New or optimized multifunctional and structural intermetallic materials are needed, e.g. for miniaturization of technological products with improved functionality even in extreme conditions or for efficient production/storage/conversion of energy carriers. For the discovery and optimization of new materials combinatorial and high-throughput experimentation methods are very effective. The materials to be investigated are deposited in the form of materials libraries by special magnetron sputter deposition methods (co-deposition, wedge-type multilayer deposition, GLAD, shadow masking). These materials libraries are subsequently processed (heat treatment in inert or reactive environment, dealloying) and characterized by high-throughput experimentation methods (automated EDX, XRD, temperature-dependent resistance and stress screening) in order to relate compositional information with structural and functional properties. The talk will cover examples of the combinatorial development of Ni- and Fe-based intermetallic materials for shape memory (Ni-Ti-X-Y, Fe-Pd-X) and other applications. Furthermore examples from nanostructured oxide systems (e.g. W-Fe-O) for solar water splitting will be discussed. The obtained results are visualized in the form of composition-structure-function diagrams.

The development of a diversity of materials with the desired functionalities and performance from a limited number of elements provided by nature has always been the challenge for chemists and materials scientists. To this end the organic chemists have done very well - for they have crafted numerous organic molecules from basic elements such as carbon, hydrogen and oxygen by changing the number and the type of atoms and their spatial relationship. Taking a lesson from such an approach, we developed a synthesis strategy capable of engineering the architecture of heterogeneous metallic nanocrystals (HMNCs) through rational and independent programming of every architecture-determining element of a HMNC, i.e. the shape and size of the component nanocrystals (NCs) and their spatial arrangement. Architectural engineering of HMNCs can generate metallic nanostructures with unprecedented diversity and structural complexity, and hence increase significantly the possibilities to create new and more varied properties for application explorations.
Our synthesis strategy takes advantage of the geometry dependent distribution of electrons in a chemical reaction system to direct the deposition of satellite NCs on specific regions of a polyhedral central NC. What guided the electron distribution is the curvature of the polyhedral NC surface, the defining feature of any polyhedron. Consequently the strategy can be used to locate satellite NCs exclusively on the corners and edges of different polyhedral central NCs. The satellite NCs can develop their own shape and size through control of crystal growth. Furthermore, this strategy can produce HMNCs consisting of two or more types of component NCs, and of different metals in various combinations.

Recently, newly high-pressure vessel was designed and fabricated in order to study high-throughput materials exploration under high-pressure and high-temperature. And, the combinatorial high-pressure vessel and process based on the hot isostatic pressing (HIP) method and succeeded multiple high-pressure experiments under 200 MPa and 500° C, simultaniously.
In this study, spinel-type MgAl₂O₄ and ZnFe₂O₄ were prepared using the combinatorial high-pressure vessel for studying solid state reactivity under 200 MPa. Starting materials used were Mg(OH)₂, AlOOH, Al(OH)₃, FeOOH and ZnO. Single phase of spinel-type MgAl₂O₄ and ZnFe₂O₄ were observed at 500° C and 300° C, respectively. These reaction were corresponded to resultes of conventional hot isostatic pressing (HIP) experiments. And, it was found that starting materials having hydroxyl group played a significant role in preparation under high-pressure.
As the above results, high-pressure reaction makes it possible to prepare stable state at high-temperature under lower temperature condition. Therefore, usage of the combinatorial high-pressure vessel and process is promising to find newly candidate magnetic materials.

Over the last few decades, the tools of materials research have become significantly more sophisticated, allowing for the rapid synthesis and characterization of large numbers of samples. As a result, materials researchers can now collect sample characterization data faster than they can analyze it. This disparity in data collection and analysis time is fueling interest in new machine learning algorithms, also known as data-mining techniques, for accelerating data processing. In this talk we will discuss a set of algorithms that can be utilized to quickly sort data from combinatorial libraries spanning large composition phase spaces. We will also show how crystal structure data from crystallographic databases can be used to improve the performance of these algorithms. This talk will focus on X-ray diffraction data obtained from Fe-Ga-Pd, Fe-Co-Ni, Fe-Co-Mo, and Fe-Co-W thin-film ternary composition spreads.

New X-ray free-electron laser (XFEL) sources that create X-ray pulses of unprecedented brilliance open up new possibilities for the structural characterization of crystalline materials. By exposing a small crystallite (from nano- to a few micrometers in size) to a single ultrafast pulse, a diffraction pattern can be obtained before the crystal is damaged. If such single-pulse diffraction patterns, collected sequentially on many randomly oriented crystallites, are combined, it is possible to determine the structure of the material accurately [1]. One of the drawbacks of this approach is that only a single position of the Ewald sphere is accessed in each pattern, so, because reflections have a finite width, the diffraction condition is not satisfied completely for any of the reflections recorded. The new XFEL source (SwissFEL), which is currently being developed at PSI (Switzerland), will provide a broad-bandpass mode with an energy bandwidth of about 4% [2]. By using the full energy range of the SwissFEL beam, a new option for structural studies of crystalline materials may become possible. The use of such an ‘extra pink&’ beam in a diffraction experiment with stationary crystallites should not only increase the number of reflection intensities that can be collected in a single shot, but also overcome the problem of ‘partial reflection&’ measurement that is inherent to the monochromatic experiment. In order to take advantage of the full SwissFEL beam for crystallographic studies, we propose a new approach, inspired by the Laue single-crystal (micro)diffraction technique and the experimental setup on BL12.3.2 at the Advanced Light Source. Diffraction patterns for 100 randomly oriented stationary crystallites of the MFI-type zeolite ZSM-5 were simulated assuming several energy bandwidth values and 2 detector positions. These necessarily sparse diffraction patterns could be indexed using a pattern recognition algorithm. The number of accessible reflections that are not affected by the partial reflection intensity problem increases significantly with bandwidth. On average with a 4% (0.5%) bandwidth, there are 140 (17) reflections per pattern with a 2D 1M Pilatus detector positioned at 90° (2theta;) relative to the incident beam and 46 (6) reflections with the detector at 45° (2theta;). Structure solution using the reflections from these patterns was performed using both direct methods and a dual-space method [3]. Our prime interest is in the area of inorganic and small-molecule structures, where the diffraction patterns are sparse, but this new approach could also be of benefit to the protein community. We believe that the ‘extra pink&’ beam mode option offers a clear opportunity to ease the data acquisition and subsequent evaluation in femtosecond time-resolved experiments at an XFEL facility.
[1] Chapman H. N. et al. Nature, 470, 73 (2011).
[2] Patterson B. D. et al. New J. Phys. 12, 035012 (2010).
[3] Dejoie C., et al. J. Appl. Cryst. 46, 791 (2013).

Accurate characterization of nanoparticle catalysts and electrocatalysts is crucial for understanding the fundamental relationship between their structure and catalytic function. The essential key required for achieving this goal is the development of more sophisticated structural tools and methods. One method that is particularly amenable to combined theoretical and experimental studies is extended X-ray absorption fine structure (EXAFS) spectroscopy, and particularly its application to the analysis of small metallic nanoparticles. One limitation of this technique is that it provides averaged information about the structure and dynamics of the nearest environment of all absorbing atoms in the nanoparticle ensemble. This can lead to erroneous structural determinations. Because of this, advanced techniques for EXAFS analysis beyond the traditional fitting procedures are a topic of great interest to researchers. Here we present a methodology for analysis of nanoparticles by EXAFS which provides much more detailed structural information by validation of density functional theory molecular dynamics (DFT-MD) simulations by comparing experimentally and theoretically produced EXAFS signals. Varying amounts of thiol were added to the surface of dendrimer-encapsulated Au nanoparticles with an average of 147 atoms in order to systematically tune the nanoparticle disorder. Additionally, an analogous system was investigated with DFT-MD simulations, and these simulations were used to produce theoretical EXAFS signals that were used for comparison to experimental data. By validating the theoretical model in this way we can infer previously unknown details of structure and dynamics of the nanoparticles. This study demonstrates the benefit of a combined experimental/theoretical approach over standard EXAFS fitting techniques for determining the structural parameters of metallic nanoparticle catalysts, and shows that DFT-MD simulations accurately depict complex experimental systems in which we have control over nanoparticle disorder.

The anatase structure of titanium dioxide displays appealing properties as negative electrode in lithium ion batteries. Several approaches have been used in order to improve the performance of this compound including downsizing the solid particle and chemical doping. For the latter, fluorine has been widely used and very recently, our group has developed a new chemical method to prepare a highly fluorinated compound. The charge deficiency induced by the substitution of oxygen by fluorine is compensated by the formation of a cation vacancy every four substitutions. The fluorine content deduced from elemental analysis and synchrotron diffraction revealed the existence of more than 20% cation vacancies. The local environment of fluorine has been probed by ¹⁹F NMR spectroscopy. Three different coordination modes for the fluorine have been detected.
The structural complexity of this compound led us to use a mixed ab initio and classical modeling and simulation approach. After deriving a classical polarizable force-field from DFT simulations, at the target concentration we screen in a step-wise fashion a large number of possible configurations differing in the positioning of the vacancies and of the F atoms. After each step only the 10% lowest energy configurations are retained for the following step. Using the polarizable classical force-field, we start from the evaluation of the energy of ~1.5 × 10⁵ configurations and then perform 0 K geometry optimizations and 0 K cell optimizations. Finally we temper the configurations from 0 K to 300 K through NVT molecular dynamics simulations. The surviving configurations are retained for further analyses including NPT molecular dynamics simulations and ab initio DFT cell optimizations from which the salient structural features are extracted. The relative abundance of the different environments for the F is also assessed as a function of the screening steps.
On one hand the method allows to predict the most favorable structural arrangements of the material, on the other it permits to discern structural characteristics that can be hard to disentangle in experimental techniques.

