Effects of CF₄ Plasma Treatment on Indium Gallium Oxide and Ti-doped Indium Gallium Oxide Sensing Membranes in Electrolyte–Insulator–Semiconductors

Round 1

Reviewer 1 Report

In this manuscript, the authors reported improvement in sensitivity and linearity after CF6 plasma treatment and Ti doping in IGO sensing membrane. The pH sensitivity value of the Ti-doped Indium Gallium Oxide membrane is moderate and the drift value is also lower. This work seems very interesting and organized. Although, the current result provides enough information, the authors should describe the mechanism in more extensive way. I think the manuscript should be published after minor revision. Some minor comments and suggestions are listed below to the authors:

1. Authors mentioned n-type Si wafer in line 42, but in the schematic (fig 1) it is mentioned that the substrate is p-type Si wafer. Could the authors explain this contradiction?
2. “In the process of the reactive sputtering, IGO and Ti targets were used in the ambient of Ar:O2 at 23:2” (line 48). Could the authors explain why they use this particular ratio?
3. There are many typing errors in the manuscript, which need to be addressed. Such as, in line 16 it is written mv which is mV and the mistake is same in line 269. In the line 268 it is written PH, you can start the sentence as ‘The pH’. In Fig. 7(a) you have mentioned ‘IGO+CF4 plasma 60 secs’ you should remove 60 secs. Same mistake follows in Fig. 7(b). Sentence incomplete in Line 61. Unwanted period punctuation in line 147.
4. Did authors examined the sensitivity after longer (>60s) plasma treatment, as the sensitivity is increasing with plasma treatment time? The authors should explain why they stopped at 60s.
5. In line 181: For the calculation of pH sensitivity the capacitance value is taken at 0.4 C_max. The reviewer wants to know how did you optimize that the “C” value is to be taken at ‘0.4 C_max’?
6. Please explain in details the relation between “Ti-doping and CF4 plasma treatment could enhance crystallization, grainization and chemical bindings” and “drastically boost the membrane sensitivity” (line 269). Moreover authors are requested to correlate the multiple material analysis results to the electrical outcome i.e. sensitivity, linearity etc. in more extensive way.
7. Figure 6: the inset figures are not clear to read. Can you please increase the font size of inset figures?
8. The authors are encouraged to use more definitive sentence rather than “could be/might be” in the text. (e.g. line 15 authors wrote “The pH sensitivity of Ti-doped IGO membranes with CF4 plasma treatment could achieve as high as 60.8mv/pH”, where they already achieve it)

Author Response

1. Authors mentioned n-type Si wafer in line 42, but in the schematic (fig 1) it is mentioned that the substrate is p-type Si wafer. Could the authors explain this contradiction?

Ans: Thank the reviewer for the careful review. We found that the substrate is n-type after checked our original data. We revised the data. Thanks for the correction.

1. “In the process of the reactive sputtering, IGO and Ti targets were used in the ambient of Ar:O2 at 23:2” (line 48). Could the authors explain why they use this particular ratio?

Thank you for your constructive suggest. For the process of the reactive sputtering, we had tried the ratio of 25:0, 20:5, 23:2, respectively. After trying each of these, we find the best result is the ambient of 23:2. Since this IGO structure have the oxygen vacancies, we need appropriate oxygen to eliminate the defects, while too much oxygen may cause oxidation and degradation. Therefore, appropriate amount of oxygen can attach the best effect.

1. There are many typing errors in the manuscript, which need to be addressed. Such as, in line 16 it is written mv which is mV and the mistake is same in line 269. In the line 268 it is written PH, you can start the sentence as ‘The pH’. In Fig. 7(a) you have mentioned ‘IGO+CF4 plasma 60 secs’ you should remove 60 secs. Same mistake follows in Fig. 7(b). Sentence incomplete in Line 61. Unwanted period punctuation in line 147.

Thanks to the reviewer for the careful review. We have corrected the typos and errors mentioned by the reviewer. In addition, we checked the whole manuscript again and correct errors.

1. Did authors examined the sensitivity after longer (>60s) plasma treatment, as the sensitivity is increasing with plasma treatment time? The authors should explain why they stopped at 60s.

