Performance improvement for printed indium gallium zinc oxide thin-film transistors with a preheating process

Abstract

High performance indium gallium zinc oxide (IGZO) thin-film transistors (TFTs) were fabricated by printing and spin-coating IGZO inks as a semiconductor layer at low temperature annealing. A preheating strategy was developed, which significantly enhanced the performance of IGZO TFTs while the post-annealing temperature was kept constant at 300 °C. It was found that when the temperature of preheating on a hotplate increased from 40 °C to 275 °C, the field effect mobility improved from 0.31 cm² V⁻¹ s⁻¹ to 4.93 cm² V⁻¹ s⁻¹ for printed IGZO TFTs and from 1.44 cm² V⁻¹ s⁻¹ to 7.9 cm² V⁻¹ s⁻¹ for spin-coated IGZO TFTs. The surface roughness of the IGZO films significantly decreased by increasing the preheating temperature from 40 °C to 95 °C. In addition, the analysis of IGZO film composition revealed that an additional nitrate bidentate configuration appeared in the films with preheating at 275 °C, though the substitution of a N atom for O sub-lattice (N)_O was found in the film regardless of the preheating temperature. It was suggested that the performance enhancement was primarily attributed to the improvement in film texture brought about by the preheating strategy. Furthermore, the mobility enhancement at high preheating temperature was also related to the appearance of a bidentate configuration (M–O₂–N).

---

Introduction

Thin film transistors (TFTs) based on transparent oxide semiconductors (TOS) have attracted much attention as they are increasingly employed in backplane driving circuits for large-area display panels such as liquid crystal displays (LCDs) and organic light-emitting diodes (OLEDs).^1–4 Conventionally, high quality TOS films are deposited via vacuum deposition methods such as magnetron sputtering⁵ and pulsed laser deposition,⁶ followed by photolithography and etching processes. In the last decade, printable electronic materials including conductors,^7–9 semiconductors^10–12 and dielectrics,^13,14 were extensively studied due to the potentially simplified and low cost manufacturing process on large area and flexible substrates. There have been reports on printing TOS based TFTs. However, these printed TFTs usually require high temperature annealing in order to achieve high performance.^15,16

Recently, considerable progresses have been made in solution processed TOS-TFTs with low temperature annealing.^17–23 Among these studies, indium oxide (In₂O₃) and indium-zinc-oxide (IZO) were the most used TOS materials, because indium can enhance the electron transport in metal oxide semiconductor, hence high mobility.²⁴ Some ZnO-based thin film transistors have been produced at temperatures lower than 200 °C by using ammine-hydroxo metal aqueous solution as well.^10,25 However, large electron concentration because of the formation of oxygen vacancies in these materials resulted in poor reliability for the fabricated TFTs. In addition, crystallization at low temperature for In₂O₃ and ZnO films results in poor uniformity and repeatability.²⁶ In contrast, TFTs based on indium gallium zinc oxide (IGZO) are much better in stability, uniformity as well as reproducibility than ZnO, In₂O₃ or IZO, because of the amorphous structure of IGZO and strong suppression of oxygen vacancies formation.^3,24,26 TFTs made of vacuum sputtered IGZO film have become the industrial standard technology for making backplane of LCD or OLED display panels. However, solution processed IGZO-TFTs with low mobility were usually obtained at processing temperature lower than 350 °C,^27,28 primarily due to the mobility degradation caused by gallium doping in the metal oxide matrix.²⁴ Up to now, there are few reports on high performance solution processed IGZO-TFTs, by using susceptible metal alkoxide precursors, sophisticated combustion or complex photochemical annealing.^17,18,22,29

To enable good performance for solution processed oxides TFTs at low annealing temperature, efforts have been focused on employing new precursors such as ammine-hydroxo metal aqueous solution¹⁰ or novel annealing techniques such as combustion,¹⁷ photochemical annealing,²² O₂/O₃ annealing,²⁰ high pressure annealing,³⁰ and microwave irradiation.³¹ Recently, metal nitrate solution has emerged as a promising ink to make high performance TFTs with low temperature process. It was also found that the chemical conversion from metal nitrate into metal oxide was greatly affected by the processing procedures and parameters.^19,21,32 For example, the activation temperature of spin-coated In₂O₃ TFTs was reduced from 250 to 175 °C by replacing alcohol with water as solvent for metal nitrates.²¹ Similarly, reducing the temperature of metal nitrates solution from 60 °C to 4 °C enhanced mobility by 2–3 times.³² Both the solvent effect and solution temperature effect were attributed to the formation of H₂O molecule shared ion pairs, which can promote the thermally driven hydrolysis reaction followed by a metal hydroxide condensation, yielding a metal oxide lattice at low temperature.^21,32

