Showing papers by "Alex Zunger published in 1987"

Electronic structure of ZnS, ZnSe, ZnTe, and their pseudobinary alloys

Abstract: Using the all-electron mixed-basis approach to the density-functional formalism for crystals, we calculate from first principles the electronic structure of zinc-blende ZnS, ZnSe, and ZnTe as well as that of their ordered pseudobinary alloys ${\mathrm{Zn}}_{2}$SSe, ${\mathrm{Zn}}_{2}$SeTe, and ${\mathrm{Zn}}_{2}$STe. For the latter we use as a model a CuAu I-like structure (space group $P\overline{4}m2$), and analyze the observed optical bowing in terms of three contributions: (i) a volume deformation of the band structure due to the replacement of the lattice constants of the binary constituents by that of the alloy, (ii) a chemical-electronegativity contribution due to charge exchange in the alloy relative to its constituent binary subsystems, and (iii) a structural contribution due to the relaxation of the anion-cation bond lengths in the alloy. The total bowing effect [the sum of (i)-(iii) above] agrees well with observations, yet the present analysis suggests a physical mechanism for optical bowing which differs profoundly from that offered by the popular virtual-crystal approach. The maximum contribution of disorder to the optical bowing is calculated for $\mathrm{Zn}{\mathrm{S}}_{x}{\mathrm{Te}}_{1\ensuremath{-}x}$ using a cluster-averaging method, resulting in a reduction in the bowing of the fundamental gap. We further discuss the band structures, x-ray scattering factors, charge distribution, and deformation potentials of the binary zinc chalcogenides and their ordered alloys.

Total-energy and band-structure calculations for the semimagnetic Cd 1 − x Mn x Te semiconductor alloy and its binary constituents

Abstract: Spin-polarized, self-consistent local-spin density total-energy and band-structure calculations have been performed for CdTe, antiferromagnetic (AF) MnTe in its NiAs structure, ferromagnetic (F) ${\mathrm{CdMnTe}}_{2}$, and the hypothetical zinc-blende phase of MnTe in the F and AF spin arrangements. We find the following: (i) The alloy environment stabilizes a zinc-blende form of MnTe, hitherto unknown to exist in the phase diagram of pure MnTe. Its calculated Mn---Te bond length (2.70\ifmmode\pm\else\textpm\fi{}0.02 A\r{}) is very close to that observed in the alloy (2.73 A\r{}), but is substantially different from the Mn---Te bond length in pure (NiAs-type) MnTe (2.92 A\r{}). (ii) AF zinc-blende MnTe is more stable than F zinc-blende MnTe due to a reduced p-d repulsion in the upper valence states. (iii) F ${\mathrm{Cd}}_{1\mathrm{\ensuremath{-}}\mathrm{x}}$${\mathrm{Mn}}_{\mathrm{x}}$Te is more stable than its zinc-blende constituents CdTe + F MnTe, hence, once formed, this ordered alloy will not disproportionate. (iv) Nevertheless, AF ${\mathrm{CdMnTe}}_{2}$ is more stable than its ferromagnetic counterpart, but it is unstable relative to its constituents CdTe + AF MnTe. Hence, if F ${\mathrm{CdMnTe}}_{2}$ converts into AF ${\mathrm{CdMnTe}}_{2}$, the latter will disproportionate into antiferromagnetic domains of MnTe. (v) The band structure of F zinc-blende MnTe and F ${\mathrm{CdMnTe}}_{2}$ predicts a novel type of negative (p-d) exchange splitting, whose origins are discussed in terms of a p-d repulsion mechanism. (vi) The calculated electronic states of ${\mathrm{Cd}}_{1\mathrm{\ensuremath{-}}\mathrm{x}}$${\mathrm{Mn}}_{\mathrm{x}}$Te show a vanishing optical bowing, a Mn ${d}_{\ensuremath{\uparrow}}$ band at ${E}_{v}$-2.5 eV and explains the observed optical transitions. (vii) The fact that ${\mathrm{Cd}}_{1\mathrm{\ensuremath{-}}\mathrm{x}}$${\mathrm{Mn}}_{\mathrm{x}}$Te does exhibit localized multiplet transitions but NiAs-type MnTe does not, is explained in terms of the coexistence of covalency and low symmetry in the latter case.We discuss the electronic structures, local magnetic moments, exchange interaction coefficients, and the general features of the chemical bonds in the semimagnetic semiconductor.

