Predicting Segregation Energy in Single Atom Alloys Using Physics and Machine Learning.

Single atom alloys (SAAs) show great promise as catalysts for a wide variety of reactions due to their tunable properties, which can enhance the catalytic activity and selectivity. To design SAAs, it is imperative for the heterometal dopant to be stable on the surface as an active catalytic site. One main approach to probe SAA stability is to calculate surface segregation energy. Density functional theory (DFT) can be applied to investigate the surface segregation energy in SAAs. However, DFT is computationally expensive and time-consuming; hence, there is a need for accelerated frameworks to screen metal segregation for new SAA catalysts across combinations of metal hosts and dopants. To this end, we developed a model that predicts surface segregation energy using machine learning for a series of SAA periodic slabs. The model leverages elemental descriptors and features inspired by the previously developed bond-centric model. The initial model accurately captures surface segregation energy across a diverse series of FCC-based SAAs with various surface facets and metal-host pairs. Following our machine learning methodology, we expanded our analysis to develop a new model for SAAs formed from FCC hosts with FCC, BCC, and HCP dopants. Our final, five-feature model utilizes second-order polynomial kernel ridge regression. The model is able to predict segregation energies with a high degree of accuracy, which is due to its physically motivated features. We then expanded our data set to test the accuracy of the five features used. We find that the retrained model can accurately capture E seg trends across different metal hosts and facets, confirming the significance of the features used in our final model. Finally, we apply our pretrained model to a series of Ir- and Pd-based SAA cuboctahedron nanoparticles (NPs), ranging in size and FCC dopants. Remarkably, our model (trained on periodic slabs) accurately predicts the DFT segregation energies of the SAA NPs. The results provide further evidence supporting the use of our model as a general tool for the rapid prediction of SAA segregation energies. By creating a framework to predict the metal segregation from bulk surfaces to NPs, we can accelerate the SAA catalyst design while simultaneously unraveling key physicochemical properties driving thermodynamic stabilization of SAAs.

Correlating structural rules with electronic properties of ligand-protected alloy nanoclusters.

Thiolate protected gold nanoclusters (TPNCs) are a unique class of nanomaterials finding applications in various fields, such as biomedicine, optics, and catalysis. The atomic precision of their structure, characterized through single crystal x-ray diffraction, enables the accurate investigation of their physicochemical properties through electronic structure calculations. Recent experimental efforts have led to the successful heterometal doping of TPNCs, potentially unlocking a large domain of bimetallic TPNCs for targeted applications. However, how TPNC size, bimetallic composition, and location of dopants influence electronic structure is unknown. To this end, we introduce novel structure-property relationships (SPRs) that predict electronic properties such as ionization potential (IP) and electron affinity (EA) of AgAu TPNCs based on physically relevant descriptors. The models are constructed by first generating a hypothetical AgAu TPNC dataset of 368 structures with sizes varying from 36 to 279 metal atoms. Using our dataset calculated with density functional theory (DFT), we employed systematic analyses to unravel size, composition, and, importantly, core-shell effects on TPNC EA and IP behavior. We develop generalized SPRs that are able to predict electronic properties across the AgAu TPNC materials space. The models leverage the same three fundamental descriptors (i.e., size, composition, and core-shell makeup) that do not require DFT calculations and rely only on simple atom counting, opening avenues for high throughput bimetallic TPNC screening for targeted applications. This work is a first step toward finely controlling TPNC electronic properties through heterometal doping using high throughput computational means.

Predicting ligand removal energetics in thiolate-protected nanoclusters from molecular complexes.

