Physics constrained unsupervised deep learning for rapid, high resolution scanning coherent diffraction reconstruction.

By circumventing the resolution limitations of optics, coherent diffractive imaging (CDI) and ptychography are making their way into scientific fields ranging from X-ray imaging to astronomy. Yet, the need for time consuming iterative phase recovery hampers real-time imaging. While supervised deep learning strategies have increased reconstruction speed, they sacrifice image quality. Furthermore, these methods' demand for extensive labeled training data is experimentally burdensome. Here, we propose an unsupervised physics-informed neural network reconstruction method, PtychoPINN, that retains the factor of 100-to-1000 speedup of deep learning-based reconstruction while improving reconstruction quality by combining the diffraction forward map with real-space constraints from overlapping measurements. In particular, PtychoPINN gains a factor of 4 in linear resolution and an 8 dB improvement in PSNR while also accruing improvements in generalizability and robustness. This blend of performance and computational efficiency offers exciting prospects for high-resolution real-time imaging in high-throughput environments such as X-ray free electron lasers (XFELs) and diffraction-limited light sources.

Tuning In-Plane Magnetic Anisotropy and Interfacial Exchange Coupling in Epitaxial La2/3Sr1/3CoO3/La2/3Sr1/3MnO3 Heterostructures.

Controlling the in-plane magnetocrystalline anisotropy and interfacial exchange coupling between ferromagnetic (FM) layers plays a key role in next-generation spintronic and magnetic memory devices. In this work, we explored the effect of tuning the magnetocrystalline anisotropy of La2/3Sr1/3CoO3 (LSCO) and La2/3Sr1/3MnO3 (LSMO) layers and the corresponding effect on interfacial exchange coupling by adjusting the thickness of the LSCO layer (tLSCO). The epitaxial LSCO/LSMO bilayers were grown on (110)o-oriented NdGaO3 (NGO) substrates with a fixed LSMO (top layer) thickness of 6 nm and LSCO (bottom layer) thicknesses varying from 1 to 10 nm. Despite the small difference (∼0.2%) in lattice mismatch between the two in-plane directions, [001]o and [11̅0]o, a pronounced in-plane magnetic anisotropy was observed. Soft X-ray magnetic circular dichroism hysteresis loops revealed that for tLSCO ≤ 4 nm, the easy axes for both LSCO and LSMO layers were along the [001]o direction, and the LSCO layer was characterized by magnetically active Co2+ ions that strongly coupled to the LSMO layer. No exchange bias effect was observed in the hysteresis loops. In contrast, along the [11̅0]o direction, the LSCO and LSMO layers displayed a small difference in their coercivity values, and a small exchange bias shift was observed. As tLSCO increased above 4 nm, the easy axis for the LSCO layer remained along the [100]o direction, but it gradually rotated to the [11̅0]o direction for the LSMO layer, resulting in a large negative exchange bias shift. Therefore, we provide a way to control the magnetocrystalline anisotropy and exchange bias by tuning the interfacial exchange coupling between the two FM layers.

Two Pathways for the Degradation of Orpiment Pigment (As2S3) Found in Paintings.

Paintings are complex objects containing many different chemical compounds that can react over time. The degradation of arsenic sulfide pigments causes optical changes in paintings. The main degradation product was thought to be white arsenolite (As2O3), but previous research also showed the abundant presence of As(V) species. In this study, we investigate the influence of the presence of a medium on the degradation mechanism of orpiment (As2S3) using synchrotron radiation (SR)-based tomographic transmission X-ray microscopy, SR-based micro-X-ray fluorescence, and X-ray absorption near edge structure spectroscopy. Upon direct illumination of dry orpiment powder using UV-visible light, only the formation of As2O3 was observed. When As2S3 was surrounded by a medium and illuminated, As2O3 was only observed in the area directly exposed to light, while As(V) degradation species were found elsewhere in the medium. Without accelerated artificial light aging, As(V)(aq) species are formed and migrate throughout the medium within weeks after preparation. In both scenarios, the As(V) species form via intermediate As(III)(aq) species and the presence of a medium is necessary. As(V)(aq) species can react with available cations to form insoluble metal arsenates, which induces stress within the paint layers (leading to, e.g., cracks and delamination) or can lead to a visual change of the image of the painting.