Role of d Orbitals in Valence-Band Offsets of Common-Anion Semiconductors

Abstract: We show through all-electron first-principles electronic structure calculations of core levels that, contrary to previous expectations, the valence-band offsets in the common-anion semiconductors AlAs-GaAs and CdTe-HgTe are decided primarily by intrinsic bulk effects and that interface charge transfer has but a small effect on these quantities. The failure of previous models is shown to result primarily from their decision to omit cation d orbitals.

Ternary and multinary compounds

Abstract: These proceedings contain 81 papers grouped under the headings of Ternary systems; Device applications of ternary semiconductors; A/sup I/B/sup III/C/sub 2//sup VI/ chalcopyrite compounds; Growth and synthesis of ternary compounds; Ternary A/sup II/B/sub 2//sup III/C/sub 4//sup VI/ compounds; Spinels ordered vacancy compounds, etc; Long range ordering in pseudobinary alloys; Ternary phases; More ordering and alloys; Defect chemistry in ternaries; Magnetic ternary compounds; Non-adamantine and other ternary compounds

New optical transitions in strained Si-Ge superlattices.

Abstract: First-principles total-energy and electronic-structure calculations for Si„Ge„superlattices grown epitaxially on an (001) Si substrate reveal a nearly direct band gap despite the pronounced indirectness of its constituents. Whereas the lowest conduction-band wave function extends over both Si and Ge sublattices, the (higher energy) lowest direct state shows strong confinement to Si states. We predict that a substrate with a larger lattice constant than Si will produce a nearly direct band-gap superlattice (indirect only by 0.01 eV).

First-principles calculations of the phase diagrams of noble metals: Cu-Au, Cu-Ag, and Ag-Au.

Abstract: It is shown how the temperature-composition phase diagrams and thermodynamic properties of noble-metal alloys can be accurately reproduced by solving the three-dimensional nearest-neighbor fcc Ising model with volume-dependent interaction energies determined from the properties of the ordered phases alone. It is found that lattice relaxation effects are essential in determining order-disorder critical temperatures. This approach enables the understanding of phase diagrams in terms of the electronic properties and atomic-scale structure of the constituent ordered phases.

First-principles calculation of semiconductor-alloy phase diagrams.

Abstract: Combining first-principles self-consistent local-density total-energy calculations with the cluster variation method, we calculate the phase diagram of a semiconductor alloy. It is demonstrated that inclusion of both elastic and chemical interactions in the total-energy functional leads to new features, including the appearance in the same phase diagram of ordering and phase separation, and strain stabilization of both stable and metastable ordered phases.

Order-disorder transformation in ternary tetrahedral semiconductors

Abstract: The recently discovered order‐disorder transformations in pseudobinary semiconductor alloys AxB1−xC are shown to belong to a broader class of such transformations in AnB4−nC4 semiconducting compounds (e.g., chalcopyrites, for n=2). Strain energy, set up by the atomic size mismatch between the A–C and B–C bonds, is shown to control the nature of the state of order in chalcopyrites and pseudobinary alloys alike. These considerations lead to a classification of all bulk tetrahedral semiconductors into four classes of order‐disorder characteristics.

Thermodynamic instability of ultrathin semiconductor superlattices: The (001) (GaAs)1(AlAs)1 structure.

