Enhanced dielectric properties of alternative NO-gas-based SiO₂ films via plasma-enhanced chemical vapor deposition for high-performance indium–gallium–zinc oxide thin-film transistors

Abstract

Nitric oxide (NO) is proposed as an alternative gas to nitrous oxide (N₂O) for the formation of silicon dioxide (SiO₂) films by plasma-enhanced chemical vapor deposition. Post-thermal annealing in a N₂ atmosphere is employed to improve the current–voltage and capacitance–voltage characteristics of the films by removing charge impurities and curing defect states, thereby restoring the intrinsic dielectric properties of the NO-based SiO₂ films. The effects of the enhanced dielectric properties of the NO-based SiO₂ films on the device performance of indium–gallium–zinc oxide (IGZO) thin-film transistors (TFTs) were subsequently investigated. The IGZO TFT consisting of a NO-based SiO₂ dielectric film annealed at 300 °C in a N₂ atmosphere exhibits excellent electrical characteristics, including a low off-current, large on/off ratio, low subthreshold swing, high field-effect mobility, and threshold voltage near 0 V. Improvements in electrical stability of the IGZO TFTs against a prolonged bias stress are also achieved owing to the introduction of N₂-annealed SiO₂ dielectric films. Finally, charge-transport properties are investigated via temperature-dependent field-effect mobility analysis to determine the activation energy and interfacial trap density of states, which agree well with the improved device performance of the IGZO TFTs consisting of NO-based SiO₂ films with enhanced dielectric properties.

---

Introduction

Silicon dioxide (SiO₂) films are widely used in the semiconductor industry, including memory devices, displays, and logic circuits.¹ Several methods have been reported to fabricate SiO₂ thin films, such as atomic layer deposition, sputtering, and spray pyrolysis.^2,3 Among them, plasma-enhanced chemical vapor deposition (PECVD), a common method used to prepare SiO₂ films, offers advantages such as ease of control, high reproducibility, and low thermal budget.⁴ PECVD not only enables high deposition rates and low process temperatures, but also prevents defect formation and dopant diffusion, thereby producing high-quality films with significant technical benefits.^4,5 The precursors commonly used for preparing SiO₂ film via PECVD include silane (SiH₄), carbon dioxide, nitrous oxide (N₂O), and hydrogen (H₂).⁵ However, the conventional use of N₂O has raised concerns because it is a greenhouse gas that can exacerbate global warming.⁶ Thus, alternative gases that can replace N₂O must be identified that are compatible with the PECVD process while maintaining or improving the dielectric properties of N₂O gas-based SiO₂ films.⁷ One promising alternative gas is nitric oxide (NO), which exhibits excellent thermal stability and acts as an effective intermediate in the chain-reaction mechanism of PECVD.⁸ NO exhibits high reactivity and oxidizing properties, which can reduce the possibility of carbon contamination while allowing for high growth rates of SiO₂ films.⁹ Furthermore, NO is known to facilitate the incorporation of oxygen into the film while simultaneously influencing the stoichiometry and defect states at the interface due to its lower ionization energy compared to conventional N₂O.¹⁰

Dielectric properties are key parameters for Si-based devices that determine the device performance metrics, such as leakage current density, capacitive charge density, and interfacial characteristics.¹¹ A number of strategies, including post-thermal annealing, cold sintering, and calcination, have been employed to improve the dielectric properties of SiO₂ films.^11,12 These methods can reduce defect states, densify the dielectric film, and improve structural characteristics.^11,13,14 In particular, post-thermal annealing is a simple and effective method to restore the intrinsic dielectric properties of SiO₂ films, which can be affected by unintentional physical and/or chemical interactions during their fabrication.^11,14 Interfacial trapped charges can be released by strengthening the Si–O bonding structure, thereby enabling the formation of high-quality SiO₂ films.¹⁴ We note that SiO₂ films deposited using NO as an alternative gas must be optimized to afford dielectric properties similar to those of films deposited using N₂O as a conventional gas.⁷ Although some studies have demonstrated improvements in dielectric properties via post-thermal annealing, very few have focused on the comprehensive analysis of dielectric films deposited using an alternative gas and the effect of enhanced dielectric properties on the charge-transport properties of thin-film transistors (TFTs).^13,14 The gate dielectric film plays an important role in device performance and charge transport because it is in direct contact with the active channel and thus affects interfacial characteristics.^15–18

In this study, we demonstrate high-quality SiO₂ films synthesized using NO as an alternative gas via PECVD. We also investigated the effects of post-thermal annealing under different processing temperatures and atmospheres on the dielectric properties of the NO-based SiO₂ films through current–voltage (I–V) and capacitance–voltage (C–V) analyses. The mechanism via which post-thermal annealing enhances the dielectric properties was examined via a compositional analysis of the Si–O bonding states in the films. Subsequently, we fabricated indium–gallium–zinc oxide (IGZO) TFTs bearing NO-based SiO₂ dielectric films and investigated the effect of improved interfacial characteristics on device performance and electrical stability. Finally, we analyzed the charge-transport characteristics of the devices by performing temperature-dependent field-effect mobility (μ) measurements to verify the effect of enhanced dielectric properties of the NO-based SiO₂ films.

