Search

International audience; Using a new time-resolved cathodoluminescence system dedicated to the UV spectral range, we present a first estimate of the radiative lifetime of free excitons in hBN at room temperature. This is carried out from a single experiment giving both the absolute luminescence intensity under continuous excitation and the decay time of free excitons in the time domain. The radiative lifetime of indirect excitons in hBN is equal to 27 ns, which is much shorter than in other indirect bandgap semiconductors. This is explained by the close proximity of the electron and the hole in the exciton complex, and also by the small energy difference between indirect and direct excitons. The unusually high luminescence efficiency of hBN for an indirect bandgap is therefore semi-quantitatively understood.

The properties of semiconductors can be crucially impacted by midgap states induced by dopants, which can be native or intentionally incorporated in the crystal lattice. For Bernal-stacked bilayer graphene (BLG), which has a tunable bandgap, the existence of midgap states induced by dopants has been conjectured, but never confirmed experimentally. Here, we report scanning tunneling microscopy and spectroscopy results, supported by tight-binding calculations, that demonstrate the existence of midgap states in BLG. We show that the midgap state in BLG -- for which we demonstrate gate-tunability -- appears when the dopant is hosted on the non-dimer sublattice sites. We further evidence the presence of narrow resonances at the onset of the high energy bands (valence or conduction, depending on the dopant type) when the dopants lie on the dimer sublattice sites. These results suggest that dopants/defects can play an important role in the transport and optical properties of multilayer graphene samples, especially at energies close to the band extrema., Includes supplementary material

We study the mechanical and electronic properties of heterobilayers composed of black phosphorus (BP) on hexagonal boron nitride (hBN) and of blue phosphorus (${\mathrm{P}}_{\mathrm{blue}}$) on hBN by means of ab initio density functional theory. Emphasis is put on how the stress applied on the constituent layers impact their structural and electronic properties. For this purpose, we adopt a specific scheme of structural relaxation which allows us to distinguish between the energy cost of distorting each layer and the gain in stacking them together. In most cases we find that the BP tends to contract along the softer armchair direction, as already reported for similar structures. This contraction can attain up to 5% of strain, which might deteriorate its very good transport properties along the armchair direction. To prevent this, we propose a twisted-bilayer configuration where the largest part of the stress applies on the zigzag axis, resulting in a lower impact on the transport properties of BP. We also investigated a ${\mathrm{P}}_{\mathrm{blue}}$/hBN bilayer. A peculiar hybridization between the valence states of the two layers lets us suggest that electron-hole pairs excited in the bilayer will exhibit a mixed character, with electrons localized solely in the ${\mathrm{P}}_{\mathrm{blue}}$ layer and holes spread onto the two layers.

International audience; Black phosphorus (BP), a 2D semiconducting material of interest in electronics and photonics, exhibits physical properties characterized by strong anisotropy and band gap energy that scales with reducing layer number. However, the investigation of its intrinsic properties is challenging because thin layer BP are photo oxidized in ambient conditions and the energy of their electronic states shift in different dielectric environment. We prepared free-standing samples of few layer BP in glovebox conditions and probed the dielectric response in vacuum using Scanning Transmission Electron Microscopy and Electron Energy Loss Spectroscopy (STEM-EELS). Thresholds of the excitation energy are measured at 1.9 eV, 1.4 eV and 1.1 eV for the mono-bi-and tri-layer BP, respectively and these values are used to estimate the corresponding optical band gaps. A comparison of our results with electronic structure calculations indicates that the variation of the quasi-particle gap is larger than that of the exciton binding energy. The dispersion of the plasmons versus momentum for 1-3 layer BP and bulk BP highlights a deviation from parabolic to linear dispersion and strong anisotropic fingerprints.

