QM-Mechanism-Based Hierarchical High-Throughput in Silico Screening Catalyst Design for Ammonia Synthesis

Theoretical understanding and prediction of metal-doped CeO2 catalysts for ammonia dissociation

Ammonia plays a critical role in energy and environmental catalysis, particularly in ammonia dissociation reactions. Understanding the adsorption and dissociation of ammonia-related species on catalysts is essential for the development of new chemical reactions and high-performance catalysts. However, establishing the relationship between catalyst properties and the adsorption of dissociated species remains challenging, particularly for metal oxide catalysts. This study employs density functional theory calculations to investigate the adsorption properties of ammonia and dissociated intermediate species on metal-doped CeO2. Through a feature correlation heat map, certain descriptors, such as single atom formation energy, gaseous atom formation heat, valence band maximum, and work function, were determined to exhibit a strong linear relationship with the adsorption properties of NHx species. As deduced from the density of states properties and orbital theory, it is also found that the energy difference between the lowest unoccupied orbital of the metal and the highest occupied orbital of ammonia, has a good relationship with the adsorption energy of NH3.

Machine-Learning-Accelerated DFT Conformal Sampling of Catalytic Processes

Computational modeling of catalytic processes at gas/solid interfaces plays an increasingly important role in chemistry, enabling accelerated materials and process optimization and rational design. However, efficiency, accuracy, thoroughness, and throughput must be enhanced to maximize its practical impact. By combining interpolation of DFT energetics via highly accurate Machine-Learning Potentials with conformal techniques for building the training database, we present here an original approach (that we name Conformal Sampling of Catalytic Processes, CSCP), to accelerate and achieve an accurate and thorough sampling of novel systems by exporting existing information on a worked-out case. We use methanol decomposition (of interest in the field of hydrogen production and storage) as a test catalytic reaction. Starting from worked-out Pt-based systems, we show that after only two iterations of active-learning CSCP is able to provide reaction energy diagrams for a set of 7 diverse systems (Pd, Ni, Au, Ag, Cu, Co, Fe) leading to DFT-accuracy-level predictions. Cases exhibiting a change in adsorption sites and mechanisms are also successfully reproduced as tests of catalytic path modification. The CSCP approach thus offers itself as an operative tool to fully take advantage of accumulated information to achieve high-throughput sampling of catalytic processes.

Enabling high throughput deep reinforcement learning with first principles to investigate catalytic reaction mechanisms

Exploring catalytic reaction mechanisms is crucial for understanding chemical processes, optimizing reaction conditions, and developing more effective catalysts. We present a reaction-agnostic framework based on high-throughput deep reinforcement learning with first principles (HDRL-FP) that offers excellent generalizability for investigating catalytic reactions. HDRL-FP introduces a generalizable reinforcement learning representation of catalytic reactions constructed solely from atomic positions, which are subsequently mapped to first-principles-derived potential energy landscapes. By leveraging thousands of simultaneous simulations on a single GPU, HDRL-FP enables rapid convergence to the optimal reaction path at a low cost. Its effectiveness is demonstrated through the studies of hydrogen and nitrogen migration in Haber-Bosch ammonia synthesis on the Fe(111) surface. Our findings reveal that the Langmuir-Hinshelwood mechanism shares the same transition state as the Eley-Rideal mechanism for H migration to NH2, forming ammonia. Furthermore, the reaction path identified herein exhibits a lower energy barrier compared to that through nudged elastic band calculation.

Mutexa: A Computational Ecosystem for Intelligent Protein Engineering

Protein engineering holds immense promise in shaping the future of biomedicine and biotechnology. This Review focuses on our ongoing development of Mutexa, a computational ecosystem designed to enable "intelligent protein engineering". In this vision, researchers will seamlessly acquire sequences of protein variants with desired functions as biocatalysts, therapeutic peptides, and diagnostic proteins through a finely-tuned computational machine, akin to Amazon Alexa's role as a versatile virtual assistant. The technical foundation of Mutexa has been established through the development of a database that combines and relates enzyme structures and their respective functions (e.g., IntEnzyDB), workflow software packages that enable high-throughput protein modeling (e.g., EnzyHTP and LassoHTP), and scoring functions that map the sequence-structure-function relationship of proteins (e.g., EnzyKR and DeepLasso). We will showcase the applications of these tools in benchmarking the convergence conditions of enzyme functional descriptors across mutants, investigating protein electrostatics and cavity distributions in SAM-dependent methyltransferases, and understanding the role of nonelectrostatic dynamic effects in enzyme catalysis. Finally, we will conclude by addressing the future steps and fundamental challenges in our endeavor to develop new Mutexa applications that assist the identification of beneficial mutants in protein engineering.

