ALBERT

Notes: Two photoluminescent defects associated with Er+3-doped Si are (i) a high-temperature defect (which appears after annealing at ∼900 °C and produces five photoluminescence lines), and (ii) a low-temperature defect (which is created at lower annealing temperatures in addition to the high-temperature defect and yields four photoluminescence lines). We conclude that both defects are at Er sites of tetrahedral symmetry, one substitutional and the other interstitial, and that the crystal-field split ground-state levels of Er+3 at either site are of the symmetries (in order of increasing energy) Γ6, Γ8, Γ8, Γ7, and Γ8. We predict that the highest Γ8 level of the low-temperature defect has not yet been resolved, which is why that defect exhibits only four levels. The symmetries of the excited (initial) state levels, in order of increasing energy, are predicted to be Γ8, Γ6, Γ6, Γ8, and Γ7. We speculate that the high-temperature defect's Er is at an interstitial site, while the low-temperature defect's Er is substitutional. © 1997 American Institute of Physics.

Notes: Localized interface states, viz., generalized Tamm states, can be induced by an interfacial InSb bond in the InAs/AlSb heterojunction system, as confirmed by electronic structure calculations. The calculated energies of the interfacial Tamm states, however, are too low to account for the observed carrier concentrations in InAs quantum wells. Native defects capable of accounting for the observed carrier concentrations are identified, and their electronic structures calculated: AlSb in an AlSb layer is responsible for the semi-insulating character of thin InAs quantum wells and the n-type character of wide wells, and AsAl at an AlAs-like interface is responsible for the high values of electron concentration in the wells. The decrease of electron concentrations with temperature can be attributed to partial freezing of electrons into the shallow levels associated with ionized donors. © 1995 American Institute of Physics.

Notes: A theory of the electronic structures of periodic N1×N2 GaAs/AlxGa1−xAs superlattices grown along the [111] direction is presented. Deep levels associated with s- and p-bonded substitutional impurities in these superlattices are also predicted. It is found that: (i) in contrast with [001] superlattices, [111] superlattices are almost always direct band-gap superlattices. (ii) The [111] superlattices exhibit weaker quantum-well confinement than the corresponding [001] superlattices. (iii) As the thickness, t(GaAs), of each GaAs layer is reduced below a critical value (tc (approximately-equal-to)13 A(ring) or N1(approximately-equal-to)4 for x=0.7iii), common shallow donor impurities such as Si cease donating electrons to the conduction band and instead become deep traps. For [111] superlattices tc is smaller than the corresponding tc for [001] superlattices. The fundamental band gap and the band edges of the superlattice, and hence the ionization energies of deep levels, depend strongly on the layer thickness t(GaAs) but only weakly on t(AlxGa1−xAs). The T2- and A1-derived deep levels (of the bulk point group Td) are split and shifted, respectively, near a GaAs/AlxGa1−xAs interface: the p-like T2 level splits into an a1 ( pσ-like) level and a doubly degenerate e ( pπ -like) level of the point group for any general superlattice site (C3v), whereas the s-like A1 bulk level becomes an a1 (s-like) level of C3v. The order of magnitude of the shifts and splittings of deep levels at a GaAs/AlxGa1−xAs interface is less than 0.1 eV, depends on x, and becomes very small for impurities more than (approximately-equal-to)3 atomic planes away from an interface. These predictions are based on a periodicsuperslab calculation for unit superslabs with total thickness t(GaAs)+t(AlxGa1−xAs) as large as 65.3 A(ring) or N1+N2=20 two-atom-thick layers. The Hamiltonian is a tight-binding model in a hybrid basis that is a generalization of the Vogl model and properly accounts for the nature of interfacial bonds. The deep levels are computed usingthe Hjalmarson et al. theory [Phys. Rev. Lett. 44, 810 (1980)] and the special points method.

Notes: By incorporating a II-VI semiconductor into a strained-layer superlattice, it should be possible to overcome the effects of deep hole traps near the valence-band edge and hence to dope the semiconductor p type in many cases. This idea is illustrated for CdTe/ZnTe superlattices.

Notes: Criteria are developed for selecting a barrier material XY such that Si/XY superlattices should emit light from their Si quantum wells. GaAs is such a material for [001] superlattices. In many such superlattices, substitutional N on a Si site will be a shallow donor, not a deep trap.

Notes: Due to the fact that impurities normally change their doping characters when they undergo shallow to deep transitions or deep-to-false-valence transitions, a single defect, such as a cation on an Sb site, can explain all of the following facts for nonintentionally doped AlxGa1−xSb alloys and InAs/AlxGa1−xSb superlattices and quantum-well structures: (i) Bulk GaSb is p type; (ii) bulk AlSb is semi-insulating; (iii) InAs/AlSb superlattices with InAs quantum wells thicker than a critical thickness dc(x=1.0) are n type, where the InAs shallow–deep critical thickness function dc(x) is around (approximately-equal-to)100–(approximately-equal-to)150 A(ring) for 0.5〈x≤1.0 for InAs/AlxGa1−xSb superlattices; (iv) InAs/AlSb superlattices with InAs quantum wells thinner than dc(x=1.0) are semi-insulating. In addition, the theory predicts that Al0.5Ga0.5Sb and AlSb will be semi-insulating when nonintentionally doped, but can be converted to p type by the application of hydrostatic pressure P: P(approximately-greater-than)90 kbar and P(approximately-greater-than)150 kbar, respectively. These changes of doping character, which lie outside the conventional effective-mass theory, occur often in type-II band-alignment systems, such as InAs/AlxGa1−xSb.

Notes: It is predicted that thin quantum-well superlattices or spike superlattices of GaAs in ZnSe will produce band gaps in the yellow-green, and that (GaAs)1−x(ZnSe)x spikes will lead to green and blue-green gaps. These thin quantum-well structures should have better doping properties than ZnSe for x〈0.6.