In this letter we report a thorough analysis of the exciton dispersion in bulk hexagonal boron nitride. We solve the ab initio GW Bethe-Salpeter equation at finite $\mathbf{q}\parallel \Gamma K$, and we compare our results with recent high-accuracy electron energy loss data. Simulations reproduce the measured dispersion and the variation of the peak intensity. We focus on the evolution of the intensity, and we demonstrate that the excitonic peak is formed by the superposition of two groups of transitions that we call $KM$ and $MK'$ from the k-points involved in the transitions. These two groups contribute to the peak intensity with opposite signs, each damping the contributions of the other. The variations in number and amplitude of these transitions determine the changes in intensity of the peak. Our results contribute to the understanding of electronic excitations in this systems along the $\Gamma K$ direction, which is the relevant direction for spectroscopic measurements. They also unveil the non-trivial relation between valley physics and excitonic dispersion in h--BN, opening the possibility to tune excitonic effects by playing with the interference between transitions. Furthermore, this study introduces analysis tools and a methodology that are completely general. They suggest a way to regroup independent-particle transitions which could permit a deeper understanding of excitonic properties in any system.

International audience; We present a thorough tight-binding analysis of the band structure of a wide variety of lattices belonging to the class of honeycomb and Kagome systems including several mixed forms combining both lattices. The band structure of these systems are made of a combination of dispersive and flat bands. The dispersive bands possess Dirac cones (linear dispersion) at the six corners (K points) of the Brillouin zone although in peculiar cases Dirac cones at the center of the zone ($\ gamma$ point) appear. The flat bands can be of different nature. Most of them are tangent to the dispersive bands at the center of the zone but some, for symmetry reasons, do not hybridize with other states. The objective of our work is to provide an analysis of a wide class of so-called ligand-decorated honeycomb Kagome lattices that are observed in 2D metal-organic framework (MOF) where the ligand occupy honeycomb sites and the metallic atoms the Kagome sites. We show that the $p_x$-$p_y$ graphene model is relevant in these systems and there exists four types of flat bands: Kagome flat (singly degenerate) bands, two kinds of ligand-centered flat bands (A$_2$ like and E like, respectively doubly and singly degenerate) and metal-centered (three fold degenerate) flat bands.

International audience; Electron energy loss spectra were measured on hexagonal boron nitride single crystals employing an electron energy loss spectroscopic setup composed of an electron microscope equipped with a monochromator and an in-column filter. This setup provides high-quality energy-loss spectra and allows also for the imaging of energy-filtered diffraction patterns. These two acquisition modes provide complementary pieces of information, offering a global view of excitations in reciprocal space. As an example of the capabilities of the method we show how easily the core loss spectra at the K edges of boron and nitrogen can be measured and imaged. Low losses associated with interband and/or plasmon excitations are also measured. This energy range allows us to illustrate that our method provides results whose quality is comparable to that obtained from nonresonant x-ray inelastic scattering but with advantageous specificities such as an enhanced sensitivity at low q and a much greater simplicity and versatility that make it well adapted to the study of two-dimensional materials and related heterostructures. Finally, by comparing theoretical calculations to our measures, we are able to relate the range of applicability of ab initio calculations to the anisotropy of the sample and assess the level of approximation required for a proper simulation of our acquisition method.

Boron nitride single layer belongs to the family of two-dimensional materials whose optical properties are currently receiving considerable attention. Strong excitonic effects have already been observed in the bulk and still stronger effects are predicted for single layers. We present here a detailed study of these properties by combining ab initio calculations and a tight-binding Wannier analysis in both real and reciprocal space. Due to the simplicity of the band structure with single valence $(\ensuremath{\pi})$ and conduction $({\ensuremath{\pi}}^{*})$ bands the tight-binding analysis becomes quasiquantitative with only two adjustable parameters and provides tools for a detailed analysis of the exciton properties. Strong deviations from the usual hydrogenic model are evidenced. The ground-state exciton is not a genuine Frenkel exciton, but a very localized tightly bound one. The other ones are similar to those found in transition-metal dichalcogenides and, although more localized, can be described within a Wannier-Mott scheme.