Discovery of LaAlO3 as an efficient catalyst for two-electron water electrolysis towards hydrogen peroxide.

Electrochemical two-electron water oxidation reaction (2e-WOR) has drawn significant attention as a promising process to achieve the continuous on-site production of hydrogen peroxide (H2O2). However, compared to the cathodic H2O2 generation, the anodic 2e-WOR is more challenging to establish catalysts due to the severe oxidizing environment. In this study, we combine density functional theory (DFT) calculations with experiments to discover a stable and efficient perovskite catalyst for the anodic 2e-WOR. Our theoretical screening efforts identify LaAlO3 perovskite as a stable, active, and selective candidate for catalyzing 2e-WOR. Our experimental results verify that LaAlO3 achieves an overpotential of 510 mV at 10 mA cm-2 in 4 M K2CO3/KHCO3, lower than those of many reported metal oxide catalysts. In addition, LaAlO3 maintains a stable H2O2 Faradaic efficiency with only a 3% decrease after 3 h at 2.7 V vs. RHE. This computation-experiment synergistic approach introduces another effective direction to discover promising catalysts for the harsh anodic 2e-WOR towards H2O2.

Nanocrystallite Seeding of Metastable Ferroelectric Phase Formation in Atomic Layer-Deposited Hafnia-Zirconia Alloys.

Hafnia-based ferroelectric thin films are promising for semiconductor memory and neuromorphic computing applications. Amorphous, as-deposited, thin-film binary alloys of HfO2 and ZrO2 transform to the metastable, orthorhombic ferroelectric phase during post-deposition annealing and cooling. This transformation is generally thought to involve formation of a tetragonal precursor phase that distorts into the orthorhombic phase during cooling. In this work, we systematically study the effects of atomic layer deposition (ALD) temperature on the ferroelectricity of post-deposition-annealed Hf0.5Zr0.5O2 (HZO) thin films. Seed crystallites having interplanar spacings consistent with the polar orthorhombic phase are observed by a plan-view transmission electron microscope in HZO thin films deposited at an elevated ALD temperature. After ALD under conditions that promote formation of these nanocrystallites, high-polarization (Pr > 18 µC/cm2) ferroelectric switching is observed after rapid thermal annealing (RTA) at low temperature (350 °C). These results indicate the presence of minimal non-ferroelectric phases retained in the films after RTA when the ALD process forms nanocrystalline particles that seed subsequent formation of the polar orthorhombic phase.

Atomic-Layer-Deposited Aluminum Oxide Thin Films Probed with X-ray Scattering and Compared to Molecular Dynamics and Density Functional Theory Models.

A better understanding of amorphous aluminum oxide's structure and electronic properties is obtained through combined experimental and computational approaches. Grazing incidence X-ray scattering measurements were carried out on aluminum oxide thin films grown using thermal atomic layer deposition. The corresponding pair distribution functions (PDFs) showed structures similar to previously reported PDFs of solid-state amorphous alumina and molten alumina. Structural models based on crystalline alumina polymorphs (PDFgui) and amorphous alumina (molecular dynamics, MD) were examined for structural comparisons to the experimental PDF data. Smaller MD models were optimized and verified against larger models to allow for quantum chemical electronic structure calculations. The electronic structure of the amorphous alumina models yields additional insight into the band structure and electronic defects present in amorphous alumina that are not present in crystalline samples.

Giant Orbital Anisotropy with Strong Spin-Orbit Coupling Established at the Pseudomorphic Interface of the Co/Pd Superlattice.