Thiolate-protected metal nanoclusters (TPNCs) have attracted great interest in the last few decades due to their high stability, atomically precise structure, and compelling physicochemical properties. Among their various applications, TPNCs exhibit excellent catalytic activity for numerous reactions; however, recent work revealed that these systems must undergo partial ligand removal in order to generate active sites. Despite the importance of ligand removal in both catalysis and stability of TPNCs, the role of ligands and metal type in the process is not well understood. Herein, we utilize Density Functional Theory to understand the energetic interplay between metal-sulfur and sulfur-ligand bond dissociation in metal-thiolate systems. We first probe 66 metal-thiolate molecular complexes across combinations of M = Ag, Au, and Cu with twenty-two different ligands (R). Our results reveal that the energetics to break the metal-sulfur and sulfur-ligand bonds are strongly correlated and can be connected across all complexes through metal atomic ionization potentials. We then extend our work to the experimentally relevant [M25(SR)18]- TPNC, revealing the same correlations at the nanocluster level. Importantly, we unify our work by introducing a simple methodology to predict TPNC ligand removal energetics solely from calculations performed on metal-ligand molecular complexes. Finally, a computational mechanistic study was performed to investigate the hydrogenation pathways for SCH3-based complexes. The energy barriers for these systems revealed, in addition to thermodynamics, that kinetics favor the break of S-R over the M-S bond in the case of the Au complex. Our computational results rationalize several experimental observations pertinent to ligand effects on TPNCs. Overall, our introduced model provides an accelerated path to predict TPNC ligand removal energies, thus aiding towards targeted design of TPNC catalysts.

Atom-by-Atom Evolution of the Same Ligand-Protected Au21, Au22, Au22Cd1, and Au24 Nanocluster Series.

Atom-by-atom manipulation on metal nanoclusters (NCs) has long been desired, as the resulting series of NCs can provide insightful understanding of how a single atom affects the structure and properties as well as the evolution with size. Here, we report crystallizations of Au22(SAdm)16 and Au22Cd1(SAdm)16 (SAdm = adamantanethiolate) which link up with Au21(SAdm)15 and Au24(SAdm)16 NCs and form an atom-by-atom evolving series protected by the same ligand. Structurally, Au22(SAdm)16 has an Au3(SAdm)4 surface motif which is longer than the Au2(SAdm)3 on Au21(SAdm)15, whereas Au22Cd1(SAdm)16 lacks one staple Au atom compared to Au24(SAdm)16 and thus the surface structure is reconstructed. A single Cd atom triggers the structural transition from Au22 with a 10-atom bioctahedral kernel to Au22Cd1 with a 13-atom cuboctahedral kernel, and correspondingly, the optical properties are dramatically changed. The photoexcited carrier lifetime demonstrates that the optical properties and excited state relaxation are highly sensitive at the single atom level. By contrast, little change in both ionization potential and electron affinity is found in this series of NCs by theoretical calculations, indicating the electronic properties are independent of adding a single atom in this series. The work provides a paradigm that the NCs with continuous metal atom numbers are accessible and crystallizable when meticulously designed, and the optical properties are more affected at the single atom level than the electronic properties.

Towards elucidating structure of ligand-protected nanoclusters.

Ligand-protected metal nanoclusters (NCs) are organic-inorganic nanostructures, exhibiting high stability at specific "magic size" compositions and tunable properties that make them promising candidates for a wide range of nanotechnology-based applications. Synthesis and characterization of these nanostructures has been achieved with atomic precision, offering great opportunities to study the origin of new physicochemical property emergence at the nanoscale using theory and computation. In this Frontier article, we highlight the recent advances in the field of ligand-protected metal NCs, focusing on stability theories on monometallic and heterometal doped NCs, and NC structure prediction. Furthermore, we discuss current challenges on predicting previously undiscovered NCs and propose future steps to advance the field through applying first principles calculations, machine learning, and data-science-based approaches.

Rapid Prediction of Bimetallic Mixing Behavior at the Nanoscale.

The nanoparticle (NP) design space allows for variations in size, shape, composition, and chemical ordering. In the search for low-energy structures, this results in an extremely large search space which cannot be screened by brute force methods. In this work, we develop a genetic algorithm to predict stable bimetallic NPs of any size, shape, and metal composition. Our method predicts nanostructures in agreement with experimental trends and it captures the detailed chemical ordering of an experimental 23,196-atom FePt NP with nearly atom-by-atom accuracy. Our developed screening process is extremely fast, allowing us to generate and analyze a database of 5454 low-energy bimetallic NPs. By identifying thermodynamically stable NPs, we rationalize bimetallic mixing at the nanoscale and reveal metal-, size-, and temperature-dependent mixing behavior. Importantly, our method is applicable to any bimetallic NP size, bridging the materials gap in nanoscale simulations, and guides experimentation in the lab by elucidating stability, mixing, and detailed chemical ordering behavior of bimetallic NPs.