It is well known that thermal treatment is responsible for the solidification of solution deposited film and for chemical transformation from metal nitrate into metal oxide.^22,30 Compared with general oven heating, directly placing the substrate on a preheated hotplate can make the film temperature rise rapidly from room temperature to the preheated temperature in seconds. Therefore, it can be expected that the rapid thermal treatment with a hotplate would make the drying and structure evolution of film different from those with a moderate heating rate. Conventionally, the primary action of hotplate-based thermal treatment, also known as prebaking process carried out at temperature no more than 100 °C, is driving off the solvent in precursor films.^33,34 Recent research demonstrated that the preheating process not only can drive off solvent but also affect the performance of solution processed oxide-TFTs.^33,34 As the chemical transformation of metal nitrates takes place at 200 °C, high temperature rapid preheating may cause not only fast removal of solvent but also alter the chemical conversion process, which so far has not been fully investigated and understood.

In the present work, the influence of preheating on solution processed IGZO-TFTs has been investigated. The TFTs were fabricated by spin-coating and inkjet printing of metal nitrates aqueous solution and preheated with a hotplate at temperatures ranging from 40 °C to 275 °C, while all the post-annealing temperatures were kept at 300 °C in a muffle furnace. It was found that the performances of spin-coated and printed TFTs were significantly enhanced by increasing the preheating temperature. The mechanism of TFTs performance enhancement was studied by analyzing the IGZO film structures using different material characterization techniques including X-ray photoelectron spectroscopy (XPS), Fourier transform infrared spectroscopy (FTIR) and atomic force microscope (AFM).

Experimental

Preparation of precursor solution

The metal oxide precursor ink was prepared by dissolving In(NO₃)₃·4.5H₂O (99.5%, Sinopharm Chemical Reagent), Zn(NO₃)₂·6H₂O (98%, Acros), Ga(NO₃)₃·7.8H₂O (99.9%, trace metals basis, Sigma-Aldrich) in deionized water with a concentration of 0.2 mol L⁻¹. The as-received metal nitrates were used without any further purification. The molar ratio of the precursor solution was 3 : 0.65 : 2 (In : Ga : Zn). Before deposition of the film, the formulated precursor solution was vigorously stirred for 4 h at room temperature and filtered with pore size of 0.45 μm.

Film deposition and thin film transistor fabrication

Bottom-gate top-contact TFTs were fabricated by inkjet printing or spin-coating metal nitrates precursor films on a heavily doped silicon substrate coated with thermally grown silicon dioxide. After forming the semiconductor channels, preheating was carried out with a hotplate and then annealing in a furnace. Finally, thermally evaporated aluminum (Al) was deposited on top of the film to form the source-drain electrodes. Fig. 1 illustrates the fabrication process.

Fig. 1 The schematic illustration of the fabrication process of printed or spin-coated IGZO-based TFTs.

The thickness of IGZO film in the channel was varied by adjusting the inkjet printing or spin-coating parameters including printing drop space, number of printed layers, spinning speed and number of spun layers etc. A commercial inkjet printer (Dimatix Materials Printer DMP-2831) with a 20 pL cartridge was used for the inkjet printing deposition. The Al source and drain (S/D) electrodes were patterned by thermally evaporation through a shadow mask to have 50 μm channel length. The width-to-length (W/L) ratio of 20 was used to avoid overestimation of field effect mobility for the spin-coated TFTs.³⁵ In the case of printed TFTs, the S/D electrodes were across the printed metal oxide lines, thus the mobility overestimation would not happen for any values of W/L. For the convenience of comparison, all the drain-source current (I_ds) values in the transfer and output curves were normalized with W/L = 1.

The as-deposited metal nitrates precursor films were placed directly on a preheated hotplate and heated for 10 min, then annealed with a heating rate of 10 °C min⁻¹ to the maximum temperature of 300 °C in a muffle furnace and remaining for 2 h under an ambient atmosphere. To investigate the hotplate heating effect, different preheating temperatures were set as 40, 95, 185 and 275 °C, lower than the post-annealing temperature (300 °C). The temperature values were monitored by an infrared thermometer. The printed IGZO films without the additional hotplate preheating were regarded as having preheating of the default printer working stage temperature of 40 °C.

Materials and device characterization

The electrical performance of transistors was characterized using semiconductor analyzer (Keithley 4200) in a dark box at ambient condition. Accordingly, saturation field effect mobility (μ_sat) and threshold voltage (V_th) were extracted from the TFT transfer curves using the following expressionwith d = 300 nm thickness of the SiO₂ gate dielectrics and the dielectric constant of 3.9. The sub-threshold swing (SS) was calculated as the minimum value of the inverse slope of the plot of lg(I_DS) versus V_GS.

