An Improved Structure Enabling Hole Erase Operation When Using an IGZO Channel in a 3D NAND Flash Structure to Which COP (Cell-On-Peri) Structure Is Applied
Nupur Navlakha
Round 1
Reviewer 1 Report
The authors have introduced a concept of using a pillar/filler for improving the time for erase operation. The concept is well-explained. However, I have the following recommendations :
1. In the introduction, authors mentioned inadequate hole characteristics. Please elaborate it as it gives the sole purpose of using the suggested design.
2. Please comment on fabrication of this topology.
3. Please show a comparative analysis to show how it is better than the state-of-art and, show figure of a conventional IGZO based NAND Flash. Please site some other works using IGZO flash memory.
4. Please comment on its overall performance as a memory device. How introduction of a filler impacts the program mechanism, the retention characteristics and area. Please comment on scalability of filler.
5. Please clearly illustrate how is it different from previous work, “ Floating Filler (FF) in an Indium Gallium Zinc Oxide (IGZO) Channel Improves the Erase Performance of Vertical Channel NAND Flash with a Cell-on-Peri (COP) Structure”.
Please correct few spellings like fillar and sentences like " inadequate hole carrier characteristics". Please use technicallanguage to clearly explain the physical insights.
Author Response
Response to Reviewer 1 Comments
Thank you for the reviewer's comments and advice. I wrote the answers to the reviewer's comments and advice in the sentences below and attached the revised paper as a separate PDF file.
Point 1: In the introduction, authors mentioned inadequate hole characteristics. Please elaborate it as it gives the sole purpose of using the suggested design.
Response 1: Thanks for your point and advice. Reflecting the reviewer's opinion, we added the corresponding information to line 55-57 on page 2. Specifically, among the poor hole characteristics of IGZO, the very heavy hole effective mass and low hole mobility were explained. Because of this nature, 3D NAND Flash structures using oxide semiconductors, including IGZO, were known to be impossible to use hole erase operation.
Point 2: Please comment on fabrication of this topology.
Response 2 : In my opinion, there is probably no etching process technology that can form perfectly smooth (no topology) vertical hole strings with current semiconductor process technology. This assumption could be inferred from the cross-sectional photos (SEM, TEM, etc.) of the hole sting shown in the 3D NAND Flash research cases that have been announced until recently. Therefore, it is considered that there is a sufficient possibility for the formation of topologyrs referred to as "spike" in this paper.
Point 3: Please show a comparative analysis to show how it is better than the state-of-art and, show figure of a conventional IGZO based NAND Flash. Please site some other works using IGZO flash memory.
Response 3 : Unfortunately, as far as I know, no actual case of 3D NAND flash structure using IGZO channel has been announced by companies or research institutes such as universities. The reason is, of course, that hole erase operation is impossible in the IGZO channel, and because of this problem, 3D NAND Flash production companies such as Samsung and SK hynix tried to apply the IGZO channel, but are currently stopping research. Therefore, our group's research, including this paper, was a study to overcome this problem, and although the performance in the actual manufactured device has not yet been confirmed, we believe that the simulation confirmed the possibility. Currently, our group is manufacturing 3D NAND Flash devices with 1 to 3 layers of IGZO channels, and based on this result, we are trying to evaluate the possibility of operation in actual devices.
Point 4: Please show a comparative analysis to show how it is better than the state-of-art and, show figure of a conventional IGZO based NAND Flash. Please site some other works using IGZO flash memory.
Response 4 : First of all, it is thought that the performance of the program in the proposed structure will be almost the same as the current general 3D NAND flash structure. This is because what drives the operation in the program operation is the high voltage (~20 V) applied to the word line to be programmed, and the channel is basically GND (0 V). Therefore, even if IGZO material is used as a channel, it is thought that there is little improvement in this part, and the filler structure will also help a little to keep the entire channel at 0 V evenly, but it is judged that there will be no improvement in the speed and efficiency of program operation. Next, since reliability aspects such as endurance and retention performance are also highly dependent on the material properties of the Oxide/Nitride/Oxide(ONO) layer that performs the memory function, we assume that there will be no significant change even if the IGZO channel and "Filler" structure are applied. However, since oxide semiconductor materials such as IGZO are known to have better interfacial properties with tunnel oxide, which is the same oxide, than polysilicon channels, it is expected that improvement in this area will be possible. Finally, it is thought that the scalability of the "Filler" structure proposed in this paper does not exist significantly. The reason why our group is proposing and researching the filler structure is to enable hole erase operation in the 3D NAND Flash structure using the IGZO channel, and we do not think about other functional improvements. In reality, the filler structure itself has a problem of increasing complexity in the process, but even if you take this risk, IGZO, a material with excellent characteristics (very low leakage current, fast operating speed, temperature stability, etc.), can be used in 3D NAND Flash I think it's one of the structures that can be used.
Point 5: Please clearly illustrate how is it different from previous work, “ Floating Filler (FF) in an Indium Gallium Zinc Oxide (IGZO) Channel Improves the Erase Performance of Vertical Channel NAND Flash with a Cell-on-Peri (COP) Structure”.