We describe a suite of high-throughput experimental tools for the characterization of phase transitions in combinatorial thin film libraries for which the primary component is VO₂. Thin film libraries are prepared by pulsed laser deposition (PLD) on silicon substrates, from chemically distinct ceramic targets, e.g. V₂O₅, Ta₂O₅, WO₃, etc. The sample libraries have a continuously varying composition, measured locally by x-ray photoelectron spectroscopy (XPS) with 1.5 mm² spot size.
The metal-to-insulator phase transition at 68°C in VO₂ is accompanied by a strong infrared (IR) thermochromism, i.e. a high IR reflectivity in the high temperature state and low IR reflectivity in the low temperature state. Depression of the transition temperature by the incorporation of transition metal dopants would enable use as the functional component in ‘smart&’ building windows that selectively admit solar radiation based on ambient temperature.
We have assembled a high-throughput tool for the characterization of thermochromic phase transitions in combinatorial libraries, using the thermochromism as an indicator of the crystalline phase transition. The measurement protocol involves making many local measurements of infrared reflectance at several hundred sample locations (i.e., film compositions) and at many temperatures in the range of 15°C to 85°C.
In this way, we build up a high quality, high density dataset of phase transition temperatures in crystalline VO₂ with varying concentrations of Ta, Nb, W, and other transition metals. Self-consistency of the data is assured by the combinatorial synthesis and measurement protocols. Further, our measured transition temperatures were found to be consistent with literature results from traditional ‘one-off&’ experiments. We find that these impurities depress the VO₂ transition temperature at efficiencies of 3 °C/at % Nb, 7°C/at % Ta, and 11°C/at. % (W+Nb). Thus, smart window coatings that can transition at 25°C are possible with, for example, <4 at. % (W+Nb) doping.

Huge efforts have been directed towards the fabrication of ultra microelectrodes (UMEs) in recent years for diverse applications. UMEs are very small and provide access to electrochemical experiments previously considered impossible with conventional electrodes including measurements in highly resistive media (nonpolar solvents, polymers, gaseous interfaces etc), high-speed voltammetry and analyses in small volumes warrants the utilization of UMEs. In the past few decades, the study of SERS has been highly developed as a result of the progress of nanoscience and nanotechnology. Deposition and synthesis of nanoparticles of the coinage metals to form SERS active substrates on a macroelectrode has been in vogue for a long time now, but till date the use of UMEs in electrochemical potential dependent surface-enhanced Raman spectroscopy (SERS) to study facile surface reactions in the interface has not been reported but is essential to study the fate of various chemical species at the interface. We have successfully fabricated novel gold thin film based UME arrays by sequential, layer by layer deposition of Au (electrode material) and silica (for insulation) on a thin adhesion layer of Chromium thin films thermal evaporation and chemical vapor deposition techniques in combination on Si glass substrates. The much-needed roughened surface for plasmonic enhancement was obtained by focused ion beam lithography. The layered structure apart from providing durability the resulting UME also provides considerable surface area for the surface reactions to be effectively studied. Iron based UMEs were on the other hand fabricated using a sputtering technique and the plasmonic surface made by deposition of nanoparticles and subsequently depositing a thin film on it to protect the surface. This is known as a film over nanosphere method to produce stable substrates, which can be used repeatedly. These conductive thin film based UMEs were characterized to be electrochemically active by cyclic voltammetry in non aqueous media and then applied to study potential dependent surface enhanced Raman spectroscopy (SERS). The structural differences of adsorbed molecular species at various potentials on the electrode in non-aqueous media were successfully studied by SERS, which is arguably one of the most sensitive techniques to identify sub monolayer concentrations of Raman active molecules. Comparison of the structures of the same molecular species adsorbed on the two different UMEs at various potentials shows the difference of reactivities of the same molecule towards different metals. Extending the scope of this technique to study surface reactions of corrosive of molecules in crude oil on the surface of iron provides an insight into the mechanism of corrosion taking place in crude oil. This useful application will help in developing effective mitigation techniques of petroleum refinery corrosion, which currently results in a huge financial burden in the refineries.

We address via integration of predictive theory and experiment three of the currently leading obstacles to discovery of technologically important functional materials.
Obstacle 1 is that whereas artificial materials can be made in the laboratory via layer-by-layer (superlattice) growth, we don&’t know a priori which of the countless number of possible layer sequences and ensuing mesoscopic structures has the material functionality we need. We address this bottleneck by genetic algorithm search of the configuration with given calculated electronic property (“Modality 1 inverse design”). This is illustrated for identifying the layer sequence of InAs/GaSb with target IR absorption and the Si-Ge structure with direct, allowed band gap.
Obstacle 2 is that we do not know even the basic materials properties of thousands of materials with known crystal structures documented in the inorganic literature, making it difficult to select suitable candidates for material-specific technologies. We address this bottleneck by theoretically developing computable metrics associated with particular target functionalities and screening computationally numerous materials, identifying a small number of “best of class” subjected to iterative experiment-theory scrutiny (“Modality 2 Inverse Design”). This is illustrated for newly discovered super strong PV absorbers and novel transparent conductors.
Obstacle 3 is that there are thousands of compounds that are missing altogether from current material compilations. We address this bottleneck by using functionality-directed first-principles thermodynamics to identify stable “missing compounds” (“modality 3 Inverse design”). This is illustrated for laboratory realization of our predicted, but previously unreported ABX compounds, including TaCoSn, and the new transparent hole conductor TaIrGe.
* This work was done as part of the Center of Inverse Design collaboration whose five nodes are University of Colorado, Boulder; National Renewable Energy laboratory, Golden Colorado Northwestern University, Evanston Illinois, Oregon State University, Corvallis, Oregon, and5Stanford Linear Accelerator, Stanford, California.

We are developing techniques to effectively integrate combinatorial experimentation with high-throughput computational approaches. Theory guided combinatorial experiments can be particularly effective, but data from combinatorial libraries can also be used to help guide computations. I will address the experimental side of this integrated approach. As a key component, techniques to rapidly analyze a large amount of data from combinatorial libraries and methods to interface the combinatorial data with databases are being developed. Some data formats are particularly challenging because of their spectral or higher dimensional nature. We map the structural properties across libraries and quickly cross-reference them with known phases in crystallographic databases. The need for such a study is common among combinatorial investigation of virtually all topics. Automated scanning X-ray diffraction across entire libraries is carried out at a synchrotron beamline as well as with in-house diffractometers. We have developed a dynamic experimental/data analysis platform, where, as synchrotron diffraction data are being captured, they are processed in real time and quickly analyzed together with entries in the Inorganic Crystal Structure Database in order to construct a structural phase diagram. Examples of materials topics include rare-earth free permanent magnets and superconductors. This work is carried out in collaboration with A. G. Kusne, K. Jin, D. Kan, T. Gao, A. Mehta, M. Kramer, K. Rabe, and S. Curtarolo. This work is funded by NIST, ONR, ARO, AFOSR, and NSF.

In spirit of Materials Genome Initiative, the Materials Project (http://www.materialsproject.org) aims to leverage the information age for materials design. The goal of the Materials Project is to accelerate materials discovery and education through advanced scientific computing and innovative design methods, scale those computations to inorganic compounds and beyond, and disseminate that information and design tools to the larger materials community. The Project was launched online in October 2011 and since then we have computed and imported > 30,000 inorganic compounds into the database, which can now be freely accessed and searched over through the web interface for structural properties, local environments and coordination, XRD, electronic structure, Li-ion electrode properties, reaction energies and more. In this talk we will highlight some of the challenges and benefits in creating an open materials design environment; with example of how the data can be used to access phase stability across different chemical systems, combining data from both experiments and computations, combining different methodologies and monitoring data accuracy in high-throughput mode.

One of the prerequisites that enable accelerated materials innovation is the existence of predictive theoretical approaches and computational tools that are able to address quantitatively a range of properties of materials. This includes both predicting properties relevant for a specific application, e.g. predicting position of semiconductor band edges relative to the water oxidation and reduction levels for application in water splitting; as well as predicting materials thermodynamics and thermochemistry that provides insights into the possible growth conditions and the thermodynamic stability of materials under operating conditions. Thanks to both theoretical developments and the availability of massively parallel computers we are now able to compute a range of properties of real materials with quantitative accuracy and at reasonable computational cost, which, in return, enables searching large chemical spaces for new material solutions as well as finding ways to improve the performance of existing materials in a wide range of applications.
In this talk I will describe two such theoretical approaches recently developed at NREL that allow for quantitative predictions of: (1) ionization potentials (IPs) and electron affinities (EAs) of semiconductors and insulators, and (2) compound enthalpies of formation. In the first case our computational approach utilizes accurate many-body GW calculations for the electronic structure of bulk materials in combination with density functional theorysurface calculations (in the GGA approximation), resulting in accurate, and surface orientation dependent IPs and EAs. The power of our approach is in its broad applicability across the periodic table, also to, for ab-initio methods rather challenging, transition metal compounds. As one important application I will describe how calculated IPs and EAs can be used in searching for new water splitting materials. In the second part I will describe our recent approach for computing accurate enthalpies of formation of compounds, named FERE, that is based on GGA+U total energies for compounds and fitted elemental-phase reference energies (FERE) for the elemental phases. FERE energies are fitted to a large set of measured enthalpies of formation and serve to correct GGA+U for its inconsistent treatment of chemically different systems (e.g. solid vs. gaseous), thereby resulting in enthalpies of formation for insulating and semiconducting solids calculated close to chemical accuracy. FERE is currently the most accurate tool for computing compound enthalpies of formation that is at the same time applicable to compounds spanning large portion of the periodic table and is also computationally inexpensive and suitable for application in a high-throughput fashion. At the end I will discuss application of the FERE method as the starting point for predicting materials thermodynamic and thermochemistry as well as predicting existence of new compounds.

Ionization potentials and band offsets are important in efficient screening of candidates for new development or improvement of semiconductor heterostructure devices. However, before starting exhaustive, or high-throughput, searches on a large number of systems, such as using descriptor-based approaches, we definitely need to obtain accurate ionization potentials and band offsets that can be used as a “training set”, and figure out the intrinsic margin of error that appears in the calculations of individual systems.
We use chalcopyrite semiconductors such as CuInSe2 (CIS) and CuGaSe2 (CGS), as systems to be investigated in depth with first principles calculations based on the hybrid density functional theory. The non-polar (110) surfaces of CIS and CGS are known to stabilize when (112) and (11-2) facets form together with ordered point defects. [1,2] Such reconstruction of the surfaces would inevitably affect the ionization potentials, and indeed the ionization potentials of the (112) and (11-2) facets of CIS and CGS can change by 0.4 and 0.5 eV, respectively, depending on the surface defects.
The band offsets depend on the atomic configurations at the interface as well as whether the constituent phases are strained or unstrained (natural). The applicability of transitivity of natural band offsets, which means that the band offset between phases A and B and the band offset between phases B and C can be summed to derive the band offset of phases A and C, is crucial for high-throughput calculations, and therefore, the discrepancy between a set of transitive band offsets and the actual, non-transitive band offsets needs to be examined. These issues will be discussed using the (110) interfaces of chalcopyrite and zincblende semiconductors.
[1]. J. E. Jaffe and A. Zunger, Phys. Rev. B 64, 241304 (2001).
[2] Y. Hinuma et al., Phys. Rev. B 86, 245433 (2012).