The chemical structures of IGZO films were examined by X-ray photoelectron spectroscopy (XPS). The XPS data were recorded for spin-coated IGZO films on SiO₂/Si substrate using a monochromatic Al Kα X-ray source and a Thermo Scientific Escalab 250Xi spectrometer with an overall energy space of ΔE = 0.01 eV. All the XPS peaks were calibrated by taking C 1s reference at 284.8 eV. The surface morphology and roughness of the films were analyzed using an atomic force microscope (AFM, Veeco Nanoscope digital instruments). The film thicknesses and profile were determined with the measurement of Step profiler (Veeco, Dektak 150).

Fourier transform infrared spectroscopy (FTIR) and thermogravimetric analysis (TGA) were used to investigate thermal treatment dependence of chemical convention from the nitrate precursor to metal oxide. Because the signals from film samples were weak, the FTIR spectra were collected from powder samples instead, which were obtained by drying off the precursor solution at 40 °C for 48 h and following the same preheating and post-annealing treatment as that of the films. After the samples were prepared in pellet form by mixing the powder with KBr (Aldrich, 99%, FTIR grade), the FTIR spectra were recorded on a FTIR spectroscopy (Nicolet NEXUS 6700). The thermogravimetric analysis was performed with a heating rate of 10 °C min⁻¹ under a dry air flow at the rate of 20 mL min⁻¹.

Results and discussion

Prior to the TFT fabrication, the oxide films printed from the IGZO ink have been optimized by varying substrate temperature, nozzle temperature, repetition of printing, droplet size and spacing etc. Optimized morphology and controlled film thickness of printed IGZO channel lines were achieved with 40 °C substrate temperature and 0.2 M metal nitrate aqueous solution, as shown in Fig. S1.†

The average saturation field effect mobility and threshold voltage against the preheating temperature are represented in Fig. 2a and b, which are measured on a group of six printed IGZO-TFTs annealed at 300 °C. With the preheating temperature, the field effect mobility gradually enhanced, while the threshold voltage slightly decreased. It is worth noting that a kink structure in the plot of mobility is found around 260 °C. The sharp enhancement of mobility will be discussed below. The slight change in threshold voltage was neglected, considering the threshold voltage more than 10 V.

Fig. 2 Printed IGZO-TFT characteristics with post-annealing at 300 °C: (a and b) average saturation field effect mobility and threshold voltage of a group of six printed IGZO-TFTs against the preheating temperature; (c) transfer characteristics of IGZO-TFTs with hotplate preheating at 40, 95, 185 and 275 °C; (d) output characteristics of IGZO-TFT with hotplate preheating at 275 °C; (e and f) transfer characteristics and distribution of saturation mobility of printed IGZO-TFTs preheated at 275 °C, which were prepared on 30 separated substrates in one batch.

Representative electronic characteristics of printed IGZO TFTs annealed at 300 °C with different preheating temperatures are shown in Fig. 2c and summarized in Table 1. The field-effect mobility of TFTs was enhanced more than 10 times (from 0.31 to 4.93 cm² V⁻¹ s⁻¹) with the increase of hotplate preheating temperature from 40 °C to 275 °C. Moreover, hysteresis and sub-threshold swing (SS) performances were also improved remarkably. The printed IGZO-TFTs with hotplate preheating at 275 °C have large current on/off ratio (I_on/I_off) more than 10⁷, relatively small SS of 0.93 V dec⁻¹ and excellent output characteristics (Fig. 2d). The similar performance enhancement was also found in the spin-coated counterparts, as shown in Fig. S2 and Table S1.† The maximum mobility of 7.9 cm² V⁻¹ s⁻¹ and I_on/I_off more than 10⁸ was obtained for the spin-coated IGZO-TFTs annealed at low temperature of 300 °C, with the hotplate preheating at 275 °C. These performances are comparable or superior to the best reported solution processed IGZO-TFTs fabricated at low temperature.^17,18,22,29 Fig. 2e and f show the transfer characteristics and distribution of saturation mobility of printed IGZO-TFTs prepared on 30 separated substrates in one batch, which were preheated at 275 °C and annealed at 300 °C. The result shows that a small discrepancy of mobility is obtained in the printed IGZO-TFTs, demonstrating the good uniformity and reproducibility of devices.

