High performance enhancement-mode thin-film transistor with graphene quantum dot-decorated In₂O₃ channel layers

Abstract

Due to the quantum confinement and edge effects, there has been ongoing enthusiasm to provide deep insight into graphene quantum dots (GQDs), serving as attractive semiconductor materials. To demonstrate the potential applications of GQDs in electronic devices, this work presents solution-processed high performance GQD-decorated In₂O₃ thin-film transistors (TFTs) based on ZrO₂ as gate dielectrics. GQDs-In₂O₃/ZrO₂ TFTs with optimized doping content have demonstrated high electrical performance and low operating voltage, including a larger field-effect mobility (μ_FE) of 34.02 cm² V⁻¹ s⁻¹, a higher I_on/I_off of 4.55 × 10⁷, a smaller subthreshold swing (SS) of 0.08 V dec⁻¹, a lower interfacial trap states (D_it) of 5.84 × 10¹¹ cm⁻² and threshold voltage shift of 0.07 V and 0.12 V under positive bias stress (PBS) and negative bias stress (NBS) for 3600 s, respectively. As a demonstration of complex logic applications, a resistor-loaded unipolar inverter based on GQDs-In₂O₃/ZrO₂ has been built, demonstrating full swing characteristic and high gain of 10.63. Low-frequency noise (LFN) characteristics of GQDs-In₂O₃/ZrO₂ TFTs have been presented and it was concluded that the noise source can be attributed to the fluctuations in mobility. As a result, it can be concluded that solution-derived GDQ-optimized oxide-based TFTs will manifest potential applications in electronic devices.

---

Introduction

Graphene quantum dots (GQDs) are derivatives of graphene with nanometer size, and have common characteristics of graphene and quantum dots.¹ They have a non-zero bandgap because of quantum confinement,² resulting in their unique properties such as large surface area, excellent solubility, edge effects etc.^3–5 As a result, GQDs are becoming promising semiconductor nanomaterials for applications in biological imaging,⁶ ultrasensitive sensors,⁷ supercapacitors,^8,9 and electronic and opto-electronic devices.^3,10–13

In₂O₃, a representative of transparent n-type semiconductors with a wide band gap of 3.6 eV, is one of the most widely used materials in various fields ranging from optoelectronic devices to thin film transistors (TFTs) due to its lower processing temperature, environment friendliness and good optoelectronic properties.^14–16 Due to its potential application in next-generation display devices, In₂O₃-based oxide TFTs have attracted more attention.^17–19 Even though the attracting properties of In₂O₃ TFTs, but there exist too much oxygen vacancies (V_O) in In₂O₃ thin films, leading to the more formation of more electrons and the deteriorated devices performance.²⁰ It has reported that V_O-related defects and superabundant mobile electrons will result in higher off-current (I_off) and poor bias stress stability.^19,21

With the aim of improving the electrical stability, great progress has been made in incorporating carrier suppressor dopants into the channel layers of the In₂O₃-based TFTs. To date, In₂O₃ films doped by gallium,²² zinc,²³ tin,²⁴ yttrium,²⁵ and magnesium²⁶ have been achieved to optimize the TFTs performance. However, their preparations and applications are limited because these doped elements are usually scarce and toxicant. Due to their unique electron transport properties, GQDs may play a key role in In₂O₃-based TFTs devices.¹¹ Based on Shan's investigation, it can be noted that TFTs modified by GQDs are desirable.^27,28 In spite of the attractive electrical properties, the performance improvement mechanism of GQDs decorated In₂O₃-based TFTs is still unclear and requires further study. Therefore, based on the research perspective and further technology application, developing an effective and simple method to prepare GQDs-decorated TFTs is urgently needed. Up to now, various methods have been adopted to fabricate oxide-based TFTs, such as atomic layer deposition (ALD) and sputtering, which highly depend on vacuum-based ambient. What's more, the complicated process and high costs of fabrication inhibit its potential application. By contrast, solution-based process has been paid more attention due to its advantages, including easy operation, low cost, easy controllability, and large-scale productivity. Therefore, great efforts have been made in preparing solution-driven oxide TFTs devices.^29–31

In this work, based on a solution-based route, a simple and feasible way has been pursued to synthesize GQDs-decorated In₂O₃ thin films. Furthermore, a systematical investigation was conducted to reveal the variation in microstructure and component of GQDs-decorated In₂O₃ films with different doping level. To demonstrate the potential possibility of GQDs in CMOS logics, their applications in GQD-In₂O₃ TFTs and resistor-loaded inverters based on solution-processed ZrO_x gate dielectrics by device engineering were also demonstrated, and then evaluated carefully. Experiment results have revealed that the optimized GQD-In₂O₃ TFTs and resistor-loaded inverter with full swing characteristics have been achieved. The improvement in performance of In₂O₃-based TFTs by GQDs doping engineering has been discussed in detail.

