Deep generative priors for biomolecular 3D heterogeneous reconstruction from cryo-EM projections.

Cryo-electron microscopy has become a powerful tool to determine three-dimensional (3D) structures of rigid biological macromolecules from noisy micrographs with single-particle reconstruction. Recently, deep neural networks, e.g., CryoDRGN, have demonstrated conformational and compositional heterogeneity of complexes. However, the lack of ground-truth conformations poses a challenge to assess the performance of heterogeneity analysis methods. In this work, variational autoencoders (VAE) with three types of deep generative priors were learned for latent variable inference and heterogeneous 3D reconstruction via Bayesian inference. More specifically, VAEs with "Variational Mixture of Posteriors" priors (VampPrior-SPR), non-parametric exemplar-based priors (ExemplarPrior-SPR) and priors from latent score-based generative models (LSGM-SPR) were quantitatively compared with CryoDRGN. We built four simulated datasets composed of hypothetical continuous conformation or discrete states of the hERG K + channel. Empirical and quantitative comparisons of inferred latent representations were performed with affine-transformation-based metrics. These models with more informative priors gave better regularized, interpretable factorized latent representations with better conserved pairwise distances, less deformed latent distributions and lower within-cluster variances. They were also tested on experimental datasets to resolve compositional and conformational heterogeneity (50S ribosome assembly, cowpea chlorotic mottle virus, and pre-catalytic spliceosome) with comparable high resolution. Codes and data are available: https://github.com/benjamin3344/DGP-SPR.

Simultaneous secondary electron microscopy in the scanning transmission electron microscope with applications for in situ studies.

Scanning/transmission electron microscopy (STEM) is a powerful characterization tool for a wide range of materials. Over the years, STEMs have been extensively used for in situ studies of structural evolution and dynamic processes. A limited number of STEM instruments are equipped with a secondary electron (SE) detector in addition to the conventional transmitted electron detectors, i.e. the bright-field (BF) and annular dark-field (ADF) detectors. Such instruments are capable of simultaneous BF-STEM, ADF-STEM and SE-STEM imaging. These methods can reveal the 'bulk' information from BF and ADF signals and the surface information from SE signals for materials <200 nm thick. This review first summarizes the field of in situ STEM research, followed by the generation of SE signals, SE-STEM instrumentation and applications of SE-STEM analysis. Combining with various in situ heating, gas reaction and mechanical testing stages based on microelectromechanical systems (MEMS), we show that simultaneous SE-STEM imaging has found applications in studying the dynamics and transient phenomena of surface reconstructions, exsolution of catalysts, lunar and planetary materials and mechanical properties of 2D thin films. Finally, we provide an outlook on the potential advancements in SE-STEM from the perspective of sample-related factors, instrument-related factors and data acquisition and processing.

In-situ TEM study from the perspective of holders.

During the in situ transmission electron microscopy (TEM) observations, the diverse functionalities of different specimen holders play a crucial role. We hereby provide a comprehensive overview of the main types of holders, associated technologies and case studies pertaining to the widely employed heating and gas heating methods, from their initial developments to the latest advancement. In addition to the conventional approaches, we also discuss the emergence of holders that incorporate a micro-electro-mechanical system (MEMS) chip for in situ observations. The MEMS technology offers a multitude of functions within a single chip, thereby enhancing the capabilities and versatility of the holders. MEMS chips have been utilized in environmental-cell designs, enabling customized fabrication of diverse shapes. This innovation has facilitated their application in conducting in situ observations within gas and liquid environments, particularly in the investigation of catalytic and battery reactions. We summarize recent noteworthy studies conducted using in situ liquid TEM. These studies highlight significant advancements and provide valuable insights into the utilization of MEMS chips in environmental-cells, as well as the expanding capabilities of in situ liquid TEM in various research domains.

Defect-Induced Secondary Crystals Drive Two-Dimensional to Three-Dimensional Morphological Evolution in the Co-Self-Assembly of Polyferrocenylsilane Block Copolymer and Homopolymer.

