Search

Learning the structure of Directed Acyclic Graphs (DAGs) presents a significant challenge due to the vast combinatorial search space of possible graphs, which scales exponentially with the number of nodes. Recent advancements have redefined this problem as a continuous optimization task by incorporating differentiable acyclicity constraints. These methods commonly rely on algebraic characterizations of DAGs, such as matrix exponentials, to enable the use of gradient-based optimization techniques. Despite these innovations, existing methods often face optimization difficulties due to the highly non-convex nature of DAG constraints and the per-iteration computational complexity. In this work, we present a novel framework for learning DAGs, employing a Stochastic Approximation approach integrated with Stochastic Gradient Descent (SGD)-based optimization techniques. Our framework introduces new projection methods tailored to efficiently enforce DAG constraints, ensuring that the algorithm converges to a feasible local minimum. With its low iteration complexity, the proposed method is well-suited for handling large-scale problems with improved computational efficiency. We demonstrate the effectiveness and scalability of our framework through comprehensive experimental evaluations, which confirm its superior performance across various settings.

Precisely determining the oxidation states of metal cations within variable-valence transition metal oxides remains a significant challenge, yet it is crucial for understanding and predicting the properties of these technologically important materials. Iron oxides, in particular, exhibit a remarkable diversity of electronic structures due to the variable valence states of iron (Fe2+ and Fe3+), however, quantitative analysis using conventional X-ray photoelectron spectroscopy (XPS) is challenging because of significant overlapping of the Fe2p spectra among different oxidation states. In this study, we leverage the intriguing case of Pt supported FeO2 phase of monolayer thickness (ML) as a model system and employ Resonant Photoemission Spectroscopy (ResPES) to directly quantify the cation valence states and compositional ratios in this complex Fe oxide. Our results reveal that this ultrathin FeO2 film (Pt-O-Fe-O), contrary to the +3 valence predicted by density functional theory (DFT), consists of an equal mixture of Fe2+ and Fe3+ cations, yielding an average valence of +2.5. Structurally, FeO2 is likely derived from the Fe3O4 sublattice, featuring an octahedral Fe layer (50% Fe3+ and 50% Fe2+) bonded to upper and lower oxygen layers., Comment: 14 pages, 4 figures, in submission

Functionalization and volatilization are competing reactions during the oxidation of carbonaceous materials and are important processes in many different areas of science and technology. Here we present a combined ambient pressure X-ray photoelectron spectroscopy (APXPS) and grazing incidence X-ray scattering (GIXS) investigation of the oxidation of oleic acid ligands surrounding NaYF4 nanoparticles (NPs) deposited onto SiOx/Si substrates. While APXPS monitors the evolution of the oxidation products, GIXS provides insight into the morphology of the ligands and particles before and after the oxidation. Our investigation shows that the oxidation of the oleic acid ligands proceeds at O2 partial pressures of below 1 mbar in the presence of X-rays, with the oxidation eventually reaching a steady state in which mainly CHx and -COOH functional groups are observed. The scattering data reveal that the oxidation and volatilization reaction proceeds preferentially on the side of the particle facing the gas phase, leading to the formation of a chemically and morphologically asymmetric ligand layer. This comprehensive picture of the oxidation process could only be obtained by combining the X-ray scattering and APXPS data. The investigation presented here lays the foundation for further studies of the stability of NP layers in the presence of reactive trace gasses and ionizing radiation, and for other nanoscale systems where chemical and morphological changes happen simultaneously and cannot be understood in isolation., Comment: 18 pages, 8 figures

Hydrogen is a promising alternative to fossil fuel, however storing it efficiently poses challenges. One promising solution is to adsorb hydrogen on solid materials demonstrating quasi-molecular bonding with hydrogen. The hydrogen adsorption energy can be tuned by changing the morphology or stoichiometry of bimetallic nanoparticles. Here we used complementary techniques to unveil the chemical compositional and morphological transformation undergone by PdxNi100-x nanoparticles during H2 adsorption. Our findings reveal NiO-rich shell and Pd-rich core as confirmed by X-ray photoelectron spectroscopy, X-ray scattering, and electron energy loss spectroscopy. During hydrogen adsorption, which mainly occurs on Pd atoms, the Pd-rich core fragments into small pockets, increasing its surface area. This process is more pronounced for nanoparticles with lower Pd loading, emphasizing the role of NiO. These results shed light on the atomic changes occurring in the PdxNi100-x nanoparticles during hydrogen adsorption and can be applicable on multi-metallic systems to improve hydrogen storage properties.

