InAs/AlSb n-N isotype heterojunctions (both sides n-type, with a large conduction-band offset between the wider-gap AlSb and the narrower-gap InAs) grown by molecular beam epitaxy on lattice-mismatched GaAs substrates were shown in 1989 to rectify like Schottky diodes despite more than 107 cm−210^7\ \mathrm{cm}^{-2} threading dislocations, and capacitance-voltage profiling yielded the InAs/AlSb conduction-band offset ΔEC=1.35±0.05 eV\Delta E_C=1.35\pm0.05\ \mathrm{eV}, a value later adopted as the accepted electron barrier for this materials pair. After correction for a roughly 1 kΩ1\ \mathrm{k}\Omega series resistance, however, the forward current-voltage characteristic showed an ideality factor of about 1.8, far from the value near 1 expected for thermionic emission over the heterojunction barrier, and the authors could only hypothesize defect-path conduction related to the misfit dislocations. The subsequent literature settles the band-offset uncertainty (the 1.35 eV value is consistent with the X-valley-referenced barrier, with a Gamma-referenced quantum-well depth of about 2.1-2.3 eV), and buffer engineering has removed the GaAs/AlSb interface-spike problem as a growth constraint. What remains open is the mechanism-level question: no published study demonstrates, or refutes, the dislocation/defect-path hypothesis for the forward current in these junctions, and no quantitative relation between threading-dislocation density (or the associated deep-level burden in the AlSb barrier) and junction figures of merit (forward ideality, reverse leakage, rectification ratio, reproducibility) exists. The problem is to settle that mechanism experimentally.

Background

InAs, GaSb, and AlSb share nearly the same lattice constant of about 6.1 Å and are therefore called the 6.1 Å family; molecular beam epitaxy (MBE, a vacuum growth technique depositing one atomic layer at a time) can combine them into nearly lattice-matched heterostructures. Because the AlSb conduction band lies far above that of InAs, an InAs layer between AlSb barriers forms one of the deepest electron quantum wells available among III-V semiconductors, which underlies InAs/AlSb resonant tunneling diodes (two-terminal devices whose current peaks where a discrete quantum level aligns with the emitter band), high-electron-mobility transistors (HEMTs), and quantum cascade lasers (unipolar mid-infrared lasers operating on transitions between confined conduction-band states).

An isotype heterojunction is a junction between two semiconductors doped with the same carrier type; an n-N junction consists of an n-type, lower-gap side (InAs) and an N-type, wider-gap side (AlSb), both n-type but with different band gaps and doping. Because of the large conduction-band offset ΔEC\Delta E_C between them, electrons on the InAs side facing the AlSb barrier see a potential step closely analogous to a metal-semiconductor (Schottky) contact, so the junction rectifies (passes current far more readily in one bias direction) even though both sides are n-type. The 1989 study of Nakagawa, Kroemer, and English grew such junctions by MBE on heavily doped n+n^+-GaAs substrates and characterized them with two standard techniques. First, capacitance-voltage (C-V) profiling of the reverse-biased depletion region on the AlSb side: for a one-sided junction, 1/C21/C^2 varies linearly with voltage VV, and the intercept VintV_{\mathrm{int}} of the extrapolated line gives the band offset through ΔEC=qVint+EF2C+EF1C+kT\Delta E_C=qV_{\mathrm{int}}+E_{F2C}+E_{F1C}+kT, where qq is the elementary charge, kTkT the thermal energy, and EF2CE_{F2C}, EF1CE_{F1C} the Fermi-level depths measured from the AlSb and InAs conduction-band edges. Second, forward-bias current-voltage (I-V) measurements, conventionally summarized by the ideality factor n=(q/kT) dV/d(ln⁡I)n=(q/kT)\,dV/d(\ln I): thermionic emission or diffusion over a uniform barrier gives n≈1n\approx1, while recombination-dominated current gives n≈2n\approx2.