Table 1 Electrical parameters of the printed IGZO-TFTs (shown in Fig. 2a) annealed at 300 °C with different hotplate preheating temperatures

Preheating temperature (°C)	Field effect mobility μ (cm^2 V^−1 s^−1)	Threshold voltage Vth (V)	Sub-threshold swing SS (V dec^−1)	Current on/off ratio Ion/Ioff
40	0.31	11	1.18	>10^6
95	0.56	14	1.53	>10^6
185	1.05	18	2.04	∼10^7
275	4.93	8	0.93	>10^7

---

The contact resistance between the channel and S/D electrodes in the printed IGZO-TFTs was measured using transmission line model.³⁶ To extract the contact resistance (R_C), the total resistances (R_T) were measured as a function of channel length in the range from 50 μm to 250 μm at a 50 μm step, when the TFTs were turned on at the gate voltage of 20 V and the drain voltage of 1 V. As shown in Fig. 3, the contact resistance and channel resistance decrease as the preheating temperature increases. However, the fractions of total device resistance due to contact in the TFTs are no more than 25% for all samples. Therefore, the mobility enhancement is not mainly attributed to the contact resistance.³⁷

Fig. 3 (a) Width-normalized total device resistances (R_T) as a function of channel length in printed IGZO-TFT annealed at 300 °C for different preheating temperature. (b) The contact resistance (R_C) and fraction of total device resistance due to contact as a function of preheating temperature.

In order to understand the mechanism of the performance enhancement influenced by preheating temperature, different material analytical techniques including XPS, FTIR and AFM, were used. XPS analyses were carried out on the spin-coated IGZO films annealed at 300 °C with hotplate heating at different temperatures. Based on the multi-times XPS measurement, no obvious difference was found in elemental ratio of the IGZO films, except a slight decrease in the Ga : In molar ratio from 0.68 : 3 to 0.55 : 3 when hotplate preheating temperature increased from 40 °C to 275 °C (see ESI Table S2†). However, this slight decrease of Ga element cannot induce such large performance enhancement, because the printed IGZO-TFTs with hotplate preheating at 95 °C exhibited low mobility about 0.88 cm² V⁻¹ s⁻¹, when the In : Ga : Zn molar ratio of 3 : 0.5 : 2 was used.

Fig. 4 shows the typical O 1s spectra of IGZO films and the difference seems to be quite small for different hotplate preheating temperatures. The curves were fitted with two Gaussian peaks centred at 529.95 ± 0.02 eV and 531.35 ± 0.05 eV, respectively, which were assigned to the oxygen species in the stoichiometric oxide lattice and the non-ideal metal oxide due to the formation of oxygen vacancies and metal hydroxide (M–OH, M is metal atom) etc.^32,38 The oxide lattice content of IGZO films decreased from 47.2% to 42.3% as the hotplate heating temperature varied from 40 °C to 95 °C, but increased to 49.5% for 275 °C. It indicated that the prebaking temperature weakly influenced the quality of oxide lattice while the same annealing temperature of 300 °C was used.

Fig. 4 XPS O 1s spectra of the IGZO thin films annealed at 300 °C with hotplate preheating temperature of (a) 40 °C (b) 95 °C (c) 185 °C (d) 275 °C. The peaks of O_I (529.95 ± 0.02 eV, blue line) and O_II (531.35 ± 0.05 eV, red line) indicate the stoichiometric oxide lattice and the non-ideal metal oxide, respectively.

The N 1s spectra of the IGZO films annealed at 300 °C with different hotplate preheating temperatures are presented in Fig. 5. The results show that the metal oxide films with preheating are nitrogen-doped IGZO films irrespective of hotplate preheating temperature. As TG curves shown in Fig. S3,† gallium nitrate and zinc nitrate cannot decompose completely at 300 °C, therefore it was consequent to detect N element in the IGZO films. For all IGZO films, a strong peak located at ca. 397.0 eV with a shoulder peak at 393.4 eV, was found. The shoulder peak is attributed to the Auger electron of Ga ion. The 397.0 eV peak is due to the substitution of N atom for O sublattice i.e., the formation of metal–N bond, denoted as (N)_O.³⁹ Likewise, more interestingly an apparent peak centred at ∼404.2 eV always appears in the XPS spectra of IGZO films with hotplate preheating at 275 °C, this peak becomes very weak if the hotplate preheating temperatures are in the range of 40–185 °C. This phenomenon indicates that the hotplate preheating at 275 °C leads to a stable N-relative chemical structure in the annealed IGZO films different from those with lower preheating temperature. The N 1s peak of the NO₃⁻ usually locates at ∼407.5 eV;⁴⁰ therefore the feature at lower binding energy of 404.2 eV should correspond to nitrogen atoms with the chemical state lower than +5 (NO₃⁻).

Fig. 5 XPS N 1s spectra of the IGZO thin films annealed at 300 °C with hotplate preheating at 40 °C, 95 °C, 185 °C and 275 °C.

