First-principles work by Duan and Stampfl [Phys. Rev. B 77, 115207 (2008)] predicted that nitrogen vacancies (VNV_N) in wurtzite indium nitride bind into clusters whose per-vacancy formation energy falls from 2.15 eV (isolated vacancy) to 1.46 eV for a six-vacancy complex, with contracted In–In bonds and electron-localization signatures of locally indium-rich, metallic-like bonding, and proposed such vacancy-cluster-induced regions as the microscopic origin of the metallic indium inclusions observed in InN films. Whether this mechanism operates in real material remains unestablished. The open problem is to determine — drawing on evidence from beyond-LDA electronic-structure methods, growth-condition-dependent kinetics, and direct experimental characterization — whether VNV_N clusters actually form in grown InN and nucleate its metallic inclusions, or to establish that they do not and identify the actual formation channel. Later literature confirms the inclusions themselves (high-resolution transmission electron microscopy; anomalous Hall-effect modeling showing growth-condition control) and provides indirect, mutually contested positron-annihilation evidence on vacancy complexes, but contains no modern recomputation of VNV_N cluster binding energies, no aggregation kinetics, no study of clusters beyond six vacancies, and no atomic-resolution correlation linking inclusions to nitrogen-vacancy clustering.

Background

Indium nitride (InN) is a group-III nitride semiconductor that crystallizes in the wurtzite structure; it has an unusually narrow band gap (about 0.7 eV), high electron mobility, and a strong tendency toward native n-type conductivity and surface electron accumulation. Like all compound semiconductors, it contains native point defects — lattice sites left empty (vacancies) or occupied by the wrong species. The nitrogen vacancy, VNV_N (an empty nitrogen site), is predicted to act as a shallow donor, i.e. a defect that donates electrons to the conduction band.

Defect populations are governed by formation energies. In the supercell approach of density-functional theory (DFT), one removes atoms from a large periodic cell of the crystal and references the removed atoms to elemental reservoirs; under In-rich conditions the indium reservoir is bulk metallic indium. The formation energy EfE^f depends on the defect's charge state and on the Fermi level EFE_F (i.e. on doping). For a cluster of nn nitrogen vacancies, the binding energy per vacancy, Eb/n=Ef(isolated VN)−Ef(nVN)/n,E_b/n = E^f(\text{isolated } V_N) - E^f(nV_N)/n, measures the thermodynamic gain of aggregation: a positive, growing Eb/nE_b/n means larger clusters are progressively favored. The accuracy of such energies depends on the exchange–correlation treatment: the local-density approximation (LDA) and generalized-gradient approximation (GGA), with or without nonlinear core corrections (nlcc, a scheme treating the In 4d4d electrons as part of the frozen ionic core), can differ from all-electron or hybrid-functional calculations (hybrid functionals mix a fraction of exact exchange into GGA and largely cure the band-gap and energy-scale errors) by several tenths of an eV per vacancy.

Metallic indium inclusions — nanometer-scale regions of elemental In embedded in the semiconductor — have been directly observed in molecular-beam-epitaxy (MBE) InN by high-resolution transmission electron microscopy, and they strongly affect the optical and electrical properties. What microscopic process produces them during growth is not known.

Duan and Stampfl (2008) computed, using LDA+nlcc pseudopotential supercells (96–128 atoms; in the pseudopotential approach the deep core electrons are replaced by an effective ionic potential acting on the valence electrons, while all-electron calculations treat every electron explicitly) with all-electron GGA cross-checks, the structure, charge states, and energetics of neutral and charged VNV_N clusters with up to n=6n=6 in InN. They found that the per-vacancy formation energy decreases monotonically with nn (from 2.15 eV for the isolated vacancy to 1.46 eV for 6VN6V_N under In-rich conditions), that the binding energy per vacancy grows (0.30 eV for n=2n=2 to 0.70 eV for n=6n=6), and that several In–In distances around the clusters contract by 8–14% toward bulk-indium-like separations, with electron-localization-function (a real-space measure of where electrons concentrate and form bonds) and charge-density signatures read as metallic-like bonding in a locally indium-rich region. They proposed these vacancy-cluster-induced regions as a possible microscopic origin of the observed metallic In inclusions. Clustering was predicted to be robust in neutral and negative charge states but absent in the highest positive (p-type) charge states. Two methodological warnings accompanied the results: all-electron GGA checks lowered the absolute formation energies by about 0.5–0.6 eV and gave binding energies 2–5 times smaller (0.06–0.26 eV per vacancy for n=2n=2–4) than LDA, and the LDA band gap (0.65 eV) leaves the placement of the cluster-induced defect levels near the conduction-band edge uncertain.