Experimental

GQDs were synthesized by our reported method.³² The p-type heavily doped Si wafers were chosen as the substrates. Firstly, 0.1 M In₂O₃ and 0.1 M ZrO_x solution was obtained by dissolving 0.3 g In(NO₃)₃·xH₂O (99%) and 0.64 g ZrOCl₂·8H₂O (99%) in 10 mL 2-methoxyethanol (2-ME). Meanwhile, 1 mg, 2 mg and 3 mg GQDs were mixed with the In₂O₃ precursor solution, respectively. All the solution were subjected to 8 hour agitation and 12 hour aging. Then, the ZrO_x precursor solution was spun thrice on the Si wafers under speed of 4500 rpm and baked at 200 °C for 10 min. After UV treatment for 30 min, the as-deposited films experienced 600 °C annealing treatment for 2 h. Moreover, the pure In₂O₃ and GQDs-decorated In₂O₃ precursor solution was spun once on the ZrO_x dielectric layer under speed of 2500 rpm and baked at 150 for 10 min, respectively. The samples were annealed at 280 °C in air ambient for 1 h. Finally, 100 nm thick Al thin film was deposited by thermal evaporation to form source and drain electrodes by a shadow mask with width of 1000 μm and length of 200 μm. TFTs with pure In₂O₃ channel layer were fabricated in the same way. For convenience, the corresponding samples were denoted as GQDs-In₂O₃-1, GQDs-In₂O₃-2, and GQDs-In₂O₃-3 with the concentration of GQDs of 0.1 mg ml⁻¹, 0.2 mg ml⁻¹, and 0.3 mg ml⁻¹, respectively. The crystal structures of GQDs-In₂O₃ and In₂O₃ thin films spin-coated on Si substrates were investigated by X-ray diffractometer (XRD, Empyrean S3, Netherlands) with a Cu Kα radiation. The surface morphologies of the thin films were investigated by atomic force microscopy (AFM, 5500M, Hitachi, Japan). X-ray photoelectron spectroscopy (XPS) analyses were carried out by using ESCALAB 250Xi system (Thermo Fisher Scientific, USA) with Al Kα radiation (1486.6 eV). To study the thermal behavior of the precursor solution in air, a thermogravimetric analyzer (TGA, STA449F3) was carried out at a heating rate of 10 °C min⁻¹. The UV visible spectroscopy (UV-Vis-2550, Shimadzu) was used to measure the optical properties of GQDs-In₂O₃ and In₂O₃ thin films spin-coated on quartz substrates. The cross-sectional images of the GQDs-In₂O₃ TFT device were obtained using transmission electron microscope (TEM, JEOL-F200). The electrical performances of the integrated TFTs and resistor-loaded unipolar inverter were measured under an air ambient by a semiconductor parameter analyzer (Agilent B1500A and Keithley 2636B) in a dark box. The low frequency noise (LFN) characteristics of the GQDs-In₂O₃ and In₂O₃ TFTs was detected by semiconductor parameter analyzer (PDA FS-380). The field-effect mobility (μ_FE) of the TFTs was calculated by the following formula:where W is the channel width, L is the channel length, V_GS is the source-gate voltage, I_DS is the saturation current, and C_i is the areal capacitance of the ZrO_x dielectric layer. The interface states density (D_it) can be inferred using the following equation,where k, T, SS, and q are the Boltzman's constant, the absolute temperature, the subthreshold swing and the charge quality, respectively.

Results and discussion

To investigate the thermal behavior of In₂O₃ precursor powders with different GQDs dopingo concentrations, corresponding measurements were performed using TGA under nitrogen ambient conditions, and the result is depicted in Fig. 1. As shown in Fig. 1(a), three steps of weight loss occurring from 50 °C to 390 °C have been observed. The first step of weight loss from 50 °C to 150 °C can be attributed to the evaporation of the residual solvent from the precursor powder.³³ The second weight loss occurs between 150 °C and 260 °C, In(NO₃)₃ starts to decompose and turn into In(OH)₃, indicating the decomposition, hydrothermal, and the thermal dihydroxylation reaction of In(NO₃)₃. According to the previous reports, it can be noted that In(OH)₃ can be transformed into orthorhombic InOOH with metatable state, an intermediate product. It is worth noting that the weight decreases abruptly between 260 °C to 390 °C. This stage represents the dehydroxylation and thermal decomposition of the intermediate products of InOOH, leading to the formation of In₂O₃. After 390 °C, no evident change in the weight loss has been detected, indicating that the In₂O₃ thin films are almost completely formed. Current observations is in good agreement with previous reports that crystallization of the In₂O₃ can be ultimately achieved above 390 °C.^34,35 After doping by GQDs, the reduction in the transfer temperature has been found, suggesting that GQDs doping can reduce the temperature of In₂O₃ film preparation. This is consistent with previous reports that graphene-like materials are accompanied by strong gas release, resulting in rapid thermal expansion.³⁶

Fig. 1 (a) Thermal behavior of the pure In₂O₃ and GQDs-In₂O₃ precursor powders. (b) XRD patterns of the In₂O₃ and GQDs-In₂O₃ thin films.

As depicted in Fig. 1(b), the variation in the microstructure of In₂O₃thin films with different GQDs doping concentrations was investigated by XRD. No characteristic peaks of In₂O₃ and GQDs-In₂O₃ thin films are observed, regardless of the GQDs doping concentrations, meaning that all the films are in an amorphous state. It is well known that the amorphous films have the advantage of large area uniformity, smooth surface, high stability, low interfacial state density, which are very beneficial to improve the performance of TFT devices.³⁷

To verify the potential application of In₂O₃ and GQDs-In₂O₃ thin films in the field of transparent electronic devices, the all the films spin-coated on quartz substrates were measured by UV-Vis. As depicted in Fig. 2(a), it can be seen that all the thin films are highly transparent with an average optical transmittance of over 90% in the visible range, demonstrating its feasibility in transparent electronic devices. In addition, the transmittance of the GQDs-In₂O₃ samples increases slightly as the GQDs concentration increases, which can be attributed to the strong photoluminescence of GQDs.^38,39 Based on Tauc plot method, the optical bandgap for the GQDs-In₂O₃ thin films as a function of GQDs doping concentrations were determined and displayed in Fig. 2(b). Based on Fig. 2(b), it can be seen that the bandgap of In₂O₃ thin films increases from 3.77 eV to 4.15 eV with GQDs content increasing. As reported, the increased bandgap after GQDs doping is associated with the decreased grain size of thin films and the bund gap of GQDs due to its quantum confinement.^2,40,41

Fig. 2 (a) Optical transmittance and (b) bandgap of In₂O₃ thin films with different GQDs concentrations.