Thank you for your carefully review. Since plasma treatment have the effect of etching, plasma treatment over 60 sec will cause damaging on the surface. Therefore, the device will be drastically deteriorated. As a result, we set 60sec to be an end point.

1. In line 181: For the calculation of pH sensitivity the capacitance value is taken at 0.4 C_max. The reviewer wants to know how did you optimize that the “C” value is to be taken at ‘0.4 C_max’?

Thank you for your constructive suggest. Actually, we can either take the capacitance value of 0.3 or 0.5. However, according to previous experience, if we take the value around 0.3 to 0.5 C_max as the reference, the sensitivity can be optimized. In the meantime, we can also avoid distortion if we take the capacitance value above 0.3 C_max. Therefore, we choose the 0.4 C_max to be our baseline. In this value, the sensitivity is the best and can also avoid the distortion. If C＞0.5 C_max, the interval will became smaller, while C＜0.3 C_max, we can see the distorted curve. To achieve better sensitivity and linearity, we pick up 0.4 C_max.

_{FOR 0.3:  FOR 0.5:}

1. Please explain in details the relation between “Ti-doping and CF4 plasma treatment could enhance crystallization, grainization and chemical bindings” and “drastically boost the membrane sensitivity” (line 269). Moreover authors are requested to correlate the multiple material analysis results to the electrical outcome i.e. sensitivity, linearity etc. in more extensive way.

Ans: Since the membrane capacitance is crucial to the sensing performance, to increase the membrane capacitance values by enhancing the effective electric field passing through the membrane material is our goal. Therefore, monitoring the material properties and build connections between material characterizations and sensing device performance are our methods. On the other hand, based on possible electrolyte membrane interface model, incorporating positive charges can expel the ions in the electrolyte, and hence lower down the diffusion capacitance in the solution. Therefore, the sensitivity can be enhanced  to improve the sensing performance. On the one hand, we fix the defects in the membrane bulk by Ti doping and CF₄ plasma treatment. On the other hand, we incorporating possible charges near the surface may decrease the C_DL and enhance the sensitivity as the sensing equation reveals.

Since Ti doping and CF₄ plasma treatment can incorporate Ti atoms and F atoms into the membrane, resulting in the removal of dangling bonds and traps in the bulk and near the surface. Therefore, the film quality became improved and more uniform. Therefore, compared with the as-deposited film, the distortion of C-V curves can be mitigated, which optimize the linearity.

1. Figure 6: the inset figures are not clear to read. Can you please increase the font size of inset figures?

Thanks to the reviewer for the suggestion. We have improved the inset figure. Moreover, in our revised manuscript, we take the subfigure out and put by the original figure readable.

1. The authors are encouraged to use more definitive sentence rather than “could be/might be” in the text. (e.g. line 15 authors wrote “The pH sensitivity of Ti-doped IGO membranes with CF4 plasma treatment could achieve as high as 60.8mv/pH”, where they already achieve it)

Ans: Thanks to the reviewer for the constructive suggestion. We removed “could” and “might” and state the observation in a more definite tongue.

Author Response File: Author Response.docx

Reviewer 2 Report

The  text seems unique and very interesting. Generally I think that you could fix the quality of figure 1 and explain in label what means a.u in figure 2 and arb.units in figure 3. It is a good practice that you explain all the acronyms once in your text and this is a good point, but i think that you should put a table somewhere in your text, that compares your results with older results and explain why you think tha they are better your results from others who do around the same thing.

Finally i believe that it is better to do another figure for e and f in figure 6, as they have to do with sensitivity in y axis. The main thing that i would like to see is the comparison table.

Author Response

The text seems unique and very interesting. Generally I think that you could fix the quality of figure 1 and explain in label what means a.u in figure 2 and arb.units in figure 3. It is a good practice that you explain all the acronyms once in your text and this is a good point, but i think that you should put a table somewhere in your text, that compares your results with older results and explain why you think tha they are better your results from others who do around the same thing

We have defined a.u. (arbitrary unit) in the figure captions and context. In addition, we use consistent a.u. to illustrate after first definition.