Two complications dominated the experiment. (i) The epitaxial layers had to grow on GaAs, whose lattice constant mismatches AlSb by about 7%; the strain relaxes through misfit dislocations (line defects at the interface) that spawn threading dislocations (defects running through the grown layers) at densities exceeding 107 cm−210^7\ \mathrm{cm}^{-2} (a density estimated from transmission electron microscopy on structurally similar wafers, not directly measured on the electrically tested junctions). (ii) The ungraded GaAs/AlSb interface creates a conduction-band spike whose resistance (measured at roughly 1 kΩ1\ \mathrm{k}\Omega in these structures) is in series with the junction and must be corrected for before the junction's own I-V can be extracted. After such correction, the forward current showed n=1.80n=1.80 at room temperature over about four decades, far from ideal thermionic emission, and the authors suggested that the main current flow is carried by a defect path — their informal term for a current path tied to the threading dislocations that penetrate the AlSb barrier, scaling with the dislocation density, as distinct from trap-assisted recombination through deep levels in the barrier. The C-V analysis gave ΔEC=1.35±0.05 eV\Delta E_C=1.35\pm0.05\ \mathrm{eV}, in excellent agreement with the value predicted from the known InAs/GaSb and GaSb/AlSb offsets by transitivity (the assumption that band offsets compose like differences of band-edge energies). The AlSb conduction-band edge entering that analysis is the one relevant to the X-point valley: AlSb is an indirect-gap semiconductor whose zone-center (Gamma) conduction band lies considerably higher, a distinction that later work made explicit. The authors left undemonstrated the microscopic origin of the n≈1.8n\approx1.8 nonideality; the InAs-side band bending entering the C-V analysis is complicated by strong degeneracy (the Fermi level lying deep inside the InAs conduction band), conduction-band nonparabolicity (the deviation of the InAs conduction-band dispersion from a simple parabola), and possible quantum confinement in narrow potential notches (thin band-bending dips near the interface); and the ungraded GaAs/AlSb spike contributed the series resistance that had to be corrected for before the junction's own characteristics could be extracted.

Problem Statement

For InAs/AlSb n-N isotype heterojunctions grown by molecular beam epitaxy on lattice-mismatched GaAs substrates, with the layer structure, doping levels, and measurement conventions of the 1989 originating study as the reference point, establish the microscopic origin of the large forward-bias nonideality exhibited by the forward current-voltage characteristic after correction for series resistance. Concretely: demonstrate, or refute, by direct experiment the paper's hypothesis that the main forward current is carried by such a dislocation-associated defect path — the reference structures contain threading dislocations at densities exceeding 107 cm−210^7\ \mathrm{cm}^{-2} resulting from the roughly 7% AlSb/GaAs lattice mismatch — rather than by thermionic emission over the heterojunction barrier. Because the ungraded GaAs/AlSb interface contributes a series resistance (roughly 1 kΩ1\ \mathrm{k}\Omega as reported in the reference structures), the junction's forward characteristics must be extracted after correction for that resistance, or after it has been separately determined, so that the nonideality attributed to the junction itself is not measurement-limited. The current-carrying mechanism responsible for the nonideal forward characteristic is to be identified and shown quantitatively to account for the forward-bias nonideality (n≈1.8n\approx1.8 after series-resistance correction) observed in the reference structures, by a discriminating experimental determination; "the observed forward characteristics" means the forward I-V behavior reported by the 1989 originating study. The band-offset part of the original question is already settled by later literature (see the opening paragraph) and is not part of this problem.