Orbital anisotropy at interfaces in magnetic heterostructures has been key to pioneering spin-orbit-related phenomena. However, modulating the interface's electronic structure to make it abnormally asymmetric has been challenging because of lack of appropriate methods. Here, the authors report that low-energy proton irradiation achieves a strong level of inversion asymmetry and unusual strain at interfaces in [Co/Pd] superlattices through nondestructive, selective removal of oxygen from Co3 O4 /Pd superlattices during irradiation. Structural investigations corroborate that progressive reduction of Co3 O4 into Co establishes pseudomorphic growth with sharp interfaces and atypically large tensile stress. The normal component of orbital to spin magnetic moment at the interface is the largest among those observed in layered Co systems, which is associated with giant orbital anisotropy theoretically confirmed, and resulting very large interfacial magnetic anisotropy is observed. All results attribute not only to giant orbital anisotropy but to enhanced interfacial spin-orbit coupling owing to the pseudomorphic nature at the interface. They are strongly supported by the observation of reversal of polarity of temperature-dependent Anomalous Hall signal, a signature of Berry phase. This work suggests that establishing both giant orbital anisotropy and strong spin-orbit coupling at the interface is key to exploring spintronic devices with new functionalities.

Ultrathin ferroic HfO2-ZrO2 superlattice gate stack for advanced transistors.

With the scaling of lateral dimensions in advanced transistors, an increased gate capacitance is desirable both to retain the control of the gate electrode over the channel and to reduce the operating voltage1. This led to a fundamental change in the gate stack in 2008, the incorporation of high-dielectric-constant HfO2 (ref. 2), which remains the material of choice to date. Here we report HfO2-ZrO2 superlattice heterostructures as a gate stack, stabilized with mixed ferroelectric-antiferroelectric order, directly integrated onto Si transistors, and scaled down to approximately 20 ångströms, the same gate oxide thickness required for high-performance transistors. The overall equivalent oxide thickness in metal-oxide-semiconductor capacitors is equivalent to an effective SiO2 thickness of approximately 6.5 ångströms. Such a low effective oxide thickness and the resulting large capacitance cannot be achieved in conventional HfO2-based high-dielectric-constant gate stacks without scavenging the interfacial SiO2, which has adverse effects on the electron transport and gate leakage current3. Accordingly, our gate stacks, which do not require such scavenging, provide substantially lower leakage current and no mobility degradation. This work demonstrates that ultrathin ferroic HfO2-ZrO2 multilayers, stabilized with competing ferroelectric-antiferroelectric order in the two-nanometre-thickness regime, provide a path towards advanced gate oxide stacks in electronic devices beyond conventional HfO2-based high-dielectric-constant materials.

MLExchange: A web-based platform enabling exchangeable machine learning workflows for scientific studies.

Machine learning (ML) algorithms are showing a growing trend in helping the scientific communities across different disciplines and institutions to address large and diverse data problems. However, many available ML tools are programmatically demanding and computationally costly. The MLExchange project aims to build a collaborative platform equipped with enabling tools that allow scientists and facility users who do not have a profound ML background to use ML and computational resources in scientific discovery. At the high level, we are targeting a full user experience where managing and exchanging ML algorithms, workflows, and data are readily available through web applications. Since each component is an independent container, the whole platform or its individual service(s) can be easily deployed at servers of different scales, ranging from a personal device (laptop, smart phone, etc.) to high performance clusters (HPC) accessed (simultaneously) by many users. Thus, MLExchange renders flexible using scenarios-users could either access the services and resources from a remote server or run the whole platform or its individual service(s) within their local network.

Highly Efficient Uniaxial In-Plane Stretching of a 2D Material via Ion Insertion.

On-chip dynamic strain engineering requires efficient micro-actuators that can generate large in-plane strains. Inorganic electrochemical actuators are unique in that they are driven by low voltages (≈1 V) and produce considerable strains (≈1%). However, actuation speed and efficiency are limited by mass transport of ions. Minimizing the number of ions required to actuate is thus key to enabling useful "straintronic" devices. Here, it is shown that the electrochemical intercalation of exceptionally few lithium ions into WTe2 causes large anisotropic in-plane strain: 5% in one in-plane direction and 0.1% in the other. This efficient stretching of the 2D WTe2 layers contrasts to intercalation-induced strains in related materials which are predominantly in the out-of-plane direction. The unusual actuation of Lix WTe2 is linked to the formation of a newly discovered crystallographic phase, referred to as Td', with an exotic atomic arrangement. On-chip low-voltage (<0.2 V) control is demonstrated over the transition to the novel phase and its composition. Within the Td'-Li0.5- δ WTe2 phase, a uniaxial in-plane strain of 1.4% is achieved with a change of δ of only 0.075. This makes the in-plane chemical expansion coefficient of Td'-Li0.5-δ WTe2 far greater than of any other single-phase material, enabling fast and efficient planar electrochemical actuation.