This paper investigates the global convergence of stepsized Newton methods for convex functions with H\"older continuous Hessians or third derivatives. We propose several simple stepsize schedules with fast global convergence guarantees, up to $\mathcal {O}\left(k^{-3} \right)$. For cases with multiple plausible smoothness parameterizations or an unknown smoothness constant, we introduce a stepsize linesearch and a backtracking procedure with provable convergence as if the optimal smoothness parameters were known in advance. Additionally, we present strong convergence guarantees for the practically popular Newton method with exact linesearch., Comment: 11 pages

The global food delivery market provides many opportunities for AI-based services that can improve the efficiency of feeding the world. This paper presents the Cloud Kitchen platform as a decision-making tool for restaurants with food delivery and a simulator to evaluate the impact of the decisions. The platform contains a Technology-Specific Bridge (TSB) that provides an interface for communicating with restaurants or the simulator. TSB uses a planning domain model to represent decisions embedded in the Unified Planning Framework (UPF). Decision-making, which concerns allocating customers' orders to vehicles and deciding in which order the customers will be served (for each vehicle), is done via a Vehicle Routing Problem with Time Windows (VRPTW), an efficient tool for this problem. We show that decisions made by our platform can improve customer satisfaction by reducing the number of delayed deliveries using a real-world historical dataset.

Classical neural networks achieve only limited convergence in hard problems such as XOR or parity when the number of hidden neurons is small. With the motivation to improve the success rate of neural networks in these problems, we propose a new neural network model inspired by existing neural network models with so called product neurons and a learning rule derived from classical error backpropagation, which elegantly solves the problem of mutually exclusive situations. Unlike existing product neurons, which have weights that are preset and not adaptable, our product layers of neurons also do learn. We tested the model and compared its success rate to a classical multilayer perceptron in the aforementioned problems as well as in other hard problems such as the two spirals. Our results indicate that our model is clearly more successful than the classical MLP and has the potential to be used in many tasks and applications., Comment: This work was funded by the Horizon-Widera-2021 European Twinning project TERAIS G.A. n. 101079338. Presented at International Conference on Artificial Neural Networks (ICANN) 2023. Accepted:1.7.2023 Published:26.9.2023. Code: https://doi.org/10.5281/zenodo.10702248

Machine learning assumes a pivotal role in our data-driven world. The increasing scale of models and datasets necessitates quick and reliable algorithms for model training. This dissertation investigates adaptivity in machine learning optimizers. The ensuing chapters are dedicated to various facets of adaptivity, including: 1. personalization and user-specific models via personalized loss, 2. provable post-training model adaptations via meta-learning, 3. learning unknown hyperparameters in real time via hyperparameter variance reduction, 4. fast O(1/k^2) global convergence of second-order methods via stepsized Newton method regardless of the initialization and choice basis, 5. fast and scalable second-order methods via low-dimensional updates. This thesis contributes novel insights, introduces new algorithms with improved convergence guarantees, and improves analyses of popular practical algorithms., Comment: Dissertation thesis

In this paper, we propose the first sketch-and-project Newton method with fast $\mathcal O(k^{-2})$ global convergence rate for self-concordant functions. Our method, SGN, can be viewed in three ways: i) as a sketch-and-project algorithm projecting updates of Newton method, ii) as a cubically regularized Newton ethod in sketched subspaces, and iii) as a damped Newton method in sketched subspaces. SGN inherits best of all three worlds: cheap iteration costs of sketch-and-project methods, state-of-the-art $\mathcal O(k^{-2})$ global convergence rate of full-rank Newton-like methods and the algorithm simplicity of damped Newton methods. Finally, we demonstrate its comparable empirical performance to baseline algorithms., Comment: 10 pages

Using CoOx thin films supported on Au(111) single crystal surfaces as model catalysts for the CO oxidation reaction we show that three reaction regimes exist in response to chemical and topographic restructuring of the CoOx catalyst as a function of reactant gas phase CO/O2 stoichiometry a finding that highlights the versatility of catalysts and their evolution in response to reaction conditions. Under oxygen-lean conditions and moderate temperatures (below 150C degrees) partially oxidized films containing CoO were found to be efficient catalysts. In contrast, stoichiometric CoO films containing only Co2+ form carbonates in the presence of CO that poison the reaction below 300 C degrees. Under oxygen-rich conditions a more oxidized catalyst phase forms containing Co3+ species that is effective in a wide temperature range. Resonant photoemission spectroscopy (ResPES) revealed the unique role of Co3+ sites in catalyzing the CO oxidation. DFT calculations provided deeper insights into the pathway and free energy barriers for the reactions on these oxide phases.

