Impact of layer thickness on the operating characteristics of In2O3/ZnO heterojunction thin-film transistors featured

Item Type Article

Authors Alghamdi, Wejdan S.;Fakieh, Aiman;Faber, Hendrik;Lin, Yen- Hung;Lin, Wei-Zhi;Lu, Po-Yu;Liu, Chien-Hao;Salama, Khaled N.;Anthopoulos, Thomas D.

Citation AlGhamdi, W. S., Fakieh, A., Faber, H., Lin, Y.-H., Lin, W.-Z., Lu, P.- Y., Liu, C.-H., Salama, K. N., & Anthopoulos, T. D. (2022). Impact of layer thickness on the operating characteristics of In2O3/ZnO heterojunction thin-film transistors. Applied Physics Letters, 121(23), 233503. https://doi.org/10.1063/5.0126935

Eprint version Publisher's Version/PDF

DOI 10.1063/5.0126935

Publisher AIP Publishing

Journal Applied Physics Letters

Rights Archived with thanks to Applied Physics Letters under a Creative Commons license, details at: https://creativecommons.org/

licenses/by/4.0/

Download date 2024-01-04 21:38:01

Item License https://creativecommons.org/licenses/by/4.0/

Link to Item http://hdl.handle.net/10754/686232

Impact of layer thickness on the operating characteristics of In ₂ O ₃ /ZnO heterojunction thin-film transistors

Cite as: Appl. Phys. Lett.121, 233503 (2022);doi: 10.1063/5.0126935 Submitted: 18 September 2022

.

Accepted: 16 November 2022

.

Published Online: 5 December 2022

Wejdan S.AlGhamdi,¹ AimanFakieh,² HendrikFaber,¹ Yen-HungLin,³ Wei-ZhiLin,⁴Po-YuLu,⁴ Chien-HaoLiu,⁴ Khaled NabilSalama,² and Thomas D.Anthopoulos^1,a)

AFFILIATIONS

1King Abdullah University of Science and Technology (KAUST), KAUST Solar Center (KSC), Physical Science and Engineering Division (PES), Thuwal 23955-6900, Saudi Arabia

2King Abdullah University of Science and Technology (KAUST), Thuwal 23955-6900, Saudi Arabia

3Department of Electronic and Computer Engineering, The Hong Kong University of Science and Technology, Clear Water Bay, Kowloon, Hong Kong

4Department of Mechanical Engineering, National Taiwan University, Taipei 10617, Taiwan

Note:This paper is part of the APL Special Collection on Metal Oxide Thin-Film Electronics.

a)Author to whom correspondence should be addressed:Thomas.anthopoulos@kaust.edu.sa

ABSTRACT

Combining low-dimensional layers of dissimilar metal oxide materials to form a heterojunction structure offers a potent strategy to improve the performance and stability of thin-film transistors (TFTs). Here, we study the impact of channel layer thicknesses on the operating charac- teristics of In2O3/ZnO heterojunction TFTs prepared via sputtering. The conduction band offset present at the In2O3/ZnO heterointerface affects the device’s operating characteristics, as is the thickness of the individual oxide layers. The latter is investigated using a variety of experimental and computational modeling techniques. An average field-effect mobility (l_FE) of>50 cm²V¹s¹, accompanied by a low threshold voltage and a high on/off ratio (10⁸), is achieved using an optimal channel configuration. The highl_FEin these TFTs is found to correlate with the presence of a quasi-two-dimensional electron gas at the In2O3/ZnO interface. This work provides important insight into the operating principles of heterojunction metal oxide TFTs, which can aid further developments.

