arXiv subjects

Brett I. Dunlap

Publications and source records attributed to Brett I. Dunlap.

16 recordsLinked to original sources

arXiv2018

Correlated Dynamics in Aqueous Proton Diffusion

The aqueous proton displays an anomalously large diffusion coefficient that is up to 7 times that of similarly sized cations. There is general consensus that the proton achieves its high diffusion through the Grotthuss mechanism, whereby protons hop from one molecule to the next. A main assumption concerning the extraction of the timescale of the Grotthuss mechanism from experimental results has been that, on average, there is an equal probability for the proton to hop to any of its neighboring water molecules. Herein, we present ab initio simulations that show this assumption is not generally valid. Specifically, we observe that there is an increased probability for the proton to revert back to its previous location. These correlations indicate that the interpretation of the experimental results need to be re-examined and suggest that the timescale of the Grotthuss mechanism is significantly shorter than was previously thought.

physics.chem-ph↗

arXiv2017

General degeneracy in density functional perturbation theory

Degenerate perturbation theory from quantum mechanics is inadequate in density functional theory (DFT) because of nonlinearity in the Kohn-Sham potential. Herein, we develop the fully general perturbation theory for open-shell, degenerate systems in Kohn-Sham DFT, without assuming the presence of symmetry or equal occupation of degenerate orbitals. To demonstrate the resulting methodology, we apply it to the iron atom in the central field approximation, perturbed by an electric quadrupole. This system was chosen because it displays both symmetry required degeneracy, between the five 3\textit{d} orbitals, as well as accidental degeneracy, between the 3\textit{d} and 4\textit{s} orbitals. The quadrupole potential couples the degenerate 3\textit{d} and 4\textit{s} states, serving as an example of the most general perturbation.

cond-mat.mtrl-sci↗

arXiv2016

Degenerate Density Perturbation Theory

Fractional occupation numbers can be used in density functional theory to create a symmetric Kohn-Sham potential, resulting in orbitals with degenerate eigenvalues. We develop the corresponding perturbation theory and apply it to a system of $N_d$ degenerate electrons in a harmonic oscillator potential. The order-by-order expansions of both the fractional occupation numbers and unitary transformations within the degenerate subspace are determined by the requirement that a differentiable map exists connecting the initial and perturbed states. Using the X$α$ exchange-correlation (XC) functional, we find an analytic solution for the first-order density and first through third-order energies as a function of $α$, with and without a self-interaction correction. The fact that the XC Hessian is not positive definite plays an important role in the behavior of the occupation numbers.

cond-mat.mtrl-sci↗

arXiv2016

Energy Continuity in Degenerate Density Functional Perturbation Theory

Fractional occupation numbers can produce open-shell degeneracy in density functional theory. We develop the corresponding perturbation theory by requiring that a differentiable map connects the initial and perturbed states. The degenerate state connects to a single perturbed state which extremizes, but does not necessarily minimize or maximize, the energy with respect to occupation numbers. Using a system of three electrons in a harmonic oscillator potential, we relate the counterintuitive sign of first-order occupation numbers to eigenvalues of the electron-electron interaction Hessian.

cond-mat.mtrl-sci↗

arXiv2007

Comparative study of unscreened and screened molecular static linear polarizability in the Hartree-Fock, hybrid-density functional, and density functional models

The sum-over-states (SOS) polarizabilities are calculated within the approximate independent electron theories such as the Hartree-Fock approximation and density functional models using the eigenvalues and orbitals obtained from the self-consistent solution of single-particle equations. The SOS polarizabilities are then compared with those calculated using the finite-field method within three widely used single particle models: (1) the Hartree-Fock (HF) method, (2) the three parameter hybrid generalized gradient approximation (B3LYP), and (3) the parameter free generalized gradient approximation due to Perdew-Burke-Ernzerhof (PBE). The comparison is carried out for polarizabilities of 142 molecules calculated using the 6-311++G(d,p) orbital basis at the geometries optimized at the B3LYP/6-311G** level. The results show that the SOS method almost always overestimate the FF polarizabilities in the PBE and B3LYP models. This trend is reversed in the HF method however exception to the trend do exist. The mean absolute errors (MAE) in the screened (FF) and unscreened (SOS) polarizability are 0.78 Å$^3$ for the HF method, 1.87 Å$^3$ for the B3LYP, and 3.44 Å$^3$ for the PBE-GGA. The Hartree-Fock approximation predicts finite-field polarizabilities that are smaller than those predicted by the PBE and B3LYP models. Finally, a simple scheme is devised to obtain FF quality estimate of polarizability from those obtained using the SOS method by exploiting the observed trends.

cond-mat.other↗

arXiv2006

Dipole moments from atomic-number-dependent potentials in analytic density-functional theory

