Download Hi-Res ImageDownload to MS-PowerPointCite This:ACS Comb. Sci. 2017, 19, 1, 37-46

Research ArticleNovember 21, 2016

Automated Phase Mapping with AgileFD and its Application to Light Absorber Discovery in the V–Mn–Nb Oxide System
Click to copy article linkArticle link copied!

View Author Information

Open PDF Supporting Information (1)

ACS Combinatorial Science

Cite this: ACS Comb. Sci. 2017, 19, 1, 37–46

Click to copy citationCitation copied!

https://doi.org/10.1021/acscombsci.6b00153

Published November 21, 2016

Request reuse permissions

Abstract

Click to copy section linkSection link copied!

Rapid construction of phase diagrams is a central tenet of combinatorial materials science with accelerated materials discovery efforts often hampered by challenges in interpreting combinatorial X-ray diffraction data sets, which we address by developing AgileFD, an artificial intelligence algorithm that enables rapid phase mapping from a combinatorial library of X-ray diffraction patterns. AgileFD models alloying-based peak shifting through a novel expansion of convolutional nonnegative matrix factorization, which not only improves the identification of constituent phases but also maps their concentration and lattice parameter as a function of composition. By incorporating Gibbs’ phase rule into the algorithm, physically meaningful phase maps are obtained with unsupervised operation, and more refined solutions are attained by injecting expert knowledge of the system. The algorithm is demonstrated through investigation of the V–Mn–Nb oxide system where decomposition of eight oxide phases, including two with substantial alloying, provides the first phase map for this pseudoternary system. This phase map enables interpretation of high-throughput band gap data, leading to the discovery of new solar light absorbers and the alloying-based tuning of the direct-allowed band gap energy of MnV₂O₆. The open-source family of AgileFD algorithms can be implemented into a broad range of high throughput workflows to accelerate materials discovery.

Subjects

Keywords

Introduction

Click to copy section linkSection link copied!

Combinatorial materials science encompasses a suite of techniques for accelerated discovery of functional materials and often entails the establishment of relationships between crystal structure and materials properties. (1) As a result, high-throughput phase mapping, the generation of composition maps of phase concentrations that serve as a proxy for the solid state phase behavior and ultimately the phase diagram for a given synthesis condition, is an important tool for materials discovery, prompting the recent development of synchrotron-based X-ray diffraction (XRD) techniques to acquire 10²–10⁴ XRD patterns per day over a composition spread library. (2) More generally, the accelerated characterization of phase diagrams in high-order composition spaces is a foundational problem in materials science given that many binary phase diagrams are known but few ternary or higher-order phase diagrams have been explored due to the requisite number of experiments. Emboldened by the progress in previous decades in automated XRD analysis for small molecules, (3) materials scientists have adopted a variety of techniques in the past 10 years to automatically generate phase maps (phase concentrations as a function of composition) from XRD patterns, as recently summarized in a perspective article (4) noting the critical need for improved algorithms that appropriately capture the complexity of materials phase behavior. By developing algorithms at the forefront of artificial intelligence research, we have established a phase mapping tool that models alloying-based shifting of peaks in XRD patterns and incorporates physical constrains based on Gibbs’ phase rule. The algorithm performs factor decomposition with an elegant agility that allows users to constrain the spectral demixing based on expert knowledge of the phase behavior, yielding the powerful phase-mapping tool we hereby introduce as AgileFD.

The phase mapping tools developed to date employ a variety of intuitive approaches to the problem of formulating phase maps from a collection of XRD patterns, which generally contain a mixture of phase-pure patterns. Recent efforts have demonstrated that if an expert has sufficient knowledge of the phase map to generate training data, in particular expert-labeled XRD patterns for a variety of phases in the system, the trained machine learning model can map phase behavior. (5) An interesting feature of these supervised machine learning approaches is that phase classification is performed on each XRD pattern sequentially, whereas the remainder of the algorithms discussed below act on the ensemble of XRD patterns to demix multiphase patterns through simultaneous consideration of a wide range of phase mixtures. Clustering XRD patterns based on their similarity (6) is a powerful data reduction tool but since XRD patterns may be clustered due to either a common phase or a common phase mixture (i.e., when a set of composition samples are from the same phase field in the underlying phase diagram), the behavior of pure-clustering techniques is sensitive to the distance metric and requires delicate manual tuning. Initial efforts to address these issues by using constraint programming (CP) to inject physics-based reasoning into clustering have shown promise but have yet to be developed into an effective phase mapping algorithm for experimental data. (7) More recently, by performing clustering in a feature vector space that includes the XRD patterns and sample compositions, compositional clusters can be generated that more closely emulate phase diagrams, and this clustering can be performed quickly enough to perform on-the-fly analysis. (8)

An expert’s manual analysis of XRD patterns is most closely emulated with a reasoning-based algorithm in which constituent phases are identified by recognizing sets of peaks that coexist in connected composition regions. By expressing the phase mapping problem as a set of logical requirements within a satisfiability modulo theory (SMT) reasoning framework, we previously demonstrated automated phase mapping using pristine, synthetic data. (9) The implementation of this approach for experimental data suffers from practical limitations due to the computational expense for large data sets and the reliance on accurate peak detection in every XRD pattern. As a result, the development of scalable algorithms for (noisy) experimental data has focused on factoring a collection of XRD patterns into a small set of basis patterns (intended to represent individual phases) and the weighting coefficients or “activation” of the basis patterns for each experimental XRD pattern (intended to represent the amount of each phase present in the corresponding material). Since both XRD patterns and phase concentrations are nonnegative, the mathematical description of this approach is non-negative matrix factorization (NMF) in which the XRD patterns for the composition samples are assembled into a matrix A and the phase map solution corresponds to identifying matrices W (whose columns are the basis patterns) and H (which contains the activation of each basis pattern for each sample), such that

(1)

Recent progress in solving the phase mapping problem has thus included implementations of NMF that yield excellent reconstruction of the XRD data set and produce phase maps that adhere (as closely as possible) to the underlying materials physics.

The various NMF approaches can be succinctly summarized by considering their modeling of 3 principle properties of solid-state phase diagrams, particularly for composition libraries synthesized at a fixed temperature: (1) Gibbs’ Rule, in thermodynamic equilibrium, the number of coexisting phases in a given material can be no more than the number of components (number of cations in the case of metal oxides) it contains; (2) connectivity, the set of composition samples containing a given phase must be connected in composition space; and (3) peak shifting or alloying, the interstitial or substitutional solution of elements within a single phase can yield composition-dependent lattice constants that cause XRD patterns to shift as a function of composition. While alloying-based peak shifting can be quite complex for noncubic phases, in the present work we only consider isotropic lattice expansions that are manifested as uniform shifting of peaks in XRD patterns.

Long et al. (10) introduced NMF as a phase mapping strategy, and while this approach is computationally efficient and effective for some data sets, it does not model any of the above properties of phase diagrams, resulting in routine production of nonphysical and uninterpretable phase maps. The complement to bare NMF is CombiFD, (11) a powerful factor decomposition approach in which the matrix factorization can be performed under constraints of arbitrary complexity, including the encoding of the above phase diagram properties as hard combinatorial constraints. This approach is analogous to the logic of the SMT algorithm described above and is too computationally intensive for large data sets. GRENDEL (12) is an extension of NMF that was recently adapted for phase mapping by Kusne et al. (13) to enforce the Connectivity constraint by combining NMF with a graphical model of the composition space. GRENDEL remains computationally efficient by iterating between NMF and clustering with an objective function that combines the loss functions of these subalgorithms, which requires tuning of the relative weights of the loss functions to optimize clustering. Further development of the algorithm is required to enforce Gibbs’ Rule, and the reliance on NMF results in an inability to incorporate Peak Shifting. Indeed Peak Shifting has remained a primary challenge in phase mapping, and while powerful approaches for modeling Peak Shifting using dynamic time warping techniques have been proposed, (7, 14) they have yet to be incorporated into scalable phase mapping algorithms.

The AgileFD phase mapping algorithm is based on convolutional non-negative matrix factorization (cNMF), (15) an algorithm that is effective for demixing audio signals, which we have tailored to explicity incorporate Peak Shifting in the matrix factorization process. The key concept is to introduce multiple copies of the basis pattern for each phase, representing multiple alloys of the same phase but with different lattice constants. Gibbs’ Rule is also incorporated in AgileFD-Gibbs, an extension of AgileFD that includes constraint optimization of H using mixed integer programming (MIP). AgileFD also enables an expert user to inject knowledge of the phase map through “expert constraints”. We have yet to develop an explicit Connectivity constraint since the phase maps we have generated to date do not substantially violate this rule due to the efficacy of our Gibbs’ Rule enforcement. As discussed below, our experience has been that these minor Connectivity violations can be easily corrected by incorporating expert feedback into the solutions.

Perhaps the most powerful aspect of AgileFD is its ability to generate both unsupervised and expert-constrained solutions for sizable data sets (i.e., the data matrix A containing ∼10⁶ values corresponding to 100 s of XRD patterns each containing 1000 s of data points) in a few minutes, enabling the user to interact with solutions via novel data visualizers by altering the number of basis patterns and applying expert constraints until a satisfactory phase map is obtained. The typical analysis of a given XRD data set proceeds by running AgileFD in unsupervised mode to generate candidate phase mappings that are further tailored with constraints applied by expert analysts.

We demonstrate this process flow for a pseudoternary V–Mn–Nb oxide library in which six primary phases are automatically identified with basis patterns that closely match those of known phases. Manual identification of 2 additional, minor phases yields a slightly more refined phase map that is used to interpret high-throughput optical spectroscopy data. While solving the phase behavior of these previously unexplored compositions demonstrates the power of AgileFD, we additionally highlight its utility in the discovery of functional materials, in particular solar light absorbers.

Using the same V–Mn–Nb oxide library, we constructed a composition map of band gap energy using automated UV–vis spectroscopy (16) and Tauc analysis. (17) By correlating band gap energy with the phase concentration and shift parameter from the AgileFD-Gibbs solution, we have discovered new visible-band gap metal oxides and alloying-based tuning of the direct-allowed band gap energy. Such metal oxide semiconductors are critically missing in the quest to develop efficient solar fuel generators, (18) and the combination of automated phase and band gap mapping described herein will significantly bolster the recent acceleration in the discovery of light absorbers for solar energy applications and beyond. (19)

Algorithm and Experiments

Click to copy section linkSection link copied!

AgileFD Algorithm

In this article, we highlight the innovations of AgileFD, mainly focusing on its incorporation of physical constraints, which drastically differs from previous-reported approaches. For a comprehensive description of the mathematics of AgileFD, we refer the reader to ref. (20)

Perhaps the most substantial advancement of AgileFD for the phase mapping problem is the efficient modeling of Peak Shifting, which we enable by applying a log-transformation to the XRD patterns. Each XRD pattern can be represented as the scattering intensity (I) as a function of q, the magnitude of the X-ray scattering vector. The contraction of a crystal lattice due to alloying can be considered as a scaling of the lattice parameter (a → γa), which corresponds to the inverse scaling of the scattering vector magnitude of each Bragg peak in the XRD pattern (q_peak → γ^–1q_peak). For the present discussion, we consider lattice contraction (γ < 1) and note below that the method generalizes to lattice contraction and expansion. By performing a log-transformation of q, the peak shifting becomes (log q_peak → log γ^–1 + log q_peak), and the additive shift can be incorporated in the matrix representation of XRD patterns through a row-shifting operation that we define below. The numerical values of q are not used in matrix factorization, but the q-spacing between data points is important when modeling shifting. A consistent model of Peak Shifting is enabled by choosing a constant interval ρ in the log q space so that the abscissa of the XRD patterns is a geometric series of q values

(2)where the N XRD patterns correspond to N composition samples and are represented as the columns of the input data matrix A. By choosing a value of L and using the same range of q values as the source data, these patterns are calculated by resampling the raw XRD patterns, and we typically choose L to be the number of data points in the raw XRD patterns, which is 2082 in the present work. A smaller value of L (larger value of ρ) can be chosen to accelerate calculations at the expense of q resolution. The logarithmic resampling results in decreasing q resolution as q increases, which may prompt the use of a larger value of L (smaller value of ρ) to retain the q resolution of the raw data. As noted below, the value of ρ also determines the peak shifting resolution.

In traditional NMF, the factorization produces K basis patterns W_k, where K is the number of phases used to describe the material system. Our innovative model to account for shifting is enabled by using K × M patterns in W that consist of M shifted versions of the basis pattern of each phase. Starting with the “non-shifted” version W_k,0, the additional M – 1 shifted versions are automatically generated in each optimization step by the row-shifting operation:

(3)

Each shifted version corresponds to a lattice parameter contraction by a factor of γ = ρ^–m, which means that compared to the highest (or lowest) lattice parameter realized in the composition library, lattice contraction (or expansion) by ρ^–(M–1) (or ρ^(M–1)) can be appropriately modeled by activating the corresponding shifted version of the basis pattern. That is, phase k is activated by N × M coefficients (H_n,k,m) in the matrix H such that the total activation of phase k for sample n is

(4)

Since multiple terms in this sum can be nonzero, multiple shifted versions of each basis pattern can be used to create a pseudocontinuous model of peak shifting as long as the Bragg peaks are relatively wide (Δq_peak > ρ × q_peak), which corresponds to using a sufficiently large value of L such that each peak spans over several data points. The representative shift parameter for a given sample and phase is calculated as a weighted mean

(5)which is the expectation value of ρ^m for each sample and for each basis pattern given the corresponding distribution of activation values.

While this construction of peak shifting is quite intuitive, the pattern shifting cannot be directly incorporated into NMF, which is one of the primary motivations for using cNMF as a starting point in developing AgileFD. Candidate solutions (typically starting with random seeding of H and W) are improved in AgileFD through scalable update rules where the loss function is translated into multiplicative matrix operations that enforce gradient descent. In ref. (20) we derive AgileFD’s custom update rules for both H and W, which are derived from a loss function based on the generalized Kullback–Leibler divergence, an advancement that is critical for the phase mapping problem and is generally applicable for other source separation problems. With these lightweight update rules, this incorporation of Peak Shifting is vastly more efficient than the previously proposed time warping techniques. (7, 14) The asymptotic time complexity for each solution update is O(KLMN), or more precisely Θ(KLMN). No explicit bound exists for the number of updates needed until the completion of the gradient descent algorithm. The calculation time increases with decreasing convergence tolerance, requiring the user to define a convergence criterion that appropriately balances data reconstruction quality and convergence time, as discussed further below.

While we have developed and continue to develop a variety of update rules to customize the loss function and impose various constraints, we focus on two additional types of constraints in the present work, namely imposing additional expert constraints and enforcing Gibbs’ Rule. Expert constraints comprise a variety of methods in which an expert can inject knowledge of the phase behavior, and AgileFD provides convenient mechanisms for encoding constraints through appropriate initialization of H and W. The multiplicative update rules have the convenient property that if any matrix value is zero in a candidate solution, it will remain zero throughout AgileFD convergence, which makes incorporation of certain constraints as simple as initializing specific matrix elements to zero. For example, to restrict the amount of shifting for a select basis pattern, the user can initialize the H values corresponding to the activation of some of its shifted copies to zero. To restrict the activation of a basis pattern to a select composition region, the user can make its activation zero for all other samples. A novel aspect of AgileFD is that imposing such constraints is substantially more computationally efficient than the (more expressive) incorporation of constraints in combinatorial reasoning algorithms such as CombiFD. (11) The most commonly desired expert constraint for W is to predefine the basis pattern for a phase that is known to exist in the composition library. The corresponding basis pattern can be incorporated in the initialization of W, and by not updating that portion of the matrix, the basis pattern is “frozen” in the AgileFD solution.

Gibbs’ Rule is incorporated into AgileFD through an extension called AgileFD-Gibbs. In terms of the h_n,k values, this rule corresponds to restricting the number of nonzero h_n,k to be no greater than the number of components in the composition of sample n. The full Gibbs phase rule is actually more restrictive than the one implemented in AgileFD-Gibbs, but we find that restricting the number of coexisting phases according to the number of components is sufficient for practical use. The lack of enforcement of Gibbs’ Rule has important consequences on the solution that cannot be ameliorated ex post facto. That is, a candidate solution can be “corrected” by zeroing the smallest h_n,k values for each n until the solution satisfies Gibbs’ Rule, but this guarantees that the corresponding basis patterns are not optimal and that the reconstructed data set may be “missing” Bragg peaks that are in the source data. More generally, this simple “correction” algorithm may not generate optimal solutions even after additional optimization of the basis patterns. To more elegantly and optimally enforce Gibbs’ Rule, we combine AgileFD with a MIP algorithm that modifies a candidate solution. While W is fixed, the MIP algorithm generates a new activation matrix by finding the optimal K × M activation coefficients for each sample that adhere to Gibbs Rule, and the resulting candidate solution is then further optimized using AgileFD. The MIP algorithm is applied to each sample independently, and since the complexity of the MIP is bound by the number of allowed phases in Gibbs’ Rule, which is 3 in the present work and generally a small number, the most prominent scaling for the execution of the MIP step is its proportionality to N, the number of samples. It is worth noting the similarity between the utility of the MIP algorithm and the expert constraints noted above; the MIP-modified solution assigns the appropriate number of values in H to zero such that subsequent candidate solutions adhere to Gibbs’ Rule.

