A 1990 study of molecular-beam-epitaxy-grown Te-doped AlSb, combining deep level transient spectroscopy (DLTS) on InAs/AlSb quasi-Schottky heterojunctions with temperature-dependent Hall measurements, found a large density of deep electron traps (2.0×10^17 cm^-3 at a Te doping of 3.1×10^17 cm^-3), a best-defined level with 0.26 eV thermal emission energy, a Hall ionization energy near 100 meV interpreted as the Te DX-center energy, and inferred that the deep-center energy varies with composition across 0.4 < x ≤ 1.0 in AlxGa1−xSb, in contrast to the near-constant value in AlxGa1−xAs. The authors left open the resolution of additional low-temperature DLTS structures, the refinement of the level and apparent-capture-cross-section parameters for the binary, the experimental mapping of the Al-rich composition range, and the microscopic identification of the traps. Later work mapped the Te DX center for x ≤ 0.5, confirmed DX behavior near the binary endpoint in Te-doped AlAs0.07Sb0.93, traced the deep donors of undoped MBE AlSb to oxygen, and computed first-principles formation energies for chalcogen donors in AlSb while explicitly leaving their localized (DX) versus delocalized ground-state question unsettled. The problem asks for a microscopically identified deep-level map of the Te-doped AlSb / Al-rich AlGaSb system: the resolved trap spectrum of binary Te-doped MBE AlSb, the experimentally established deep-center energy across 0.5 < x ≤ 1.0, and a microscopic account of why such a large fraction of Te donors forms deep traps in AlSb relative to n-AlAs.

Background

Aluminium antimonide (AlSb) is a zinc-blende structure, indirect-gap semiconductor (room-temperature band gap about 1.6 eV) whose conduction-band minimum lies near the X point of the Brillouin zone rather than at the zone center. It serves as the wide-gap barrier material of InAs/AlSb quantum-well transistors and resonant tunnelling devices, and more recently of AlSb/AlGaAsSb barrier infrared detectors. Molecular beam epitaxy (MBE) is the standard vacuum-evaporation growth technique for these heterostructures. n-type doping of AlSb is difficult: silicon turns out to act as an acceptor in this material, so group-VI donors such as tellurium (Te), which substitute for Sb atoms, are the natural choice.

Deep levels are electronic trap states lying far from the band edges; in n-type material they compensate free electrons and act as slow carrier traps. Two experimental probes dominate this field. Deep level transient spectroscopy (DLTS) measures the capacitance transient of a reverse-biased junction as trapped electrons are thermally re-emitted, yielding per trap a thermal emission activation energy, an extrapolated apparent capture cross section, and a concentration. Temperature-dependent Hall effect measures the free-electron density versus temperature, from which donor ionization energies can be fitted; such fits require assumptions about the conduction-band structure (effective masses, valley multiplicity and spin degeneracy, and the ordering of the zone-center Γ and zone-boundary L and X conduction minima, which changes with alloy composition across the AlxGa1−xSb series). Because good Schottky barriers on AlSb are hard to make, the source study used InAs/AlSb n–N isotype heterojunctions (both sides n-type, the narrow-gap side written lowercase n and the wide-gap side uppercase N) as quasi-Schottky barriers for the DLTS measurement.

The archetype of a donor-related deep level is the DX center, first identified in Te- and Si-doped AlxGa1−xAs: a donor that, instead of remaining a hydrogen-like shallow state, relaxes into a strongly localized configuration with large lattice relaxation, characterized by a Franck–Condon shift (the difference between the optical and thermal ionization energies), thermally activated carrier capture, persistent photoconductivity (a metastable, light-induced conductivity increase that decays only on warming), and negative-U behaviour (the defect is most stable in a doubly occupied, charged state). In AlxGa1−xAs the DX thermal emission energy is nearly independent of composition; whether the antimonide analogues behave the same became a natural question when antimonide epitaxy matured around 1990. Before the source paper, Takeda et al. (Jpn. J. Appl. Phys. 26, L273 (1987)) had measured DLTS deep levels of 0.46–0.48 eV in AlxGa1−xSb with x ≤ 0.5.