VC 2022 Author(s). All article content, except where otherwise noted, is licensed under a Creative Commons Attribution (CC BY) license (http://

creativecommons.org/licenses/by/4.0/).https://doi.org/10.1063/5.0126935

Recently, new dynamics initiated a wide range of research and innovations beyond the traditional shrinkage of feature sizes in com- plementary metal-oxide-semiconductor (CMOS) technology, involv- ing the development of alternative materials and processes. Oxide semiconductors are considered as one of the most integral material families to drive developments in a wide variety of modern applica- tions such as the internet of things (IoT), artificial intelligence (AI), robotics, self-driving cars, and energy harvesting devices.^1–4 High interest in oxide electronics is driven by their success in the display industry⁵and by their attractive properties such as low processing temperatures, scalable manufacturing, optical transparency (>80%), and compatibility with flexible and printed electronics. Since the dem- onstration of high mobility electron gas systems in epitaxially grown AlGaN/GaN and GaAs/AlGaAs,^6,7 the quest for practical

implementations of similar systems with other materials and simpler processing schemes has been ongoing. For example, the existence of 2DEG in oxide-based materials, such as ZnO/ZnMgO, has been widely investigated.^8–11In addition, In₂O₃, ZnO, and Ga₂O₃heterointerfaces have been realized via solution-processing that features a high mobility quasi-two-dimensional electron gas (q-2DEG) formed at the heteroin- terface.^12–14The impact of the channel film thickness on thin-film transistor (TFT) performance has been reported for single layer metal oxides, including ZnO, In2O3, IGZO, and ITGO as well as for the het- erojunction a-IGZO/a-IGZO-high Indium content system.^15–22 In general, these devices’ threshold voltage (VTH) tends to shift to more negative potentials when the channel thickness increases due to the increase in the number of free carriers, whereas the effect of the field effect mobility (lFE) varies with the thickness of single layers, where it

improved in ZnO TFTs with thicker films and showed a degenerate behavior for the In₂O₃film thicker than 11.5 nm, which still deserves more investigation. Fora-IGZO, the mobility remained independent for the single layers and gradually improved for the heterojunction of a-IGZO/(a-IGZO-high In-content) with a thicker top layer of a higher indium content.

Here, we fabricated In2O3/ZnO heterojunction TFTs via sputter- ing and studied the effect of the individual layer thicknesses on the electron distribution within the channel and its impact on the electrical performance of the resulting TFTs. By combining a complementary gamut of material/device characterization and simulation tools, we demonstrated optimally designed In2O3/ZnO heterojunction TFTs with an average electron mobility of>50 cm²V¹s¹. This value is significantly higher than those obtained for TFTs based on the individ- ual In₂O₃and ZnO layers.

We fabricated metal oxide TFTs using heavily doped silicon (Si^þþ) as the bottom gate featuring a 50-nm-thick SiO₂gate dielectric with a geometric capacitance (C_i) of 69.06 nF/cm². The In₂O₃/ZnO heterojunction channel was grown using radio frequency (RF) sputter- ing at a 0.1 A˚ /s deposition rate with a gas-flow ratio of 10:05 and 20:00 Ar:O for In2O3and ZnO, respectively. Both sputtering targets are 3 in.

in diameter, 99.9% purity, 1/8 in. thick, and bonded to a 1/8 in. OFHC backing plate. The gas flow pressure was set to 5 mTorr, and targets were pre-sputtered for 15 min. Each layer was annealed at 200C for 1 h in ambient air following deposition. A 60-nm-thick aluminum layer was thermally evaporated to form the top source and drain elec- trodes (S–D) via a shadow mask. In addition to single layers of In2O3

and ZnO with thicknesses between 5 and 30 nm, different In₂O₃/ZnO heterojunctions with varying layer thicknesses were fabricated and studied [Fig. 1(a)]. Individual oxide layers with thicknesses below 5 nm have not been considered due to the difficulty in growing contin- uous layers. The microstructural properties of the In₂O₃/ZnO hetero- junctions were studied using cross-sectional high-resolution transmission electron microscopy (HR-TEM). As shown inFig. 1(b), the In₂O₃/ZnO heterojunction is polycrystalline with a well-defined interface between the two metal oxides. Analysis of the TEM images reveals that the In2O3layer features a lattice d-spacing of3 A˚ , which corresponds to the [222] crystal plane. This was additionally con- firmed by grazing incident diffraction (GID) [Fig. 1(c)] to be the most dominant plane in the In2O3film. The ZnO layer shows a lattice spac- ing of2.6 A˚ , corresponding to the [101] crystal plane of the hexago- nal wurtzite structure [Fig. 1(c)].