To understand the evolution of the chemical bonding states of In₂O₃ thin film as function of GQDs doping concentrations, XPS measurements were carried out and displayed in Fig. 3. All the XPS spectra deconvolution was performed by Shirley background subtraction using a Voigt function convoluting Gaussian and Lorentzian functions. Fig. 3(a) shows the O 1s XPS spectra for the In₂O₃ and GQDs-In₂O₃ samples and the binding energies of the O 1s peaks are deconvoluted into three subpeaks centered at 529.9, 531, and 531.8 eV, which can be assigned to the oxygen in oxide lattice (M–O), the oxygen in oxygen vacancies (V_O), and the oxygen in hydroxyl species (M–OH), respectively.^42,43 Compared with In₂O₃ sample, the XPS feature of GQDs-In₂O₃ samples shifts toward the higher binding energy due to the M–OH oxygen atoms are less negatively charged than those in oxides.⁴⁴ Based on the XPS results, the fraction of oxygen-related components in In₂O₃ and GQDs-In₂O₃ samples was quantitatively discussed and illustrated in Fig. 3(b). It is noteworthy that with increasing GQDs concentration from 0 to 0.3 mg ml⁻¹, the fraction of M–O contents increased from 40.0% to 48% and the fraction of M–OH species from 33% to 27% have been observed, which indicates that GQDs doping helps to convert the M–OH to M–O content. The V_O composition increases from 27% to 28% and then decreases to 25%. This phenomenon is caused by insufficient GQDs element incorporated into In₂O₃, so there is not enough GQDs to inhibit the oxygen vacancy caused by In₂O₃. Which is consistent with our previous investigations.³³ Fig. 3(c) demonstrates the high-resolution C 1s peaks of the GQDs-In₂O₃ samples with different GQDs doping concentrations, which can be deconvoluted into four peaks corresponding to the C–C bond in aromatic rings (284.8 eV), C–N bond (285.2 eV), C–O bond (286.4 eV), and C=O bond (288.6 eV).^32,45 Based on the evolution of the fraction of carbon-related components (Fig. 3(d)), it is found that the contents of C–O and C=O remain relatively stable and no obvious change has been detected. Meanwhile, the fraction of the C–C slightly decreases and C–N species slightly increases with increasing the GQDs concentration.

Fig. 3 XPS spectra of (a) O 1s, (c) C 1s, and semiquantitative analyses of (b) oxygen-related, (d) carbon-related components in In₂O₃ thin films with different GQDs concentration.

To confirm the possibility of solution-driven ZrO_x dielectric thin film annealed at 600 °C as the dielectric layer, Al/ZrO_x/Si MOS capacitor devices were fabricated and the dielectric properties and leakage current behavior were analyzed by an impedance analyzer. Fig. 4 displays the areal capacitance as a function of frequency in the range from 20 to 100 kHz. As shown in Fig. 4(a), the areal capacitance of ZrO_x thin films annealed at 600 °C is measured to 301 nF cm⁻² at 20 Hz and the calculated dielectric constant is 5.78. Compared to some reported values, the larger area capacitance at 20 Hz has been found, indicating the formation of ZrO_x gate dielectric thin film with high quality. For field-effect devices, it is well-known that the dielectric with large area capacitance can induce more carriers at the interface between the semiconductor and the dielectric, thereby reducing the operating voltage. Meanwhile, a weaker frequency dispersion of capacitance for 600 °C-annealed ZrO_x films has been observed, indicating the formation of denser metal-oxide bonding and much lower defect densities in ZrO_x films such as hydroxyl group or oxygen vacancies in thin films. This will be of benefit to reduce the leakage current because the conduction paths in dielectrics are mainly caused by hydroxyl and grain boundaries. Fig. 4(b) shows the leakage current characteristics of MOS capacitors and a relatively low leakage current density (7.5 × 10⁻⁸ A cm⁻² at 2 V), which is comparable to that of SiO₂, has been found. The lower leakage current density comes from the gradually decomposition of residuals and reduction in the defect density. The dielectric layers possessing sufficiently low leakage current ensures that TFT devices have low OFF-state current and can operate at low voltage. AFM morphology of ZrO_x dielectric films annealed at 600 °C is shown in the inset of Fig. 4(b) and RMS of 0.58 nm has been observed. Generally, the preferred smooth surface is conducive to the rapid transmission of carriers in narrow channel without suffering from electronic scattering and the formation of lower leakage current. Indeed, the smooth surface results from the reduction of the interface charge traps and dangling bonds, leading to the production of an excellent interface between the channel layer and the dielectric layer, which will contribute to high performance TFTs. Moreover, the achievement of dielectrics with small RMS will be of benefit to obtain high quality channel layer.⁴⁶

Fig. 4 (a) Frequency-dependent areal capacitance of the ZrO_x thin films annealed at 600 °C. (b) Leakage current density of the ZrO_x thin films annealed at 600 °C. Inset shows the AFM image of ZrO_x thin film annealed at 600 °C.

The cross-sectional transmission electron microscopy (TEM) images of the complete devices, for example, with GQDs-In₂O₃-2 TFT shown in Fig. 5(a), confirm the expected thickness of each layer. The images also reveal excellent homogeneity of the films with uniform thickness. The scanning transmission electron microscopy (STEM) images obtained with elemental mappings (Fig. 5(b)–(f)) show high distribution uniformity of the elements with clear demarcation across the layers. All images show a clear boundary case, with no obvious interfaces spreading out or separating from each other, which helps to minimize the trapped charge. Based on STEM-ABF images shown in Fig. 5(h) and (i), it can be noted that many quantum dots are distributed in In₂O₃ films. To confirm the exact content of the QDs, the HRTEM images of the enlarged selected area marked with a red box, displayed in Fig. 5(k), shows that the materials have clear discernible lattice structures and the C–C bond is close to the ideal C–C bond for graphite.⁴⁷ According to 2D fast Fourier transform (FFT) patterns (Fig. 5(i)), the esthetic hexagonal honeycomb structure with the bond lengths of 1.42 Å for C–C/C=C has been observed,⁴⁸ revealing that the GQDs have been successfully incorporated into In₂O₃ thin films.

Fig. 5 (a) (g) (j) Cross-sectional and zoomed-in area HRTEM images of the TFT with GQDs-In₂O₃/ZrO_x/Si gate dielectrics. (b–f) Elemental mapping images across the cross section of the GQDs-In₂O₃-2 TFT. (b) Cross-sectional and zoomed-in area HRTEM images of the TFT. Inset, the FFT pattern. (h–i) STEM-ABF images of the GQDs-In₂O₃-2 TFT. (k) HRTEM images of the enlarged selected area marked with a red box. (l) The corresponding 2D FFT images.