Bottom-up fabrication protocols for uniform 3D hierarchical structures in solution are rare. We report two different approaches to fabricate uniform 3D spherulites and their precursors using mixtures of poly(ferrocenyldimethylsilane) (PFS) block copolymer (BCP) and PFS homopolymer (HP). Both protocols are designed to promote defects in 2D assemblies that serve as intermediate structures. In a multistep seeded growth protocol, we add the BCP/HP mixture to (1D) rod-like PFS micelles in a selective solvent as first-generation seeds. This leads to 2D platelet structures. If this step is conducted at a high supersaturation, secondary crystals form on the basal surface of these platelets. Co-crystallization and rapid crystallization of BCP/HP promote the formation of defects that act as nucleation sites for secondary crystals, resulting in multilayer platelets. This is the key step. The multilayer platelets serve as second-generation seeds upon subsequent addition of BCP/HP blends and, with increasing supersaturation, lead to the sequential formation of uniform (3D) hedrites, sheaves, and spherulites. Similar structures can also be obtained by a simple one-pot direct self-assembly (heating-cooling-aging) protocol of PFS BCP/HP blends. In this case, for a carefully chosen but narrow temperature range, PFS HPs nucleate formation of uniform structures, and the annealing temperature regulates the supersaturation level. In both protocols, the competitive crystallization kinetics of HP/BCP affects the morphology. Both protocols exhibit broad generality. We believe the morphological transformation from 2D to 3D structures, regulated by defect formation, co-crystallization, and supersaturation levels, could apply to various semicrystalline polymers. Moreover, the 3D structures are sufficiently robust to serve as recoverable carriers for nanoparticle catalysts, exhibiting valuable catalytic activity and opening new possibilities for applications requiring exquisite 3D structures.

Peltier cooling for the reduction of carbon contamination in scanning electron microscopy.

We used a novel Peltier anticontamination device (PAC) to reduce carbon contamination upon electron beam irradiation in scanning electron microscopy through a reduction of hydrocarbon molecules in the specimen chamber. Unlike liquid-nitrogen based cold traps, the PAC operates free of user maintenance and is suitable for lengthy imaging sessions without degradation of the anticontamination performance. Its performance as an alternative cold trap method provides considerable reduction of electron beam-assisted carbon build-up. We compared the thickness of carbon contamination deposited upon prolonged electron beam scans with the PAC system on and off. Topographical structures of the carbon build-up were characterized using atomic force microscopy. We report that under identical beam parameters, thickness of the carbon contamination was reduced by over 79 % for area scans (1.2 × 1.2 µm2), and by two orders of magnitude for stationary point scans when the PAC cooling mode is engaged.

Polyferrocenylsilane Block Copolymer Spherulites in Dilute Solution.

Self-assembly of block copolymers (BCP) into uniform 3D structures in solution is an extremely rare phenomenon. Furthermore, the investigation of general prerequisites for fabricating a specific uniform 3D structure remains unknown and challenging. Here, through a simple one-pot direct self-assembly (heating and cooling) protocol, we show that uniform spherulite-like structures and their precursors can be prepared with various poly(ferrocenyldimethylsilane) (PFS) BCPs in a variety of polar and non-polar solvents. These structures all evolve from elongated lamellae into hedrites, sheaf-like micelles, and finally spherulites as the annealing temperature and supersaturation degree are increased. The key feature leading to this growth trajectory is the formation of secondary crystals by self-nucleation on the surface of early-elongated lamellae. We identified general prerequisites for fabricating PFS BCP spherulites in solution. These include corona/PFS core block ratios in the range of 1-5.5 that favor the formation of 2D structures as well as the development of secondary crystals on the basal faces of platelets at early stages of the self-assembly. The one-pot direct self-assembly provides a general protocol to form uniform spherulites and their precursors consisting of PFS BCPs that match these prerequisites. In addition, we show that manipulation of various steps in the direct self-assembly protocol can regulate the size and shape of the structures formed. These general concepts show promise for the fabrication and optimization of spherulites and their precursors from semicrystalline BCPs with interesting optical, electronic, or biomedical properties using the one-pot direct self-assembly protocol.

Microstructure-Dependent Thermal Stability of Super-Tetragonal Nanocomposite Films through In Situ TEM/EELS Study.