To understand how the thickness of each layer impacts the heter- ojunction characteristics, we modeled the electron density using COMSOL Multiphysics^VR simulation software.Figures 2(a)–2(f)show the simulation results for different In₂O₃/ZnO heterojunction configu- rations fromFig. 1(a)at equilibrium. The various physical parameters, such as Fermi energy level (E_F) and electron density (N_e), used in modeling were adopted from our previous studies.^23,24 Combining In₂O₃and ZnO results in an increase in the electron density across the interface. Interestingly, a higher electron concentration resides in In2O3at the heterointerface’s vicinity. This interface forms a type-II heterojunction where electrons migrate from the conduction band of ZnO to that of In2O3, leading to the formation of a q2DEG.¹²For a more quantitative view of the different situations, we extracted the electron density gradient along the two layers of each sample, starting from the ZnO top surface [following the dashed line inFig. 2(a)].

Figure 3shows the electron density profiles across the In₂O₃/ ZnO heterojunction extracted from the data inFig. 2from the top sur- face of the channel to the bottom. A sharp increase inN_eat the hetero- junction is observed and coincides with the precise position of the heterointerface. The peak value ofNeacross all heterojunctions consid- ered remains comparable and around 2.810¹⁸cm³ [Fig. 3(a)].

Importantly,Nereaches a maximum within the In2O3side [Ne(In)] of the interface. In contrast, the minimum values exist on the ZnO side close to the heterointerface [Fig. 3(b)].

Interestingly, the value ofN_ewithin the bulk of the ZnO layer [N_e(Zn)] increases with increasing ZnO thickness. This is not the case for In2O3, where [Ne(In)] remains relatively constant for all heterojunc- tions considered [Fig. 3(b)]. The electron concentration at the top ZnO surface (depth¼0 nm) and at the bottom of In2O3(i.e., maxi- mum depth), denoted bulk values below, are summarized inFig. 3(c).

While the bulk N_e(In) remains relatively constant, N_e(Zn) steadily increases and reaches a value of 10¹⁸cm³, which is the starting parameter value considered for ZnO in modeling. We attribute this to the electron migration from ZnO to the In2O3layer. In the case of the 5 nm ZnO, this electron transfer process removes the majority of elec- trons andNe(Zn)reaches a minimum value [Fig. 3(c)]. As the thickness of ZnO increases,Ne(Zn)appears to saturate at a maximum value. A similar but opposite effect is observed when the thickness of In2O3is varied with ZnO remaining constant at 5 nm [Fig. 3(d)]. Here,N_e(In) appears to reach its highest value when In₂O₃is 5 nm. This finding is not surprising considering the smaller volume of the ultra-thin In₂O₃ layer, which accepts all electrons transferred from the ZnO’s conduc- tion band. As the In2O3layer gets thicker, the transferred electrons

FIG. 1.(a) Schematic of the transistor architecture and the thickness variations of the studied In2O3/ZnO heterojunction channels. (b) Cross-sectional TEM of an In2O3/ZnO (5/7 nm) heterojunction deposited atop the Si^þþ/SiO2wafer. (c) Grazing incident diffraction (GID) data for 20-nm-thin ZnO and In2O3films.

remain close to the heterointerface and do not diffuse further into the bulk. As a result,Ne(In)remains relatively unaffected.

Next, the electrical characteristics of TFTs based on single layer In2O3, ZnO, and heterojunction In2O3/ZnO channels were measured under an inert atmosphere. Representative transfer characteristics (i.e., drain current, IDvs gate voltage, VG) for all TFTs measured in satura- tion (a drain voltage, V_D, of 20 V) are shown inFig. 4. All devices exhibit an n-channel behavior with varying performance that depends on the channel configuration. The statistical data of the corresponding l_FEand VTHfor both the single layer and heterojunction TFTs are summarized inFig. 5and Tables S1 and S2 (thesupplementary mate- rial). TFTs based on 5-nm-thin In2O3show uniform performance with V_TH3 V andl_FE32 cm²V¹s¹. In transistors featuring thicker In₂O₃channels, the concentration of free electrons increases, which shifts the VTHto11 and beyond20 V for 10 and 30 nm- thick In2O3channels, respectively. It is worth noticing that single-layer In2O3TFTs exhibit a more significant variation in their performance compared to optimized heterojunction devices.