To investigate the feasibility of the GQDs-In₂O₃ thin films as the channel layers in TFTs, In₂O₃/ZrO_x and GQDs-In₂O₃/ZrO_x TFTs with bottom-gate and top-contact architectures were fabricated and shown in Fig. 6. The typical output characteristic curves of the In₂O₃/ZrO_x and GQDs-In₂O₃-2/ZrO_x TFTs at a V_GS of 1.5 V are depicted in Fig. 6(a) and (b). It can be noted that these devices represent typical n-channel conduction behavior with clear pinch-off voltage and current saturation. Compared to GQDs-In₂O₃-2/ZrO_x TFT, the output curves of In₂O₃/ZrO_x TFT demonstrate the shifting trend towards right side (Fig. 6(a)). In addition, the higher gate voltage V_GS is, the more the output curve right shifts, which can be attributed to the fact that the ohmic contact between Al electrodes and GQDs-In₂O₃-2 thin films is much better than that of In₂O₃ thin films. It can be observed that GQDs-In₂O₃-2 TFT represents a saturation current of 28.2 μA at 2.0 V, which is nearly 6 times larger than that of In₂O₃/ZrO_x TFT. The obvious increase of on-current level of TFTs doped by GQDs can result from the relatively higher mobility of GQDs-In₂O₃ layers compared to In₂O₃ channel layers. The GQDs-embedded In₂O₃ TFTs demonstrate reasonably high on-current characteristics and we expect them to show stable unipolar switching characteristics. Fig. 6(c) shows the transfer characteristics of the In₂O₃ TFT in comparison with those of the three reference TFTs with different GQDs doping concentrations. The current between the source and the drain was measured at a static drain-source voltage of 2 V while the gate voltage changed from −1 V to 4 V.

Fig. 6 Output cure of (a) In₂O₃ TFT and (b) GQDs-In₂O₃-2 TFT. (c) Transfer characteristics of the In₂O₃ and GQDs-In₂O₃ TFTs. (d) Square-root of the transfer characteristics of the In₂O₃ and GQDs-In₂O₃ TFTs. The insets in (a) and (b) show the schematic illustration of the bottom-gate and top-contact In₂O₃-based TFTs.

To investigate the evolution of the electrical properties of the In₂O₃ TFTs as a function of GQDs doping concentrations, key electrical parameters, including the field-effect mobility (μ_FE) and the threshold voltage (V_TH), are calculated from the slope of I_DS^1/2 vs. V_GS. All extracted TFTs performance parameters as a function of GQDs doping concentration are summarized in Table 1. It has been found that compared with pure In₂O₃ TFT, the μ_FE of GQDs-In₂O₃ TFTs increases obviously. It has been reported that the carrier concentration of oxide semiconductors can be increased obviously with C incorporation.⁴⁹ The increase of the μ_FE of GQDs-In₂O₃ TFTs can be attributed to the higher electron mobility of graphene.^50,51 However, excessive doping level of GQDs leads to the reduced μ_FE. As a result, it can be concluded that GQDs plays an important role in improving the mobility of TFTs device and the doping level should be controlled accurately. The extracted changes in V_TH reveals that increasing the doping level of GQDs will lead to the increased positive V_TH. This case can be attributed to the fact that the excess carrier concentration of In₂O₃ will be compensated by GQDs and remarkably shift V_TH towards the positive values necessary for enhancement mode (E-mode) device characteristics has been observed. Based on Table 1, it can be also noted that compared to In₂O₃ TFTs, GQDs-In₂O₃ TFTs has a higher I_on/I_off, which can be due to the increased I_on and the decreased I_off. With increasing the GQDs doping level, the oxygen vacancy and free carrier concentration decrease. As a result, I_on/I_off initially increases because of the reduced I_off. However, further increasing the GQDs doping concentration leads to the degradation of the device performance, resulting from the increased trap states near the GQDs-In₂O₃/ZrO_x TFTs interface.

Table 1 Electrical parameters of In₂O₃ TFT and GQDs-In₂O₃ TFTs

Samples	μFE (cm^2 V^−1 s^−1)	Ion/Ioff	SS (V dec^−1)	Dit (cm^−2)	VTH (V)
In2O3	19.46	2.57 × 10^6	0.11	1.60 × 10^12	0.92
GQDs-In2O3-1	24.26	4.40 × 10^6	0.10	1.11 × 10^12	0.98
GQDs-In2O3-2	34.02	4.55 × 10^7	0.08	5.84 × 10^11	1.22
GQDs-In2O3-3	28.04	5.79 × 10^6	0.12	2.03 × 10^12	0.81

---

Another electrical parameter, the subthreshold swing (SS), is calculated to be 0.11, 0.10, 0.08, and 0.12 V dec⁻¹ for In₂O₃-based TFTs with different GQDs doping level. As we know, the SS value depends on the traps located in channel/dielectric interface and the lower SS can be expected to achieve a high operation speed and low power consumption. The small SS value of 80 mV dec⁻¹ for GQDs-In₂O₃-2 sample is close to the theoretical limit (60 mV dec⁻¹), which can originate from the large areal capacitance of the ZrO_x dielectric layer and the electronic–clean interface between the channel layer and ZrO_x.⁵² Based on the extracted SS, the D_it values of all In₂O₃-based TFTs with different GQDs doping content have been determined and listed in Table 1. Compared to In₂O₃-TFT, the reduction in D_it values for GQDs-In₂O₃ TFTs has been observed, indicating the reduction of the defect states and the formation of high quality GQDs-In₂O₃/ZrO_x interface. With its large critical current density and high saturation speed, graphene has great potential for high-speed electronics.⁵³ Appropriate GQDs doping may reduce the defect density at the interface between the channel layer and the gate dielectric layer, resulting in its small values of SS and D_it for GQDs-In₂O₃ TFT. The observed smallest D_it in GQDs-In₂O₃-2 is not only beneficial to carrier transport in the interface region, but also to the operation stability. However, for GQDs-In₂O₃-3 TFTs, the sharp increased D_it has been observed, which is in agreement with the evolution of the SS value. Excessive GQDs doping into In₂O₃ will introduce a large number of oxygen-containing functional groups, which will act as the trap states and trap mobile carriers, leading to the degraded the performance parameter of TFTs.