Big Data in a Nano World: A Review on Computational, Data-Driven Design of Nanomaterials Structures, Properties, and Synthesis

The recent rise of computational, data-driven research has significant potential to accelerate materials discovery. Automated workflows and materials databases are being rapidly developed, contributing to high-throughput data of bulk materials that are growing in quantity and complexity, allowing for correlation between structural-chemical features and functional properties. In contrast, computational data-driven approaches are still relatively rare for nanomaterials discovery due to the rapid scaling of computational cost for finite systems. However, the distinct behaviors at the nanoscale as compared to the parent bulk materials and the vast tunability space with respect to dimensionality and morphology motivate the development of data sets for nanometric materials. In this review, we discuss the recent progress in data-driven research in two aspects: functional materials design and guided synthesis, including commonly used metrics and approaches for designing materials properties and predicting synthesis routes. More importantly, we discuss the distinct behaviors of materials as a result of nanosizing and the implications for data-driven research. Finally, we share our perspectives on future directions for extending the current data-driven research into the nano realm.

Biomacromolecule-Assisted Screening for Reaction Discovery and Catalyst Optimization

Reaction discovery and catalyst screening lie at the heart of synthetic organic chemistry. While there are efforts at de novo catalyst design using computation/artificial intelligence, at its core, synthetic chemistry is an experimental science. This review overviews biomacromolecule-assisted screening methods and the follow-on elaboration of chemistry so discovered. All three types of biomacromolecules discussed─enzymes, antibodies, and nucleic acids─have been used as "sensors" to provide a readout on product chirality exploiting their native chirality. Enzymatic sensing methods yield both UV-spectrophotometric and visible, colorimetric readouts. Antibody sensors provide direct fluorescent readout upon analyte binding in some cases or provide for cat-ELISA (Enzyme-Linked ImmunoSorbent Assay)-type readouts. DNA biomacromolecule-assisted screening allows for templation to facilitate reaction discovery, driving bimolecular reactions into a pseudo-unimolecular format. In addition, the ability to use DNA-encoded libraries permits the barcoding of reactants. All three types of biomacromolecule-based screens afford high sensitivity and selectivity. Among the chemical transformations discovered by enzymatic screening methods are the first Ni(0)-mediated asymmetric allylic amination and a new thiocyanopalladation/carbocyclization transformation in which both C-SCN and C-C bonds are fashioned sequentially. Cat-ELISA screening has identified new classes of sydnone-alkyne cycloadditions, and DNA-encoded screening has been exploited to uncover interesting oxidative Pd-mediated amido-alkyne/alkene coupling reactions.

Reaction Mechanisms, Kinetics, and Improved Catalysts for Ammonia Synthesis from Hierarchical High Throughput Catalyst Design

The Haber-Bosch (HB) process is the primary chemical synthesis technique for industrial production of ammonia (NH₃) for manufacturing nitrate-based fertilizer and as a potential hydrogen carrier. The HB process alone is responsible for over 2% of all global energy usage to produce more than 160 million tons of NH₃ annually. Iron catalysts are utilized to accelerate the reaction, but high temperatures and pressures of atmospheric nitrogen gas (N₂) and hydrogen gas (H₂) are required. A great deal of research has aimed at increased performance over the last century, but the rate of progress has been slow. This Account focuses on determining the atomic-level reaction mechanism for HB synthesis of NH₃ on the Fe catalysts used in industry and how to use this knowledge to suggest greatly improved catalysts via a novel paradigm of catalyst rational design.We determined the full reaction mechanism on the two most active surfaces for the HB process, Fe(111) and Fe(211)R. We used density functional theory (DFT) to predict the free-energy barriers for all 12 important reactions and the 34 most important 2 × 2 surface configurations. Then we incorporated the mechanism into kinetic Monte Carlo (kMC) simulations run for several hours of real time to predict turnover frequencies (TOFs). The predicted TOFs are within experimental error, indicating that the predicted barriers are within 0.04 eV of experiment.With this level of accuracy, we are poised to use DFT to improve the catalyst. Rather than forming bulk alloys with uniform concentration, we aimed at finding additives that strongly prefer near-surface sites so that minor amounts of the additive might lead to dramatic improvements. However, even for a single additive, the combinations of surface species and reactions multiplies significantly, with ∼48 reaction steps to examine and nearly 100 surface configurations per 2 × 2 site. To make it practical to examine tens of dopant candidates, we developed the hierarchical high-throughput catalysis screening (HHTCS) approach, which we applied to both the Fe(111) and Fe(211) surfaces. For HHTCS, we identified the most important 4 reaction steps out of 12 for the two surfaces to examine >50 dopant cases, where we required performance at each step no worse than for pure Fe. With HHTCS, the computational cost is about 1% of that for doing the full reaction mechanism, allowing us to do ≈50 cases in about 1/2 the time it took to do pure Fe(111). The new leads identified with HHTCS are then validated with full mechanistic studies.For Fe(111), we predict three high-performance dopants that strongly prefer the second layer: Co with a rate 8 times higher, Ni with a rate 16 times higher, and Si with a rate 43 times higher, at 400 °C and 20 atm. We also found four dopants that strongly prefer the top layer and improve performance: Pt or Rh 3 times faster and Pd or Cu 2 times faster. For Fe(211), the best dopant was found to be second-layer Co with a rate 3 times faster than that for the undoped surface.The DFT/kMC data were used to predict reshaping of the catalyst particles under reaction conditions and how to tune dopant content so as to maximize catalytic area and thus activity. Finally, we show how to validate our mechanistic modeling via a comparison between theoretical and experimental operando spectroscopic signatures.

