Wednesday 2:00 PM

March 1, 2017

Room: 11A

Location: San Diego Convention Ctr

While the Ti-O binary is a fundamental building block to many multi-component phase diagrams of technologically important alloys, it still has many unknowns and has a huge scope for further discovery of new oxides. Ti is unique among metals in that it forms an immense number of stable and metastable oxides. The crystal structures and compositions of Ti-O compounds remain poorly understood. We have performed a first-principles based cluster expansion study of phase stability in the Ti-O binary and the Ti-Al-O ternary to establish the relative stability among the many titanium oxides. A large fraction of the oxides in the Ti-O binary are orderings over the interstitial and substitutional sites of FCC, HCP and the omega phase. In addition to confirming the HCP-based ordered TiOx phases, we unravel new vacancy ordering tendencies over the rocksalt form of TiO and predict the stability of an omega phase based titanium monoxide.

Predicting novel materials can be challenging due to the time-consuming effort necessitated by experimental searches or the resources and approximations utilized by theoretical ones. To overcome these obstacles, we have developed a framework that combines machine learning and density functional theory (DFT) in a unique way to achieve rapid materials discovery. We use experimental data to determine stable structures and DFT to identify unstable ones. This approach avoids relying on DFT to predict stability, since theoretically predicted, zero-temperature formation enthalpies of stable compounds are often distributed above the convex hull. Instead, we rely on simulations to predict instability, a less challenging task. Utilizing these classified structures and their constituent elemental attributes, we predict new stable compounds with several machine learning techniques. We have applied this methodology to the spinel phase space for the discovery of new materials and are working with experimental collaborators to synthesize some top candidates.

While hugely successful for materials with ordered crystal structures, ab initio modeling of substitutionally disordered alloys with a multitude of elements remains a challenging task. Recently, a novel small set of ordered structures (SSOS) method has been developed that allows extremely efficient ab initio modeling of random multicomponent alloys. The SSOS method can enable rapid high-throughput computational screening of a large number of multicomponent alloy chemistries on a computer, which may lead to the discovery of novel materials. In this presentation, we will give a brief introduction to the SSOS method. We will demonstrate the usefulness of the SSOS method by applying it to identify new high entropy alloy compositions that can be experimentally synthesized and to provide better understanding of the chemistry and crystal structure of the metallic fission-product precipitates (white inclusions) in irradiated uranium nuclear fuels.

The extraordinary properties (high strength and high hardness) of boron carbide originate from the collection of polymorphs that comprise fabricated samples. However, it suffers from amorphization, a pressure-induced phenomena that causes catastrophic loss in crystallinity and strength. To mitigate this deleterious phenomenon, we have adopted a combined computational and experimental approach in which we have defined new groups of equivalent polymorphs based on quantum-mechanical simulations. Raman spectra were developed using density functional perturbation theory and the trends in lattice parameters, total energy, and Raman spectra are explained in terms of the unique covalent-bonding environments of boron-carbide polymorphs. The presentation will address Raman spectra, polymorphism, cage-space in the crystal structure, influence of grain size and design issues relevant to fabrication of amorphization-resistant, superhard boron carbide.

Mechanical strength, high melting temperature, and oxidation and creep resistance make nickel based superalloys an ideal candidate for high temperature industrial applications. First-principles DFT calculations combined with statistical mechanics enable not only a prediction of the stability of known phases, but also facilitate the discovery of previously unknown ordered phases. First-principles effective Hamiltonians combined with Monte Carlo simulations yield free energies that predict phase stability at elevated temperatures. Applied to the Ni-Al binary and Ni-Al-Cr ternary, we discover a family of previously unknown ordered phases that are stable at compositions where the alloy transitions from FCC to BCC. We explore dynamical instabilities at compositions where martensitic transformations are observed, and identify convenient strain order parameters to track instabilities along the Bain path. Ordering of vacancies and ternary alloying additions are investigated with the cluster expansion approach. The resulting free energy description enables the calculation of high temperature oxidation phase diagrams.

Thermodynamic properties of hight temperature phase of NiTi and PtTi are calculated, using our recently developed P4 method. The austenite phase of NiTi and PtTi exhibit harmonic phonon instabilities, making the standard lattice dynamics approaches insufficient to calculate their free energy. In the P4 method, we propose to explore the potential energy surface by discrete sampling of local minima, surrounding the high-symmetry time-averaged structure, via a lattice gas Monte Carlo approach and by a continuous sampling via a harmonic lattice dynamic approach in the vicinity of each local minima. The simple extension of the proposed P4 method to solid solutions makes the calculation of technologically important compounds like NiTi and PtTi computationally feasible.

Harmonic and anharmonic phonon calculations based on the density functional theory are getting popular and have become usual tools for our research since computational results that can be satisfied are often obtained with respect to acceptable computational demand due to recent computational power and developments of calculation methods and those software. However when we want to run tens of thousands of phonon calculations, e.g., for constructing database of physical properties, we need intelligent control of workflow made of different types of calculations, which must be automated. We will present our researches, that were achieved with automated, on structural-phase-transition pathway search, systematic calculation of lattice thermal conductivity, and phonon database, and we will discuss how we use the data as obtained.

Aluminum-Silicon (Al-Si) alloys have been vastly modified through minute additions of Strontium, transforming the morphology from coarse and plate-like to fine and fibrous; though the mechanisms involved still elude scientists today. Understanding the liquid structure of a material is key to uncovering the mechanisms behind the changing morphology. This project is a first principal study into the liquid state of Al-Si, through Density Functional Theory, using the Vienna Ab-initio Simulation Package (VASP). A solution to the many-bodied Schrödinger equation is computed, from which structural information can be uncovered and compared with experimental results found through neutron and synchrotron diffraction. Through an understanding of the fundamental nature of Aluminum-Silicon alloys captured through experimentation as well as computational modelling, a better grasp of solidification characteristics that influence the morphology can be achieved. Such knowledge is useful for refinement of mechanical properties for use in industry.

The overall product engineering process is a complicated iterative task that requires many points of interaction between the various components of the design process. Uncertainty quantification (UQ) is one such interaction/design point that is often overlooked or under-appreciated in most computational materials design frameworks. This talk presents the idea of Envelopes of Uncertainty, novel graphical representations of expected output values calculated from systematic variations of input value uncertainties. An easy to read visualization of the model sensitivity and effective design space of the material emerges, enabling effective communication between the various members of the design team. This UQ approach was applied to a computational framework that prescribes heat treatments based on desired size distributions of precipitates. Inputs were separated into two groups i) precipitation model parameters and ii) processing variables, based on their end-use relevance to material scientists and process engineers, respectively.

Knowing the thermal conductivity of a material is important whenever thermal processes are involved in a physical property (such as thermoelectricity) or in an industrial process. It is interesting to predict the thermal properties of a material by computer simulations before going to phases of synthesis and characterization that are often long and expensive. Therefore we propose an original and reliable method of determining by parameterized numerical simulations the thermal conductivity of a material by eliminating the existing problems in the current methods. We extend the procedure developed for glasses [1] based on a “cold” and a “hot” plate in the simulation box, to crystalline materials. The materials under study are crystalline iron oxide (mimicking the bulk of an automobile engine) as well as liquids (PAO, PAG,...) supposed to model the main component of a lubricant. [1] Philippe Jund and Rémi Jullien, Phys. Rev. B 59, 13707 (1999)