International audience; Scanning tunneling microscopy experiments have been performed to measure the local electron injection in nitrogen-doped graphene on SiC(000¯1) and were successfully compared to ab initio calculations. In graphene, a gaplike feature is measured around the Fermi level due to a phonon-mediated tunneling channel. At nitrogen sites, this feature vanishes due to an increase of the elastic channel that is allowed because of symmetry breaking induced by the nitrogen atoms. A large conductance enhancement by a factor of up to 500 was measured at the Fermi level by comparing local spectroscopy at nitrogen sites and at carbon sites. Nitrogen doping can therefore be proposed as a way to improve tunnel-electron injection in graphene.

cited By 20; International audience; Many potential applications of graphene require either the possibility of tuning its electronic structure or the addition of reactive sites on its chemically inert basal plane. Among the various strategies proposed to reach these objectives, nitrogen doping, i.e., the incorporation of nitrogen atoms in the carbon lattice, leads in most cases to a globally n-doped material and to the presence of various types of point defects. In this context, the interactions between chemical dopants in graphene have important consequences on the electronic properties of the systems and cannot be neglected when interpreting spectroscopic data or setting up devices. In this report, the structural and electronic properties of complex doping sites in nitrogen-doped graphene have been investigated by means of scanning tunneling microscopy and spectroscopy, supported by density functional theory and tight-binding calculations. In particular, based on combined experimental and simulation works, we have systematically studied the electronic fingerprints of complex doping configurations made of pairs of substitutional nitrogen atoms. Localized bonding states are observed between the Dirac point and the Fermi level in contrast with the unoccupied state associated with single substitutional N atoms. For pyridinic nitrogen sites (i.e., the combination of N atoms with vacancies), a resonant state is observed close to the Dirac energy. This insight into the modifications of electronic structure induced by nitrogen doping in graphene provides us with a fair understanding of complex doping configurations in graphene, as it appears in real samples. © 2015 American Chemical Society.

International audience; Hexagonal boron nitride (h-BN) and graphite are structurally similar but with very different properties. Their combination in graphene-based devices meets now a huge research focus, and it becomes particularly important to evaluate the role played by crystalline defects in them. In this work, the cathodoluminescence (CL) properties of hexagonal boron nitride crystallites are reported and compared to those of nanosheets mechanically exfoliated from them. First the link between the presence of structural defects and the recombination intensity of bound-excitons, the so-called D series, is confirmed. Low defective h-BN regions are further evidenced by CL spectral mapping (hyperspectral imaging), allowing us to observe new features in the near-band-edge region, tentatively attributed to phonon replica of exciton recombinations. Second the h-BN thickness was reduced down to six atomic layers, using mechanical exfoliation, as evidenced by atomic force microscopy. Even at these low thicknesses, the luminescence remains intense and exciton recombination energies are not strongly modified with respect to the bulk, as expected from theoretical calculations indicating extremely compact excitons in h-BN.

cited By 8; International audience; Using scanning tunnelling microscopy and spectroscopy, we investigated the atomic and electronic structure of nitrogen-doped single walled carbon nanotubes synthesized by chemical vapor deposition. The insertion of nitrogen in the carbon lattice induces several types of point defects involving different atomic configurations. Spectroscopic measurements on semiconducting nanotubes reveal that these local structures can induce either extended shallow levels or more localized deep levels. In a metallic tube, a single doping site associated with a donor state was observed in the gap at an energy close to that of the first van Hove singularity. Density functional theory calculations reveal that this feature corresponds to a substitutional nitrogen atom in the carbon network. © 2013 American Chemical Society.

Optimized growth of single wall carbon nanotubes requires full knowledge of the actual state of the catalyst nanoparticle and its interface with the tube. Using tight binding based atomistic computer simulations, we calculate carbon adsorption isotherms on nanoparticles of nickel, a typical catalyst, and show that carbon solubility increases for smaller nanoparticles that are either molten or surface molten under experimental conditions. Increasing carbon content favors the dewetting of Ni nanoparticles with respect to $s{p}^{2}$ carbon walls, a necessary property to limit catalyst encapsulation and deactivation. Grand canonical Monte Carlo simulations of the growth of tube embryos show that wetting properties of the nanoparticles, controlled by carbon solubility, are of fundamental importance to enable the growth, shedding new light on the growth mechanisms.

