Abstract: The properties of alternative metals to Cu and W for interconnect applications are reviewed based on first-principles simulations and benchmarked in terms of intrinsic bulk resistivity and electromigration.

Abstract: RuCp2 (ruthenocene) and RuCpPy (cyclopentadienyl pyrrolyl ruthenium) complexes are used in ruthenium (Ru) atomic layer deposition (ALD) but exhibit a markedly different reactivity with respect to the substrate and co-reactant. In search of an explanation, we report here the results of a comparative study of the heterolytic and homolytic dissociation enthalpy of these two ruthenium complexes, making use of either density functional theory (DFT) or multiconfigurational perturbation theory (CASPT2). While both methods predict distinctly different absolute dissociation enthalpies, they agree on the relative values between both molecules. A reduced heterolytic dissociation enthalpy is obtained for RuCpPy compared to RuCp2, although the difference obtained from CASPT2 (19.9 kcal/mol) is slightly larger than the one obtained with any of the DFT functionals (around 17 kcal/mol). Both methods also agree on the more pronounced stability of the Cp- ligand in RuCpPy than in RuCp2 (by around 9 kcal/mol with DFT and by 6 kcal/mol with CASPT2).

Abstract: The application of coreshell SiSiO2 nanowires as nanoelectronic devices strongly depends on their structure, which is difficult to tune precisely. In this work, we investigate the formation of the coreshell nanowires at the atomic scale, by reactive molecular dynamics simulations. The occurrence of two temperature-dependent oxidation mechanisms of ultra-small diameter Si-NWs is demonstrated. We found that control over the Si-core radius and the SiOx (x ≤ 2) oxide shell is possible by tuning the growth temperature and the initial Si-NW diameter. Two different structures were obtained, i.e., ultrathin SiO2 silica nanowires at high temperature and Si core|ultrathin SiO2 silica nanowires at low temperature. The transition temperature is found to linearly decrease with the nanowire curvature. Finally, the interfacial stress is found to be responsible for self-limiting oxidation, depending on both the initial Si-NW radius and the oxide growth temperature. These novel insights allow us to gain control over the exact morphology and structure of the wires, as is needed for their application in nanoelectronics.

Abstract: The electronic properties of a set of intrinsic and extrinsic point defects in gamma-Al2O3 are investigated using quasiparticle calculations within the G(0)W(0) approximation. We find that the electronic signature of atomic vacancies lie deep in the band gap, close to the top of the valence band edge. The introduction of C, Si, and N impurities induces defective levels that are located close to the conduction band edge and near the middle of the band gap of the oxide. The comparison with electrical measurements reveals that the energy levels of some of these defects match with the electronic fingerprint of the defects reported in gamma-Al2O3 based nonvolatile memories. (C) 2010 American Institute of Physics. [doi:10.1063/1.3507385]

Abstract: Continued miniaturization of metal-oxide-semiconductor field-effect transistors (MOSFETs) requires an ever-decreasing thickness of the gate oxide. The structure of ultrathin silicon oxide films, however, critically depends on the oxidation mechanism. Using reactive atomistic simulations, we here demonstrate how the oxidation mechanism in hyperthermal oxidation of such structures may be controlled by the oxidation temperature and the oxidant energy. Specifically, we study the interaction of hyperthermal oxygen with energies of 15 eV with thin SiOx (x ≤ 2) films with a native oxide thickness of about 10 Å. We analyze the oxygen penetration depth probability and compare with results of the hyperthermal oxidation of a bare Si(100){2 × 1} (c-Si) surface. The temperature-dependent oxidation mechanisms are discussed in detail. Our results demonstrate that, at low (i.e., room) temperature, the penetrated oxygen mostly resides in the oxide region rather than at the SiOx|c-Si interface. However, at higher temperatures, starting at around 700 K, oxygen atoms are found to penetrate and to diffuse through the oxide layer followed by reaction at the c-Si boundary. We demonstrate that hyperthermal oxidation resembles thermal oxidation, which can be described by the DealGrove model at high temperatures. Furthermore, defect creation mechanisms that occur during the oxidation process are also analyzed. This study is useful for the fabrication of ultrathin silicon oxide gate oxides for metal-oxide-semiconductor devices as it links parameters that can be straightforwardly controlled in experiment (oxygen temperature, velocity) with the silicon oxide structure.