Experimental

Fabrication of the NO-based SiO₂ dielectric films and IGZO TFTs

The NO-based SiO₂ films were deposited on a highly doped p-type Czochralski wafer (resistivity = <0.0005 Ω cm; thickness = 525 μm) via PECVD. The substrate was cleaned using the standard Radio Corporation of America method with 10% hydrofluoric acid to remove the native oxide and then dried using N₂ gas to remove the residue. Fig. 1a shows a schematic of the PECVD process utilized in this study. PECVD was performed at a frequency of 13.56 MHz and power of 100 W. The pressure inside the chamber was 0.38 Torr, and the deposition temperature was 250 °C. The [NO]/[SiH₄] gas ratio was controlled from 5 to 50. The N₂O and NO flow rates were 25 to 250 sccm and 18 to 177 sccm for a gas ratio of 5 to 50, respectively, considering the conversion factor between N₂O and NO. Post-thermal annealing of the NO-based SiO₂ films was performed by heating them in a vacuum chamber with either 99% N₂ or 99% O₂ at temperatures ranging from 200 to 400 °C for 1 h. The optical image of a SiO₂ film as deposited on a 4-inch wafer shown in Fig. 1b confirms the successful formation of a wafer-scale SiO₂ film with a clean surface by PECVD. Fig. 1c shows a schematic of the fabrication process for the IGZO TFTs. The SiO₂ films were cleaned with acetone and isopropyl alcohol for 10 min each, and surface residues were removed by UV-ozone treatment for 30 min. Subsequently, a 10-nm-thick IGZO film was deposited on the SiO₂ dielectric film by radio-frequency magnetron sputtering using an IGZO (In : Ga : Zn = 1 : 1 : 1) target at room temperature, followed by channel patterning via wet etching. Post-thermal annealing under different process conditions was conducted prior to the formation of the IGZO channels to enable the comparative investigation of the effects of enhanced dielectric properties on device performance. Finally, aluminum (thickness: 75 nm) source/drain electrodes were formed using electron-beam evaporation to complete the fabrication of a bottom-gate top-contact IGZO TFT.

Fig. 1 (a) Schematic of the fabrication process for the SiO₂ films based on NO alternative gas based on PECVD. (b) Optical image of as-deposited SiO₂ on a 4-inch wafer. (c) Schematic of IGZO-TFT fabrication with alternative NO-gas-based SiO₂ dielectric films.

Material and electrical characterization

The refractive index (RI) of the film was determined using ellipsometry (J.A. Woollam; VASE System, USA). The morphological characteristics of the films were investigated using an X-ray reflectometer (XRR; D8 DISCOVER, BRUKER, USA) and atomic force microscopy (AFM; XE-150, Park Systems, South Korea). Their thickness was measured using a scanning electron microscope (SEM; S-4800, Hitachi, Japan). The chemical bonding state of Si and O in the films was characterized using X-ray photoelectron spectroscopy (XPS; K-Alpha+, Thermo Fisher Scientific, USA) and Fourier transform infrared (FTIR; Nicolet5700, Thermo Electron, USA). The electrical characteristics of the IGZO TFTs were measured using a semiconductor parameter analyzer (Agilent 4155B, Keysight, USA). Areal capacitance was measured using a precision LCR meter (Agilent 4284 A) under an applied voltage of 1 V at 1 kHz. Analyses of temperature-dependent μ were conducted in a vacuum chamber using liquid N₂ at a pressure of approximately 1 Pa and temperatures of 90–300 K.