In this work, we consider the problem of minimizing the sum of Moreau envelopes of given functions, which has previously appeared in the context of meta-learning and personalized federated learning. In contrast to the existing theory that requires running subsolvers until a certain precision is reached, we only assume that a finite number of gradient steps is taken at each iteration. As a special case, our theory allows us to show the convergence of First-Order Model-Agnostic Meta-Learning (FO-MAML) to the vicinity of a solution of Moreau objective. We also study a more general family of first-order algorithms that can be viewed as a generalization of FO-MAML. Our main theoretical achievement is a theoretical improvement upon the inexact SGD framework. In particular, our perturbed-iterate analysis allows for tighter guarantees that improve the dependency on the problem's conditioning. In contrast to the related work on meta-learning, ours does not require any assumptions on the Hessian smoothness, and can leverage smoothness and convexity of the reformulation based on Moreau envelopes. Furthermore, to fill the gaps in the comparison of FO-MAML to the Implicit MAML (iMAML), we show that the objective of iMAML is neither smooth nor convex, implying that it has no convergence guarantees based on the existing theory.

In this paper, we present the first stepsize schedule for Newton method resulting in fast global and local convergence guarantees. In particular, a) we prove an $O\left( \frac 1 {k^2} \right)$ global rate, which matches the state-of-the-art global rate of cubically regularized Newton method of Polyak and Nesterov (2006) and of regularized Newton method of Mishchenko (2021) and Doikov and Nesterov (2021), b) we prove a local quadratic rate, which matches the best-known local rate of second-order methods, and c) our stepsize formula is simple, explicit, and does not require solving any subproblem. Our convergence proofs hold under affine-invariance assumptions closely related to the notion of self-concordance. Finally, our method has competitive performance when compared to existing baselines, which share the same fast global convergence guarantees.

The employees health and wellbeing are an actual topic in our fast-moving world. The employers losing money when their employees suffer from different health problems and cannot work. The major problem is the spinal pain caused by the poor sitting posture on the office chair. This paper deals with the proposal and realization of the system for the detection of incorrect sitting positions. The smart chair has six flexible force sensors. The Internet of Things (IoT) node based on Arduino connects these sensors into the system. The system detects wrong seating positions and notifies the users. In advance, we develop a mobile application to receive those notifications. The user gets feedback about sitting posture and additional statistical data. We defined simple rules for processing the sensor data for recognizing wrong sitting postures. The data from smart chairs are collecting by a private cloud solution from QNAP and are stored in the MongoDB database. We used the Node-RED application for whole logic implementation., Comment: 19 pages, 13 figures, 3 tables, article in journal

Chemical modifications such as intercalation can be used to modify surface properties or to further functionalize the surface states of topological insulators. Using ambient pressure X-ray photoelectron spectroscopy, we report copper migration in $\text{Cu}_{x}\text{Bi}_2\text{Se}_3$, which occurs on a timescale of hours to days after initial surface cleaving. The increase in near-surface copper proceeds along with the oxidation of the sample surface and large changes in the selenium content. These complex changes are further modelled with core-level spectroscopy simulations, which suggest a composition gradient near the surface which develops with oxygen exposure. Our results shed light on a new phenomenon that must be considered for intercalated topological insulators$\unicode{x2014}$and intercalated materials in general$\unicode{x2014}$that surface chemical composition can change when specimens are exposed to ambient conditions.

