I. Knezevic

Electronic transport in nanometre-scale silicon-on-insulator membranes

The widely used ‘silicon-on-insulator’ (SOI) system consists of a layer of single-crystalline silicon supported on a silicon dioxide substrate. When this silicon layer (the template layer) is very thin, the assumption that an effectively infinite number of atoms contributes to its physical properties no longer applies, and new electronic, mechanical and thermodynamic phenomena arise, distinct from those of bulk silicon. The development of unusual electronic properties with decreasing layer thickness is particularly important for silicon microelectronic devices, in which (001)-oriented SOI is often used. Here we show—using scanning tunnelling microscopy, electronic transport measurements, and theory—that electronic conduction in thin SOI(001) is determined not by bulk dopants but by the interaction of surface or interface electronic energy levels with the ‘bulk’ band structure of the thin silicon template layer. This interaction enables high-mobility carrier conduction in nanometre-scale SOI; conduction in even the thinnest membranes or layers of Si(001) is therefore possible, independent of any considerations of bulk doping, provided that the proper surface or interface states are available to enable the thermal excitation of ‘bulk’ carriers in the silicon layer.

Ballistic to diffusive crossover of heat flow in graphene ribbons

Heat flow in nanomaterials is an important area of study, with both fundamental and technological implications. However, little is known about heat flow in two-dimensional devices or interconnects with dimensions comparable to the phonon mean free path. Here we find that short, quarter-micron graphene samples reach ~35% of the ballistic thermal conductance limit up to room temperature, enabled by the relatively large phonon mean free path (~100 nm) in substrate-supported graphene. In contrast, patterning similar samples into nanoribbons leads to a diffusive heat-flow regime that is controlled by ribbon width and edge disorder. In the edge-controlled regime, the graphene nanoribbon thermal conductivity scales with width approximately as ~W(1.8)(0.3), being about 100 W m(-1) K(-1) in 65-nm-wide graphene nanoribbons, at room temperature. These results show how manipulation of two-dimensional device dimensions and edges can be used to achieve full control of their heat-carrying properties, approaching fundamentally limited upper or lower bounds.

Electron transport in silicon nanowires: The role of acoustic phonon confinement and surface roughness scattering

We investigate the effects of electron and acoustic phonon confinements on the low-field electron mobility of thin, gated, square silicon nanowires (SiNWs), surrounded by SiO2. We employ a self-consistent Poisson–Schrodinger–Monte Carlo solver that accounts for scattering due to acoustic phonons (confined and bulk), intervalley phonons, and the Si/SiO2 surface roughness. The wires considered have cross sections between 3×3 and 8×8 nm2. For larger wires, the dependence of the mobility on the transverse field from the gate is pronounced, as expected. At low transverse fields, where phonon scattering dominates, scattering from confined acoustic phonons results in about a 10% decrease in the mobility with respect to the bulk phonon approximation. As the wire cross section decreases, the electron mobility drops because the detrimental increase in both electron-acoustic phonon and electron-surface roughness scattering rates overshadows the beneficial volume inversion and subband modulation. For wires thinner th...

Lattice thermal conductivity of graphene nanoribbons: Anisotropy and edge roughness scattering

We present a calculation of the thermal conductivity of graphene nanoribbons GNRs, based on solving the Boltzmann transport equation with the full phonon dispersions, a momentum-dependent model for edge roughness scattering, as well as three-phonon and isotope scattering. The interplay between edge roughness scattering and the anisotropy of the phonon dispersions results in thermal conduction that depends on the chiral angle of the nanoribbon. Lowest thermal conductivity occurs in the armchair direction and highest in zig-zag nanoribbons. Both the thermal conductivity and the degree of armchair/zig-zag anisotropy depend strongly on the width of the nanoribbon and the rms height of the edge roughness, with the smallest and most anisotropic thermal conductivities occurring in narrow GNRs with rough edges.

Electron Mobility in Silicon Nanowires

The low-field electron mobility in rectangular silicon nanowire (SiNW) transistors was computed using a self-consistent Poisson-Schroumldinger-Monte Carlo solver. The behavior of the phonon-limited and surface-roughness-limited components of the mobility was investigated by decreasing the wire width from 30 to 8 nm, the width range capturing a crossover between two-dimensional and one-dimensional electron transport. The phonon-limited mobility, which characterizes transport at low and moderate transverse fields, is found to decrease with decreasing wire width due to an increase in the electron-phonon wavefunction overlap. In contrast, the mobility at very high transverse fields, which is limited by surface roughness scattering, increases with decreasing wire width due to volume inversion. The importance of acoustic phonon confinement is also discussed briefly

