A 2003 device-simulation study by Hossain and coworkers explained the degraded performance of polycrystalline ZnO thin-film transistors (TFTs) by grain boundaries (GBs) carrying localized trap states, assumed to form symmetric Gaussian donor-like and acceptor-like distributions with peak energy E1=1.7E_1=1.7 eV and decay width E2=0.25E_2=0.25 eV, with defect-free crystallites and a defect-free ZnO/insulator interface, and extracted a total GB areal trap density NtN_t of order 101210^{12} cm−2^{-2} by matching simulated field-effect mobility to grain-boundary count. The authors acknowledged this trap distribution as too simple and crude and proposed as next steps the analysis of the trap distribution through temperature-dependent TFT measurements compared with simulation, and the testing of chemical modification of GBs. The open problem is the experimental, energy-resolved determination of the trap distribution limiting transport in polycrystalline ZnO TFT channels — including whether traps are truly GB-localized, whether the donor/acceptor distributions are symmetric Gaussians with the assumed parameters — together with independent validation of the NtN_t extraction method and quantification of how chemical GB modification changes the trap distribution.

Background

A thin-film transistor (TFT) is a field-effect transistor in which the conducting channel is a thin semiconductor film deposited on (or under) a gate insulator; it is the switching device of large-area and transparent electronics. Zinc oxide (ZnO), a wide-band-gap (3.4 eV) transparent semiconductor, was one of the first channel materials for transparent TFTs, but early polycrystalline devices showed field-effect mobilities of only about 1–10 cm2^2/Vs, far below single-crystal values, with threshold shifts and subthreshold degradation attributed to grain boundaries. A grain boundary (GB) is the interface between two crystallites of different orientation; mismatched bonding there creates localized electronic states in the band gap that can trap channel carriers. When acceptor-like traps (neutral when empty, negatively charged when occupied) capture electrons, the GB becomes negatively charged and the adjacent crystallites depleted, producing two back-to-back depletion regions known as a double Schottky barrier; carriers cross it by thermionic emission (thermal excitation over the barrier). For polycrystalline silicon TFTs this picture was formalized in the Levinson method, which extracts the total GB areal trap density from the Arrhenius-like dependence of the drain current on gate voltage.

Hossain et al. (2003) transferred this framework to nanocrystalline ZnO. In a two-dimensional drift-diffusion device simulation, equally spaced GBs of atomic thickness perpendicular to the current were given trap states with Gaussian energy distributions, Nga(E)=Ntaexp⁡{−[(E1a−E)/E2a]2}N_{ga}(E)=N_{ta}\exp\{-[(E_{1a}-E)/E_{2a}]^2\} for acceptor-like states and analogously for donor-like states (positively charged when empty, neutral when occupied; occupation handled by Shockley–Read–Hall capture/emission statistics, the standard phenomenology of trap occupancy rates). They assumed a perfectly symmetric and idealized case, Nta=Ntd=NtN_{ta}=N_{td}=N_t, E1a=E1d=E1E_{1a}=E_{1d}=E_1, E2a=E2d=E2E_{2a}=E_{2d}=E_2 (here NtN_t is the source's total areal grain-boundary trap density, in cm−2^{-2}, and this single parameter sets the peak amplitude of both Gaussians rather than their energy integral), with the peak E1=1.7E_1=1.7 eV (essentially midgap, since energies are referenced to the valence-band edge and the band gap is 3.4 eV) and width E2=0.25E_2=0.25 eV both treated as assumed material constants, and with the crystallites and the ZnO/Si3_3N4_4 gate-insulator interface declared completely defect-free. The simulations showed how gate bias suppresses the GB barrier near the insulator interface (forming the channel there while barriers persist deep in the film), how depletion regions of neighboring nanometer-scale grains overlap (unlike in polycrystalline Si), and, by comparing simulated field-effect mobility versus effective GB count with measured mobilities of 1–10 cm2^2/Vs, yielded Nt≈1012N_t\approx10^{12} cm−2^{-2} — an order of magnitude also reported contemporaneously for rf-sputtered ZnO TFTs by Carcia et al. The authors thus offered a way to estimate GB trap densities from routine electrical data without spectroscopy, while explicitly acknowledging that the assumed trap distribution in the gap is too simple and crude.