Results and discussion

Material characteristics of the NO-based SiO₂ films

Fig. 2a shows the deposition rate and RI of the SiO₂ films as a function of the [NO]/[SiH₄] gas ratio. The refractive index was measured at a wavelength of 630 nm to verify whether the deposited film exhibits SiO₂-like characteristics. Fig. S1 (ESI†) shows the refractive index for the SiO₂ films synthesized with [NO]/[SiH₄] gas ratio = 5, 10, 20, 30, 40, and 50 as a function of wavelength, which confirmed the film's SiO₂-like optical properties. As the [NO]/[SiH₄] gas ratio increased from 5 to 50, the deposition rate increased from 20 to 33 Å s⁻¹, whereas the RI decreased from 1.60 to 1.46. This is due to the increased concentration of NO as an oxidant, which promotes the generation of oxygen radicals (O*) through its dissociation.¹⁹ The increased number of reactant radicals resulted in an increase in the deposition rate.¹⁹ The deposition rate increased, and RI significantly decreased when the gas ratio was increased from 5 to 10. Typically, as the gas ratio increases, the amount of oxygen in the system increases, thereby increasing the deposition rate and decreasing RI.^20,21 In addition, because NO is reactive, an increase in this gas can rapidly increase the deposition rate.²² However, no significant changes in the deposition rate or RI were observed in this study when the gas ratio was increased from 30 to 50, indicating that the effect of the gas ratio on the films became saturated. Fig. 2b shows the experimental and simulated XRR profiles of the NO-based SiO₂ film with a gas ratio of 50. The film thickness, density, and roughness were determined to be ∼90 nm, ∼2.25 g cm⁻³, and ∼0.37 nm, respectively. These values are similar to those of previously reported SiO₂ films based on conventional N₂O, indicating that NO is a suitable alternative gas for the effective fabrication of SiO₂ films by PECVD.²³ Notably, the simulated XRR curve, which was extracted from the structural properties of the thin films using the interface formula and recursive Parratt formalism, matched the experimental curve well.²³Fig. 2c shows the capacitance values of the NO-based SiO₂ films synthesized with different [NO]/[SiH₄] gas ratios. When the gas ratio was 10, a capacitance of 47.7 pF was obtained. As the gas ratio increased, the capacitance of the films decreased. When the gas ratio was 50, the capacitance dropped to 20.5 pF. This finding may be attributed to the increase in binding energy with increasing gas ratio, indicating a transition of the film structure from the SiON to the SiO_x phase.⁷ The dielectric constant of the films was determined using the following equation:⁷where C is the capacitance of the film, A is the area of the electrode, d is the thickness of the dielectric film, and ε₀ is the vacuum permittivity. As shown in the inset of Fig. 2d, the thickness of the NO gas-based SiO₂ film was determined by SEM to be ∼90 nm, which is identical to that obtained from the XRR analysis. Fig. 2d shows a reduction in dielectric constant from 11.9 to 5.1 as the gas ratio increased from 10 to 50. As the oxygen content increases, the nitrogen content in the film decreases, leading to the formation of a more stoichiometric SiO_x structure.²⁴ Notably, the dielectric constant of the NO gas-based SiO₂ films was higher than that of conventional N₂O-based SiO₂ films (∼3.9). This result can be attributed to the incomplete Si–O–Si bonding structure of the SiO₂ film.²⁵ A relatively low Si–O–Si binding energy can induce oxygen vacancies, which affect the intrinsic dielectric properties of the PECVD-based SiO₂ films by increasing the repulsion between Si atoms and the Si bond length.²⁵

Fig. 2 (a) Deposition rates and refractive indices of NO-gas-based SiO₂ films for different [NO]/[SiH₄] gas ratios. (b) Experimental and simulated curves of XRR profiles. (c) C–V characteristics and (d) dielectric constants of NO-gas-based SiO₂ films prepared with different [NO]/[SiH₄] gas ratios: the inset shows a cross-sectional SEM image of the NO-gas-based SiO₂ film prepared by PECVD.

Electrical characteristics of NO gas-based SiO₂ films annealed at different temperatures under different atmospheres

Fig. 3a and b show the electrical characteristics of the N₂O- and NO-based SiO₂ films, respectively. For fair comparison, the N₂O-based SiO₂ films have the same thickness of ∼90 nm as the NO-based SiO₂ films (Fig. S2, ESI†). Based on the J–E measurements, the as-deposited N₂O- and NO-based SiO₂ films exhibited current densities of 5.25 × 10⁻⁹ and 2.65 × 10⁻⁸ A cm⁻², respectively. In addition, the as-deposited N₂O- and NO-based SiO₂ films exhibited capacitances of 38 and 42 nF cm⁻², respectively, obtained from the C–V measurements. Relatively high leakage current density and capacitance were obtained when N₂O was replaced with NO during the fabrication of the PECVD-based SiO₂ films. The NO-based SiO₂ films with relatively incomplete Si–O–Si bonds have defect states, including oxygen vacancies, which result in a higher leakage current than in the N₂O-based SiO₂ films.²⁵ To improve the dielectric properties of the proposed SiO₂ films, we performed post-thermal annealing under various processing conditions. As shown in Fig. 3c, the post-thermally annealed SiO₂ films exhibited a leakage current of 7.61 × 10⁻¹⁰ A cm⁻², which is lower than that of the as-deposited film. Leakage currents in SiO₂ films are induced by conduction through impurities or trapped charges in trap sites.²⁶ Thermal energy during the annealing process can rearrange unstable bonding structures to densify the film and remove trapped charges at trap sites, thereby improving the dielectric properties of the NO-based SiO₂ films.²⁷

Fig. 3 (a) J–E and (b) C–V characteristics of the PECVD-based SiO₂ films deposited using conventional N₂O and alternative NO gases. (c) J–E and (d) C–V characteristics of the PECVD-based SiO₂ films annealed at 300 °C under various atmospheres. (e) J–E and (f) C–V characteristics of PECVD-based SiO₂ films annealed at temperatures of 200, 300, and 400 °C.