With an AgileFD or AgileFD-Gibbs solution in hand, the collection of basis patterns W_k,0 and the composition maps of h_n,k and s_n,k provide an intuitive visualization of the phase behavior. Further quantification of phase concentrations and lattice parameters require identification of the crystal structure for each basis pattern, which was performed in the present work through search and match with the entries in the International Center for Diffraction Data (ICDD) database using DIFFRAC.SUITE EVA software. Using relative intensity ratios and tabulated Bragg peaks from the ICDD entries (see Supporting Information for details), the relative total scattering intensity per mole of metal (ϑ_k) was calculated for each phase, enabling the relative molar activation of each phase in each sample to be calculated as

(6)where the integral over the basis pattern is analogous to the integration over the ICDD pattern. The k phase fractions for each sample are then provided through normalization of these relative molar activations:

(7)which is the solution to the phase map problem under the approximation that the total relative scattering factor of each phase matches that of the ICDD pattern and does not vary substantially with alloy composition within the phase.

To provide a more intuitive visualization of the weighted shift parameters, s_n,k, they are converted to lattice expansion parameters γ_n,k = s_n,k^–1. By determining the lattice expansion parameter γ_k,ICDD that corresponds to the shift of W_k,0 that best matches the respective ICDD pattern, the relative shift of each phase in each sample compared to its ICDD entry is the ratio of γ_n,k to γ_k,ICDD.

Library Synthesis

The continuous composition spread of V, Mn, and Nb was synthesized by reactive magnetron cosputtering in the presence of O₂ and Ar gas using elemental sources arranged symmetrically with respect to the 100 mm Si/SiO₂ substrate. The deposition proceeded for about 10 h with the RF power of the V, Mn, Nb source fixed at 150, 115, and 81 W, respectively. The spatial variation in deposition rate from each source resulted in a continuous thin film with composition gradient of the order of 0.5 at. % mm^–1. The variation in the deposition rates among different sources resulted in thickness variation across the library within a factor of 2 of the center thickness of 400 nm. The as-deposited oxide composition library was oxidized in air at 883 K for 1 h, producing the (V–Mn–Nb)O_x library for which oxygen composition is unknown and samples are represented by their cation composition.

Composition and Structure Measurements and Analysis

XRD data was acquired using a custom HiTp setup incorporated into the bending-magnet beamline 1–5 of the Stanford Synchrotron Radiation Light source (SSRL) at SLAC National Accelerator Laboratory. A detailed description of the experiment is provided by Gregoire et al., (2a) and the characterization of the (V–Mn–Nb)O_x library employed a monochromated 13 keV source in reflection scattering geometry with an XRD image detector (Princeton Quad-RO 4320). Using a 1 mm² footprint, N = 317 samples were acquired on a square grid with 4.5 mm pitch and within a radius of 45 mm from the center of the 100 mm-diameter library. Diffraction images were processed into XRD patterns, I_n(q), using WxDiff software (21) and further processed with a custom background subtraction algorithm using cubic splines.

The XRF measurements were performed on an EDAX Orbis Micro-XRF system (EDAX Inc., Mahwah, NJ) with an X-ray beam approximately 1 mm in diameter. The V K, Mn K and Nb K XRF intensities were extracted from the Orbis software and converted to normalized V–Mn–Nb compositions using relative sensitivity factors calibrated at substrate center via a separate composition measurement. The calibration composition was measured on an Oxford Instruments X-Max 80 mm² energy dispersive X-ray spectroscopy (EDS; Oxford Instrument, Concord, MA) detector on a FEI Nova NanoSEM 450 (FEI, Hillsboro, OR). The absolute composition uncertainty for this EDS measurement is 10 at. %, and the XRF-determined relative compositions, which enable composition-property maps, have approximately 1 at. % resolution. (4, 22)

Optical Characterization

Optical characterization of the (V–Mn–Nb)O_x library was performed on a custom HiTp diffuse reflectance (DR) instrument built to analyze light absorber thin films on opaque substrates. The computer-automated experiment employed a 200 W Hg(Xe) lamp (Newport/Oriel Apex) and an integrating sphere (Ocean Optics ISP-50-8-R-GT) fiber-coupled to a spectrometer (Spectral Products, Inc. SM303), with further details provided by Mitrovic at et al. (16) DR spectra were acquired on a square grid of 1521 positions with a pitch of 2.032 mm, spot size of approximately 1 mm, and integration time of 0.025 s per spectrum, and typically three spectra were averaged to produce the DR spectrum for each sample. The absorption coefficient (α) was calculated up to a factor of a spectral scattering factor, which we approximate to be energy-independent, using the Kubelka–Munk radiative transfer model (23) which enabled calculation of the normalized Tauc property

(8)where n is the Tauc exponent whose value is 1/2 for analysis of direct-allowed (DA) electronic transitions. The plots of DA Tauc property (TP^DA) versus photon energy (hν) were not only automatically generated but also automatically interpreted using a recently developed algorithm (17b) that mimics expert judgment. The algorithm either estimates the DA band gap energy or determines that the band gap cannot be confidently estimated from the Tauc plot, which can occur when the band gap energy is beyond the range of the spectrometer, the sample is not sufficiently absorbing, or the direct-allowed band gap signature is convoluted by the presence of multiple phases with comparable contributions to the DR signal. We note that for discovery of photoabsorbers, indirect-allowed (IA) transitions are of primary interest only when the IA band gap is substantially lower than DA band gap. For the DA results shown here, no IA band gaps were identified at energies more than 0.3 eV below the direct-allowed gap.

The optical and XRF measurements were performed on the same set of library samples. The XRD measurements were performed on a coarser grid of library positions, and the composition for each XRD sample was calculated using linear interpolation in the Cartesian library position space. The results of XRD analysis (γ_n,k and P_n,k) were interpolated to the optical and XRF samples through linear interpolation in the ternary composition space. The composition spread library and compositions of the samples characterized are shown in Figure 1.

Figure 1. White light image of the (V–Mn–Nb)O_x composition library on 100 mm Si/SiO₂ substrate is shown with elemental labels indicating the orientation of the sputter deposition sources. The grid of library locations for XRD (blue) and both XRF and optical characterization (green) are shown along with an additional plot showing the XRF-determined compositions of these points.

Results and Discussion

Click to copy section linkSection link copied!

The phase behavior of the (V–Mn–Nb)O_x library was analyzed using AgileFD-Gibbs in unsupervised mode and then with additional expert constraints. Manual visualization of the XRD patterns revealed substantial Peak Shifting, prompting our use of M = 10 shifted copies of each basis pattern, which corresponds to a maximum lattice expansion of 1.0072 with respect to the basis pattern with the lowest lattice constant. Several values of K were attempted, revealing that the K = 6 solution appeared physically sound and that higher values of K neither substantially improved the data reconstruction nor produced basis patterns that were distinctive from those of the K = 6 solution. A more objective determination of appropriate values for parameters such as K and M is the subject of ongoing research and is beyond the scope of the present work.

Using the basis patterns from the K = 6 solution, we performed phase matching and identified the first 6 phases in Table 1. That is, upon fixing the values of K and M, the unsupervised AgileFD-Gibbs algorithm produced 6 basis patterns that match those of known phases despite substantial overlap of the basis patterns, which is a testament to the excellent source separation enabled by the update rules and Gibbs rule constraint. To demonstrate how the identification of 6 phases was enabled by the Peak Shifting feature of AgileFD, we show the AgileFD M = 1 and AgileFD-Gibbs M = 10 solutions in Figure 2. With M = 1, the dimension of the activation matrix H is equivalent to that of NMF, motivating our labeling of this solutions as “∼NMF”, but the algorithms are not equivalent because of both the log transformation and gradient descent with custom update rules used in AgileFD. While some basis patterns and their phase distributions are similar between the AgileFD-Gibbs and ∼NMF solutions, the lack of peak shift modeling has an important consequence in the ∼NMF solution. Inspection of the k = 5 basis pattern in Figure 2 reveals that while the AgileFD-Gibbs solution is an excellent match to the ICDD pattern, the ∼NMF pattern is a very poor match. Instead of separating this ICDD phase, the data is better reconstructed in the ∼NMF solution by using 2 basis patterns (k = 2 and 5) to model the k = 2 basis pattern of the AgileFD-Gibbs solution. The k = 2 and 5 basis patterns in the ∼NMF solution are essentially shifted versions of each other and activated in complementary portions of the composition space. The ∼NMF solution fails to identify the k = 5 phase from the AgileFD-Gibbs solution as a direct consequence of the inability to model Peak Shifting.

Figure 2. Six-phase solutions for the (V–Mn–Nb)O_x library with algorithm parameters noted in the left-hand labels. The ∼NMF solution is the AgileFD solution with the peak shifting removed by setting M = 1. The basis patterns are plotted along with the ICDD patterns listed in Table 1. The map of each phase is shown as a composition plot where the point size represents the phase fraction P_n,k and the color represents the relative lattice constant compared to the respective basis pattern, which is aligned to the best-match with the ICDD pattern.

Table 1. Primary Phases in the AgileFD-Gibbs Solutions for K = 6 and 8 Along with Their ICDD Entries^a

k, phase index	formula unit (crystal system)	ICDD entry number	relative total scattering intensity per mole of metal (ϑ_k)
0	Mn₂O₃ (cubic)	01-071-0636	923.2
1	V_2.38Nb_10.7O_32.7 (orthorhombic)	01-079-8393	2195.7
2	MnNb₂O₆ (orthorhombic)	01-072-0484	2195.2
3	Mn₃V₂O₈ (unknown)	00-039-0091	739.2 (26)
4	MnV₂O₆ (monoclinic)	01-072-1837	911.8
5	Mn₃O₄ (tetragonal)	01-080-0382	964.5
6	NbVO₅ (orthorhombic)	00-046-0046	1706.2 (27)
7	V₂O₅ (orthorhombic)	00-041-1426	678.0

ICDD entries are overlaid in the basis pattern plots of Figures 2 and 3. The ϑ_k values used to calculate phase fractions are also shown.

Additional subtle but important difference exist between the ∼NMF and AgileFD-Gibbs solution, and perhaps the illustrative example for the present discussion is that the AgileFD-Gibbs solution adheres to the connectivity requirement much more closely than the ∼NMF solution even though this constraint was not imposed. The enforcement of the Gibbs rule has important implications for optimizing not only the phase distributions but also the basis patterns. It is important to note that the quality of the phase map solution lies in its explanation of physically meaningful phase behavior, which requires (at a minimum) that the basis patterns represent solid state phases with plausible phase distributions in composition space. Considering the reconstruction error for each sample to be the sum of the absolute residuals (E), then with the total intensity in the XRD pattern (T), the fractional reconstruction accuracy can be calculated as 1 – E/T. Averaged over the samples, the ∼NMF and AgileFD-Gibbs solutions have 73% and 79% accuracy, respectively, but the improvements in the AgileFD-Gibbs solution are not sufficiently represented by this increase in reconstruction accuracy. Indeed a challenging aspect of the phase demixing problem is that a multitude of solutions provide similar reconstruction accuracy, requiring the incorporation of physical constraints to yield a solution that is both accurate and physically meaningful.

The AgileFD-Gibbs solution in Figure 2 reveals that the lattice constant for several phases varies systematically through the composition space, a strong indicator of alloying. Upon careful inspection of the AgileFD-Gibbs solution we also noticed some opportunities to further improve the phase map via expert constraints in AgileFD-Gibbs. We proceeded to manually inspect the compositions closest to the end-member binary oxides. For Mn oxide the AgileFD-Gibbs solution produced 2 different Mn oxide phases (k = 0 and 5). We have commonly observed the coexistence of these phases in Mn-containing libraries of metal oxides due to the lack of a strong thermodynamic differentiation under the library synthesis conditions (Gibbs free energies of formation of −359 and −356 kJ mol^–1 per Mn atom for Mn₂O₃ and Mn₃O₄, respectively). Since the library compositions contain substantial amounts of Nb and V that may alloy into these phases, the alloyed metal oxides may indeed coexist in thermodynamic equilibrium in this ternary composition space. Through manual inspection of XRD patterns containing these phases, we confirmed that both phases are present and in fact they often coexist. It is worth noting that without prior knowledge of the XRD patterns of pure phases, any source separation algorithm cannot robustly produce phase-pure basis patterns for phases that always coexists. For the Mn oxides, AgileFD-Gibbs provides sufficient separation of the phases that they were identified in ICDD search and match, and we enforced complete separation by seeding W with simulated basis patterns obtained by convoluting the ICDD patterns with a Gaussian peak shape (σ = 0.13 nm^–1). Inspection of the most Nb-rich samples revealed that, as suggested by Figure 2, the library samples are far enough from the Nb end-member composition that no Nb oxides are observed. Inspection of the V-rich composition revealed the presence of 2 minor phases, V₂O₅ and NbVO₅, whose relatively weak intensity and existence in only a small fraction of the composition samples gives their signals little contribution to the system-wide loss function. While AgileFD-Gibbs may factor these phases with an extended data set that includes compositions closer to the stoichiometries of these phases, we amended this shortcoming of the source data by introducing and freezing a new basis pattern for both V₂O₅ and NbVO₅. Since these phases were only observed in V-rich compositions, we additionally seeded the corresponding values of H with zeros for all samples with V concentration less than 0.45. The expanded K = 8 solution with expert constraints on four phases was generated using AgileFD-Gibbs, producing the solution shown in Figure 3.

Figure 3. AgileFD-Gibbs solution with K = 8, M = 10, and expert constraints applied to k = 0, 5, 6, 7. The map of each phase is shown as a composition plot where the point size represents the phase fraction P_n,k and the color represents the relative lattice constant, as described in Figure 2.

The addition of 2 phases and expert constraints enabled more accurate modeling of every phase, leading to slight changes (compared to the AgileFD-Gibbs solution of Figure 2) in the basis patterns and phase maps even for the phases that were not constrained. The agreement between k = 1, 2, 4 and the respective basis patterns is remarkable, and the agreement for k = 3 is quite good but not perfect. Upon manual inspection of samples with high activation of this phase, we believe this basis pattern accurately models the complex phase behavior in this composition region. It appears that the Mn₃V₂O₈ coexists with minor fractions of 2 polytypes of Mn₂V₂O₇. We previously reported (24) on the coexistence of these polytypes in thin film (V–Mn)O_x libraries and did not pursue further disambiguation in the phase map solution since all 3 of these manganese vanadates only appear as minor phases since the composition samples are all beyond ∼10 at. % of the V–Mn binary line.

With the phase map of Figure 3 in hand, we turn to the optical characterization data to investigate if the composition library contains any metal oxide phases of interest for solar light absorption. While the automated Tauc analysis algorithm did not identify a DA transition for some samples, DA band gap energies (E_g^DA) were estimated for 1329 samples and are mapped in composition space in Figure 4. While comparisons of the band gap data and phase map reveal a number of interesting structure–property relationships, for the present purposes we provide a detailed interpretation of MnV₂O₆, corresponding to k = 4 in Figure 3. It is worth noting that for mixed-phase samples, the majority phase does not necessarily provide the majority contribution to the Tauc plot due to possible differences in the optical absorption strength and band gap energies of different phases. To infer which composition samples exhibit a band gap representative of MnV₂O₆, we show the band gap energy as a function of MnV₂O₆ phase fraction in Figure 5, in which the sample points are colored according to the relative lattice constant from the k = 4 phase map of Figure 3. With low phase fraction, the band gap energy is most likely due to a different phase, so inspecting the convergence of the band gap energy as the phase fraction increases provides an indication of the phase fraction at which MnV₂O₆ is providing the major contribution to the band gap measurement. Figure 5 reveals that samples with phase fraction in excess of 0.8 exhibit a band gap in the 1.8–2.08 range, and that the variation within this range is correlated with the lattice parameter.

Figure 4. Composition map of the DA band gap energy from automated Tauc analysis is shown for 1329 composition samples.

Figure 5. DA band gap energy is plotted against the k = 4 phase fraction from Figure 3, demonstrating that when the phase fraction of MnV₂O₆ is above 0.8, the band gap value systematically varies with the relative lattice parameter.

To further investigate this relationship between band gap and lattice parameter in MnV₂O₆, we extract the samples with phase purity in excess of 0.8 and visualize in Figure 6 both the original XRD patterns and the relative lattice parameter from AgileFD-Gibbs. This visualization provides a more detailed understanding of the phase behavior in this composition region. At the lowest V concentration (composition V_0.19Mn_0.38Nb_0.43O_x), the lattice parameter is approximately 1.009 times larger than the ICDD entry and as the V concentration is increased to 0.5, the relative lattice constant systematically lowers and trends toward 1 as the cation composition gets closer to the Mn_0.33V_0.67 composition of stoichiometric MnV₂O₆. The composition library does not extend to this composition, and in the composition region noted in Figure 6, there is a phase boundary at V concentration of 0.5 where the alloying of MnV₂O₆ does not continue into more V-rich compositions and this phase becomes mixed with NbVO₅. The highlighted composition region in Figure 6 contains phase-pure MnV₂O₆ (within the detectability limit of other phases) and exhibits alloying-based peak shifting with respect to both Mn and Nb, suggesting that the Mn concentration varies by ±4 at. % from the phase stoichiometry and that Nb is soluble up to approximately 25 at. %. For this noncubic crystal structure, the lattice constant determined by AgileFD-Gibbs is an “average” lattice constant but not necessarily the mean of the three lattice parameters. Figure 6a shows that most peaks shift uniformly with some indication of nonisotropic lattice expansion that we do not model in the present work and may provide insights into the elemental site substitutions of this oxide alloy.

