Microscopic origin of the deep subgap states in amorphous InGaZnO4 and a physically validated full-subgap density-of-states model for a-IGZO thin-film transistors

Amorphous InGaZnO4 ($a$-IGZO) is the workhorse channel material of oxide thin-film transistors (TFTs) in display backplanes and, increasingly, in dynamic random-access memory (DRAM) development, and its subgap density of states (DOS) is the central materials parameter controlling trapping, subthreshold swing, turn-on voltage, stability, and off-state leakage. The foundational device model of Hsieh, Kamiya, Nomura, Hosono, and Wu (Appl. Phys. Lett. 92, 133503, 2008) extracted a two-component subgap DOS (exponential tail plus Gaussian deep states) valid only from the conduction-band edge $E_C$ down to $E_C-0.8$ eV and left four coupled issues open: the microscopic identity and formation mechanism of the deep gap states (oxygen deficiency suggested but unconfirmed), the unconstrained DOS outside the sampled window, the unmodeled off-state leakage, and the unresolved separation of gate-insulator interface traps from bulk traps. Later literature has largely closed the energy-window and interface/bulk-separation issues and has produced off-leakage models that assume deep donor states. The surviving open core is the definitive, experimentally confirmed microscopic identification of the deep subgap states — with competing oxygen-vacancy, zinc-vacancy, hydrogen-related, and cation-undercoordination assignments still coexisting — and its quantitative integration into a validated full-subgap DOS device model.

Background

A thin-film transistor (TFT) is a field-effect transistor in which the channel is a thin semiconductor film deposited on an insulating substrate, used as the switching element in large-area electronics such as display backplanes. Amorphous oxide semiconductors (AOSs), of which amorphous InGaZnO4 ($a$-IGZO) is the archetype, differ from covalent amorphous semiconductors such as hydrogenated amorphous silicon in that their conduction bands arise from spatially extended metal $ns$ orbitals, so they retain high electron mobilities even in the amorphous phase. As in any noncrystalline semiconductor, however, disorder and defects introduce electronic states inside the band gap: exponential band-tail states descending from the conduction-band edge $E_C$ (the mobility edge above which carriers move quasi-freely) and, deeper in the gap, broadly distributed defect states. This energy distribution is called the subgap density of states (DOS), and it governs carrier trapping, the Fermi-level position, the subthreshold swing $S$ (the gate voltage required per decade of drain-current change below threshold), the turn-on voltage, long-term stability under bias and illumination stress, and off-state leakage. It is therefore the key input for TCAD (technology computer-aided design) device simulation of TFT circuits.

Hsieh, Kamiya, Nomura, Hosono, and Wu (Appl. Phys. Lett. 92, 133503, 2008) constructed the first compact TCAD model for $a$-IGZO TFTs: a constant field-independent electron mobility combined with a two-component acceptor-like subgap DOS, namely an exponential tail $g_{\mathrm{exp}}(E)=N_{TA}\exp[-(E_C-E)/W_{TA}]$ and a monotonically decreasing Gaussian deep-gap component $g_G(E)=N_{GA}\exp[-(E_0-E)/W_{GA}]^2$ with the center fixed at $E_0=E_C$. Fitting measured transfer and output characteristics of pulsed-laser-deposited depletion-type (deposited at oxygen partial pressure 5.2 Pa, background electron concentration $4.1\times10^{17}\ \mathrm{cm^{-3}}$) and enhancement-type (6.5 Pa, $1\times10^{15}\ \mathrm{cm^{-3}}$) devices with 40 nm channels and 100 nm SiO2 gate insulators yielded, respectively, $N_{TA}=2.3\times10^{18}\ \mathrm{cm^{-3}eV^{-1}}$, $W_{TA}=0.14$ and 0.08 eV, $N_{GA}=4.5\times10^{16}$ and $3.2\times10^{16}\ \mathrm{cm^{-3}eV^{-1}}$, $W_{GA}=1.5$ eV, and mobilities of 11.88 and $8.57\ \mathrm{cm^2V^{-1}s^{-1}}$; the small subthreshold swings (0.224 and 0.104 V/decade) confirmed that $a$-IGZO's subgap DOS is orders of magnitude below that of amorphous silicon. The model, however, left four coupled issues open, which define this problem. First, the microscopic physical identity and formation mechanism of the deep gap states were undetermined; oxygen deficiency was suggested but not experimentally confirmed as the dominant origin. Second, the extracted DOS is constrained by the data only from $E_C$ down to roughly $E_C-0.8$ eV, the range swept by the channel Fermi level under the applied biases, so the DOS deeper in the gap — which controls off-state leakage and deep trapping — was unconstrained. Third, off-state currents and leakage mechanisms were not modeled. Fourth, interface trap states at the gate-insulator/$a$-IGZO boundary were implicitly subsumed into the bulk subgap DOS rather than separately quantified.