Despite their high computation and communication costs, Newton-type methods remain an appealing option for distributed training due to their robustness against ill-conditioned convex problems. In this work, we study ommunication compression and aggregation mechanisms for curvature information in order to reduce these costs while preserving theoretically superior local convergence guarantees. We prove that the recently developed class of three point compressors (3PC) of Richtarik et al. [2022] for gradient communication can be generalized to Hessian communication as well. This result opens up a wide variety of communication strategies, such as contractive compression} and lazy aggregation, available to our disposal to compress prohibitively costly curvature information. Moreover, we discovered several new 3PC mechanisms, such as adaptive thresholding and Bernoulli aggregation, which require reduced communication and occasional Hessian computations. Furthermore, we extend and analyze our approach to bidirectional communication compression and partial device participation setups to cater to the practical considerations of applications in federated learning. For all our methods, we derive fast condition-number-independent local linear and/or superlinear convergence rates. Finally, with extensive numerical evaluations on convex optimization problems, we illustrate that our designed schemes achieve state-of-the-art communication complexity compared to several key baselines using second-order information.

Near total reflection regime has been widely used in X-ray science, specifically in grazing incidence small angle X-ray scattering and in hard X-ray photoelectron spectroscopy. In this work, we introduce some practical aspects of using near total reflection in ambient pressure X-ray photoelectron spectroscopy and apply this technique to study chemical concentration gradients in a substrate/photoresist system. Experimental data are accompanied by X-ray optical and photoemission simulations to quantitatively probe the photoresist and the interface with the depth accuracy of ~1 nm. Together, our calculations and experiments confirm that near total reflection X-ray photoelectron spectroscopy is a suitable method to extract information from buried interfaces with highest depth-resolution, which can help address open research questions regarding our understanding of concentration profiles, electrical gradients, and charge transfer phenomena at such interfaces. The presented methodology is especially attractive for solid/liquid interface studies, since it provides all the strengths of a Bragg-reflection standing-wave spectroscopy without the need of an artificial multilayer mirror serving as a standing wave generator, thus dramatically simplifying the sample synthesis., Comment: 13 pages, 4 figures Supplemental Information

Using angle-resolved photoemission spectroscopy, we show the direct evidence of charge transfer between adsorbed molecules and metal substrate, i.e. chemisorption of CO on Pt(111) and Pt-Sn/Pt(111) 2x2 surfaces. The observed band structure shows a unique signature of charge transfer as CO atoms are adsorbed,revealing the roles of specific orbital characters participating in the chemisorption process. As the coverage of CO increases, the degree of charge transfer between CO and Pt shows clear difference to that of Pt-Sn. With comparison to DFT calculation results, the observed distinct features in the band structure are interpreted as backdonation bonding states of Pt molecular orbital to the 2{\pi} orbital of CO. Furthermore, the change in the surface charge concentration, measured from the Fermi surface area, shows Pt surface has a larger charge concentration change than Pt-Sn surface upon CO adsorption. The difference in the charge concentration change between Pt and Pt-Sn surfaces reflects the degree of electronic effects during CO adsorption on Pt-Sn., Comment: 15 pages, 9 figures, 2 tables

By adjusting the incidence angle of incoming X-ray near the critical angle of X-ray total reflection, the photoelectron intensity is strongly modulated due to the variation of X-ray penetration depth. Photoelectron spectroscopy (PES) combining with near total reflection (NTR) exhibit tunable surface sensitivity, providing depth-resolved information. In this review article, we first describe the experimental setup and specific data analysis process. We then review three different examples which show the broad application of this method. The emphasis is on its applications to correlated oxide heterostructures, especially quantitative depth analyses of compositions and electronic states. In a last part, we discussed the limitations of this technique, mostly in terms of range of samples which can be studied., Comment: 28 pages, 5 figures, review paper on NTR HAXPES

When a three-dimensional material is constructed by stacking different two-dimensional layers into an ordered structure, new and unique physical properties can emerge. An example is the delafossite PdCoO2, which consists of alternating layers of metallic Pd and Mott-insulating CoO2 sheets. To understand the nature of the electronic coupling between the layers that gives rise to the unique properties of PdCoO2, we revealed its layer-resolved electronic structure combining standing-wave X-ray photoemission spectroscopy and ab initio many-body calculations. Experimentally, we have decomposed the measured valence band spectrum into contributions from Pd and CoO2 layers. Computationally, we find that many-body interactions in Pd and CoO2 layers are highly different. Holes in the CoO2 layer interact strongly with charge-transfer excitons in the same layer, whereas holes in the Pd layer couple to plasmons in the Pd layer. Interestingly, we find that holes in states hybridized across both layers couple to both types of excitations (charge-transfer excitons or plasmons), with the intensity of photoemission satellites being proportional to the projection of the state onto a given layer. This establishes satellites as a sensitive probe for inter-layer hybridization. These findings pave the way towards a better understanding of complex many-electron interactions in layered quantum materials.

