Session H1: Computational Discovery and Design of Novel Materials VI

Computational Design of Nanostructured Thermoelectrics

Creating nanostructures within alloyed bulk thermoelectric materials can greatly decrease the lattice thermal conductivity of the material and thereby increase the thermoelectric efficiency of these materials. However, the rational design of thermoelectric alloys with even larger figures of merit will require a quantitative knowledge of the electronic and thermal properties and phase stability of nanostructured semiconductor materials. Here, we show how first-principles based calculations can reveal the intricate but tractable relationships between properties for optimization of thermoelectric performance. The integrated optimization includes a multipronged strategy: 1) significant reduction of the lattice thermal conductivity with multi-scale hierarchical architecturing, 2) large enhancement of Seebeck coefficients with intra-matrix electronic band convergence engineering, and 3) control of the carrier mobility with band alignment between host and second phases. These techniques can simultaneously enhance the power factor and reduce the lattice thermal conductivity, thereby leading to high efficiency thermoelectric materials.   

An Automated Ab Initio Framework for Identifying New Ferroelectrics

Ferroelectric materials have a wide-range of technological applications including non-volatile RAM and optoelectronics. In this work, we present an automated first-principles search for ferroelectrics. We integrate density functional theory, crystal structure databases, symmetry tools, workflow software, and a custom analysis toolkit to build a library of known and proposed ferroelectrics. We screen thousands of candidates using symmetry relations between nonpolar and polar structure pairs. We use two search strategies 1) polar-nonpolar pairs with the same composition and 2) polar-nonpolar structure type pairs. Results are automatically parsed, stored in a database, and accessible via a web interface showing distortion animations and plots of polarization and total energy as a function of distortion. We benchmark our results against experimental data, present new ferroelectric candidates found through our search, and discuss future work on expanding this search methodology to other material classes such as anti-ferroelectrics and multiferroics.   

Computational and Experimental Design of Functional Deep-Ultraviolet Non-Linear Optical Materials

Non-linear optical (NLO) materials are of intense interest owing to their ability to generate coherent radiation at a variety of difficult to access wavelengths. However, designing materials to access the deep ultraviolet (DUV, $\lambda <$ 200 nm) region remains a significant challenge, as such a compound must exhibit a number of characteristics including a non-centrosymmetric crystal structure, wide band gap (E$_g >$ 6.2 eV), large second harmonic generation (SHG) coefficients, and moderate birefringence. In this work, we use first-principles density functional theory calculations and introduce new structural metrics to disentangle the contributions of crystal-chemistry and NLO-active structural units to the properties of several newly synthesized DUV NLO materials. We find that the presence of stereo-active lone pair cations (such as Pb$^{2+}$) and triply bidentate cations enhance the SHG response, while polyhedral units serve to lift inversion symmetry in the crystal structures. We anticipate that the targeted design approach applied here can be harnessed for the discovery of advanced optical materials at other important regions of the electromagnetic spectrum.   

Data Mining for 3D Organic Dirac Materials

The study of Dirac materials, i.e. materials where the low-energy fermionic excitations behave as massless Dirac particles has been of ongoing interest for more than two decades. Such massless Dirac fermions are characterized by a linear dispersion relation with respect to the particle momentum. A combined study using group theory and data mining within the Organic Materials Database leads to the discovery of stable Dirac-point nodes and Dirac line-nodes within the electronic band structure in the class of 3-dimensional organic crystals. The nodes are protected by crystalline symmetry. As a result of this study, we present band structure calculations and symmetry analysis for previously synthesized organic materials. In all these materials, the Dirac nodes are well separated within the energy and located near the Fermi surface, which opens up a possibility for their direct experimental observation.   