Subsequent work built an extensive toolkit for this problem. Hard x-ray photoelectron spectroscopy (HAXPES — x-ray photoelectron spectroscopy with high-energy photons giving bulk-sensitive probing of valence-region electronic structure) directly observed subgap states in $a$-IGZO films. Photonic capacitance–voltage techniques extract donor- and acceptor-like subgap DOS by illuminating the TFT with sub-bandgap photons while sweeping gate bias. Ultrabroadband photoconduction (UBPC — measuring the photoconductive response over photon energies spanning the whole subgap, typically 0.3–3.5 eV) recovers the full-subgap DOS from TFTs themselves. On the theory side, density-functional theory (DFT) calculations on model amorphous structures (often with the DFT+U correction for localized states) compute candidate defect levels, and defect formation energies indicate which defects are likely present. The question is how these threads combine into a physically identified, validated DOS model.

Problem Statement

Establish the microscopic physical identity and formation mechanism of the deep subgap (gap) states in amorphous InGaZnO4, and use it to complete the physically validated subgap-DOS-based device model of $a$-IGZO thin-film transistors that Hsieh et al. (2008) left unfinished. Specifically: (1) determine which atomic-scale defects — oxygen vacancies (with their local-structure-dependent levels), zinc vacancies, hydrogen-related states, or undercoordinated cations/oxygen — actually generate the deep gap states measured in the subgap density of states (the energy distribution of electronic states inside the band gap), and confirm this experimentally rather than only by first-principles assignment, including the formation mechanisms (deposition conditions, oxygen partial pressure, annealing, hydrogen incorporation) that control their density; (2) constrain the subgap DOS over the entire band gap rather than only from $E_C$ down to $E_C-0.8$ eV, so that deep trapping and recombination-relevant states are quantified; (3) incorporate off-state currents and leakage mechanisms into the device model, with the deep states microscopically identified rather than assumed; and (4) separate gate-insulator interface traps from bulk subgap states and quantify each. The common objective is a single validated model in which the measured full-subgap DOS is explained defect-by-defect and predicts TFT transfer/output characteristics, subthreshold swing, turn-on voltage, off-state leakage, and stress stability. Of these four coupled parts, (2) and (4) have been substantially addressed by later techniques (ultrabroadband photoconduction and photonic I-V/C-V extraction; channel-thickness-based interface/bulk decomposition), and (3) has been modeled only on the assumption of deep donor states of unconfirmed identity; the open core that remains is part (1) — an experimentally confirmed microscopic identification of the deep states, valid across the process space of real $a$-IGZO devices, together with its quantitative integration into the validated device model above. Here 'experimentally confirmed' means the microscopic assignment is supported by an independent experimental probe, not by first-principles calculation alone.