Based on the J–E curves, we confirmed that the currents of the PECVD-based SiO₂ films were significantly suppressed in a similar manner regardless of the annealing atmosphere. These findings may be attributed to reductions in charge impurities, which can affect the dielectric quality of the SiO₂ films, after post-thermal annealing.^4,14Fig. 3d shows the C–V characteristics of the prepared SiO₂ films after post-thermal annealing under various atmospheres. Unlike the J–E curves, the C–V characteristics of the films were affected by the processing atmosphere. The decrease in capacitance of the PECVD-based SiO₂ films to 37.7 nF cm⁻² after N₂ annealing confirmed that post-thermal annealing in a N₂ atmosphere provided the greatest restoration, compatible with that of thermally grown SiO₂ films.^4,13Fig. 3e shows the optimal annealing temperature under an N₂ atmosphere. Thermal annealing at 200 °C resulted in a leakage current density lower than that before annealing but higher than those obtained at 300 and 400 °C. Similar currents were obtained when the films were annealed at 300 and 400 °C. These results indicate that the effect of N₂ annealing on the J–E characteristics was limited at 400 °C. We analyzed the breakdown electric strength of the NO-based SiO₂ films before and after N₂ annealing, as shown in Fig. S3 (ESI†). The breakdown electric field of the as-deposited SiO₂ film was approximately 8 MV cm⁻¹, which is comparable to that of the 100 nm-thick SiO₂ film in a previously reported study.²⁸ The breakdown of the N₂-annealed SiO₂ film was not observed even when 100 V was applied, which is the voltage limit of the measurement equipment. As depicted in Fig. 3f, thermal annealing at 200 °C resulted in a higher capacitance (39.3 nF cm⁻²) than those obtained at 300 and 400 °C. It is very likely that similar capacitance values were obtained when the films were annealed at 300 and 400 °C. The dielectric constant of the N₂-annealed SiO₂ films was determined to be ∼3.9, which is close to the reported value of 3.9.^29,30 Based on these results, 300 °C was selected as the optimal temperature for post-thermal annealing in an N₂ atmosphere.

Mechanism of the improvement in dielectric properties by N₂ annealing

We investigated the mechanism behind the improved dielectric properties of the PECVD-based SiO₂ films after N₂ annealing. Fig. 4a shows the FTIR profiles of the NO-based SiO₂ films before and after N₂ annealing. A peak associated with the asymmetric stretching vibration of the Si–O–Si bond appeared at 1060 cm⁻¹, and its intensity increased after N₂ annealing.³¹ This result indicates an increase in not only the Si–O–Si bond angle but also the quantity or strength of Si–O bonds following post-thermal annealing.^4,14,31 Additionally, the surface diffusion induced by N₂ annealing generates a tensile stress in the SiO₂ film, mitigating compressive stress and reducing the formation of micropores.³² The subsequent rearrangement of trapped charges affects the dielectric properties of the NO-based SiO₂ films.³³Fig. 4b shows the O 1s spectra of the proposed SiO₂ films before and after N₂ annealing. Prior to annealing, the Si–O–Si peak was observed at 532.88 eV.³³ After N₂ annealing, the binding energy of this peak shifted to 532.43 eV, and its intensity increased accordingly, thereby validating the rearrangement of trapped charges in the NO-based SiO₂ films.³⁴ The enhancement of the Si–O–Si peak intensity is due to the stabilization of the unstable bonding structure through thermal energy during the post-thermal annealing process.³⁴ It is presumed that post-thermal annealing in an N₂ atmosphere can effectively cure defect states (or remove charge impurities), thereby restoring the intrinsic dielectric properties of the NO-based SiO₂ films.^33,34Fig. 4c shows the Si 2p spectra of the prepared SiO₂ films; here, only one chemical structure, which decomposes into the Si oxidation state Si⁴⁺, was observed.^34,35 This peak was observed at 103.73 eV before N₂ annealing and shifted to 103.38 eV with an increase in intensity after N₂ annealing. Furthermore, the full width at half maximum decreased from 1.98 to 1.60 eV, indicating the formation of a dense and complete surface morphology after N₂ annealing.¹⁴ The Si 2p binding energy related to SiO₂ is known to vary with the oxygen composition.³⁶ Typically, stoichiometric SiO₂ films, which are characterized by Si atoms surrounded by four O atoms, exhibit a binding energy of 103.3 eV.³⁷ The peak observed at 103.38 eV confirms that the intrinsic material properties of the NO-based SiO₂ films were restored after optimal post-thermal annealing.³⁷ Fig. S4a and b (ESI†) show the N 1s and wide-XPS spectra of the NO-based SiO₂ films before and after N₂ annealing, respectively. The N 1s contents of both the as-deposited and N₂-annealed SiO₂ films were determined to be within ∼1 at%, indicating that very little nitrogen was contained in the film, and there is no significant difference between before and after N₂ annealing. Fig. 4d and e show the morphological characteristics of the as-deposited SiO₂ and N₂-annealed SiO₂ films, respectively, obtained from AFM measurements. The root-mean-square roughness was 0.583 nm for the as-deposited NO-based SiO₂ film and 0.443 nm for the N₂-annealed NO-based SiO₂ film, indicating that the morphological characteristics were improved by the proposed N₂ annealing. This can be attributed to the surface diffusion by thermal energy during the annealing process, which improved the dielectric properties of the NO-based SiO₂ films.³⁸

Fig. 4 (a) FT-IR spectra and (b) and (c) O 1s and Si 2p spectra of as-deposited and N₂-annealed SiO₂ films. AFM images of (d) as-deposited SiO₂ and (e) N₂-annealed SiO₂ films.