A refraction correction for buried interfaces applied to in situ grazing-incidence X-ray diffraction studies on Pd electrodes.

In situ characterization of electrochemical systems can provide deep insights into the structure of electrodes under applied potential. Grazing-incidence X-ray diffraction (GIXRD) is a particularly valuable tool owing to its ability to characterize the near-surface structure of electrodes through a layer of electrolyte, which is of paramount importance in surface-mediated processes such as catalysis and adsorption. Corrections for the refraction that occurs as an X-ray passes through an interface have been derived for a vacuum-material interface. In this work, a more general form of the refraction correction was developed which can be applied to buried interfaces, including liquid-solid interfaces. The correction is largest at incidence angles near the critical angle for the interface and decreases at angles larger and smaller than the critical angle. Effective optical constants are also introduced which can be used to calculate the critical angle for total external reflection at the interface. This correction is applied to GIXRD measurements of an aqueous electrolyte-Pd interface, demonstrating that the correction allows for the comparison of GIXRD measurements at multiple incidence angles. This work improves quantitative analysis of d-spacing values from GIXRD measurements of liquid-solid systems, facilitating the connection between electrochemical behavior and structure under in situ conditions.

Characterization data of an (AlFeNiTiVZr)1-xCrx multi-principal element alloy continuous composition spread library.

The data provided in this article is related to the research article entitled "Phase stabilization and oxidation of a continuous composition spread multi-principal element (AlFeNiTiVZr)1-xCrx alloy" [1]. This data article describes the high-throughput synthesis and characterization processes of an (AlFeNiTiVZr)1-xCrx alloy system. Continuous composition spread (CCS) thin-film libraries were synthesized by co-depositing an AlFeNiTiVZr metal alloy target and Cr target via magnetron sputtering. Post-processing was performed on the sample libraries with a vacuum anneal at 873 K and an air anneal at 873 K. Compositional data was determined via WDS in order to verify parameters provided by an in-house sputter model. Crystallographic data was captured via synchrotron diffraction and diffractograms were compared as a function of the change in Cr concentration. These measurements were taken in order to observe phase behavior after oxidation throughout the composition library. Furthermore, vibrational spectrographic data is provided of the oxidized library to show surface speciation along the composition gradient of the alloy system. The structural and oxidative behavior of the (AlFeNiTiVZr)1-xCrx alloy can be analysed using the data provided in this article. Additionally, this characterization dataset can be utilized in machine learning algorithms for determining important features and parameters for future hypothesis generation of functional multi-principal element alloys (MPEAs).

Oxidation State and Surface Reconstruction of Cu under CO2 Reduction Conditions from In Situ X-ray Characterization.

The electrochemical CO2 reduction reaction (CO2RR) using Cu-based catalysts holds great potential for producing valuable multi-carbon products from renewable energy. However, the chemical and structural state of Cu catalyst surfaces during the CO2RR remains a matter of debate. Here, we show the structural evolution of the near-surface region of polycrystalline Cu electrodes under in situ conditions through a combination of grazing incidence X-ray absorption spectroscopy (GIXAS) and X-ray diffraction (GIXRD). The in situ GIXAS reveals that the surface oxide layer is fully reduced to metallic Cu before the onset potential for CO2RR, and the catalyst maintains the metallic state across the potentials relevant to the CO2RR. We also find a preferential surface reconstruction of the polycrystalline Cu surface toward (100) facets in the presence of CO2. Quantitative analysis of the reconstruction profiles reveals that the degree of reconstruction increases with increasingly negative applied potentials, and it persists when the applied potential returns to more positive values. These findings show that the surface of Cu electrocatalysts is dynamic during the CO2RR, and emphasize the importance of in situ characterization to understand the surface structure and its role in electrocatalysis.