We propose ZeroSARAH -- a novel variant of the variance-reduced method SARAH (Nguyen et al., 2017) -- for minimizing the average of a large number of nonconvex functions $\frac{1}{n}\sum_{i=1}^{n}f_i(x)$. To the best of our knowledge, in this nonconvex finite-sum regime, all existing variance-reduced methods, including SARAH, SVRG, SAGA and their variants, need to compute the full gradient over all $n$ data samples at the initial point $x^0$, and then periodically compute the full gradient once every few iterations (for SVRG, SARAH and their variants). Note that SVRG, SAGA and their variants typically achieve weaker convergence results than variants of SARAH: $n^{2/3}/\epsilon^2$ vs. $n^{1/2}/\epsilon^2$. Thus we focus on the variant of SARAH. The proposed ZeroSARAH and its distributed variant D-ZeroSARAH are the \emph{first} variance-reduced algorithms which \emph{do not require any full gradient computations}, not even for the initial point. Moreover, for both standard and distributed settings, we show that ZeroSARAH and D-ZeroSARAH obtain new state-of-the-art convergence results, which can improve the previous best-known result (given by e.g., SPIDER, SARAH, and PAGE) in certain regimes. Avoiding any full gradient computations (which are time-consuming steps) is important in many applications as the number of data samples $n$ usually is very large. Especially in the distributed setting, periodic computation of full gradient over all data samples needs to periodically synchronize all clients/devices/machines, which may be impossible or unaffordable. Thus, we expect that ZeroSARAH/D-ZeroSARAH will have a practical impact in distributed and federated learning where full device participation is impractical., Comment: 28 pages

In this work, we consider the optimization formulation of personalized federated learning recently introduced by Hanzely and Richt\'arik (2020) which was shown to give an alternative explanation to the workings of local {\tt SGD} methods. Our first contribution is establishing the first lower bounds for this formulation, for both the communication complexity and the local oracle complexity. Our second contribution is the design of several optimal methods matching these lower bounds in almost all regimes. These are the first provably optimal methods for personalized federated learning. Our optimal methods include an accelerated variant of {\tt FedProx}, and an accelerated variance-reduced version of {\tt FedAvg}/Local {\tt SGD}. We demonstrate the practical superiority of our methods through extensive numerical experiments., Comment: NeurIPS 2020

Recent advances in the theoretical understanding of SGD led to a formula for the optimal batch size minimizing the number of effective data passes, i.e., the number of iterations times the batch size. However, this formula is of no practical value as it depends on the knowledge of the variance of the stochastic gradients evaluated at the optimum. In this paper we design a practical SGD method capable of learning the optimal batch size adaptively throughout its iterations for strongly convex and smooth functions. Our method does this provably, and in our experiments with synthetic and real data robustly exhibits nearly optimal behaviour; that is, it works as if the optimal batch size was known a-priori. Further, we generalize our method to several new batch strategies not considered in the literature before, including a sampling suitable for distributed implementations., Comment: Accepted to the 12th Annual Workshop on Optimization for Machine Learning (OPT2020)

The discovery of a two-dimensional electron system (2DES) at the interfaces of perovskite oxides such as LaAlO3 and SrTiO3 has motivated enormous efforts in engineering interfacial functionalities with this type of oxide heterostructures. However, its fundamental origins are still not understood, e.g. the microscopic mechanisms of coexisting interface conductivity and magnetism. Here we report a comprehensive spectroscopic investigation of the depth profile of 2DES-relevant Ti 3d interface carriers using depth- and element-specific techniques, standing-wave excited photoemission and resonant inelastic scattering. We found that one type of Ti 3d interface carriers, which give rise to the 2DES are located within 3 unit cells from the n-type interface in the SrTiO3 layer. Unexpectedly, another type of interface carriers, which are polarity-induced Ti-on-Al antisite defects, reside in the first 3 unit cells of the opposing LaAlO3 layer (~10 {\AA}). Our findings provide a microscopic picture of how the localized and mobile Ti 3d interface carriers distribute across the interface and suggest that the 2DES and 2D magnetism at the LaAlO3/SrTiO3 interface have disparate explanations as originating from different types of interface carriers., Comment: 28 pages, 4 figures; Supplemental Materials 19 pages

