Exploring the relationship between vibrational mode locality and coupling using constrained optimization.

The use of alternate coordinate systems as a means to improve the efficiency and accuracy of anharmonic vibrational structure analysis has seen renewed interest in recent years. While normal modes (which diagonalize the mass-weighted Hessian matrix) are a typical choice, the delocalized nature of this basis makes it less optimal when anharmonicity is in play. When a set of modes is not designed to treat anharmonicity, anharmonic effects will contribute to inter-mode coupling in an uncontrolled fashion. These effects can be mitigated by introducing locality, but this comes at its own cost of potentially large second-order coupling terms. Herein, a method is described which partially localizes vibrations to connect the fully delocalized and fully localized limits. This allows a balance between the treatment of harmonic and anharmonic coupling, which minimizes the error that arises from neglected coupling terms. Partially localized modes are investigated for a range of model systems including a tetramer of hydrogen fluoride, water dimer, ethene, diphenylethane, and stilbene. Generally, partial localization reaches ∼75% of maximal locality while introducing less than ∼30% of the harmonic coupling of the fully localized system. Furthermore, partial localization produces mode pairs that are spatially separated and thus weakly coupled to one another. It is likely that this property can be exploited in the creation of model Hamiltonians that omit the coupling parameters of the distant (and therefore uncoupled) pairs.

Optimizing Vibrational Coordinates To Modulate Intermode Coupling.

The choice of coordinate system strongly affects the convergence properties of vibrational structure computations. Two methods for efficient generation of improved vibrational coordinates are presented and justified by analysis of a model anharmonic two-mode Hessian and numerical computations on polyatomic molecules. To produce optimal coordinates, metrics which quantify off-diagonal couplings over a grid of Hessian matrices are minimized through unitary rotations of the vibrational basis. The first proposed metric minimizes the total squared off-diagonal coupling, and the second minimizes the total squared change in off-diagonal coupling. In this procedure certain anharmonic modes tend to localize, for example X-H stretches. The proposed methods do not rely on prior fitting of the potential energy, vibrational structure computations, or localization metrics, so they are unique from previous vibrational coordinate generation algorithms and are generally applicable to polyatomic molecules. Fitting the potential to the approximate n-mode representation in the optimized bases for all-trans polyenes shows that off-diagonal anharmonic couplings are substantially reduced by the new choices of coordinate system. Convergence of vibrational energies is examined in detail for ethylene, and it is shown that coupling-optimized modes converge in vibrational configuration interaction computations to within 1 cm(-1) using only 3-mode couplings, where normal modes require 4-mode couplings for convergence. Comparison of the vibrational configuration interaction convergence with respect to excitation level for the two proposed metrics shows that minimization of the total off-diagonal coupling is most effective for low-cost vibrational structure computations.

Orbitals with intermediate localization and low coupling: spanning the gap between canonical and localized orbitals.

Localized orbitals are representations of electronic structure, which are easier to interpret than delocalized, canonical orbitals. While unitary transformations from canonical orbitals into localized orbitals have long been known, existing techniques maximize localization without regard to coupling between orbitals. Especially in conjugated π spaces, orbitals are collapsed by unitary localization procedures into nonintuitive, strongly interacting units. Over-localization decreases interpretability, results in large values of interorbital coupling, and gives unmeaningful diagonal Fock energies. Herein, we introduce orbitals of intermediate localization that span between canonical and fully localized orbitals. To within a specified error, these orbitals preserve the diagonal nature of the Fock matrix while still introducing significant locality. In systems composed of molecular fragments, π spaces can be localized into weakly coupled units. Importantly, as the weakly coupled orbitals separate, highly coupled orbitals maintain their expected structure. The resulting orbitals therefore correspond well to chemical intuition and maintain accurate orbital energies, making this procedure unique among existing orbital localization techniques. This article focuses on the formation and physical analysis of orbitals that smoothly connect the known fully delocalized and fully localized limits.

Modeling the elastic energy of alloys: Potential pitfalls of continuum treatments.

Some issues that arise when modeling elastic energy for binary alloys are discussed within the context of a Keating model and density-functional calculations. The Keating model is a simplified atomistic formulation based on modeling elastic interactions of a binary alloy with harmonic springs whose equilibrium length is species dependent. It is demonstrated that the continuum limit for the strain field are the usual equations of linear elasticity for alloys and that they correctly capture the coarse-grained behavior of the displacement field. In addition, it is established that Euler-Lagrange equation of the continuum limit of the elastic energy will yield the same strain field equation. This is the same energy functional that is often used to model elastic effects in binary alloys. However, a direct calculation of the elastic energy atomistic model reveals that the continuum expression for the elastic energy is both qualitatively and quantitatively incorrect. This is because it does not take atomistic scale compositional nonuniformity into account. Importantly, this result also shows that finely mixed alloys tend to have more elastic energy than segregated systems, which is the exact opposite of predictions made by some continuum theories. It is also shown that for strained thin films the traditionally used effective misfit for alloys systematically underestimate the strain energy. In some models, this drawback is handled by including an elastic contribution to the enthalpy of mixing, which is characterized in terms of the continuum concentration. The direct calculation of the atomistic model reveals that this approach suffers serious difficulties. It is demonstrated that elastic contribution to the enthalpy of mixing is nonisotropic and scale dependent. It is also shown that such effects are present in density-functional theory calculations for the Si-Ge system. This work demonstrates that it is critical to include the microscopic arrangements in any elastic model to achieve even qualitatively correct behavior.

Filling of hole arrays with InAs quantum dots.

Focused ion beams are used to pattern GaAs(001) surfaces with an array of nanometer-deep holes upon which deposition of InAs results in quantum dot formation at the hole location. Experiments show that the size and quantity of quantum dots formed depend on growth parameters, and ion dose, which affects the size and shape of the resulting holes. Quantum dots fabricated in this fashion have a photoluminescence peak at 1.28 eV at 77 K, indicating that the ion irradiation due to patterning does not destroy their optical activity. Kinetic Monte Carlo simulations that include elastic relaxation qualitatively model the growth of dots in nanometer-deep holes, and demonstrate that growth temperature, depth of the holes, and the angle of the hole sidewalls strongly influence the number of quantum dots that form at their perimeter.