Predicting crystal form stability under real-world conditions.

The physicochemical properties of molecular crystals, such as solubility, stability, compactability, melting behaviour and bioavailability, depend on their crystal form1. In silico crystal form selection has recently come much closer to realization because of the development of accurate and affordable free-energy calculations2-4. Here we redefine the state of the art, primarily by improving the accuracy of free-energy calculations, constructing a reliable experimental benchmark for solid-solid free-energy differences, quantifying statistical errors for the computed free energies and placing both hydrate crystal structures of different stoichiometries and anhydrate crystal structures on the same energy landscape, with defined error bars, as a function of temperature and relative humidity. The calculated free energies have standard errors of 1-2 kJ mol-1 for industrially relevant compounds, and the method to place crystal structures with different hydrate stoichiometries on the same energy landscape can be extended to other multi-component systems, including solvates. These contributions reduce the gap between the needs of the experimentalist and the capabilities of modern computational tools, transforming crystal structure prediction into a more reliable and actionable procedure that can be used in combination with experimental evidence to direct crystal form selection and establish control5.

Ab Initio-Based Bond Order Potential for Arsenene Polymorphs Developed via Hierarchical Reinforcement Learning.

Arsenene, a less-explored two-dimensional material, holds the potential for applications in wearable electronics, memory devices, and quantum systems. This study introduces a bond-order potential model with Tersoff formalism, the ML-Tersoff, which leverages multireward hierarchical reinforcement learning (RL), trained on an ab initio data set. This data set covers a spectrum of properties for arsenene polymorphs, enhancing our understanding of its mechanical and thermal behaviors without the complexities of traditional models requiring multiple parameter sets. Our RL strategy utilizes decision trees coupled with a hierarchical reward strategy to accelerate convergence in high-dimensional continuous search spaces. Unlike the Stillinger-Weber approach, which demands separate formalisms for buckled and puckered forms, the ML-Tersoff model concurrently captures multiple properties of the two polymorphs by effectively representing the local environment, thereby avoiding the need for different atomic types. We apply the ML model to understand the mechanical and thermal properties of the arsenene polymorphs and nanostructures. We observe an inverse relationship between the critical strain and temperature in arsenene. Thermal conductivity calculations in nanosheets show good agreement with ab initio data, reflecting a decrease in thermal conductivity attributable to increased anharmonic effects at higher temperatures. We also apply the model to predict the thermal behavior of arsenene nanotubes.

Atomic Cross-Talk at the Interface: Enhanced Lubricity and Wear and Corrosion Resistance in Sub 2 nm Hybrid Overcoats via Strengthened Interface Chemistry.

Friction, wear, and corrosion remain the major causes of premature failure of diverse systems including hard-disk drives (HDDs). To enhance the areal density of HDDs beyond 1 Tb/in2, the necessary low friction and high wear and corrosion resistance characteristics with sub 2 nm overcoats remain unachievable. Here we demonstrate that atom cross-talk not only manipulates the interface chemistry but also strengthens the tribological and corrosion properties of sub 2 nm overcoats. High-affinity (HA) atomically thin (∼0.4 nm) interlayers (ATIs, XHA), namely Ti, Si, and SiNx, are sandwiched between the hard-disk media and 1.5 nm thick carbon (C) overlayer to develop interface-enhanced sub 2 nm hybrid overcoats that consistently outperform a thicker conventional commercial overcoat (≥2.7 nm), with the C/SiNx bilayer overcoat bettering all other <2 nm thick overcoats. These hybrid overcoats can enable the development of futuristic 2-4 Tb/in2 areal density HDDs and can transform various moving-mechanical-system based technologies.

Development of Chemical Kinetics Models from Atomistic Reactive Molecular Dynamics Simulations: Application to Iso-octane Combustion and Rubber Ablative Degradation.

