MedeA® Application Notes for Electronics

Patrick Soukiassian, Erich Wimmer Edvige Celasco, Claudia Giallombardo, Simon Bonanni, Luca Vattuone, Letizia Savio, Anontio Tejeda, Mathieu Silly, Marie D’angelo, Fausto Sirotti, Mario Rocca

Nanostructuring a surface is a key and mandatory engineering step toward advances in nanotechnology. A team of french and italian scientists and of a franco-american company has just shown that hydrogen/deuterium (H/D) induces the opening of nanotunnels below the surface of an advanced semiconductor, silicon carbide (SiC). Such a finding is an especially interesting one, particularly in views of the specific properties of SiC. These investigations have been performed using advanced experimental tools such as synchrotron radiation and vibrational spectroscopy techniques, and state-of-the-art theoretical simulations. Depending on the H/D SiC surface exposures, these nanotunnels undergo through a sequence of semiconducting/metallic/semiconducting transitions. Such nanotunnels open very promising prospects toward applications in electronic, chemistry, storage, sensors and biotechnology.

The electronic structures of the metallic and insulating phases of VO₂ are calculated using density functional theory and modern hybrid functionals as recently implemented in MedeA-VASP. Strongly contrasting previous calculations as based on local or semilocal approximations, which missed the insulating behavior of the low-temperature phase, these new calculations accurately reproduce the optical band gap and thus bring to an end a fifty-year old controversy on the origin of the metal-insulator transition of VO₂.

Accurate Band Gaps of Correlated Transition-Metal Oxides from MedeA-VASP with Hybrid-Functionals

We demonstrate the capabilities of MedeA with selected examples, focusing on the lattice thermal
conductivity using forcefield methods as implemented in the software environment. The thermal conductivity is of high interest in different fields. In thermoelectrics, materials are sought with a high electrical conductivity combined with a low thermal conductivity as can be found in doped semiconductors with a high density of states near the band edges. In the present paper, we investigate the thermal conductivity of Si-Ge alloys and discuss the influence of defects, and disorder. All the computations are done using MedeA's LAMMPS and Thermal Conductivity modules, with the Reverse Non-Equilibrium Molecular Dynamics (RNEMD) approach.

Lattice thermal conductivity calculations in Si-Ge

The contact resistance between metals and semiconductors in nanoelectronic devices is mainly determined by the Schottky barrier. Controlling the Schottky barrier height (SBH) hence means being able to manipulate the contact resistance. and thereby to reduce the energy consumption as well as the heat production of electronic devices. While so far phenomenological considerations were able to determine the SBH only at a qualitative level, a parameter-free quantitative evaluation is possible via atomistic simulations using the MedeA software platform.. This application note illustrates the calculation and modification of the SBH for a system of direct technological importance, namely a NiSi/Si contact. Furthermore, the effect of dopant atoms on the SBH, which are needed to tune the SBH for minimal contact resistance of n- and p-doped semiconductors, is investigated. Reduction of the Schottky barrier height is achieved by doping with Ba. This note also demonstrates how the preferred positions of dopant elements can be determined with S as example.

Prediction of Schottky Barrier in Electronic Devices

The energy band structure of InAs is computed with the HSE06 hybrid functional using MedeA-VASP. The calculations give a direct band gap of 0.35 eV, which coincides with the experimental value. The Γ-L separation in the lowest conduction band is computed to be 1.11 eV. Screened-exchange FLAPW calculations reported a value of 1.21 eV. The earlier handbook value was 0.74 eV. More recent experiments reported 1.1±0.05 eV, which is perfectly consistent with the present calculations.

Energy Band Structure of InAs with MedeA-VASP

In the design of substrates for the epitaxial growth of InxGa1-xN alloys it is useful to know the elastic properties of the semiconductor as a function of composition. This application note shows the use of MedeA with VASP 5.2 and the mechanical-thermal (MT) module in computing these properties. Judging by the results for the binaries GaN and InN, the level of accuracy is comparable with that achieved in experiments.

Structure and Elastic Properties of (In,Ga)N

First-principles calculations reveal a three-fold increase in the Young’s modulus of graphite as it is lithiated (C→LiC₆). A linear expression is determined that describes the approximate stiffness of Li intercalated graphite as a function of loading which may lead to greatly improved continuum models of electrode deformation and failure.

Graphite Electrode Elastic Properties upon Li Intercalation

A key process in the semiconductor manufacturing is the reactive adsorption of molecules such as
silane (SiH4) and dichlorosilane (SiCl2H2) on the surfaces of silicon wafers. This case study
demonstrates the calculation of the geometry of a silane molecule on a reconstructed Si(001)

Dissociation of SiH₄ on Si(001) Surface

First-principles computations correctly describe the ferroelectric distortions and the macroscopic polarization of BaTiO₃ in agreement with experiment. Computations of the vibrational properties (phonons) reveal that a cubic perovskite structure of BaTiO₃ becomes stable under compression of the lattice. This demonstrates the usefulness of first-principles calculations in the design and optimization of ferroelectric materials.

Ferroelectric Properties of BaTiO₃

The work function of the metal gate in a CMOS stack depends on the composition and structure of the interfaces. This is demonstrated here for the case of a Si-HfO2-W stack by introducing a Hf vacancy at the Si/HfO2 interface. At a concentration of 1.2 vacancies per nm2 the work functions is increased by 500 meV.

Modeling work function changes in CMOS stacks containing HfO₂ high-k dielectrics

Elemental germanium is a semiconductor with a measured indirect band gap of 0.66 eV. Using a hybrid functional as implemented in VASP 5.2, the computed value is 0.66 eV while standard density functional approaches incorrectly predict Ge to have no band gap. Other features of the band structure such as the direct gap at Γ are also well reproduced by the current level of theory, namely 0.8 eV (measured) and 0.73 eV (computed), thus demonstrating the reliability of this level of approach in predicting energy band structures. This sets the stage for using computations to modify the band structure for example by uniaxial strain to meet specific design criteria.

Energy band structure of germanium