Smart microstructure design in nanocomposite films allows us to tailor physical properties such as ferroelectricity and thermal stability to broaden applications of next-generation electronic devices. Here, we study the thermal stability of self-assembled PbTiO3 (PTO)/PbO nanocomposite films with nano-spherical and nanocolumnar microstructures by utilizing an environmental transmission electron microscopy (TEM) combined with electron energy loss spectroscopy (EELS). The real-time study reveals that the microstructure-dependent interphase strain has an effect on the stabilization of the tetragonal phase. Compared to the nano-spherical configuration, the nanocomposite film with the nanocolumnar microstructure can maintain the giant tetragonality of ∼1.20 up to 450 °C, and the tetragonal phase is predicted to be stable at elevated temperatures > 600 °C. Moreover, the temperature-dependent EELS further demonstrates the sensitivity of the chemical bonding of Pb and Ti with O to the PTO lattice distortion, correlating the structural variation and electronic properties at different temperatures. Such in situ heating TEM study provides insights into the thermal stability of nanocomposites with different microstructures and facilitates the advancement of power electronics applications in harsh environments.

New black indium oxide-tandem photothermal CO2-H2 methanol selective catalyst.

It has long been known that the thermal catalyst Cu/ZnO/Al2O3(CZA) can enable remarkable catalytic performance towards CO2 hydrogenation for the reverse water-gas shift (RWGS) and methanol synthesis reactions. However, owing to the direct competition between these reactions, high pressure and high hydrogen concentration (≥75%) are required to shift the thermodynamic equilibrium towards methanol synthesis. Herein, a new black indium oxide with photothermal catalytic activity is successfully prepared, and it facilitates a tandem synthesis of methanol at a low hydrogen concentration (50%) and ambient pressure by directly using by-product CO as feedstock. The methanol selectivities achieve 33.24% and 49.23% at low and high hydrogen concentrations, respectively.

Automatic quantification and classification of microplastics in scanning electron micrographs via deep learning.

Microplastics quantification and classification are demanding jobs to monitor microplastic pollution and evaluate the potential health risks. In this paper, microplastics from daily supplies in diverse chemical compositions and shapes are imaged by scanning electron microscopy. It offers a greater depth and finer details of microplastics at a wider range of magnification than visible light microscopy or a digital camera, and permits further chemical composition analysis. However, it is labour-intensive to manually extract microplastics from micrographs, especially for small particles and thin fibres. A deep learning approach facilitates microplastics quantification and classification with a manually annotated dataset including 237 micrographs of microplastic particles (fragments or beads) in the range of 50 µm-1 mm and fibres with diameters around 10 µm. For microplastics quantification, two deep learning models (U-Net and MultiResUNet) were implemented for semantic segmentation. Both significantly outmatched conventional computer vision techniques and achieved a high average Jaccard index over 0.75. Especially, U-Net was combined with object-aware pixel embedding to perform instance segmentation on densely packed and tangled fibres for further quantification. For shape classification, a fine-tuned VGG16 neural network classifies microplastics based on their shapes with high accuracy of 98.33%. With trained models, it takes only seconds to segment and classify a new micrograph in high accuracy, which is remarkably cheaper and faster than manual labour. The growing datasets may benefit the identification and quantification of microplastics in environmental samples in future work.

The role of cooling rate in crystallization-driven block copolymer self-assembly.

Self-assembly of crystalline-coil block copolymers (BCPs) in selective solvents is often carried out by heating the mixture until the sample appears to dissolve and then allowing the solution to cool back to room temperature. In self-seeding experiments, some crystallites persist during sample annealing and nucleate the growth of core-crystalline micelles upon cooling. There is evidence in the literature that the nature of the self-assembled structures formed is independent of the annealing time at a particular temperature. There are, however, no systematic studies of how the rate of cooling affects self-assembly. We examine three systems based upon poly(ferrocenyldimethylsilane) BCPs that generated uniform micelles under typical conditions where cooling took pace on the 1-2 h time scale. For example, several of the systems generated elongated 1D micelles of uniform length under these slow cooling conditions. When subjected to rapid cooling (on the time scale of a few minutes or faster), branched structures were obtained. Variation of the cooling rate led to a variation in the size and degree of branching of some of the structures examined. These changes can be explained in terms of the high degree of supersaturation that occurs when unimer solutions at high temperature are suddenly cooled. Enhanced nucleation, seed aggregation, and selective growth of the species of lowest solubility contribute to branching. Cooling rate becomes another tool for manipulating crystallization-driven self-assembly and controlling micelle morphologies.