References

  1. A. Nakagawa, H. Kroemer, and J. H. English, "Electrical properties and band offsets of InAs/AlSb n-N isotype heterojunctions grown on GaAs," Applied Physics Letters 54(19), 1893-1895 (1989). DOI: 10.1063/1.101233, https://doi.org/10.1063/1.101233
  2. 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(15), 1551-1553 (1990). DOI: 10.1063/1.103350, https://doi.org/10.1063/1.103350
  3. G. Tuttle, H. Kroemer, and J. H. English, "Effects of interface layer sequencing on the transport properties of InAs/AlSb quantum wells: Evidence for antisite donors at the InAs/AlSb interface," Journal of Applied Physics 67(6), 3032-3037 (1990). DOI: 10.1063/1.345426, https://doi.org/10.1063/1.345426
  4. J. R. Soderstrom, D. H. Chow, and T. C. McGill, "InAs/AlSb double-barrier structure with large peak-to-valley current ratio: a candidate for high-frequency microwave devices," IEEE Electron Device Letters 11(1), 27-29 (1990). DOI: 10.1109/55.46920, https://doi.org/10.1109/55.46920
  5. R. G. Dandrea and C. B. Duke, "Calculation of InAs/AlSb(001) band offsets: Effect of strain and interfacial atomic structure," Applied Physics Letters 63(13), 1795-1797 (1993). DOI: 10.1063/1.110665, https://doi.org/10.1063/1.110665
  6. H. Kroemer and E. Hu, "Semiconducting and Superconducting Physics and Devices in the InAs/AlSb Materials System," in Nanotechnology, edited by G. L. Timp (Springer, New York, 1999), pp. 629-688. DOI: 10.1007/978-1-4612-0531-9_15, https://doi.org/10.1007/978-1-4612-0531-9_15
  7. I. Vurgaftman, J. R. Meyer, and L. R. Ram-Mohan, "Band parameters for III-V compound semiconductors and their alloys," Journal of Applied Physics 89(11), 5815-5875 (2001). DOI: 10.1063/1.1368156, https://doi.org/10.1063/1.1368156
  8. C. R. Bolognesi, "Antimonide-based high-speed electronics: a transistor perspective," Conference Proceedings. 14th Indium Phosphide and Related Materials Conference (IPRM 2002), IEEE, pp. 55-58 (2002). DOI: 10.1109/iciprm.2002.1014102, https://doi.org/10.1109/iciprm.2002.1014102
  9. H. Kroemer, "The 6.1 Å family (InAs, GaSb, AlSb) and its heterostructures: a selective review," Physica E 20(3-4), 196-203 (2004). DOI: 10.1016/j.physe.2003.08.003, https://doi.org/10.1016/j.physe.2003.08.003 (open-access full text at https://web.ece.ucsb.edu/Faculty/kroemer/pubs/14_04_6_1Family.pdf)
  10. J. Devenson, D. Barate, O. Cathabard, R. Teissier, and A. N. Baranov, "Very short wavelength (λ = 3.1-3.3 μm) quantum cascade lasers," Applied Physics Letters 89(19), 191115 (2006). DOI: 10.1063/1.2387473, https://doi.org/10.1063/1.2387473
  11. F. Wang, Y. Jia, S.-F. Li, and Q. Sun, "First-principles calculation of the 6.1 Å family bowing parameters and band offsets," Journal of Applied Physics 105(4), 043101 (2009). DOI: 10.1063/1.3072688, https://doi.org/10.1063/1.3072688
  12. G. Moschetti, H. Zhao, P.-Å. Nilsson, S. Wang, A. Kalabukhov, G. Dambrine, S. Bollaert, L. Desplanque, X. Wallart, and J. Grahn, "Anisotropic transport properties in InAs/AlSb heterostructures," Applied Physics Letters 97(24), 243510 (2010). DOI: 10.1063/1.3527971, https://doi.org/10.1063/1.3527971
  13. J. Zhang, H. Lv, H. Ni, Z. Niu, and Y. Zhang, "Temperature dependence on the electrical and physical performance of InAs/AlSb heterojunction and high electron mobility transistors," Chinese Physics B 27(9), 097201 (2018). DOI: 10.1088/1674-1056/27/9/097201, https://doi.org/10.1088/1674-1056/27/9/097201
  14. Q. Wei, H. Wang, X. Zhao, and J. Zhao, "Electron mobility anisotropy in (Al,Ga)Sb/InAs two-dimensional electron gases epitaxied on GaAs (001) substrates," Journal of Semiconductors 43(7), 072101 (2022). DOI: 10.1088/1674-4926/43/7/072101, https://doi.org/10.1088/1674-4926/43/7/072101