Abstract: First-principles total-energy pseudopotential and all-electron calculations predict (001) (GaAs)i(A1As)1 and (CdTe)~(HgTe)~ superlattices to be intrinsically unstable towards disproportionation into compounds. This instability is traced to unfavorable charge redistribution in the system. Many disordered' or artificially ordered semiconductor systems are manifestly metastable in temperature and composition ranges in which they are usually characterized and utilized. Such are disordered GaAs Sb~ alloys (grown in the range of thermodynamic immiscibility of GaAs and GaSb), and ordered (A'"B )t (C2 )„alloys (judged by their equilibrium phase diagrams to spinodally decompose). Metastable systems come to exist through kinetic rather than thermodynamic control, e.g. , by nonequilibrium growth techniques. ' They owe their thermal stability' to large reorientation activation barriers, small thermodynamic driving forces, ' and exceedingly low diAusion coefficients at laboratory temperatures. ' Despite extensive study, it is as yet unclear whether artificial semiconductor superlattices (AC) (BC)„are thermodynamically (intrinsically) stable or metastable. Current understanding can be characterized as follows. Disordered (D) isovalent alloys A„Bt-„Care known to have positive enthalpies of mixing hH (I), so that at a sufticiently low temperature T, the negative entropy term — T,BS is overwhelmed by the positive dH, leading eventually to disproportionation. Most contemporary theoretical models analyze this instability of A„Bj C alloys via models that do not distinguish them from ordered compounds A~B„C~+„ofthe same composition. Such are, e.g. , elastic models which attribute dH & 0 to the destabilizing role of microscopic strain associated with a mismatch h, a between lattice constants of AC and BC. Since thin superlattices (AC)~(BC)„are most naturally regarded as ordered compounds — e.g. , an m =n =1 superlattice in the (001) orientation is crystallographically identical to an ABC2 compound with the simple tegragonal p4m2 space groups 6 (having a CuAu-I-like A-B sublattice) — these models would judge both alloys and superlattices (having nearly the same ha) intrinsically unstable at low temperatures. However, Srivastava, Martins, and Zunger demonstrated that hH ) 0 does not require ordered (0) phases to be unstable too because (i) a chemical energy term, neglected by other models, may render AH negative, and (ii) coherently ordered arrangements of bonds can reduce strain imposed by bond-length mismatch better than do disordered arrangements. Perhaps the best-studied superlattice — (GaAs) (AIAs)„— exhibits, however, a delicate energy balance: It has a nearly vanishing h, a =R&1A, — R~,A, =0.0009 A at growth temperatures — 800 K and consequently a nearly vanishing hH (hence, ordering offers but a small reduction in strain), yet Al differs (slightly) from Ga in electronegativity (hence, charge transfer may stabilize the system). This delicacy is highlighted by the disparate views on stability of the (GaAs)1(AIAs) t superlattice. Kuan et al. , having observed ordered GaAlAs2 even in spontaneous growth, characterized it as the thermodynamic equilibrium state of Ga Al~ As, as did Petroff for the layer-by-layer-grown superlattice. On the other hand, Phillips' suggested that this phase was intrinsically unstable but stabilized via pinning by oxygen impurities, and Ourmazd and Bean" suggested that it was stable only because of extrinsic substrate strain eAects. Theoretical estimates for the formation enthalpy of the ordered (0) superlattice similarly range (referring all energies to a primitive cell of four atoms) between stability (AHo = — 1.5 meV, obtained from empirical tight binding i2 or AHO — 20 meV, from a cluster calculation after optimization of bond lengths' ) and instability (AHo=+9. 2 meV in a recent calculation using relativistic pseudopotentials ' ).

Effect of chemical and elastic interactions on the phase diagrams of isostructural solids.

Abstract: It is shown how the introduction of volume-dependent elastic interactions into lattice models of order-disorder transformations, in addition to the familiar constant-volume interactions (analogous to Ising ''spin energies''), leads to qualitatively new features in a binary A/sub x/B/sub 1-x/ phase diagram.

Electronic structure of M3ISb-type filled tetrahedral semiconductors.