The aim of this lecture is to give an overlook about methods developed in infinite (bulk) and semi-infinite (surface) metallic materials and some tracks to extend them to finite size systems. In this framework we will first study the effect of bond breaking and dimension lowering on electronic structure, at surfaces of pure metals (surface states, atomic level shifts, reconstructions and relaxations) and in monometallic clusters. Then we will illustrate the influence of chemical ordering on electronic structure (and vice versa) by considering firstly bulk alloys (diagonal versus off-diagonal disorder) and then bimetallic surfaces (stress effect induced by either surface segregation or epitaxial growth). These two approaches will then naturally be combined in the peculiar case of nanoalloys. The methods will be developed following two main goals. The first one is to determine local electronic densities of states (LDOS), the knowledge of which is essential to the understanding and the analysis of nano-objects. The second one is to derive from these LDOS energetic models well suited to both the degree of complexity of the systems under study (bulk and surface crystalline structure, chemical ordering, …) and their implementation in numerical simulations (Molecular Dynamics, Monte Carlo). The different sections of the lecture will be illustrated by examples issued from studies performed on systems which can be considered as archetypal in the nano-alloy community, such as CoPt, CoAu and CuAg.

Although significant efforts have been directed towards a selective single wall carbon nanotube synthesis, the resulting diameter and chirality distributions are still too broad and their control remains a challenge. Progress in this direction requires an understanding of the mechanisms leading to the chiral selectivity reported by some authors. Here, we focus on one possible such mechanism and investigate the healing processes of defective tubes, at the atomic scale. We use tight-binding Monte Carlo simulations to perform a statistical analysis of the healing of a number of defective tubes. We study the role of temperature as a primary factor to overcome the energy barriers involved by healing, as well as the role of the metal catalyst. Using both electron diffraction patterns and local characterizations, we show that the healing proceeds first along the tube axis, before spreading laterally, and observe the competition between two or more chiralities. The resulting picture is that no chirality seems to be favored by the healing mechanisms, implying that the reported chiral preference should result from other sources.

The healing of graphene grown from a metallic substrate is investigated using tight-binding Monte Carlo simulations. At temperatures (ranging from 1000 to 2500 K), an isolated graphene sheet can anneal a large number of defects suggesting that their healings are thermally activated. We show that in the presence of a nickel substrate we obtain a perfect graphene layer. The nickel−carbon chemical bonds keep breaking and reforming around defected carbon zones, providing a direct interaction, necessary for the healing. Thus, the action of Ni atoms is found to play a key role in the reconstruction of the graphene sheet by annealing defects.

International audience; We present a tight-binding potential for transition metals, carbon, and transition-metal carbides, which has been optimized through a systematic fitting procedure. A minimal basis, including the s and p electrons of carbon and the d electrons of the transition metal, is used to obtain a transferable tight-binding model of the carbon-carbon, metal-metal, and metal-carbon interactions applicable to binary systems. The Ni-C system is more specifically discussed. The successful validation of the potential for different atomic configurations indicates a good transferability of the model and makes it a good choice for atomistic simulations sampling a large configuration space. This approach appears to be very efficient to describe interactions in systems containing carbon and transition-metal elements. By way of example, we present results concerning the epitaxial growth of graphene sheets on (111) Ni surfaces, as well as the catalytic nucleation of carbon nanotubes.

The nucleation of carbon nanotubes on small nickel clusters is studied using a tight binding model coupled to grand canonical Monte Carlo simulations. This technique follows the conditions of the synthesis of carbon nanotubes by chemical vapor deposition. The possible formation of a carbon cap on the catalyst particle is studied as a function of the carbon chemical potential, for particles of different size, either crystalline or disordered. We show that these parameters strongly influence the structure of the cap/particle interface which in turn will have a strong effect on the control of the structure of the nanotube. In particular, we discuss the presence of carbon on surface or in subsurface layers.

Cathodoluminescence imaging and spectroscopy experiments on a single bamboo-like boron nitride nanotube are reported. Imaging experiments show that the luminescence is located all along the nanotube. Spectroscopy experiments point out the important role of dimensionality in this one dimensional object.