Based on above analyses, it can be concluded that compared with pure In₂O₃ TFT, solution-derived GQDs-In₂O₃-2 TFT exhibits impressive enhancement-mode device performance with a relatively smaller V_TH of 1.22 V, an extraordinary on/off current ratio (>10⁷), and a higher peak μ_FE of 32.02 cm² (V⁻¹ s⁻¹). These excellent electrical properties are due to the structural flexibility of graphene.⁵⁴ To our best knowledge, such a mobility is the highest compared with previous reports on In₂O₃-based TFTs via doping engineering route in recent years (indicated in Table 2). All these excellent device performances have guaranteed the potential application of GQDs-In₂O₃ TFTs in e-book and TFT-LCD. Particularly, all these parameters were obtained at a low operating voltage of 4 V, which is significant for low-power and green electronics.

Table 2 Recent advances in In₂O₃-based TFTs doped by different element

Samples	Mobility (cm^2 V^−1 s^−1)	Ion/Ioff	SS (V dec^−1)	Year	Ref.
Zn–In2O3	14.7	10^7	0.1	2015	55
W–In2O3	15.3	10^7	0.07	2016	17
Mg–In2O3	13.7	10^7	0.85	2017	26
Y–In2O3	14.4	10^6	0.41	2019	20
Ca– In2O3	15.1	10^7	0.08	2020	56
C–In2O3	20.4	10^6	0.37	2021	57
Zr–Al–In2O3	2.3	10^4	1.61	2021	58
GQDs-In2O3	34.02	10^7	0.08	This paper

---

Since the TFT is the basic unit of integrated circuit, the operational stability of the TFT is a key electrical performance parameter, which determines their commercial applications such as the backplane of AMLCDs and AMOLEDs. The threshold voltage shift of TFTs will cause the decrease of display brightness and directly affect the display effect. To investigate the bias stability of the In₂O₃ and GQDs-In₂O₃-2 TFTs, positive and negative bias stress (PBS and NBS) measurements were carried out by applying a constant gate bias of 2.0 V while maintaining source and drain electrodes grounded. Fig. 7(a) and (b) display the transfer characteristics of the In₂O₃ and GQDs-In₂O₃-2 TFTs as a function of the stress time up to 3600 s. It can be seen that the TFT device with pure In₂O₃ channel layer suffers from positive threshold voltage shift (ΔV_TH) of 0.69 V during PBS duration (Fig. 7(a)), indicating that more electrons are trapped near/at the dielectric and In₂O₃ interface. In contrast, the GQDs-In₂O₃-2 TFT demonstrates an improved PBS-induced stability with a smaller ΔV_TH of 0.07 V up to 3600 s (Fig. 7(c)), indicating that appropriate GQDs doping suppressed the trapped charge in the film due to low interface states at the dielectric and channel interface, confirmed by previous D_it characterization. In addition, the observed parallel V_TH shift and negligible change in the SS value indicates that no additional defect states are created at the channel/dielectric interface under bias stressing. Fig. 7(b) and (d) present the evolution of the transfer curves of the In₂O₃ and GQDs-In₂O₃-2 TFTs under NBS condition. The obvious left shift of the transfer curves in NBS has been observed, which may be caused by the release of electrons from oxygen defects in the active layer under the continuous gate bias stress. Compared to pure In₂O₃ TFT, the superior bias stability of the GQDs-In₂O₃-2 TFT device is preserved in cases of NBS testing. The negative V_TH shifts after NBS application is 0.32 V for device with In₂O₃ channel layer (Fig. 7(b)). Conversely, the ΔV_TH value for the GQDs-In₂O₃-2 TFT is reduced to ∼0.12 V (Fig. 7(d)). These results suggest that the charge trapping centers in the dielectric bulk region or ZrO_x/In₂O₃ interfaces can be diminished by appropriate GQDs doing.

Fig. 7 Transfer curves under PBS test of (a) In₂O₃ TFT and (c) GQDs-In₂O₃-2 TFT. Transfer curves under NBS test of (b) In₂O₃ TFT and (d) GQDs-In₂O₃-2 TFT. Positive (e) and negative (f) threshold voltage shift as a function of stress time for In₂O₃ TFTs and GQDs-In₂O₃-2 TFTs.

To confirm the dominant charge trapping mechanism causing the bias stress-induced the ΔV_TH in TFTs, the relationship between the ΔV_TH and the stress time has been summarized in Fig. 7(e) and (f). Results show that the time dependence of ΔV_TH in the In₂O₃-based TFTs device follows a stretched exponential equation, originating from the charge being trapped at the dielectric/channel interface or getting injected into the gate dielectric. The time dependence of ΔV_TH reflected by the stretched exponential model can be described as follows:

where t represents the stress time, ΔV_TH0 is the value of ΔV_TH at infinite time, τ is the constant characteristic trapping time of carriers, and β is the stretched exponential exponent, respectively. The stretched exponential model well describes the measured ΔV_TH, revealing that the charge trapping is the dominant mechanism of ΔV_TH at the ZrO_x/In₂O₃-based TFTs interface. Similar experiment results have been observed in our previous InGaZnO-based TFTs.⁵⁹

To further explain the mechanism of ΔV_TH in the transfer curves, energy band diagrams of the GQDs-In₂O₃ TFTs device under PBS and NBS have been depicted in Fig. 8. When PBS measurement is performed in air atmosphere, excessive electrons will be accumulated in the GQDs-In₂O₃ channel. The surrounding oxygen molecules can capture electrons (acceptor, O₂(g) + e⁻ →O₂(s)) from the conduction band and then adsorb on the TFT device. Therefore, the O₂ adsorption in the channel layer can deplete the electron carriers and the buildup of negative space charge O₂⁻ easily will repel conduction electrons in the GQDs-In₂O₃ channel,⁶⁰ leading to the positive V_TH shift and the bending downwards of the energy band, as shown in Fig. 8(a) (E_f: Fermi energy level, E_c: conduction band, E_v: valence band, E_i: intermediate energy band). During the NBS process, the origination of NBS instability can also be attributed to the moisture absorption mechanism where this phenomenon can be expressed by the reaction equation H₂O(g) + h⁺ → H₂O⁺(s).⁶¹ Owing to the abundance of water molecular diffusing from the air ambient, this reaction will keep towards the right side and produce the mass of free electrons as demonstrated in Fig. 8(b), therefore the negative V_TH shift has been detected. Under NBS, electrons will be depleted near the interface leading the energy band bending upwards.

