G. Malcolm Stocks, Oak Ridge National Laboratory
Bruce N. Harmon, Ames Laboratory/Iowa State University
Michael Weinert, Brookhaven National Laboratory

Research Objectives

The objective is to develop first-principles quantum mechanical methods for addressing materials problems microscopically, especially the relationship between technical magnetic properties and microstructure. Towards this goal are major problems involving microstructure and magnetism independently.

Computational Approach

A number of first-principles techniques are used to perform fundamental studies of the atomistic, electronic, and magnetic structure of metals, alloys, and semiconductors. The goal is to understand the influence on properties of microstructural defects and inhomogeneities. First principles methods include an iterative pseudopotential (IP) method, locally self-consistent multiple scattering (LSMS) and layer Korringa-Kohn-Rostoker (LKKR) methods, and tight-binding molecular dynamics (TBMD). In addition, first principles spin dynamics is being developed as a fundamental theory of the magnetic properties of metals and alloys.

Accomplishments

Non-collinear magnetic structure at interfaces: The LSMS method was used to calculate the magnetic structure at interfaces between Cu and permalloy (Py). Multiple non-collinear metastable magnetic configurations with similar energies were found. The non-collinearity deduced from these calculations explains the observed trend in saturation magnetization as a function of permalloy layer thickness. The same method was used to study the magnetic structure of Cu-Ni interfaces using very large cell models (883 atoms) obtained from independent deposition modeling. Taken together, the work on Cu-Py and Cu-Ni interfaces demonstrates that the constrained local moment theory implemented in the LSMS can provide accurate descriptions of the magnetic structure at technologically important magnetic interfaces.

Magnetic structure at interfaces: Distribution of magnetic moments in a Cu-Ni multilayer calculated using the first-principles LSMS method. Ni atoms are coded blue and Cu atoms green. The size of the magnetic moments is color-coded onto the arrows. Red arrows indicate Ni atoms whose magnetic moment is near the Ni bulk value. The magnetic moment is suppressed on atoms at the Ni-Cu interface. The atomic positions were obtained from Monte Carlo modeling of the deposition process. The structural model contains 883 atoms per unit cell.

Atomistic simulations of martensitic transformations: Large-scale simulations of both the high-temperature body-centered-cubic (bcc) phase of Zr, as well as the transformation from bcc to hexagonal-closest-packed (hcp) have been performed. These simulations allow for a full calculation of the dynamic structure factor and how the scattering evolves as the system transforms from the bcc to the hcp phase. The actual atomic arrangements are observed during the transformation process, in order to learn about the nucleation and growth processes. The calculated X-ray scattering of the high-temperature phase shows the characteristic streaking of the Bragg peaks observed in experiments. By examining the scattering as a function of time, it has been shown that the streaking in the high temperature phase mimics the coherent movement of Bragg peaks that occurs during the transformation. This demonstrates that the observed anomalous scattering is caused by coherent fluctuations towards the hcp phase. The development of microstructure has also been observed as the transition progresses.

Dislocation-nucleated twin boundaries in hcp metals: One of the fundamental questions in materials science is to understand the nucleation and growth of twin boundaries, and the competition between twinning and slip deformation modes. These issues can play an important role in the ductility of materials and are currently being studied in hcp metals, where the ability to twin makes Zr and Ti very ductile even at low temperatures. Conversely, materials such as Mg and Be do not twin, and are brittle. Large-scale atomistic simulations have shown that twin boundaries may be nucleated via dislocation cores. One such observation was of a dislocation in an hcp metal that had dissociated into a large twin nucleus, with a small partial dislocation at the bottom of the simulation cell. These two dislocations are connected by a stacking fault, making the arrangement difficult to move. Instead, tension along the c-axis has the effect of causing the twinned region to grow. This arrangement has not been seen in previous simulations, due to the small simulation sizes used previously. This provides a microscopic explanation for the observed fact that systems that twin under c-axis compression also twin under tension.

First-principles calculations of Mo₅Si₃ and Ti₅Si₃: Intermetallic M₅Si₃-type silicides have been of great interest as potential candidates for high-temperature materials operating above 1500 C. First-principles calculations for the phase stability, bonding mechanism and elastic properties of this material have been performed. Results are in good agreement with experimental measurements. An extensive study of the lattice parameters of C, B, N, and O-doped D₈₈-Ti₅Si₃ has been made. In collaboration with the experimental group of M. Akinc at Ames Lab, it was demonstrated that the calculated heats of formation and the variation of lattice constants and interatomic distances compare well with experimental data.

	Atomistic simulations of martensitic transformations: Final simulation cell obtained during a simulation of a transformation of Zr from the high-temperature bcc phase to the low-temperature hcp phase. The atoms are colored according to potential energy (with blue being the highest). The domain walls are evident. The system develops long domains along the [111] direction as the transition progresses. This allows three different hcp domains to form.

Core level shifts in metallic alloys: Experimentally, core level binding energy shifts can be measured using electron spectroscopy for chemical analysis (ESCA). Chemical shifts and their distribution about their mean have been calculated for three alloy systems, CuPd, CuZu, and AgPd. The calculations were based on large supercell models of the disordered phase that contain hundreds of atoms and were performed using first principles order-N LSMS method. Results were compared with predictions based on the commonly used ESCA potential model that relates the core shifts to charge transfer. Since the charge transfer is also obtained in the first principles calculations, this allows detailed testing of the ESCA model. While first principles calculations provide reliable predictions for the chemical shifts in the alloys, the relationship between chemical shifts and the charge transfer do not agree with the ESCA potential model.

Significance

The availability of powerful and accurate first-principles techniques permits the study of quantum interatomic interactions on a length scale not previously accessible, opening up the possibility of relating these fundamental interatomic interactions to the strength, ductility, transport and magnetic properties of materials. Applied to magnetic materials, these techniques should help establish the foundations for understanding the relationship between the technical magnetic properties (permeability, coercivity, remenance) of magnets and microstructure.

Publications

G. M. Stocks, B. Ujfalussy, X. Wang, D. M. C. Nicholson, W. A. Shelton, Y. Wang, and B. L. Gyorffy, "Towards a constrained local moment model for first principles spin dynamics," Phil. Mag. B 78, 665 (1998).

J. S. Faulkner, Y. Wang, and G. M. Stocks, "Core level chemical shifts in metallic alloys," Phys. Rev. Letters 81, 1905 (1998).

J. R. Morris, Z. Y. Lu, D. Ring, J. B. Xiang, C. Z. Wang, K. M. Ho, and C.-L. Fu, "First-principles determination of the = 13 {510} symmetric tilt boundary structure in silicon and germanium," Phys. Rev. B 58, 11241 (1998)