Device performance of IGZO TFTs prepared from NO-based SiO₂ dielectric films without and with N₂ annealing

Fig. 5a shows the transfer characteristics of IGZO TFTs prepared from N₂O- and NO-based SiO₂ dielectric films. The IGZO TFT prepared from the NO-based SiO₂ film exhibited a μ of 4.3 cm² (V s)⁻¹, a threshold voltage (V_th) of −5 V, an on/off ratio of 3 × 10⁶, and a subthreshold swing (S.S.) of 1.23 V dec⁻¹. By contrast, the IGZO TFT prepared from the N₂O-based SiO₂ film showed a μ of 6.8 cm² (V s)⁻¹, a V_th of −3 V, an on/off ratio of 1.58 × 10⁸, and an S.S. of 0.49 V dec⁻¹. Fig. 5b compares the transfer characteristics between an IGZO TFT prepared from a NO-based SiO₂ dielectric film annealed at 300 °C in an N₂ atmosphere and that prepared from an as-deposited NO-based SiO₂ dielectric film. In this study, post-thermal annealing was performed on the gate dielectric layer prior to IGZO sputtering to maintain the material properties of the channel film, thereby allowing for a comparative investigation of how the enhanced dielectric properties of the NO-based SiO₂ films affect the device performance of the resulting IGZO TFTs. Improvements in the off-state current, V_th, on/off ratio, and field-effect mobility owing to enhancements in interfacial characteristics between the gate dielectric and channel layers in the IGZO TFT prepared from a NO-based SiO₂ film with enhanced dielectric properties were observed.³⁹ Fig. S5 (ESI†) shows the hysteresis characteristics of the IGZO TFTs with NO-based SiO₂ films before and after N₂ annealing. The IGZO TFT with an as-deposited SiO₂ film exhibited a hysteresis window of approximately 1.3 V, whereas that after N₂ annealing was almost 0 V, which can be attributed to the suppressed charge trapping at the interface between the gate dielectric and channel layers.⁴⁰ The reduction of defect states in the gate dielectric film and/or improved interfacial characteristics are presumed to suppress the leakage current and improve S.S. significantly, resulting in enhanced IGZO TFT performance. The interfacial trap states that hinder the charge transport of the IGZO TFT can be significantly reduced after annealing, which not only improved the S.S. performance but also enhanced the field-effect mobility.⁴¹ Significantly, the field-effect mobilities of the IGZO TFTs after annealing can be improved by suppressing the trapping effect induced in the trap states.⁴² The charge trapping induced in the defect states hinders the favorable switching in the subthreshold region, thus increasing the S.S. value.⁴³ The N₂ annealing proposed in this study reduced the density of trap sites, which can suppress the trapping effect and thereby improve the switching properties with low S.S. values.⁴⁴ The key parameters of on/off ratio, μ, V_th, and S.S. obtained from 15 IGZO TFTs fabricated from as-deposited and N₂-annealed SiO₂ films in different batches at different times were analyzed to further investigate the reliability of the proposed method. As shown in Fig. 5c–f, on average, an on/off ratio of 4.4, μ of 4.1 cm² (V s)⁻¹, V_th of −5.3 V, and S.S. of 1.0 V dec⁻¹ were observed in the IGZO TFTs prepared from an as-deposited NO-based SiO₂ dielectric film. By contrast, on average, an on/off ratio of 6.4, μ of 6.1 cm² (V s)⁻¹, V_th of −1.8 V, and S.S. of 0.7 V dec⁻¹ were observed in the IGZO TFTs prepared from an N₂-annealed NO-based SiO₂ dielectric film. These findings indicate that the performance of the IGZO TFTs was significantly improved by enhanced interface between the gate dielectric and channel films as well as the suppressed defect states.⁴⁵

Fig. 5 (a) Transfer curves of IGZO TFTs with PECVD-based SiO₂ dielectric films based on conventional N₂O and alternative NO gases. (b) Transfer curves of IGZO TFTs with alternative gas-based SiO₂ dielectric films before and after N₂ annealing. Statistical plots of (c) log (on/off current ratio), (d) field-effect mobility in a linear region, (e) V_th, and (f) S.S. obtained from 15 IGZO-TFTs with NO-based SiO₂ dielectric films as deposited and after N₂ annealing.

Charge-transport properties of IGZO TFTs prepared from NO-based SiO₂ dielectric films without and with N₂ annealing

