Influence of N₂/O₂ Partial Pressure Ratio during Channel Layer Deposition on the Temperature and Light Stability of a-InGaZnO TFTs

Round 1

Reviewer 1 Report

Authors investigated the effect of adding nitrogen during the deposition process of the a-IGZO  channel layer and found that moderate doping is required for performance enhancement. Although results are sound and clearly presented, I recommend the paper for publication after the following revisions:

 Please include values of parameters( e.g. of Pn/o , Vth, and Er) in the abstract.

Replace "back gate structure" with "bottom gate" or "inverted staggered" in experiments section.

The word "dramatically" seems to be inappropriate for a decrease in dVth from 1.2 V to 3.0 V. Please consider removing it from abstract and text.

It is common practice to arrange figure legends chronologically from the initial characteristics followed by increasing stress time (Figure 2).

On page 5 line 135, "exiting" should be "existing". Revision of the English language style and writing is required in other areas of the manuscript. Please check carefully.

Author Response

Reviewer 1

Authors investigated the effect of adding nitrogen during the deposition process of the a-IGZO channel layer and found that moderate doping is required for performance enhancement. Although results are sound and clearly presented, I recommend the paper for publication after the following revisions:

Please include values of parameters (e.g. of Pn/o, Vth, and Eτ) in the abstract.

Response:

Thank you very much for your suggestion. We have added the values of P_N/O, Vth, and Eτ in the abstract.

Replace "back gate structure" with "bottom gate" or "inverted staggered" in experiments section.

Response:

We have replaced the back gate structure with inverted staggered in experiments section.

The word "dramatically" seems to be inappropriate for a decrease in dVth from 1.2 V to 3.0 V. Please consider removing it from abstract and text.

Response:

Thank you very much for your suggestion. We have removed the word “dramatically” in the abstract and text.

It is common practice to arrange figure legends chronologically from the initial characteristics followed by increasing stress time (Figure 2).

Response:

We have revised the figure legends chronologically from the initial characteristics followed by increasing stress time in figure 2.

On page 5 line 135, "exiting" should be "existing". Revision of the English language style and writing is required in other areas of the manuscript. Please check carefully.

Response:

Thank you very much for you to point out a spelling mistake. We have replaced “exiting” by existing in the manuscript. Meanwhile, the English language style and writing are checked by a native English speaking colleague.

Author Response File: Author Response.docx

Reviewer 2 Report

This paper reported a reliability study for N-doped IGZO-TFT and includes important information for the researcher in this field. Some description and correction are required for publication.

1.       The TFT performance, especially in mobility, is rather poor compared with the standard IGZO-TFT. IGZO-TFTs usually shows the mobility more than 10 cm²/Vs, though this paper has only 2.2 in the conventional condition. The author should address that this value can be evaluated as a reference since the device having poor performances can be easily improved by small differences in the fabrication process. In addition, the explanation of the deviation of each performance is mandatory.

2.       Usually, oxygen deficiency region has no oxygen. Why can you assign the Ov although there is no oxygen at the region?

3.       In Fig. 5(a)-(c), the position of symbols of O_I and O_III are incorrect. O_I and O_III should be located at 530.1 eV and 532.4 eV, respectively, according to this manuscript.

4.       Fig. 5(d): Showing of N1s for the film with undoped and 40% P_N/O condition is required as well, in order to discuss a relation with N-bonding state, Ov, and reliability.

Author Response

Reviewer 2

This paper reported a reliability study for N-doped IGZO-TFT and includes important information for the researcher in this field. Some description and correction are required for publication.

1. The TFT performance, especially in mobility, is rather poor compared with the standard IGZO-TFT. IGZO-TFTs usually shows the mobility more than 10 cm²/Vs, though this paper has only 2.2 in the conventional condition. The author should address that this value can be evaluated as a reference since the device having poor performances can be easily improved by small differences in the fabrication process. In addition, the explanation of the deviation of each performance is mandatory.