Binary Cu2-xS Templates Direct the Formation of Quaternary Cu2ZnSnS4 (Kesterite, Wurtzite) Nanocrystals.

Kesterite Cu2ZnSnS4 (k-CZTS) nanocrystals have received attention for their tunable optoelectronic properties, as well as the earth abundance of their constituent atoms. However, the phase-pure synthesis of these quaternary NCs is challenging due to their polymorphism, as well as the undesired formation of related binary and ternary impurities. A general synthetic route to tackle this complexity is to pass through intermediate template nanocrystals that direct subsequent cation exchange toward the desired quaternary crystalline phase, particularly those that are thermodynamically disfavored or otherwise synthetically challenging. Here, working within this model multinary system, we achieve control over the formation of three binary copper sulfide polymorphs, cubic digenite (Cu1.8S), hexagonal covellite (CuS), and monoclinic djurleite (Cu1.94S). Controlled experiments with Cu0 seeds show that selected binary phases can be favored by the identity and stoichiometry of the sulfur precursor alone under otherwise comparable reaction conditions. We then demonstrate that the nature of the Cu2-xS template dictates the final polymorph of the CZTS nanocrystal products. Through digenite, the cation exchange reaction readily yields the k-CZTS phase due to its highly similar anion sublattice. Covellite nanocrystals template the k-CZTS phase but via major structural rearrangement to digenite that requires elevated temperatures in the absence of a strong reducing agent. In contrast, we show that independently synthesized djurleite nanorods template the formation of the wurtzite polymorph (w-CZTS) but with prominent stacking faults in the final product. Applying this refined understanding to the standard one-pot syntheses of k- and w-CZTS nanocrystals, we identify that these reactions are each effectively templated by binary intermediates formed in situ, harnessing their properties to guide the overall synthesis of phase-pure quaternary materials. Our results provide tools for the careful development of tailored nanocrystal syntheses in complex polymorphic systems.

Gold-in-copper at low *CO coverage enables efficient electromethanation of CO2.

The renewable-electricity-powered CO2 electroreduction reaction provides a promising means to store intermittent renewable energy in the form of valuable chemicals and dispatchable fuels. Renewable methane produced using CO2 electroreduction attracts interest due to the established global distribution network; however, present-day efficiencies and activities remain below those required for practical application. Here we exploit the fact that the suppression of *CO dimerization and hydrogen evolution promotes methane selectivity: we reason that the introduction of Au in Cu favors *CO protonation vs. C-C coupling under low *CO coverage and weakens the *H adsorption energy of the surface, leading to a reduction in hydrogen evolution. We construct experimentally a suite of Au-Cu catalysts and control *CO availability by regulating CO2 concentration and reaction rate. This strategy leads to a 1.6× improvement in the methane:H2 selectivity ratio compared to the best prior reports operating above 100 mA cm-2. We as a result achieve a CO2-to-methane Faradaic efficiency (FE) of (56 ± 2)% at a production rate of (112 ± 4) mA cm-2.

Uniform 1D Micelles and Patchy & Block Comicelles via Scalable, One-Step Crystallization-Driven Block Copolymer Self-Assembly.

Fiber-like (1D) core-crystalline micelles of uniform length can be obtained in protocols involving multiple steps from block copolymers (BCPs) in which crystallization of the core-forming polymer drives the self-assembly. Here we report a systematic study that shows that adding small amounts (<5 w/w%) of a homopolymer corresponding to the core-forming block of the BCP enables uniform 1D micelles (mean lengths Ln = 0.6 to 9.7 µm) to be obtained in a single step, simply by heating the mixture in a selective solvent followed by slow cooling. A series of poly(ferrocenyldimethylsilane) (PFS) BCPs with different corona-forming blocks and different compositions as well as PFS homopolymers of different lengths were examined. Dye labeling and confocal fluorescence microscopy showed that the homopolymer ends up in the center of the micelle, signaling that it served as the initial seed for epitaxial micelle growth. The rate of unimer addition was strongly enhanced by the length of the PFS block, and this enabled more complex structures to be formed in one-pot self-assembly experiments from mixtures of two or three BCPs with different PFS block lengths. Furthermore, BCP mixtures that included PFS-b-PI (PI = polyisoprene) and PFS-b-PDMS with similar PFS block lengths resulted in simultaneous addition to growing micelles, resulting in a patchy block that could be visualized by staining the vinyl groups of the PI with Pt nanoparticles. This approach also enabled scale up, so that uniform 1D micelles of controlled architecture can be obtained at concentrations of 10 w/w % solids or more.