Two neighboring interpretive frameworks matter for judging this problem. First, for disordered semiconductors the subgap density of states (DOS) — the density of trap states as a function of energy within the gap, often split into shallow exponentially decaying band-tail states and deeper states — is commonly extracted from temperature-dependent transfer characteristics, and transport is then described by multiple trapping and release (MTR), in which carriers are repeatedly immobilized in these traps and re-released thermally, rather than by discrete GB barriers. Second, local probes such as scanning tunneling microscopy/spectroscopy (STM/STS) can measure the local DOS at individual GBs directly. Whether polycrystalline ZnO TFT transport is governed by GB-localized deep traps with thermionic crossing, as the 2003 model assumes, or by spatially distributed tail and interface states with MTR/percolation conduction, remains the central unresolved question behind the source's acknowledged simplifications.

Problem Statement

Determine, experimentally and with energy resolution, the distribution of trap states that limits carrier transport in the channel of polycrystalline ZnO thin-film transistors, and thereby validate, refine, or replace the grain-boundary trap model and trap-density extraction method of Hossain et al. (2003). Concretely: (i) establish whether the subgap traps are localized at grain boundaries as assumed, or distributed within the crystallites or at the ZnO/gate-insulator interface, both of which the 2003 model declared defect-free; (ii) determine the actual trap energy distribution — its shape and any asymmetry between donor-like and acceptor-like components — versus the assumed symmetric Gaussians with peak energy E1=1.7E_1=1.7 eV and decay width E2=0.25E_2=0.25 eV; (iii) validate the extraction of the total GB areal trap density NtN_t (order 1012^{12} cm−2^{-2}) from measured field-effect mobility and grain-boundary count independently of the assumed trap model and simulated device parameters; (iv) quantify how controlled chemical modification of the grain boundaries changes the trap distribution and the resulting device characteristics; and (v) adjudicate between the GB-thermionic picture and the competing distributed tail-state/interface-trap (multiple trapping and release) interpretations of the same temperature-dependent data. All five parts (i)–(v) are required, and each is a separately resolvable sub-question.