Figure 6. (a) For the 26 samples with phase fraction of MnV₂O₆ in excess of 0.8, the series of XRD patterns (in the q-range with primary ICDD peaks) are arranged according to the V concentration. The top 10 patterns show a small amount of NbVO₅, and the bottom 16 patterns exhibit systematic shifting of the MnV₂O₆ peaks with respect to V concentration. (b) The relative lattice constant for the 26 samples is shown and the 16 samples with no NbVO₅ are indicated by a gray region. (c) The 26 samples are shown in a composition plot with end-members V′ = V_0.61Mn_0.29Nb_0.09O_x, Mn′ = V_0.43Mn_0.47Nb_0.1O_x, and Nb′ = V_0.43Mn_0.29Nb_0.28O_x. The composition gray composition region contains the same 16 samples as that in b. The data for each sample is colored according to its composition in all 3 plots.

To investigate the impact of alloying on band gap energy, we consider the samples in the composition region highlighted in Figure 6. As shown in Figure 7, the band gap energy of these samples spans a nearly 0.2 eV range and systematically increases with increasing relative lattice parameter, which is highly correlated with composition as shown in Figure 6. We note that while the band gaps of these oxide alloys have not been previously reported, the trend in Figure 7 is commensurate with reported direct-allowed band gap value of 1.8–1.95 eV for MnV₂O₆ (where the relative lattice parameter is 1). (25) This alloying-induced variation in the band gap energy is well-studied in semiconductors based on main-group elements but is rarely observed in metal oxides, particularly in the 1.8–2.1 eV range that is of primary interest for photoanodes in solar hydrogen generators and other applications. (18) While substantial additional research is required to fully understand the alloying and band gap tuning in V–Mn–Nb oxides, the results presented in Figures 2–7 highlight the excellent performance of the AgileFD suite of algorithms and their enabling ability to conduct materials science research in high order composition spaces.

Figure 7. (a) Band gap values from the composition region containing high phase purity MnV₂O₆ (see Figure 6) are plotted against the relative lattice parameter, demonstrating alloying-based tuning of the band gap energy. (b) Sample compositions using the same composition-color scale revealing that among these samples, the highest band gap energy and lattice parameters are found with the most Nb-rich compositions. (c) Representative Tauc plots for 3 samples with line colors matching the samples’ colors in b. The band gap energies produced by the automated Tauc algorithm are listed in the legend.

Since the matrix factorization in AgileFD proceeds via gradient descent from randomized initial matrices (other than the matrix elements set according to expert constraints), different randomizations may require different calculation times to reach the convergence criterion and may yield different solutions. It is important to ensure that a solution is “stable” by comparing the results from several random initializations, as was done for all solutions presented in the present work. Solutions equivalent (within negligible variations related to random initialization) to that of Figure 3 can be generated using single or multithreaded processing with convergence times of approximately 1200 and 370 s, respectively (Intel x5690 3.46 GHZ processor), which in the multithreaded case is approximately equally distributed among the initial matrix factorization, the MIP solution for applying Gibbs’ Rule, and the subsequent matrix factorization. For this publication-quality solution, a stringent convergence criterion was used wherein the matrix updates proceeded until the change in the loss function dropped below 0.008%. It is worth noting that for initial exploration of data, a typical convergence threshold of 0.05% can be used, in which case the AgileFD matrix factorization completes in approximately 50 s for this sizable data set with KLMN = 52.8 million matrix elements.

In closing, we note some limitations of phase mapping techniques and algorithms that may be addressed in future research. It is worth noting that sputter deposition is a nonthermal synthesis technique, and even with postdeposition annealing of a library such as that described here, deviations from thermodynamic equilibrium may persist. As noted above, the coexistence of Mn₂O₃ and Mn₃O₄ may be an example of such nonequilibrium phase behavior, requiring that constraints based on Gibbs’ phase rule must be applied with care. To address this potential issue we routinely generate an AgileFD-Gibbs solution with an additional allowed phase (4 coexisting phases for ternary compositions) for comparison. Concerning the modeling of alloying-based peak shifting, AgileFD algorithm is somewhat limited as it is intended to track peak shifting resulting from isotropic lattice distortions. More complex alloying-based distortions may result in different Bragg peaks shifting by different amounts and even in different directions. Such peaks shifting can only be modeled in AgileFD by splitting the pattern of a phase into subpatterns representing sets of peaks that shift together, and through recognition that these “sub” basis patterns are activated over the same composition ranges, the full pattern can be reconstructed in post processing. This approach to modeling peak shifting is not commensurate with Gibbs’ phase rule and is not a sufficiently general model for alloying in noncubic systems, motivating further algorithm development. Finally, we noted above that adherence to the connectivity constraint is a general requirement for meaningful solutions, which is not enforced in AgileFD. In the V–Mn–Nb oxide system described here, we demonstrated that the AgileFD-Gibbs algorithm sufficiently addresses the connectivity shortcomings of the ∼NMF algorithm, but indeed a connectivity constraint is a desirable extension of AgileFD that will be addressed in future work. The source data can be found in the Supporting Information and at http://www.udiscover.it/resources/data/. The source code will be hosted at http://www.udiscover.it/resources/software/.

Summary

Click to copy section linkSection link copied!

We introduce AgileFD, the first scalable phase mapping algorithm for combinatorial XRD data that models alloying-based peak shifting and imposes Gibbs’ phase rule. The importance of encoding these properties of solid state phase diagrams into a source separation algorithm is discussed using the (V–Mn–Nb)O_x system where unsupervised mapping of six phases is directly enabled by these capabilities of AgileFD. Several of the 317 XRD patterns from the composition library contain small signals from minor phases that are not amenable to unsupervised factorization but can be modeled by AgileFD through straightforward injection of expert knowledge. The resulting 8-phase solution reveals alloying in ternary oxide phases such as MnV₂O₆. By combining the AgileFD solution with band gap energies obtained from automated Tauc analysis of high throughput UV–vis spectroscopy data, we identify band gap tuning of nearly 0.2 eV as a function of lattice parameter and V composition in the energy range of interest for solar applications. The identification of this family of promising solar light absorbers was enabled through a multidisciplinary effort in which materials-motivated advancements of computer science techniques produced powerful new algorithms for accelerating materials discovery.

Supporting Information

Click to copy section linkSection link copied!

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acscombsci.6b00153.

XRD and composition data for the V–Mn–Nb oxide system and additional source data (ZIP)

co6b00153_si_001.zip (67.34 MB)

Terms & Conditions

Most electronic Supporting Information files are available without a subscription to ACS Web Editions. Such files may be downloaded by article for research use (if there is a public use license linked to the relevant article, that license may permit other uses). Permission may be obtained from ACS for other uses through requests via the RightsLink permission system: http://pubs.acs.org/page/copyright/permissions.html.

Author Information

Click to copy section linkSection link copied!

Corresponding Authors
- Carla P. Gomes - Department of Computer Science, Cornell University, Ithaca, New York 14850, United States; Email: [email protected]
- John M. Gregoire - Joint Center for Artificial Photosynthesis, California Institute of Technology, Pasadena California 91125, United States; http://orcid.org/0000-0002-2863-5265; Email: [email protected]
Authors
- Santosh K. Suram - Joint Center for Artificial Photosynthesis, California Institute of Technology, Pasadena California 91125, United States
- Yexiang Xue - Department of Computer Science, Cornell University, Ithaca, New York 14850, United States
- Junwen Bai - Zhiyuan College, Shanghai Jiao Tong University, Shanghai, China
- Ronan Le Bras - Department of Computer Science, Cornell University, Ithaca, New York 14850, United States
- Brendan Rappazzo - Department of Computer Science, Cornell University, Ithaca, New York 14850, United States
- Richard Bernstein - Department of Computer Science, Cornell University, Ithaca, New York 14850, United States
- Johan Bjorck - Department of Computer Science, Cornell University, Ithaca, New York 14850, United States
- Lan Zhou - Joint Center for Artificial Photosynthesis, California Institute of Technology, Pasadena California 91125, United States
- R. Bruce van Dover - Department of Materials Science and Engineering, Cornell University, Ithaca, New York 14850, United States
Notes
The authors declare no competing financial interest.

Acknowledgment

Click to copy section linkSection link copied!

The experimental work was performed in the Joint Center for Artificial Photosynthesis (JCAP), a DOE Energy Innovation Hub supported through the Office of Science of the U.S. Department of Energy under Award No. DE-SC0004993. The algorithm development is supported by NSF awards CCF-1522054 and CNS-0832782 (Expeditions), CNS-1059284 (Infrastructure), and IIS-1344201 (INSPIRE); and ARO award W911-NF-14-1-0498. Use of the Stanford Synchrotron Radiation Lightsource, SLAC National Accelerator Laboratory, is supported by the U.S. Department of Energy, Office of Science, Office of Basic Energy Sciences under Contract No. DE-AC02-76SF00515. The authors thank Apurva Mehta and Douglas G. Van Campen for assistance with collection of synchrotron XRD data.

References

Click to copy section linkSection link copied!

This article references 27 other publications.