Fig. 8 Bias instability mechanism for GQDs-In₂O₃ TFT under (a) PBS and (b) NBS.

To investigate the compatibility of In₂O₃-based TFTs devices for circuit integration, their applications in inverter were further explored because the inverters are widely considered as the fundamental element which can be fabricated based on E-mode devices. In this work, the resistance-loaded inverters are constructed based on pure In₂O₃ and GQDS-In₂O₃-2 TFTs and a resistor of 10 MΩ and displayed in the inset of Fig. 9(c) and (d). As demonstrated in Fig. 9(a) and (b), the typical voltage transfer characteristic (VTC) curves at certain voltages (V_DD) from 0.5 to 2.5 V with 0.5 V step have been found, indicating that a full swing of inverter characteristic is resulted with the input and output signals effectively rectified. The corresponding voltage gain (defined as −∂V_OUT/∂V_IN) is found to be 9.31 in maximum when V_DD comes up to 2.5 V (Fig. 9(c)). When replacing In₂O₃ with GDQs-In₂O₃-2, the inverter characteristics can be substantially enhanced with a maximum voltage gain of 10.63 under the same V_DD, as concluded in Fig. 9(e). Current extracted result is obviously superior to the reported inverters constructed by a single N-TFT, and even to some complementary inverters constructed by N-TFT and P-TFT.^62,63 The high voltage gain of inverter benefits from high mobility, low off-state current, reliable stability and small subthreshold swing of TFT device. As reported, graphene plays an important role in the common source inverter due to its advantages such as het small transconductance, large drain conductance, and small back-gate capacitance.⁶⁴ Another important performance parameter, the transition width, which defines the borders of the transition region, is also extracted and show in Fig. 9(f). It can be note that compared to the In₂O₃ TFT, the reduction in transition width has been observed after GQDS doping. In addition, the as-calculated maximum transition width of 0.40 V is lower than those previously reported resistor-loaded inverters based on In₂O₃ (2 V) and InGaZnO (3.8 V) transistors.^65,66 The narrow transition width is expected to improve the responds characteristics of logic circuits.

Fig. 9 VTC curves of (a) In₂O₃ TFT and (b) GQDs-In₂O₃-2 TFT. Voltage gain of the inverter (c) In₂O₃ TFT and (d) GQDs-In₂O₃-2 TFT at different applied voltages. Insets in (c) and (d) are circuit schematic of the resistor-loaded inverters. Gain (e) and transition width (f) curves of In₂O₃ and GQDs-In₂O₃-2 TFTs as a function of VDD.

Finally, low-frequency noise (LFN) measurements were carried out at room temperature for the pristine In₂O₃ and GQDs-In₂O₃-2 TFTs to investigate the reduction of the interface trap density and improvement in the stability of the In₂O₃ TFTs with GQDs doping with a fixed gate overdrive voltage (V_GS − V_TH) of 1.0 V together with V_DS of 2.0 V, as shown in Fig. 10(a). The 1/f noise curves were obtained from the linear region of the curves. Based on the experimental results, it can be noted that it is in full accord with the classical 1/f noise theory.⁶⁷ Observably, compared with the pure In₂O₃ TFT, the normalized drain current noise spectral density (S_ID/I_D²) of GQDs-In₂O₃-2 TFT presents a decreasing trend, indicating that the average trap density and interface trap density for the In₂O₃-based TFT are reduced after GQDs doping. To further confirm the main source of LFN of In₂O₃-based TFTs, normalized S_ID/I_D² properties were estimated and analyzed, as demonstrated in Fig. 10(b). The experimental parameters for S_ID/I_D² measurement are as follows: V_GS − V_TH = (1–0.92) V, V_DS = 2 V, and f = 20 Hz. The analysis of the results shows that the slopes of the curves for the pure In₂O₃ and GQDs-In₂O₃-2 are close to −1, indicating that the noise source mainly results from the fluctuations in mobility.⁶⁸ The evolution law of the normalized drain current noise spectral density full conforms to previous electrical characterizations. As a result, it can be inferred that appropriate doping of GQDs is conductive to reducing trap density and improving the electrical properties and stability of In₂O₃ TFTs.

Fig. 10 (a) Typical LFN curves and (b) normalized S_ID/I_D² with different bias of the In₂O₃ and GQDs-In₂O₃-2 TFTs.

Conclusion

In conclusion, this work reports a novel and effective solution-based method to prepare In₂O₃ TFTs with different GQDs doping amounts. Experimental results have found that GQDs-In₂O₃-2 TFT has exhibited excellent electrical performance, including a large μ_FE of 34.02 cm² V⁻¹ s⁻¹, a high I_on/I_off of 4.55 × 10⁷, a small SS of 0.08 V dec⁻¹ and a small threshold voltage shift of 0.07 V and 0.12 V under PBS and NBS after 3600 s, respectively. On the basis of the GQDs-In₂O₃-2 TFT, inverter with a gain of 10.63 has been successfully constructed, demonstrating full swing characteristics. LFN measurements have concluded that the noise source can come from the fluctuations in mobility. The enhanced electrical performance and reliability of In₂O₃ TFTs can be ascribed to the appropriate incorporation of GQDs into the In₂O₃ channel layer, leading to the reduction of the interface trap density and oxygen vacancy states. All in all, the solution-derived GDQs-decorated oxide-based TFTs are promising for the development of transparent low-cost oxide thin film electronics.