We have checked all the acronyms and defined all the acronyms. Thanks to the reviewer for the constructive comments.

We have included a table to compare this research previous reports in the revised manuscript.

Apparently, Ti doping plus plasma treatment can optimize the sensing performance.

Table 1. Comparison of CF4 plasma treated Ti-doped IGO with previous literatures

Compared with previous works related to CF4 plasma treatment and IGO membrane as shown in Table 1, Ti doped IGO with plasma treatment had the highest sensitivity and linearity, indicating that Ti doping plus plasma treatment on IGO membrane can optimize the sensing performance

Finally i believe that it is better to do another figure for e and f in figure 6, as they have to do with sensitivity in y axis. The main thing that i would like to see is the comparison table.

Thank you for your advice. We will present our current figures for this very research. A comparison table is shown in Table 1 in the revised manuscript.

Author Response File: Author Response.docx

Reviewer 3 Report

The article titled “Effects of CF4 Plasma Treatment on Indium Gallium Oxide and Ti-doped Indium Gallium Oxide Sensing Membrane in Electrolyte-Insulator-Semiconductor” by Kao et al. is an interesting work dealing with improving the sensitivity, reliability and stability of the EIS membranes through improvements in the crystallization and reduction in dangling bonds using Ti – doping and CF4 plasma treatment. I would suggest following changes to be made in the manuscript:

1. The ‘Abstract’ should be improved by adding the crux of the background information on the EIS membranes and what is state of the art in the current technology when it comes to Ti – doped IGO EIS membranes. Also, it is suggested that the authors include some tangible piece of data based on their quantification of the material properties similar to the sensitivity information of 60.8 mV/pH that they have mentioned. This information could be the thickness of the membrane, stability in terms of time and reliability in terms of number of tests.
2. The ‘Introduction’ section is too short. The authors need to significantly add a lot more background on the EIS membranes and the technology of the ISFETs pertaining to sensing. Also, typically the last paragraph of ‘Introduction’ is supposed to be the current work and what is the primary goal and the expectations from the work by way of hypothesizing the process. Most of these things are missing from the section.
3. In the ‘Materials and Methods’ section, on lines 42 – 43, the authors need to mention the source from where they got the IGO films and the 4 – inch n – type (100) silicon wafer. Similarly, the information on thermal oxidation equipment is missing. What was the size of the PCB and the thickness of the silver gel?
4. Figures 1, 2, 5b, 5c, 6 and 7 look very similar to a figures already published by same authors – “Kao, C., Liu, C.S., Xu, C.Y. et al.Ti-doped indium gallium oxide electrolyte–insulator–semiconductor membranes for multiple ions and solutes detectors. J Mater Sci: Mater Electron 30, 20596–20604 (2019).” The authors should provide compelling evidence as to how their submitted figures are different than the ones already published.  
5. The equipment used for quantifying the deposited films, XRD, XPS, SIMS, AFM, CV is completely missing from the materials and methods section.
6. The authors needs to explain the SIMS fluorine penetration depths with relevant values and how does it demonstrate the diffusion of fluorine ions.
7. In Figure 6, panel (e) is missing the label. Also, the authors need to explain Figures 6e and 6f in a better way especially the linearity of the sensor.
8. ‘Conclusion’ section should also have the primary results related to the stability and reliability of the synthesized EIS membrane from the results pertaining to SIMS, AFM, XRD and CV.

Author Response

The article titled “Effects of CF4 Plasma Treatment on Indium Gallium Oxide and Ti-doped Indium Gallium Oxide Sensing Membrane in Electrolyte-Insulator-Semiconductor” by Kao et al. is an interesting work dealing with improving the sensitivity, reliability and stability of the EIS membranes through improvements in the crystallization and reduction in dangling bonds using Ti – doping and CF4 plasma treatment. I would suggest following changes to be made in the manuscript:

1.The ‘Abstract’ should be improved by adding the crux of the background information on the EIS membranes and what is state of the art in the current technology when it comes to Ti – doped IGO EIS membranes. Also, it is suggested that the authors include some tangible piece of data based on their quantification of the material properties similar to the sensitivity information of 60.8 mV/pH that they have mentioned. This information could be the thickness of the membrane, stability in terms of time and reliability in terms of number of tests.