The source paper (Nakagawa et al., Appl. Phys. Lett. 57, 1551 (1990)) then examined the x = 1 endpoint, MBE-grown Te-doped AlSb, and found: several electron traps between 90 and 300 K, not all well resolved, the best-defined one with a 0.26 eV thermal emission energy, much shallower than Takeda's values at lower x; a Hall ionization energy of 100 meV, interpreted as the DX-center energy of Te in AlSb; and a trap concentration of 2.0×10^17 cm^-3 at a Te doping of 3.1×10^17 cm^-3 — roughly two thirds of the donors are deep traps, far more than in comparably doped n-AlAs. The authors inferred a compositional dependence of the deep-center energy across 0.4 < x ≤ 1.0, in contrast to AlxGa1−xAs, and identified as unfinished business: the separation and characterization of the additional low-temperature DLTS structures, the refinement of the Hall-derived level and of the apparent capture cross section (whose extraction depends on the assumed conduction-band parameters), the systematic measurement of the deep-center energy toward the binary endpoint, and the microscopic identification of the trap species.

Subsequent work clarified much of the low-aluminium half of this picture. The Parma group established the Te DX center in MBE AlxGa1−xSb: the DX occupancy level (the Fermi-level position at which the center changes charge state) enters the band gap at x ≈ 0.25 and deepens with increasing x up to 0.5, tracking the X edge for x ≥ 0.30 and lying resonant with, and close to, the L edge for x ≤ 0.20; photoionization measurements gave optical ionization energies of 0.84–0.95 eV and Franck–Condon shifts of 0.70–0.74 eV, confirming large lattice relaxation; and a shallow Te state tied to the L minima coexists with the DX state. Near the binary endpoint, the closest study is on Te-doped AlAs0.07Sb0.93, where improved isothermal capacitance transient spectroscopy (a DLTS variant that records full transients versus time in a single temperature scan) found a single DX-center trap with an emission-level energy of 278 meV, a capture barrier (the activation energy impeding capture of a free electron) of 137 meV and a high-temperature capture cross section of 1.3×10^-12 cm^2. For undoped MBE AlSb, deep donors were traced to oxygen incorporated mainly from the Sb source; for bulk Se-doped AlSb, electron paramagnetic resonance gave evidence for a negative-U DX center; and first-principles calculations find S, Se and Te on Sb sites to be deep donors while explicitly acknowledging that local-density functionals cannot settle the localized (DX) versus delocalized ground state of these chalcogen donors. What no study has provided is a resolved trap inventory for binary Te-doped MBE AlSb, any experimental deep-center data for AlxGa1−xSb between x = 0.5 and 1, or a microscopic account of why so large a fraction of Te donors in MBE AlSb becomes deep traps.

Problem Statement

Establish the deep-level systematics of the Te donor and of the associated defects in MBE-grown AlSb and Al-rich AlxGa1−xSb, as left open by the 1990 source study. This comprises three linked facets of one objective.

(1) Resolve the complete electron-trap spectrum of Te-doped MBE AlSb. In the source experiment several traps appeared between 90 and 300 K, of which only the best-defined (0.26 eV thermal emission energy) could be characterized; the additional low-temperature DLTS structures remain to be separated, with their emission energies, apparent capture cross sections and concentrations extracted, and each trap assigned to a microscopic species (Te DX center, oxygen-related donor, or native defect such as an antisite).