Fourier-transform infrared (FTIR) spectroscopy was performed to have an insight into the chemical structure of the products after various thermal processes. To enhance signal intensity, powder samples obtained by the same thermal process as the films, were used in the measurement, instead of the films. Fig. 6a and b show the FTIR absorption of samples with hotplate preheating at different temperatures, before and after 300 °C post-annealing, respectively. Before post-annealing, the powders obtained by hotplate heating at 40 °C and 95 °C, presented typical infrared absorption characteristics of metal nitrates. The broad peaks around 3500 cm⁻¹ are assigned to the hydroxyl group (–OH) stretching vibrations of M–OH and water molecule. The peak at ∼1620 cm⁻¹ is attributed to –OH bending vibrations of water molecules.⁴¹ The weak peak at 1770 cm⁻¹ might be assigned to the out-of-phase bending vibrations of the flanking water molecules.⁴² The absorption bands around 1388 and 833 cm⁻¹ correspond to the asymmetrical stretching vibration and the symmetrical out-of-plane bending vibration of NO₃⁻, respectively.⁴³ The symmetric stretch vibration of nitrates located at 1000–1050 cm⁻¹, however is infrared inactive. Therefore, the medium-intensity peak at ∼1015 cm⁻¹ in spectra of samples with hotplate heating at 40–95 °C is associated with the bending vibrations of –OH bonded to metal atoms.

Fig. 6 FTIR adsorption spectra of the IGZO power obtained by: (a) hotplate heating at 40 °C, 95 °C, 185 °C and 275 °C, respectively, without further annealing; (b) hotplate preheating at 95 °C, 185 °C and 275 °C with further annealing at 300 °C; (c) schematic representation of monodentate, bidentate, (O–O)-bridged and O-bridged coordinated nitrate structures bonding on metal atoms.

With the hotplate heating temperature increasing to 185 °C or above, water molecules and –OH groups in samples decreased greatly, so the relative absorption peaks became weak (3500 and 1620 cm⁻¹) or even disappeared (1770 and ∼1015 cm⁻¹). Meanwhile, it could also be observed that the peak intensity at 1388 cm⁻¹ was weakened and the absorption of 833 cm⁻¹ almost disappeared, which indicates that the thermal decomposition of nitrates has taken place at temperature as low as 185 °C. Moreover, the peak due to M–O at 509 cm⁻¹ red shifted to 447 cm⁻¹, demonstrating the formation of metal–oxygen–metal bonds (M–O–M).⁴⁴ In addition, new absorption peaks at 1517 and 1572 cm⁻¹ appeared in the samples with hotplate heating at 185 °C and 275 °C before further annealing, respectively. Based on the previous calculated and experimental results, nitrate species are likely bound primarily with metal atoms in different stable configurations, including monodentate, bidentate and (O–O)-bridged configurations (Fig. 6c).^43,45 The IR absorption peaks at 1517 and 1572 cm⁻¹ are from the monodentate and bidentate coordinated nitrate on metal atoms, respectively.

After post annealing at 300 °C, the absorption at 1572 cm⁻¹ became more apparent for sample preheated at 275 °C, but not intensive yet for samples preheated at 95 °C and 185 °C, as shown in Fig. 6b. It indicates that the formation of bidentate structure depends on the hotplate preheating temperature if annealed at low temperature of 300 °C. The appearance of this FTIR peak is very similar to the N 1s peak at 404.2 eV in XPS spectra. Therefore, it is suggested that the IR absorption at 1572 cm⁻¹ and N 1s XPS peak at 404.2 eV are originated from the same nitrate structure: bidentate configuration bound with metal atoms (M–O₂–N). In the structure, chemical state of N atom is lower than +5, due to its two O atoms bonding with metal atoms. Therefore, the bidentate configuration (M–O₂–N) can conform to the XPS feature of N 1s at 404.2 eV.

It is well known that the chemical kinetics of some reactions depend on reaction conditions. In the present work, the hotplate heating leads to far larger temperature rising rate (100 °C in less than 1 min), compared with conventional furnace heating with rate less than 100 °C min⁻¹. When directly heating the precursor films on a hotplate at 275 °C, different chemical kinetics such as hydrolysis, decomposition and dehydroxylation might take place and the bidentate configuration (M–O₂–N) was produced. As mentioned above, the metal nitrates are hard to completely convert into oxides at 300 °C. Therefore, the bidentate configuration still remained in the final annealed film.

As shown in Fig. 2c and Table 1, the TFT performances with printed IGZO films annealed at 300 °C are improved moderately when the hotplate preheating temperature increases from 40 °C to 95 °C and then 185 °C, with the mobility of 0.31, 0.56 and 1.05 cm² V⁻¹ s⁻¹ in sequence. From the XPS and FTIR results, it can be seen that these IGZO films have similar chemical structure. In addition, as discussed above, the metal ratio change (Table S2†) in these films is too little to result in such large performance change. Accordingly, the performance improvement in this range of temperature should be mainly attributed to the differences in film texture.