Abstract: First-principles band-structure and total-energy calculations have been performed for Li/sub 3/Sb, K/sub 3/Sb, and Cs/sub 3/Sb in the cubic D0/sub 3/ structure. The structure of these M/sub 3//sup I/A/sup V/ octet semiconductors consists of a face-centered-cubic lattice A/sup V/ in which three metal M/sup I/ atoms occupy the interstitial sites, thereby completing the octet shell. The equilibrium lattice parameters and bulk moduli calculated from the total energy agree well with available experimental data. The semiconducting character of K/sub 3/Sb and Cs/sub 3/Sb is attributed to p-d repulsion between the anion p states and the unoccupied metal d states. The calculated charge density shows weak covalent bonding between like atoms and ionic bonding between unlike atoms. The covalency decreases and the ionicity increases with increasing metal atomic number, as the M/sup +/ cores dilate the lattice. We find that the electronic structure of these compounds can be understood qualitatively by considering a skeleton of Sb/sup 3-/ anions whose separation is determined by the otherwise nearly inert alkali ions M/sup +/. The trends in the electronic properties in the series M = Li, K, and Cs then reflect the perturbations exerted by different M/sup +/ ions through (I) charge transfer, (II) p-dmore » hybridization, and (III) relativistic effects.« less

Calculation of the valence band offsets of common-anion semiconductor heterojunctions from core levels: The role of cation d orbitals

Abstract: The valence band offsets of the common‐anion CdTe–HgTe, CdTe–ZnTe, ZnTe–HgTe, and GaAs–AlAs semiconductor pairs are calculated from the core level energies. The good agreement obtained with experiment for lattice‐matched systems and a simple electrostatic model analysis suggest interface dipoles to have only a small effect. Furthermore, the microscopic origin of the failure of the common‐anion rule in lattice‐matched systems is identified: it is found that participation of cation d orbitals (neglected by tight‐binding and pseudopotential approaches alike) in the valence band maxima is responsible for much of the band offset in these systems.

Transition Metal Impurities in Semiconductors

Abstract: The study of impurities and defects in semiconductors is of fundamental interest and is important for technological applications. This monograph is a first attempt to generalise experimental data and theoretical interpretation about the nature and behaviour of impurity atoms of transition metals in semiconductors. The nature of impurities and changes in their electronic structure are analysed. The molecualr orbital approach is followed extensively in the theoretical interpretation, with particular emphasis on crystal field splitting, electron paramagnetic resonance and optical absorption spectoscopies. Coverage of experimental data is extensive with more the 300 references to the literature. This is a translation of a Russian text published in 1983. The authors have updated the content for the English language edition. This book will be of interest to scientists and engineers in solid state physics and chemistry, materials science and electronic engineering. It should also be useful for postgraduate students in these fields.

Common-anion rule and its limits: Photoemission studies of CuInxGa1-xSe2-Ge and CuxAg1-xInSe2-Ge interfaces.

Abstract: Synchrotron-radiation photoemission data show that the valence-band discontinuities of CuIn x Ga1 − x Se2-Ge and Cu x Ag1 − x InSe2-Ge interfaces are independent of x within the experimental accuracy of 0.1 eV. We argue that this result is consistent with the Wei-Zunger explanation of the breakdown of the common-anion rule, which is based on a substantial role of the cation d states in determining the valence-band-edge position.

New Ordering-Induced Optical Transitions in Strained Sige Superlattices

Abstract: First principles total energy and electronic structure calculations for SinGen superlattices grown epitaxially on a Si (001) substrate reveal a nearly direct band gap despite the pronounced indirectness of the Si0.5Ge0.5alloy. While the new direct superlattice transitions have their origin in folded indirect transitions of the random alloy, their transition matrix elements are considerably enhanced by superlattice atomic relaxation and the superlattice ordering potential. For a superlattice grown on a substrate with an lager lattice contant than we predict a nearly direct band gap.

Reply to "Comment on 'Atomic structure and ordering in semiconductor alloys' "

Abstract: The claim by Podg\'orny and Czyz\ifmmode \dot{}\else \.{}\fi{}yk (preceding comment) that their simplified empirical alloy model captures the essence of the first-principles model of the present authors is shown to be unfounded.