References

  1. F. M. Hossain, J. Nishii, S. Takagi, A. Ohtomo, T. Fukumura, H. Fujioka, H. Ohno, H. Koinuma, M. Kawasaki, 'Modeling and simulation of polycrystalline ZnO thin-film transistors', Journal of Applied Physics 94 (12), 7768–7777 (2003). DOI: 10.1063/1.1628834, https://doi.org/10.1063/1.1628834
  2. P. F. Carcia, R. S. McLean, M. H. Reilly, G. Nunes, Jr., 'Transparent ZnO thin-film transistor fabricated by rf magnetron sputtering', Applied Physics Letters 82 (7), 1117–1119 (2003). DOI: 10.1063/1.1553997, https://doi.org/10.1063/1.1553997
  3. F. Torricelli, J. R. Meijboom, E. Smits, A. K. Tripathi, M. Ferroni, S. Federici, G. H. Gelinck, L. Colalongo, Z. M. Kovács-Vajna, D. de Leeuw, E. Cantatore, 'Transport Physics and Device Modeling of Zinc Oxide Thin-Film Transistors Part I: Long-Channel Devices', IEEE Transactions on Electron Devices 58 (8), 2610–2619 (2011). DOI: 10.1109/TED.2011.2155910, https://doi.org/10.1109/TED.2011.2155910
  4. F. Torricelli, E. C. P. Smits, J. R. Meijboom, A. K. Tripathi, G. H. Gelinck, L. Colalongo, Z. M. Kovács-Vajna, D. M. de Leeuw, E. Cantatore, 'Transport Physics and Device Modeling of Zinc Oxide Thin-Film Transistors—Part II: Contact Resistance in Short Channel Devices', IEEE Transactions on Electron Devices 58 (9), 3025–3033 (2011). DOI: 10.1109/TED.2011.2159929, https://doi.org/10.1109/TED.2011.2159929
  5. E. M. Likovich, R. Jaramillo, K. J. Russell, S. Ramanathan, V. Narayanamurti, 'Scanning tunneling microscope investigation of local density of states in Al-doped ZnO thin films', Physical Review B 83, 075430 (2011). DOI: 10.1103/PhysRevB.83.075430, https://doi.org/10.1103/PhysRevB.83.075430
  6. P. Thapaliya, W. Lu, R. Jha, 'Electrical Characteristics of RF Sputtered ZnO/HfO2 Interfaces in Transparent Thin Film Transistors', MRS Online Proceedings Library 1792, article mrss15-2137545 (2015). DOI: 10.1557/opl.2015.648, https://doi.org/10.1557/opl.2015.648
  7. P. B. Pillai, M. M. De Souza, 'Transport mechanism in sub 100 °C processed high mobility polycrystalline ZnO transparent thin film transistors', 2015 IEEE International Electron Devices Meeting (IEDM), 28.1.1–28.1.4 (2015). DOI: 10.1109/IEDM.2015.7409785, https://doi.org/10.1109/IEDM.2015.7409785
  8. J.-L. Wu, H.-Y. Lin, P.-H. Kuo, B.-Y. Su, S.-Y. Chu, Y.-C. Chen, S.-Y. Liu, C.-C. Chang, C.-J. Wu, 'Effect of UV-Ozone Treatment on the Performance of ZnO TFTs Fabricated by RF Sputtering Deposition Technique', IEEE Transactions on Electron Devices 61 (5), 1403–1409 (2014). DOI: 10.1109/TED.2014.2309636, https://doi.org/10.1109/TED.2014.2309636
  9. A. Tanpure, M. J. Alam, P. Murkute, Sushama Sushama, J. Saha, S. Chakrabarti, 'Grain boundary modeling and simulation of ZnO thin film transistor and its correlation with UV ozone annealing', Proceedings of SPIE 11092, 80 (2019). DOI: 10.1117/12.2530325, https://doi.org/10.1117/12.2530325
  10. N. Soufyane, N. Sengouga, M. Labed, A. Meftah, 'Temperature Dependent Poly Crystalline Zinc Oxide Thin Film Transistor Characteristics', Transactions on Electrical and Electronic Materials 22 (5), 711–716 (2021). DOI: 10.1007/s42341-021-00288-2, https://doi.org/10.1007/s42341-021-00288-2
  11. M. Yoon, 'Subthreshold Conduction of Disordered ZnO-Based Thin-Film Transistors', Micromachines 14 (8), 1596 (2023). DOI: 10.3390/mi14081596, https://doi.org/10.3390/mi14081596
  12. R. A. Chapman, R. A. Rodriguez-Davila, I. Mejia, M. Quevedo-Lopez, 'Nanocrystalline ZnO TFTs Using 15-nm Thick Al2O3 Gate Insulator: Experiment and Simulation', IEEE Transactions on Electron Devices 63 (10), 3936–3943 (2016). DOI: 10.1109/TED.2016.2597284, https://doi.org/10.1109/TED.2016.2597284
  13. Y.-M. Zhou, Y.-G. He, A.-X. Lu, Q. Wan, 'Simulation of grain boundary effect on characteristics of ZnO thin film transistor by considering the location and orientation of grain boundary', Chinese Physics B 18 (9), 3966–3969 (2009). DOI: 10.1088/1674-1056/18/9/057, https://doi.org/10.1088/1674-1056/18/9/057
  14. M. Wang, Z. Gong, 'An Explicit Analytical Solution to the Grain Boundary Barrier Height in Undoped Polycrystalline Semiconductor Thin-Film Transistors', IEEE Transactions on Electron Devices 61 (6), 2078–2084 (2014). DOI: 10.1109/TED.2014.2318709, https://doi.org/10.1109/TED.2014.2318709
  15. S. Jaiswal, J. Singh, M. Goswami, K. Kandpal, 'Impact of deep and tail trap states on the electrical performance of double-gate ZnO thin film transistors', Semiconductor Science and Technology 40 (1), 015003 (2025; published online 4 December 2024). DOI: 10.1088/1361-6641/ad95b6, https://doi.org/10.1088/1361-6641/ad95b6
  16. S. Singh, S. Jaiswal, M. Goswami, K. Kandpal, 'Grain boundary-induced threshold voltage shift in dual-gate ZnO TFTs: an analytical and simulation approach', Journal of Computational Electronics 24 (6), 186 (2025). DOI: 10.1007/s10825-025-02428-w, https://doi.org/10.1007/s10825-025-02428-w
  17. M. A. Domínguez, S. Alcantara, S. Soto, 'Physically-based simulation of zinc oxide thin-film transistors: Contact resistance contribution on density of states', Solid-State Electronics 120, 41–46 (2016). DOI: 10.1016/j.sse.2016.03.006, https://doi.org/10.1016/j.sse.2016.03.006