We performed temperature-dependent μ measurements to investigate the effect of enhanced dielectric properties on the charge-transport characteristics of the IGZO TFTs. Fig. 6a shows the Arrhenius plots of IGZO TFTs prepared from NO-based SiO₂ dielectric films without and with N₂ annealing at measurement temperatures ranging from 90 to 300 K. An increase in μ with increasing measurement temperature was observed, indicating that thermally activated transport is dominant in the IGZO TFTs.¹⁵Fig. 6b shows the activation energy (E_a) determined from the slope of the Arrhenius plots by fitting the logarithm of the thermally activated μ to the following equation:⁴⁶where k is the Boltzmann constant and T is the temperature. E_a represents the minimum energy required for charge carriers to move to the transport level through defect states. Based on the activation energy analysis, the density of trap states, which affects the charge transport properties and field-effect mobility, can be determined.⁴⁷ The E_a of the IGZO TFT prepared from the as-deposited SiO₂ dielectric film was determined to be 47.6 meV, while that of the IGZO TFT prepared from the N₂-annealed SiO₂ dielectric film was determined to be 38.1 meV. A low E_a indicates that the Fermi level is significantly shifted toward the delocalized mobility edge, resulting in improved charge transport.¹⁵ To further investigate the effect of improvements in dielectric properties on device performance, we determined the DOS of the IGZO TFTs using E_a as follows:⁴⁶where N(E) is the DOS of the band gap and d is the effective thickness of the deposition layer. As seen in Fig. 6c, the DOS of the IGZO TFTs increased exponentially in the conduction band. Most charge carriers in the IGZO TFTs are in the tail state, and a small number of charges can move to the transport energy level.⁴⁶ The DOS of the IGZO TFT prepared from the as-deposited SiO₂ dielectric film ranged from 47.6 to 117.0 meV (9.32 × 10¹⁸–6.41 × 10¹⁹ eV⁻¹ cm⁻³) below the conduction band, while that of the IGZO TFT prepared from the N₂-annealed SiO₂ dielectric film ranged from 38.0 to 107.8 meV (5.33 × 10¹⁸–3.17 × 10¹⁹ eV⁻¹ cm⁻³) below the conduction band. Fig. 6d shows the interfacial trap density (D_it) values of the IGZO TFTs prepared from NO-based SiO₂ dielectric films with and without N₂ annealing; this property is an essential factor in identifying the interfacial characteristics between the channel and dielectric layers.⁴⁵D_it was determined from the S.S. of the films using the following equation:⁴⁵

Fig. 6 (a) Comparison of Arrhenius plots as a function of measuring temperature (90–300 K) and (b) activation energy as a function of gate voltage in IGZO TFTs with NO-based SiO₂ dielectric films as deposited and after N₂ annealing (c) DOS as a function of energy below the conduction band and (d) interfacial trap density (D_it) in IGZO TFTs with NO-based SiO₂ dielectric films as deposited and after N₂ annealing.

The D_it of the IGZO TFT prepared from the as-deposited SiO₂ dielectric film was determined to be 4.62 × 10¹² cm⁻² eV⁻¹, while that of the IGZO TFT prepared from the N₂-annealed SiO₂ dielectric film was determined to be 1.59 × 10¹² cm⁻² eV⁻¹. A decrease in D_it is in good agreement with improved interfacial characteristics between the IGZO channel and NO-based SiO₂ dielectric film by N₂ annealing, thus supporting improvements in device performance and charge transport.⁴⁵ The conduction band offset between SiO₂ (low-k dielectric) and IGZO is relatively large compared to that between a high-k dielectric and IGZO, thus suppressing the interface dipole-induced effect.⁴⁸ For this reason, the effects of shifts in the conduction band offset or changes in interface dipoles on charge injection, carrier confinement, or threshold voltage stability in the long term are not dominant in this study.

Electrical stability of IGZO TFTs prepared from NO-based SiO₂ dielectric films with and without N₂ annealing

Next, we investigated the effect of N₂ annealing on the electrical stability of IGZO TFTs prepared from the NO-based SiO₂ dielectric films by performing positive bias stress (PBS) and negative bias stress (NBS) tests to verify the effect of improved interfacial characteristics on device stability. Fig. 7a and b show the respective transfer characteristics of IGZO TFTs prepared with the as-deposited NO-based SiO₂ dielectric films under a PBS and NBS applied for 20 000 s on a linear scale. V_th shifted positively under the PBS but shifted negatively under the NBS. Typically, positive shifts in V_th under a PBS are attributed to charge trapping at the interface between the channel and dielectric layers or charge injection into the gate dielectric layer.^49,50 On the other hand, negative shifts in V_th under an NBS are attributed to charge trapping at the interface or oxygen vacancies acting as a shallow donor state.^45,50Fig. 7c and d show the respective transfer characteristics of IGZO TFTs prepared with the N₂-annealed NO-based SiO₂ dielectric films under a PBS and NBS applied for 20 000 s on a linear scale. The IGZO TFTs demonstrated a relatively small ΔV_th under both the PBS and NBS, which can be attributed to the suppression of charge trapping owing to the restoration of the intrinsic dielectric properties of the PECVD-based SiO₂ films.⁴⁵ A V_th shift of approximately 13.3 V was observed in the IGZO TFT prepared with the as-deposited NO-based SiO₂ dielectric film after the PBS test, but a V_th shift of only 6.0 V was achieved when the device was prepared from an N₂-annealed NO-based SiO₂ dielectric film (Fig. 7e). Moreover, V_th negatively shifted to −7.2 V in the IGZO TFT prepared with the as-deposited NO-based SiO₂ dielectric film after the NBS test, whereas a V_th shift of only −3 V was observed in the device prepared from the N₂-annealed NO-based SiO₂ dielectric film (Fig. 7f). The instability of the V_th shift under a prolonged electrical bias stress was greatly suppressed in IGZO TFTs prepared from the N₂-annealed NO-based SiO₂ films. This improvement in electrical stability indicates that the trap states for charge trapping were effectively reduced by N₂ annealing.³²

Fig. 7 Transfer characteristics of IGZO TFTs with NO-based SiO₂ dielectric films before N₂ annealing as a function of stress time during a (a) PBS and (b) NBS and after N₂ annealing as a function of stress time during a (c) PBS and (d) NBS. Comparison of ΔV_th as deposited and after N₂ annealing as a function of stress time for the (e) PBS and (f) NBS.

