Search

Zigzag edges of neutral armchair-oriented Graphene Nano-Ribbons show states strongly localized at those edges. They behave as free radicals that can capture electrons during processing, increasing ribbon's stability. Thus, charging and its consequences should be investigated.Total energy calculations of finite ribbons using spin polarized Density Functional Theory (DFT) show that ribbon's charging is feasible. Energies for Pariser-Parr-Pople (PPP) model Hamiltonian are compatible with DFT allowing the study of larger systems. Results for neutral ribbons indicate: i) the fundamental gap of spin polarized (non polarized) solutions is larger (smaller) than experimental data, ii) the ground state is spin polarized, a characteristic still not observed experimentally. Total energy of GNRs decreases with the number of captured electrons reaching a minimum for a number that mainly depends on zigzag edges size. The following changes with respect to neutral GNRs are noted: i) the ground state is not spin polarized, ii) fundamental gap is in-between that of spin polarized and non polarized solutions of neutral ribbons, iii) while in neutral ribbons valence and conduction band onsets vs. the fundamental gap, linearly and symmetrically approach mid-gap with slope 0.5, charging induces Fermi level pinning, i.e., the slopes of the valence and conduction bands being about 0.1 and 0.9, in agreement with experiment., Comment: 15 pages containing seven figures, changes in notation and presentation, unpublished

Controlling the forbidden gap of graphene nano-ribbons (GNR) is a major challenge that has to be attained if this attractive material has to be used in micro- and nano-electronics. Using an unambiguous notation {m,n}-GNR, where m (n) is the number of six carbon rings in the arm-chair (zig-zag) directions, we investigate how varies the HOMO-LUMO gap when the size of the GNR is varied by increasing either m or n, while keeping the other variable fixed. It is shown that no matter whether charge- or spin-density-waves solutions are considered, the gap varies smoothly when n is kept fixed whereas it oscillates when the opposite is done, posing serious difficulties to the control of the gap. It is argued that the origin of this behavior is the fact that excess or defect charges or magnetic moments are mostly localized at zig-zag edges., Comment: 7 pages, 7 figures

We investigate the size scaling of the conductance of surface disordered graphene sheets of width W and length L. Metallic leads are attached to the sample ends across its width. At E ~ 0, the conductance scales with the system size as follows: i) For constant W/L, it remains constant as size is increased, at a value which depends almost lineally on that ratio; this scaling allows the definition of a conductivity value that results similar to the experimental one. ii) For fixed width, the conductance decreases exponentially with length L, both for ordered and disordered samples. Disorder reduces the exponential decay, leading to a higher conductance. iii) For constant length, conductance increases linearly with width W, a result that is exclusively due to the tails of the states of the metallic wide contact. iv) The average conductance does not show an appreciable dependence on magnetic field. Away from E = 0, the conductance shows the behavior expected in two-dimensional systems with surface disorder, i.e., ballistic transport., Comment: 4 pages, 5 figures, RevTex

We present a state-of-the-art first-principles analysis of electronic transport in a Pt nanocontact in the presence of H2 which has been recently reported by Smit et al. in Nature 419, 906 (2002). Our results indicate that at the last stages of the breaking of the Pt nanocontact two basic forms of bridge involving H can appear. Our claim is, in contrast to Smit et al.'s, that the main conductance histogram peak at G approx 2e^2/h is not due to molecular H2, but to a complex Pt2H2 where the H2 molecule dissociates. A first-principles vibrational analysis that compares favorably with the experimental one also supports our claim ., Comment: 5 pages, 3 figures

Here we report on several anomalies in quantum transport at the band center of a bipartite lattice with vacancies that are surely due to its chiral symmetry, namely: no weak localization effect shows up, and, when leads have a single channel the transmission is either one or zero. We propose that these are a consequence of both the chiral symmetry and the large number of states at the band center. The probability amplitude associated to the eigenstate that gives unit transmission ressembles a classical trajectory both with or without vacancies. The large number of states allows to build up trajectories that elude the blocking vacancies explaining the absence of weak localization., Comment: 5 pages, 5 figures

Keldysh formalism is used to get the current-voltage characteristic of a small system of interacting electrons described by a Hubbard model coupled to metallic wires. The numerical procedure is checked recovering well-known results for an Anderson impurity. When larger interacting regions are considered quite different results are obtained depending on whether the Hubbard part is half-filled or not. At half-filling the existence of a gap charge manifests itself making current exponentially small as a function both of the number of interacting sites and the value of U. The behavior changes at large voltages above the gap energy when activated charge transport takes place. On the contrary, for filling factors other than half, current goes through the interacting system suffering just a small amount of scattering at both connections. Conductance depends slightly on U and much more on the filling factor but not on the length of the interacting region., Comment: 14 twocolumn pages including 15 figures, RevTex4, amended version

We discuss the key steps that have to be followed to calculate coherent quantum transport in molecular and atomic-scale systems, making emphasis on the ab-initio Gaussian Embedded Cluster Method recently developed by the authors. We present various results on a simple system such as a clean Au nanocontact and the same nanocontact in the presence of hydrogen that illustrate the applicability of this method in the study and interpretation of a large range of experiments in the field of molecular electronics., Comment: Lecture notes of tutorial given at the Summer School Nicolas Cabrera on "Molecular Electronics" held in Miraflores (Spain), September 2002