Access to reliable thermodynamic data is a key component of materials discovery, where density functional theory (DFT), often used with the generalized gradient approximation (GGA) to the exchange-correlation functional, serves as the standard first-principles tool. However, due to its tendency to over-delocalize electrons, GGA fails to describe certain inorganic materials with strong electron correlations, such as first-row transition metal oxides (TMOs). Modifying GGA with an on-site Hubbard-U term is a common and computationally feasible remedy that forces correlated electrons to localize, but it also requires the parameter U to be specified a priori. We develop a framework for predicting valence and ligand dependent U values for GGA+U based thermochemical calculations by formulating a relation between U and experimental reaction energies. Applying this method, we calculate such material-specific U values for binary oxides and fluorides of multivalent metals M = Ti, V, Cr, Mn, Fe, Co and Ni, and also provide total energy compensation factors that establish compatibility among phases of M treated with different U values. We validate the transferability of U values and energy compensations from binary compounds with certain oxidation states and ligands to other compounds with similar environments by showing that experimental formation enthalpies of more than 50 ternary TMOs are reproduced with a remarkably small mean absolute error of ~19 meV/atom. We further present potential applications of the method in calculating redox potentials of Li-ion batteries and phase stabilities in mixed-valence TMO systems. Our method can be readily implemented in high-throughput databases for first-principles materials thermodynamics.

The Materials Genome Initiative relies extensively on high-throughput ab-initio computing to determine the basic properties of materials. High-throughput approaches have been very successful for properties such as energy, battery voltage, band structure, Seebeck coefficient. These are mostly properties that require a single, or a small number of sequential tasks and calculations. In this presentation I will discuss the extension of high-throughput computing to more difficult properties such as ionic diffusion. Methods such as the Nudged Elastic Band Method to determine activation energies are often difficult to converge, making their high-throughput application difficult. Ab initio molecular dynamics (MD) on the other hand is computing intensive and costly to perform on a large number of compounds. As part of a search for fast ion conductors we have developed a statistical MD-based approach to evaluate the probability that a compound displays fast diffusion for lithium. I will show some success of this approach in identifying fast ionic conductors.

Accelerating materials innovation is important in all areas of modern technology, but it is particularly pressing in the field of renewable energy. Photovoltaic (PV) devices, also known as solar cells, is one of the most promising sustainable energy generation technologies. In this talk, I will discuss the progress towards development of a three-step combinatorial strategy for accelerating materials innovation in the field of inorganic thin film solar cells and exemplify this strategy by recent results of combinatorial materials and device development.
The starting point of accelerated development of solar cell absorbers is combinatorial screening of the candidate materials and alloys for their ability to (1) absorb sunlight and (2) facilitate extraction of charge carriers. In this areas, I will summarize the results of recent materials design in (a) enhancement of optical absorption and electrical transport in Cu2O through combinatorial alloying with ZnO and Cu2Se, (b) improvement of electrical charge transport properties of Cu3N by combinatorial temperature and pressure gradient control, and (c) optimization of both optical and electrical properties in ternary Cu-M-S materials (M = Sn, Sb) using various combinatorial strategies.
Beyond meeting the simple absorption/transport requirements, accelerated innovation in solar cell absorbers requires optimization with respect to the truly limiting factors of the PV technology. For example, the light absorption, quantified by appropriate band gap and large absorption coefficient, is often not the limiting factor for solar cell operation. Instead, other materials properties, such as minority carrier lifetime, band offsets, microstructure and so on are much more important for the solar cell performance. I will present the recent improvements in spatially resolved high-throughput measurements of these critical PV absorber properties on combinatorial sample libraries of Cu-M-S and other thin film PV absorbers.
The ultimate goal of accelerated innovation in inorganic thin film solar cell absorbers is incorporations of the “champion” materials into working photovoltaic device prototypes. The traditional approach of identification of the champion material and then integration of it into a device prototype is prone to usual concerns of scalability and transferability of combinatorial results. On the example of the Cu-based solar cell absorber materials discussed above, I will demonstrate the progress towards development of the approach, where the combinatorial design of photovoltaic absorbers with respect to materials properties is performed in parallel with accelerated innovation at the solar cell prototype level with respect to realistic device operation metrics, such as short circuit current, open circuit voltage, fill factor, and energy conversion efficiency.
This research is supported by the U. S. Department of Energy under Contract No. DE-AC36-08GO28308 to NREL

The High Throughput Experimentation (HTE) project of the Joint Center for Artificial Photosynthesis (JCAP, http://solarfuelshub.org/) performs accelerated discovery of new earth-abundant photoabsorbers and electrocatalysts. Through collaboration within the DOE solar fuels hub and with the broader research community, the new materials will be utilized in devices that efficiently convert solar energy, water and carbon dioxide into transportation fuels. JCAP-HTE builds high-throughput pipelines for the synthesis, screening and characterization of photoelectrochemical materials. In addition to a summary of these pipelines, we will describe several new screening instruments for high throughput (photo-)electrochemical measurements. These instruments are not only optimized for screening against solar fuels requirements, but also provide new tools for the broader combinatorial materials science community. We will also describe the high throughput discovery, follow-on verification, and device implementation of a new quaternary metal oxide catalyst. This rapid technology development from discovery to device implementation is a hallmark of the multi-faceted JCAP research effort.
This material is based upon work performed by the Joint Center for Artificial Photosynthesis, a DOE Energy Innovation Hub, supported through the Office of Science of the U.S. Department of Energy under Award Number DE-SC0004993.

The optimization of magnetron sputtering deposition procedures to synthesize samples with a desired thickness and composition is a time intensive and costly process. This is particularly the case for high-throughput experiments where initial composition searches are subsequently refined to produce narrow composition regions exhibiting optimal figures of merit. Predictive modeling can improve the efficiency of this procedure by lowering the number of samples required to optimize deposition conditions. In this work, a semi-empirical approach to modeling the thickness of thin-film samples deposited via magnetron sputtering has been developed. Using the dimensionless angular atomic flux, as derived by Yamamura, and a measured deposition rate at a point in space for a single experimental condition, the model predicts the deposition profile for off-center, tilted, and planar DC sputtering sources as well as shadowing effects from gun chimneys used in most state-of-the-art sputtering systems.
The modeling algorithm was validated by comparing its results with experimental deposition rates obtained from a state-of-the-art sputtering system. The sputtering system was equipped with five sputtering sources in an off-center geometry, a DC power supply, in situ gun tilting and a chimney assembly consisting of a lower ground shield and a removable gas chimney. Simulations were run for gun-tilts ranging from 0° to 31.3° from the vertical with and without the gas chimney installed. The correlation between the predicted and experimental angular dependence of the sputtering deposition rate was found to have an average magnitude of the relative error of 4.14% ± 3.02% for a 0° to 31.3° gun tilt range without the gas chimney, and 2.12% ± 1.71% for a 17.7° to 31.3° gun tilt range with the gas chimney. Furthermore, the model was utilized to predict composition distributions for Fe-Cr-Al ternary composition spread samples deposited by a combination of DC and RF power sources. Here, the model guided the selection of deposition parameters, such that the composition spread of a single sample could be varied from large portions of ternary phase space to focusing on only the Fe-based bcc composition region. The breakdown of the model due to the time-varying energy supplied by RF power sources, as well as empirical and theoretical methods for improving the model performance under these conditions will also be discussed.

Copper nitride (Cu₃N) is an Earth-abundant semiconducting material with potential as a photovoltaic absorber. Cu₃N forms in a linearly bonded crystal structure and is thermodynamically unstable (has a positive heat of formation). Nevertheless, thin films and bulk samples of Cu₃N can be synthesized by a variety of methods and are shelf-stable for at least a year. Cu₃N has much less stringent growth requirements than traditional inorganic absorbers, and simple reactive sputtering can produce high quality thin films with reasonable transport properties. In reactive sputtering of metals, a working gas reacts with a metal target to produce a compound film. The chemical potential of the working gas, which is a function of temperature and pressure, is an important parameter in film growth. Traditionally, modulation of partial pressure or flow rate is used to change the working gas chemical potential. However, this method is inherently low-throughput as only a single setting is used throughout a deposition. Substrate temperature has been varied in a combinatorial manner, though often with the primary intent of encouraging crystallinity.
In this work, Cu₃N films were synthesized by radio-frequency sputtering of a copper target in an argon and nitrogen atmosphere with a nitrogen plasma source. In order to optimize growth conditions, the temperature of the substrate and the distance between the target and substrate were varied in a combinatorial manner. This produced growth conditions with different chemical potentials of nitrogen at each point on the substrate: As temperature increased, the chemical potential of nitrogen decreased because emission of N₂ was encouraged. As target-substrate distance increased, the chemical potential of nitrogen decreased as the target plasma spread out and atomic nitrogen underwent gas-phase recombination. At temperatures below 190 °C and small target-substrate distances, <00L> oriented films were produced. Higher temperatures and longer target-substrate distances produced polycrystalline films, some of which contained copper metal inclusions. Grain size was not a monotonic function of temperature as would be expected for a thermodynamically stable material. The morphology of the films depended strongly on the chemical potential of nitrogen: the largest grains were observed in a region of optimal chemical potential with grain size falling as the chemical potential of nitrogen increased or decreased.
These results provide optimized growth conditions for Cu₃N that will be incorporated into lab scale photovoltaic devices. Further, this high-throughput method of changing anion chemical potential is not limited to nitrides but rather is a promising route to accelerate the development of a wide range of materials systems.
This research is supported by the U.S. Department of Energy, Office of Energy Efficiency and Renewable Energy, as a part of a Next Generation PV II project within the SunShot initiative.