References

1. H.-H. Hsieh, T. Kamiya, K. Nomura, H. Hosono, and C.-C. Wu, 'Modeling of amorphous InGaZnO4 thin film transistors and their subgap density of states,' Appl. Phys. Lett. 92, 133503 (2008). DOI: 10.1063/1.2857463, https://doi.org/10.1063/1.2857463 (source paper)
2. M. Kimura, T. Nakanishi, K. Nomura, T. Kamiya, and H. Hosono, 'Trap densities in amorphous-InGaZnO4 thin-film transistors,' Appl. Phys. Lett. 92, 133512 (2008). DOI: 10.1063/1.2904704, https://doi.org/10.1063/1.2904704 (independent contemporary trap-density extraction from low-frequency capacitance-voltage measurement)
3. T. Kamiya, K. Nomura, M. Hirano, and H. Hosono, 'Electronic structure of oxygen deficient amorphous oxide semiconductor a-InGaZnO(4-x): Optical analyses and first-principle calculations,' Phys. Status Solidi C 5, 3098 (2008). DOI: 10.1002/pssc.200779300, https://doi.org/10.1002/pssc.200779300 (companion DFT/optical study of oxygen-deficiency states in a-IGZO)
4. K. Nomura, T. Kamiya, H. Yanagi, E. Ikenaga, K. Yang, K. Kobayashi, M. Hirano, and H. Hosono, 'Subgap states in transparent amorphous oxide semiconductor, In-Ga-Zn-O, observed by bulk sensitive x-ray photoelectron spectroscopy,' Appl. Phys. Lett. 92, 202117 (2008). DOI: 10.1063/1.2927306, https://doi.org/10.1063/1.2927306 (HAXPES observation of subgap DOS near the valence band)
5. T. Kamiya, K. Nomura, and H. Hosono, 'Subgap states, doping and defect formation energies in amorphous oxide semiconductor a-InGaZnO4 studied by density functional theory,' Phys. Status Solidi A 207, 1698 (2010). DOI: 10.1002/pssa.200983772, https://doi.org/10.1002/pssa.200983772 (DFT: oxygen deficiencies form deep occupied states near the valence-band maximum or shallow donors depending on local structure)
6. I.-J. Kang and C. H. Park, 'Microscopic Property of Amorphous Semiconductor Metal Oxide InGaZnO4 and Role of O-deficiency,' arXiv:1108.2086 (2011). DOI: 10.48550/arXiv.1108.2086, https://arxiv.org/abs/1108.2086 (DFT: O-deficiency does not create a donor level near the conduction-band minimum)
7. E. K.-H. Yu, S. Jun, D. H. Kim, and J. Kanicki, 'Density of states of amorphous In-Ga-Zn-O from electrical and optical characterization,' J. Appl. Phys. 116, 154505 (2014). DOI: 10.1063/1.4898567, https://doi.org/10.1063/1.4898567 (subgap DOS model combining UV-Vis optical absorption and multi-frequency C-V)
8. G. Wakimura, Y. Yamauchi, T. Matsuoka, and Y. Kamakura, 'Mechanism of off-leakage current in InGaZnO thin-film transistors,' 2014 IEEE International Meeting for Future of Electron Devices, Kansai (IMFEDK) (2014). DOI: 10.1109/imfedk.2014.6867062, https://doi.org/10.1109/imfedk.2014.6867062
9. G. Wakimura, Y. Yamauchi, and Y. Kamakura, 'Simulation and modeling of off-leakage current in InGaZnO thin-film transistors,' J. Adv. Simul. Sci. Eng. 2, 201 (2015). DOI: 10.15748/jasse.2.201, https://doi.org/10.15748/jasse.2.201 (TCAD off-leakage model based on deep donor-like states; microscopic origin of those states left unresolved)
10. S. Sallis, N. F. Quackenbush, D. S. Williams, M. Senger, J. C. Woicik, B. E. White, and L. F. J. Piper, 'Deep subgap feature in amorphous indium gallium zinc oxide: Evidence against reduced indium,' Phys. Status Solidi A 212, 1471 (2015). DOI: 10.1002/pssa.201431806, https://doi.org/10.1002/pssa.201431806 (evidence against In+ as the deep subgap feature; supports undercoordinated oxygen)
11. W. H. Han and K. J. Chang, 'Subgap States near the Conduction-Band Edge Due to Undercoordinated Cations in Amorphous In-Ga-Zn-O and Zn-Sn-O Semiconductors,' Phys. Rev. Applied 6, 044011 (2016). DOI: 10.1103/PhysRevApplied.6.044011, https://doi.org/10.1103/PhysRevApplied.6.044011 (DFT: undercoordinated cations as origin of near-CB subgap states)
12. J. Jia, A. Suko, Y. Shigesato, T. Okajima, K. Inoue, and H. Hosomi, 'Evolution of Defect Structures and Deep Subgap States during Annealing of Amorphous In-Ga-Zn Oxide for Thin-Film Transistors,' Phys. Rev. Applied 9, 014018 (2018). DOI: 10.1103/PhysRevApplied.9.014018, https://doi.org/10.1103/PhysRevApplied.9.014018 (annealing evolution of deep subgap states)