Response:

Thank you very much for your question. It has been demonstrated that carrier transport of a-IGZO TFTs is largely through the overlap of heavy post-transition metal cations ns orbitals [1*-2*]. As a result, a-IGZO TFTs can exhibit high carrier mobility (>10 cm²/Vs). However, high density localized states originating from O_v exist within the bandgap of a-IGZO active layer due to the disordered amorphous nature, which would considerably affect the carrier transport properties of a-IGZO TFTs. For example, it has been reported that the field-effect mobility (μ_FE) is reduced from 12 to 0.1 cm²/V·s when the oxygen partial pressure (P_O2) is increased from 0.3 to 1.5 mTorr, which is attributed to the Fermi level pinning below the conduction band edge induced by the increase of subgap density of states [3*]. Thus, in this work, a low value of μ_FE can be caused by the high density O_v (35%) existing within the active layer based on XPS analysis. Meanwhile, it is found that the μ_FE is increased when O_v is decreased by moderate N doping within the a-IGZO film. In contrast, the O_v is increased by excess N doping within the a-IGZO film. Correspondingly, the μ_FE of TFT is degraded. These result mean that the change of μ_FE is associated with the O_v in the a-IGZO film. In addition, although μ_FE is an important parameter to evaluate the performance of TFTs, the motivation of this study is to improve the electrical stability of a-IGZO TFTs by suppressing the O_v. Thus, the authors think that the low value of μ_FE can be evaluated as a reference for this study.

Finally, the change of performance parameters has been explained in the manuscript. For example, the improved electrical properties for a-IGZO TFT fabricated with 20 % P_N/O can be determined by the decrease of trap density in the device active region.In contrast, the V_th and SS of the a-IGZO TFT with 40 % P_N/O are increased, which indicates that the new trap states are generated by excess N doping.

[1*] Nomura, K.; Ohta, H.; Takagi, A.; Kamiya, T.; Hirano, M.; Hosono, H. Room-temperature fabrication of transparent flexible thin-film transistors using amorphous oxide semiconductors. Nature. 2004, 432, 488.

[2*] Hsieh, H. H; Kamiya, T.; Nomura, K.; Hosono, H.; Wu, C. C. Modeling of amorphous InGaZnO 4 thin film transistors and their subgap density of states. Appl. Phys. Lett. 2008, 92, 133503.

[3*] Hung, M. P.; Chare, C.; Nag, M.; Meux, A. J.; Genoe, J.; Steudel, S. Effect of High Oxygen Partial Pressure on Carrier Transport Mechanism in a-InGaZnO TFTs. IEEE Electron Device Lett. 2018, 65, 2833.

2. Usually, oxygen deficiency region has no oxygen. Why can you assign the Ov although there is no oxygen at the region?

Response:

Actually, there is no oxygen ions in oxygen deficiency (O_V) region. But the characterized area of a-IGZO thin film is ~ 300μm × 300μm by x-ray photoelectron spectroscopy (XPS) in this work. Therefore, the O 1s XPS spectra of the a-IGZO thin film is macro-statistics analysis. Because the oxygen deficient region in oxide lattices will affect the whole O 1s spectra of the a-IGZO thin film, the oxygen vacancies can be decomposed from O 1s peak. Meanwhile, the relative amount of O_V in a-IGZO film can be calculated by the proportion of peak area O_V to the whole area O 1s.

3. In Fig. 5(a)-(c), the position of symbols of O_I and O_III are incorrect. O_I and O_III should be located at 530.1 eV and 532.4 eV, respectively, according to this manuscript.

Response:

Thank you very much for you to point out this mistake. We have revised the position of symbols of O_I and O_III in Fig. 5(a)-(c).

4. Fig. 5(d): Showing of N1s for the film with undoped and 40% P_N/O condition is required as well, in order to discuss a relation with N-bonding state, O_v, and reliability.

Response:

Thank you very much for your question. In fact, there is no N 1s peak in undoped a-IGZO thin film due to without incorporating N into a-IGZO thin film. The N 1s spectrum of the a-IGZO: 40 % P_N/O film is also characterized, as shown in Fig.1*. Compared to the a-IGZO: 20 % P_N/O film, the relative amount of N-Ga existing in a-IGZO: 40 % P_N/O film is increased from 34% to 38%, which agrees with the increase of P_N/O. However, it is found that the O_v is increased when the P_N/O increases to 40 % P_N/O, which is because the heavy N doping in a-IGZO film will suppress the bonding of O and Ga. Correspondingly, the reliability of the TFT is degraded. This result suggests that the reliability of a-IGZO TFTs is associated with the O_v in the active layer. Therefore, the N 1s spectrum of the a-IGZO can only reveal the N-bonding properties in a-IGZO:N thin film, can’t discuss a relation with N-bonding state, O_v, and reliability.

Fig.1* N 1s XPS spectra of a-IGZO film grown with 40 % P_N/O.

Author Response File: Author Response.docx