Session U42: Nitrides, Carbides, Oxides: Surfaces and Properties

Electronic structure of nitride surfaces

Knowledge of surface reconstructions and the corresponding surface electronic structure is important to control growth, since Fermi-level pinning can affect defect creation and incorporation. In addition, surface states can play an important role in devices, for instance in high-electron mobility transistors where the surface acts as a source of electrons for the channel. In the case of InN a very high, and thus far unexplained, electron accumulation has been observed on all polar surfaces. We have addressed these issues by performing a systematic computational study of reconstructed GaN and InN surfaces in various orientations, including (11-20) ($a$ plane) and (10-10) ($m$ plane), as well as the polar (0001) (+$c)$ and (000-1) (-$c)$ planes. The calculations are based on density-functional theory, combined with an extensively tested approach for correcting the band-gap error through use of modified pseudopotentials. For GaN we identify the microscopic origins of the experimentally observed Fermi-level pinning. For InN we find that on polar surfaces occupied surface states occur above the conduction-band minimum, thus explaining the observed electron accumulation. We predict an absence of electron accumulation on \textit{nonpolar} surfaces grown at moderate In/N ratios.   

Electronic structures and work functions of InN(0001) films

InN films have attracted great attention with its recently discovered band gap, 0.7eV, and evidence for p-type doping\footnote{R. E. Jones, et al, Phys. Rev. Lett, {\bf 96}, 125505 (2006)}. We have studied theoretically the electronic structures, surfaces, and work functions of InN films using the highly precise FLAPW method\footnote{Wimmer, Krakauer, Weinert, Freeman, Phys. Rev. B, {\bf 24}, 864 (1981)}. The passivation with pseudo-hydrogens\footnote{K. Shiraishi, J. Phys. Soc. Jpn, {\bf 59}, 3455 (1990)} has also been applied to the surfaces of InN(0001) films for comparison with the electronic structure of the ideal InN(0001) films. We compare the work functions of InN films with other wurtzite materials such as ZnO, GaN, and AlN, which we have also calculated. We discuss the mechanism of the structural transition\footnote{C. L. Freeman, et al, Phys. Rev. Lett, {\bf 96}, 066102 (2006)} with layer thickness for the very thin InN(0001) films, for which we have found that the ideal InN(0001) films of the wurtzite structure, up to 4 bilayers, optimize to the graphitic- like structure. We then discuss the relationship between the dipoles and the surfaces (work functions) of the InN(0001) films, and the possibilities of their p-type doping.   

HREELS, AES, and LEED of InN(000-1): Surface structure and electron accumulation

InN layers grown by high pressure chemical vapor deposition (HPCVD) have been studied using several surface sensitive techniques. Following argon sputtering and atomic hydrogen cleaning (AHC), Auger electron spectroscopy showed that surface contaminants had been removed and a 1x1 hexagonal low energy electron diffraction pattern demonstrated that the InN surface was well ordered. HREEL spectra of the atomic hydrogen cleaned layer show a Fuchs-Kliewer surface phonon at 560 cm$^{-1}$ and adsorbate loss peaks at 3260 and 870 cm$^{-1}$ assigned to N-H stretching and bending vibrations, respectively. These assignments are confirmed by isotopic shifts using deuterium. No surface In-H vibrations are observed indicating N-H termination of the surface and the film is N-polar. HREEL spectra also showed a broad loss features due to conduction band plasmon excitations. The plasmon excitation shifted to higher energy as the incident electron energy (and therefore the penetration depth) was decreased. This shift indicates that the surface has a higher plasma frequency than bulk of the InN layer, which in turn indicates the presence of a surface electron accumulation layer.   

In-situ X-ray Studies of MOCVD Growth of InN

One of the fundamental issues in the continued development of III-nitride semiconductor alloys is to understand incorporation of indium. Our approach is to use real-time x-ray scattering and fluorescence as \textit{in situ} probes during growth by MOCVD. We observe the equilibrium condensation boundaries for elemental In and InN as a function of temperature and trimethylindium supply, which allow us to determine the effective activities of In and N at the sample surface. We find that the partial pressures of both hydrogen and ammonia in the ambient have strong effects on the activities. We also observe strong effects of the substrate on condensation, including an oscillatory regime, indicating that surface reactions are important. Work supported by the U.S. Dept. of Energy contract DE-AC02-06CH11357.   