A novel oxide semiconductor material is designed and realized by exploiting both structure-property and composition-structure relationships. While transition metal oxides often have desirable chemical properties, e.g., stability in aqueous solution for photo-electrocatalysis, they usually do not have desirable semiconducting properties. A promising new route towards semiconducting transition metal oxides was recently identified in our study on binary and ternary d5 oxides [1,2]. The key finding of Ref. [1] was that undesirable self-trapping of holes occurs in MnO due to a Mn^+II/Mn^+III transition in an octahedral coordination of Mn, but not in a tetrahedral coordination. Thus, a hypothetical zinc-blende or wurtzite (WZ) structure of MnO could bring about a desirable band transport mechanism instead of the undesired small polaron mechanism that impedes hole transport in the normal rocksalt (RS) structure.

In order to induce a transition from octahedral to tetrahedral coordination in MnO, we considered theoretically the isovalent alloying with ZnO and aliovalent alloying with GaN, i.e., compounds that prefer the tetrahedral over the octahedral coordination. In order to predict the critical compositions at which the transition into a wurtzite structure occurs, we employed total-energy calculations in the random-phase-approximation [3], which are able to overcome problems to predict the structure of MnO as experienced in conventional density functional approaches [4]. Those critical compositions are predicted at x = 0.38 and x=0.15 for (MnO)_1-x(ZnO)_x and (MnO)_1-x(GaN)_x alloys, respectively. Further calculations for MnO-ZnO alloys of the band-structure, optical properties, and the band alignment with the vacuum level (ionization potential) suggest that such alloys could be interesting materials for solar water splitting.
In the experimental part of this work, we have synthesized MnO-ZnO alloys by combinatorial pulsed laser deposition. For deposition temperatures up to about 300° C and alloy compositions above x = 0.4, we were able to realize wurtzite structure alloys without rocksalt impurity phases, whereas at higher temperatures a phase separation between RS and WZ structures occurs, especially above 500° C. Optical measurements confirm the strong reduction of the (optical) band gap associated with the RS to WZ transition, and are consistent with the predicted value of E_g = 2.1 eV at a x = 0.5 alloy composition. The undoped MnO-ZnO alloy films are electrically insulating, but we were able to achieve electron concentrations in the 10¹⁹ cm^-3 range by means of Ga doping.
[1] H. Peng and S. Lany, Phys. Rev. B 85, 201202(R) (2012).
[2] H. Peng et al., Adv. Funct. Mater. DOI: 10.1002/adfm.201300807.
[3] H. Peng and S. Lany, Phys. Rev. B 87, 174113 (2013).
[4] A. Schrön, C. Rödl, and F. Bechstedt, Phys. Rev. B 82, 165109 (2010).
Supported by the US Department of Energy, Office of Basic Energy Sciences as part of an Energy Frontier Research Center.

When building physical models (truncated expansions, force-fields, etc.), one often employs the widely-accepted intuition that the physics is determined by a few dominant terms. This reductionist paradigm is limited because the intuition for identifying the those terms often does not exist or is difficult to develop. A "Bayesian compressive sensing" technique provides simple, general, and computationally efficient solution to this challenge. Combined with the high-throughput first principles approach to materials, BCS makes it possible to automatically build models for binary alloy models without human monitoring. Furthermore, the method can automatically adjust to generate the simplest model (fewest terms) that meets a specific accuracy requirement. Beyond the alloy applications, our BCS code can readily be applied to other model building problems. One merely needs a basis representation and training data to build a model in just a few seconds.

Control of errors over the whole range of structures is essential when we combine a large set of density functional theory calculations and the cluster expansion method for predicting the ground-state structures and configurational thermodynamics of multicomponent systems. An optimal CE is generally constructed from the DFT results of many ordered structures that are sampled from the total population of structures. Although the total population of structures is mainly composed of structures near a random structure, minority structures that are far from a random structure are important for constructing CE with controlled accuracy. It should be emphasized that ground-state structures are usually included in minority structures. We here propose a procedure based on the cluster analysis of the structure population, which can adequately take into account the errors of minority structures as well as those of random structures.
The cross validation (CV) score has been widely accepted as a quantity for controlling the accuracy of the CE. For the accurate evaluation of the CV score, an optimal set of many DFT structures is necessary. In such a case, however, the errors of minority structures are only a minor part of the CV score. When the errors of minority structures are much larger than the average error, they tend to be underestimated in the CV score. In order to estimate the accuracy of a wide range of structures including minority structures, the cluster analysis of the structure population (CASP) is introduced. CASP enables us to classify structures of similar correlation functions into the same group. The usefulness of the procedure is demonstrated by applying it to configurational behaviors of MgAl2O4 spinel.
In multicomponent ionic systems with configurations of heterovalent ions such as MgAl2O4 spinel, contributions of long-range effective cluster interactions (ECIs) to the configurational energetics are not negligible, which is ascribed to long-range electrostatic interactions. CE only with short-range ECIs in such systems leads to systematic errors. A typical problem with such errors can be seen in Monte Carlo (MC) simulations since much larger supercells than DFT cells are used. The prediction errors for the long-period structures beyond the DFT cell in addition to those for the short-period structures within the DFT cell need to be carefully examined in order to control the accuracy of CE. In the present study, we quantitatively discuss the contribution of the truncation of long-range ECIs to the accuracy of CE. Two kinds of systems, namely the point-charge spinel lattice and the real spinel MgAl2O4 crystal, are examined. The prediction error of the long-period structures is satisfactorily improved both by increasing the number of pairs and by considering an additional effective screened electrostatic energy.

Multiphase nanomaterials offer an effective approach for the design and development of bulk solids with optimal mechanical properties. Most of the current research in this field is focused on devising novel materials or new synthesis methods but only a limited effort has been directed at establishing a robust and versatile modeling tool that would be capable of guiding synthetic processes toward optimal bulk properties. Additionally, the common practices in modeling efforts largely involve methods based on classical continuum mechanics, which suffer from the inherent limitation that the spatial derivatives required by the partial differential equations do not, by definition, exist at crack tips or along crack surfaces. Therefore, the basic mathematical structure of the formulation breaks down whenever a crack appears in a body. Recently, a new theory, peridynamics, has been introduced for predicting the fracture behavior of material systems with dissimilar constituent materials. Interfaces between dissimilar materials have their own properties and damage can propagate when and where it is energetically favorable for it to do so. This feature allows modeling of damage initiation and propagation at multiple sites, with arbitrary paths inside the material, without resorting to special crack growth criteria. Nanostructured materials have significantly larger fraction of interfaces; hence, PD theory provides the ability for realistic computational modeling of fracture and failure in these materials.
Based on the foregoing, this study presents a combined experimental and computational methodology for design and development of nanostructured materials with superior mechanical properties. The work involves collaborative work between synthesis, testing and computational modeling. The methodology will be demonstrated on glass-ceramics synthesized with nanoscale inclusions of variable size, shape, volume fraction, moduli as well as variable interfacial properties between the matrix and inclusions. The synthesis effort will be guided by peridynamics modeling; optimum values of the aforementioned variables will be predicted by the simulations. Mechanical testing in the form of indentation and impact will be used for validation purposes. The findings of this study will help provide a direct method for the design of novel high-strength materials.

Substances which have anisotropic crystal structures and physical properties require high degree of orientation to generate their intrinsic functionalities for improving them into practical materials. Especially in high-T_c superconductors which have highly anisotropic crystal structure and physical properties, an increase in the misorientation angle between two grains, even for c-oriented materials, leads to a serious decrease in intergrain critical current density. Therefore, improving cuprate superconductor materials for practical applications requires not only the formation of a c-axis oriented microstructure but also the alignment of grains along the a- and b-axes directions parallel to the superconducting CuO₂ planes.
Magnetic alignment by application of modulated rotation magnetic field (MRF) to magnetically anisotropic crystals is a new method of crystal alignment which realizes tri-axial crystal orientation without epitaxial growth. In this magneto-scientific method, magnetic anisotropy in materials is the most important factor for determing degrees of orientation and orientation axes. In this study, we focused on RE-based (RE = rare earth ions) cuprate superconductors, which require high degrees of orientation for application and could be substituted for almost all RE in its RE site. Stoichiometric and twin-free REBa₂Cu₄O₈(RE124) superconductors were chosen to distinguish the effect of single ion anisotropy of rare earth ions.
RE124 single crystals were grown by the flux method with optimization of growth temperature. These compounds were obtained for RE = Y, Nd, Sm, Eu, Gd, Dy, Ho, Er, and Tm. The RE124 crystals were ball-milled and aligned in epoxy resin under several MRF conditions. Their magnetization axes and degrees of grain-orientation were evaluated using XRD and rocking curves.
Magnetic tri-axial alignment was realized for all the obtained RE124 samples. Their magnetization axes depended on the type of RE ions. Furthermore, among the changes in the degrees of orientation for the three different MRF conditions, it was found that tri-axial single-ion magnetic anisotropies of heavy RE³⁺ ions were higher than magnetic anisotropies generated by Cu-O networks and RE³⁺ions.

In the present work, we report the prediction of oxygen diffusivities in fcc nickel from first-principles calculations, associated with large-scale simulation techniques. While the commonly recognized mechanism for oxygen diffusion in fcc phase is of interstitial type with the minimum energy pathway (MEP) of octahedral to tetrahedral to octahedral, our results show that considering this MEP alone could underestimate the diffusivities by 2~3 magnitudes and the existence of vacancy has a significant effect on the migration barrier and leads the predicted diffusivities consistent with available experimental data. Our first-principles results show that at high temperatures, the vacancy concentration is comparable to the oxygen solubility and there is a high opportunity that oxygen and vacancy can get together because the strong bonding effect between them, which is demonstrated by the large bonding energy and the redistribution of charge density.