Molecular dipole moments of analytic density-functional theory are investigated. The effect of element-dependent exchange potentials on these moments are examined by comparison with conventional quantum-chemical methods and experiment for the subset of the extended G2 set of molecules that have nonzero dipole moment. Fitting the Kohn-Sham [Phys. Rev. 140, A1133 (1965)] potential itself makes a mean absolute error of less than 0.1 Debye. Variation of alpha (Slater's [Phys. Rev. 81, 385 (1951)] exchange parameter) values has far less effect on dipole moments than on energies. It is argued that in variable alpha methods one should choose the smaller of the two rather than the geometric mean of the two alpha values for the heteroatomic part of the linear-combination-atomic-orbital density. Calculations on the dipole moment of NH2(CH)24NO2 are consistent with earlier calculations and show that varying the differences between alpha values for atoms with different atomic numbers has only short-ranged electrostatic effects.

cond-mat.other↗

arXiv2006

Fully analytic implementation of density functional theory for efficient calculations on large molecules

Fullerene like cages and naonotubes of carbon and other inorganic materials are currently under intense study due to their possible technological applications. First principle simulations of these materials are computationally challenging due to large number of atoms. We have recently developed a fast, variational and fully analytic density functional theory (ADFT) based model that allows study of systems larger than those that can be studied using existing density functional models. Using polarized Gaussian basis sets (6-311G**) and ADFT, we optimize geometries of large fullerenes, fullerene-like cages and nanotubes of carbon, boron nitride, and aluminum nitride containing more than two thousand atoms. The calculation of C2160 using nearly 39000 orbital basis functions is the largest calculation on any isolated molecule reported to-date at this level of theory, and it includes full geometry optimization. The electronic structure related properties of the inorganic cages and other carbon fulerenes have been studied.

cond-mat.mtrl-sci↗

arXiv2006

Efficient quantum-chemical geometry optimization and the structure of large icosahedral fullerenes

Geometry optimization is efficient using generalized Gaunt coefficients, which significantly limit the amount of cross differentiation for multi-center integrals of high-angular-momentum solid-harmonic basis sets. We parameterize the fully analytic formulation of density functional theory (ADFT), called the Slater-Roothaan method, developed in our group to give the exact geometry of C60 fullerene. The parametrized ADFT is subsequently used to optimize geometries of most stable C240, C540, C960, C1500 and C2160 icosahedral fullerenes. The calculations are all electron, the orbital basis set includes d functions and the exchange-correlation-potential basis set includes f functions. The calculation of C2160 fullerene employed about 39000 basis functions and is the largest calculation reported on any isolated molecule to-date. The evolution of interatomic distance and atomization energy from C60 to graphite has been investigated.

cond-mat.mtrl-sci↗

arXiv2005

Slater's exchange paramters for analytic and variational xalpha calculations

We treat the local exchange parameter, alpha, of the Slater's Xalpha method as a fitting parameter and determine its value by constraining some physical quantity obtained from the self-consistent solution of the Xalpha method to be equal to its "exact" value. Thus, the alpha values that reproduce the exact atomic energies are obtained for four different combinations of basis sets. A similar set of alpha values that is independent of basis set is obtained from numerical calculation. The alpha values are obtained for the atoms H through Cl. The performance of the analytic Slater-Roothaan (SR) method, that permits calculations to be performed accurate to machine precision, is examined using these set of alpha values for atomization energies of molecules belonging to the G2 set. The calculations indicate that the average absolute and mean errors are larger than those obtained using the Hartree-Fock alpha values in the SR method, but smaller than those for the standard Hartree-Fock theory or Kohn-Sham local density approximation. Alternative set of alpha values is determined by matching the highest occupied eigenvalue of the SR method to negative of the first ionization potential. Finally, the alpha values for the diatomic molecules of these atoms that yield the "exact" atomization energy are obtained. We find that the alpha values for the diatomic molecules show much larger deviation than what is observed for the atomic alpha values. The results obtained are useful in the development of the analytic Slater-Roothaan Xalpha method for molecular calculations.

cond-mat.other↗

arXiv2005

The limitations of Slater's element-dependent exchange functional from analytic density functional theory

Our recent formulation of the analytic and variational Slater-Roothaan (SR) method, which uses Gaussian basis sets to variationally express the molecular orbitals, electron density and the one body effective potential of density functional theory, is reviewed. Variational fitting can be extended to the resolution of identity method,where variationality then refers to the error in each two electron integral and not to the total energy. It is proposed that the appropriate fitting functions be charge neutral and that all ab initio energies be evaluated using two-center fits of the two-electron integrals. The SR method has its root in the Slater's Xalpha method and permits an arbitrary scaling of the Slater-Gaspar-Kohn-Sham exchange-correlation potential around each atom in the system. Of several ways of choosing the scaling factors (Slater's exchange parameters), two most obvious are the Hartree-Fock (HF), alpha_HF, values and the exact atomic, alpha_EA, values. The performance of this simple analytic model with both sets for atomization energies of G2 set of 148 molecules is better than the local density approximation or the HF theory, although the errors in atomization energy are larger than the target chemical accuracy. To improve peformance for atomization energies, the SR method is reparametrized to give atomization energies of 148 molecules to be comparbale to those obtained by one of the most widely used generalized gradient approximations. The mean absolute error in ionization potentials of 49 atoms and molecules is about 0.5 eV and that in bond distances of 27 molecules is about 0.02 Angstrom. The overall good performance of the computationally efficient SR method using any reasonable set of alpha values makes it a promising method for study of large systems.