We present a first-principles numerical implementation of Landauer formalism for electrical transport in nanostructures characterized down to the atomic level. The novelty and interest of our method lies essentially on two facts. First of all, it makes use of the versatile Gaussian98 code, which is widely used within the quantum chemistry community. Secondly, it incorporates the semi-infinite electrodes in a very generic and efficient way by means of Bethe lattices. We name this method the Gaussian Embedded Cluster Method (GECM). In order to make contact with other proposed implementations, we illustrate our technique by calculating the conductance in some well-studied systems such as metallic (Al and Au) nanocontacts and C-atom chains connected to metallic (Al and Au) electrodes. In the case of Al nanocontacts the conductance turns out to be quite dependent on the detailed atomic arrangement. On the contrary, the conductance in Au nanocontacts presents quite universal features. In the case of C chains, where the self-consistency guarantees the local charge transfer and the correct alignment of the molecular and electrode levels, we find that the conductance oscillates with the number of atoms in the chain regardless of the type of electrode. However, for short chains and Al electrodes the even-odd periodicity is reversed at equilibrium bond distances., Comment: 14 pages, two-column format, submitted to PRB

Kubo formula is used to get the scaling behavior of the static conductance distribution of wide wires showing pure non-diagonal disorder. Following recent works that point to unusual phenomena in some circumstances, scaling at the band center of wires of odd widths has been numerically investigated. While the conductance mean shows a decrease that is only proportional to the inverse square root of the wire length, the median of the distribution exponentially decreases as a function of the square root of the length. Actually, the whole distribution decays as the inverse square root of the length except close to G=0 where the distribution accumulates the weight lost at larger conductances. It accurately follows the theoretical prediction once the free parameter is correctly fitted. Moreover, when the number of channels equals the wire length but contacts are kept finite, the conductance distribution is still described by the previous model. It is shown that the common origin of this behavior is a simple Gaussian statistics followed by the logarithm of the E=0 wavefunction weight ratio of a system showing chiral symmetry. A finite value of the two-dimensional conductance mean is obtained in the infinite size limit. Both conductance and the wavefunction statistics distributions are given in this limit. This results are consistent with the 'critical' character of the E=0 wavefunction predicted in the literature., Comment: 10 pages, 9 figures, RevTeX macro

Building upon traditional quantum chemistry calculations, we have implemented an {\em ab-initio} method to study the electrical transport in nanocontacts. We illustrate our technique calculating the conductance of C$_{60}$ molecules connected in various ways to Al electrodes characterized at the atomic level. Central to a correct estimate of the electrical current is a precise knowledge of the local charge transfer between molecule and metal which, in turn, guarantees the correct positioning of the Fermi level with respect to the molecular orbitals. Contrary to our expectations, ballistic transport seems to occur in this system., Comment: 4 pages in two-column format

The magnetoconductance G in chaotic quantum dots at medium/high magnetic fluxes Phi is calculated by means of a tight binding Hamiltonian on a square lattice. Chaotic dots are simulated by introducing diagonal disorder on surface sites of L x L clusters. It is shown that when the ratio W/L is sufficiently large, W being the leads width, G increases steadily showing a maximum at a flux Phi_max ~ W. Bulk disordered ballistic cavities (with an amount of impurities proportional to L) does not show this effect. On the other hand, for magnetic fluxes larger than that for which the cyclotron radius is of the order of L/2, the average magnetoconductance inceases almost linearly with the flux with a slope proportional to W^2, shows a maximum and then decreases stepwise. These results closely follow a theory proposed by Beenakker and van Houten to explain the magnetoconductance of two point contacts in series., Comment: RevTeX including six postscript figures

Magnetotransport in chaotic quantum dots at low magnetic fields is investigated by means of a tight binding Hamiltonian on L x L clusters of the square lattice. Chaoticity is induced by introducing L bulk vacancies. The dependence of weak localization on the Fermi energy, dot size and leads width is investigated in detail and the results compared with those of previous analyses, in particular with random matrix theory predictions. Our results indicate that the dependence of the critical flux Phi_c on the square root of the number of open modes, as predicted by random matrix theory, is obscured by the strong energy dependence of the proportionality constant. Instead, the size dependence of the critical flux predicted by Efetov and random matrix theory, namely, Phi_c ~ sqrt{1/L}, is clearly illustrated by the present results. Our numerical results do also show that the weak localization term significantly decreases as the leads width W approaches L. However, calculations for W=L indicate that the weak localization effect does not disappear as L increases., Comment: RevTeX, 8 postscript figures included

Kubo formula is used to get the d.c conductance of a statistical ensemble of two-dimensional clusters of the square lattice in the presence of standard diagonal disorder, a uniform magnetic field and random magnetic fluxes. Working within a one-band tight-binding approach the calculation is quite general. The shape of the cluster is rectangular with ideal leads attached to opposite corners. Both geometrical characteristics and physical parameters can be easily selected. The output is just the conductance of a system of given parameters or a statistical ensemble of conductances measured for different disorder realizations., Comment: 14 pages, one table, no figures, RevTeX style

The elastic mean free path of carriers in a recently introduced model of quantum chaotic billiards in two and three dimensions is calculated. The model incorporates surface roughness at a microscopic scale by randomly choosing the atomic levels at the surface sites between -W/2 and W/2. Surface roughness yields a mean free path l that decreases as L/W^2 as W increases, L being the linear size of the system. But this diminution ceases when the surface layer begins to decouple from the bulk for large enough values of W, leaving more or less unperturbed states on the bulk. Consequently, the mean free path shows a minimum of about L/2 for W of the order of the band width. Energy fluctuations reflect the behavior of the mean free path. At small energy scales, strong level correlations manifest themselves by small values of the number of levels variance Sigma^2(E) that are close to Random Matrix Theory (RMT) in all cases. At larger energy scales, fluctuations are below the logarithmic behavior of RMT for l > L, and above RMT value when l < L., Comment: 8 twocolumn pages, seven figures, revtex and epsf macros. To be published in Physical Review B.