Shaped femtosecond near IR laser irradiation is explored as a general methodology to produce metal nanoparticles from metal precursor solutions. Initial studies of the formation and transformations of gold nanoparticle in aqueous solution were used as model processes to evaluate the effects of laser parameters, solvent media and surfactants to achieve control of metal nanoparticle formation. The addition of polymer surfactants like polyethylene glycol (PEG) results in significantly faster and more efficient metal ion reduction than observed for aqueous, surfactant free systems. This acceleration in the reduction rate is attributed to the in situ generation of reactive radical species from PEG fragmentation. Photo-reduction for aqueous solutions of Au(III) in the presence of PEG results in relatively small narrowly dispersed spherical gold nanoparticles compared to relatively large well-formed crystalline nanoparticles that are observed in the absence of surfactants. Varying the concentration of PEG is an effective approach to tune the diameter and size distribution from 3.9±0.7 nm to 11±2.4 nm for Au nanoparticles produced by laser processing. Systematic variation of the reaction parameters provides strategies to achieve size and morphology selectivity. Results of detailed mechanistic studies and the scope of transition metal nanoparticle preparation using shaped femtosecond near IR laser irradiation will be presented.

The growth and characterization of single crystals by the Czochralski method is being carried out to generate YAG crystals with well-defined chemistry to investigate the efficacy of using tetravalent species to control point defect chemistry. The ultimate focus is to affect the relative ratio of radiative to nonradiative transitions. The compositions investigated were identified using statistical learning techniques that take into account crystal structure, composition, and defects. Chemistry, crystallographic defects, and point defects of the crystals will be described. The results of anion and cation tracer diffusion studies will also be described.

Lightweight porous foams have been of particular interest in recent years, since they have a very unique set of properties which can be significantly different from their solid parent materials. These properties arise from their random porous structure which is generated through specialized foaming techniques. The materials of particular interest in this study are open cell metallic foams which are subjected to compressive loads. The objective of this project was to integrate simulations and experiment to determine whether a relationship exists between the relative density (porous density/bulk density) and the mechanical properties of Al foams. Modeling the foams themselves is a particular challenge as they possess no regular solid structure and therefore cannot be modeled as such. Generating a realistic random porous model is difficult, hence a method was created for the purpose of this study. Previously, simplifications to a regular open structure have been used to gain a basic idea of how a foamed material will perform; however the models typically consist of a regular repeating unit cell, and some of the foam&’s true nature is lost in this approximation. Models of the foam were generated using a combination of an open source software, Voro++, and MATLAB. A Finite Element Method (FEM)-based software, COMSOL Multiphysics 4.3, was used to simulate the mechanical behavior of Al foam structures under compressive loads ranging from 1-100 MPa. From these simulated structures, the maximum von Mises stress, volumetric strain, and other properties were calculated. These simulation results were compared against data from compression experiments performed using the Instron Universal Testing Machine (IUTM) on ERG Duocel open cell Al foams with 4-6% relative density. CES EduPack software, a materials design program, was also used to estimate the mechanical properties of open cell foams for values not available experimentally, and for comparison purposes. The method developed in this study can be used to generate other open cell foam models of any density and pore size distribution (independent of the material) for use in future FEM simulations. Overall results indicated that a combination of experiments and simulations, as well as versatile models, can be used to calculate structure-property relationships and to predict yielding and failure, which may help in the pursuit of simulation-based design of metallic foams. In the future, more robust modeling and simulation techniques need to be explored, as well as investigating closed cell Al foams and different porous geometries (nm to mu;m). This study can help to improve the current methods of characterizing porous materials and enhance knowledge about their properties for alternative energy applications, while promoting their design through integrated approaches.

The goal of this study was to evaluate and predict the mechanical properties of sandwich structures, consisting of two 6061 Al sheets and a polyurethane foam core, as a function of three different foam densities (62 kg/m³, 160 kg/m³, and 320 kg/m³). The sandwich structure dimensions were 12 mm x 25 mm x 390 mm, and load levels ranged from 2000-3200 N as in previous independent experiments. Computer simulations and experiments were carried out in this project in order to create an efficient procedure for future design considerations. A Finite Element Method (FEM) program, COMSOL Multiphysics 4.3, was used to predict the critical buckling load and bending modulus of the sandwich structure when subjecting it to uniaxial compression loads and a three-point bending test. The results were compared with previous independent experimental findings. From the FEM simulations, other mechanical properties were analyzed including the maximum von Mises stress, displacement, and volumetric strain among others. An experimental three-point bending test was also conducted using a Instron Universal Testing Machine to measure the mechanical properties of the sandwich structure. Predicted results showed that the critical buckling load increases as foam density increases, implying that the density and strength of the foam in the sandwich structure have a direct relationship, and affect the overall properties of the combined foam-sheet system. A design-based software, CES EduPack, was also used to estimate the mechanical properties of other similar sandwich structures. In brief, simulations and experimentation combined can be a powerful tool for the study, prediction and design of new or advanced materials. The original motivation for this project was to evaluate the possible use of this sandwich structure in a Small Formula One Car for the Society of Automotive Engineers (SAE) at UC Merced. Future work will involve assessment on the feasibility of these sandwich structures as vehicle impact absorbers in particular. This project has also implications in areas such as allowing future automobiles to be lighter and safer, which in return would increase fuel efficiency.

One pathway for achieving materials with new or improved properties is imaginative chemical modification. This has been particularly exploited in perovskites with the chemical formula ABO3 (with A and B being cations) - materials that have found solutions in diverse technological applications in fields ranging from electronics to sensors to catalysis. Owing to the intrinsic capability of the perovskite structure to host ions of various sizes across the Periodic Table, a wide range of dopants can be, and have been, successfully accommodated in the ABO3 structures. Within the class of ABO3 perovskite-type materials, BaTiO3 (BTO) is perhaps the most widely used and studied multifunctional ceramic material.
In the present study, an extensive assessment of the physicochemical factors that control the behavior of dopants in BaTiO3 has been performed using high-throughput first-principles density functional theory (DFT) computations. Dopants spanning the Periodic Table - 44 in total - including K-As, Rb-Sb, and Cs-Bi were considered, and have allowed us to reveal previously unknown correlations, chemical trends, and the interplay between stability, chemistry, and electrical activity. In order to understand the dominant factors that control the dopant formation energy, we have used a newly developed machine learning method supplemented with a genetic search algorithm to determine the specific properties of the dopant atoms that most control the computed formation energy. The considered properties included the ionic size, oxidation state, electronegativity, polarizability, ionization potential, and electron affinity. Our machine learning strategy (which use the DFT data as input) lead us to the conclusion that the most important factors that determine dopant formation energy in BaTiO3 are the dopant ionic size and oxidation state, and that all other properties play an insignificant role. The present development will allow for the identification of “Hume-Rothery” type rules that govern materials behavior.

Atomic layer deposition (ALD) is a well-known deposition technique that is based on sequential and self-limiting surface reaction. Due to the good uniformity and conformality, ALD technique is widely used in many commercial areas. In the atomistic view of the initial reaction of ALD, the precursors adsorbed on the surface, react with the functional chemical species of the surface via bond breakings and new bond makings, and formed reaction products. In order to react with the surface, adsorption energy of the precursor on the surface should be higher than its reaction energy barrier; otherwise, the precursor would desorb rather than react with the surface. However, the relationship between the adsorption energy and reaction energy barrier could not be connected to experimental works directly, because the enthalpy and entropy at operating temperatures and pressures were not considered carefully. In order to overcome the gap between the theoretical and experimental results, the contribution of entropy and enthalpy on initial ALD reaction should be considered. There are three different methods for calculating the entropy and enthalpy; Sackur-Tetrode equation, thermodynamics data and vibrational frequency calculation. The entropy of a monoatomic gas can be calculated by Sackur-Tetrode equation. The entropy contribution of small molecules was well matched with the experimental data. However, the calculational results for complex molecules were not matched with experimental ones. The experimentally measured thermodynamic data can used to consider these issues. However, the thermodynamic data were limited to the small molecules and not available for currently used ALD precursors. The vibrational frequency calculation is another approach to consider entropy and enthalpy. In this work, we investigated the contribution of enthalpy and entropy effect on the initial reaction of ALD using the three approaches. By comparing initial reaction of silicon precursors with a hydroxyl terminated silicon (001) surface, we will further elucidate the initial reaction mechanism of ALD.

Robust experimental confirmation of the initial design principles used to guide materials selection is a critical step in a materials-by-design approach. In this work, we demonstrate how synchrotron based resonant elastic X-ray diffraction (REXD) were used to quantify the anti-site defect concentrations and confirm the role of intentional cation-site-occupancy disorder as a design principle for p-type doping in cobalt oxide based spinels. Further, we find that the cation anti-site disorder present in the as-deposited thin film samples grown at T asymp; 350 °C is equivalent to that expected for equilibrium grown samples at T asymp; 2000 °C. The resultant concept of using an effective disorder temperature to computationally mimic the site-occupancy disorder found in non-equilibrium grown samples may enable more accurate predictions of electronic properties of such materials.
In particular, we confirm the predicted cation disorder and doping physics in A2BO4 and demonstrate that intrinsic cation disorder improve electrical properties of Doping Type 2 spinels (a special class of normal spinels having not significant hole killers, not even cation defects)[i] in agreement with our prior predictions. This was made by investigating the impact of cation disorder on the electrical properties of biaxially textured Co2ZnO4 and Co2NiO4 thin films, using a combination of experiment and theory. Cation disorder measurements by REXD along with conductivity measurements before and after post-deposition annealing show that conductivity decreases with disorder in the case of Co2ZnO4. This confirms predictions via theory that anti-site disorder is a net hole producer for Doping Type 2 spinels typified by Co2ZnO4.[ii] In the case of Co2NiO4, an inverse spinel, conductivity increases with disorder. To explain the opposite changes of the conductivity with cation disorder for these two materials, we determine the changes in enthalpy and electronic structure as a function of the disorder using first principles calculations. Our results demonstrate that Co2NiO4 behaves as half-metal whereas Co2ZnO4 behaves as semiconductor providing an explanation of the opposite trends of conductivity as a function of disorder. We also introduce a concept of effective temperature that provides a way to quantify the disorder due to non-equilibrium growth, and to establish a link between theoretical thermodynamic models and thin film growth techniques.
These new insights into the role of cation-site-occupancy disorder underpin a new Design Principle whereby non-equilibrium growth is used to create beneficial disorder.
This work was supported through the Center for Inverse Design, an Energy Frontier Research Center funded by the U.S. Department of Energy, Office of Science, Office of Basic Energy Sciences.
[i]Paudel et al, Adv. Funct. Mat. 21, 4493 (2011).
[ii]Perkins et al, Phys. Rev. B, 84, 205207 (2011).