Data-driven discovery of new Dirac semimetal materials

In recent years, a significant amount of materials property data from high-throughput computations based on density functional theory (DFT) and the application of database technologies have enabled the rise of data-driven materials discovery. In this work, we initiate the extension of the data-driven materials discovery framework to the realm of topological semimetal materials and to accelerate the discovery of novel Dirac semimetals. We implement current available and develop new workflows to data-mine the Materials Project database for novel Dirac semimetals with desirable band structures and symmetry protected topological properties. This data-driven effort relies on the successful development of several automatic data generation and analysis tools, including a workflow for the automatic identification of topological invariants and pattern recognition techniques to find specific features in a massive number of computed band structures. Utilizing this approach, we successfully identified more than 15 novel Dirac point and Dirac nodal line systems that have not been theoretically predicted or experimentally identified. This work is supported by the Materials Project Predictive Modeling Center through the U.S. Department of Energy, Office of Basic Energy Sciences, Materials Sciences and Engineering Division, under Contract No. DE-AC02-05CH11231.   

Designing nanostructured mixed Heusler systems for improved thermoelectrics

The search for new and more efficient thermoelectric materials has been largely focused on the improvement of the thermoelectric figure of merit by forming nanostructures in a host matrix. Special interest has been directed at the Heusler compounds, especially due to their favorable electrical properties and the possibility of lowering their thermal conductivity via nanostructuring. Aiming to guide future experiments, we predict the possibility of forming nanostructured systems between full and half Heusler host matrices, and other Heusler compounds using density functional theory calculations. Our prediction is based on estimating the solvus between two Heusler compounds, the matrix and the nanostructured compound, using their mixing energy and considering the pairs within a mixing energy interval that favors nanostructuring. Screening the Heusler pairs using this approach gave 25 matrix/nanostructured compound pairs that have not been previously considered as nanostructured thermoelectrics. In addition, based on the mixing energies we argue that different types of Heusler compounds, e.g. half and full Heusler, would favor formation of stable interfaces with low intermixing between the phases, and Heusler compounds of the same type would form nanocomposites or solid solutions.   

Design a giant 3D quantum spin Hall insulator with double pervoskites

We propose a new approach to find three-dimensional topological insulators (TIs) in which the spin-orbit coupling (SOC) can more effectively generate a band gap. The band gap of conventional TI such as Bi2Se3 is mainly limited by two factors, the strength of SOC and, from electronic structure perspective, the band gap when SOC is absent. While the former is an atomic property, we find that the latter can be minimized in a generic rock-salt lattice model in which band touching at the Fermi level along with band inversion takes place in the absence of SOC. Thus, giant-gap TIs or TIs comprised of lighter elements are expected. The model applies to a class of double perovskites A2BiXO6 (A $=$ Ca, Sr, Ba; X $=$ Br, I) and the band gap is predicted up to 0.55 eV, much larger than known pristine 3D TIs. Besides, the doped compounds might turn into topological superconductors at low temperature since the Bloch states near the Fermi level are unaltered by the dopant at A-site. First-principle calculations considering realistic surface indicate that the Dirac cones are stabilized if proper termination is chosen. The mechanism is general and may open a new vista for future TI-based electronic devices.   

High-Throughput study of chemical substitution in clay minerals

High-throughput (HT) DFT computations facilitates the understanding and the design of materials with novel properties. In this work, we use our HT infrastructure, AFLOW$\pi$, to compute the electronic structure and related properties for mineral in the clay family: lizardite (Mg$_{3}$(Si$_{2}$O$_{5}$)(OH)$_{4}$), talc (Mg$_{3}$(Si$_{2}$O$_{5})$_{2}$(OH)$_{2}$), kaolinite(Al$_{2}$(Si$_{2}$O$_{5}$)(OH)$_{4}$) and pyrophyllite (Al$_{2}$(Si$_{2}$O$_{5})$_{2}$(OH)$_{2}$). Using these four prototypes, we studied the effect of chemical substitutions in 48 different compositions. We computed the formation energies, optimal lattice parameters, elastic constants and the band structures using ACBN0, a pseudo hybrid Hubbard density functional, all of which is incorporated in the AFLOW$\pi$ framework. One main result shows that Ni-substituted lizardite (Ni$_{3}$(Si$_{2}$O$_{5}$)(OH)$_{4}$) is structurally stable and is a promising candidate in spintronic applications as spin filter.   