High-Throughput Characterization of (FexCo1-x)3O4 Thin-Film Composition Spreads.

Thin-film continuous composition spreads of Fe-Co-O were fabricated by reactive cosputtering from elemental Fe and Co targets in reactive Ar/O2 atmosphere using deposition temperatures ranging from 300 to 700 °C. Fused silica and platinized Si/SiO2 strips were used as substrates. Ti and Ta were investigated as adhesion layer for Pt and the fabrication of the Fe-Co-O films. The thin-film composition spreads were characterized by high-throughput electron-dispersive X-ray spectroscopy, X-ray diffraction, X-ray photoelectron spectroscopy, atomic force microscopy, scanning electron microscopy, and optical transmission spectroscopy. The Fe-content ranged from 28 to 72 at. %. The spinel phases Fe2CoO4 and FeCo2O4 could be synthesized and stabilized at all deposition temperatures with a continuous variation in spinel composition in between. The dependence of the film surface microstructure on the deposition temperature and the composition was mapped. Moreover, the band gap values, ranging from 2.41 eV for FeCo2O4 to 2.74 eV for Fe2CoO4, show a continuous variation with the composition.

On-the-fly closed-loop materials discovery via Bayesian active learning.

Active learning-the field of machine learning (ML) dedicated to optimal experiment design-has played a part in science as far back as the 18th century when Laplace used it to guide his discovery of celestial mechanics. In this work, we focus a closed-loop, active learning-driven autonomous system on another major challenge, the discovery of advanced materials against the exceedingly complex synthesis-processes-structure-property landscape. We demonstrate an autonomous materials discovery methodology for functional inorganic compounds which allow scientists to fail smarter, learn faster, and spend less resources in their studies, while simultaneously improving trust in scientific results and machine learning tools. This robot science enables science-over-the-network, reducing the economic impact of scientists being physically separated from their labs. The real-time closed-loop, autonomous system for materials exploration and optimization (CAMEO) is implemented at the synchrotron beamline to accelerate the interconnected tasks of phase mapping and property optimization, with each cycle taking seconds to minutes. We also demonstrate an embodiment of human-machine interaction, where human-in-the-loop is called to play a contributing role within each cycle. This work has resulted in the discovery of a novel epitaxial nanocomposite phase-change memory material.

Exploring the First High-Entropy Thin Film Libraries: Composition Spread-Controlled Crystalline Structure.

Thin films of two types of high-entropy oxides (HEOs) have been deposited on 76.2 mm Si wafers using combinatorial sputter deposition. In one type of the oxides, (MgZnMnCoNi)Ox, all the metals have a stable divalent oxidation state and similar cationic radii. In the second type of oxides, (CrFeMnCoNi)Ox, the metals are more diverse in the atomic radius and valence state, and have good solubility in their sub-binary and ternary oxide systems. The resulting HEO thin films were characterized using several high-throughput analytical techniques. The microstructure, composition, and electrical conductivity obtained on defined grid maps were obtained for the first time across large compositional ranges. The crystalline structure of the films was observed as a function of the metallic elements in the composition spreads, that is, the Mn and Zn in (MgZnMnCoNi)Ox and Mn and Ni in (CrFeMnCoNi)Ox. The (MgZnMnCoNi)Ox sample was observed to form two-phase structures, except single spinel structure was found in (MgZnMnCoNi)Ox over a range of Mn > 12 at. % and Zn < 44 at. %, while (CrFeMnCoNi)Ox was always observed to form two-phase structures. Composition-controlled crystalline structure is not only experimentally demonstrated but also supported by density function theory calculation.

Combinatorial Exploration and Mapping of Phase Transformation in a Ni-Ti-Co Thin Film Library.