Nanoparticles have intriguing effects on materials (magnetic, optic, catalytic, etc.), which are closely related to the following properties; particle size, morphology, exposed crystal facets, and crystal structure. To achieve tunable nanoparticle performance, fine control over nanoparticle properties including particle size and morphology is required. This in turn requires an understanding of the level of participation of each major synthesis conditions on the final nanoparticle morphology. Unfortunately, multiple parameters are usually involved in nanoparticle synthesis procedures, and significant amounts of time and costs are required to determine the influence of each of the parameters by conventional one-at-a-time experiment methods. These synthesis parameters include reaction temperature, rate of temperature ramping, reaction time, concentration of metal precursors, and ratio of the concentrations of the metal precursors and the capping agent. Each of those parameters can influence the particle size of the CoFe nanoparticles, further there may be interactions between parameters. To determine the influence of all those factors, a complex multi-dimensional phase-space must be explored systematically. Design of Experiment (DoE) has been widely applied in material optimization due to its ability to create systematic studies and provide detailed statistical analysis of the relationship between the manipulated variables and the final condition of reaction products.
To demonstrate the power of DoE in providing a comprehensive understanding of how synthesis conditions affect nanoparticle size, we have applied DoE methodologies to the synthesis of CoFe nanoparticles, an interesting catalyst for NaBH4 hydrolysis. We observed that the synthesis temperature and ratio of the capping agent concentration to that of the metal precursors had the most significant influence on the final particle size. The effect of changing the precursor can be understood as it relates the nucleation speed and thus produces smaller nanoparticles when increased. Conversely, increasing the capping agent decreases the nanoparticle growth rate by blocking the particle surface from synthesis solution. Moving forward it will be beneficial to explore these more noteworthy variables at multiple levels. This will then enable the synthesis to be further optimized, using more thorough knowledge of the local minima.

Equilibrium phase diagram of α-Fe2O3 (hematite) - α-Al2O3 (corundum) system shows the presence of the orthorhombic AlFeO3(AFO). Both α-Fe2O3 and α-Al2O3 crystallize in the rhombohedral structure (R-3c) while the orthorhombic phase AlFeO3 is isostructural with the polar GaFeO3 belonging to a space group Pna21. Because GaFeO3-type structure is polar structure, GaFeO3-type AlFeO3 is expected to exhibit ferroelectricity. However, this expectation has not been confirmed by experiment until now. Because it is difficult to obtain the GaFeO3-type AlFeO3 by the conventional solid reaction technique due to the range of stability of AlFeO3 in a phase diagram is very limited with regard to the temperature ( > 1318 deg.) and the composition.
Pulsed laser deposition (PLD) technique could be used in a stabilization of metastable phases that could not be synthesized using the traditional solid state chemistry routes.In this work, we demonstrate to grow metastable GaFeO3-type AlFeO3 employing PLD technique by selecting optimum substrates.
The GaFeO3-type AlFeO3 film was epitaxially grown on (111)SrTiO3 substrates. The results of PFM measurements revealed ferroelectricity in the GaFeO3-type AlFeO3 films. We also tried to prepare Al1-xFexO3 with x = 0.25 and 0.75, and measured physical properties. AlFeO3 with the GaFeO3-type structure is one of the ferroelectric candidates, which is composed of environmentally friendly and ubiquitous elements. Detail of the thin film structure, magnetism, ferroelectricity will be reported in the presentation.

Recent studies on gold nanorods (GNRs) have identified the presence of Au310 and Au520 surfaces on their tip. These novel stepped surfaces have unique binding properties that can explain the high reactivity at tip of GNRs. The present contribution presents first principles density functional theory simulations of Au310 and Au520 surfaces with alkanethiol ligands and 4prime;-(12-mercaptoalkyloxy) biphenyl-4-carbonitriles (CBO(CH2)12SH). The simulation results were compared with conventional Au100, Au110 and Au111 surfaces. Due to the tight gold packing, the sloped Au111 surfaces have the highest binding energy with alkanethiolates at 41.38 kcal/mol. The Au310 have the lowest binding energy of 21.51 kcal/mol. The trends of ligand binding energy correspond exactly with ligand surface modulus of elasticity. Au111 surfaces have the highest modulus of elasticity of 134 GPa, and Au310 have the smallest modulus of elasticity of only 32 GPa. The average modulus of elasticity for other 3 surfaces is 74 ±8 GPa and average binding energy is 28.5 ±1 kcal/mol, and both of them agree very well with the experimental value of 64 ±10 GPa for molecular gold and of 29.5 kcal/mol for alkanethiolates binding energy. The sloped Au520 have a higher binding energy than flat Au520 surface, 31 kcal/mol versus 25 kcal/mol. The sloped surfaces at GNR tip are able to distribute the ligand induced strain with minimum surface reorganization due to high modulus of elasticity and it can provide extra free space at the top of ligands to reduce ligand to ligand strain. This is why GNR tip can have a higher reactivity. The binding of CBO(CH2)12SH also demonstrate this. These longer ligands have very bulky head groups, the result is a small ligand cant angle of 20o and a larger curvature to space the bulky head group. The curvature is measure via the dihedral angle the carbon backbone formed with the gold surface. It shows that as the carbon backbone move away from the surface, ligand curvature increases. Our earlier experimental results have indicated that GNR decorated with even mixture of CBO(CH2)12SH and alkanethiolates interact better with liquid crystals than GNR with pure alkanethiolates. The simulation of mix SAM with 5CB have shown that the curvature of CBO(CH2)12SH is not affected by the 5CB. The only difference is the ligand cant angle for 11CBO increased by 5o in the presence of 5CB. This is not the case for Au100 surface with the shorter hexanthiolate ligand, the ligand cant angles have increase from 50 o to 60o and more importantly the 5CB forced a cis-conformation formed with the carbon backbone. This simulation results showed that it would take too much strain for the 5CB to interact effectively with the hexanthiolate. The simulation results provide a comprehensive quantitative understanding of GNRs in how to incorporate them into novel material systems that include liquid crystals.

I will discuss novel theoretical methodologies and computational techniques that can simultaneously address the localized (molecular) and delocalized (solid-state) electronic structure of materials in a seamlessly integrated computational framework, an essential requirement for the study of systems with stronger electron localization and correlation. In particular, I will describe a straightforward, non-iterative projection scheme that can represent the electronic ground state of a periodic system on a finite atomic-orbital-like basis, up to a predictable number of electronic states, with controllable accuracy and at a negligible computational cost, an essential requirement for the design of efficient algorithms for electronic structure simulations of realistic material systems and massive high-throughput investigations.
Such approach is essential in a gamut of applications, and I will illustrate its capabilities using a few examples form some of the following: calculation of the electronic states on ultra dense k-space grids for accurate Brillouin zone (BZ) integrations; calculation of exact-exchange integrals; evaluation of quantum transport properties; efficient calculation of electron-phonon interactions; construction of model Hamiltonians for correlated-electrons and magnetic systems.

We present here the ab initio random structure searching (AIRSS) method and how it can be applied to study solids in the context of materials discovery. AIRSS relies on generating random structures and relaxing them within the framework of density-functional-theory (DFT). The method is simple, requiring only a small set of parameters that can easily be connected to the physics of the system of interest, and efficient, allowing for high-throughput first-principles calculations on modern parallel architectures. Our method can identify the stable ground state structure and also the adjacent low-energy metastable states. This makes the method particularly well suited to analyze systems where knowledge of all possible states is crucial, as for instance for structures trapped in higher-energy states during growth or processing. We apply the method to novel compounds combining metals and III-V semiconductors that have recently seen great interest in the search for a replacement of conventional metal silicides for low-resistance contacts in Si-based electronic devices. We include a detailed discussion of the low-energy structures and their respective physical properties of one such promising new metal/semiconductor compound. Comparison to TEM studies allows us to put further constraints on our search results.

In order to design new materials with magnetic components, e.g. steels or CrN-based hard coatings, it is necessary to understand their high temperature paramagnetic phase in which they are typically synthesized [1]. Unfortunately, first-principles calculations has this far been unable to simultaneously treat the relevant magnetic, vibrational and structural excitations present at high temperature. Here we present a first-principles framework capable of calculating thermodynamic and magnetic properties of magnetic materials in their high-temperature paramagnetic state. It is based on a supercell formulation and carried through using standard plane-wave ab-initio codes. In this way it matches well with the current efforts within the high-throughput computational community and could be a valuable tool to accelerate the design of new magnetic materials.
Calculations of temperature dependent vibrational spectra, Gibbs free energies and corresponding phase stabilities are based on a combination of two of our recent developments in the field of quantum molecular dynamics: the disordered local moments-molecular dynamics (DLM-MD) [2,3] and the temperature dependent effective potential (TDEP) method [4]. We demonstrate its ability by modeling the phonon spectra and pressure-temperature dependent phase transition of the hard coating material CrN and the mixing thermodynamics of Cr1-xAlxN.
We also demonstrate that our supercell framework is able to calculate the effects of vibrations and structural disorder on the magnetic exchange interactions, needed to predict magnetic critical temperatures from first-principles. We use our approach to derive the complex magnetic interactions of amorphous CrN [5].
[1] B. Alling, T. Marten, and I. A. Abrikosov, Nature Materials 9, 283 (2010)
[2] P. Steneteg, B. Alling, and I. A. Abrikosov, PRB 85, 144404 (2012)
[3] B. Alling, L. Hultberg, L. Hultman, and I. A. Abrikosov, APL 102, 031910 (2013)
[4] O. Hellman, I. A. Abrikosov, and S. I. Simak, PRB 84, 180301(R) (2011)
[5] A. Lindmaa, R. Lizaraga, E. Holmström, I. A. Abrikosov, and B. Alling, arXiv:1306.3107

A theory-led approach to understand the properties of real materials and devise better ones is a long-standing vision. Two central issues to be addressed are (i) real (experimental) materials can be complex, (ii) the underlying level of theory must be reliable and feasible for the problem at hand. This talk discusses the concepts and recent advances of our first-principles approach to (i) and (ii), embodied in the FHI-aims [1] all-electron electronic structure package for materials and nanoscale systems. The package provides reliable, affordable all-electron numbers up to large systems (1,000s of atoms), periodic and cluster systems on equal footing, a seamless treatment from light to heavy elements, excellent scalability on massively parallel hardware, advanced functionals, and more. One area where these ingredients are critical is nanoscale interface design. We discuss the growth of large-area graphene films on SiC semiconductor substrates: Why mono- and bilayer graphene films can be grown by Si sublimation on the Si side but not (with the same quality) on the C side of this substrate, and how defects and strain in MBE-grown graphene films are related.
[1] http://aims.fhi-berlin.mpg.de; V. Blum et al., Computer Physics Communications 180, 2175-2196 (2009).