The surface morphology of printed and spin-coated IGZO films annealed at 300 °C with different hotplate preheating temperatures were characterized using atomic force microscope (AFM), and the images are shown in Fig. 7 and S4,† respectively. The printed film with hotplate preheating at 40 °C has high root-mean-square (RMS) roughness about 0.61 nm, while the RMS roughness was significantly reduced to 0.28 nm by elevating the prebake temperature to 95 °C. Kim et al. also reported the surface roughness decreases for spin-coated, photo-chemically annealed IGZO films with the rise of the prebaking temperature from 35 °C to 100 °C.³³ The difference in surface morphology suggested that there were different “drying kinetics” as different preheating temperatures are used, because other fabrication conditions are the same. Based on the FTIR results (Fig. 6), it is known that the crystal water in metal nitrates is not removed completely at 40 °C temperature under ambient condition. The water would cause the oxide film texture depending on preheating process.

Fig. 7 AFM images of surface morphology of the inkjet-printed films annealed at 300 °C in furnace for 2 h with variable hotplate preheating temperatures: (a) 40 °C (RMS = 0.61 nm), (b) 95 °C (RMS = 0.28 nm), (c) 185 °C (RMS = 0.28 nm), (d) 275 °C (RMS = 0.26 nm).

Firstly, the metal nitrates of high solubility can dissolve in the crystal water at higher temperature. For example, zinc nitrate hexahydrate that is solid at 20 °C can become zinc nitrate solution at 40 °C or above, because it has high solubility in water about 327 g/100 mL at 40 °C. As a result, the as-deposited precursor film would be unstable and easy to deform upon temperature rising. If the precursor films were directly put on the hotplate preheated at higher temperature, the removal of water in the films would be faster, reducing the deformation. Therefore, the produced IGZO films have more homogeneous internal texture of film and smoother surface morphology.

The surface film morphology does not directly influence electronic properties of the top-contact bottom-gate TFTs, because the electron transport takes place at the interfaces between the channel and dielectric layer.¹⁴ However, the larger surface roughness means more inhomogeneous film texture and higher probability of discontinuity in the IGZO film when its thickness is less than 20 nm. Accordingly, the printed IGZO films preheated at 40 °C with high roughness resulted in the worst electronic properties, whereas using hotplate heating at higher temperature, the devices performance were improved. By using spin-coating process, the surface of all the films became smoother (Fig. S4†), suggesting the improved continuity of IGZO films, so the IGZO-TFT performances were better than their printed counterparts.

Secondly, it is well known that the fast evaporation of water would drive nanoparticle aggregation due to the change of surface tension when nanoparticle suspension was dried.⁴⁶ Similarly, we suggest herein that the faster water loss may also induce the film compacter, which can be responsible for the performance improvement with the preheating temperature. Therefore, the performance is improved with further increasing the hotplate preheating temperature above 95 °C, although the surface roughness was almost unchanged in both printed and spin-coated IGZO films.

As seen in Fig. 2a, a kink structure is found around 240–275 °C in the plot of mobility against the preheating temperature. The mobility was enhanced sharply in this temperature range, suggesting there was another mechanism responsible for the performance improvement, besides the film texture effect. As discussed above, the signal from bidentate configuration (M–O₂–N) becomes more apparent when the preheating temperature is 275 °C. In some previous literature, it has been found that N-doping can improve the performance of oxide TFTs, by passivation of the defects and/or dangling bonds to reduce the charge-trap density in the films.^39,44,47,48 Therefore, it is reasonable to suggest that the mobility enhancement from 240 °C to 275 °C was also attributed to an unknown mechanism which is closely related to the formation of the bidentate configuration. The underlying mechanism of the bidentate configuration (M–O₂–N) is not clear, as we did not find any relevant literature about how the bidentate configuration influence the carrier transfer or energy band of oxide semiconductor. It needs further experimental and theoretical research. In addition, the IGZO films with 275 °C hotplate preheating have the smoother surface, highest ratio of stoichiometric oxide lattice, slightly lower Ga ion, resulting in the best TFT performances.

Conclusions

High performance solution processed IGZO TFTs at low annealing temperature of 300 °C have been achieved by employing a simple hotplate preheating strategy. The TFTs made by inkjet printing and spin-coating of metal nitride inks went through hotplate heating at 275 °C prior to 300 °C furnace annealing, which resulted in field-effect mobility enhancement of more than 10 times from 0.31 cm² V⁻¹ s⁻¹ to 4.93 cm² V⁻¹ s⁻¹ for the printed TFTs and 5 times from 1.44 cm² V⁻¹ s⁻¹ to 7.9 cm² V⁻¹ s⁻¹ for spin-coated TFTs, respectively. High I_on/I_off ratios of more than 10⁸ were also achieved for spin-coated IGZO TFTs and more than 10⁷ for inkjet-printed IGZO TFTs. Thorough investigations on film composition and surface roughness with and without hotplate preheating were carried out. The results suggested that the performance enhancement was primarily attributed to the film texture improvement. The mobility enhancement at high preheating temperature was also related to the appearance of bidentate configuration. In addition, a slight decrease of Ga element ratio was observed in the films preheated at higher temperature, which to some extent may also contribute to TFT mobility enhancement.