Modeling the complex chemical phenomena resulting from multiple active species and long-chain polymers is limited by uncertainties in the reaction rate parameters, which increase rapidly with the number of active species and/or reaction pathways. Reactive molecular dynamics simulations have the potential to help obtain in-depth information on the chemical reactions that occur between the polymer (e.g., ablative material) and the multiple active species in an aggressive environment. In this work, we demonstrate that molecular dynamics (MD) simulations using the ReaxFF interatomic potential can be used to obtain the reaction kinetics of complex reaction pathways at high temperatures. We report two recently developed tools, namely, MolfrACT and KinACT, designed to extract chemical kinetic pathways by postprocessing reactive MD simulation data. The pathway extraction is based on a new algorithm, Consistent Reaction Stoichiometry via an Iterative Scheme (CReSIS), for the automated extraction of reactions and kinetics from MD trajectories. As a validation of the methodology, we first report the kinetic analysis and mechanisms for the high-temperature combustion of iso-octane. The observed reaction pathways are consistent with experimental models. In addition, we compare the activation energies of select iso-octane combustion pathways with experimental data and show that nanosecond timescale molecular dynamics simulations are sufficient to obtain realistic estimates of activation energies for different fuel consumption reaction pathways at high temperatures. The framework developed here can potentially be combined with time-series forecasting and machine learning methods to further reduce the computational complexity of transient molecular dynamics simulations, yet yielding realistic chemical kinetics information. Subsequently, the CReSIS scheme applied to ethylene-propylene-diene-monomer (EPDM) rubber ablative reveals that H2O, C2H4, and HCHO are the major products during the initial stages of the polymer degradation in high-temperature oxidative environments. While prior work involving ReaxFF simulations is restricted to overall rates of formation of any species, we extract kinetic information for individual reaction pathways. In this paper, we present several reaction pathways observed during the EPDM rubber degradation into the dominant products and report the pathway-specific reaction rates. Arrhenius analysis reveals that the dominant reaction pathway activation energies for the formation of water, ethylene, and formaldehyde are 34.42, 27.26, and 6.37 kcal/mol, respectively. In contrast, the activation energies for the overall formation (across all reaction pathways) of these products are in the 40-50 kcal/mol range.

Slippery and Wear-Resistant Surfaces Enabled by Interface Engineered Graphene.

Friction and wear remain the primary cause of mechanical energy dissipation and system failure. Recent studies reveal graphene as a powerful solid lubricant to combat friction and wear. Most of these studies have focused on nanoscale tribology and have been limited to a few specific surfaces. Here, we uncover many unknown aspects of graphene's contact-sliding at micro- and macroscopic tribo-scales over a broader range of surfaces. We discover that graphene's performance reduces for surfaces with increasing roughness. To overcome this, we introduce a new type of graphene/silicon nitride (SiNx, 3 nm) bilayer overcoats that exhibit superior performance compared to native graphene sheets (mono and bilayer), that is, display the lowest microscale friction and wear on a range of tribologically poor flat surfaces. More importantly, two-layer graphene/SiNx bilayer lubricant (<4 nm in total thickness) shows the highest macroscale wear durability on tape-head (topologically variant surface) that exceeds most previous thicker (∼7-100 nm) overcoats. Detailed nanoscale characterization and atomistic simulations explain the origin of the reduced friction and wear arising from these nanoscale coatings. Overall, this study demonstrates that engineered graphene-based coatings can outperform conventional coatings in a number of technologies.

Three-Dimensional Integrated X-ray Diffraction Imaging of a Native Strain in Multi-Layered WSe2.