Discovering Catalytic Reaction Networks Using Deep Reinforcement Learning from First-Principles

Determining the reaction pathways, which is central to illustrating the working mechanisms of a catalyst, is severely hindered by the high complexity of the reaction and the extreme scarcity of the data. Here, we develop a novel artificial intelligence framework integrating deep reinforcement learning (DRL) techniques with density functional theory simulations to automate the quantitative search and evaluation on the complex catalytic reaction networks from zero knowledge. Our framework quantitatively transforms the first-principles-derived free energy landscape of the chemical reactions to a DRL environment and the corresponding actions. By interacting with this dynamic environment, our model evolves by itself from scratch to a complete reaction path. We demonstrate this framework using the Haber-Bosch process on the most active Fe(111) surface. The new path found by our framework has a lower overall free energy barrier than the previous study based on domain knowledge, demonstrating its outstanding capability in discovering complicated reaction paths. Looking forward, we anticipate that this framework will open the door to exploring the fundamental reaction mechanisms of many catalytic reactions.

Mechanistic aspects of ammonia synthesis on Ta3N5 surfaces in the presence of intrinsic nitrogen vacancies

A possible dinitrogen activation and the ammonia synthesis mechanism were studied on the (100), (010) and (001) surfaces of Ta₃N₅ that contain intrinsic nitrogen vacancies. The study suggests that intrinsic nitrogen vacancies can become catalytic centers for the ammonia synthesis reaction on Ta₃N₅via a Langmuir-Hinshelwood mechanism, which may explain the moderate production of ammonia at high temperatures. In the proposed mechanism, dinitrogen is activated in a peculiar side on a sandwich-like configuration between two surface Ta atoms. Calculation of reaction activation barriers suggests that the mechanism proceeds via moderate barriers but some elementary reaction steps involve the strong adsorption of ammonia which appears to poison the surface catalytic sites on Ta₃N₅.

Reduction of N2 to Ammonia by Phosphate Molten Salt and Li Electrode: Proof of Concept Using Quantum Mechanics

Electrochemical routes provide an attractive alternative to the Haber-Bosch process for cheaper and more efficient ammonia (NH₃) synthesis from N₂ while avoiding the onerous environmental impact of the Haber-Bosch process. We prototype a strategy based on a eutectic mixture of phosphate molten salt. Using quantum-mechanics (QM)-based reactive molecular dynamics, we demonstrate that lithium nitride (Li₃N) produced from the reduction of nitrogen gas (N₂) by a lithium electrode can react with the phosphate molten salt to form ammonia. We extract reaction kinetics of the various steps from QM to identify conditions with favorable reaction rates for N₂ reduction by a porous lithium electrode to form Li₃N followed by protonation from phosphate molten salt (Li₂HPO₄-LiH₂PO₄ mixture) to selectively form NH₃.

Controlling the Shapes of Nanoparticles by Dopant-Induced Enhancement of Chemisorption and Catalytic Activity: Application to Fe-Based Ammonia Synthesis