Spherulite-Like Micelles.

One-dimensional (1D) and 2D structures by crystallization-driven self-assembly of block copolymers (BCPs) can form fascinating hierarchical structures through secondary self-assembly. But examples of 3D structures formed via hierarchical self-assembly are rare. Here we report seeded growth experiments in decane of a poly(ferrocenyldimethylsilane) BCP with an amphiphilic corona forming block in which lenticular platelets grow into classic spherulite-like uniform colloidally stable structures. These 3D objects are spherically symmetric on the exterior, but asymmetric near the core, where there is a more open structure consisting of sheaf-like leaves. The most remarkable aspect of these experiments is that growth stops at different stages of growth process, depending upon how much unimer is added in the seeded growth step. The system provides a model for studying spherulitic growth where real-time observations on their growth at different stages remains challenging.

Highly Efficient Antimicrobial Activity of CuxFeyOz Nanoparticles against Important Human Pathogens.

The development of innovative antimicrobial materials is crucial in thwarting infectious diseases caused by microbes, as drug-resistant pathogens are increasing in both number and capacity to detoxify the antimicrobial drugs used today. An ideal antimicrobial material should inhibit a wide variety of bacteria in a short period of time, be less or not toxic to normal cells, and the fabrication or synthesis process should be cheap and easy. We report a one-step microwave-assisted hydrothermal synthesis of mixed composite CuxFeyOz (Fe2O3/Cu2O/CuO/CuFe2O) nanoparticles (NPs) as an excellent antimicrobial material. The 1 mg/mL CuxFeyOz NPs with the composition 36% CuFeO2, 28% Cu2O and 36% Fe2O3 have a general antimicrobial activity greater than 5 log reduction within 4 h against nine important human pathogenic bacteria (including drug-resistant bacteria as well as Gram-positive and Gram-negative strains). For example, they induced a >9 log reduction in Escherichia coli B viability after 15 min of incubation, and an ~8 log reduction in multidrug-resistant Klebsiella pneumoniae after 4 h incubation. Cytotoxicity tests against mouse fibroblast cells showed about 74% viability when exposed to 1 mg/mL CuxFeyOz NPs for 24 h, compared to the 20% viability for 1 mg/mL pure Cu2O NPs synthesized by the same method. These results show that the CuxFeyOz composite NPs are a highly efficient, low-toxicity and cheap antimicrobial material that has promising potential for applications in medical and food safety.

Publisher Correction: Copper adparticle enabled selective electrosynthesis of n-propanol.

An amendment to this paper has been published and can be accessed via a link at the top of the paper.

Conversion of Metal-Organic Cage to Ligand-Free Ultrasmall Noble Metal Nanocluster Catalysts Confined within Mesoporous Silica Nanoparticle Supports.

Supported ultrasmall noble metal nanocluster-based (UNMN-based) catalysts are one of the most important classes of solid materials for heterogeneous catalysis. In this work, we present a novel strategy for the controlled synthesis of ligand-free UNMN nanocatalysts based on in situ reduction of a palladium-based (Pd-based) metal-organic cage (MOC) confined within monosized, thiol-modified mesoporous silica nanoparticle (MSN) supports. By taking advantage of the high mutual solubility of MOCs and MSNs in DMSO and the strong interactions between the thiol-modified MSN pore wall and MOC surface, a good dispersion of MOC molecules was achieved throughout the MSN support. The close correspondence of the MSN pore diameter (ca. 5.0 nm) with the diameter of the MOC (ca. 4.0 nm) confines MOC packing to approximately a monolayer. Based on this spatial constraint and electrostatic binding of the MOC to the thiol-modified MSN pore surface, in situ MOC reduction followed by metal atom diffusion, coalescence, and anchoring on the active sites resulted in ligand-free Pd-based UNMNs of approximately 0.9 ± 0.2 nm in diameter decorating the MSN pore surfaces. Control experiments of the reduction of a conventional palladium source or the reduction of free, unconstrained cages in solution under the same conditions only produced large metal nanocrystals (NP, >2 nm), confirming the importance of confined reduction to achieve a highly catalytically active surface. In light of this strategy, two catalytic experiments including the reaction of 4-nitrophenol to 4-aminophenol and the Suzuki C-C coupling reaction show superior catalytic activity of the engineered MSN-supported UNMN nanocatalysts compared to their free form and state of the art commercial catalysts. We believe that our new strategy will open new avenues for artificially designed UNMN-inspired nanoarchitectures for wide applications.