1
Green, M. L.; Takeuchi, I.; Hattrick-Simpers, J. R. Applications of high throughput (combinatorial) methodologies to electronic, magnetic, optical, and energy-related materials J. Appl. Phys. 2013, 113 (23) 231101– 231101 DOI: 10.1063/1.4803530
Google Scholar
1
Applications of high throughput (combinatorial) methodologies to electronic, magnetic, optical, and energy-related materials
Green, Martin L.; Takeuchi, Ichiro; Hattrick-Simpers, Jason R.
Journal of Applied Physics (Melville, NY, United States) (2013), 113 (23), 231101/1-231101/53CODEN: JAPIAU; ISSN:0021-8979. (American Institute of Physics)
A review. High throughput (combinatorial) materials science methodol. is a relatively new research paradigm that offers the promise of rapid and efficient materials screening, optimization, and discovery. The paradigm started in the pharmaceutical industry but was rapidly adopted to accelerate materials research in a wide variety of areas. High throughput expts. are characterized by synthesis of a "library" sample that contains the materials variation of interest (typically compn.), and rapid and localized measurement schemes that result in massive data sets. Because the data are collected at the same time on the same "library" sample, they can be highly uniform with respect to fixed processing parameters. This article critically reviews the literature pertaining to applications of combinatorial materials science for electronic, magnetic, optical, and energy-related materials. High throughput methodologies will facilitate commercialization of novel materials for these critically important applications. Despite the overwhelming evidence presented in this paper that high throughput studies can effectively inform com. practice, in our perception, it remains an underutilized research and development tool. Part of this perception may be due to the inaccessibility of proprietary industrial research and development practices, but clearly the initial cost and availability of high throughput lab. equipment plays a role. Combinatorial materials science has traditionally been focused on materials discovery, screening, and optimization to combat the extremely high cost and long development times for new materials and their introduction into commerce. Going forward, combinatorial materials science will also be driven by other needs such as materials substitution and exptl. verification of materials properties predicted by modeling and simulation, which have recently received much attention with the advent of the Materials Genome Initiative. Thus, the challenge for combinatorial methodol. will be the effective coupling of synthesis, characterization and theory, and the ability to rapidly manage large amts. of data in a variety of formats. (c) 2013 American Institute of Physics.
>> More from SciFinder ^®
https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADC%252BC3sXpsFSjsbo%253D&md5=ad8a63053af7c06469a6d89f74296c49
2
(a) Gregoire, J. M.; Van Campen, D. G.; Miller, C. E.; Jones, R.; Suram, S. K.; Mehta, A. High Throughput Synchrotron X-ray Diffraction for Combinatorial Phase Mapping J. Synchrotron Radiat. 2014, 21 (6) 1262– 1268 DOI: 10.1107/S1600577514016488
Google Scholar
2a
High-throughput synchrotron X-ray diffraction for combinatorial phase mapping
Gregoire, J. M.; Van Campen, D. G.; Miller, C. E.; Jones, R. J. R.; Suram, S. K.; Mehta, A.
Journal of Synchrotron Radiation (2014), 21 (6), 1262-1268CODEN: JSYRES; ISSN:1600-5775. (International Union of Crystallography)
Discovery of new materials drives the deployment of new technologies. Complex technol. requirements demand precisely tailored material functionalities, and materials scientists are driven to search for these new materials in compositionally complex and often non-equil. spaces contg. three, four or more elements. The phase behavior of these high-order compn. spaces is mostly unknown and unexplored. High-throughput methods can offer strategies for efficiently searching complex and multi-dimensional material genomes for these much needed new materials and can also suggest a processing pathway for synthesizing them. However, high-throughput structural characterization is still relatively under-developed for rapid material discovery. Here, a synchrotron X-ray diffraction and fluorescence expt. for rapid measurement of both X-ray powder patterns and compns. for an array of samples in a material library is presented. The expt. is capable of measuring more than 5000 samples per day, as demonstrated by the acquisition of high-quality powder patterns in a bismuth-vanadium-iron oxide compn. library. A detailed discussion of the scattering geometry and its ability to be tailored for different material systems is provided, with specific attention given to the characterization of fiber textured thin films. The described prototype facility is capable of meeting the structural characterization needs for the first generation of high-throughput material genomic searches.
>> More from SciFinder ^®
https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADC%252BC2cXhvVakur7P&md5=8c3849a28e4f06cfe5a28cb37c5bc9e6
(b) Gregoire, J. M.; Dale, D.; Kazimirov, A.; DiSalvo, F. J.; van Dover, R. B. High energy x-ray diffraction/x-ray fluorescence spectroscopy for high-throughput analysis of composition spread thin films Rev. Sci. Instrum. 2009, 80 (12) 123905– 123905 DOI: 10.1063/1.3274179
Google Scholar
2b
High energy x-ray diffraction/x-ray fluorescence spectroscopy for high-throughput analysis of composition spread thin films
Gregoire, John M.; Dale, Darren; Kazimirov, Alexander; Di Salvo, Francis J.; van Dover, R. Bruce
Review of Scientific Instruments (2009), 80 (12), 123905/1-123905/6CODEN: RSINAK; ISSN:0034-6748. (American Institute of Physics)
A technique for simultaneous acquisition of diffraction images and fluorescence spectra on a continuous compn. spread thin film using a 60 keV x-ray source is presented. Subsequent noninteractive data processing provides maps of the diffraction profiles, thin film fiber texture, and compn. Even for highly textured films, our diffraction technique provides detection of diffraction from each family of Bragg reflections, which affords direct comparison of the measured profiles with powder patterns of known phases. These techniques are important for high throughput combinatorial studies as they provide structure and compn. maps which may be correlated with performance trends within an inorg. library. (c) 2009 American Institute of Physics.
>> More from SciFinder ^®
https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADC%252BC3cXltFym&md5=78948653251ba38b1b2facd1a396e9b5
3
Barr, G.; Dong, W.; Gilmore, C. J. PolySNAP: a computer program for analysing high-throughput powder diffraction data J. Appl. Crystallogr. 2004, 37 (4) 658– 664 DOI: 10.1107/S0021889804011173
Google Scholar
3
PolySNAP: a computer program for analyzing high-throughput powder diffraction data
Barr, Gordon; Dong, Wei; Gilmore, Christopher J.
Journal of Applied Crystallography (2004), 37 (4), 658-664CODEN: JACGAR; ISSN:0021-8898. (Blackwell Publishing Ltd.)
In high-throughput crystallog. expts., it is possible to measure over 1000 powder diffraction patterns on related compds., often polymorphs or salts, in less than one week. The anal. of these patterns poses a difficult statistical problem. A computer program is presented that can analyze such data, automatically sort the patterns into related clusters or classes, characterize each cluster and identify any unusual samples contg., for example, unknown or unexpected polymorphs. Mixts. may be analyzed quant. if a database of pure phases is available. A key component of the method is a set of visualization tools based on dendrograms and pie charts, as well as principal-component anal. and metric multidimensional scaling as a source of three-dimensional score plots. The procedures were incorporated into the computer program PolySNAP, which is available com. from Bruker-AXS.
>> More from SciFinder ^®
https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADC%252BD2cXlvVKmtrs%253D&md5=d85fe634ce16c26e0515b1908f3cafc9
4
Hattrick-Simpers, J. R.; Gregoire, J. M.; Kusne, A. G. Perspective: Composition–structure–property mapping in high-throughput experiments: Turning data into knowledge APL Mater. 2016, 4 (5) 053211 DOI: 10.1063/1.4950995
Google Scholar
There is no corresponding record for this reference.
5
(a) Bunn, J. K.; Han, S.; Zhang, Y.; Tong, Y.; Hu, J.; Hattrick-Simpers, J. R. Generalized machine learning technique for automatic phase attribution in time variant high-throughput experimental studies J. Mater. Res. 2015, 30 (7) 879– 889 DOI: 10.1557/jmr.2015.80
Google Scholar
5a
Generalized machine learning technique for automatic phase attribution in time variant high-throughput experimental studies
Bunn, Jonathan Kenneth; Han, Shizhong; Zhang, Yan; Tong, Yan; Hu, Jianjun; Hattrick-Simpers, Jason R.
Journal of Materials Research (2015), 30 (7), 879-889CODEN: JMREEE; ISSN:2044-5326. (Cambridge University Press)
Phase identification is an arduous task during high-throughput processing expts., which can be exacerbated by the need to reconcile results from multiple measurement techniques to form a holistic understanding of phase dynamics. Here, we demonstrate AutoPhase, a machine learning algorithm, which can identify the presence of the different phases in spectral and diffraction data. The algorithm uses training data to det. the characteristic features of each phase present and then uses these features to evaluate new spectral and diffraction data. AutoPhase was used to identify oxide phase growth during a high-throughput oxidn. study of NiAl bond coats that used x-ray diffraction, Raman, and fluorescence spectroscopic techniques. The algorithm had a min. overall accuracy of 88.9% for unprocessed data and 98.4% for postprocessed data. Although the features selected by AutoPhase for phase attribution were distinct from those of topical experts, these results show that AutoPhase can substantially increase the throughput high-throughput data anal.
>> More from SciFinder ^®
https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADC%252BC2MXotVSmu70%253D&md5=acff9ce14d534920e6ee32c7e1f9915c
(b) Bunn, J. K.; Hu, J.; Hattrick-Simpers, J. R. Semi-Supervised Approach to Phase Identification from Combinatorial Sample Diffraction Patterns JOM 2016, 68 (8) 2116– 2125 DOI: 10.1007/s11837-016-2033-8
Google Scholar
There is no corresponding record for this reference.
6
Long, C. J.; Hattrick-Simpers, J.; Murakami, M.; Srivastava, R. C.; Takeuchi, I.; Karen, V. L.; Li, X. Rapid structural mapping of ternary metallic alloy systems using the combinatorial approach and cluster analysis Rev. Sci. Instrum. 2007, 78 (7) 072217 DOI: 10.1063/1.2755487
Google Scholar
6
Rapid structural mapping of ternary metallic alloy systems using the combinatorial approach and cluster analysis
Long, C. J.; Hattrick-Simpers, J.; Murakami, M.; Srivastava, R. C.; Takeuchi, I.; Karen, V. L.; Li, X.
Review of Scientific Instruments (2007), 78 (7), 072217/1-072217/6CODEN: RSINAK; ISSN:0034-6748. (American Institute of Physics)
A procedure for the rapid identification of structural phases in thin film compn. spread expts. which map large fractions of compositional phase diagrams of ternary alloy systems was developed. An inhouse scanning x-ray microdiffractometer is used to obtain x-ray spectra from 273 different compns. on a single compn. spread library. A cluster anal. software is then used to sort the spectra into groups to discover rapidly the distribution of phases in the ternary diagram. The most representative pattern of each group is then compared to a database of known structures to identify known phases. With this method, the arduous anal. and classification of hundreds of spectra is decreased to a much shorter anal. of only a few spectra.
>> More from SciFinder ^®
https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADC%252BD2sXpt1Kmsbs%253D&md5=fb6550e039f8df5c2f03594f2532e23f
7
Lebras, R.; Damoulas, T.; Gregoire, J. M.; Sabharwal, A.; Gomes, C. P.; van Dover, R. B. Constraint Reasoning and Kernel Clustering for Pattern Decomposition With Scaling Proc. 17th Intl. Conf. on Principles and Practice of Constraint Programming 2011, 6876, 508– 522 DOI: 10.1007/978-3-642-23786-7_39
Google Scholar
There is no corresponding record for this reference.
8
Kusne, A. G.; Gao, T.; Mehta, A.; Ke, L.; Nguyen, M. C.; Ho, K.-M.; Antropov, V.; Wang, C.-Z.; Kramer, M. J.; Long, C.; Takeuchi, I. On-the-fly machine-learning for high-throughput experiments: search for rare-earth-free permanent magnets Sci. Rep. 2014, 4, 6367 DOI: 10.1038/srep06367
Google Scholar
8
On-the-fly machine-learning for high-throughput experiments: search for rare-earth-free permanent magnets
Kusne, Aaron Gilad; Gao, Tieren; Mehta, Apurva; Ke, Liqin; Nguyen, Manh Cuong; Ho, Kai-Ming; Antropov, Vladimir; Wang, Cai-Zhuang; Kramer, Matthew J.; Long, Christian; Takeuchi, Ichiro
Scientific Reports (2014), 4 (), 6367/1-6367/7CODEN: SRCEC3; ISSN:2045-2322. (Nature Publishing Group)
Advanced materials characterization techniques with ever-growing data acquisition speed and storage capabilities represent a challenge in modern materials science, and new procedures to quickly assess and analyze the data are needed. Machine learning approaches are effective in reducing the complexity of data and rapidly homing in on the underlying trend in multi-dimensional data. Here, we show that by employing an algorithm called the mean shift theory to a large amt. of diffraction data in high-throughput experimentation, one can streamline the process of delineating the structural evolution across compositional variations mapped on combinatorial libraries with minimal computational cost. Data collected at a synchrotron beamline are analyzed on the fly, and by integrating exptl. data with the inorg. crystal structure database (ICSD), we can substantially enhance the accuracy in classifying the structural phases across ternary phase spaces. We have used this approach to identify a novel magnetic phase with enhanced magnetic anisotropy which is a candidate for rare-earth free permanent magnet.
>> More from SciFinder ^®
https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADC%252BC2MXktlSmsrY%253D&md5=471edc300cc8d21571dd1dc40de9f1a3
9
Ermon, S.; Le Bras, R.; Gomes, C. P.; Selman, B.; van Dover, R. B. SMT-Aided Combinatorial Materials Discovery Proceedings of the 15th International Conference on Theory and Applications of Satisfiability Testing (SAT) 2012, 7317, 172– 185 DOI: 10.1007/978-3-642-31612-8_14
Google Scholar
There is no corresponding record for this reference.
10
Long, C. J.; Bunker, D.; Li, X.; Karen, V. L.; Takeuchi, I. Rapid identification of structural phases in combinatorial thin-film libraries using x-ray diffraction and non-negative matrix factorization Rev. Sci. Instrum. 2009, 80 (10) 103902– 103902 DOI: 10.1063/1.3216809
Google Scholar
10
Rapid identification of structural phases in combinatorial thin-film libraries using x-ray diffraction and non-negative matrix factorization
Long, C. J.; Bunker, D.; Li, X.; Karen, V. L.; Takeuchi, I.
Review of Scientific Instruments (2009), 80 (10), 103902/1-103902/6CODEN: RSINAK; ISSN:0034-6748. (American Institute of Physics)
The authors apply a technique called non-neg. matrix factorization (NMF) to the problem of analyzing hundreds of x-ray microdiffraction (μXRD) patterns from a combinatorial materials library. An inhouse scanning x-ray microdiffractometer is used to obtain μXRD patterns from 273 different compns. on a single compn. spread library. NMF is then used to identify the unique μXRD patterns present in the system and quantify the contribution of each of these basis patterns to each exptl. diffraction pattern. As a baseline, the results of NMF are compared to the results obtained using principal component anal. The basis patterns found using NMF are then compared to ref. patterns from a database of known structural patterns in order to identify known structures. As an example system, the authors explore a region of the Fe-Ga-Pd ternary system. The use of NMF in this case reduces the arduous task of analyzing hundreds of μXRD patterns to the much smaller task of identifying only nine μXRD patterns. (c) 2009 American Institute of Physics.
>> More from SciFinder ^®
https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADC%252BD1MXht1ajtr7J&md5=e7a3fcd7caef1689db7f14e53c766c84
11
Ermon, S; Le Bras, R.; Suram, S. K.; Gregoire, J. M.; Gomes, C. P.; Selman, B.; Van Dover, R. B. Pattern Decomposition with Complex Combinatorial Constraints: Application to Materials Discovery AAAI'15 Proceedings of the Twenty-Ninth Conference on Artificial Intelligence (AAAI-15) 2015, 636– 643
Google Scholar
There is no corresponding record for this reference.
12
Massoudifar, P.; Rangarajan, A.; Zare, A.; Gader, P. An integrated graph cuts segmentation and piece-wise convex unmixing approach for hyperspectral imaging. Presented at the 6th Workshop on Hyperstectral Image and Signal Processing (WHISPERS): Evolution in Remote Sensing, 24–27 June 2014, Lausanne, Switzerland; http://www.ieee-whispers.com/images/techprograms/whispers_2014_technical_program.pdf.
Google Scholar
There is no corresponding record for this reference.
13
Kusne, A. G.; Keller, D.; Anderson, A.; Zaban, A.; Takeuchi, I. High-throughput determination of structural phase diagram and constituent phases using GRENDEL Nanotechnology 2015, 26 (44) 444002 DOI: 10.1088/0957-4484/26/44/444002
Google Scholar
There is no corresponding record for this reference.
14
Baumes, L. A.; Moliner, M.; Nicoloyannis, N.; Corma, A. A reliable methodology for high throughput identification of a mixture of crystallographic phases from powder X-ray diffraction data CrystEngComm 2008, 10 (10) 1321– 1321 DOI: 10.1039/b812395k
Google Scholar
There is no corresponding record for this reference.
15
Smaragdis, P. Non-negative Matrix Factor Deconvolution; Extraction of Multiple Sound Sources from Monophonic Inputs International Conference on Independent Component Analysis and Signal Separation 2004, 3195, 494– 499 DOI: 10.1007/978-3-540-30110-3_63
Google Scholar
There is no corresponding record for this reference.
16
Mitrovic, S.; Cornell, E. W.; Marcin, M. R.; Jones, R. J.; Newhouse, P. F.; Suram, S. K.; Jin, J.; Gregoire, J. M. High-throughput on-the-fly scanning ultraviolet-visible dual-sphere spectrometer Rev. Sci. Instrum. 2015, 86 (1) 013904 DOI: 10.1063/1.4905365
Google Scholar
16
High-throughput on-the-fly scanning ultraviolet-visible dual-sphere spectrometer
Mitrovic, Slobodan; Cornell, Earl W.; Marcin, Martin R.; Jones, Ryan J. R.; Newhouse, Paul F.; Suram, Santosh K.; Jin, Jian; Gregoire, John M.
Review of Scientific Instruments (2015), 86 (1), 013904/1-013904/5CODEN: RSINAK; ISSN:0034-6748. (American Institute of Physics)
The authors have developed an on-the-fly scanning spectrometer operating in the UV-visible and near-IR that can simultaneously perform transmission and total reflectance measurements at the rate better than 1 sample/s. High throughput optical characterization is important for screening functional materials for a variety of new applications. The authors demonstrate the utility of the instrument for screening new light absorber materials by measuring the spectral absorbance, which is subsequently used for deriving band gap information through Tauc plot anal. (c) 2015 American Institute of Physics.
>> More from SciFinder ^®
https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADC%252BC2MXovFSksA%253D%253D&md5=fb329f3f16e133a5f0155c395df30c15
17
(a) Suram, S. K.; Newhouse, P. F.; Zhou, L.; Van Campen, D. G.; Mehta, A.; Gregoire, J. M. High Throughput Light Absorber Discovery, Part 2: Establishing Structure-Band Gap Energy Relationships ACS Comb. Sci. 2016, 18, 682 DOI: 10.1021/acscombsci.6b00054
Google Scholar
17a
High Throughput Light Absorber Discovery, Part 2: Establishing Structure-Band Gap Energy Relationships
Suram, Santosh K.; Newhouse, Paul F.; Zhou, Lan; Van Campen, Douglas G.; Mehta, Apurva; Gregoire, John M.
ACS Combinatorial Science (2016), 18 (11), 682-688CODEN: ACSCCC; ISSN:2156-8944. (American Chemical Society)
Combinatorial materials science strategies have accelerated materials development in a variety of fields, and these strategies were extended to enable structure-property mapping for light absorber materials, particularly in high order compn. spaces. High throughput optical spectroscopy and synchrotron x-ray diffraction are combined to identify the optical properties of Bi-V-Fe oxides, leading to the identification of Bi4V1.5Fe0.5O10.5 as a light absorber with direct band gap near 2.7 eV. The strategic combination of exptl. and data anal. techniques includes automated Tauc anal. to est. band gap energies from the high throughput spectroscopy data, providing an automated platform for identifying new optical materials.
>> More from SciFinder ^®
https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADC%252BC28XhsFKmsLnL&md5=03fc49c785f336b9a026ea22092d2efa
(b) Suram, S. K.; Newhouse, P. F.; Gregoire, J. M. High Throughput Light Absorber Discovery, Part 1: An Algorithm for Automated Tauc Analysis ACS Comb. Sci. 2016, 18, 673 DOI: 10.1021/acscombsci.6b00053
Google Scholar
17b
High Throughput Light Absorber Discovery, Part 1: An Algorithm for Automated Tauc Analysis
Suram, Santosh K.; Newhouse, Paul F.; Gregoire, John M.
ACS Combinatorial Science (2016), 18 (11), 673-681CODEN: ACSCCC; ISSN:2156-8944. (American Chemical Society)
High-throughput experimentation provides efficient mapping of compn.-property relations, and its implementation for the discovery of optical materials enables advancements in solar energy and other technols. In a high throughput pipeline, automated data processing algorithms are often required to match exptl. throughput, and the authors present an automated Tauc anal. algorithm for estg. band gap energies from optical spectroscopy data. The algorithm mimics the judgment of an expert scientist, which is demonstrated through its application to a variety of high throughput spectroscopy data, including the identification of indirect or direct band gaps in Fe2O3, Cu2V2O7, and BiVO4. The applicability of the algorithm to est. a range of band gap energies for various materials is demonstrated by a comparison of direct-allowed band gaps estd. by expert scientists and by automated algorithm for 60 optical spectra.
>> More from SciFinder ^®
https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADC%252BC28XhsFKmsL%252FJ&md5=497ad87216da8e3fc75cf827e5f1a329
18
Hu, S.; Xiang, C.; Haussener, S.; Berger, A. D.; Lewis, N. S. An analysis of the optimal band gaps of light absorbers in integrated tandem photoelectrochemical water-splitting systems Energy Environ. Sci. 2013, 6 (10) 2984– 2984 DOI: 10.1039/c3ee40453f
Google Scholar
18
An analysis of the optimal band gaps of light absorbers in integrated tandem photoelectrochemical water-splitting systems
Hu, Shu; Xiang, Chengxiang; Haussener, Sophia; Berger, Alan D.; Lewis, Nathan S.
Energy & Environmental Science (2013), 6 (10), 2984-2993CODEN: EESNBY; ISSN:1754-5706. (Royal Society of Chemistry)
The solar-to-hydrogen (STH) efficiency limits, along with the max. efficiency values and the corresponding optimal band gap combinations, have been evaluated for various combinations of light absorbers arranged in a tandem configuration in realistic, operational water-splitting prototypes. To perform the evaluation, a current-voltage model was employed, with the light absorbers, electrocatalysts, soln. electrolyte, and membranes coupled in series, and with the directions of optical absorption, carrier transport, electron transfer and ionic transport in parallel. The c.d. vs. voltage characteristics of the light absorbers were detd. by detailed-balance calcns. that accounted for the Shockley-Queisser limit on the photovoltage of each absorber. The max. STH efficiency for an integrated photoelectrochem. system was found to be ∼31.1% at 1 Sun (=1 kW m-2, air mass 1.5), fundamentally limited by a matching photocurrent d. of 25.3 mA cm-2 produced by the light absorbers. Choices of electrocatalysts, as well as the fill factors of the light absorbers and the Ohmic resistance of the soln. electrolyte also play key roles in detg. the max. STH efficiency and the corresponding optimal tandem band gap combination. Pairing 1.6-1.8 eV band gap semiconductors with Si in a tandem structure produces promising light absorbers for water splitting, with theor. STH efficiency limits of >25%.
>> More from SciFinder ^®
https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADC%252BC3sXhsV2it7fE&md5=cd660988f30bdee6369817f75f6e5de3
19
(a) Zhou, L.; Yan, Q.; Shinde, A.; Guevarra, D.; Newhouse, P. F.; Becerra-Stasiewicz, N.; Chatman, S. M.; Haber, J. A.; Neaton, J. B.; Gregoire, J. M. High Throughput Discovery of Solar Fuels Photoanodes in the CuO–V2O5 System Adv. Energy Mater. 2015, 5, 1500968 DOI: 10.1002/aenm.201500968
Google Scholar
There is no corresponding record for this reference.
(b) Thienhaus, S.; Naujoks, D.; Pfetzing-Micklich, J.; Konig, D.; Ludwig, A. Rapid identification of areas of interest in thin film materials libraries by combining electrical, optical, X-ray diffraction, and mechanical high-throughput measurements: a case study for the system Ni-Al ACS Comb. Sci. 2014, 16 (12) 686– 94 DOI: 10.1021/co5000757
Google Scholar
19b
Rapid Identification of Areas of Interest in Thin Film Materials Libraries by Combining Electrical, Optical, X-ray Diffraction, and Mechanical High-Throughput Measurements: A Case Study for the System Ni-Al
Thienhaus, S.; Naujoks, D.; Pfetzing-Micklich, J.; Koenig, D.; Ludwig, A.
ACS Combinatorial Science (2014), 16 (12), 686-694CODEN: ACSCCC; ISSN:2156-8944. (American Chemical Society)
The efficient identification of compositional areas of interest in thin film materials systems fabricated by combinatorial deposition methods is essential in combinatorial materials science. We use a combination of compositional screening by EDX together with high-throughput measurements of elec. and optical properties of thin film libraries to det. efficiently the areas of interest in a materials system. Areas of interest are compns. which show distinctive properties. The crystallinity of the thus detd. areas is identified by X-ray diffraction. Addnl., by using automated nanoindentation across the materials library, mech. data of the thin films can be obtained which complements the identification of areas of interest. The feasibility of this approach is demonstrated by using a Ni-Al thin film library as a ref. system. The obtained results promise that this approach can be used for the case of ternary and higher order systems.
>> More from SciFinder ^®
https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADC%252BC2cXhvVGls7%252FM&md5=a6006b56ec7d83f96073e2bde9944e6e
(c) Subramaniyan, A.; Perkins, J. D.; O’Hayre, R. P.; Ginley, D. S.; Lany, S.; Zakutayev, A. Non-equilibrium synthesis, structure, and opto-electronic properties of Cu2–2x Zn x O alloys J. Mater. Sci. 2015, 50 (3) 1350– 1357 DOI: 10.1007/s10853-014-8695-0
Google Scholar
19c
Non-equilibrium synthesis, structure, and opto-electronic properties of Cu2-2xZnxO alloys
Subramaniyan, Archana; Perkins, John D.; O'Hayre, Ryan P.; Ginley, David S.; Lany, Stephan; Zakutayev, Andriy
Journal of Materials Science (2015), 50 (3), 1350-1357CODEN: JMTSAS; ISSN:0022-2461. (Springer)
Alloying in traditional semiconductors is a well-established method to tune the electronic structure and the materials properties, but this technique is less common for oxides. Here, we present results on the non-equil. alloying of the prototypical semiconductor Cu2O with ZnO synthesized via high-throughput RF magnetron sputtering. It is demonstrated that the Zn solid soly. in Cu2O structure can be increased up to 17 at.% in the substrate temp. range 210-270 °C; this upper bound est. of the soly. limit is much higher than that at equil. (sub at. percent range). The preferential orientation in the film changes from (200) to (111) with increasing Zn concn., but the lattice parameter and the grain size (80-180 nm) remains const. Incorporation of Zn into Cu2O increases the optical absorption fourfold at the band gap (2.1 eV) and reduces the p-type elec. cond. by an order of magnitude. The ability to synthesize phase pure Cu2-2xZnxO alloys with Zn solid soly. in excess of the thermodn. limit with tunable structural and optoelectronic properties demonstrates the potential of non-equil. growth to overcome the soly. limits in oxide thin films and the promise of such alloys for optoelectronic applications.
>> More from SciFinder ^®
https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADC%252BC2cXhvFGrsLfL&md5=06d0cfea7de99372a3591d0a05697a9b
20
Xue, Y.; Bai, J.; Le Bras, R.; Rappazzo, B.; Bernstein, R.; Bjorck, J.; Longpre, L.; Suram, S. K.; van Dover, R. B.; Gregoire, J.; Gomes, C. P. Phase-Mapper: An AI Platform to Accelerate High Throughput Materials Discovery. 2016, arXiv:1610.00689. arXiv.org e-Print archive. http://adsabs.harvard.edu/abs/2016arXiv161000689X.
Google Scholar
There is no corresponding record for this reference.
21
Mannsfeld, S. C.; Tang, M. L.; Bao, Z. Thin film structure of triisopropylsilylethynyl-functionalized pentacene and tetraceno[2,3-b]thiophene from grazing incidence X-ray diffraction Adv. Mater. 2011, 23 (1) 127– 31 DOI: 10.1002/adma.201003135
Google Scholar
21
Thin Film Structure of Triisopropylsilylethynyl-Functionalized Pentacene and Tetraceno[2,3-b]thiophene from Grazing Incidence X-Ray Diffraction
Mannsfeld, Stefan C. B.; Tang, Ming Lee; Bao, Zhenan
Advanced Materials (Weinheim, Germany) (2011), 23 (1), 127-131CODEN: ADVMEW; ISSN:0935-9648. (Wiley-VCH Verlag GmbH & Co. KGaA)
The thin films of TIPS-thiotetracene and TIPS-pentacene grown on SiO2 exhibit a mol. packing that is nearly identical to that in the resp. bulk crystal structure and there is no indication of polymorphism or thin-film phases are obsd. for both arom. core materials of pentacene and thiotetracene.
>> More from SciFinder ^®
https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADC%252BC3cXhs1WntbrK&md5=7e6866afd703e3679d316828847bb5d3
22
Le Bras, R.; Bernstein, R.; Gregoire, J. M.; Suram, S. K.; Gomes, C. P.; Selman, B.; van Dover, R. B., A Computational Challenge Problem in Materials Discovery: Synthetic Problem Generator and Real-World Datasets. Presented at the Twenty-Eighth Conference on Artificial Intelligence (AAAI-14), Québec City, Canada, 2014.
Google Scholar
There is no corresponding record for this reference.
23
Kubelka, P. New Contributions to the Optics of Intensely Light-Scattering Materials. Part I J. Opt. Soc. Am. 1948, 38 (5) 448– 457 DOI: 10.1364/JOSA.38.000448
Google Scholar
23
New contributions to the optics of intensely light-scattering materials
KUBELKA P
Journal of the Optical Society of America (1948), 38 (5), 448-57 ISSN:0030-3941.
There is no expanded citation for this reference.
>> More from SciFinder ^®
https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A280%3ADyaH1c%252FmtVKgug%253D%253D&md5=1bfe837bfc7f2fb868e165ee748212d1
24
Yan, Q.; Li, G.; Newhouse, P. F.; Yu, J.; Persson, K. A.; Gregoire, J. M.; Neaton, J. B. Mn2V2O7: An Earth Abundant Light Absorber for Solar Water Splitting Adv. En Mater. 2015, 5 (8) 1401840 DOI: 10.1002/aenm.201401840
Google Scholar
24
Mn2V2O7: An Earth Abundant Light Absorber for Solar Water Splitting
Yan, Qimin; Li, Guo; Newhouse, Paul F.; Yu, Jie; Persson, Kristin A.; Gregoire, John M.; Neaton, Jeffrey B.
Advanced Energy Materials (2015), 5 (8), 1401840/1-1401840/6CODEN: ADEMBC; ISSN:1614-6840. (Wiley-Blackwell)
This paper describes the Mn2V2O7, earth abundant light absorber for solar water splitting.
>> More from SciFinder ^®
https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADC%252BC2MXmslyls7g%253D&md5=01e00d166f34150853937d8fc4a82b72
25
Zhang, W.; Shi, L.; Tang, K.; Liu, Z. Synthesis, surface group modification of 3D MnV2O6 nanostructures and adsorption effect on Rhodamine B Mater. Res. Bull. 2012, 47 (7) 1725– 1733 DOI: 10.1016/j.materresbull.2012.03.038
Google Scholar
25
Synthesis, surface group modification of 3D MnV2O6 nanostructures and adsorption effect on Rhodamine B
Zhang, Wanqun; Shi, Lei; Tang, Kaibin; Liu, Zhongping
Materials Research Bulletin (2012), 47 (7), 1725-1733CODEN: MRBUAC; ISSN:0025-5408. (Elsevier Ltd.)
Highly uniform 3D MnV2O6 nanostructures modified by oxygen functional groups (COO) were successfully prepd. in large quantities by an approach involving prepn. of vanadyl ethylene glycolate as the precursor. The growth and self-assembly of MnV2O6 nanobelts and nanorods could be readily tuned by additive species and quantities, which brought different morphologies and sizes to the final products. With a focus on the regulation of structure, the formation process of 3D architectures of MnV2O6 by self-assembly of nanobelts was followed by field emission SEM (FE-SEM) and X-ray diffraction (XRD). The consecutive processes of vanadyl ethylene glycolate and benzoyl peroxide assisted formation of layered structure Mn0.5V2O5·nH2O, growth of aligned MnV2O6 nanobelts, and oriented assembly were proposed for the growth mechanism. The band gap vs. different morphol. was also studied. Optical characterization of these MnV2O6 with different morphologies showed direct bandgap energies at 1.8-1.95 eV. The adsorption properties of 3D MnV2O6 nanostructures synthesized under different conditions were investigated through the removal test of Rhodamine B in aq. water, and the 3D nanostructures synthesized with 30 g L-1 benzoyl peroxide showed good adsorption capability of Rhodamine B.
>> More from SciFinder ^®
https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADC%252BC38Xls1WmsLk%253D&md5=fbf26298dccf393956d87f55cefe1cbe
26
This ICDD entry does not have relative scattering information. The average value from Mn₂V₂O₇ polytypes (01-073-6361, 01-089-0483) is used as a proxy.
There is no corresponding record for this reference.
27
This ICDD entry does not have relative scattering information. The average value from V₂O₅ (00-041-1426) and Nb₂O₅ (04-007-2424) is used as a proxy.
There is no corresponding record for this reference.