Ans: Thanks to the reviewer for the constructive comments. We revised the abstract accordingly as follows.

In this research, Indium Gallium Oxide, which is an optical material, was used for 50 nm electrolyte-insulator-semiconductor membrane. Compared with conventional treatment such as annealing, Ti doping and CF₄ plasma treatment were incorporated in the fabrication of the film. Because of the effective treatment of doping and plasma treatment, the defects were mitigated and the membrane capacitance was boosted. Therefore, the pH sensitivity can be increased up to 60.8mV/pH. In addition, the hysteresis voltage could be improved down to 2.1mV and the drift voltage can be suppressed as low as 0.23 mv/hr. The IGO-based membranes are promising for future high sensitivity and stability devices integrated with optical applications. 

2.The ‘Introduction’ section is too short. The authors need to significantly add a lot more background on the EIS membranes and the technology of the ISFETs pertaining to sensing. Also, typically the last paragraph of ‘Introduction’ is supposed to be the current work and what is the primary goal and the expectations from the work by way of hypothesizing the process. Most of these things are missing from the section.

Ans: Thanks to the reviewer for the insightful review. We added a lot more background information pertaining to sensing as the reviewer commented.

Since the membrane capacitance is crucial to the sensing performance, to increase the membrane capacitance values by enhancing the effective electric field passing through the membrane material is our goal. Therefore, monitoring the material properties and build connections between material characterizations and sensing device performance are our methods. On the other hand, based on possible electrolyte membrane interface model, incorporating positive charges can expel the ions in the electrolyte, and hence lower down the diffusion capacitance in the solution. Therefore, the sensitivity can be enhanced  to improve the sensing performance. On the one hand, we fix the defects in the membrane bulk by Ti doping and CF₄ plasma treatment. On the other hand, we incorporating possible charges near the surface may decrease the C_DL and enhance the sensitivity as the sensing equation reveals.

3.In the ‘Materials and Methods’ section, on lines 42 – 43, the authors need to mention the source from where they got the IGO films and the 4 – inch n – type (100) silicon wafer. Similarly, the information on thermal oxidation equipment is missing. What was the size of the PCB and the thickness of the silver gel?

Ans: Thank the reviewer for the question. We purchased the IGO target and the Si wafers from Well-being. Enterprise Co. Ltd.  Taiwan. We added the information in the context. We purchased the Si wafers with wet oxides on top of them so we don’t know the equipment of the company but we label the company source. The PCB size is 8.5*1.5(cm²). The thickness of gel is around 5µm.

4.Figures 1, 2, 5b, 5c, 6 and 7 look very similar to a figures already published by same authors – “Kao, C., Liu, C.S., Xu, C.Y. et al.Ti-doped indium gallium oxide electrolyte–insulator–semiconductor membranes for multiple ions and solutes detectors. J Mater Sci: Mater Electron 30, 20596–20604 (2019).” The authors should provide compelling evidence as to how their submitted figures are different than the ones already published.  

Ans: Thanks to the reviewer for the comment. If you zoom in the previous paper [Ref. 1] and this paper, you will find the previous paper discussed the annealing effects but this research investigated the effect of plasma treatment. The two treatments are totally differently although the material analyses are similar. Furthermore, we found that the plasma treatment can enhance the sensing performance stronger than the anneal effects. That is the highest sensitivity can be achieved by plasma treatment and Ti doping. In the future, we may combine the two effects.

Apparently, Ti doping plus plasma treatment can optimize the sensing performance.

Table 1. Comparison of CF4 plasma treated Ti-doped IGO with previous literatures

Compared with previous works related to CF4 plasma treatment and IGO membrane as shown in Table 1, Ti doped IGO with plasma treatment had the highest sensitivity and linearity, indicating that Ti doping plus plasma treatment on IGO membrane can optimize the sensing performance

Compared with the paper mentioned by the reviewer, our device performance had further improvements. 

5.The equipment used for quantifying the deposited films, XRD, XPS, SIMS, AFM, CV is completely missing from the materials and methods section.