Copper adparticle enabled selective electrosynthesis of n-propanol.

The electrochemical reduction of carbon monoxide is a promising approach for the renewable production of carbon-based fuels and chemicals. Copper shows activity toward multi-carbon products from CO reduction, with reaction selectivity favoring two-carbon products; however, efficient conversion of CO to higher carbon products such as n-propanol, a liquid fuel, has yet to be achieved. We hypothesize that copper adparticles, possessing a high density of under-coordinated atoms, could serve as preferential sites for n-propanol formation. Density functional theory calculations suggest that copper adparticles increase CO binding energy and stabilize two-carbon intermediates, facilitating coupling between adsorbed *CO and two-carbon intermediates to form three-carbon products. We form adparticle-covered catalysts in-situ by mediating catalyst growth with strong CO chemisorption. The new catalysts exhibit an n-propanol Faradaic efficiency of 23% from CO reduction at an n-propanol partial current density of 11 mA cm-2.

Lipid-Coated Mesoporous Silica Nanoparticles for the Delivery of the ML336 Antiviral to Inhibit Encephalitic Alphavirus Infection.

Venezuelan equine encephalitis virus (VEEV) poses a major public health risk due to its amenability for use as a bioterrorism agent and its severe health consequences in humans. ML336 is a recently developed chemical inhibitor of VEEV, shown to effectively reduce VEEV infection in vitro and in vivo. However, its limited solubility and stability could hinder its clinical translation. To overcome these limitations, lipid-coated mesoporous silica nanoparticles (LC-MSNs) were employed. The large surface area of the MSN core promotes hydrophobic drug loading while the liposome coating retains the drug and enables enhanced circulation time and biocompatibility, providing an ideal ML336 delivery platform. LC-MSNs loaded 20 ± 3.4 µg ML336/mg LC-MSN and released 6.6 ± 1.3 µg/mg ML336 over 24 hours. ML336-loaded LC-MSNs significantly inhibited VEEV in vitro in a dose-dependent manner as compared to unloaded LC-MSNs controls. Moreover, cell-based studies suggested that additional release of ML336 occurs after endocytosis. In vivo safety studies were conducted in mice, and LC-MSNs were not toxic when dosed at 0.11 g LC-MSNs/kg/day for four days. ML336-loaded LC-MSNs showed significant reduction of brain viral titer in VEEV infected mice compared to PBS controls. Overall, these results highlight the utility of LC-MSNs as drug delivery vehicles to treat VEEV.

Nonlinear fracture toughness measurement and crack propagation resistance of functionalized graphene multilayers.

Despite promising applications of two-dimensional (2D) materials, one major concern is their propensity to fail in a brittle manner, which results in a low fracture toughness causing reliability issues in practical applications. We show that this limitation can be overcome by using functionalized graphene multilayers with fracture toughness (J integral) as high as ~39 J/m2, measured via a microelectromechanical systems-based in situ transmission electron microscopy technique coupled with nonlinear finite element fracture analysis. The measured fracture toughness of functionalized graphene multilayers is more than two times higher than graphene (~16 J/m2). A linear fracture analysis, similar to that previously applied to other 2D materials, was also conducted and found to be inaccurate due to the nonlinear nature of the stress-strain response of functionalized graphene multilayers. A crack arresting mechanism of functionalized graphene multilayers was experimentally observed and identified as the main contributing factor for the higher fracture toughness as compared to graphene. Molecular dynamics simulations revealed that the interactions among functionalized atoms in constituent layers and distinct fracture pathways in individual layers, due to a random distribution of functionalized carbon atoms in multilayers, restrict the growth of a preexisting crack. The results inspire potential strategies for overcoming the relatively low fracture toughness of 2D materials through chemical functionalization.