Conflicts of interest

There are no conflicts of interests or patent pending.

Acknowledgements

The authors acknowledge the support from National Natural Science Foundation of China (11774001), Anhui Project (Z010118169).

References

1. J. H. Shen, Y. H. Zhu, X. L. Yang and C. Z. Li, Chem. Commun., 2012, 48, 3686.
2. L. S. Li and X. Yan, J. Phys. Chem. Lett., 2010, 1, 2572.
3. D. Pan, J. Zhang, Z. Li and M. Wu, Adv. Mater., 2010, 22, 734.
4. L. L. Li, J. Ji, R. Fei, C. Z. Wang, Q. Lu, J. R. Zhang, L. P. Jiang and J. J. Zhu, Adv. Funct. Mater., 2012, 22, 2971.
5. Y. Shin, J. Lee, J. Yang, J. Park, K. Lee, S. Kim, Y. Park and H. Lee, Small, 2014, 10, 866.
6. Y. Wang, L. Zhang, R. P. Liang, J. M. Bai and J. D. Qiu, Anal. Chem., 2013, 85, 9148.
7. Z. S. Qian, X. Y. Shan, L. J. Chai, J. J. Ma, J. R. Chen and H. Feng, Biosens. Bioelectron., 2014, 60, 64.
8. W. W. Liu, Y. Q. Feng, X. B. Yan, J. T. Chen and Q. J. Xue, Adv. Funct. Mater., 2013, 23, 4111.
9. W. W. Liu, X. B. Yan, J. T. Chen, Y. Q. Feng and Q. J. Xue, Nanoscale, 2013, 5, 6053.
10. P. O. Anikeeva, J. E. Halpert, M. G. Bawendi and V. Bulovic, Nano Lett., 2007, 7, 2196.
11. L. A. Ponomarenko, F. Schedin, M. I. Katsnelson, R. Yang, E. W. Hill, K. S. Novoselov and A. K. Geim, Science, 2008, 320, 356.
12. H. Tetsuka, R. Asahi, A. Nagoya, K. Okamoto, I. Tajima, R. Ohta and A. Okamoto, Adv. Mater., 2012, 24, 5333.
13. K. A. Ritter and J. W. Lyding, Nat. Mater., 2009, 8, 235.
14. B. Y. Su, S. Y. Chu, Y. D. Juang and S. Y. Liu, J. Alloys Compd., 2013, 580, 10.
15. S. Y. Han, G. S. Herman and C. H. Chang, J. Am. Chem. Soc., 2011, 133, 5166.
16. G. X. Jiang, A. Liu, G. X. Liu, C. D. Zhu, Y. Meng, B. Shin, E. Fortunato, R. Martins and F. K. Shan, Appl. Phys. Lett., 2016, 109, 183508.
17. A. Liu, G. Liu, H. Zhu, B. Shin, E. Fortunato, R. Martins and F. K. Shan, J. Mater. Chem. C, 2016, 4, 4478.
18. U. Ozgur, D. Hofstetter and H. Morkoc, Proc. IEEE, 2012, 98, 1255.
19. S. Parthiban and J. Y. Kwon, J. Mater. Res., 2014, 29, 1585.
20. J. Li, Y. H. Zhou, W. Q. Zhu, J. H. Zhang and Z. L. Zhang, Mater. Sci. Semicond. Process., 2019, 93, 201.
21. L. Lan, N. Xiong, P. Xiao, M. Li, H. Xu, R. Yao, S. Wen and J. Peng, Appl. Phys. Lett., 2013, 102, 242102.
22. T. Hong, H. J. Jeong, H. M. Lee, S. H. Choi, J. H. Lim and J. S. Park, ACS Appl. Mater. Interfaces, 2021, 13, 28493.
23. S. Sugumaran, M. N. Bin Ahmad, M. F. Jamlos, C. S. Bellan, S. Chandran and M. Sivaraj, Opt. Mater., 2016, 54, 67.
24. K. Jang, J. Raja, Y. J. Lee, D. Kim and J. Yi, IEEE Electron Device Lett., 2013, 34, 1151.
25. J. W. Hennek, M. G. Kim, M. G. Kanatzidis, A. Facchetti and T. J. Marks, J. Am. Chem. Soc., 2012, 134, 9593.
26. C. Y. Zhao, J. Li, D. Y. Zhong, C. X. Huang, J. H. Zhang, X. F. Li, X. Y. Jiang and Z. L. Zhang, IEEE Trans. Electron Devices, 2017, 64, 2216.
27. H. H. Zhu, A. Liu, Y. Xu, F. K. Shan, A. Li, J. Wang, W. Yang, C. Barrow and J. Liu, Carbon, 2015, 88, 225.
28. H. H. Zhu, A. Liu, F. K. Shan, W. Yang, W. Zhang, D. Li and J. Liu, Carbon, 2016, 100, 201.
29. J. H. Park, Y. B. Yoo, K. H. Lee, W. S. Jang, J. Y. Oh, S. S. Chae, H. W. Lee, S. W. Han and H. K. Baik, ACS Appl. Mater. Interfaces, 2013, 5, 8067.
30. B. Yang, G. He, Y. Zhang, C. Zhang, Y. Xia, F. Alam and J. Cui, IEEE Trans. Electron Devices, 2019, 66, 3479.
31. W. Wang, G. He, L. Wang, X. Xu and Y. Zhang, IEEE Trans. Electron Devices, 2021, 68, 4437.
32. X. F. Xu, F. H. Gao, X. H. Bai, F. C. Liu, W. J. Kong and M. Li, Materials, 2017, 10, 1328.
33. C. Zhang, G. He, B. Yang, Y. F. Xia and Y. C. Zhang, IEEE Trans. Electron Devices, 2019, 66, 3854.
34. M. C. Nguyen, M. Jang, D. H. Lee, H. J. Bang, M. Lee, J. K. Jeong, H. Yang and R. Choi, Sci. Rep., 2016, 6, 25079.
