Session E10: Other

A Convex Hull Algorithm for Finding New Superalloys

In computational materials science identifying new stable phases is a primary strategy for developing new materials. Constitutional phase diagrams can be calculated from first-principles to narrow the scope of experimental searches. In complicated materials, those that contain more than two components, identifying these phases from many calculations is a convex hull finding problem. We have developed a framework that implements the qhull code[1] and automatically generates convex hulls for ternary intermetallic systems. Using this framework we examined 2224 systems in which we identified 75 new superalloys that have not been reported in experimental literature. Discovering new superalloys such as these has potential to revolutionize materials of the 21$^{st}$ century. [1]www.qhull.org   

Understanding Cu migration in CdTe solar cells

In this work, we report on development of one-dimensional (1D) finite-difference and two-dimensional (2D) finite-element diffusion-reaction simulators to investigate mechanisms behind Cu-related metastabilities observed in CdTe solar cells. To achieve such capability, the simulators solve reaction-diffusion equations for the defect states in time-space domain self-consistently with the free carrier transport. Evolution of concentration profile for an arbitrary defect (including drift, diffusion and reaction) is simulated by solving reaction-diffusion equation. In our 2D FEM scheme, anisotropic diffusion model for a single grain boundary is utilized to simulate fast movement of defects in the grain boundaries. For a finite difference mesh with a symmetric grain boundary, this model gives identical discretization to the traditional Fisher model. Results of 1-D and 2-D simulations have been compared to verify the accuracy of solutions. The simulation results from this study give us a deeper understanding of the role of Cu on the performance of CdTe solar cells.   

A Time-Decomposition Method and Applications to Physical Problems

Physical phenomena such as wave propagation and singularity formation can be described with partial differential equations (PDEs). ~In many cases, such PDEs originate in physical, hyperbolic systems. ~Often, analytic solutions are not available. ~In this talk, we present a time-parallel numerical method to solve such PDEs with physically relevant results. ~The model problems we apply this method to are taken from relativistic quantum mechanical systems including the Klein-Gordon equation, the Dirac equation and the semilinear wave equation. ~   

Improving Band Interpolation and Band Integration in DFT

A multitude of energy band interpolation and integration methods have been developed as a means to improve the accuracy and reduce computer run times of material's properties calculations. Each method adjusts one or more of the following parameters: the location of the sampling points, the interpolation basis functions, the conditions imposed on the interpolation and the integration technique employed. Common problems encountered when interpolating energy bands include handling band crossings, band kissings, and cusps. Our approach fits the energy bands with splines, finds the Fermi level with ``adaptive cubes", integrates the bands using Gaussian quadrature, employing various smoothings at the Fermi level, and extrapolates the smoothed integrals of the spline representation to the band integral. We attempted to fix band crossings and kissings with band character decomposition; however, the resolution it provided was insufficient to distinguish one band from another in all cases. We discuss the potential of the tight-binding pseudo-atomic orbitals (PAO) basis to resolve band crossing and kissing issues.   

How T-symmetry decides the laws thermodynamics

The "arrow of time" is synonymous with the second law of thermodynamics. For the first time simulations show T-symmetry (the symmetry of transforming $t$ to $-t$) is required for entropy to reach its true maximum value. Moreover, preserving the symmetry is required to preserve the zeroth law. The simulations use an Ising-like model attributed to Creutz. Additional degrees of freedom flip binary spins deterministically rather than with Monte Carlo techniques. It is analytically demonstrable that the added quanta satisfy or violate T-symmetry, depending on the particle interactions. Thus we see for the first time the laws of thermodynamics are only valid when T-symmetry is preserved. This provides theoretical confirmation for neutral kaon decay and testable predictions for electric dipole moment experiments. Because low-entropy states still evolve to states of higher entropy when T-symmetry is violated, but only reach the true maximum when it is satisfied, the simulations constitute a new solution of Loschmidt's Paradox: entropy increases only in the future, despite the time-reversal invariance of classical mechanics. The coarse-graining solution to the paradox from the Ehrenfests is unneeded in this more fundamental approach.   

A Monte Carlo approach to study transport in silicon heterojunction solar cells.

The device performance of an amorphous silicon (a-Si)/crystalline silicon (c-Si) solar cell depends strongly on the interfacial transport properties of the device. The energy of the photogenerated carriers at the barrier strongly depends on the strength of the inversion at the heterointerface and their collection requires interaction with the defects present in the intrinsic amorphous silicon buffer layer. In this work we present a theoretical model to study the transport through the heterointerface by applying an ensemble Mone Carlo (EMC) and a kinetic Monte Carlo (KMC). The EMC studies carrier behavior at the heterointerface whereas the KMC method allows us to simulate the interaction of discrete carriers with discrete defects. This method allows us to study defect transport which takes place on a time scale which is too long for traditional ensemble Monte Carlo's to analyze. We calculate the injection and extraction of carriers via defects by calculating transition rates, i.e. probability of transition to defect states within the intrinsic amorphous silicon barrier. The KMC results allow us to quantitatively study the properties of the heterointerface barrier in terms of how it affects transport.