This wide variation in the conductivity in the bulk In2O3has been attributed to the higher concentration of the oxygen vacan- cies.²⁵On the other hand, ZnO single layers show averagel_FEof FIG. 2.Electron density distribution (color bar in units of log[cm³]) of the hetero-

junction channel, modeled at equilibrium for In2O3/ZnO thickness variations of 5/5 (a), 10/5 (b), 30/5 (c), 10/10 (d), 5/10 (e), and 5/30 nm (f). Note that each panel’s height corresponds to the sum of the individual layer thicknesses.

FIG. 3.Extracted electron density profiles across the In2O3/ZnO bilayer heterojunction from top (top ZnO surface) to bottom (bottom of the In2O3layer). (a) The evolution ofNe

across heterojunctions featuring three different ZnO layer thicknesses, while In2O3remains fixed to 5 nm. (b) Evolution ofNeacross three heterojunctions featuring three differ- ent In2O3layer thicknesses, while ZnO remains fixed to 5 nm. (c) and (d) The electron densities in the “bulk” of the individual layers corresponding to the samples considered in (a) and (b), respectively.

1.6 cm²V¹s¹and V_TH5 V for both 5 and 10 nm channel thick- nesses, whilel_FEslightly decreased for a 30 nm ZnO channel due to the increased density of trap states, which is manifested as an increase in the subthreshold swing, SS (Tables S1 and S2). Overall, ZnO single layers show more consistent and uniform results than

In₂O₃[Figs. 5(a)and5(b)], suggesting that ZnO has a higher toler- ance than In₂O₃against ambient exposure. The latter has been quan- titatively studied by Lanyet al., which found that the surface carrier sheet density for In2O3dominates the conductivity of the 150-nm- thick epitaxial film.²⁵

With the exception of the 30/5 nm In₂O₃/ZnO TFTs, all heterojunc- tion devices show an average saturationl_FEin excess of 50 cm²V¹s¹ (maximum value>60 cm²V¹s¹), on/off ratios of10⁸, and negligi- ble operational hysteresis. Higherl_FEvalues have been observed before in the solution-processed In2O3/ZnO TFT^12,13and were attributed to the q-2DEG formed at the In₂O₃/ZnO heterojunction. The peak values ofN_e determined via simulations inFigs. 2and3support the existence of the q-2DEG confined in the vicinity of the heterointerface in agreement with previously reported experimental findings of higherNe.^12,13

The V_THof TFTs based on fixed 5-nm-thin In₂O₃at the bottom and a top layer of ZnO with varying thicknesses of 5, 10, and 30 nm shifts gradually toward a more positive V_G[Fig. 5(b), the bottom plot]

in a similar trend observed for the single layers ZnO-based TFTs [Fig.

5(b), the top plot]. Just like in the case of single-layer TFTs, an increase in the bottom In₂O₃thickness leads to strong V_THshifts toward nega- tive bias in In₂O₃/ZnO heterostructure TFTs. To some extent, this can be counterbalanced by increasing the thickness of ZnO atop. Analysis of the data of TFTs with 10/5 and 10/10 nm In2O3/ZnO channels reveals that the average VTHmoves toward zero. However, these TFTs exhibit wide parameter spreads where the V_THfluctuates fromþ3 to 6 V within a single batch of devices [Fig. 5(b), the bottom plot]. We attribute this to the presence of stoichiometric/structural defects in the thicker oxide layers and to the poorer quality of the formed FIG. 4.The transfer characteristics of (a) the single layer TFTs based on ZnO or

In2O3at different thicknesses and (b) the heterojunction In2O3/ZnO TFTs with differ- ent thickness variations. All devices were measured at VD¼20 V with a channel width W of 1 mm and a channel length L of 30lm.