Heterometal-Doped M23 (M = Au/Ag/Cd) Nanoclusters with Large Dipole Moments.

Dipole moment (µ) is a critical parameter for molecules and nanomaterials as it affects many properties. In metal-thiolate (SR) nanoclusters (NCs), µ is commonly low (0-5 D) compared to quantum dots. Herein, we report a doping strategy to give giant dipoles (∼18 D) in M23 (M = Au/Ag/Cd) NCs, falling in the experimental trend for II-VI quantum dots. In M23 NCs, high µ is caused by the Cd-Br bond and the arrangement of heteroatoms along the C3 axis. Strong dipole-dipole interactions are observed in crystalline state, with energy exceeding 5 kJ/mol, directing a "head-to-tail" alignment of Au22-nAgnCd1(SR)15X (SR = adamantanethiolate) dipoles. The alignment can be controlled by µ via doping. The optical absorption peaks of M23 show solvent polarity-dependent shifts (∼25 meV) with negative solvatochromism. Detailed electronic structures of M23 are revealed by density functional theory and time-dependent DFT calculations. Overall, the doping strategy for obtaining large dipole moments demonstrates an atomic-level design of clusters with useful properties.

Structure-property relationships on thiolate-protected gold nanoclusters.

Since their discovery, thiolate-protected gold nanoclusters (Au n (SR) m ) have garnered a lot of interest due to their fascinating properties and "magic-number" stability. However, models describing the thermodynamic stability and electronic properties of these nanostructures as a function of their size are missing in the literature. Herein, we employ first principles calculations to rationalize the stability of fifteen experimentally determined gold nanoclusters in conjunction with a recently developed thermodynamic stability theory on small Au nanoclusters (≤102 Au atoms). Our results demonstrate that the thermodynamic stability theory can capture the stability of large, atomically precise nanoclusters, Au279(SR)84, Au246(SR)80, and Au146(SR)57, suggesting its applicability over larger cluster size regimes than its original development. Importantly, we develop structure-property relationships on Au nanoclusters, connecting their ionization potential and electron affinity to the number of gold atoms within the nanocluster. Altogether, a computational scheme is described that can aid experimental efforts towards a property-specific, targeted synthesis of gold nanoclusters.

Haploinsufficiency leads to neurodegeneration in C9ORF72 ALS/FTD human induced motor neurons.

An intronic GGGGCC repeat expansion in C9ORF72 is the most common cause of amyotrophic lateral sclerosis (ALS) and frontotemporal dementia (FTD), but the pathogenic mechanism of this repeat remains unclear. Using human induced motor neurons (iMNs), we found that repeat-expanded C9ORF72 was haploinsufficient in ALS. We found that C9ORF72 interacted with endosomes and was required for normal vesicle trafficking and lysosomal biogenesis in motor neurons. Repeat expansion reduced C9ORF72 expression, triggering neurodegeneration through two mechanisms: accumulation of glutamate receptors, leading to excitotoxicity, and impaired clearance of neurotoxic dipeptide repeat proteins derived from the repeat expansion. Thus, cooperativity between gain- and loss-of-function mechanisms led to neurodegeneration. Restoring C9ORF72 levels or augmenting its function with constitutively active RAB5 or chemical modulators of RAB5 effectors rescued patient neuron survival and ameliorated neurodegenerative processes in both gain- and loss-of-function C9ORF72 mouse models. Thus, modulating vesicle trafficking was able to rescue neurodegeneration caused by the C9ORF72 repeat expansion. Coupled with rare mutations in ALS2, FIG4, CHMP2B, OPTN and SQSTM1, our results reveal mechanistic convergence on vesicle trafficking in ALS and FTD.