Mechano-electronic Superlattices in Silicon Nanoribbons

Significant new mechanical and electronic phenomena can arise in single-crystal semiconductors when their thickness reaches nanometer dimensions, where the two surfaces of the crystal are physically close enough to each other that what happens at one surface influences what happens at the other. We show experimentally that, in silicon nanomembranes, through-membrane elastic interactions cause the double-sided ordering of epitaxially grown nanostressors that locally and periodically highly strains the membrane, leading to a strain lattice. Because strain influences band structure, we create a periodic band gap modulation, up to 20% of the band gap, effectively an electronic superlattice. Our calculations demonstrate that discrete minibands can form in the potential wells of an electronic superlattice generated by Ge nanostressors on a sufficiently thin Si(001) nanomembrane at the temperature of 77 K. We predict that it is possible to observe discrete minibands in Si nanoribbons at room temperature if nanostressors of a different material are grown.

X-valley leakage in GaAs-based midinfrared quantum cascade lasers: A Monte Carlo study

We present a detailed Monte Carlo simulation of electron transport incorporating both Γ- and X-valley states in GaAs-based quantum cascade lasers (QCLs). Γ states are calculated using the K⋅p method, while X states are obtained within the effective mass framework. All the relevant electron-phonon, electron-electron, and intervalley scattering mechanisms are included. We investigate the X-valley leakage in two equivalent-design GaAs/AlGaAs QCLs with 33% and 45% Al-barrier compositions. We find that the dominant X-valley leakage path in both laser structures is through interstage X→X intervalley scattering, leading to a parallel leakage current JX. The magnitude of JX depends on the temperature and occupation of the X subbands, which are populated primarily by the same-stage scattering from the Γ-continuum (Γc) states. At 77 K, JX is small up to very high fields in both QCLs. However, at room temperature the 33% QCL shows a much higher JX than the 45% QCL even at low fields. The reason is that in the 33% QC...

Influence of surface chemical modification on charge transport properties in ultrathin silicon membranes.

Ultrathin silicon-on-insulator, composed of a crystalline sheet of silicon bounded by native oxide and a buried oxide layer, is extremely resistive because of charge trapping at the interfaces between the sheet of silicon and the oxide. After chemical modification of the top surface with hydrofluoric acid (HF), the sheet resistance drops to values below what is expected based on bulk doping alone. We explain this behavior in terms of surface-induced band structure changes combined with the effective isolation from bulk properties created by crystal thinness.

Full-dispersion Monte Carlo simulation of phonon transport in micron-sized graphene nanoribbons

We simulate phonon transport in suspended graphene nanoribbons (GNRs) with real-space edges and experimentally relevant widths and lengths (from submicron to hundreds of microns). The full-dispersion phonon Monte Carlo simulation technique, which we describe in detail, involves a stochastic solution to the phonon Boltzmann transport equation with the relevant scattering mechanisms (edge, three-phonon, isotope, and grain boundary scattering) while accounting for the dispersion of all three acoustic phonon branches, calculated from the fourth-nearest-neighbor dynamical matrix. We accurately reproduce the results of several experimental measurements on pure and isotopically modified samples [S. Chen et al., ACS Nano 5, 321 (2011);S. Chen et al., Nature Mater. 11, 203 (2012); X. Xu et al., Nat. Commun. 5, 3689 (2014)]. We capture the ballistic-to-diffusive crossover in wide GNRs: room-temperature thermal conductivity increases with increasing length up to roughly 100 μm, where it saturates at a value of 5800 W/m K. This finding indicates that most experiments are carried out in the quasiballistic rather than the diffusive regime, and we calculate the diffusive upper-limit thermal conductivities up to 600 K. Furthermore, we demonstrate that calculations with isotropic dispersions overestimate the GNR thermal conductivity. Zigzag GNRs have higher thermal conductivity than same-size armchair GNRs, in agreement with atomistic calculations.

Terahertz conductivity of doped silicon calculated using the ensemble Monte Carlo/finite-difference time-domain simulation technique

We present terahertz-frequency characterization of doped silicon via a multiphysics numerical technique that couples ensemble Monte Carlo (EMC) simulation of carrier transport and a finite-difference time-domain (FDTD) solver of Maxwell’s curl equations. We elucidate the importance of rigorous enforcement of Gauss’s law, in order to avoid unphysical charge buildup and enhance solver accuracy. The calculated complex conductivity of doped bulk silicon shows excellent agreement with available experimental data. This comprehensive microscopic simulator is a valuable predictive tool in the terahertz frequency range, where experimental data are scarce and the Drude model inadequate.