The catalytic growth of single-wall carbon nanotubes is investigated by high-resolution transmission electron microscopy. The similarities between the samples synthesized from different techniques suggest a common growth mechanism based on a vapor-liquid-solid model. Quantum-molecular-dynamics simulations support a root growth mechanism where carbon atoms are incorporated into the tube base by a diffusion-segregation process.

The study by electron microscopy of the fringes due to planar defects such as antiphase boundaries (APB) is not a new problem, and the equations describing this phenomenon are well-known, but the effect on electron diffraction of the broadening with temperature of these boundaries close to order-disorder transitions is a new field of investigation. We show here that a careful study of the behaviour of the fringes generated by the APB in dark field images is a good method to analyze the different order-disorder transitions. In particular we describe in detail a quantitative method that allows us to determine the correlation length ξ below a second order transition.

The formation of single-wall carbon nanotubes requires the presence of a metal catalyst whatever the technique used. To understand their growth mechanism, we have initiated a theoretical study in order to simulate the segregation and the self-organisation of the carbon atoms at the surface of metallic nanoparticles. We have developed a semi-empirical model based on tight-binding method to describe binary systems. Using this model, we have investigated the effect of carbon atoms on the structural properties of (100) and (111) metallic surfaces. Those studies show different behaviours of the surface as a function of the carbon coverage. We have also studied the first stages of the nucleation process using Monte Carlo simulation in the grand canonical ensemble and demonstrated the role played by the metal in the dissolution-segregation process of carbon atoms.; La production de nanotubes monofeuillets de carbone requiert la présence d'un catalyseur métallique quel que soit le mode de synthèse. Pour comprendre leur mécanisme de croissance, nous avons entrepris une étude théorique visant à rendre compte des phénomènes de ségrégation et d'auto-organisation du carbone à la surface des particules métalliques. Ce travail a abouti à la mise en place d'un modèle semi-phénoménologique reposant sur la méthode des liaisons fortes. A l'aide de notre modèle énergétique, nous avons cherché à comprendre l'effet du carbone sur les propriétés structurales des surfaces métalliques (100) et (111). Nous avons ainsi pu mettre en évidence des comportements différents suivant le taux d couverture en carbone des surfaces. Nous avons également procédé à l'étude du mécanisme de germination-croissance à travers des simulations grand canonique Monte Carlo qui montrent le rôle joué par le métal dans le processus de dissolution-ségrégation d'atomes de carbone.

The formation of single-wall carbon nanotubes requires the presence of a metal catalyst whatever the technique used. To understand their growth mechanism, we have initiated a theoretical study in order to simulate the segregation and the self-organisation of the carbon atoms at the surface of metallic nanoparticles. We have developed a semi-empirical model based on tight-binding method to describe binary systems. Using this model, we have investigated the effect of carbon atoms on the structural properties of (100) and (111) metallic surfaces. Those studies show different behaviours of the surface as a function of the carbon coverage. We have also studied the first stages of the nucleation process using Monte Carlo simulation in the grand canonical ensemble and demonstrated the role played by the metal in the dissolution-segregation process of carbon atoms., La production de nanotubes monofeuillets de carbone requiert la présence d'un catalyseur métallique quel que soit le mode de synthèse. Pour comprendre leur mécanisme de croissance, nous avons entrepris une étude théorique visant à rendre compte des phénomènes de ségrégation et d'auto-organisation du carbone à la surface des particules métalliques. Ce travail a abouti à la mise en place d'un modèle semi-phénoménologique reposant sur la méthode des liaisons fortes. A l'aide de notre modèle énergétique, nous avons cherché à comprendre l'effet du carbone sur les propriétés structurales des surfaces métalliques (100) et (111). Nous avons ainsi pu mettre en évidence des comportements différents suivant le taux d couverture en carbone des surfaces. Nous avons également procédé à l'étude du mécanisme de germination-croissance à travers des simulations grand canonique Monte Carlo qui montrent le rôle joué par le métal dans le processus de dissolution-ségrégation d'atomes de carbone.