Abstract: In this letter, we address the bipolar resistive switching phenomenology in scaled TiN\HfO2\Hf cells. By means of stack engineering using a thin Al2O3 layer inserted either at the TiN\HfO2 or at the Hf\HfO2 interface, we demonstrate that the reset operation takes place close to the TiNanode. Due to the increase of the oxygen-vacancy profile from the TiN to the Hf interface, the filament-confining and wide band-gap Al2O3 layer should indeed be engineered at the interface with the TiN electrode in order to further improve the switching control and to allow reaching larger state resistances. (C) 2012 The Electrochemical Society. [DOI: 10.1149/2.003204ssl] All rights reserved.

Abstract: Modeling the oxidation process of silicon nanowires through reactive force field based molecular dynamics simulations suggests that the formation of Si epoxide defects occurs both at the Si/SiOx interface and at the nanowire surface, whereas for flat surfaces, this defect is experimentally observed to occur only at the interface as a result of stress. In this paper, we argue that the increasing curvature stabilizes the defect at the nanowire surface, as suggested by our density functional theory calculations. The latter can have important consequences for the opto-electronic properties of thin silicon nanowires, since the epoxide induces an electronic state within the band gap. Removing the epoxide defect by hydrogenation is expected to be possible but becomes increasingly difficult with a reduction of the diameter of the nanowires.

Abstract: We have characterized theoretically the work-function modifications of the (111) surfaces of gold and silver upon deposition of self-assembled monolayers based on methanethiol and trifluoromethanethiol. A comparative analysis is made between the experimental results and those obtained from two widely used approaches based on density functional theory. The contributions to the total work-function modifications are estimated on the basis of two decomposition schemes of the thiol molecules that have been proposed in the literature. The contributions are found to differ significantly between the two approaches, as do the corresponding adsorption energies.

Abstract: This study presents first-principles results on the electronic functionalization of silicene and germanene monolayers by means of chemisorption of adatom species H, Li, F, Sc, Ti, V. Three general adatom-monolayer configurations are considered, each having its distinct effect on the electronic structure, yielding metallic or semiconducting dispersions depending on the adatom species and configuration. The induced bandgap is a (in)direct F gap ranging from 0.2 to 2.3 eV for both silicene and germanene. In general the alternating configuration was found to be the most energetically stable. The boatlike and chairlike conformers are degenerate with the former having anisotropic effective carrier masses. The top configuration leads to the planar monolayer and predominately to a gapped dispersion. The hollow configuration with V adatoms retains the Dirac cone, but with strong orbital planar hybridization at the Fermi level. We also observe a planar surface state the Fermi level for the latter systems. (C) 2013 Elsevier B.V. All rights reserved.

Abstract: Because of its excellent properties in nanotechnology applications, atomic layer deposition of ruthenium (Ru) has been the subject of numerous experimental studies. Recently, two different Ru precursors were compared for plasma-enhanced atomic layer deposition (PEALD) of Ru, and their reactivity was found to be different. Inhibition was observed for bis(ethylcyclopentadienyl)ruthenium (Ru(EtCp)(2)), while nearly linear growth behavior was observed for (methylcyclopentadienyl-pyrrolyl)ruthenium (Ru(MeCp)Py). To understand this difference in reactivity, we investigate the adsorption of RuCp, and RuCpPy (i.e., without substituents) on a TiN surface using calculations based on periodic boundary conditions density functional theory (DFT) combined with experiments based on Rutherford backscattering spectroscopy (RBS). The calculations demonstrate that the RuCpPy precursor chemisorbs on the TiN(100) surface while the RuCp2 precursor only physisorbs. We propose a reaction mechanism for the chemisorption of RuCpPy. The area density of the calculated RuCpPy surface species is compared with the experimental values from RBS. The impact of a H-plasma is also investigated. The DFT calculations and experimental results from RBS provide insight into the adsorption processes of the RuCpPy and RuCp2 precursors on the TiN(100) surface.

Abstract: The electronic and vibrational properties of 2D honeycomb structures of molybdenum disulfide (MoS2) subjected to strain have been investigated using first-principles calculations based on density functional theory. We have studied the evolution of the electronic properties of bulk and layered MoS2, going down from a few layers up to a mono-layer, and next investigated the effect of bi-axial strain on their electronic structure and vibrational frequencies. Both for tensile and compressive biaxial strains, the shrinking of the energy band-gap of MoS2 with increasing level of applied strain is observed and a transition limit of the system from semiconducting to metallic is predicted to occur for strains in the range of 8-10%. We also found a progressive downshift (upshift) of both the E-2g(1) and A(1g) Raman active modes with increasing level of applied tensile (compressive) strain. Interestingly, significant changes in the curvature of the conduction and valence band near their extrema upon the application of strain are also predicted, with correlated variations of the electron and hole effective masses. These changes present interesting possibilities for engineering the electronic properties of 2D structures of MoS2. (C) 2012 Elsevier B.V. All rights reserved.