Emerging two-dimensional (2-D) materials such as transition-metal dichalcogenides show great promise as viable alternatives for semiconductor and optoelectronic devices that progress beyond silicon. Performance variability, reliability, and stochasticity in the measured transport properties represent some of the major challenges in such devices. Native strain arising from interfacial effects due to the presence of a substrate is believed to be a major contributing factor. A full three-dimensional (3-D) mapping of such native nanoscopic strain over micron length scales is highly desirable for gaining a fundamental understanding of interfacial effects but has largely remained elusive. Here, we employ coherent X-ray diffraction imaging to directly image and visualize in 3-D the native strain along the (002) direction in a typical multilayered (∼100-350 layers) 2-D dichalcogenide material (WSe2) on silicon substrate. We observe significant localized strains of ∼0.2% along the out-of-plane direction. Experimentally informed continuum models built from X-ray reconstructions trace the origin of these strains to localized nonuniform contact with the substrate (accentuated by nanometer scale asperities, i.e., surface roughness or contaminants); the mechanically exfoliated stresses and strains are localized to the contact region with the maximum strain near surface asperities being more or less independent of the number of layers. Machine-learned multimillion atomistic models show that the strain effects gain in prominence as we approach a few- to single-monolayer limit. First-principles calculations show a significant band gap shift of up to 125 meV per percent of strain. Finally, we measure the performance of multiple WSe2 transistors fabricated on the same flake; a significant variability in threshold voltage and the "off" current setting is observed among the various devices, which is attributed in part to substrate-induced localized strain. Our integrated approach has broad implications for the direct imaging and quantification of interfacial effects in devices based on layered materials or heterostructures.

Ultrafast Three-Dimensional X-ray Imaging of Deformation Modes in ZnO Nanocrystals.

Imaging the dynamical response of materials following ultrafast excitation can reveal energy transduction mechanisms and their dissipation pathways, as well as material stability under conditions far from equilibrium. Such dynamical behavior is challenging to characterize, especially operando at nanoscopic spatiotemporal scales. In this letter, we use X-ray coherent diffractive imaging to show that ultrafast laser excitation of a ZnO nanocrystal induces a rich set of deformation dynamics including characteristic "hard" or inhomogeneous and "soft" or homogeneous modes at different time scales, corresponding respectively to the propagation of acoustic phonons and resonant oscillation of the crystal. By integrating the 3D nanocrystal structure obtained from the ultrafast X-ray measurements with a continuum thermo-electro-mechanical finite element model, we elucidate the deformation mechanisms following laser excitation, in particular, a torsional mode that generates a 50% greater electric potential gradient than that resulting from the flexural mode. Understanding of the time-dependence of these mechanisms on ultrafast scales has significant implications for development of new materials for nanoscale power generation.

Ultrafast Three-Dimensional Integrated Imaging of Strain in Core/Shell Semiconductor/Metal Nanostructures.

Visualizing the dynamical response of material heterointerfaces is increasingly important for the design of hybrid materials and structures with tailored properties for use in functional devices. In situ characterization of nanoscale heterointerfaces such as metal-semiconductor interfaces, which exhibit a complex interplay between lattice strain, electric potential, and heat transport at subnanosecond time scales, is particularly challenging. In this work, we use a laser pump/X-ray probe form of Bragg coherent diffraction imaging (BCDI) to visualize in three-dimension the deformation of the core of a model core/shell semiconductor-metal (ZnO/Ni) nanorod following laser heating of the shell. We observe a rich interplay of radial, axial, and shear deformation modes acting at different time scales that are induced by the strain from the Ni shell. We construct experimentally informed models by directly importing the reconstructed crystal from the ultrafast experiment into a thermo-electromechanical continuum model. The model elucidates the origin of the deformation modes observed experimentally. Our integrated imaging approach represents an invaluable tool to probe strain dynamics across mixed interfaces under operando conditions.

Thermal transport across a substrate-thin-film interface: effects of film thickness and surface roughness.

Using molecular dynamics simulations and a model AlN-GaN interface, we demonstrate that the interfacial thermal resistance R(K) (Kapitza resistance) between a substrate and thin film depends on the thickness of the film and the film surface roughness when the phonon mean free path is larger than film thickness. In particular, when the film (external) surface is atomistically smooth, phonons transmitted from the substrate can travel ballistically in the thin film, be scattered specularly at the surface, and return to the substrate without energy transfer. If the external surface scatters phonons diffusely, which is characteristic of rough surfaces, R(K) is independent of film thickness and is the same as R(K) that characterizes smooth surfaces in the limit of large film thickness. At interfaces where phonon transmission coefficients are low, the thickness dependence is greatly diminished regardless of the nature of surface scattering. The film thickness dependence of R(K) is analogous to the well-known fact of lateral thermal conductivity thickness dependence in thin films. The difference is that phonon-boundary scattering lowers the in-plane thermal transport in thin films, but it facilitates thermal transport from the substrate to the thin film.