Ans: Thank you for your meticulous review. The XRD equipment is Bruker D8 Discover with Oxford Cryo Drive 3.0. The XPS equipment is VG scientific Microlab 350. The SIMS equipment is Germany ION-TOF-SIMSV. The CV equipment is keysight 4284 A precision LCR Meter 20 Hz to 1 Hz. The AFM equipment is Bruker Dimension Icon.

6.The authors needs to explain the SIMS fluorine penetration depths with relevant values and how does it demonstrate the diffusion of fluorine ions.

Ans: Thanks to the reviewer for the constructive comment. Apparently, as for the undoped films, fluorine concentration became 700 times near the surface and the penetration depth was around 20 nm for the plasma treatment samples with various treatment time. As for the Ti-doped sample as shown in Fig.4(b), the fluorine plasma treatment with 60 seconds can further push the penetration depth up to 25 nm. CF₄ plasma with a longer treatment time of 60s and Ti doping can activate the fluorine atoms to move further downward.

7.In Figure 6, panel (e) is missing the label. Also, the authors need to explain Figures 6e and 6f in a better way especially the linearity of the sensor.

Ans: Thank you for your carefully review. We have fixed the panel(e) in the paper.

Since Ti doping and CF₄ plasma treatment can incorporate Ti atoms and F atoms into the membrane, resulting in the removal of dangling bonds and traps in the bulk and near the surface. Therefore, the film quality became improved and more uniform. Therefore, compared with the as-deposited film, the distortion of C-V curves can be mitigated, which optimize the linearity. In addition, Fig 6(e) has been fixed.

8.‘Conclusion’ section should also have the primary results related to the stability and reliability of the synthesized EIS membrane from the results pertaining to SIMS, AFM, XRD and CV.

  Ans: Thanks to the reviewer for the constructive comments. We revised the conclusion accordingly.

In addition, the hysteresis voltage and the drift voltage both dropped to 2.1mv and 0.32mv/hr, respectively indicating the device are more reliable and stable with the multi-treatment. The improvements comes from the enhancement of crystallization as revealed by XRD and AFM and the fluorine atom penetration as shown by the SIMS data. Therefore, the electric field can more effective pass through the membrane, and the capacitance and sensitivity increase.

Author Response File: Author Response.docx

Round 2

Reviewer 3 Report

The authors have made significant improvements to the manuscript. I would still suggest one minor revision to the abstract. The abstract does read well except it is still missing the background information. Typically, it would start with the subject, in this case ‘Electrolyte-Insulator-Semiconductor sensors’. For e.g. it may start something like this, “Electrolyte-insulator-semiconductor sensors are the most basic type of ISFET membranes. Such membranes are used in applications such as………………. Currently, some of the most popular techniques for synthesizing such sensors are………………. However, there are certain limitations on such techniques such as……………………… Here we come up with a novel design in which an optical material, Indium Gallium oxide is used for a 50 nm EIS membrane…….” After this the authors can continue with what they already have in the abstract.

Author Response

Thanks to the reviewer for the constructive comment and helpful suggestion. The abstract has been revised based on the reviewer’s comment as follows.

Electrolyte-insulator-semiconductor (EIS) sensors, used in applications such as pH sensing and sodium ion sensing, are the most basic type of ion-sensitive field-effect transistor (ISFET) membranes. Currently, some of the most popular techniques for synthesizing such sensors are chemical vapor deposition, reactive sputtering and sol-gel deposition. However, there are certain limitations on such techniques such as reliability concerns and isolation problems. In this research, a novel design of EIS membrane consisting of an optical material of Indium Gallium Oxide (IGO) was demonstrated. Compared with conventional treatment such as annealing, Ti doping and CF₄ plasma treatment were incorporated in the fabrication of the film. Because of the effective treatment of doping and plasma treatment, the defects were mitigated and the membrane capacitance was boosted. Therefore, the pH sensitivity can be increased up to 60.8 mV/pH. In addition, the hysteresis voltage could be improved down to 2.1mV and the drift voltage can be suppressed as low as 0.23 mV/hr. The IGO-based membranes are promising for future high sensitivity and stability devices integrated with optical applications.

Author Response File: Author Response.docx