Lutetium dodecaboride LuB12 is a simple weak-coupling BCS superconductor with critical temperature Tc = 0.42 K, whilst ZrB12 is a strong-coupling BCS superconductor with the highest critical temperature Tc = 6.0 K among this group of materials. In case of lutetium substitution by zirconium ions in LuB12 the crossover from weak- to strong-coupling superconductor can be studied. We have investigated the evolution of critical temperature Tc and critical field Hc in high-quality single crystalline superconducting samples of Lu(1-x)Zr(x)B12 (0 =< x =< 0.45) by measuring magnetic ac susceptibility between 1 K and 50 mK. To obtain this kind of experimental data, a new susceptometer was designed, constructed and tested, which can work in a wide temperature range of 0.05 K - 3 K in 3He-4He dilution refrigerator. The measurements with this new susceptometer revealed how Tc(x) and Hc(x) increases with increasing concentration of zirconium in Lu(1-x)Zr(x)B12 solid solutions as well as how their superconducting phase diagram develops., Comment: 6 pages, 4 figures, CSMAG conference contribution

Based on accurate X-ray structure analysis of GdB6 over the temperature range 85-300 K it has been shown that anomalously strong charge carrier scattering in the quantum diffusion regime of charge transport in this compound arises due to the formation of (i) dynamically coupled Gd3+ dimers of about 3.3 {\AA} in size with an energy of quasi-local oscillations ~ 7 - 8 meV and (ii) dynamic charge stripes along the [001] direction of the cubic lattice. It has been revealed that anharmonic approximation is useful when analyzing the static and dynamic components of the atomic displacement parameters of gadolinium. The barrier height of double-well potential of Gd3+ ions was determined both from low-temperature heat capacity measurements and from the electron density distribution reconstructed from X-ray data., Comment: 13 pages, 5 figures

Structure differences of isotopically different dodecaborides LuNB12 (N = 10, 11, natural) and their impact on thermal and charge transport characteristics of the crystals have been first discovered. Atomic displacement parameters (ADPs) of Lu and B atoms are described in terms of the Einstein and Debye models, respectively. Characteristic Einstein and Debye temperatures are calculated directly from the x-ray data and corresponding ADPs are separated into temperature dependent and temperature independent components. The first component is a measure of thermal atomic vibrations whereas the second one is a sum of zero vibrations and static shifts of some atoms from their crystallographic positions. Such a local disordering is more expressed in LunatB12 with 10B : 11B = 1 : 4 judging both from the large static ADP components and the Schottky anomalies in the heat capacity. Crystal structures are refined in Fm3-m group but certain distortions of the ideal cubic unit-cell values are observed in all three crystals under study due to cooperative Jahn-Teller effect. The distortions are mainly trigonal or mainly tetragonal depending on the isotope composition. Low symmetry distribution of electron density reveals itself in LunatB12 in the form of the dynamic charge stripes oriented in selected directions close to some of <110>. The large static ADP components of LunatB12 are surprisingly combined with high conductivity of the crystal. One may suppose the static shifts (defects) are centers of pinning facilitating formation of additional conductive channels., Comment: 20 pages, 9 figures, 4 Tables

X-ray standing-wave photoemission experiments involving multilayered samples are emerging as unique probes of the buried interfaces that are ubiquitous in current device and materials research. Such data require for their analysis a structure optimization process comparing experiment to theory that is not straightforward. In this work, we present a new computer program for optimizing the analysis of standing-wave data, called SWOPT, that automates this trial-and-error optimization process. The program includes an algorithm that has been developed for computationally expensive problems: so-called black-box simulation optimizations. It also includes a more efficient version of the Yang X-ray Optics Program (YXRO) [Yang, S.-H., Gray, A.X., Kaiser, A.M., Mun, B.S., Sell, B.C., Kortright, J.B., Fadley, C.S., J. Appl. Phys. 113, 1 (2013)] which is about an order of magnitude faster than the original version. Human interaction is not required during optimization. We tested our optimization algorithm on real and hypothetical problems and show that it finds better solutions significantly faster than a random search approach. The total optimization time ranges, depending on the sample structure, from minutes to a few hours on a modern laptop computer, and can be up to 100x faster than a corresponding manual optimization. These speeds make the SWOPT program a valuable tool for realtime analyses of data during synchrotron experiments.