13. J. Bang, S. Matsuishi, and H. Hosono, 'Hydrogen anion and subgap states in amorphous In-Ga-Zn-O thin films for TFT applications,' Appl. Phys. Lett. 110, 232105 (2017). DOI: 10.1063/1.4985627, https://doi.org/10.1063/1.4985627 (direct evidence for anionic hydrogen bonding to metal cations contributing subgap states)
14. S. Choi, J.-Y. Kim, J. Rhee, H. Kang, S. Park, D. M. Kim, S.-J. Choi, and D. H. Kim, 'Method to Extract Interface and Bulk Trap Separately Over the Full Sub-Gap Range in Amorphous InGaZnO Thin-Film Transistors by Using Various Channel Thicknesses,' IEEE Electron Device Letters 40, 574 (2019). DOI: 10.1109/led.2019.2898217, https://doi.org/10.1109/led.2019.2898217 (decomposition of interface trap density and bulk subgap DOS over the full subgap)
15. K. Ide, K. Nomura, H. Hosono, and T. Kamiya, 'Electronic Defects in Amorphous Oxide Semiconductors: A Review,' Phys. Status Solidi A 216, 1800372 (2019). DOI: 10.1002/pssa.201800372, https://doi.org/10.1002/pssa.201800372 (review: many AOS defects are oxygen- and hydrogen-related; instability the most serious issue)
16. K. T. Vogt, C. E. Malmberg, J. C. Buchanan, G. W. Mattson, G. M. Brandt, D. B. Fast, P. H.-Y. Cheong, J. F. Wager, and M. W. Graham, 'Ultrabroadband density of states of amorphous In-Ga-Zn-O,' Phys. Rev. Research 2, 033358 (2020). DOI: 10.1103/PhysRevResearch.2.033358, https://doi.org/10.1103/PhysRevResearch.2.033358 (UBPC full-subgap DOS with DFT classification of oxygen- and zinc-vacancy peaks)
17. J. T. Jang, D. Ko, S.-J. Choi, D. M. Kim, and D. H. Kim, 'Observation of Hydrogen-Related Defect in Subgap Density of States and Its Effects Under Positive Bias Stress in Amorphous InGaZnO TFT,' IEEE Electron Device Letters 42, 708 (2021). DOI: 10.1109/led.2021.3066624, https://doi.org/10.1109/led.2021.3066624 (hydrogen-related defect observed within the extracted subgap DOS)
18. G. W. Mattson, K. T. Vogt, J. F. Wager, and M. W. Graham, 'Illuminating Trap Density Trends in Amorphous Oxide Semiconductors with Ultrabroadband Photoconduction,' Adv. Funct. Mater. 33, 2300742 (2023). DOI: 10.1002/adfm.202300742, https://doi.org/10.1002/adfm.202300742 (seven oxygen deep donor vacancy peaks correlated with photoluminescence across 15 TFT types)
19. S. H. Han, H. Kim, J.-H. Bae, S.-J. Choi, D. H. Kim, and D. M. Kim, 'Photovoltaic Effect De-Embedded Photonic C-V Characterization of Subgap Density of States in Amorphous Oxide Semiconductor Thin-Film Transistors,' IEEE Trans. Electron Devices 71, 6795 (2024). DOI: 10.1109/TED.2024.3469161, https://doi.org/10.1109/TED.2024.3469161 (refined full-range subgap DOS extraction)
20. H. Kim, S. H. Han, H. Jeong, H. Yang, S.-J. Choi, D. H. Kim, J.-H. Bae, and D. M. Kim, 'Photonic I-V Technique With Photogating Effect for Extended Extraction of Subgap Density of States in Amorphous Oxide Semiconductor TFTs,' IEEE Trans. Electron Devices 72, 2751 (2025). DOI: 10.1109/TED.2025.3547707, https://doi.org/10.1109/TED.2025.3547707 (extended-energy-range DOS extraction validated against TCAD)
21. H. U. Huzaibi, S.-T. Han, and M. Zhang, 'Model representation in amorphous metal oxide thin-film transistors: a critical review,' npj Flexible Electronics 10, 13 (2026). DOI: 10.1038/s41528-025-00515-3, https://doi.org/10.1038/s41528-025-00515-3 (open-access critical review of a-MO TFT modeling; cites the source paper)
22. M. J. Mattsson, J. Lee, C. E. Malmberg, J. Parker, K. T. Vogt, H. Kim, M. Hong, P. Yun, D. Ha, T. Lee, P. H.-Y. Cheong, J. F. Wager, and M. W. Graham, 'Shallow Trap States Control Electrical Performance of Amorphous Oxide Semiconductor Thin-Film Transistors,' Adv. Funct. Mater. e78029 (2026); preprint arXiv:2602.06329. DOI: 10.1002/adfm.78029, https://doi.org/10.1002/adfm.78029 (also available as arXiv:2602.06329, https://arxiv.org/abs/2602.06329; full-subgap DoS to within 0.1 eV of the mobility edge with DFT+U assignment of observed peaks)