A key experimental probe in this area is positron annihilation spectroscopy, in which implanted positrons become trapped at open-volume defects and annihilate with electrons; the annihilation radiation encodes the open volume and chemical environment of the trap, making the technique sensitive to vacancy-type defects. As discussed below, theory indicates that pure nitrogen-vacancy complexes, being locally electron-rich rather than open-volume defects, are hard or impossible to see this way — one of several reasons the experimental status of VNV_N clusters remains unsettled.

Problem Statement

Determine whether nitrogen-vacancy clusters actually form in grown indium nitride and whether they are the microscopic origin of the metallic indium inclusions observed in InN samples — or establish that they are not, identifying the formation channel that instead operates. The four directions of evidence below all bear on this single question. They are complementary evidence rather than four separately required acceptance criteria: a decisive result in direction (iii) — identifying VNV_N clusters in as-grown InN and linking them to the observed inclusions, or excluding them and discriminating the formation channel that instead operates — resolves the task, whereas results in (i) and (ii) establish only whether clustering is thermodynamically and kinetically possible and do not by themselves show that it occurs in grown material. Direction (iv) is an optional extension, not a required deliverable.

(i) Reliable thermodynamics of VNV_N clusters: formation and binding energies of nVNnV_N complexes (n=2n=2 up to 6, in their relevant charge states) computed beyond the original nlcc-LDA treatment — for example at a hybrid-functional or comparably validated level of theory, with converged supercells and controlled treatment of charged states in this narrow-gap host — establishing whether the per-vacancy binding genuinely persists and grows with nn (and whether Ef/nE^f/n saturates), or was substantially an artifact of LDA (the source's own all-electron GGA checks already gave binding energies 2–5 times smaller than LDA, while preserving the qualitative clustering trend).

(ii) Kinetic feasibility: whether nitrogen-vacancy diffusion and aggregation can occur under realistic growth and annealing conditions — migration barriers beyond the single computed c-axis monomer path, across directions and charge states, and the aggregation pathway from isolated vacancies to bound clusters under the non-equilibrium conditions of plasma-assisted MBE (the roles of growth mode, polarity, temperature, In:N ratio, and doping in enabling or suppressing clustering).

(iii) Direct experimental identification: characterization of as-grown InN (for example atomic-resolution transmission electron microscopy correlated with local electronic/chemical spectroscopy, or probe-level identification validated by ab initio fingerprints) that either identifies VNV_N clusters and links vacancy-cluster-rich, locally In-rich metallic regions to the observed inclusions, or excludes this mechanism by discriminating it from indium-vacancy-based complexes and from heterogeneous nucleation of metallic In at dislocations, surfaces, and layer interfaces.

(iv) Cluster growth and phase separation: whether and how the predicted vacancy clustering can evolve into true metallic indium precipitates. Aggregates larger than six vacancies lie beyond the source's computed range and may be examined as an explicitly announced extension, not a required deliverable.