cond-mat.other↗

arXiv2005

Electronic structure of fullerene-like cages and finite nanotubes of aluminum nitride

We report density functional study of alternate fullerene-like cage structures and finite closed capped single-wall nanotubes of aluminum nitride. The cages and nanotubes studied are modeled as Al24N24, Al28N28, Al32N32, Al36N36, Al48N48, and Al96N96. The structure optimization and calculation of the electronic structure, vertical ionization potential, and the electron affinity are performed at the all electron level by the analytic Slater-Roothaan method, using polarized Gaussian basis set of double zeta quality. All structures are energetically stable with binding energy of about 10-11 eV per AlN pair. For the larger Al96N96, the fullerene like cage is energetically less favorable than the two-shell cluster that has Al24N24 as an inner shell. The vertical ionization potential and the electronic affinity are in the range 6.7-6.9 eV and 1.5-2.0 eV, respectively. The binding energy show systematic increase with increase in the length of (4,4) nanotube. The energy band gap, determined using the "Delta SCF" method show that these structures are characterized by a fairly large band gap about 4-5 eV, which is however smaller than the gap for the corresponding boron nitride structures.

cond-mat.mtrl-sci↗

arXiv2005

Accurate molecular energies by extrapolation of atomic energies using an analytic quantum mechanical model

Using a new analytic quantum mechanical method based on Slater's Xalpha method, we show that a fairly accurate estimate of the total energy of a molecule can be obtained from the exact energies of its constituent atoms. The mean absolute error in the total energies thus determined for the G2 set of 56 molecules is about 16 kcal/mol, comparable to or better than some popular pure and hybrid density functional models.

cond-mat.mtrl-sci↗

arXiv2005

Theoretical infra-red, Raman, and Optical spectra of the B36N36 cage

The B36N36 fullerene-like cage structure was proposed as candidate structure for the single-shell boron-nitride cages observed in electron-beam irradiation experiment. We have performed all electron density functional calculations, with large polarized Gaussian basis sets, on the B36N36 cage. We show that the cage is energetically and vibrationally stable. The infra-red, Raman and optical spectra are calculated. The predicted spectra, in combination with experimentally measured spectra, will be useful in conclusive assignment of the proposed B36N36 cage. The vertical and adiabatic ionization potentials as well as static dipole polarizability are also reported.

physics.atm-clus↗

arXiv2004

On optimal values of alpha for the analytic Hartree-Fock-Slater method

We have examined the performance of the analytic Hartree-Fock-Slater (HFS) method for various alpha (Slater's exchange parameter) values and empiricaly determined the optimal alpha value by minimizing the mean absolute error (MAE) in atomization energies of the G2 set of molecules. At the optimal alpha the HFS method's performance is far superior with the MAE of 14 kcal/mol than that of the local density approximation (MAE ~ 36 kcal/mol) or the Hartree-Fock theory (MAE ~ 78 kcal/mol). The HFS exchange functional with alpha = 0.7091 is found to perform significantly better than the Kohn-Sham exchange functional for equally weighted atoms H-Kr. We speculate that use of this alpha value may be useful in parametrization of empirical exchange-correlation functionals.

cond-mat.other↗

arXiv2004

Electronic structure, vibrational stability, infra-red, and Raman spectra of B24N24 cages

We examine the vibrational stability of three candidate structures for the B24N24 cage and report their infra-red (IR) and Raman spectra. The candidate structures considered are a round cage with octahedral O symmetry, a cage with S_4 symmetry that satisfies the isolated square rule, and a cage of S_8 symmetry, which combines the caps of the (4,4) nanotube, and contains two extra squares and octagons. The calculations are performed within density functional theory, at the all electron level, with large basis sets, and within the generalized gradient approximation. The vertical ionization potential (VIP) and static dipole polarizability are also reported. The S_4 and S_8 cages are energetically nearly degenerate and are favored over the O cage which has six extra octagons and squares. The IR and Raman spectra of the three clusters show notable differences providing thereby a way to identify and possibly synthesize the cages.

physics.atm-clus↗

arXiv2004

Are hemispherical caps of boron-nitride nanotubes possible?

We report all-electron, density-functional calculations with large Gaussian polarization basis set of the recently synthesized octahedral B24N24 cage that is perfectly round by symmetry, and boron-nitride (BN) clusters that its existence might suggest. We consider whether it is energetically possible that the two halves of this round cage could cap the BN nanotubes, modeled by B28N28 and B32N32. The energetics show that BN nanotubes with such round caps, are only slightly less favorable than the BN clusters containing six squares as the only defects in the otherwise perfect hexagonal lattice. A larger B96N96 octahedral cage formed from B24N24 by adding sufficient hexagons to isolate all squares is not very favorable energetically. The squares protrude noticeably from its otherwise round surface.

physics.atm-clus↗