Because of the crucial roles of native defects and dopants in the properties of semiconductors, a fair amount of experimental research has been devoted to their characterization in previous decades. However, the understanding of the native defects and dopants is still limited, particularly at the atomistic and electronic level. One difficulty partly results from the fact that they often show complex behavior associated with off-symmetric atomic relaxation and electron localization, in addition to the variety in defect species and charge states. Accurate and systematic understanding and prediction of such behavior of native defects and dopants are required for an efficient screening of oxide and nitride semiconductors. In this talk, I will present some new insights into their atomistic and electronic structures from first-principles calculations. Examples include local octahedral rotation around the O vacancy and the formation of off-centered Ti antisite defects in SrTiO3 [1, 2], multiple configurations of the O vacancy in BaTiO3 [3], the formation of a dopant-native defect complex in cubic BN [4], and the doping of hexagonal BN with a layered structure, for which dilute intercalation of donors or acceptors is proposed to be effective [5].
[1] M. Choi, F. Oba, Y. Kumagai, and I. Tanaka, Adv. Mater. 25, 86 (2013).
[2] M. Choi, F. Oba, and I. Tanaka, Phys. Rev. Lett. 103, 185502 (2009).
[3] M. Choi, F. Oba, and I. Tanaka, Appl. Phys. Lett. 98, 172901 (2011).
[4] R. Ishikawa, N. Shibata, F. Oba, T. Taniguchi, S. D. Findlay, I. Tanaka, and Y. Ikuhara, Phys. Rev. Lett. 110, 065504 (2013).
[5] F. Oba, A. Togo, I. Tanaka, K. Watanabe, and T. Taniguchi, Phys. Rev. B 81, 075125 (2010).

Fe-Si binary alloy has a variety of applications due to its excellent magnetic and mechanical properties. Some drastic change in mechanical properties as Si concentration increasing from 2-3 to 5-6wt.% has been intriguing since a long time ago. To understand the mechanism of the Si concentration dependence of the mechanical properties, especially a drop down of ductility at 4-5wt.%Si, electronic structure calculations incorporated the phonon vibration effect and thermal electrons excitation are performed to Si-doped bcc-Fe alloy up to 7.0wt.%Si. Using stress-strain method, various elastic properties such as bulk modulus, shear modulus and elastic constants have been evaluated based on the high accurate free energies. The calculations confirmed a non-monotonic change of the elastic properties with Si concentration, found a ductile to brittle transition according to an empirical rule as Si content increasing beyond 4.2wt.%, which is agree with the well known experimental results. Analysis of density of states, magnetic moment and force constants revealed the difference in Fe-Si bond characteristic when Si concentration is crossing 4.6wt%, and suggest a relation of the mechanical properties to the magnetovolume effect of this alloy. Further CVM calculation at finite temperature clarified the influence of the D03 ordering on mechanical properties with Si concentration's increasing to 5.0wt%. The investigation final reached at a conclusion that the interplay between magnetic properties and structure ordering plays essential role in the sharp change of various mechanical properties at 4.6-5.6wt.%Si region in Fe-Si alloy.

As the lightest metallic structural materials, Mg alloys hold great promises to considerably reduce the weight of transportation vehicles and improve their fuel efficiency. Impurity diffusion coefficients are critical properties for Mg alloy design. However, experimental measurements of impurity diffusion coefficients in Mg are scarce in the literature. In the present work, first-principles calculations based on Density Functional Theory (DFT) were used to calculate the full impurity diffusion coefficients for a large series of substitutional elements in hexagonal closed packed (hcp) Mg as a function of temperature using an 8-frequency model. Minimum energy pathways for impurity diffusion and the saddle point configurations during solute migration were calculated with the climbing image nudged elastic band method. Vibrational properties were obtained using Debye-Grüneisen model with input parameters directly calculated from first-principles. The recently developed PBEsol exchange-correlation (XC) functional was used in the present work, which is able to get good results on both vacancy formation energies and equilibrium properties. Excellent agreements of the calculated diffusion coefficients compared with available experimental data were obtained. The established diffusion database forms the foundations to accelerate the design process of high-performance Mg alloys in the context of Materials Genome Initiative. The computational methodology developed herein can be extended to efficiently predict diffusion coefficients in other crystalline materials.

The enhancement in thermoelectric figure of merit (ZT) due to nanostructuring bulk thermoelectric materials has been a profitable paradigm in recent years, giving rise to thermoelectric materials with ZT above 2, such as PbTe-SrTe. Nanostructures in these materials form through phase separation, which is governed by the thermodynamics of mixing the binary constituent compounds. Unfortunately, the phase diagrams of the (II,IV)-VI, (IV,IV)-VI, and IV-(VI,VI) pseudo-binary alloys are not well characterized, and discovering which systems will phase separate is often reliant on an Edisonian, trial and error approach. Recently, we have determined that the phase separation tendencies in the (IV,IV)-VI and IV-(VI,VI) rocksalt semiconductor alloys are due largely to lattice mismatch strain. In particular, the critical temperature of the miscibility gaps (T_MG) in these systems scale with the lattice mismatch squared (a_mis²) between the constituent compounds of the systems. The strain energies which provide the driving force for phase separation also cause the nanostructures to form incoherently, scattering electrons more so than coherent nanostructures would. Thus coherent nanostructures should give a higher ZT than incoherent nanostructures. Here, we use high-throughput first principles calculations to search for semiconductor systems which defy this trend of T_MG prop; a_mis², to find promising candidates for new coherently nanostructured thermoelectric materials. To perform this high-throughput search, we calculate the dilute-mixing energies of 136 pseudo-binary systems on the rocksalt lattice. The lattice mismatch between binary rocksalt constituents are found from DFT lattice parameters, and critical miscibility gap temperatures are found by fitting the dilute mixing energies to a regular solution model. We then screen these calculations to find systems which (i) contain a material with good thermoelectric properties in the bulk, (ii) have a small a_mis² but large T_MG, and (iii) rocksalt as the ground state for at least one of the constituents. Our high-throughput screen is validated as we correctly identify PbTe-SrTe as a good system for forming coherent nanostructures, i.e., a system with a large T_MG despite having a small a_mis². In addition, we identify several promising candidate compounds to alloy with the IV-VI chalcogenides to form ZT-enhancing nanostructures.

Noncentrosymmetric (NCS) materials are under intense investigation because their fascinating properties including piezoelectricity provide a wide range of technical applications. Perovskite oxides without inversion symmetry, such as Pb(Zr,Ti)O3, are widely used in capacitors and transducers. In these perovskite oxides, polar distortions are driven by second-order Jahn-Teller active cations such as Ti4+, Zr4+, or lone pair electrons as in Pb2+ resulting in their lack of inversion symmetry. These limited chemistries limit the abundance of NCS perovskite oxides. On the other hand, most perovskite oxides exhibit the instabilities for oxygen octahedral rotations (OORs). Although inversion symmetry is not only broken by the OORs, the recent theoretical studies have suggested strategies for inversion symmetry breaking by the OORs in conjunction with some types of layered structures such as Ruddlesden-Popper (RP) phase [1].
Here we report the noncentrosymmetry induced by the OORs in A-site-ordered RP phase NaRETiO4 (RE: rare earth), in which the NaO and REO rocksalt layers are alternately stacked across the TiO2 perovskite layer. This series of layered perovskites has been experimentally reported to be centrosymmetric (CS) so far. We have revisited this series by means of synchrotron x-ray diffraction (SXRD), second harmonic generation (SHG), and first-principles phonon calculations. As a result, we have found that their inversion symmetry is broken by a type of TiO6 OOR pattern for NaRETiO4 (RE=Y, Sm-Ho) at room temperature while NaLaTiO4 and NaNdTiO4, which show no OORs, are CS.
The phonon calculations have shown that highly symmetric structure without any OOR belonging to space group P4/nmm (#129) is dynamically stable for NaLaTiO4. The calculations have revealed that the dynamically stable structures for NaSmTiO4 and NaYTiO4 have the OOR pattern represented by a-a0c0/a0a-c0 in Glazer notation with space group P-421m (#113), in which the OOR axes are orthogonal to those in the neighboring perovskite layers. The structure is NCS in contrast to the CS structure with the a-a-c0 OOR pattern belonging to space group Pbcm (#57), which have been reported previously for these compounds. Polycrystalline samples were prepared via conventional solid-state reactions. Noncentrosymmetry has been confirmed by SHG for NaRETiO4 (RE=Y, Sm-Ho) at room temperature. The SXRD patterns have shown the superlattice reflections corresponding to the unit-cell doubling accompanied by the OORs for the NCS compounds at room temperature. Temperature dependence of SHG intensity and SXRD have revealed the CS P4/nmm to NCS P-421m second-order phase transition around 450°C for NaSmTiO4. Thus the novel NSC compounds have been uncovered via the experimental and theoretical approaches.
This work was partially supported by the Penn State NSF-MRSEC Center for Nanoscale Science (DMR-0820404).
References
[1] N. A. Benedek et al., J. Solid State Chem. 195, 11 (2012).

Magnetic materials underpin a vast and diverse range of modern technologies, going from data storage to energy production and use. However, the choice of magnets for mainstream applications is limited to a few dozens and the development of a new high-performance magnetic compound is a long and often unpredictable process. Here we describe a systematic pathway to the discovery of novel magnetic materials for multiple applications, which demonstrates an unprecedented throughput and speed up in the discovery process. We have constructed a massive electronic structures library for Heusler alloys containing 236,856 materials. We have then extracted those magnetic compounds with specific electronic properties, such as half-metallicity and large magnetization density, and finally established whether these can be fabricated at thermodynamical equilibrium. Based on our analysis we have identified 249 stable new intermetallic Heuslers, including 23 new magnets. Our work paves the way for large scale design of novel magnetic materials at unprecedented
speed.