Conclusions

The electrical characteristics of PECVD-based SiO₂ films prepared using NO gas as an alternative to N₂O gas were investigated, and the device performance of IGZO TFTs fabricated from these films was analyzed. The I–V and C–V characteristics of the NO-based SiO₂ films were evaluated, and the effects of post-thermal annealing under different processing conditions were investigated. IGZO TFTs consisting of NO-based SiO₂ dielectric films with and without N₂ annealing were fabricated, and the origin of the improvements in device performance underlying the charge-transport mechanism was investigated by performing temperature-dependent μ measurements and material characterization. The enhancements in dielectric properties and/or interfacial characteristics between the channel and dielectric layers following post-thermal annealing indicated the potential application of alternative gas-based SiO₂ films as a replacement for conventional gas-based SiO₂ films in the semiconductor industry.

Author contributions

S. P. conducted the synthesis, characterization, and electrical measurements as well as wrote the original article. E. K., Y. J., Y. K., supported the experiments and investigated the analysis as well as reviewed the article. T. H., J. Y., and Y. K. directed and supervised the entire experiments and contributed to the analysis of data while developing and revising the manuscript. S. P and E. K. contributed equally to this work. All authors read and approved the final manuscript.

Data availability

Data will be made available on request.

Conflicts of interest

The authors declare no competing interest.

Acknowledgements

This work was supported by the Technology Innovation Program of Ministry of Trade, Industry and Energy (MOTIE) (project no. RS-2023-00266568).