(2) Determine experimentally the deep-center energy of Te across the Al-rich half of the alloy system, 0.5 < x ≤ 1.0, connecting the mapped low-x regime (DX occupancy level deepening with x up to 0.5) with the binary endpoint (Hall ionization energy ≈ 100 meV, DLTS emission energy 0.26 eV). The deep-center energy here is the position of the Te DX level — its thermal ionization energy relative to the relevant conduction-band edge — and the composition dependence of this single quantity is the target; the Hall ionization energy and the DLTS thermal emission energy are distinct observables through which it is to be established, and they must be reported and reconciled rather than confused with one another. The source paper could only suggest a composition-dependent emission energy across 0.4 < x ≤ 1.0 — in contrast to the near-constant value in AlxGa1−xAs — from these two endpoints; the Parma group's DX occupancy-level studies (Baraldi et al. 1999, samples with 0.25 ≤ x ≤ 0.50; Ghezzi et al. 2002, with L- and X-edge references stated) have since mapped the deep-center energy up to x = 0.5, leaving 0.5 < x ≤ 1.0 as the unmapped portion of the source's range.

(3) Identify microscopically the traps responsible for the large deep-trap fraction in Te-doped MBE AlSb (trap concentration 2.0×10^17 cm^-3 at a Te doping of 3.1×10^17 cm^-3 in the source sample, i.e. a deep-trap-to-donor ratio far larger than in comparably doped n-AlAs) — which traps are Te-related DX centers, which are oxygen-related or native defects — and explain that contrast with n-AlAs.

These facets jointly constitute one problem: a microscopically identified deep-level map of the Te-doped AlSb / Al-rich-AlGaSb material system.