Atomic and Electronic Structures of Oxygen on the $\beta-$Si$_3$N$_4$ $(10\overline{1}0)$ Surface

The desirable mechanical and physical properties of Si$_3$N$_4$ ceramics in high temperature applications are hindered by their intrinsic brittleness. Doping Si$_3$N$_4$ with rare-earth oxides has long been known to overcome this limitation creating a tougher material. Precise information about the microscopic origin of this empirical observation has, however, been lacking for many years. In this study, we present {\em ab initio} calculations for the structural stability of $\beta-$Si$_3$N$_4$ $(10\overline{1}0)$ surfaces in the presence of different oxygen concentrations. Two different $(10 \overline{1}0)$ surface terminations, the ``open ring" and the ``half surface",\footnote{J. C. Idrobo {\em et al}., Phys. Rev. B {\bf 72}, 241301(R) (2005).} are investigated in detail using an asymmetric slab. We find that the Si-O bond plays the most important role in the structural stability and passivation of the surface. The theoretical results are analyzed in connection with recent electron microscopy studies on the interface.\footnote{A. Ziegler {\em et al.}, Science {\bf 306}, 1768 (2004); N. Shibata {\em et al.}, Nature {\bf 428}, 730 (2004); G. B. Winkelman {\em et al.}, Phil. Mag. Lett. {\bf 84}, 755 (2004).}   

Growth and Structure of ZrSiN Thin Films

A series of Zr$_{1-x}$Si$_{x}$N thin films were grown on r-plane sapphire substrates using rf magnetron co-sputtering of Zr and Si targets in a N$_{2}$/Ar plasma. The films were grown at 200$^{o}$C and also post-deposition annealed to 900$^{o}$C in vacuum. Pure ZrN grows with high quality (100) epitaxy on r-sapphire as demonstrated by x-ray diffraction reflectivity and pole figure analysis. When small amounts of Si are added into the lattice, the films become strained as evidenced by a continual increase in the lattice parameter (up to a 6{\%} for x=0.12) and become polycrystalline. Higher amounts of Si cause the structure to become amorphous and the films become much rougher. X-ray photoelectron spectroscopy measurements show large shape changes in the N and Zr core levels as the alloy composition changes, whereas the Si peaks exhibit negligible change. UV-visible optical absorption measurements show a direct correlation between the location of the absorption edge and Zr-Si ratio.   

Passivation of 4H-SiC Silicon surface

The material properties of the silicon carbide (SiC) 4H polytype are ideally suited for use in metal-oxide-semiconductor field-effect transistors (MOSFETs) operating under high temperature, high power conditions. Currently, the development of lateral SiC MOSFETs is hindered by excessively small field-effect mobilities that are typically measured in these devices. The cause for such small mobilities is believed to be directly related to the very large density of traps measured at the 4H-SiC/SiO2 interface. Recently, oxidation processing in the presence of nitrogen or in the presence of metals, has been shown to improve the mobility of 4H-SiC MOSFETs. However, there is no clear consensus on the physical mechanisms involved in improvement of the 4H-SiC/SiO2 interface. We use \textit{/ab-initio/} density functional theory to study passivation of the 4H-SiC Si surface by nitrogen, oxygen, aluminum, and sodium.   

A DFT study of low index polar surfaces: the case of SiC and ZnO

With the advent of nanostructured devices, it appears evident that the role of surface and interface effects may prevail on bulk properties and determine the physical characteristics of the material at the nanoscale: in particular, the understanding of the electronic properties of nanosized structures demands for a proper accurate treatment. Here we report on first principles density functional calculations of the structural and electronic properties of the so called ``non-polar'' low index surfaces of hexagonal silicon carbide (SiC) and zinc oxide (ZnO), specifically focussing on surface polarity. We first provide an accurate analysis of the macroscopic polarization field as a function of the hexagonality and ionicity of the compound, and than describe in details the properties of the relaxed surfaces. Our predictions nicely compare with recent experimental data on similar SiC and ZnO surfaces: in particular we obtain good agreement between the perpendicular surface dipole component and the experimental work functions values. Moreover, for the first time, we highlight the presence of a strong in-plane dipole component related to dangling bond rearrangement at the surface, which opposes the bulk spontaneous polarization. The decaying behaviour of this dipole inside the slab shows that the presence of surfaces deeply change the polar properties of structures of few nanometers size, while bulk polarization is recovered for thicker systems. Given the importance of surface dipoles in adsorption and functionalization processes, we finally analyze the local surface electric field by employing a simple polar model molecule (HF) as a probe: we describe the interaction in terms of the potential energy surface (PES) experienced by the molecule. This allows us to give a complete description of the surface polarity and to draw conclusions on the most likely adsorption sites of charged and polar species.   