Cited By

Click to copy section linkSection link copied!

Explore this article's citation statements on scite.ai

This article is cited by 64 publications.

Tyler B. Martin, Debra J. Audus. Emerging Trends in Machine Learning: A Polymer Perspective. ACS Polymers Au 2023, 3 (3) , 239-258. https://doi.org/10.1021/acspolymersau.2c00053
Peter A. Beaucage, Tyler B. Martin. The Autonomous Formulation Laboratory: An Open Liquid Handling Platform for Formulation Discovery Using X-ray and Neutron Scattering. Chemistry of Materials 2023, 35 (3) , 846-852. https://doi.org/10.1021/acs.chemmater.2c03118
Meghna Srivastava, John M. Howard, Tao Gong, Mariama Rebello Sousa Dias, Marina S. Leite. Machine Learning Roadmap for Perovskite Photovoltaics. The Journal of Physical Chemistry Letters 2021, 12 (32) , 7866-7877. https://doi.org/10.1021/acs.jpclett.1c01961
Ryoji Katsube, Kei Terayama, Ryo Tamura, Yoshitaro Nose. Experimental Establishment of Phase Diagrams Guided by Uncertainty Sampling: An Application to the Deposition of Zn–Sn–P Films by Molecular Beam Epitaxy. ACS Materials Letters 2020, 2 (6) , 571-575. https://doi.org/10.1021/acsmaterialslett.0c00104
Zhaoyang Zhao, Ying Jin, Peng Shi, Yanpeng Xue, Bingbing Zhao, Yanpeng Zhang, Feifei Huang, Peng Bi, Qingrui Wang. An Improved High-Throughput Data Processing Based on Combinatorial Materials Chip Approach for Rapid Construction of Fe–Cr–Ni Composition-Phase Map. ACS Combinatorial Science 2019, 21 (12) , 833-842. https://doi.org/10.1021/acscombsci.9b00149
Joshua P. McClure, Jonathan Boltersdorf, David R. Baker, Thomas G. Farinha, Nicholas Dzuricky, Cesar E. P. Villegas, Alexandre R. Rocha, Marina S. Leite. Structure–Property-Performance Relationship of Ultrathin Pd–Au Alloy Catalyst Layers for Low-Temperature Ethanol Oxidation in Alkaline Media. ACS Applied Materials & Interfaces 2019, 11 (28) , 24919-24932. https://doi.org/10.1021/acsami.9b01389
Hui Xing, Bingbing Zhao, Yujie Wang, Xiaoyi Zhang, Yang Ren, Ningning Yan, Tieren Gao, Jindong Li, Lanting Zhang, Hong Wang. Rapid Construction of Fe–Co–Ni Composition-Phase Map by Combinatorial Materials Chip Approach. ACS Combinatorial Science 2018, 20 (3) , 127-131. https://doi.org/10.1021/acscombsci.7b00171
Santosh K. Suram, Sean W. Fackler, Lan Zhou, Alpha T. N’Diaye, Walter S. Drisdell, Junko Yano, and John M. Gregoire . Combinatorial Discovery of Lanthanum–Tantalum Oxynitride Solar Light Absorbers with Dilute Nitrogen for Solar Fuel Applications. ACS Combinatorial Science 2018, 20 (1) , 26-34. https://doi.org/10.1021/acscombsci.7b00143
Akira Miura, Tsukasa Hokimoto, Masanori Nagao, Takashi Yanase, Toshihiro Shimada, and Kiyoharu Tadanaga . Prediction of Ternary Liquidus Temperatures by Statistical Modeling of Binary and Ternary Ag–Al–Sn–Zn Systems. ACS Omega 2017, 2 (8) , 5271-5282. https://doi.org/10.1021/acsomega.7b00784
Fang Ren, Ronald Pandolfi, Douglas Van Campen, Alexander Hexemer, and Apurva Mehta . On-the-Fly Data Assessment for High-Throughput X-ray Diffraction Measurements. ACS Combinatorial Science 2017, 19 (6) , 377-385. https://doi.org/10.1021/acscombsci.7b00015
Khurram Shahzad, Andrei Ionut Mardare, Achim Walter Hassel. Accelerating materials discovery: combinatorial synthesis, high-throughput characterization, and computational advances. Science and Technology of Advanced Materials: Methods 2024, 4 (1) https://doi.org/10.1080/27660400.2023.2292486
Kentaro Kutsukake, Takefumi Kamioka, Kota Matsui, Ichiro Takeuchi, Takashi Segi, Takuo Sasaki, Seiji Fujikawa, Masamitu Takahasi. Feature extraction and spatial imaging of synchrotron radiation X-ray diffraction patterns using unsupervised machine learning. Science and Technology of Advanced Materials: Methods 2024, 4 (1) https://doi.org/10.1080/27660400.2024.2336402
A. Gilad Kusne, Austin McDannald, Brian DeCost. Learning material synthesis–process–structure–property relationship by data fusion: Bayesian co-regionalization N-dimensional piecewise function learning. Digital Discovery 2024, 3 (11) , 2211-2225. https://doi.org/10.1039/D4DD00048J
Abhijith S. Parackal, Rhys E. A. Goodall, Felix A. Faber, Rickard Armiento. Identifying crystal structures beyond known prototypes from x-ray powder diffraction spectra. Physical Review Materials 2024, 8 (10) https://doi.org/10.1103/PhysRevMaterials.8.103801
Qingmeng Li, Rongchang Xing, Linshan Li, Haodong Yao, Liyuan Wu, Lina Zhao. Synchrotron radiation data-driven artificial intelligence approaches in materials discovery. Artificial Intelligence Chemistry 2024, 2 (1) , 100045. https://doi.org/10.1016/j.aichem.2024.100045
Yimeng Min, Ming-Chiang Chang, Shufeng Kong, John M. Gregoire, R. Bruce van Dover, Michael O. Thompson, Carla P. Gomes. Physically Informed Graph-Based Deep Reasoning Net for Efficient Combinatorial Phase Mapping. 2023, 392-399. https://doi.org/10.1109/ICMLA58977.2023.00061
Maria Alexandra Theodorescu, Jacob Furst, Daniela Stan Raicu, Roselyne Tchoua. Leveraging Interpretable Features and Clustering Agreements for Assistive Materials Discovery. 2023, 71-78. https://doi.org/10.1109/TransAI60598.2023.00021
David Rolnick, Priya L. Donti, Lynn H. Kaack, Kelly Kochanski, Alexandre Lacoste, Kris Sankaran, Andrew Slavin Ross, Nikola Milojevic-Dupont, Natasha Jaques, Anna Waldman-Brown, Alexandra Sasha Luccioni, Tegan Maharaj, Evan D. Sherwin, S. Karthik Mukkavilli, Konrad P. Kording, Carla P. Gomes, Andrew Y. Ng, Demis Hassabis, John C. Platt, Felix Creutzig, Jennifer Chayes, Yoshua Bengio. Tackling Climate Change with Machine Learning. ACM Computing Surveys 2023, 55 (2) , 1-96. https://doi.org/10.1145/3485128
Keishu Utimula, Masao Yano, Hiroyuki Kimoto, Kenta Hongo, Kousuke Nakano, Ryo Maezono. Feature Space of XRD Patterns Constructed by an Autoencoder. Advanced Theory and Simulations 2023, 6 (2) https://doi.org/10.1002/adts.202200613
Akihiro Yamashita, Takahiro Nagata, Shinjiro Yagyu, Toru Asahi, Toyohiro Chikyow. Direct feature extraction from two-dimensional X-ray diffraction images of semiconductor thin films for fabrication analysis. Science and Technology of Advanced Materials: Methods 2022, 2 (1) , 23-37. https://doi.org/10.1080/27660400.2022.2029222
Dan Guevarra, Lan Zhou, Matthias H. Richter, Aniketa Shinde, Di Chen, Carla P. Gomes, John M. Gregoire. Materials structure–property factorization for identification of synergistic phase interactions in complex solar fuels photoanodes. npj Computational Materials 2022, 8 (1) https://doi.org/10.1038/s41524-022-00747-1
Lei Zhang, Zhenyu Li. Machine learning for materials classifications from images. Journal of Physics: Conference Series 2022, 2369 (1) , 012081. https://doi.org/10.1088/1742-6596/2369/1/012081
Lei Zhang, Shaofeng Shao. Image-based machine learning for materials science. Journal of Applied Physics 2022, 132 (10) https://doi.org/10.1063/5.0087381
Guo-Xu Zhang, Yajie Song, Wei Zhao, Hanwen An, Jiajun Wang. Machine learning-facilitated multiscale imaging for energy materials. Cell Reports Physical Science 2022, 3 (9) , 101008. https://doi.org/10.1016/j.xcrp.2022.101008
Philip Ball. Materials innovation from quantum to global. Nature Materials 2022, 21 (9) , 962-967. https://doi.org/10.1038/s41563-022-01350-x
Yuta Suzuki. Automated Data Analysis for Powder X-Ray Diffraction Using Machine Learning. Synchrotron Radiation News 2022, 35 (4) , 9-15. https://doi.org/10.1080/08940886.2022.2112496
Haotong Liang, Valentin Stanev, Aaron Gilad Kusne, Yuto Tsukahara, Kaito Ito, Ryota Takahashi, Mikk Lippmaa, Ichiro Takeuchi. Application of machine learning to reflection high-energy electron diffraction images for automated structural phase mapping. Physical Review Materials 2022, 6 (6) https://doi.org/10.1103/PhysRevMaterials.6.063805
Dane Morgan, Ghanshyam Pilania, Adrien Couet, Blas P. Uberuaga, Cheng Sun, Ju Li. Machine learning in nuclear materials research. Current Opinion in Solid State and Materials Science 2022, 26 (2) , 100975. https://doi.org/10.1016/j.cossms.2021.100975
Sun Young Park, Byeong-Kook Son, Jiyoung Choi, Hongkeun Jin, Kyungbook Lee. Application of machine learning to quantification of mineral composition on gas hydrate-bearing sediments, Ulleung Basin, Korea. Journal of Petroleum Science and Engineering 2022, 209 , 109840. https://doi.org/10.1016/j.petrol.2021.109840
Naman Jain, Akarsh Verma, Shigenobu Ogata, M. R. Sanjay, Suchart Siengchin. Application of Machine Learning in Determining the Mechanical Properties of Materials. 2022, 99-113. https://doi.org/10.1007/978-981-19-6278-3_5
Biao Wu, Haihui Zhang, Yuanxun Zhou, Lanting Zhang, Hong Wang. A Novel Approach for the Rapid Construction of the Composition-Phase Map Via Bayesian Strategies. SSRN Electronic Journal 2022, 78 https://doi.org/10.2139/ssrn.4022930
Lars Banko, Phillip M. Maffettone, Dennis Naujoks, Daniel Olds, Alfred Ludwig. Deep learning for visualization and novelty detection in large X-ray diffraction datasets. npj Computational Materials 2021, 7 (1) https://doi.org/10.1038/s41524-021-00575-9
Valentin Stanev, Kamal Choudhary, Aaron Gilad Kusne, Johnpierre Paglione, Ichiro Takeuchi. Artificial intelligence for search and discovery of quantum materials. Communications Materials 2021, 2 (1) https://doi.org/10.1038/s43246-021-00209-z
Phillip M. Maffettone, Aidan C. Daly, Daniel Olds. Constrained non-negative matrix factorization enabling real-time insights of in situ and high-throughput experiments. Applied Physics Reviews 2021, 8 (4) https://doi.org/10.1063/5.0052859
Rohit Batra, Le Song, Rampi Ramprasad. Emerging materials intelligence ecosystems propelled by machine learning. Nature Reviews Materials 2021, 6 (8) , 655-678. https://doi.org/10.1038/s41578-020-00255-y
Daniel B. Miracle, Mu Li, Zhaohan Zhang, Rohan Mishra, Katharine M. Flores. Emerging Capabilities for the High-Throughput Characterization of Structural Materials. Annual Review of Materials Research 2021, 51 (1) , 131-164. https://doi.org/10.1146/annurev-matsci-080619-022100
Jin-Woong Lee, Woon Bae Park, Minseuk Kim, Satendra Pal Singh, Myoungho Pyo, Kee-Sun Sohn. A data-driven XRD analysis protocol for phase identification and phase-fraction prediction of multiphase inorganic compounds. Inorganic Chemistry Frontiers 2021, 8 (10) , 2492-2504. https://doi.org/10.1039/D0QI01513J
Keishu Utimula, Genki I. Prayogo, Kousuke Nakano, Kenta Hongo, Ryo Maezono. Stochastic Estimations of the Total Number of Classes for a Clustering having Extremely Large Samples to be Included in the Clustering Engine. Advanced Theory and Simulations 2021, 4 (5) https://doi.org/10.1002/adts.202000301
Phillip M. Maffettone, Lars Banko, Peng Cui, Yury Lysogorskiy, Marc A. Little, Daniel Olds, Alfred Ludwig, Andrew I. Cooper. Crystallography companion agent for high-throughput materials discovery. Nature Computational Science 2021, 1 (4) , 290-297. https://doi.org/10.1038/s43588-021-00059-2
Changrong Lin, Yijun Wang, Fulan Zhong, Huiling Yu, Yurong Yan, Songping Wu. Carbon materials for high-performance potassium-ion energy-storage devices. Chemical Engineering Journal 2021, 407 , 126991. https://doi.org/10.1016/j.cej.2020.126991
Jin-Woong Lee, Woon Bae Park, Jin Hee Lee, Satendra Pal Singh, Kee-Sun Sohn. A deep-learning technique for phase identification in multiphase inorganic compounds using synthetic XRD powder patterns. Nature Communications 2020, 11 (1) https://doi.org/10.1038/s41467-019-13749-3
Steven B. Torrisi, Matthew R. Carbone, Brian A. Rohr, Joseph H. Montoya, Yang Ha, Junko Yano, Santosh K. Suram, Linda Hung. Random forest machine learning models for interpretable X-ray absorption near-edge structure spectrum-property relationships. npj Computational Materials 2020, 6 (1) https://doi.org/10.1038/s41524-020-00376-6
Yuta Suzuki, Hideitsu Hino, Takafumi Hawai, Kotaro Saito, Masato Kotsugi, Kanta Ono. Symmetry prediction and knowledge discovery from X-ray diffraction patterns using an interpretable machine learning approach. Scientific Reports 2020, 10 (1) https://doi.org/10.1038/s41598-020-77474-4
Nobuaki Kikkawa, Akitoshi Suzumura, Kazutaka Nishikawa, Shin Tajima, Seiji Kajita. Extraction of component bases from mixed spectra using non-negative matrix factorization with dissimilarity regularization. Chemometrics and Intelligent Laboratory Systems 2020, 206 , 104096. https://doi.org/10.1016/j.chemolab.2020.104096
BL DeCost, JR Hattrick-Simpers, Z Trautt, AG Kusne, E Campo, ML Green. Scientific AI in materials science: a path to a sustainable and scalable paradigm. Machine Learning: Science and Technology 2020, 1 (3) , 033001. https://doi.org/10.1088/2632-2153/ab9a20
Keishu Utimula, Rutchapon Hunkao, Masao Yano, Hiroyuki Kimoto, Kenta Hongo, Shogo Kawaguchi, Sujin Suwanna, Ryo Maezono. Machine‐Learning Clustering Technique Applied to Powder X‐Ray Diffraction Patterns to Distinguish Compositions of ThMn 12 ‐Type Alloys. Advanced Theory and Simulations 2020, 3 (7) https://doi.org/10.1002/adts.202000039
Mitsutaro Umehara, Helge S. Stein, Dan Guevarra, Paul F. Newhouse, David A. Boyd, John M. Gregoire. Analyzing machine learning models to accelerate generation of fundamental materials insights. npj Computational Materials 2019, 5 (1) https://doi.org/10.1038/s41524-019-0172-5
Yuta Suzuki, Hideitsu Hino, Masato Kotsugi, Kanta Ono. Automated estimation of materials parameter from X-ray absorption and electron energy-loss spectra with similarity measures. npj Computational Materials 2019, 5 (1) https://doi.org/10.1038/s41524-019-0176-1
Felipe Oviedo, Zekun Ren, Shijing Sun, Charles Settens, Zhe Liu, Noor Titan Putri Hartono, Savitha Ramasamy, Brian L. DeCost, Siyu I. P. Tian, Giuseppe Romano, Aaron Gilad Kusne, Tonio Buonassisi. Fast and interpretable classification of small X-ray diffraction datasets using data augmentation and deep neural networks. npj Computational Materials 2019, 5 (1) https://doi.org/10.1038/s41524-019-0196-x
Alfred Ludwig. Discovery of new materials using combinatorial synthesis and high-throughput characterization of thin-film materials libraries combined with computational methods. npj Computational Materials 2019, 5 (1) https://doi.org/10.1038/s41524-019-0205-0
Edwin Soedarmadji, Helge S. Stein, Santosh K. Suram, Dan Guevarra, John M. Gregoire. Tracking materials science data lineage to manage millions of materials experiments and analyses. npj Computational Materials 2019, 5 (1) https://doi.org/10.1038/s41524-019-0216-x
Arghya Bhowmik, Ivano E. Castelli, Juan Maria Garcia-Lastra, Peter Bjørn Jørgensen, Ole Winther, Tejs Vegge. A perspective on inverse design of battery interphases using multi-scale modelling, experiments and generative deep learning. Energy Storage Materials 2019, 21 , 446-456. https://doi.org/10.1016/j.ensm.2019.06.011
Rama K. Vasudevan, Kamal Choudhary, Apurva Mehta, Ryan Smith, Gilad Kusne, Francesca Tavazza, Lukas Vlcek, Maxim Ziatdinov, Sergei V. Kalinin, Jason Hattrick-Simpers. Materials science in the artificial intelligence age: high-throughput library generation, machine learning, and a pathway from correlations to the underpinning physics. MRS Communications 2019, 9 (3) , 821-838. https://doi.org/10.1557/mrc.2019.95
Carla P. Gomes, Bart Selman, John M. Gregoire. Artificial intelligence for materials discovery. MRS Bulletin 2019, 44 (7) , 538-544. https://doi.org/10.1557/mrs.2019.158
Shijing Sun, Noor T.P. Hartono, Zekun D. Ren, Felipe Oviedo, Antonio M. Buscemi, Mariya Layurova, De Xin Chen, Tofunmi Ogunfunmi, Janak Thapa, Savitha Ramasamy, Charles Settens, Brian L. DeCost, Aaron G. Kusne, Zhe Liu, Siyu I.P. Tian, Ian Marius Peters, Juan-Pablo Correa-Baena, Tonio Buonassisi. Accelerated Development of Perovskite-Inspired Materials via High-Throughput Synthesis and Machine-Learning Diagnosis. Joule 2019, 3 (6) , 1437-1451. https://doi.org/10.1016/j.joule.2019.05.014
Carla P. Gomes, Junwen Bai, Yexiang Xue, Johan Björck, Brendan Rappazzo, Sebastian Ament, Richard Bernstein, Shufeng Kong, Santosh K. Suram, R. Bruce van Dover, John M. Gregoire. CRYSTAL: a multi-agent AI system for automated mapping of materials’ crystal structures. MRS Communications 2019, 9 (2) , 600-608. https://doi.org/10.1557/mrc.2019.50
John M. Howard, Elizabeth M. Tennyson, Bernardo R.A. Neves, Marina S. Leite. Machine Learning for Perovskites' Reap-Rest-Recovery Cycle. Joule 2019, 3 (2) , 325-337. https://doi.org/10.1016/j.joule.2018.11.010
Valentin Stanev, Velimir V. Vesselinov, A. Gilad Kusne, Graham Antoszewski, Ichiro Takeuchi, Boian S. Alexandrov. Unsupervised phase mapping of X-ray diffraction data by nonnegative matrix factorization integrated with custom clustering. npj Computational Materials 2018, 4 (1) https://doi.org/10.1038/s41524-018-0099-2
Philip Ball. Learning from the big picture. Nature Materials 2018, 17 (12) , 1062-1062. https://doi.org/10.1038/s41563-018-0238-7
John M. Gregoire, David A. Boyd, Dan Guevarra, Joel A. Haber, Ryan Jones, Kevin Kan, Martin Marcin, Paul F. Newhouse, Aniketa Shinde, Edwin Soedarmadji, Santosh K. Suram, Lan Zhou. High Throughput Experimentation for the Discovery of Water Splitting Materials. 2018, 305-340. https://doi.org/10.1039/9781788010313-00305
Logan Ward, Muratahan Aykol, Ben Blaiszik, Ian Foster, Bryce Meredig, James Saal, Santosh Suram. Strategies for accelerating the adoption of materials informatics. MRS Bulletin 2018, 43 (9) , 683-689. https://doi.org/10.1557/mrs.2018.204
Shaobo Li, Zheng Xiong, Jianjun Hu. Inferring phase diagrams from X-ray data with background signals using graph segmentation. Materials Science and Technology 2018, 34 (3) , 315-326. https://doi.org/10.1080/02670836.2017.1389116
Fang Ren, Travis Williams, Jason Hattrick-Simpers, Apurva Mehta. On-the-fly segmentation approaches for x-ray diffraction datasets for metallic glasses. MRS Communications 2017, 7 (3) , 613-620. https://doi.org/10.1557/mrc.2017.76
Junwen Bai, Johan Bjorck, Yexiang Xue, Santosh K. Suram, John Gregoire, Carla Gomes. Relaxation Methods for Constrained Matrix Factorization Problems: Solving the Phase Mapping Problem in Materials Discovery. 2017, 104-112. https://doi.org/10.1007/978-3-319-59776-8_9

Download PDF

Get e-Alerts

ACS Combinatorial Science

Cite this: ACS Comb. Sci. 2017, 19, 1, 37–46

Click to copy citationCitation copied!

https://doi.org/10.1021/acscombsci.6b00153

Published November 21, 2016

Request reuse permissions

Article Views

Altmetric

Citations

Learn about these metrics

Article Views are the COUNTER-compliant sum of full text article downloads since November 2008 (both PDF and HTML) across all institutions and individuals. These metrics are regularly updated to reflect usage leading up to the last few days.

Citations are the number of other articles citing this article, calculated by Crossref and updated daily. Find more information about Crossref citation counts.

The Altmetric Attention Score is a quantitative measure of the attention that a research article has received online. Clicking on the donut icon will load a page at altmetric.com with additional details about the score and the social media presence for the given article. Find more information on the Altmetric Attention Score and how the score is calculated.