LaCrO$_3$ (LCO) / SrTiO$_3$ (STO) heterojunctions are intriguing due to a polar discontinuity along (001), two distinct and controllable interface structures [(LaO)$^+$/(TiO$_2$)$^0$ and (SrO)$^0$/(CrO$_2$)$^-$], and interface-induced polarization. In this study, we have used soft- and hard x-ray standing-wave excited photoemission spectroscopy (SW-XPS) to generate a quantitative determination of the elemental depth profiles and interface properties, band alignments, and the depth distribution of the interface-induced built-in potentials in the two constituent oxides. We observe an alternating charged interface configuration: a positively charged (LaO)$^+$/(TiO$_2$)$^0$ intermediate layer at the LCO$_\textbf{top}$/STO$_\textbf{bottom}$ interface and a negatively charged (SrO)$^0$/(CrO$_2$)$^-$ intermediate layer at the STO$_\textbf{top}$/LCO$_\textbf{bottom}$ interface. Using core-level SW data, we have determined the depth distribution of species, including through the interfaces, and these results are in excellent agreement with scanning transmission electron microscopy and electron energy loss spectroscopy (STEM-EELS) mapping of local structure and composition. SW-XPS also enabled deconvolution of the LCO-contributed and STO- contributed matrix-element-weighted density of states (MEWDOSs) from the valence band (VB) spectra for the LCO/STO superlattice (SL). Monitoring the VB edges of the deconvoluted MEWDOS shifts with a change in probing profile, the alternating charge- induced built-in potentials are observed in both constituent oxides. Finally, using a two-step simulation approach involving first core-level binding energy shifts and then valence-band modeling, the built-in potential gradients across the SL are resolved in detail and represented by the depth distribution of VB edges., Comment: Main text: 29 pages, 5 figures; Supplementary Information: 20 pages, 10 figures

Recent developments in environmental and liquid cells equipped with electron transparent graphene windows have enabled traditional surface science spectromicroscopy tools, such as X-ray photoelectron spectroscopy (XPS), photoemission electron microscopy (PEEM), and scanning electron microscopy (SEM) to be applied to study solid-liquid and liquid-gas interfaces. Here, we focus on the experimental implementation of PEEM to probe electrified graphene-liquid interfaces using electrolyte-filled microchannel arrays as a new sample platform. We demonstrate the important methodological advantage of these multi-sample arrays: they enable the combination of the wide field of view hyperspectral imaging capabilities from PEEM with the use of powerful data mining algorithms to reveal spectroscopic and temporal behaviors at the level of the individual microsample or the entire array ensemble, Comment: 14 pages 6 figures 47 references

The dilute magnetic semiconductors have promise in spin-based electronics applications due to their potential for ferromagnetic order at room temperature, and various unique switching and spin-dependent conductivity properties. However, the precise mechanism by which the transition-metal doping produces ferromagnetism has been controversial. Here we have studied a dilute magnetic semiconductor (5% manganese-doped gallium arsenide) with Bragg-reflection standing-wave hard X-ray angle-resolved photoemission spectroscopy, and resolved its electronic structure into element- and momentum- resolved components. The measured valence band intensities have been projected into element-resolved components using analogous energy scans of Ga 3d, Mn 2p, and As 3d core levels, with results in excellent agreement with element-projected Bloch spectral functions and clarification of the electronic structure of this prototypical material. This technique should be broadly applicable to other multi-element materials.