35. Z. B. Zhuang, Q. Peng, J. F. Liu, X. Wang and Y. D. Li, Inorg. Chem., 2007, 46, 5179.
36. S. Stankovich, D. A. Dikin, R. D. Piner, K. A. Kohlhaas, A. Kleinhammes, Y. Jia, Y. Wu, S. T. Nguyen and R. S. Ruoff, Carbon, 2007, 45, 1558.
37. S. Hong, J. Park, K. Heejun, Y. G. Kim and H. J. Kim, J. Inf. Display, 2016, 17, 93.
38. J. M. Yuan, R. Zhao, Z. J. Wu, W. Li and X. G. Yang, Small, 2018, 14, 1703714.
39. X. Wu, L. Ma, S. Sun, K. Jiang, L. Zhang, Y. Wang, H. Zeng and H. Lin, Nanoscale, 2018, 10, 1532.
40. R. J. Hong, J. B. Huang, H. B. He, Z. X. Fan and J. D. Shao, Appl. Surf. Sci., 2005, 242, 346.
41. S. Neubeck, L. A. Ponomarenko, F. Freitag, A. J. M. Giesbers, U. Zeitler, S. V. Morozov, P. Blake, A. K. Geim and K. S. Novoselov, Small, 2010, 6, 1469.
42. Y. Jeong, K. Song, T. Jun, S. Jeong and J. Moon, Thin Solid Films, 2011, 519, 6164.
43. J. S. Lee, Y. J. Kwack and W. S. Choi, ACS Appl. Mater. Interfaces, 2013, 5, 11578.
44. G. Liu, A. Liu, H. H. Zhu, B. Shin, E. Fortunato, R. Martins, Y. Wang and F. K. Shan, Adv. Funct. Mater., 2015, 25, 2564.
45. Y. Li, Y. Hu, Y. Zhao, G. Shi, L. Deng, Y. Hou and L. Qu, Adv. Mater., 2011, 23, 776.
46. L. Zhu, G. He, J. Lv, E. Fortunato and R. Martins, RSC Adv., 2018, 8, 16788.
47. S. Kim, S. W. Hwang, M. K. Kim, D. Y. Shin, D. H. Shin, C. O. Kim, S. B. Yang, J. H. Park, E. Hwang, S. H. Choi, G. Ko, S. Sim, C. Sone, H. J. Choi, S. Bae and B. H. Hong, ACS Nano, 2012, 6, 8203.
48. F. Joucken, F. Frising and R. Sporken, Carbon, 2015, 83, 48.
49. S. Akbar, S. K. Hasanain, M. Abbas, S. Ozcan, B. Ali and S. I. Shah, Solid State Commun., 2011, 151, 17.
50. K. S. Novoselov, A. K. Geim, S. V. Morozov, D. Jiang, Y. Zhang, S. V. Dubonos, I. V. Grigorieva and A. A. Firsov, Science, 2004, 306, 666.
51. J. Feng, H. Dong, B. Pang, Y. Chen, L. Yu and L. Dong, J. Mater. Chem. C, 2019, 7, 237.
52. L. Zhu, G. He, W. D. Li, B. Yang, E. Fortunato and R. Martins, Adv. Electron. Mater., 2018, 4, 1800100.
53. L. Liao, Y. C. Lin, M. Bao, R. Cheng, J. Bai, Y. Liu, Y. Qu, K. L. Wang, Y. Huang and X. Duan, Nature, 2010, 467, 305.
54. A. H. Castro Neto, F. Guinea, N. M. R. Peres, K. S. Novoselov and A. K. Geim, Rev. Mod. Phys., 2009, 81, 109.
55. F. Shan, A. Liu, G. X. Liu, Y. Meng, E. Fortunato and R. Martins, J. Disp. Technol., 2015, 11, 541.
56. C. Zhang, G. He, W. H. Wang, Y. C. Zhang, Y. F. Xia and B. Yang, IEEE Trans. Electron Devices, 2020, 67, 4238.
57. R. Kobayashi, T. Nabatame, T. Onaya, A. Ohi, N. Ikeda, T. Nagata, K. Tsukagoshi and A. Ogura, Jpn. J. Appl. Phys., 2021, 60, 030903.
58. D. Yao, X. Xiong, X. Fu, Z. Xu, H. Ning, D. Luo, H. Tang, H. Zheng, R. Yao and J. Peng, Surf. Interfaces, 2021, 27, 101459.
59. Y. C. Zhang, Y. J. Lin, G. He, B. H. Ge and W. J. Liu, ACS Appl. Electron. Mater., 2020, 2, 3728.
60. J. K. Jeong, H. W. Yang, J. H. Jeong, Y.-G. Mo and H. D. Kim, Appl. Phys. Lett., 2008, 93, 123508.
61. P.-T. Liu, Y.-T. Chou and L.-F. Teng, Appl. Phys. Lett., 2009, 95, 233504.
62. Y. J. Jeong, T. K. An, D.-J. Yun, L. H. Kim, S. Park, Y. Kim, S. Nam, K. H. Lee, S. H. Kim, J. Jang and C. E. Park, ACS Appl. Mater. Interfaces, 2016, 8, 5499.
63. S. K. Lee, S. M. H. Kabir, B. K. Sharma, B. J. Kim, J. H. Cho and J. H. Ahn, Nanotechnology, 2014, 25, 014002.
64. N. Harada, K. Yagi, S. Sato and N. Yokoyama, Appl. Phys. Lett., 2010, 96, 012102.
65. L. Petti, H. Faber, N. Münzenrieder, G. Cantarella, P. A. Patsalas, G. Troter and T. D. Anthopoulos, Appl. Phys. Lett., 2015, 106, 092105.
66. I. T. Cho, J. W. Lee, J. M. Park, W. S. Cheong, C. S. Hwang, J. S. Kwak, H. Cho, H. I. Kwon, H. Shin and B. G. Park, IEEE Electron Device Lett., 2012, 33, 1726.
67. L. K. J. Vandamme and F. N. Hooge, IEEE Trans. Electron Devices, 2008, 55, 3070.
68. T. C. Fung, G. Baek and J. Kanicki, J. Appl. Phys., 2010, 108, 074518.

---