Abstract
High Resolution Image
Download MS PowerPoint Slide
Figure 1
Figure 1. White light image of the (V–Mn–Nb)O_x composition library on 100 mm Si/SiO₂ substrate is shown with elemental labels indicating the orientation of the sputter deposition sources. The grid of library locations for XRD (blue) and both XRF and optical characterization (green) are shown along with an additional plot showing the XRF-determined compositions of these points.
High Resolution Image
Download MS PowerPoint Slide
Figure 2
Figure 2. Six-phase solutions for the (V–Mn–Nb)O_x library with algorithm parameters noted in the left-hand labels. The ∼NMF solution is the AgileFD solution with the peak shifting removed by setting M = 1. The basis patterns are plotted along with the ICDD patterns listed in Table 1. The map of each phase is shown as a composition plot where the point size represents the phase fraction P_n,k and the color represents the relative lattice constant compared to the respective basis pattern, which is aligned to the best-match with the ICDD pattern.
High Resolution Image
Download MS PowerPoint Slide
Figure 3
Figure 3. AgileFD-Gibbs solution with K = 8, M = 10, and expert constraints applied to k = 0, 5, 6, 7. The map of each phase is shown as a composition plot where the point size represents the phase fraction P_n,k and the color represents the relative lattice constant, as described in Figure 2.
High Resolution Image
Download MS PowerPoint Slide
Figure 4
Figure 4. Composition map of the DA band gap energy from automated Tauc analysis is shown for 1329 composition samples.
High Resolution Image
Download MS PowerPoint Slide
Figure 5
Figure 5. DA band gap energy is plotted against the k = 4 phase fraction from Figure 3, demonstrating that when the phase fraction of MnV₂O₆ is above 0.8, the band gap value systematically varies with the relative lattice parameter.
High Resolution Image
Download MS PowerPoint Slide
Figure 6
Figure 6. (a) For the 26 samples with phase fraction of MnV₂O₆ in excess of 0.8, the series of XRD patterns (in the q-range with primary ICDD peaks) are arranged according to the V concentration. The top 10 patterns show a small amount of NbVO₅, and the bottom 16 patterns exhibit systematic shifting of the MnV₂O₆ peaks with respect to V concentration. (b) The relative lattice constant for the 26 samples is shown and the 16 samples with no NbVO₅ are indicated by a gray region. (c) The 26 samples are shown in a composition plot with end-members V′ = V_0.61Mn_0.29Nb_0.09O_x, Mn′ = V_0.43Mn_0.47Nb_0.1O_x, and Nb′ = V_0.43Mn_0.29Nb_0.28O_x. The composition gray composition region contains the same 16 samples as that in b. The data for each sample is colored according to its composition in all 3 plots.
High Resolution Image
Download MS PowerPoint Slide
Figure 7
Figure 7. (a) Band gap values from the composition region containing high phase purity MnV₂O₆ (see Figure 6) are plotted against the relative lattice parameter, demonstrating alloying-based tuning of the band gap energy. (b) Sample compositions using the same composition-color scale revealing that among these samples, the highest band gap energy and lattice parameters are found with the most Nb-rich compositions. (c) Representative Tauc plots for 3 samples with line colors matching the samples’ colors in b. The band gap energies produced by the automated Tauc algorithm are listed in the legend.
High Resolution Image
Download MS PowerPoint Slide
References
This article references 27 other publications.
1. 1
  Green, M. L.; Takeuchi, I.; Hattrick-Simpers, J. R. Applications of high throughput (combinatorial) methodologies to electronic, magnetic, optical, and energy-related materials J. Appl. Phys. 2013, 113 (23) 231101– 231101 DOI: 10.1063/1.4803530
  1
  Applications of high throughput (combinatorial) methodologies to electronic, magnetic, optical, and energy-related materials
  Green, Martin L.; Takeuchi, Ichiro; Hattrick-Simpers, Jason R.
  Journal of Applied Physics (Melville, NY, United States) (2013), 113 (23), 231101/1-231101/53CODEN: JAPIAU; ISSN:0021-8979. (American Institute of Physics)
  A review. High throughput (combinatorial) materials science methodol. is a relatively new research paradigm that offers the promise of rapid and efficient materials screening, optimization, and discovery. The paradigm started in the pharmaceutical industry but was rapidly adopted to accelerate materials research in a wide variety of areas. High throughput expts. are characterized by synthesis of a "library" sample that contains the materials variation of interest (typically compn.), and rapid and localized measurement schemes that result in massive data sets. Because the data are collected at the same time on the same "library" sample, they can be highly uniform with respect to fixed processing parameters. This article critically reviews the literature pertaining to applications of combinatorial materials science for electronic, magnetic, optical, and energy-related materials. High throughput methodologies will facilitate commercialization of novel materials for these critically important applications. Despite the overwhelming evidence presented in this paper that high throughput studies can effectively inform com. practice, in our perception, it remains an underutilized research and development tool. Part of this perception may be due to the inaccessibility of proprietary industrial research and development practices, but clearly the initial cost and availability of high throughput lab. equipment plays a role. Combinatorial materials science has traditionally been focused on materials discovery, screening, and optimization to combat the extremely high cost and long development times for new materials and their introduction into commerce. Going forward, combinatorial materials science will also be driven by other needs such as materials substitution and exptl. verification of materials properties predicted by modeling and simulation, which have recently received much attention with the advent of the Materials Genome Initiative. Thus, the challenge for combinatorial methodol. will be the effective coupling of synthesis, characterization and theory, and the ability to rapidly manage large amts. of data in a variety of formats. (c) 2013 American Institute of Physics.
  >> More from SciFinder ^®
  https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADC%252BC3sXpsFSjsbo%253D&md5=ad8a63053af7c06469a6d89f74296c49
2. 2
  (a) Gregoire, J. M.; Van Campen, D. G.; Miller, C. E.; Jones, R.; Suram, S. K.; Mehta, A. High Throughput Synchrotron X-ray Diffraction for Combinatorial Phase Mapping J. Synchrotron Radiat. 2014, 21 (6) 1262– 1268 DOI: 10.1107/S1600577514016488
  2a
  High-throughput synchrotron X-ray diffraction for combinatorial phase mapping
  Gregoire, J. M.; Van Campen, D. G.; Miller, C. E.; Jones, R. J. R.; Suram, S. K.; Mehta, A.
  Journal of Synchrotron Radiation (2014), 21 (6), 1262-1268CODEN: JSYRES; ISSN:1600-5775. (International Union of Crystallography)
  Discovery of new materials drives the deployment of new technologies. Complex technol. requirements demand precisely tailored material functionalities, and materials scientists are driven to search for these new materials in compositionally complex and often non-equil. spaces contg. three, four or more elements. The phase behavior of these high-order compn. spaces is mostly unknown and unexplored. High-throughput methods can offer strategies for efficiently searching complex and multi-dimensional material genomes for these much needed new materials and can also suggest a processing pathway for synthesizing them. However, high-throughput structural characterization is still relatively under-developed for rapid material discovery. Here, a synchrotron X-ray diffraction and fluorescence expt. for rapid measurement of both X-ray powder patterns and compns. for an array of samples in a material library is presented. The expt. is capable of measuring more than 5000 samples per day, as demonstrated by the acquisition of high-quality powder patterns in a bismuth-vanadium-iron oxide compn. library. A detailed discussion of the scattering geometry and its ability to be tailored for different material systems is provided, with specific attention given to the characterization of fiber textured thin films. The described prototype facility is capable of meeting the structural characterization needs for the first generation of high-throughput material genomic searches.
  >> More from SciFinder ^®
  https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADC%252BC2cXhvVakur7P&md5=8c3849a28e4f06cfe5a28cb37c5bc9e6
  (b) Gregoire, J. M.; Dale, D.; Kazimirov, A.; DiSalvo, F. J.; van Dover, R. B. High energy x-ray diffraction/x-ray fluorescence spectroscopy for high-throughput analysis of composition spread thin films Rev. Sci. Instrum. 2009, 80 (12) 123905– 123905 DOI: 10.1063/1.3274179
  2b
  High energy x-ray diffraction/x-ray fluorescence spectroscopy for high-throughput analysis of composition spread thin films
  Gregoire, John M.; Dale, Darren; Kazimirov, Alexander; Di Salvo, Francis J.; van Dover, R. Bruce
  Review of Scientific Instruments (2009), 80 (12), 123905/1-123905/6CODEN: RSINAK; ISSN:0034-6748. (American Institute of Physics)
  A technique for simultaneous acquisition of diffraction images and fluorescence spectra on a continuous compn. spread thin film using a 60 keV x-ray source is presented. Subsequent noninteractive data processing provides maps of the diffraction profiles, thin film fiber texture, and compn. Even for highly textured films, our diffraction technique provides detection of diffraction from each family of Bragg reflections, which affords direct comparison of the measured profiles with powder patterns of known phases. These techniques are important for high throughput combinatorial studies as they provide structure and compn. maps which may be correlated with performance trends within an inorg. library. (c) 2009 American Institute of Physics.
  >> More from SciFinder ^®
  https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADC%252BC3cXltFym&md5=78948653251ba38b1b2facd1a396e9b5
3. 3
  Barr, G.; Dong, W.; Gilmore, C. J. PolySNAP: a computer program for analysing high-throughput powder diffraction data J. Appl. Crystallogr. 2004, 37 (4) 658– 664 DOI: 10.1107/S0021889804011173
  3
  PolySNAP: a computer program for analyzing high-throughput powder diffraction data
  Barr, Gordon; Dong, Wei; Gilmore, Christopher J.
  Journal of Applied Crystallography (2004), 37 (4), 658-664CODEN: JACGAR; ISSN:0021-8898. (Blackwell Publishing Ltd.)
  In high-throughput crystallog. expts., it is possible to measure over 1000 powder diffraction patterns on related compds., often polymorphs or salts, in less than one week. The anal. of these patterns poses a difficult statistical problem. A computer program is presented that can analyze such data, automatically sort the patterns into related clusters or classes, characterize each cluster and identify any unusual samples contg., for example, unknown or unexpected polymorphs. Mixts. may be analyzed quant. if a database of pure phases is available. A key component of the method is a set of visualization tools based on dendrograms and pie charts, as well as principal-component anal. and metric multidimensional scaling as a source of three-dimensional score plots. The procedures were incorporated into the computer program PolySNAP, which is available com. from Bruker-AXS.
  >> More from SciFinder ^®
  https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADC%252BD2cXlvVKmtrs%253D&md5=d85fe634ce16c26e0515b1908f3cafc9
4. 4
  Hattrick-Simpers, J. R.; Gregoire, J. M.; Kusne, A. G. Perspective: Composition–structure–property mapping in high-throughput experiments: Turning data into knowledge APL Mater. 2016, 4 (5) 053211 DOI: 10.1063/1.4950995
  There is no corresponding record for this reference.
5. 5
  (a) Bunn, J. K.; Han, S.; Zhang, Y.; Tong, Y.; Hu, J.; Hattrick-Simpers, J. R. Generalized machine learning technique for automatic phase attribution in time variant high-throughput experimental studies J. Mater. Res. 2015, 30 (7) 879– 889 DOI: 10.1557/jmr.2015.80
  5a
  Generalized machine learning technique for automatic phase attribution in time variant high-throughput experimental studies
  Bunn, Jonathan Kenneth; Han, Shizhong; Zhang, Yan; Tong, Yan; Hu, Jianjun; Hattrick-Simpers, Jason R.
  Journal of Materials Research (2015), 30 (7), 879-889CODEN: JMREEE; ISSN:2044-5326. (Cambridge University Press)
  Phase identification is an arduous task during high-throughput processing expts., which can be exacerbated by the need to reconcile results from multiple measurement techniques to form a holistic understanding of phase dynamics. Here, we demonstrate AutoPhase, a machine learning algorithm, which can identify the presence of the different phases in spectral and diffraction data. The algorithm uses training data to det. the characteristic features of each phase present and then uses these features to evaluate new spectral and diffraction data. AutoPhase was used to identify oxide phase growth during a high-throughput oxidn. study of NiAl bond coats that used x-ray diffraction, Raman, and fluorescence spectroscopic techniques. The algorithm had a min. overall accuracy of 88.9% for unprocessed data and 98.4% for postprocessed data. Although the features selected by AutoPhase for phase attribution were distinct from those of topical experts, these results show that AutoPhase can substantially increase the throughput high-throughput data anal.
  >> More from SciFinder ^®
  https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADC%252BC2MXotVSmu70%253D&md5=acff9ce14d534920e6ee32c7e1f9915c
  (b) Bunn, J. K.; Hu, J.; Hattrick-Simpers, J. R. Semi-Supervised Approach to Phase Identification from Combinatorial Sample Diffraction Patterns JOM 2016, 68 (8) 2116– 2125 DOI: 10.1007/s11837-016-2033-8
  There is no corresponding record for this reference.
6. 6
  Long, C. J.; Hattrick-Simpers, J.; Murakami, M.; Srivastava, R. C.; Takeuchi, I.; Karen, V. L.; Li, X. Rapid structural mapping of ternary metallic alloy systems using the combinatorial approach and cluster analysis Rev. Sci. Instrum. 2007, 78 (7) 072217 DOI: 10.1063/1.2755487
  6
  Rapid structural mapping of ternary metallic alloy systems using the combinatorial approach and cluster analysis
  Long, C. J.; Hattrick-Simpers, J.; Murakami, M.; Srivastava, R. C.; Takeuchi, I.; Karen, V. L.; Li, X.
  Review of Scientific Instruments (2007), 78 (7), 072217/1-072217/6CODEN: RSINAK; ISSN:0034-6748. (American Institute of Physics)
  A procedure for the rapid identification of structural phases in thin film compn. spread expts. which map large fractions of compositional phase diagrams of ternary alloy systems was developed. An inhouse scanning x-ray microdiffractometer is used to obtain x-ray spectra from 273 different compns. on a single compn. spread library. A cluster anal. software is then used to sort the spectra into groups to discover rapidly the distribution of phases in the ternary diagram. The most representative pattern of each group is then compared to a database of known structures to identify known phases. With this method, the arduous anal. and classification of hundreds of spectra is decreased to a much shorter anal. of only a few spectra.
  >> More from SciFinder ^®
  https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADC%252BD2sXpt1Kmsbs%253D&md5=fb6550e039f8df5c2f03594f2532e23f
7. 7
  Lebras, R.; Damoulas, T.; Gregoire, J. M.; Sabharwal, A.; Gomes, C. P.; van Dover, R. B. Constraint Reasoning and Kernel Clustering for Pattern Decomposition With Scaling Proc. 17th Intl. Conf. on Principles and Practice of Constraint Programming 2011, 6876, 508– 522 DOI: 10.1007/978-3-642-23786-7_39
  There is no corresponding record for this reference.
8. 8
  Kusne, A. G.; Gao, T.; Mehta, A.; Ke, L.; Nguyen, M. C.; Ho, K.-M.; Antropov, V.; Wang, C.-Z.; Kramer, M. J.; Long, C.; Takeuchi, I. On-the-fly machine-learning for high-throughput experiments: search for rare-earth-free permanent magnets Sci. Rep. 2014, 4, 6367 DOI: 10.1038/srep06367
  8
  On-the-fly machine-learning for high-throughput experiments: search for rare-earth-free permanent magnets
  Kusne, Aaron Gilad; Gao, Tieren; Mehta, Apurva; Ke, Liqin; Nguyen, Manh Cuong; Ho, Kai-Ming; Antropov, Vladimir; Wang, Cai-Zhuang; Kramer, Matthew J.; Long, Christian; Takeuchi, Ichiro
  Scientific Reports (2014), 4 (), 6367/1-6367/7CODEN: SRCEC3; ISSN:2045-2322. (Nature Publishing Group)
  Advanced materials characterization techniques with ever-growing data acquisition speed and storage capabilities represent a challenge in modern materials science, and new procedures to quickly assess and analyze the data are needed. Machine learning approaches are effective in reducing the complexity of data and rapidly homing in on the underlying trend in multi-dimensional data. Here, we show that by employing an algorithm called the mean shift theory to a large amt. of diffraction data in high-throughput experimentation, one can streamline the process of delineating the structural evolution across compositional variations mapped on combinatorial libraries with minimal computational cost. Data collected at a synchrotron beamline are analyzed on the fly, and by integrating exptl. data with the inorg. crystal structure database (ICSD), we can substantially enhance the accuracy in classifying the structural phases across ternary phase spaces. We have used this approach to identify a novel magnetic phase with enhanced magnetic anisotropy which is a candidate for rare-earth free permanent magnet.
  >> More from SciFinder ^®
  https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADC%252BC2MXktlSmsrY%253D&md5=471edc300cc8d21571dd1dc40de9f1a3
9. 9
  Ermon, S.; Le Bras, R.; Gomes, C. P.; Selman, B.; van Dover, R. B. SMT-Aided Combinatorial Materials Discovery Proceedings of the 15th International Conference on Theory and Applications of Satisfiability Testing (SAT) 2012, 7317, 172– 185 DOI: 10.1007/978-3-642-31612-8_14
  There is no corresponding record for this reference.
10. 10
  Long, C. J.; Bunker, D.; Li, X.; Karen, V. L.; Takeuchi, I. Rapid identification of structural phases in combinatorial thin-film libraries using x-ray diffraction and non-negative matrix factorization Rev. Sci. Instrum. 2009, 80 (10) 103902– 103902 DOI: 10.1063/1.3216809
  10
  Rapid identification of structural phases in combinatorial thin-film libraries using x-ray diffraction and non-negative matrix factorization
  Long, C. J.; Bunker, D.; Li, X.; Karen, V. L.; Takeuchi, I.
  Review of Scientific Instruments (2009), 80 (10), 103902/1-103902/6CODEN: RSINAK; ISSN:0034-6748. (American Institute of Physics)