We investigated the metal-insulator transition for epitaxial thin films of the perovskite CaFeO3, a material with a significant oxygen ligand hole contribution to its electronic structure. We find that biaxial tensile and compressive strain suppress the metal-insulator transition temperature. By combining hard X-ray photoelectron spectroscopy, soft X-ray absorption spectroscopy, and density functional calculations, we resolve the element-specific changes to the electronic structure across the metal-insulator transition. We demonstrate that the Fe electron valence undergoes no observable change between the metallic and insulating states, whereas the O electronic configuration undergoes significant changes. This strongly supports the bond-disproportionation model of the metal-insulator transition for CaFeO3 and highlights the importance of ligand holes in its electronic structure. By sensitively measuring the ligand hole density, however, we find that it increases by ~5-10% in the insulating state, which we ascribe to a further localization of electron charge on the Fe sites. These results provide detailed insight into the metal-insulator transition of negative charge transfer compounds and should prove instructive for understanding metal-insulator transitions in other late transition metal compounds such as the nickelates., Comment: Minor typographic changes made

A major remaining challenge in the superconducting cuprates is the unambiguous differentiation of the composition and electronic structure of the CuO$_2$ layers and those of the intermediate layers. The large c axis for these materials permits employing soft x-ray (930.3 eV) standing wave (SW) excitation in photoemission that yields atomic layer-by-atomic layer depth resolution of these properties. Applying SW photoemission to Bi$_2$Sr$_2$CaCu$_2$O$_{8+\delta}$ yields the depth distribution of atomic composition and the layer-resolved densities of states. We detect significant Ca presence in the SrO layers and oxygen bonding to three different cations. The layer-resolved valence electronic structure is found to be strongly influenced by the supermodulation structure--as determined by comparison to DFT calculations, by Ca-Sr intermixing, and by the Cu 3d-3d Coulomb interaction, further clarifying the complex interactions in this prototypical cuprate. Measurements of this type for other quasi-two-dimensional materials with large-c represent a promising future direction., Comment: 22 pages, 5 figures, plus Supplemental Material (15 pages)

We have investigated the electronic structure of the dilute magnetic semiconductor (DMS) $Ga_{0.98}Mn_{0.02}P$ and compared it to that of an undoped $GaP$ reference sample, using hard X-ray photoelectron spectroscopy (HXPS) and hard X-ray angle-resolved photoemission spectroscopy (HARPES) at energies of about 3 keV. We present experimental data, as well as theoretical calculations, in order to understand the role of the Mn dopant in the emergence of ferromagnetism in this material. Both core-level spectra and angle-resolved or angle-integrated valence spectra are discussed. In particular, the HARPES experimental data are compared to free-electron final-state model calculations and to more accurate one-step photoemission theory. The experimental results show differences between $Ga_{0.98}Mn_{0.02}P$ and $GaP$ in both angle-resolved and angle-integrated valence spectra. The $Ga_{0.98}Mn_{0.02}P$ bands are broadened due to the presence of Mn impurities that disturb the long-range translational order of the host $GaP$ crystal. Mn-induced changes of the electronic structure are observed over the entire valence band range, including the presence of a distinct impurity band close to the valence-band maximum of the DMS. These experimental results are in good agreement with the one-step photoemission calculations, and a prior HARPES study of $Ga_{0.97}Mn_{0.03}As$ and $GaAs$ (Gray et al. Nature Materials 11, 957 (2012)), demonstrating the strong similarity between these two materials. The Mn 2p and 3s core-level spectra also reveal an essentially identical state in doping both $GaAs$ and $GaP$., Comment: 37 pages (double spaced) 10 figures

Functionalization of polyoxotungstates with organoarsonate co-ligands enabling surface decoration was explored for the triangular cluster architectures of the composition [CoII9(H2O)6(OH)3(p-RC6H4AsVO3)2({\alpha}-PV2WVI15O56)3]25-({Co9(P2W15)3}, R = H or NH2), isolated as Na25[Co9(OH)3(H2O)6(C6H5AsO3)2(P2W15O56)3]86H2O (Na-1) and Na25[Co9(OH)3(H2O)6(H2NC6H4AsO3)2(P2W15O56)3]86H2O (Na-2). The axially oriented para-aminophenyl groups in 2 facilitate the formation of self-assembled monolayers on gold surfaces, and thus provide a viable molecular platform for charge transport studies of magnetically functionalized polyoxometalates. The title systems were isolated and characterized in the solid state and in aqueous solutions, and on metal surfaces. Using conducting tip atomic force microscopy (C-AFM), the energies of {Co9(P2W15)3} frontier molecular orbitals in the surface-bound state were found to directly correlate with cyclic voltammetry data in aqueous solution., Comment: Full paper with figures and supporting information