First-principles high-throughput screening of novel materials requires a materials informatics infrastructure able to automatically prepare and execute HPC calculations on large classes of materials, to monitor calculation progress, and to store, retrieve and analyze complex data. In order to speed up computational discovery and optimization, we have developed a powerful flexible environment that integrates these capabilities and is adaptable to diverse applications and use cases. The new platform called AiiDA ("Automated Interactive Infrastructure and Database") combines database storage with user-defined grid-enabled computational workflows. We demonstrate how automated AiiDA workflows make computational design efforts faster, easier, and fully integrated with data collection and community sharing.

In the present work we propose a novel computational approach to explore the broad configurational space of possible interfaces formed from known crystal structures to find new hetrostructure materials with potentially interesting properties. In the series of subsequent steps with increasing complexity and accuracy, the vast number of possible combinations is narrowed down to a limited set of the most promising and chemically compatible candidates. This systematic screening encompasses (i) establishing the geometrical compatibility along multiple crystallographic orientations of two (or more) materials, (ii) simple functions eliminating configurations with unfavorable interatomic steric conflicts, (iii) application of empirical and semi-empirical potentials estimating approximate energetics and structures, (iv) use of DFT based quantum-chemical methods to ascertain the final optimal geometry and stability of the interface in question. We also demonstrate the flexibility and efficiency of our approach depending on the size of the investigated structures as well on the size of the search space. The representative results from our search protocol will be presented for selected materials including semiconductors, transition metal systems, and oxides.

Lattice thermal conductivity, the dominant contribution to total thermal conductivity in semiconductors at room temperature, is an extremely important material property for many applications. For instance, high thermal conductivities are desirable if a compound is to be used for heat dissipation; in contrast, low values of this variable are required to achieve high efficiencies in thermoelectric energy scavenging. In either case, computational methods can supplement experiment in looking for better materials, at a fraction of the cost and providing additional insight into the mechanisms of heat conduction.
In pure (i.e., not alloyed) semiconductor compounds, lattice thermal conductivity is limited mainly by three-phonon scattering. The probability amplitudes of these anharmonic processes can be computed from knowledge of the second- and third-order interatomic force constants (IFCs) of a compound. Then, thermal conductivity is obtained from the solution of the Boltzmann transport equation for phonons. Although within the reach of current ab-initio method and software, these calculations are extremely computationally intensive, even for relatively simple unit cells. This fact makes thermal conductivity calculations on large libraries of materials infeasible, and particularly ill-suited for a direct high-throughput approach.
In this talk two alternatives with lower computational demands are presented. The first is based on the observation that third-order IFCs have some degree of transferability between compounds with the same crystal structure. Thus, an estimation of the thermal conductivity can be obtained using compound-specific second-order constants and a transferable set of third-order ones. The second approach is based on exactly computing the thermal conductivity for a relatively small subset of the materials, building a data set of more easily obtainable descriptors for all of them and training a machine-learning model to predict the thermal conductivities of the rest. The machine-learning algorithm chosen for this work is random-forest regression, and the class of materials used as benchmark was derived from the 78,768 half-Heusler compounds in the aflowlib.org repository. The performance of both approaches to the problem is discussed from different angles, and physical insights are extracted from the results as to which qualities make a half Heusler likely to have a low thermal conductivity. Since the usefulness of this class of compounds as thermoelectrics is acknowledged to be limited precisely because of the relatively high values of this variable, it is concluded that there is probably a lot of room for the discovery of better thermoelectrics among half Heuslers, and some practical criteria are proposed to this end.

The materials discovery process can be significantly expedited and simplified if we can learn effectively from past knowledge. Indeed, the ability to make predictions or decisions by automated learning from the past has found astonishing success in the cognitive sciences. Along similar lines, in the present work a similarity-based machine learning method supplemented with a genetic search algorithm is developed for materials property predictions. We show that such a learning method can effectively determine the most relevant attributes of the building blocks of a system that controls specific material properties, and can thus allow us perform targeted and efficient chemical space searches.
In the present study, a family of C, Si, Ge and Sn-based hybrid polymer systems is considered as an example, with the initial property dataset generated using high throughput density functional theory (DFT) computations. For the polymeric chains composed of seven possible building blocks studied in this work, a chemo-structural attribute or fingerprint vector is defined. This vector contains the number fraction of all possible building blocks of type i, number fraction of i-j pairs, and number fraction of i-j-k triplets. We first use our machine learning approach to determine the proper components of the fingerprint vector that most control a certain property. Trained machine learning models then predict a plethora of properties (including energetic, structural, elastic, optical, and dielectric properties) of an enormous number of new polymers within the same family, and reveal hidden correlations between properties, at negligible cost with high fidelity. The strategy developed in this work allows us to systematically explore and mine vast chemical spaces, and can significantly accelerate the discovery of new application specific materials.

We are confronted with urgent needs for development of various kinds of functional materials with high performance and sufficiently low cost to keep competitiveness in industry and also to tackle inevitable global environmental issues such as sustainable resource and energy supply for society. In practice, it requires lots of trial-and-error experiments for many years to realize just one material substitution in a system. Therefore, computational prediction based on ab-initio calculations, as it proves to be more efficient and sometimes even more accurate than experiment, becomes a major component in the development procedure of new materials. To perform the materials design based on ab-initio calculations more in general, however, we are faced with the following two major problems: (i) representative structures to be optimized for functional materials should be known in advance, and (ii) transport properties such as diffusions are often difficult to be efficiently predicted because the computational effort for ab-initio MD calculations exceeds current hardware capability. Here we develop a materials design platform that consists of hierarchy integration of data mining and simulations. The search space of materials are first selected from database by specifying empirical/theoretical descriptors, and then expanded to new materials deductively. Optimizer for target properties includes ab-initio calculations and forcefield molecular dynamics; for the latter, we have developed an algorithm suitable for automated creation of the FF parameters in solids. Methodological development demonstrated here is promising to develop key functional materials in industry.

> Cultural objects and artworks are unique subject matter for scientific research since their material integrity must be preserved whilst yielding information about their structure and compositional make-up. Information about the composition and chemistry is invaluable to inform decisions regarding treatment, preservation strategies, storage, transport and display conditions as well as informing art historical knowledge. Cultural materials also comprise the broadest expanse of materials space and the inherent material complexity and diversity provide additional challenges for their analysis and interpretation. While many ancient materials survive through history, new approaches for the exploration and analysis of multivariate data are being developed that significantly improve the ability to answer these questions. The work presented will cover a range of examples illustrating how analytical data from X-ray and spectroscopic techniques can be interrogated using multivariate approaches to reveal insights into Cultural materials.

Accelerated discovery and design of materials requires integration of high throughput experimentation (HTE) and directed research strategies. While HTE strategies develop methodologies to rapidly synthesize and quantify the performance of vast arrays of combinatorial libraries of materials, the success of directed research efforts depends on understanding several properties of a material such as microstructure, crystal structure, surface composition etc. It is accepted that in-depth analysis of every material synthesized via HTE route is out of scope for rapid discovery of materials and only the best performing compositions are usually selected for further directed research. However, this approach ignores information regarding chemistry (composition and phase information)-performance trends that could be harnessed to identify pathways for accelerated discovery. In this presentation, I shall discuss a materials informatics based approach that provides a pathway to down select materials synthesized using HTE for solar fuels applications for further examination by directed research efforts.

Ten years have passed to the open public a free web-based material database MatNavi. Now, MatNavi is composed of eleven materials databases, Polymer DB (PoLyInfo), Inorganic DB (AtomWork), Computational Electronic Structure DB (CompES), Database of Promising Adsorbents for Decontamination of Radioactive Substance (READS), Neutron Transmutation DB (NeuTran), Interfacial Thermal Conductance DB (ITC), Diffusion DB (Kakusan), Superconducting Materials DB (SuperCon), Metallic Materials DB (Kinzoku), CCT Diagram DB (CCTD) and online versions of the structural materials datasheets on creep, fatigue, corrosion, and space use materials strength. MatNavi not only supports the foundations of materials science education in basic knowledge on crystals, phase diagrams, but also provides support for the development of new materials for building a safe and secure society and the selection and use of the optimum materials to researchers and engineers in private companies.
In the future, NIMS will continue to expand the data in MatNavi, and will implement a user-friendly system for disseminating useful information for material developers and users. In particular, to increase the data in the electronic structure(CompES) and Crystal Structural (AtomWork) to support the material design of such as electronic structure of semiconductors. Electronic structure data of the CompES from the results of first-principles calculations by VASP. MatNavi is possible to contribute to the features of Materials Informatics, to promote the development of new materials.

Twenty-five years ago the desktop computer started becoming ubiquitous in the scientific lab. Researchers were delighted with its ability to both control instrumentation and acquire data on a single system, but they were not completely satisfied. There were often gaps in knowledge that they thought might be gained if they just had more data and they could get the data faster. Computer technology has evolved in keeping with Moore&’s Law meeting those desires, however those improvements have of late become a double-edged sword for researchers. Computers are now capable of producing high speed data streams containing terabytes of information; capabilities that evolved faster than envisioned last century. Software to handle large scientific data sets has not kept up. How much information might be lost through accidental mismanagement or how many discoveries are missed through data overload are now vital questions. An important new task in most scientific disciplines involves developing methods to address those issues and to create the software that can handle large data sets with an eye towards scalability. This software must create archived, indexed, and searchable data from heterogeneous instrumentation for the implementation of a strong data-driven materials development strategy. At the National Center for Photovoltaics in the National Renewable Energy Laboratory, we began development a few years ago on a Laboratory Information Management System (LIMS) designed to handle lab-wide scientific data acquisition, management, processing and mining needs for physics and materials science data, and with a specific focus towards future scalability for new equipment or research focuses. We will present the decisions, processes, and problems we went through while building our LIMS system for materials research, its current operational state and our steps for future development.