used. In this manuscript, we exploit a high-dimensional neural network potential (HDNNP) on Pt clusters of sizes from 6 to 20 as one example. Our standard level of energy calculation is DFT GGA (PBE) using a plane wave basis set. We introduce an approximate but fast level with the PBE functional and a minimal atomic orbital basis set, and then, a more accurate but expensive level, using a hybrid functional or nonlocal vdW functional and a plane wave basis set, is reliably predicted by learning the difference with HDNNP. The results show that such a differential approach (named ΔHDNNP) can deliver very accurate predictions (error

Unlike nanostructured metal catalysts, supported single-atom catalysts (SACs) contain only atomically dispersed metal atoms, hinting at much more pronounced metal-support effects. Herein, we take a series of polyoxometalate-supported Pt catalysts as examples to quantitatively investigate the stability of Pt atoms on oxide supports and how the Pt-support interaction influences the catalytic performance. For this entire series, we show that the Pt atoms prefer to stay at a 4-fold hollow site of one polyoxometalate molecule and that the least adsorption energy to obtain sintering-resistant Pt SACs is 5.50 eV, which exactly matches the cohesive energy of bulk Pt metal. Further, we compared their catalytic performance in several hydrogenation reactions and simulated the reaction pathways of propene hydrogenation by density functional theory (DFT) calculations. Both experimental and theoretical approaches suggest that despite the Pt1-support interactions being different, the reaction pathways of various Pt1-polyoxometalate catalysts are very similar and their effective reaction barriers are close to each other and as low as 24 kJ/mol, indicating the possibility of obtaining SACs with improved stability without compromising activity. DFT calculations show that all reaction elementary steps take place only on the Pt atom without involving neighboring O atoms and that hydrogenation proceeds from the molecularly adsorbed H2 species. Pt SACs give a weaker H2 adsorption energy than Pt clusters or surfaces, resulting in small adsorption equilibrium constants and small apparent activation barriers, which agree between experiment and theory.

Reactivity studies on catalytic transition metal clusters are usually performed on a single global minimum structure. With the example of a Pt13 cluster under a pressure of hydrogen, we show from first-principle calculations that low energy metastable structures of the cluster can play a major role for catalytic reactivity and that hence consideration of the global minimum structure alone can severely underestimate the activity. The catalyst is fluxional with an ensemble of metastable structures energetically accessible at reaction conditions. A modified genetic algorithm is proposed to comprehensively search for the low energy metastable ensemble (LEME) structures instead of merely the global minimum structure. In order to reduce the computational cost of density functional calculations, a high dimensional neural network potential is employed to accelerate the exploration. The presence and influence of LEME structures during catalysis is discussed by the example of H covered Pt13 clusters for two reactions of major importance: hydrogen evolution reaction and methane activation. The results demonstrate that although the number of accessible metastable structures is reduced under reaction condition for Pt13 clusters, these metastable structures can exhibit high activity and dominate the observed activity due to their unique electronic or structural properties. This underlines the necessity of thoroughly exploring the LEME structures in catalysis simulations. The approach enables one to systematically address the impact of isomers in catalysis studies, taking into account the high adsorbate coverage induced by reaction conditions.

Many forest ecosystems have experienced recent declines in productivity; however, in some alpine regions, tree growth and forest expansion are increasing at marked rates. Dendrochronological analyses at the upper limit of alpine forests in the Tibetan Plateau show a steady increase in tree growth since the early 1900s, which intensified during the 1930s and 1960s, and have reached unprecedented levels since 1760. This recent growth acceleration was observed in small/young and large/old trees and coincided with the establishment of trees outside the forest range, reflecting a connection between the physiological performance of dominant species and shifts in forest distribution. Measurements of stable isotopes (carbon, oxygen, and nitrogen) in tree rings indicate that tree growth has been stimulated by the synergistic effect of rising atmospheric CO2 and a warming-induced increase in water and nutrient availability from thawing permafrost. These findings illustrate the importance of considering soil-plant-atmosphere interactions to understand current and anticipate future changes in productivity and distribution of forest ecosystems.