Role of neutral impurity scattering in the analysis of Hall data from ZnO

Zinc oxide is a wide band-gap semiconductor with bright UV emission. To determine donor and acceptor concentrations affecting electrical properties in n-type ZnO crystals, the relaxation time approximation has been used to analyze mobility ($\mu )$ and carrier concentration data measured from 80 to 400 K. Five scattering mechanisms are included: polar-optical-phonon, piezoelectric potential, deformation potential, ionized impurity, and neutral impurity (NI) scattering. The NI scattering is often ignored but plays an important role in limiting the total $\mu $. By including NI scattering, the experimental deformation potential E$_{1}$ = 3.8 eV can be used. Temperature dependences of the intrinsic Hall r factor and intrinsic $\mu $ are determined. At 300 K, ``pure'' ZnO has an electron $\mu $ of about 210 cm$^{2}$/Vs. Analysis of Hall data from commercial hydrothermally and CVT-grown n-type ZnO crystals is presented. Donor and acceptor concentrations from Hall data are compared with those estimated using infrared absorption and EPR data. Intrinsic hole mobility in p-type ZnO is also discussed. This work was supported by NSF Grant No. DMR-0508140.   

Direct observation of zinc vacancies and oxygen vacancies in an electron-irradiated ZnO crystal

Electron paramagnetic resonance (EPR) has been used to monitor zinc vacancies and oxygen vacancies in a ZnO crystal irradiated near room temperature with 1.5 MeV electrons. Out-of-phase detection at 30 K greatly enhances the EPR signals from these vacancies. After electron irradiation, but before illumination, Fe$^{3+}$ ions and nonaxial singly ionized zinc vacancies are observed. Illumination with 325 nm light at low temperature produces spectra from singly ionized oxygen vacancies, neutral zinc vacancies, and axial zinc vacancies. The light also produces spectra from zinc vacancies with an adjacent hydrogen (an OH$^{-}$ ion). The response of the irradiated crystal to illumination wavelengths out to 750 nm is described. Wavelengths shorter than 600 nm convert the Fe$^{3+}$ ions to Fe$^{2+}$ ions and convert the neutral oxygen vacancies to singly ionized oxygen vacancies. Warming above 130 K in the dark reverses the effect of the illuminations. This work was supported by NSF Grant DMR-0508140. One of the authors (SME) acknowledges support from the West Virginia STEM Fellowship Program.   

Tunable UV-Luminescent MgZnO Nanoalloys

Mg(x)Zn(1-x)O alloys are promising wide-bandgap semiconductors for optoelectronic applications, and also of considerable interest from a fundamental viewpoint. The environmentally friendly chemical composition and the deep excitonic level $\sim $ 60 and 90 meV of ZnO and MgO respectively make it an excellent candidate for high-efficiency next generation ultraviolet light sources. These optical alloys may enable the tuning of the bandgap and the luminescence at the range of $\sim $ 3.0 for ZnO of the wurtzite structure up to $\sim $ 7 eV for the MgO of the rocksalt structure. We will present studies on the photoluminescence and Raman properties of Mg(x)Zn(1-x)O nanocrystallites of average size $\sim $ 30 nm that were synthesized via the thermal decomposition method. For the studied composition range of 0-26{\%} Mg, the room temperature UV-PL was found to be tuned by $\sim $ 0.3 eV towards the UV-spectral range. For that composition range the first-order LO Raman mode was found to exhibit a significant blueshift of $\sim $ 33 cm$^{-1}$ indicating that a good solid solution was achieved at the nanoscale. At higher composition ranges a PL blue shift of at least 1.3 eV was achieved. Issues such as excitonic emissions, alloy spectral broadening, and phonon symmetry will be presented   