References

1. A. I. Kingon, J. P. Maria and S. K. Streiffer, Nature, 2000, 406, 1032–1038.
2. T. Usui, C. A. Donnelly, M. Logar, R. Sinclair, J. Schoonman and F. B. Prinz, Acta Mater., 2013, 61, 7660–7670.
3. J. K. Saha, A. Ali, R. N. Bukke, Y. G. Kim, M. M. Islam and J. Jang, IEEE Trans. Electron Devices, 2021, 68, 1063–1069.
4. C. E. Vianan, N. I. Morimoto and O. Bonnaud, Microelectron. Reliab., 2000, 40, 613–616.
5. M. I. Alayo, I. Pereyra and M. N. P. Carreño, Thin Solid Films, 1998, 332, 40–45.
6. P. Yang, K. W. Tang, H. Yang, C. Tong, L. Zhang, D. Y. F. Lai, Y. Hong, L. Tan, W. Zhu and C. Tang, J. Hydrol., 2023, 617, 128876.
7. Y. Jeong, M. P. Nguyen, J. K. Song, Y. S. Kim, Y. B. Chung, W. S. Jeon, J. Jo, Y. Kim, D. P. Pham and J. Yi, Opt. Mater., 2024, 148, 114970.
8. L. C. Xu, Y. Wo, M. E. Meyerhoff and C. A. Siedlecki, Acta Biomater., 2017, 51, 53–65.
9. A. D. French, K. P. Hobbs, R. M. Cox and I. J. Arnquist, Analyst, 2024, 149, 5812–5820.
10. J. H. Lee, C. H. Jeong, J. T. Lim, N. G. Jo, S. J. Kyung and G. Y. Yeom, J. Korean Phys. Soc., 2005, 46, 890–894.
11. K. Omri, I. Najehm and L. El-Mir, Ceram. Int., 2016, 42, 8940–8948.
12. H. Guo, A. Baker, J. Guo and C. A. Randall, J. Am. Ceram. Soc., 2016, 99, 3489–3507.
13. K. Tachiki, M. Kaneko, T. Kobayashi and T. Kimoto, Appl. Phys. Express, 2020, 13, 121002.
14. M. S. Haque, H. A. Naseem and W. D. Brown, Thin Solid Films, 1997, 308–309, 68–73.
15. S. J. Park and T. J. Ha, IEEE Trans. Electron Devices, 2023, 70, 99–104.
16. W. B. Lee, Y. S. Kim and J. S. Park, J. Inf. Disp., 2024, 25, 179–185.
17. J. K. Saha and J. Jang, ACS Nano, 2024, 18, 30484–30496.
18. S. J. Park and T. J. Ha, Thin Solid Films, 2020, 708, 138113.
19. A. Kumagai, K. Ishibashi, G. Xu, M. Tanaka, H. Nogami and O. Okada, Vacuum, 2002, 66, 317–322.
20. C. C. Wang, D. S. Wuu, S. Y. Lien, Y. S. Lin, C. Y. Liu, C. H. Hsu and C. F. Chen, Int. J. Photoenergy, 2012, 2012, 890284.
21. Y. T. Cheng, J. J. Ho, W. J. Lee, S. Y. Tsai, Y. A. Lu, J. J. Liou, S. H. Chang and K. L. Wang, Int. J. Photoenergy, 2010, 268035.
22. V. Blagojevic, E. Flaim, M. J. Y. Jarvis, G. K. Koyanagi and D. K. Bohme, Int. J. Mass Spectrom., 2006, 249–250, 385–391.
23. G. Ceriola, F. Iacona, F. L. Via, V. Raineri, E. Bontempi and L. E. Depero, J. Electrochem. Soc., 2001, 148, F221.
24. C. Y. Chou, C. H. Lin, W. H. Chen, B. J. Li and C. Y. Liu, Thin Solid Films, 2020, 709, 138198.
25. K. Hirose, H. Nohira, K. Azuma and T. Hattori, Prog. Surf. Sci., 2007, 82, 3–54.
26. M. Sometani, R. Hasunuma, M. Ogino, H. Kuribayashi, Y. Sugahara and K. Yamabe, Jpn. J. Appl. Phys., 2009, 48, 05DB03.
27. C. Liu, X. Li, L. X. Qian, J. Tian and X. Zhang, APL Mater., 2024, 12, 081111.
28. T. Kim, C. Oh, S. H. Park, J. W. Lee, S. I. Lee and B. S. Kim, AIP Adv., 2021, 11, 115126.
29. C. Y. Ng, T. P. Chen, L. Ding, Y. Liu, M. S. Tse, S. Fung and Z. L. Dong, Appl. Phys. Lett., 2006, 88, 063103.
30. J. McPherson, J. Y. Kim, A. Shanware and H. Mogul, Appl. Phys. Lett., 2003, 82, 2121–2123.
31. M. Vishwas, K. N. Rao, A. R. Phani, K. V. A. Gowda and R. P. S. Chakradhar, Spectrochim. Acta, Part A, 2011, 78, 695–699.
32. M. C. Kwan, K. H. Chemg, P. T. Lai and C. M. Che, Solid-State Electron., 2007, 51, 77–88.
33. S. Li, J. Xu, L. Wang, N. Yang, X. Ye, X. Yuan, H. Xiang, C. Liu and H. Li, Mater. Sci. Semicond. Process., 2020, 106, 104777.
34. E. Desbiens, R. Dolbec and M. A. E. Khakani, J. Vac. Sci. Technol., A, 2002, 20, 1157–1161.
35. J. Xu, S. Li, W. Zhang, S. Yan, C. Liu, X. Yuan, X. Ye and H. Li, Appl. Surf. Sci., 2021, 554, 144889.
36. F. Rochet, C. Poncey, G. Dufour, H. Roulet, C. Guillot and F. Sirotti, J. Non-Cryst. Solids, 1997, 216, 148–155.
37. D. Kim, K. M. Kim, H. Han, J. Lee, D. Ko, K. R. Park, K. B. Jang, D. Kim, J. S. Forrester, S. H. Lee, J. C. Kim and S. Mhin, Sci. Rep., 2022, 12.
38. S. Pandey, P. Kothari, S. Verma and K. J. Rangra, J. Mater. Sci.: Mater. Electron., 2017, 28, 760–767.
39. M. H. Cho, C. H. Choi, H. J. Seul, H. C. Cho and J. K. Jeong, ACS Appl. Mater. Interfaces, 2021, 13, 16628–16640.
40. Y. Li, X. Huang, C. Liao, R. Wang, S. Zhang, L. Zhang and R. Huang, Solid-State Electron., 2022, 197, 108459.
41. J. H. Bae and Y. Choi, Solid-State Electron., 2012, 72, 44–47.
42. S. J. Park, S. R. Park, J. Na, W. S. Jeon, Y. Kang, S. Ham, Y. Kim, Y. B. Chung and T. J. Ha, J. Mater. Chem. C, 2024, 12, 19071–19077.
43. S. G. Jeong, H. J. Jeong and J. S. Park, IEEE Trans. Electron Devices, 2021, 68, 1670–1675.
44. S. J. Park and T. J. Ha, Appl. Surf. Sci., 2023, 622, 156959.
45. S. J. Park, S. R. Park, W. S. Jeon, J. M. Na, J. H. Lim, S. Ham, Y. B. Chung and T. J. Ha, J. Mater. Chem. C, 2024, 12, 11361–11367.
46. Y. Mei, P. J. Diemer, M. R. Niazi, R. K. Hallani, K. Jarolimek, C. S. Day, C. Risko, J. E. Anthony, A. Amassian and O. D. Jurchescu, Proc. Natl. Acad. Sci. U. S. A., 2017, 114, E6739–E6748.
47. J. A. Merlo and C. D. Frisbie, J. Phys. Chem. B, 2004, 108, 19169–19179.
48. H. Kim, K. Im, J. Park, T. Khim, H. Hwang, S. Kim, S. Lee, M. Song, P. Choi, J. Song and B. Choi, IEEE Electron Device Lett., 2020, 41, 737–740.
49. S. J. Park and T. J. Ha, IEEE Electron Device Lett., 2023, 44, 642–645.
50. H. J. Oh, Y. S. Kim, H. J. Jeong, S. Lee, J. S. Park and J. S. Park, J. Inf. Disp., 2024, 25, 295–303.

---

Footnotes

† Electronic supplementary information (ESI) available. See DOI: https://doi.org/10.1039/d5tc00213c

‡ These authors contributed equally to this work

---