FIG. 5.(a) Statistical overview of saturation mobility extracted from 10 to 20 devices fabricated in three different batches each with different channel lengths for single layer channel TFTs based on ZnO and In2O3(top plot) and In2O3/ZnO heterojunction TFTs (bottom plot). (b) Corresponding threshold voltage (VTH) of the same set of devices based on single layer channels (In2O3, ZnO) and the In2O3/ZnO heterojunction TFTs.

heterointerface. Higher uniformity is realized at 5/5 and 5/10 nm of In₂O₃/ZnO devices, respectively. The best parameter uniformity is obtained when the thickness of the top ZnO layer is set between 5 and 10 nm. The layer also serves as additional protection against oxygen and moisture in the air and other impurities that could react with the channel and lower its stability.

To summarize, we have shown that the presence of a higher con- centration of free electrons at the In2O3/ZnO interface is related to the high electron mobility measured in these heterostructure TFTs.

Although the electron concentration in the channel in these devices appears to be largely independent of the thicknesses of the In2O3and ZnO layers employed, the threshold voltage was found to be strongly affected, allowing for an optimization of the individual layer thickness.

The optimal metal oxide thicknesses were identified to be 5 nm for In2O3 and between 5 and 10 nm for the ZnO layer atop.

Heterostructure TFTs with an optimized channel design were shown to exhibit the highestl_FEof>50 cm²V¹s¹, an on/off ratio of10⁸, and the lowest device parameter spread.

See thesupplementary materialfor the summary of key operating characteristics of single-layer and heterojunction channel TFTs featur- ing sputtered ZnO and In₂O₃ semiconducting layers of varying thicknesses.

This publication is based upon the work supported by the King Abdullah University of Science and Technology (KAUST) Office of Sponsored Research (OSR) under Award No. OSR-2018-CARF/CCF- 3079. We are grateful for assistance from the KAUST Solar Center and KAUST Core labs. Additionally, we would like to thank Dr. Cheng Sheng Lin of Pitotech Co., Ltd. (Taiwan) for useful suggestions and assistance in device modeling and simulation.

AUTHOR DECLARATIONS Conflict of Interest

The authors have no conflicts to disclose.

Author Contributions

Wejdan S. AlGhamdi:Data curation (lead); Formal analysis (lead);

Investigation (lead); Methodology (lead); Writing – original draft (lead).Aiman Fakieh:Data curation (equal); Formal analysis (equal);

Investigation (equal); Writing – original draft (equal).Hendrik Faber:

Data curation (equal); Formal analysis (equal); Writing – original draft (equal).Yen-Hung Lin:Formal analysis (equal); Investigation (equal);

Methodology (equal).Wei-Zhi Lin:Data curation (equal); Formal anal- ysis (equal); Investigation (equal).Po-Yu Lu: Data curation (equal);

Formal analysis (equal); Investigation (equal).Chien-Hao Liu: Data curation (equal); Formal analysis (equal); Supervision (equal).Khaled N Salama:Funding acquisition (supporting); Supervision (supporting).

Thomas D. Anthopoulos:Conceptualization (lead); Funding acquisi- tion (lead); Project administration (lead); Resources (lead); Supervision (lead); Writing – original draft (lead).

DATA AVAILABILITY

The data that support the findings of this study are available from the corresponding author upon reasonable request.

REFERENCES

1M. Patel, H.-S. Kim, and J. Kim,Adv. Electron. Mater.1(11), 1500232 (2015).

2J. Jiang, Y. Li, J. Liu, X. Huang, C. Yuan, and X. W. Lou,Adv. Mater.24(38), 5166 (2012).

3S. R. Bhalerao, D. Lupo, and P. R. Berger,Mater. Sci. Semicond. Process.139, 106354 (2022).

4H. Zhang, Z. Wang, Z. Wang, B. He, M. Chen, M. Qi, Y. Liu, J. Xin, and L.

Wei,Front. Optoelectron.15(1), 2 (2022).

5M. Batzill and U. Diebold,Prog. Surf. Sci.79(2–4), 47 (2005).

6O. Ambacher, J. Smart, J. R. Shealy, N. G. Weimann, K. Chu, M. Murphy, W.

J. Schaff, L. F. Eastman, R. Dimitrov, L. Wittmer, M. Stutzmann, W. Rieger, and J. Hilsenbeck,J. Appl. Phys.85(6), 3222 (1999).