References

  1. X. M. Duan and C. Stampfl, "Nitrogen vacancies in InN: Vacancy clustering and metallic bonding from first principles", Physical Review B 77, 115207 (2008). DOI: 10.1103/PhysRevB.77.115207, https://doi.org/10.1103/PhysRevB.77.115207
  2. X. M. Duan and C. Stampfl, "Vacancies and interstitials in indium nitride: Vacancy clustering and molecular bondlike formation from first principles", Physical Review B 79, 174202 (2009). DOI: 10.1103/PhysRevB.79.174202, https://doi.org/10.1103/PhysRevB.79.174202
  3. T. P. Bartel, C. Kisielowski, P. Specht, T. V. Shubina, V. N. Jmerik, and S. V. Ivanov, "High resolution transmission electron microscopy of InN", Applied Physics Letters 91, 101908 (2007). DOI: 10.1063/1.2779843, https://doi.org/10.1063/1.2779843
  4. T. A. Komissarova, M. A. Shakhov, V. N. Jmerik, T. V. Shubina, R. V. Parfeniev, S. V. Ivanov, X. Wang, and A. Yoshikawa, "Abnormal magnetic-field dependence of Hall coefficient in InN epilayers", Applied Physics Letters 95, 012107 (2009). DOI: 10.1063/1.3167823, https://doi.org/10.1063/1.3167823
  5. T. A. Komissarova, V. N. Jmerik, S. V. Ivanov, and P. Paturi, "Detection of metallic In nanoparticles in InGaN alloys", Applied Physics Letters 99, 072107 (2011). DOI: 10.1063/1.3626039, https://doi.org/10.1063/1.3626039
  6. T. A. Komissarova, P. Wang, P. Paturi, X. Wang, and S. V. Ivanov, "Influence of MBE growth modes and conditions on spontaneous formation of metallic In nanoparticles and electrical properties of InN matrix", Journal of Crystal Growth 478, 216–219 (2017). DOI: 10.1016/j.jcrysgro.2017.09.010, https://doi.org/10.1016/j.jcrysgro.2017.09.010
  7. T. A. Komissarova, V. N. Jmerik, and S. V. Ivanov, "Spontaneous Formation of Indium Clusters in InN Epilayers Grown by Molecular-Beam Epitaxy", Technical Physics Letters 44, 149–152 (2018). DOI: 10.1134/S1063785018020220, https://doi.org/10.1134/S1063785018020220
  8. A. Uedono, H. Nakamori, K. Narita, J. Suzuki, X. Wang, S.-B. Che, Y. Ishitani, A. Yoshikawa, and S. Ishibashi, "Vacancy-type defects in Mg-doped InN probed by means of positron annihilation", Journal of Applied Physics 105, 054507 (2009). DOI: 10.1063/1.3075907, https://doi.org/10.1063/1.3075907
  9. C. Rauch, I. Makkonen, and F. Tuomisto, "Towards experimental identification of vacancy complexes in InN", physica status solidi (a) 208, 1548–1550 (2011); preprint arXiv:1104.2498, "Identifying vacancy complexes in compound semiconductors with positron annihilation spectroscopy: a case study of InN". DOI: 10.1002/pssa.201001160, https://doi.org/10.1002/pssa.201001160 ; arXiv:1104.2498, https://arxiv.org/abs/1104.2498
  10. C. Rauch, F. Tuomisto, A. Vilalta-Clemente, B. Lacroix, P. Ruterana, S. Kraeusel, B. Hourahine, and W. J. Schaff, "Defect evolution and interplay in n-type InN", Applied Physics Letters 100, 091907 (2012). DOI: 10.1063/1.3688038, https://doi.org/10.1063/1.3688038
  11. A. Uedono, S. Ishibashi, T. Watanabe, X. Q. Wang, S. T. Liu, G. Chen, L. W. Sang, M. Sumiya, and B. Shen, "Vacancy-type defects in InxGa1−xN alloys probed using a monoenergetic positron beam", Journal of Applied Physics 112, 014507 (2012). DOI: 10.1063/1.4732141, https://doi.org/10.1063/1.4732141
  12. A. Janotti, J. L. Lyons, and C. G. Van de Walle, "Hybrid functional calculations of native point defects in InN", physica status solidi (a) 209, 65–70 (2012). DOI: 10.1002/pssa.201100216, https://doi.org/10.1002/pssa.201100216
  13. A. Terentjevs, A. Catellani, and G. Cicero, "Nitrogen vacancies at InN (1̄100) surfaces: A theoretical study", Applied Physics Letters 96, 171901 (2010). DOI: 10.1063/1.3409224, https://doi.org/10.1063/1.3409224
  14. R. Hrytsak, P. Kempisty, E. Grzanka, M. Leszczynski, and M. Sznajder, "DFT study on point defects migration through the pseudomorphic and lattice-matched InN/GaN interfaces", Computational Materials Science 186, 110239 (2021). DOI: 10.1016/j.commatsci.2020.110039, https://doi.org/10.1016/j.commatsci.2020.110039
  15. E. Shin, Y. Kang, and S. H. Jeon, "Nitrogen Complex-Driven Vacancy Cluster in Group-III Nitrides", ACS Applied Materials & Interfaces 16, 64244–64252 (2024). DOI: 10.1021/acsami.4c13648, https://doi.org/10.1021/acsami.4c13648
  16. E. Shin, Y. Kang, and S. H. Jeon, "Thermodynamic Role of Nitrogen Complex in Vacancy Cluster of Group-III Nitrides", The Journal of Physical Chemistry C 129, 8372–8381 (2025). DOI: 10.1021/acs.jpcc.5c00874, https://doi.org/10.1021/acs.jpcc.5c00874