Molecular dynamics investigation of nanoscale cavitation dynamics.

We use molecular dynamics simulations to investigate the cavitation dynamics around intensely heated solid nanoparticles immersed in a model Lennard-Jones fluid. Specifically, we study the temporal evolution of vapor nanobubbles that form around the solid nanoparticles heated over ps time scale and provide a detail description of the following vapor formation and collapse. For 8 nm diameter nanoparticles we observe the formation of vapor bubbles when the liquid temperature 0.5-1 nm away from the nanoparticle surface reaches ∼90% of the critical temperature, which is consistent with the onset of spinodal decomposition. The peak heat flux from the hot solid to the surrounding liquid at the bubble formation threshold is ∼20 times higher than the corresponding steady state critical heat flux. Detailed analysis of the bubble dynamics indicates adiabatic formation followed by an isothermal final stage of growth and isothermal collapse.

Curvature induced phase stability of an intensely heated liquid.

We use non-equilibrium molecular dynamics simulations to study the heat transfer around intensely heated solid nanoparticles immersed in a model Lennard-Jones fluid. We focus our studies on the role of the nanoparticle curvature on the liquid phase stability under steady-state heating. For small nanoparticles we observe a stable liquid phase near the nanoparticle surface, which can be at a temperature well above the boiling point. Furthermore, for particles with radius smaller than a critical radius of 2 nm we do not observe formation of vapor even above the critical temperature. Instead, we report the existence of a stable fluid region with a density much larger than that of the vapor phase. We explain the stability in terms of the Laplace pressure associated with the formation of a vapor nanocavity and the associated effect on the Gibbs free energy.

Liquid phase stability under an extreme temperature gradient.

Using nonequilibrium molecular dynamics simulations, we subject bulk liquid to a very high-temperature gradient and observe a stable liquid phase with a local temperature well above the boiling point. Also, under this high-temperature gradient, the vapor phase exhibits condensation into a liquid at a temperature higher than the saturation temperature, indicating that the observed liquid stability is not caused by nucleation barrier kinetics. We show that, assuming local thermal equilibrium, the phase change can be understood from the thermodynamic analysis. The observed elevation of the boiling point is associated with the interplay between the "bulk" driving force for the phase change and surface tension of the liquid-vapor interface that suppresses the transformation. This phenomenon is analogous to that observed for liquids in confined geometries. In our study, however, a low-temperature liquid, rather than a solid, confines the high-temperature liquid.

Learning in continuous action space for developing high dimensional potential energy models.

Reinforcement learning (RL) approaches that combine a tree search with deep learning have found remarkable success in searching exorbitantly large, albeit discrete action spaces, as in chess, Shogi and Go. Many real-world materials discovery and design applications, however, involve multi-dimensional search problems and learning domains that have continuous action spaces. Exploring high-dimensional potential energy models of materials is an example. Traditionally, these searches are time consuming (often several years for a single bulk system) and driven by human intuition and/or expertise and more recently by global/local optimization searches that have issues with convergence and/or do not scale well with the search dimensionality. Here, in a departure from discrete action and other gradient-based approaches, we introduce a RL strategy based on decision trees that incorporates modified rewards for improved exploration, efficient sampling during playouts and a "window scaling scheme" for enhanced exploitation, to enable efficient and scalable search for continuous action space problems. Using high-dimensional artificial landscapes and control RL problems, we successfully benchmark our approach against popular global optimization schemes and state of the art policy gradient methods, respectively. We demonstrate its efficacy to parameterize potential models (physics based and high-dimensional neural networks) for 54 different elemental systems across the periodic table as well as alloys. We analyze error trends across different elements in the latent space and trace their origin to elemental structural diversity and the smoothness of the element energy surface. Broadly, our RL strategy will be applicable to many other physical science problems involving search over continuous action spaces.