Combinatorial synthesis and high-throughput characterization of a Ni-Ti-Co thin film materials library are reported for exploration of reversible martensitic transformation. The library was prepared by magnetron co-sputtering, annealed in vacuum at 500 °C without atmospheric exposure, and evaluated for shape memory behavior as an indicator of transformation. Composition, structure, and transformation behavior of the 177 pads in the library were characterized using high-throughput wavelength dispersive spectroscopy (WDS), X-ray photoelectron spectroscopy (XPS), X-ray diffraction (XRD), and four-point probe temperature-dependent resistance (R(T)) measurements. A new, expanded composition space having phase transformation with low thermal hysteresis and Co > 10 at. % is found. Unsupervised machine learning methods of hierarchical clustering were employed to streamline data processing of the large XRD and XPS data sets. Through cluster analysis of XRD data, we identified and mapped the constituent structural phases. Composition-structure-property maps for the ternary system are made to correlate the functional properties to the local microstructure and composition of the Ni-Ti-Co thin film library.

Controlling Magnetization Vector Depth Profiles of La0.7Sr0.3CoO3/La0.7Sr0.3MnO3 Exchange Spring Bilayers via Interface Reconstruction.

The La0.7Sr0.3CoO3-δ/La0.7Sr0.3MnO3-δ (LSCO/LSMO) bilayer system is an ideal perovskite oxide platform for investigating interface reconstruction and its effect on their magnetic properties. Previous studies have shown that LSCO can separate into magnetic sublayers, which possess distinct trends as the total LSCO thickness increases. In this study, we used polarized neutron reflectometry to quantify changes in the magnetic and chemical depth profiles, and it confirms the formation of ∼12 Å-thick interfacial LSCO and LSMO layers, characterized by a decreased nuclear scattering length density compared to the bulk of the layers. This decrease is attributed to the combined effects of oxygen vacancy formation and interfacial charge transfer, which lead to magnetically active Co2+ ions with ionic radii larger than the Co3+/Co4+ ions typically found in bulk LSCO or single-layer films. The interfacial magnetization values, as well as Co2+ ion and oxygen vacancy concentrations, depend strongly on the LSCO layer thickness. These results highlight the sensitive interplay of the cation valence states, oxygen vacancy concentration, and magnetization at interfaces in perovskite oxide multilayers, demonstrating the potential to tune their functional properties via careful design of their structure.

Structural and photoelectrochemical properties in the thin film system Cu-Fe-V-O and its ternary subsystems Fe-V-O and Cu-V-O.

Thin-film material libraries in the ternary and quaternary metal oxide systems Fe-V-O, Cu-V-O, and Cu-Fe-V-O were synthesized using combinatorial reactive co-sputtering with subsequent annealing in air. Their compositional, structural, and functional properties were assessed using high-throughput characterization methods. Prior to the investigation of the quaternary system Cu-Fe-V-O, the compositions (Fe61V39)Ox and (Cu52V48)Ox with promising photoactivity were identified from their ternary subsystems Fe-V-O and Cu-V-O, respectively. Two Cu-Fe-V-O material libraries with (Cu29-72Fe4-27V22-57)Ox and (Cu11-55Fe27-73V12-34)Ox composition spread were investigated. Seven mixed ternary and quaternary phase regions were identified: I (α-Cu3FeV6O26/FeVO4), II (Cu5V2O10/FeVO4/α-Cu3Fe4V6O26), III (Cu5V2O10), IV (Cu5V2O10/FeVO4, V (FeVO4/γ-Cu2V2O7/α-Cu3Fe4V6O26), VI (ß-Cu2V2O7/α-Cu3Fe4V6O26/FeVO4), and VII (ß-Cu3Fe4V6O26/FeVO4). In the investigated composition range, two photoactive regions, (Cu53Fe7V40)Ox and (Cu45Fe21V34)Ox, were identified, exhibiting 103 µA/cm2 and 108 µA/cm2 photocurrent density for the oxygen evolution reaction at 1.63 V vs reversible hydrogen electrode, respectively. The highest photoactive region (Cu45Fe21V34)Ox comprises the dominant α-Cu3Fe4V6O24 phase and minor FeVO4 phase. This photoactive region corresponds to having an indirect bandgap of 1.87 eV and a direct bandgap of 2.58 eV with an incident photon-to-current efficiency of 30% at a wavelength of 310 nm.