  The authors apply a technique called non-neg. matrix factorization (NMF) to the problem of analyzing hundreds of x-ray microdiffraction (μXRD) patterns from a combinatorial materials library. An inhouse scanning x-ray microdiffractometer is used to obtain μXRD patterns from 273 different compns. on a single compn. spread library. NMF is then used to identify the unique μXRD patterns present in the system and quantify the contribution of each of these basis patterns to each exptl. diffraction pattern. As a baseline, the results of NMF are compared to the results obtained using principal component anal. The basis patterns found using NMF are then compared to ref. patterns from a database of known structural patterns in order to identify known structures. As an example system, the authors explore a region of the Fe-Ga-Pd ternary system. The use of NMF in this case reduces the arduous task of analyzing hundreds of μXRD patterns to the much smaller task of identifying only nine μXRD patterns. (c) 2009 American Institute of Physics.
  >> More from SciFinder ^®
  https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADC%252BD1MXht1ajtr7J&md5=e7a3fcd7caef1689db7f14e53c766c84
11. 11
  Ermon, S; Le Bras, R.; Suram, S. K.; Gregoire, J. M.; Gomes, C. P.; Selman, B.; Van Dover, R. B. Pattern Decomposition with Complex Combinatorial Constraints: Application to Materials Discovery AAAI'15 Proceedings of the Twenty-Ninth Conference on Artificial Intelligence (AAAI-15) 2015, 636– 643
  There is no corresponding record for this reference.
12. 12
  Massoudifar, P.; Rangarajan, A.; Zare, A.; Gader, P. An integrated graph cuts segmentation and piece-wise convex unmixing approach for hyperspectral imaging. Presented at the 6th Workshop on Hyperstectral Image and Signal Processing (WHISPERS): Evolution in Remote Sensing, 24–27 June 2014, Lausanne, Switzerland; http://www.ieee-whispers.com/images/techprograms/whispers_2014_technical_program.pdf.
  There is no corresponding record for this reference.
13. 13
  Kusne, A. G.; Keller, D.; Anderson, A.; Zaban, A.; Takeuchi, I. High-throughput determination of structural phase diagram and constituent phases using GRENDEL Nanotechnology 2015, 26 (44) 444002 DOI: 10.1088/0957-4484/26/44/444002
  There is no corresponding record for this reference.
14. 14
  Baumes, L. A.; Moliner, M.; Nicoloyannis, N.; Corma, A. A reliable methodology for high throughput identification of a mixture of crystallographic phases from powder X-ray diffraction data CrystEngComm 2008, 10 (10) 1321– 1321 DOI: 10.1039/b812395k
  There is no corresponding record for this reference.
15. 15
  Smaragdis, P. Non-negative Matrix Factor Deconvolution; Extraction of Multiple Sound Sources from Monophonic Inputs International Conference on Independent Component Analysis and Signal Separation 2004, 3195, 494– 499 DOI: 10.1007/978-3-540-30110-3_63
  There is no corresponding record for this reference.
16. 16
  Mitrovic, S.; Cornell, E. W.; Marcin, M. R.; Jones, R. J.; Newhouse, P. F.; Suram, S. K.; Jin, J.; Gregoire, J. M. High-throughput on-the-fly scanning ultraviolet-visible dual-sphere spectrometer Rev. Sci. Instrum. 2015, 86 (1) 013904 DOI: 10.1063/1.4905365
  16
  High-throughput on-the-fly scanning ultraviolet-visible dual-sphere spectrometer
  Mitrovic, Slobodan; Cornell, Earl W.; Marcin, Martin R.; Jones, Ryan J. R.; Newhouse, Paul F.; Suram, Santosh K.; Jin, Jian; Gregoire, John M.
  Review of Scientific Instruments (2015), 86 (1), 013904/1-013904/5CODEN: RSINAK; ISSN:0034-6748. (American Institute of Physics)
  The authors have developed an on-the-fly scanning spectrometer operating in the UV-visible and near-IR that can simultaneously perform transmission and total reflectance measurements at the rate better than 1 sample/s. High throughput optical characterization is important for screening functional materials for a variety of new applications. The authors demonstrate the utility of the instrument for screening new light absorber materials by measuring the spectral absorbance, which is subsequently used for deriving band gap information through Tauc plot anal. (c) 2015 American Institute of Physics.
  >> More from SciFinder ^®
  https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADC%252BC2MXovFSksA%253D%253D&md5=fb329f3f16e133a5f0155c395df30c15
17. 17
  (a) Suram, S. K.; Newhouse, P. F.; Zhou, L.; Van Campen, D. G.; Mehta, A.; Gregoire, J. M. High Throughput Light Absorber Discovery, Part 2: Establishing Structure-Band Gap Energy Relationships ACS Comb. Sci. 2016, 18, 682 DOI: 10.1021/acscombsci.6b00054
  17a
  High Throughput Light Absorber Discovery, Part 2: Establishing Structure-Band Gap Energy Relationships
  Suram, Santosh K.; Newhouse, Paul F.; Zhou, Lan; Van Campen, Douglas G.; Mehta, Apurva; Gregoire, John M.
  ACS Combinatorial Science (2016), 18 (11), 682-688CODEN: ACSCCC; ISSN:2156-8944. (American Chemical Society)
  Combinatorial materials science strategies have accelerated materials development in a variety of fields, and these strategies were extended to enable structure-property mapping for light absorber materials, particularly in high order compn. spaces. High throughput optical spectroscopy and synchrotron x-ray diffraction are combined to identify the optical properties of Bi-V-Fe oxides, leading to the identification of Bi4V1.5Fe0.5O10.5 as a light absorber with direct band gap near 2.7 eV. The strategic combination of exptl. and data anal. techniques includes automated Tauc anal. to est. band gap energies from the high throughput spectroscopy data, providing an automated platform for identifying new optical materials.
  >> More from SciFinder ^®
  https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADC%252BC28XhsFKmsLnL&md5=03fc49c785f336b9a026ea22092d2efa
  (b) Suram, S. K.; Newhouse, P. F.; Gregoire, J. M. High Throughput Light Absorber Discovery, Part 1: An Algorithm for Automated Tauc Analysis ACS Comb. Sci. 2016, 18, 673 DOI: 10.1021/acscombsci.6b00053
  17b
  High Throughput Light Absorber Discovery, Part 1: An Algorithm for Automated Tauc Analysis
  Suram, Santosh K.; Newhouse, Paul F.; Gregoire, John M.
  ACS Combinatorial Science (2016), 18 (11), 673-681CODEN: ACSCCC; ISSN:2156-8944. (American Chemical Society)
  High-throughput experimentation provides efficient mapping of compn.-property relations, and its implementation for the discovery of optical materials enables advancements in solar energy and other technols. In a high throughput pipeline, automated data processing algorithms are often required to match exptl. throughput, and the authors present an automated Tauc anal. algorithm for estg. band gap energies from optical spectroscopy data. The algorithm mimics the judgment of an expert scientist, which is demonstrated through its application to a variety of high throughput spectroscopy data, including the identification of indirect or direct band gaps in Fe2O3, Cu2V2O7, and BiVO4. The applicability of the algorithm to est. a range of band gap energies for various materials is demonstrated by a comparison of direct-allowed band gaps estd. by expert scientists and by automated algorithm for 60 optical spectra.
  >> More from SciFinder ^®
  https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADC%252BC28XhsFKmsL%252FJ&md5=497ad87216da8e3fc75cf827e5f1a329
18. 18
  Hu, S.; Xiang, C.; Haussener, S.; Berger, A. D.; Lewis, N. S. An analysis of the optimal band gaps of light absorbers in integrated tandem photoelectrochemical water-splitting systems Energy Environ. Sci. 2013, 6 (10) 2984– 2984 DOI: 10.1039/c3ee40453f
  18
  An analysis of the optimal band gaps of light absorbers in integrated tandem photoelectrochemical water-splitting systems
  Hu, Shu; Xiang, Chengxiang; Haussener, Sophia; Berger, Alan D.; Lewis, Nathan S.
  Energy & Environmental Science (2013), 6 (10), 2984-2993CODEN: EESNBY; ISSN:1754-5706. (Royal Society of Chemistry)
  The solar-to-hydrogen (STH) efficiency limits, along with the max. efficiency values and the corresponding optimal band gap combinations, have been evaluated for various combinations of light absorbers arranged in a tandem configuration in realistic, operational water-splitting prototypes. To perform the evaluation, a current-voltage model was employed, with the light absorbers, electrocatalysts, soln. electrolyte, and membranes coupled in series, and with the directions of optical absorption, carrier transport, electron transfer and ionic transport in parallel. The c.d. vs. voltage characteristics of the light absorbers were detd. by detailed-balance calcns. that accounted for the Shockley-Queisser limit on the photovoltage of each absorber. The max. STH efficiency for an integrated photoelectrochem. system was found to be ∼31.1% at 1 Sun (=1 kW m-2, air mass 1.5), fundamentally limited by a matching photocurrent d. of 25.3 mA cm-2 produced by the light absorbers. Choices of electrocatalysts, as well as the fill factors of the light absorbers and the Ohmic resistance of the soln. electrolyte also play key roles in detg. the max. STH efficiency and the corresponding optimal tandem band gap combination. Pairing 1.6-1.8 eV band gap semiconductors with Si in a tandem structure produces promising light absorbers for water splitting, with theor. STH efficiency limits of >25%.
  >> More from SciFinder ^®
  https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADC%252BC3sXhsV2it7fE&md5=cd660988f30bdee6369817f75f6e5de3
19. 19
  (a) Zhou, L.; Yan, Q.; Shinde, A.; Guevarra, D.; Newhouse, P. F.; Becerra-Stasiewicz, N.; Chatman, S. M.; Haber, J. A.; Neaton, J. B.; Gregoire, J. M. High Throughput Discovery of Solar Fuels Photoanodes in the CuO–V2O5 System Adv. Energy Mater. 2015, 5, 1500968 DOI: 10.1002/aenm.201500968
  There is no corresponding record for this reference.
  (b) Thienhaus, S.; Naujoks, D.; Pfetzing-Micklich, J.; Konig, D.; Ludwig, A. Rapid identification of areas of interest in thin film materials libraries by combining electrical, optical, X-ray diffraction, and mechanical high-throughput measurements: a case study for the system Ni-Al ACS Comb. Sci. 2014, 16 (12) 686– 94 DOI: 10.1021/co5000757
  19b
  Rapid Identification of Areas of Interest in Thin Film Materials Libraries by Combining Electrical, Optical, X-ray Diffraction, and Mechanical High-Throughput Measurements: A Case Study for the System Ni-Al
  Thienhaus, S.; Naujoks, D.; Pfetzing-Micklich, J.; Koenig, D.; Ludwig, A.
  ACS Combinatorial Science (2014), 16 (12), 686-694CODEN: ACSCCC; ISSN:2156-8944. (American Chemical Society)
  The efficient identification of compositional areas of interest in thin film materials systems fabricated by combinatorial deposition methods is essential in combinatorial materials science. We use a combination of compositional screening by EDX together with high-throughput measurements of elec. and optical properties of thin film libraries to det. efficiently the areas of interest in a materials system. Areas of interest are compns. which show distinctive properties. The crystallinity of the thus detd. areas is identified by X-ray diffraction. Addnl., by using automated nanoindentation across the materials library, mech. data of the thin films can be obtained which complements the identification of areas of interest. The feasibility of this approach is demonstrated by using a Ni-Al thin film library as a ref. system. The obtained results promise that this approach can be used for the case of ternary and higher order systems.
  >> More from SciFinder ^®
  https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADC%252BC2cXhvVGls7%252FM&md5=a6006b56ec7d83f96073e2bde9944e6e
  (c) Subramaniyan, A.; Perkins, J. D.; O’Hayre, R. P.; Ginley, D. S.; Lany, S.; Zakutayev, A. Non-equilibrium synthesis, structure, and opto-electronic properties of Cu2–2x Zn x O alloys J. Mater. Sci. 2015, 50 (3) 1350– 1357 DOI: 10.1007/s10853-014-8695-0
  19c
  Non-equilibrium synthesis, structure, and opto-electronic properties of Cu2-2xZnxO alloys
  Subramaniyan, Archana; Perkins, John D.; O'Hayre, Ryan P.; Ginley, David S.; Lany, Stephan; Zakutayev, Andriy
  Journal of Materials Science (2015), 50 (3), 1350-1357CODEN: JMTSAS; ISSN:0022-2461. (Springer)
  Alloying in traditional semiconductors is a well-established method to tune the electronic structure and the materials properties, but this technique is less common for oxides. Here, we present results on the non-equil. alloying of the prototypical semiconductor Cu2O with ZnO synthesized via high-throughput RF magnetron sputtering. It is demonstrated that the Zn solid soly. in Cu2O structure can be increased up to 17 at.% in the substrate temp. range 210-270 °C; this upper bound est. of the soly. limit is much higher than that at equil. (sub at. percent range). The preferential orientation in the film changes from (200) to (111) with increasing Zn concn., but the lattice parameter and the grain size (80-180 nm) remains const. Incorporation of Zn into Cu2O increases the optical absorption fourfold at the band gap (2.1 eV) and reduces the p-type elec. cond. by an order of magnitude. The ability to synthesize phase pure Cu2-2xZnxO alloys with Zn solid soly. in excess of the thermodn. limit with tunable structural and optoelectronic properties demonstrates the potential of non-equil. growth to overcome the soly. limits in oxide thin films and the promise of such alloys for optoelectronic applications.
  >> More from SciFinder ^®
  https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADC%252BC2cXhvFGrsLfL&md5=06d0cfea7de99372a3591d0a05697a9b
20. 20
  Xue, Y.; Bai, J.; Le Bras, R.; Rappazzo, B.; Bernstein, R.; Bjorck, J.; Longpre, L.; Suram, S. K.; van Dover, R. B.; Gregoire, J.; Gomes, C. P. Phase-Mapper: An AI Platform to Accelerate High Throughput Materials Discovery. 2016, arXiv:1610.00689. arXiv.org e-Print archive. http://adsabs.harvard.edu/abs/2016arXiv161000689X.
  There is no corresponding record for this reference.
21. 21
  Mannsfeld, S. C.; Tang, M. L.; Bao, Z. Thin film structure of triisopropylsilylethynyl-functionalized pentacene and tetraceno[2,3-b]thiophene from grazing incidence X-ray diffraction Adv. Mater. 2011, 23 (1) 127– 31 DOI: 10.1002/adma.201003135
  21
  Thin Film Structure of Triisopropylsilylethynyl-Functionalized Pentacene and Tetraceno[2,3-b]thiophene from Grazing Incidence X-Ray Diffraction
  Mannsfeld, Stefan C. B.; Tang, Ming Lee; Bao, Zhenan
  Advanced Materials (Weinheim, Germany) (2011), 23 (1), 127-131CODEN: ADVMEW; ISSN:0935-9648. (Wiley-VCH Verlag GmbH & Co. KGaA)
  The thin films of TIPS-thiotetracene and TIPS-pentacene grown on SiO2 exhibit a mol. packing that is nearly identical to that in the resp. bulk crystal structure and there is no indication of polymorphism or thin-film phases are obsd. for both arom. core materials of pentacene and thiotetracene.
  >> More from SciFinder ^®
  https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADC%252BC3cXhs1WntbrK&md5=7e6866afd703e3679d316828847bb5d3
22. 22
  Le Bras, R.; Bernstein, R.; Gregoire, J. M.; Suram, S. K.; Gomes, C. P.; Selman, B.; van Dover, R. B., A Computational Challenge Problem in Materials Discovery: Synthetic Problem Generator and Real-World Datasets. Presented at the Twenty-Eighth Conference on Artificial Intelligence (AAAI-14), Québec City, Canada, 2014.
  There is no corresponding record for this reference.
23. 23
  Kubelka, P. New Contributions to the Optics of Intensely Light-Scattering Materials. Part I J. Opt. Soc. Am. 1948, 38 (5) 448– 457 DOI: 10.1364/JOSA.38.000448
  23
  New contributions to the optics of intensely light-scattering materials
  KUBELKA P
  Journal of the Optical Society of America (1948), 38 (5), 448-57 ISSN:0030-3941.
  There is no expanded citation for this reference.
  >> More from SciFinder ^®
  https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A280%3ADyaH1c%252FmtVKgug%253D%253D&md5=1bfe837bfc7f2fb868e165ee748212d1
24. 24
  Yan, Q.; Li, G.; Newhouse, P. F.; Yu, J.; Persson, K. A.; Gregoire, J. M.; Neaton, J. B. Mn2V2O7: An Earth Abundant Light Absorber for Solar Water Splitting Adv. En Mater. 2015, 5 (8) 1401840 DOI: 10.1002/aenm.201401840
  24
  Mn2V2O7: An Earth Abundant Light Absorber for Solar Water Splitting
  Yan, Qimin; Li, Guo; Newhouse, Paul F.; Yu, Jie; Persson, Kristin A.; Gregoire, John M.; Neaton, Jeffrey B.
  Advanced Energy Materials (2015), 5 (8), 1401840/1-1401840/6CODEN: ADEMBC; ISSN:1614-6840. (Wiley-Blackwell)
  This paper describes the Mn2V2O7, earth abundant light absorber for solar water splitting.
  >> More from SciFinder ^®
  https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADC%252BC2MXmslyls7g%253D&md5=01e00d166f34150853937d8fc4a82b72
25. 25
  Zhang, W.; Shi, L.; Tang, K.; Liu, Z. Synthesis, surface group modification of 3D MnV2O6 nanostructures and adsorption effect on Rhodamine B Mater. Res. Bull. 2012, 47 (7) 1725– 1733 DOI: 10.1016/j.materresbull.2012.03.038
  25
  Synthesis, surface group modification of 3D MnV2O6 nanostructures and adsorption effect on Rhodamine B
  Zhang, Wanqun; Shi, Lei; Tang, Kaibin; Liu, Zhongping
  Materials Research Bulletin (2012), 47 (7), 1725-1733CODEN: MRBUAC; ISSN:0025-5408. (Elsevier Ltd.)
  Highly uniform 3D MnV2O6 nanostructures modified by oxygen functional groups (COO) were successfully prepd. in large quantities by an approach involving prepn. of vanadyl ethylene glycolate as the precursor. The growth and self-assembly of MnV2O6 nanobelts and nanorods could be readily tuned by additive species and quantities, which brought different morphologies and sizes to the final products. With a focus on the regulation of structure, the formation process of 3D architectures of MnV2O6 by self-assembly of nanobelts was followed by field emission SEM (FE-SEM) and X-ray diffraction (XRD). The consecutive processes of vanadyl ethylene glycolate and benzoyl peroxide assisted formation of layered structure Mn0.5V2O5·nH2O, growth of aligned MnV2O6 nanobelts, and oriented assembly were proposed for the growth mechanism. The band gap vs. different morphol. was also studied. Optical characterization of these MnV2O6 with different morphologies showed direct bandgap energies at 1.8-1.95 eV. The adsorption properties of 3D MnV2O6 nanostructures synthesized under different conditions were investigated through the removal test of Rhodamine B in aq. water, and the 3D nanostructures synthesized with 30 g L-1 benzoyl peroxide showed good adsorption capability of Rhodamine B.
  >> More from SciFinder ^®
  https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADC%252BC38Xls1WmsLk%253D&md5=fbf26298dccf393956d87f55cefe1cbe
26. 26
  This ICDD entry does not have relative scattering information. The average value from Mn₂V₂O₇ polytypes (01-073-6361, 01-089-0483) is used as a proxy.
  There is no corresponding record for this reference.
27. 27
  This ICDD entry does not have relative scattering information. The average value from V₂O₅ (00-041-1426) and Nb₂O₅ (04-007-2424) is used as a proxy.
  There is no corresponding record for this reference.
Supporting Information
Supporting Information
The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acscombsci.6b00153.
- XRD and composition data for the V–Mn–Nb oxide system and additional source data (ZIP)
- co6b00153_si_001.zip (67.34 MB)
Terms & Conditions