Simulated Scattering of Coherent X-rays from Dynamic GaN Crystal Surfaces

New techniques using coherent x-rays promise to reveal qualitatively new aspects of the arrangements and dynamics of atomic-scale features during materials synthesis. We present analysis of time-dependent speckle in the scattering of coherent x-rays from crystal surfaces, calculated from kinetic Monte Carlo simulations of atomic dynamics both at equilibrium in the growth environment at high temperature, and during non-equilibrium crystal growth. For equilibrium surfaces, standard single-$q$ time correlation functions reveal the rates and power-law exponents of the q dependence for adatom/vacancy diffusion and surface step dynamics. For non-equilibrium surfaces undergoing growth, we calculate two-time correlation functions, and investigate how they can elucidate mechanisms of island nucleation and growth. During layer-by-layer growth, oscillatory correlations at integer monolayers appear at $q < q_{max}$, where $q_{max}$ is the position of the island diffuse scattering peak. During 3-dimensional growth, correlations at non-integer-monolayer times appear at $q > q_{max}$. We will discuss the physical origins of these phenomena, to illustrate how coherent x-ray measurements of complex space/time correlations will be sensitive to atomic-scale mechanisms.   

Monovacancy Properties From Atomistic Simulations Based on OpenKIM

A longtime goal of scientists is to be able to calculate the properties of materials from their structures accurately and efficiently. And atomistic simulation with good interatomic potentials has its unique position in the trade-off between these two targets. Different models may be suitable for different situations and our study focus on vacancy, the simplest and the most common point defects. In order to assess how each interatomic potential model performs in vacancy-related simulations, we calculate the most important monovacancy properties for all the elements and all the simple crystal structures predicted by the interatomic potential models available on OpenKIM at this time. These results can provide useful information for selecting interatomic models. We also examine how these properties depend on each other and other elemental properties. In particular, we shall report on relationships between the vacancy formation energy, the migration energy, the surface energy and the elastic constants, and how these relationships depend on the class of interatomic potential.   

Data assimilation based on 4DVar for structural materials

Data assimilation (DA) is a computational technique to integrate numerical simulation models and observation data based on Bayesian statistics. One key issue is the implementation of DA in massive simulation models under the constraints of limited computation time and resources. We propose a new DA methodology based on the four-dimensional variation method (4DVar) for massive models that produces optimum estimates and their uncertainties within the reasonable computational limitations. The uncertainties are given as diagonal elements of the inverse of Hessian matrix, which is the covariance matrix of a Gaussian that approximates the posterior distribution in the neighborhood of the optimum. Conventional algorithms for deriving the Hessian inverse require $O(CN^2 + N^3)$ computations and $O(N^2)$ memory, where $N$ is the dimension of the model and $C$ is the number of computations needed to simulate time series. The proposed method using a second-order adjoint method allows us to directly evaluate the diagonal elements of the Hessian inverse without computing all elements. This drastically reduces the number of computations to $O(C)$ and the amount of memory to $O(N)$. We report an initial result when our method is applied to practical experimental data of structural materials.   

Advanced understanding of paper coating structure and its relationship to coating performance

Paper coatings have been utilized to improve paper performance for decades, for example, for improving brightness, opacity, gloss, stiffness, ink acceptance, printability, and smoothness. In thermal paper, coatings are employed to impart improved thermal, morphological and mechanical properties often through the incorporation of hollow spheres into the coating film. Hollow sphere pigments having well controlled size and narrow size distribution provide a unique opportunity to model and study the particle packing phenomena and its effect on coating film strength, smoothness and thermal properties. This talk introduces a multi-dimensional modeling approach in paper coating modeling, with a special interest in microscopic mechanistic model. Based on these approaches we have seen that control of binary packing and local structure lead to improved and balanced coating mechanical property and thermal conductivity. This will enable us to achieve better coating materials using guided design of single particle geometry and formulation.