Stability, Electronic and Optical Properties of In$_2$O$_3$(ZnO)$_n$ Alloys

(In$_2$O$_3$)-(ZnO) heterostructural alloy has recently become a promising transparent conducting oxides (TCO) because it possesses combined physical properties such as excellent optical transmission, high electrical conductivity, chemical and thermal stability, and good film smoothness. However, the origin of these superior properties is not well understood. In this work, using first-principles methods, we have investigated the structural stability, electronic, and optical properties of In$_2$O$_3$(ZnO)$_n$ $(n = 1 - 5)$. In$_2$O$_3 $(ZnO)$_n$ forms layered hexagonal or rhombohedral In$_2$O$_3$(ZnO)$_n$ superlattices, which are isostructural with LuFeO$_3$(ZnO)$_n$. The deformed In$_2$O$_3$ layer is highly strained and plays an important role in determining the stability of the system. The calculated band structures show that these alloys has a direct band gap at the $\Gamma$-point and typical features as other known n-type TCOs. The calculated small effective mass for these materials is consistent with the high electron mobility for these system. The optical properties of these alloys are calculated and compared with that of In$_2$O$_3$ and ZnO.   

Carrier Dynamics and Emission Efficiency in Sulfur-doped ZnO Powders

In previous work [\textit{Nano Lett. }\textbf{6}, 1126 (2006)] it was found that sulfur-doping ZnO micro- and nanostructures dramatically enhanced the broadband, visible wavelength defect emission centered at $\sim $2.5 eV (500 nm), while quenching the band edge ultraviolet emission. The effects of sulfur-doping on carrier dynamics and integrated emission efficiency are further characterized here by studying the time-resolved photoluminescence of band edge and defect emitters as a function of sulfur doping concentration, temperature, and excitation intensity in ZnO powders. The dynamics can be understood in terms of a rate equation model which describes energy transfer between band edge and radiative defect levels, as well as nonradiative centers. The potential application of these materials for efficient visible wavelength phosphors will be discussed.   

Origin of high-density two-dimensional electron gas in ZnO/ZnMgO heterostructures.

We performed a self-consistent calculation of electronic states in ZnO/ZnMgO multiple quantum wells (MQWs). In ZnO/ZnMgO MQWs, the charges induced by spontaneous and/or piezoelectric polarizations at the heterointerfaces play an important role in determining the optical properties. By comparing the optical transition energies between the calculations and experiments, the polarization charge density was determined. In addition, the band bending effects caused by ionized impurities in the structure were found to be crucial for wider ZnO well thicknesses. Therefore, the electronic states in ZnO/ZnMgO MQWs was calculated by changing the thickness of the ZnO layer, Lw (1-8 nm), the sheet polarization charge,$\sigma $, and donor concentrations in the ZnO, Nw, and in the ZnMgO, Nb. We also calculated the two-dimensional electron gas (2DEG) concentration in a thick ZnO layer grown on a ZnO (5 nm)/ZnMgO (5 nm) MQW buffer layer by using the same parameters in order to validate the calculation. The 2DEG concentration was successfully explained by the calculation. The calculation indicates that the 2DEG concentration is mainly determined by the donor concentration in the ZnMgO barrier layer.   

Anisotropic thermal expansion in wurtzite materials: an ab-initio calculation for ZnO

Many optoelectronic devices are based on wurtzite materials, e.g., ZnO, GaN, and SiC. Contrary to cubic structures, scarce experimental data have been reported on the linear thermal expansion coefficients of anisotropic materials. To our knowledge, no first principles calculations have been reported for the anisotropic thermal expansion coefficients. We report here two different approaches for first principles calculations of these coefficients based on the lattice dynamics obtained in the quasiharmonic approximation from the \textit{ab initio} electronic band structure. The first method relies on thermodynamic relations for the entropy and the phonon density of states. The second approach requires the explicit calculation of Gr\"{u}neisen parameters in the irreducible Brillouin zone. The two methods are applied to wurtzite ZnO and the obtained expansion coefficients are in excellent agreement with those derived from x-ray diffraction data taken with synchrotron radiation at the beamline ID31 of the ESRF. The calculations also provide the so-called zero-point contribution to the lattice parameters, which is also anisotropic and of interest in the analysis of the temperature dependence of electronic gaps.