We showed recently that the catalytic efficiency of ammonia synthesis on Fe-based nanoparticles (NP) for Haber-Bosch (HB) reduction of N₂ to ammonia depends very dramatically on the crystal surface exposed and on the doping. In turn, the stability of each surface depends on the stable intermediates present during the catalysis. Thus, under reaction conditions, the shape of the NP is expected to evolve to optimize surface energies. In this paper, we propose to manipulate the shape of the nanoparticles through doping combined with chemisorption and catalysis. To do this, we consider the relationships between the catalyst composition (adding dopant elements) and on how the distribution of the dopant atoms on the bulk and facet sites affects the shape of the particles and therefore the number of active sites on the catalyst surfaces. We use our hierarchical, high-throughput catalyst screening (HHTCS) approach but extend the scope of HHTCS to select dopants that can increase the catalytically active surface orientations, such as Fe-bcc(111), at the expense of catalytically inactive facets, such as Fe-bcc(100). Then, for the most promising dopants, we predict the resulting shape and activity of doped Fe-based nanoparticles under reaction conditions. We examined 34 possible dopants across the periodic table and found 16 dopants that can potentially increase the fraction of active Fe-bcc(111) vs inactive Fe-bcc(100) facets. Combining this reshaping criterion with our HHTCS estimate of the resulting catalytic performance, we show that Si and Ni are the most promising elements for improving the rates of catalysis by optimizing the shape to decrease reaction barriers. Then, using Si dopant as a working example, we build a steady-state dynamical Wulff construction of Si-doped Fe bcc nanoparticles. We use nanoparticles with a diameter of ∼10 nm, typical of industrial catalysts. We predict that doping Si into such Fe nanoparticles at the optimal atomic content of ∼0.3% leads to rate enhancements by a factor of 56 per nanoparticle under target HB conditions.

Si-Doped Fe Catalyst for Ammonia Synthesis at Dramatically Decreased Pressures and Temperatures

The Haber-Bosch (HB) process combining nitrogen (N₂) and hydrogen (H₂) into ammonia (NH₃) gas plays an essential role in the synthesis of fertilizers for food production and many other commodities. However, HB requires enormous energy resources (2% of world energy production), and the high pressures and temperatures make NH₃ production facilities very expensive. Recent advances in improving HB catalysts have been incremental and slow. To accelerate the development of improved HB catalysts, we developed a hierarchical high-throughput catalyst screening (HHTCS) approach based on the recently developed complete reaction mechanism to identify non-transition-metal (NTM) elements from a total set of 18 candidates that can significantly improve the efficiency of the most active Fe surface, Fe-bcc(111), through surface and subsurface doping. Surprisingly, we found a very promising subsurface dopant, Si, that had not been identified or suggested previously, showing the importance of the subsurface Fe atoms in N₂ reduction reactions. Then we derived the full reaction path of the HB process for the Si doped Fe-bcc(111) from QM simulations, which we combined with kinetic Monte Carlo (kMC) simulations to predict a ∼13-fold increase in turnover frequency (TOF) under typical extreme HB conditions (200 atm reactant pressure and 500 °C) and a ∼43-fold increase in TOF under ideal HB conditions (20 atm reactant pressure and 400 °C) for the Si-doped Fe catalyst, in comparison to pure Fe catalyst. Importantly, the Si-doped Fe catalyst can achieve the same TOF of pure Fe at 200 atm/500 °C under much milder conditions, e.g. at a much decreased reactant pressure of 20 atm at 500 °C, or alternatively at temperature and reactant pressure decreased to 400 °C and 60 atm, respectively. Production plants using the new catalysts that operate under such milder conditions could be much less expensive, allowing production at local sites needing fertilizer.

Reaction mechanism and kinetics for ammonia synthesis on the Fe(211) reconstructed surface

To provide guidelines to accelerate the Haber-Bosch (HB) process for synthesis of ammonia from hydrogen and nitrogen, we used Quantum Mechanics (QM) to determine the reaction mechanism and free energy reaction barriers under experimental reaction conditions (400 °C and 20 atm) for all 10 important surface reactions on the Fe(211) reconstructed (Fe(211)R) surface. These conditions were then used in full kMC modeling for 30 minutes to attain steady state. We find that the stable surface under Haber-Bosch conditions is the missing row 2 × 1 reconstructed surface (211)R and that the Turn Over Frequency (TOF) is 18.7 s^-1 per 2 × 2 surface site for 1.5 Torr NH₃ pressure, but changes to 3.5 s^-1 for 1 atm, values close (within 6%) to the ones on Fe(111). The experimental ratio between (211) and (111) rates at low (undisclosed) NH₃ pressure was reported to be 0.75. The excellent agreement with experiment on two very different surfaces and reaction mechanisms is a testament of the accuracy of QM modeling. In addition, our kinetic analysis indicates that Fe(211)R is more active than Fe(111) at high pressure, close to HB industrial conditions, and that (211)R is more abundant than (111) via a steady-state Wulff construction under HB conditions. Thus, at variance with common thinking, we advocate the Fe(211)R surface as the catalytically active phase of pure iron ammonia synthesis catalyst under HB industrial conditions.

2D oxides on metal materials: concepts, status, and perspectives

Two-dimensional oxide-on-metal materials: concepts, methods, and link to technological applications, with 5 subtopics: structural motifs, robustness, catalysis, ternaries, and nanopatterning.