Acknowledgements

This work was supported by the project of the Major Research plan of the Nation Natural Science Foundation of China (Grant No. 91123034), Strategic Priority Research Program of the Chinese Academy of Sciences (Grant No. XDA09020201), and Project supported by National Science and Technology Ministry (Grant No. 2012BAF13B05-402).

References

1. K. Nomura, H. Ohta, A. Takagi, T. Kamiya, M. Hirano and H. Hosono, Nature, 2004, 432, 488–492.
2. P. Gorrn, M. Sander, J. Meyer, M. Kroger, E. Becker, H. H. Johannes, W. Kowalsky and T. Riedl, Adv. Mater., 2006, 18, 738.
3. J. S. Park, W. J. Maeng, H. S. Kim and J. S. Park, Thin Solid Films, 2012, 520, 1679–1693.
4. H. Xu, D. X. Luo, M. Li, M. Xu, J. H. Zou, H. Tao, L. F. Lan, L. Wang, J. B. Peng and Y. Cao, J. Mater. Chem. C, 2014, 2, 1255–1259.
5. H. Yabuta, M. Sano, K. Abe, T. Aiba, T. Den, H. Kumomi, K. Nomura, T. Kamiya and H. Hosono, Appl. Phys. Lett., 2006, 89, 112123.
6. A. Suresh, P. Gollakota, P. Wellenius, A. Dhawan and J. F. Muth, Thin Solid Films, 2008, 516, 1326–1329.
7. K. Woo, C. Bae, Y. Jeong, D. Kim and J. Moon, J. Mater. Chem., 2010, 20, 3877–3882.
8. C. Yang, C. P. Wong and M. M. F. Yuen, J. Mater. Chem. C, 2013, 1, 4052–4069.
9. Z. Chen, X. C. Qin, T. Zhou, X. Z. Wu, S. S. Shao, M. L. Xie and Z. Cui, J. Mater. Chem. C, 2015, 3, 11464–11470.
10. S. T. Meyers, J. T. Anderson, C. M. Hung, J. Thompson, J. F. Wager and D. A. Keszler, J. Am. Chem. Soc., 2008, 130, 17603–17609.
11. W. Y. Xu, J. W. Zhao, L. Qian, X. Y. Han, L. Z. Wu, W. C. Wu, M. S. Song, L. Zhou, W. M. Su, C. Wang, S. H. Nie and Z. Cui, Nanoscale, 2014, 6, 1589–1595.
12. L. Qian, W. Y. Xu, X. F. Fan, C. Wang, J. H. Zhang, J. W. Zhao and Z. Cui, J. Phys. Chem. C, 2013, 117, 18243–18250.
13. X. Z. Wu, Z. Chen, T. Zhou, S. S. Shao, M. L. Xie, M. S. Song and Z. Cui, RSC Adv., 2015, 5, 20924–20930.
14. W. Tang, J. H. Li, J. Q. Zhao, W. M. Zhang, F. Yan and X. J. Guo, IEEE Electron Device Lett., 2015, 36, 950–952.
15. D. H. Lee, S. Y. Han, G. S. Herman and C. H. Chang, J. Mater. Chem., 2009, 19, 3135–3137.
16. S. Jeong, J. Y. Lee, S. S. Lee, S. W. Oh, H. H. Lee, Y. H. Seo, B. H. Ryu and Y. Choi, J. Mater. Chem., 2011, 21, 17066–17070.
17. M. G. Kim, M. G. Kanatzidis, A. Facchetti and T. J. Marks, Nat. Mater., 2011, 10, 382–388.
18. K. K. Banger, Y. Yamashita, K. Mori, R. L. Peterson, T. Leedham, J. Rickard and H. Sirringhaus, Nat. Mater., 2011, 10, 45–50.
19. G. X. Liu, A. Liu, H. H. Zhu, B. Shin, E. Fortunato, R. Martins, Y. Q. Wang and F. K. Shan, Adv. Funct. Mater., 2015, 25, 2564–2572.
20. S. Y. Han, G. S. Herman and C. H. Chang, J. Am. Chem. Soc., 2011, 133, 5166–5169.
21. Y. H. Hwang, J. S. Seo, J. M. Yun, H. Park, S. Yang, S. H. K. Park and B. S. Bae, NPG Asia Mater., 2013, 5, e45.
22. Y. H. Kim, J. S. Heo, T. H. Kim, S. Park, M. H. Yoon, J. Kim, M. S. Oh, G. R. Yi, Y. Y. Noh and S. K. Park, Nature, 2012, 489, 128–U191.
23. B. D. Ahn, H. J. Jeon, J. Z. Sheng, J. Park and J. S. Park, Semicond. Sci. Technol., 2015, 30, 064001.