7T. J. Drummond, W. T. Masselink, and H. Morkoc,Proc. IEEE74(6), 773 (1986).

8T. Edahiro, N. Fujimura, and T. Ito,J. Appl. Phys.93(10), 7673 (2003).

9J. D. Ye, S. Pannirselvam, S. T. Lim, J. F. Bi, X. W. Sun, G. Q. Lo, and K. L.

Teo,Appl. Phys. Lett.97(11), 111908 (2010).

10H. Tampo, H. Shibata, K. Matsubara, A. Yamada, P. Fons, S. Niki, M.

Yamagata, and H. Kanie,Appl. Phys. Lett.89(13), 132113 (2006).

11J. Falson, D. Maryenko, Y. Kozuka, A. Tsukazaki, and M. Kawasaki,Appl. Phys.

Express4(9), 091101 (2011).

12Y. H. Lin, H. Faber, J. G. Labram, E. Stratakis, L. Sygellou, E. Kymakis, N. A.

Hastas, R. Li, K. Zhao, A. Amassian, N. D. Treat, M. McLachlan, and T. D.

Anthopoulos,Adv. Sci.2(7), 1500058 (2015).

13H. Faber, S. Das, Y.-H. Lin, N. Pliatsikas, K. Zhao, T. Kehagias, G.

Dimitrakopulos, A. Amassian, P. A. Patsalas, and T. D. Anthopoulos,Sci. Adv.

3(3), e1602640 (2017).

14D. Khim, Y.-H. Lin, S. Nam, H. Faber, K. Tetzner, R. Li, Q. Zhang, J. Li, X.

Zhang, and T. D. Anthopoulos,Adv. Mater.29(19), 1605837 (2017).

15I. Isakov, H. Faber, A. D. Mottram, S. Das, M. Grell, A. Regoutz, R. Kilmurray, M. A. McLachlan, D. J. Payne, and T. D. Anthopoulos,Adv. Electron. Mater.

6(11), 2000682 (2020).

16Q. Jiang, Y. Yang, Y. Zhang, X. Liu, and Z. Li,Phys. Status Solidi B254(9), 1600648 (2017).

17Y. Caglar, M. Caglar, S. Ilican, S. Aksoy, and F. Yakuphanoglu, J. Alloys Compd.621, 189 (2015).

18Y. Li, Y. L. Pei, R. Q. Hu, Z. M. Chen, Y. Zhao, Z. Shen, B. F. Fan, J. Liang, and G. Wang,Curr. Appl. Phys.14(7), 941 (2014).

19D.-H. Kim, H.-S. Cha, and H.-S. Jeong, S.-H. Hwang, and H.-I. Kwon, Electronics10(11), 1295 (2021).

20A. Alian, M. A. Pourghaderi, Y. Mols, M. Cantoro, T. Ivanov, N. Collaert, and A. Thean, paper presented at the 2013 IEEE International Electron Devices Meeting, Washington, DC, December, 2013.

21D.-Y. Jeon,Jpn. J. Appl. Phys.59(1), 014001 (2020).

22D. Koretomo, S. Hamada, Y. Magari, and M. Furuta, Materials13(8), 1935 (2020).

23E. Yarali, H. Faber, E. Yengel, A. Seitkhan, K. Loganathan, G. T. Harrison, B.

Adilbekova, Y. Lin, C. Ma, Y. Firdaus, and T. D. Anthopoulos,Adv. Electron.

Mater.6(6), 2000028 (2020).

24Y.-H. Lin, Y. Han, A. Sharma, W. S. AlGhamdi, C.-H. Liu, T.-H. Chang, X.-W.

Xiao, W.-Z. Lin, P.-Y. Lu, A. Seitkhan, A. D. Mottram, P. Pattanasattayavong, H. Faber, M. Heeney, and T. D. Anthopoulos,Adv. Mater.34(3), 2104608 (2022).

25S. Lany, A. Zakutayev, T. O. Mason, J. F. Wager, K. R. Poeppelmeier, J. D.

Perkins, J. J. Berry, D. S. Ginley, and A. Zunger,Phys. Rev. Lett.108(1), 016802 (2012).