Most electronic Supporting Information files are available without a subscription to ACS Web Editions. Such files may be downloaded by article for research use (if there is a public use license linked to the relevant article, that license may permit other uses). Permission may be obtained from ACS for other uses through requests via the RightsLink permission system: http://pubs.acs.org/page/copyright/permissions.html.

Automated Phase Mapping with AgileFD and its Application to Light Absorber Discovery in the V–Mn–Nb Oxide SystemClick to copy article linkArticle link copied!

ACS Combinatorial Science

Abstract

Introduction

Algorithm and Experiments

AgileFD Algorithm

Library Synthesis

Composition and Structure Measurements and Analysis

Optical Characterization

Figure 1

Results and Discussion

Figure 2

Figure 3

Figure 4

Figure 5

Figure 6

Figure 7

Summary

Supporting Information

Terms & Conditions

Author Information

Acknowledgment

References

Cited By

ACS Combinatorial Science

Article Views

Altmetric

Citations

Recommended Articles

Abstract

Figure 1

Figure 2

Figure 3

Figure 4

Figure 5

Figure 6

Figure 7

References

Supporting Information

Supporting Information

Terms & Conditions

Automated Phase Mapping with AgileFD and its Application to Light Absorber Discovery in the V–Mn–Nb Oxide System
Click to copy article linkArticle link copied!