References

  1. A. Nakagawa, J. J. Pekarik, H. Kroemer, and J. H. English, 'Deep levels in Te-doped AlSb grown by molecular beam epitaxy', Applied Physics Letters 57, 1551–1553 (1990). DOI: 10.1063/1.103350, https://doi.org/10.1063/1.103350
  2. Y. Takeda, X. C. Gong, Y. Zhu, and A. Sasaki, 'Determination of Al Composition and DLTS Measurements of AlxGa1−xSb on GaSb Substrate', Japanese Journal of Applied Physics 26, L273 (1987). DOI: 10.1143/JJAP.26.L273, https://doi.org/10.1143/JJAP.26.L273
  3. A. Baraldi, F. Colonna, G. Covucci, C. Ghezzi, R. Magnanini, A. Parisini, L. Tarricone, A. Bosacchi, and S. Franchi, 'Control of the n-type doping in AlxGa1−xSb: DX-center behavior of the Te impurity', Materials Science and Engineering B 44, 70–73 (1997). DOI: 10.1016/S0921-5107(96)01811-9, https://doi.org/10.1016/S0921-5107(96)01811-9
  4. A. Baraldi, C. Ghezzi, R. Magnanini, A. Parisini, L. Tarricone, and S. Zerbini, 'Photoionization cross-section of the DX center in Te-doped AlxGa1−xSb', Journal of Applied Physics 83, 491–496 (1998). DOI: 10.1063/1.366720, https://doi.org/10.1063/1.366720
  5. A. Baraldi, C. Ghezzi, A. Parisini, R. Magnanini, L. Tarricone, and S. Franchi, 'Occupancy level of the DX center in Te-doped AlxGa1−xSb', Journal of Applied Physics 85, 256–263 (1999). DOI: 10.1063/1.369438, https://doi.org/10.1063/1.369438
  6. C. Ghezzi, R. Magnanini, A. Parisini, S. Franchi, E. Gombia, and R. Mosca, 'Determination of the occupancy level of the DX center resonant with the conduction band in Te-doped AlGaSb', Journal of Applied Physics 91, 3015 (2002). DOI: 10.1063/1.1445495, https://doi.org/10.1063/1.1445495
  7. C. Ghezzi, R. Magnanini, A. Parisini, E. Gombia, and R. Mosca, 'Effects of a resonant shallow state on the electrical properties of Te-doped AlxGa1−xSb', Journal of Applied Physics 93, 9743–9748 (2003). DOI: 10.1063/1.1574671, https://doi.org/10.1063/1.1574671
  8. D. K. Johnstone, Y. K. Yeo, R. L. Hengehold, and G. W. Turner, 'Deep level capture barrier in molecular beam epitaxial grown AlAsySb1−y measured by isothermal capacitance transient spectroscopy', Applied Physics Letters 71, 506–508 (1997). DOI: 10.1063/1.119592, https://doi.org/10.1063/1.119592
  9. A. Furukawa and S. Ideshita, 'Origin of deep donors in AlSb grown by molecular beam epitaxy', Journal of Applied Physics 75, 5012 (1994). DOI: 10.1063/1.355741, https://doi.org/10.1063/1.355741
  10. P. Stallinga, W. Walukiewicz, E. R. Weber, P. Becla, and J. Lagowski, 'Electron-paramagnetic-resonance study of Se-doped AlSb: Evidence for negative U of the DX center', Physical Review B 52, R8609 (1995). DOI: 10.1103/PhysRevB.52.R8609, https://doi.org/10.1103/PhysRevB.52.R8609
  11. P. Becla, A. Witt, J. Lagowski, and W. Walukiewicz, 'Large photoinduced persistent optical absorption in selenium doped AlSb', Applied Physics Letters 67, 395–397 (1995). DOI: 10.1063/1.114640, https://doi.org/10.1063/1.114640
  12. Jun Shen, Herb Goronkin, John D. Dow, and Shang Yuan Ren, 'Tamm states and donors at InAs/AlSb interfaces', Journal of Applied Physics 77, 1576 (1995). DOI: 10.1063/1.358910, https://doi.org/10.1063/1.358910
  13. Paul Erhart, Daniel Åberg, and Vincenzo Lordi, 'Extrinsic point defects in aluminum antimonide', Physical Review B 81, 195216 (2010). DOI: 10.1103/PhysRevB.81.195216, https://doi.org/10.1103/PhysRevB.81.195216; preprint arXiv:1009.0242, https://arxiv.org/abs/1009.0242
  14. V. Ya. Aleshkin, V. I. Gavrilenko, D. M. Gaponova, A. V. Ikonnikov, K. V. Marem'yanin, S. V. Morozov, Yu. G. Sadofyev, S. R. Johnson, and Y.-H. Zhang, 'Spectra of persistent photoconductivity in InAs/AlSb quantum-well heterostructures', Semiconductors 39, 22–26 (2005). DOI: 10.1134/1.1852637, https://doi.org/10.1134/1.1852637
  15. Yasutomo Kajikawa, 'Hopping Conduction in DX-Center-Related Impurity Bands in AlxGa1−xSb, AlxGa1−xAs, and GaAs1−xPx', Journal of Electronic Materials 54, 727–746 (published online 13 November 2024). DOI: 10.1007/s11664-024-11525-8, https://doi.org/10.1007/s11664-024-11525-8
  16. Sergey Yu. Karpov, 'A Universal Model for DX-Center Binding Energy in Cubic III–V Compounds', physica status solidi (b) 258, 2000596 (2021). DOI: 10.1002/pssb.202000596, https://doi.org/10.1002/pssb.202000596
  17. Philip C. Klipstein, 'Perspective on III–V barrier detectors', Applied Physics Letters 120, 060502 (2022). DOI: 10.1063/5.0084100, https://doi.org/10.1063/5.0084100
  18. Saroj Kumar Patra, Thanh-Nam Tran, Lasse Vines, Ilia Kolevatov, Edouard Monakhov, and Bjørn-Ove Fimland, 'Dopant incorporation in Al0.9Ga0.1As0.06Sb0.94 grown by molecular beam epitaxy', Journal of Crystal Growth 463, 116–122 (2017). DOI: 10.1016/j.jcrysgro.2017.01.035, https://doi.org/10.1016/j.jcrysgro.2017.01.035
  19. A. T. Newell, R. A. Carrasco, J. V. Logan, C. P. Hains, D. Shima, G. Balakrishnan, D. Maestas, C. P. Morath, and P. T. Webster, 'Unintentional doping and native defects in the AlGaAsSb barrier of nBn photodetectors and their effect on detector dark current', Journal of Applied Physics 139, 084504 (2026). DOI: 10.1063/5.0313715, https://doi.org/10.1063/5.0313715