Efficient Crystal Structure Prediction for Structurally Related Molecules with Accurate and Transferable Tailor-Made Force Fields.

Crystal structure prediction (CSP) is generally used to complement experimental solid form screening and applied to individual molecules in drug development. The fast development of algorithms and computing resources offers the opportunity to use CSP earlier and for a broader range of applications in the drug design cycle. This study presents a novel paradigm of CSP specifically designed for structurally related molecules, referred to as Quick-CSP. The approach prioritizes more accurate physics through robust and transferable tailor-made force fields (TMFFs), such that significant efficiency gains are achieved through the reduction of expensive ab initio calculations. The accuracy of the TMFF is increased by the introduction of electrostatic multipoles, and the fragment-based force field parameterization scheme is demonstrated to be transferable for a family of chemically related molecules. The protocol is benchmarked with structurally related compounds from the Bromodomain and Extraterminal (BET) domain inhibitors series. A new convergence criterion is introduced that aims at performing only as many ab initio optimizations of crystal structures as required to locate the bottom of the crystal energy landscape within a user-defined accuracy. The overall approach provides significant cost savings ranging from three- to eight-fold less than the full-CSP workflow. The reported advancements expand the scope and utility of the underlying CSP building blocks as well as their novel reassembly to other applications earlier in the drug design cycle to guide molecule design and selection.

Formation and Nature of Carbon-Containing Tribofilms.

Minimizing friction and wear at a rubbing interface continues to be a challenge and has resulted in the recent surge toward the use of coatings such as diamond-like carbon (DLC) on machine components. The problem with the coating approach is the limitation of coating wear life. Here, we report a lubrication approach in which lubricious, wear-protective carbon-containing tribofilms can be self-generated and replenishable, without any surface pretreatment. Such carbon-containing films were formed under modest sliding conditions in a lubricant consisting of cyclopropanecarboxylic acid as an additive dissolved in polyalphaolefin base oil. These tribofilms show the same Raman D and G signatures that have been interpreted to be due to the presence of graphite- or DLC films. Our experimental measurements and reactive molecular dynamics simulations demonstrate that these tribofilms are in fact high-molecular weight hydrocarbons acting as a solid lubricant.

Machine learning a bond order potential model to study thermal transport in WSe2 nanostructures.

Nanostructures of transition metal di-chalcogenides (TMDCs) exhibit exotic thermal, chemical and electronic properties, enabling diverse applications from thermoelectrics and catalysis to nanoelectronics. The thermal properties of these nanoscale TMDCs are of particular interest for thermoelectric applications. Thermal transport studies on nanotubes and nanoribbons remain intractable to first principles calculations whereas existing classical molecular models treat the two chalcogen layers in a monolayer with different atom types; this imposes serious limitations in studying multi-layered TMDCs and dynamical phenomena such as nucleation and growth. Here, we overcome these limitations using machine learning (ML) and introduce a bond order potential (BOP) trained against first principles training data to capture the structure, dynamics, and thermal transport properties of a model TMDC such as WSe2. The training is performed using a hierarchical objective genetic algorithm workflow to accurately describe the energetics, as well as thermal and mechanical properties of a free-standing sheet. As a representative case study, we perform molecular dynamics simulations using the ML-BOP model to study the structure and temperature-dependent thermal conductivity of WSe2 tubes and ribbons of different chiralities. We observe slightly higher thermal conductivities along the armchair direction than zigzag for WSe2 monolayers but the opposite effect for nanotubes, especially of smaller diameters. We trace the origin of these differences to the anisotropy in thermal transport and the restricted momentum selection rules for phonon-phonon Umpklapp scattering. The developed ML-BOP model is of broad interest and will facilitate studies on nucleation and growth of low dimensional WSe2 structures as well as their transport properties for thermoelectric and thermal management applications.

Active site localization of methane oxidation on Pt nanocrystals.