24. E. Fortunato, P. Barquinha and R. Martins, Adv. Mater., 2012, 24, 2945–2986.
25. Y. H. Lin, H. Faber, K. Zhao, Q. X. Wang, A. Amassian, M. McLachlan and T. D. Anthopoulos, Adv. Mater., 2013, 25, 4340–4346.
26. J. S. Park, H. Kim and I. D. Kim, J. Electroceram., 2014, 32, 117–140.
27. J. M. Kwon, J. Jung, Y. S. Rim, D. L. Kim and H. J. Kim, ACS Appl. Mater. Interfaces, 2014, 6, 3371–3377.
28. S. Jeong, J. Y. Lee, S. S. Lee, Y. H. Seo, S. Y. Kim, J. U. Park, B. H. Ryu, W. Yang, J. Moon and Y. Choi, J. Mater. Chem. C, 2013, 1, 4236–4243.
29. K. Everaerts, L. Zeng, J. W. Hennek, D. I. Camacho, D. Jariwala, M. J. Bedzyk, M. C. Hersam and T. J. Marks, ACS Appl. Mater. Interfaces, 2013, 5, 11884–11893.
30. Y. S. Rim, W. H. Jeong, D. L. Kim, H. S. Lim, K. M. Kim and H. J. Kim, J. Mater. Chem., 2012, 22, 12491–12497.
31. T. Jun, K. Song, Y. Jeong, K. Woo, D. Kim, C. Bae and J. Moon, J. Mater. Chem., 2011, 21, 1102–1108.
32. K. H. Lee, J. H. Park, Y. B. Yoo, W. S. Jang, J. Y. Oh, S. S. Chae, K. J. Moon, J. M. Myoung and H. K. Baik, ACS Appl. Mater. Interfaces, 2013, 5, 2585–2592.
33. J. Kim, H. H. Park, Y. H. Kim and S. K. Park, Sci. Adv. Mater., 2015, 7, 1525–1529.
34. P. K. Nayak, M. N. Hedhili, D. Cha and H. N. Alshareef, ACS Appl. Mater. Interfaces, 2013, 5, 3587–3590.
35. K. Okamura, D. Nikolova, N. Mechau and H. Hahn, Appl. Phys. Lett., 2009, 94, 183503.
36. Y. Shimura, K. Nomura, H. Yanagi, T. Kamiya, M. Hirano and H. Hosono, Thin Solid Films, 2008, 516, 5899–5902.
37. M. A. Reyes-Martinez, A. J. Crosby and A. L. Briseno, Nat. Commun., 2015, 6, 6948.
38. M. Haming, A. Issanin, D. Walker, H. von Seggern, W. Jagermann and K. Bonrad, J. Phys. Chem. C, 2014, 118, 12826–12836.
39. B. Yao, D. Z. Shen, Z. Z. Zhang, X. H. Wang, Z. P. Wei, B. H. Li, Y. M. Lv, X. W. Fan, L. X. Guan, G. Z. Xing, C. X. Cong and Y. P. Xie, J. Appl. Phys., 2006, 99, 123510.
40. S. Park, K. H. Kim, J. W. Jo, S. Sung, K. T. Kim, W. J. Lee, J. Kim, H. J. Kim, G. R. Yi, Y. H. Kim, M. H. Yoon and S. K. Park, Adv. Funct. Mater., 2015, 25, 2807–2815.
41. Y. Shen and P. Y. Wu, J. Phys. Chem. B, 2003, 107, 4224–4226.
42. J. M. Headrick, J. C. Bopp and M. A. Johnson, J. Chem. Phys., 2004, 121, 11523–11526.
43. J. Baltrusaitis, J. Schuttlefield, J. H. Jensen and V. H. Grassian, Phys. Chem. Chem. Phys., 2007, 9, 4970–4980.
44. B. D. Ahn, H. J. Jeon and J. S. Park, ACS Appl. Mater. Interfaces, 2014, 6, 9228–9235.
45. M. M. Angelini, R. J. Garrard, S. J. Rosen and R. Z. Hinrichs, J. Phys. Chem. A, 2007, 111, 3326–3335.
46. G. Derkachov, K. Kolwas, D. Jakubczyk, M. Zientara and M. Kolwas, J. Phys. Chem. C, 2008, 112, 16919–16923.
47. J. Raja, K. Jang, N. Balaji, W. Choi, T. T. Trinh and J. Yi, Appl. Phys. Lett., 2013, 102, 083505.
48. X. M. Huang, C. F. Wu, H. Lu, F. F. Ren, D. J. Chen, R. Zhang and Y. D. Zheng, Appl. Phys. Lett., 2013, 102, 193505.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c6ra01776b

---