Abstract: First-principles simulations were employed to gain atomistic insights on the working principles of amorphous HfO2 based Resistive Random Access Memory stack: the nature of the defect responsible for the switching between the High and Low Resistive States has been unambiguously identified to be the substoichiometric Hf sites (commonly called oxygen vacancy-V-O) and the kinetics of the process have been investigated through the study of O diffusion. Also the role of each material layer in the TiN/HfO2/Hf/TiN RRAM stack and the impact of the deposition techniques have been examined: metallic Hf sputtering is needed to provide an oxygen exchange layer that plays the role of defect buffer. TiN shall be a good defect barrier for O but a bad defect buffer layer. A possible scenario to explain the device degradation (switching failure) mechanism has been proposed – the relaxation of the metastable amorphous phase towards crystalline structure leads to denser, more structured cluster that can increase the defect migration barriers. (C) 2013 Elsevier B.V. All rights reserved.

Abstract: Two-dimensional group-IV lattices silicene and germanene are known to share many of graphene's remarkable mechanical and electronic properties. Due to the out-of-plane buckling of the former materials, there are more means of electronic funtionalization, e.g. by applying uniaxial strain or an out-of-plane electric field. We consider monolayer hexagonal Sn (stanene) as an ideal candidate to feasibly implement and exploit graphene physics for nanoelectronic applications: with increased out-of-plane buckling and sizable spin-orbit coupling it lends itself to improved Dirac cone engineering. We investigate the ballistic charge transport regime of armchair Sn nanoribbons, classified according to the ribbon width W = {3m – 1, 3m, 3m + 1} with integer m. We study transport through (non-magnetic) armchair ribbons using a combination of density functional theory and non-equilibrium Green's functions. Sn ribbons have earlier current onsets and carry currents 20% larger than C/Si/Ge-nanoribbons as the contact resistance of these ribbons is found to be comparable. ((c) 2014 WILEY-VCH Verlag GmbH &Co. KGaA, Weinheim)

Abstract: Atomic, layer deposition,(ALD of ruthenium using two ruthenium precursors, i.e., Ru(C5H5)(2) (RuCp2) and Ru(C5H5)(C4H4N) (RuCpPy), is studied using density functional theory. By investigating the reaction mechanisms On bare ruthenium surfaces, i.e., (001), (101), and (100), and H-terminated surfaces, an atomistic insight in the Ru ALD is provided. The calculated results show that on the Ru surfaces both RuCp2 and RuCpPy an undergo dehydrogenation and ligand dissociation reactions. RuCpPy is more reactive than RuCp2. By forming a, strong, bond between N of Py and Ru of the surface, RuCpPy can easily chemisorb on the surfaces. The reactions of RuCp2,On the Surfaces are less favorable the adsorption is not strong enough This could be a,factor contributing to the higher growth-per-cycle of Ru using RuCpPy, as observed experimentally. By Studying, the adsorption on H-terminated Ru surfaces, We showed that H Can prevent the adsorption of the precursors, thus inhibiting the growth of Ru. Our calculations indicate that the H content on the surface can have an impact on the growth-per-cycle. Finally, our simulations also demonstrate large impacts of the surface structure on the reaction mechanisms. Of the three surfaces, the (100) surface, which is the less stable and has a zigzag surface structure, is also the most reactive one.

Abstract: By using first-principles simulations, we investigate the interaction of a 2D silicon layer with two classes of chalcogenide-layered compounds, namely MoX2 and GaX (X = S, Se, Te). A rather weak (van der Waals) interaction between the silicene layers and the chalcogenide layers is predicted. We found that the buckling of the silicene layer is correlated to the lattice mismatch between the silicene layer and the MoX2 or GaX template. The electronic properties of silicene on these different templates largely depend on the buckling of the silicene layer: highly buckled silicene on MoS2 is predicted to be metallic, while low buckled silicene on GaS and GaSe is predicted to be semi-metallic, with preserved Dirac cones at the K points. These results indicate new routes for artificially engineering silicene nanosheets, providing tailored electronic properties of this 2D layer on non-metallic substrates. These non-metallic templates also open the way to the possible integration of silicene in future nanoelectronic devices.