High catalytic efficiency in metal nanocatalysts is attributed to large surface area to volume ratios and an abundance of under-coordinated atoms that can decrease kinetic barriers. Although overall shape or size changes of nanocatalysts have been observed as a result of catalytic processes, structural changes at low-coordination sites such as edges, remain poorly understood. Here, we report high-lattice distortion at edges of Pt nanocrystals during heterogeneous catalytic methane oxidation based on in situ 3D Bragg coherent X-ray diffraction imaging. We directly observe contraction at edges owing to adsorption of oxygen. This strain increases during methane oxidation and it returns to the original state after completing the reaction process. The results are in good agreement with finite element models that incorporate forces, as determined by reactive molecular dynamics simulations. Reaction mechanisms obtained from in situ strain imaging thus provide important insights for improving catalysts and designing future nanostructured catalytic materials.

Silicon compatible Sn-based resistive switching memory.

Large banks of cheap, fast, non-volatile, energy efficient, scalable solid-state memories are an increasingly essential component for today's data intensive computing. Conductive-bridge random access memory (CBRAM) - which involves voltage driven formation and dissolution of Cu or Ag filaments in a Cu (or Ag) anode/dielectric (HfO2 or Al2O3)/inert cathode device - possesses the necessary attributes to fit the requirements. Cu and Ag are, however, fast diffusers and known contaminants in silicon microelectronics. Herein, employing a criterion for electrode metal selection applicable to cationic filamentary devices and using first principles calculations for estimating diffusion barriers in HfO2, we identify tin (Sn) as a rational, silicon CMOS compatible replacement for Cu and Ag anodes in CBRAM devices. We then experimentally fabricate Sn based CBRAM devices and demonstrate very fast, steep-slope memory switching as well as threshold switching, comparable to Cu or Ag based devices. Furthermore, time evolution of the cationic filament formation along with the switching mechanism is discussed based on time domain measurements (I vs. t) carried out under constant voltage stress. The time to threshold is shown to be a function of both the voltage stress (Vstress) as well as the initial leakage current (I0) through the device.

Quantitative Observation of Threshold Defect Behavior in Memristive Devices with Operando X-ray Microscopy.

Memristive devices are an emerging technology that enables both rich interdisciplinary science and novel device functionalities, such as nonvolatile memories and nanoionics-based synaptic electronics. Recent work has shown that the reproducibility and variability of the devices depend sensitively on the defect structures created during electroforming as well as their continued evolution under dynamic electric fields. However, a fundamental principle guiding the material design of defect structures is still lacking due to the difficulty in understanding dynamic defect behavior under different resistance states. Here, we unravel the existence of threshold behavior by studying model, single-crystal devices: resistive switching requires that the pristine oxygen vacancy concentration reside near a critical value. Theoretical calculations show that the threshold oxygen vacancy concentration lies at the boundary for both electronic and atomic phase transitions. Through operando, multimodal X-ray imaging, we show that field tuning of the local oxygen vacancy concentration below or above the threshold value is responsible for switching between different electrical states. These results provide a general strategy for designing functional defect structures around threshold concentrations to create dynamic, field-controlled phases for memristive devices.

Ab Initio-Based Bond Order Potential to Investigate Low Thermal Conductivity of Stanene Nanostructures.

We introduce a bond order potential (BOP) for stanene based on an ab initio derived training data set. The potential is optimized to accurately describe the energetics, as well as thermal and mechanical properties of a free-standing sheet, and used to study diverse nanostructures of stanene, including tubes and ribbons. As a representative case study, using the potential, we perform molecular dynamics simulations to study stanene's structure and temperature-dependent thermal conductivity. We find that the structure of stanene is highly rippled, far in excess of other 2-D materials (e.g., graphene), owing to its low in-plane stiffness (stanene: ∼ 25 N/m; graphene: ∼ 480 N/m). The extent of stanene's rippling also shows stronger temperature dependence compared to that in graphene. Furthermore, we find that stanene based nanostructures have significantly lower thermal conductivity compared to graphene based structures owing to their softness (i.e., low phonon group velocities) and high anharmonic response. Our newly developed BOP will facilitate the exploration of stanene based low dimensional heterostructures for thermoelectric and thermal management applications.