Response 5 : The "Floating Filler (FF)" structure mentioned by the reviewer is an application of the "GIDL erase method" currently used in 3D NAND Flash structures. Specifically, it was a method of forming hole carriers by the GIDL phenomenon in the polysilicon channel that exists below the upper BL contact and transferring these hole carriers to the IGZO channel through the filler. However, in the case of the GIDL erase method, it is known that it takes a lot of time (~ 1 ms) to form hole carriers by the GIDL phenomenon in actual device operation, and as confirmed by simulation in this paper, when using the IGZO channel, it is necessary to Since there is no need to form hole carriers, this method was proposed and verified. When the proposed structure is applied, the erase operation speed is expected to be relatively fast because the voltage is immediately delivered to the filler, and then a small number of hole carriers already present in the filler are delivered to the IGZO channel.
Comments on the Quality of English Language
Please correct few spellings like fillar and sentences like " inadequate hole carrier
Response on the Quality of English Language : Thanks for your point and advice. We have corrected the points you pointed out and marked them in red.
Author Response File: 
Author Response.pdf
Reviewer 2 Report
This manuscript outlines a novel IGZO-Filler (IF) structure for Cell-On-Peri (Cell-Peri) in 3D NAND Flash memory, which improves upon the existing IGZO-Pillar (IP) structure. The primary novelty in the proposed approach lies in leveraging the 'poor hole property' of IGZO, indicating an ingenious inverse application of this characteristic. The manuscript details the technical achievements of the proposed structure, such as the satisfactory erase operation with the sole presence of hole carriers in thin fillers, maintaining a low leakage current, and overcoming temperature stability problems. It is commendable that the paper backs up these claims with simulation results. In general, the manuscript successfully communicates the proposed structure's novelty, potential, and validated performance improvements. I have the following suggestions and comments that hope the authors could address:
1. In Table 3 comparison table, it could be better if the authors can provide more quantifiable data for leakage current and self-boosting performance rather than very little bad/little bad. It would give the readers better understanding of the potential trad-off in the proposed technology.
2. In the conclusion part, it would be great if the authors can discuss some of the feasibility for applying the proposed technology to real memory product to highlight the impact of this paper.
Author Response
Response to Reviewer 2 Comments
Thank you for the reviewer's comments and advice. I wrote the answers to the reviewer's comments and advice in the sentences below and attached the revised paper as a separate PDF file.
Point 1: In Table 3 comparison table, it could be better if the authors can provide more quantifiable data for leakage current and self-boosting performance rather than very little bad/little bad. It would give the readers better understanding of the potential trad-off in the proposed technology.
Response 1: Thanks for your point and advice. Reflecting the reviewer's opinion, we added the correct data sets to the items in Table 3.
Point 2: In the conclusion part, it would be great if the authors can discuss some of the feasibility for applying the proposed technology to real memory product to highlight the impact of this paper.
Response 2 : Thanks for your point and advice. We added improvements to the application of the proposed structure reflecting the reviewer's comment at the end of the conclusion. However, as of now, since this part has been verified by only simulation, we will proceed with verification by actual device results in the future.
Author Response File: Author Response.pdf
Reviewer 3 Report
Referee report
The article “A improved structure that enables hole erase operation in 3D NAND flash using indium-gallium-zinc-oxide (IGZO) channels and Cell-On-Peri (COP) structures” is devoted to very actual problem - improving the properties of non-volatile memory. The authors proposed IGZO-Filler (IF) structure. Metal oxides, which are wide gap semiconductors (as well as semiconductor suboxides) are now very popular for use as materials for non-volatile memory, not only for flash memory, but also for memristors, see for example - Ivan D. Yushkov et. al. Memristors Based on Many-layer Non-stoichiometric Germanosilicate Glass Films. Electronics, 12, 873, 2023, DOI: https://doi.org/10.3390/electronics12040873. The authors have chosen an adequate and widely tested program (Synopsys Sentaurus TCAD devices simulator) for the numerical simulation of electronic devices. The article contains new data and will be interesting for researchers and technologists. The article can be published, иге the author should take into account my comments.
Comment:
1) The physical mechanism for improving erasure will be described in more detail. It is clear that due to the accumulation of holes at the IGZO-tunnel oxide interface, the voltage drop across the IGZO decreases (Figure 1d compared to Figure 1c). However, this should also lead to an increase in the potential difference between the metal gate and the tunnel oxide (this difference is not visible comparing the Figures 1d and 1c), since the total voltage drop between the metal gate and the p+ polysilicon layer is unchanged and amounts to 20 Volts. In addition, it is not clear why (comparing the Figures 1d and 1c) the band offset between p+ polysilicon and IGZO is changed? The band offset is determined only by the materials of the heterostructure and does not depend on the charge at the interface.
2) All abbreviations should be explained, for example – GIDL, CMP, CTF. Authors should use consistent abbreviations, for example for Cell-On-Peri – COP, not (Cell-Peri). Line 301 10-15 A - it is necessary to use superscript to indicate the degree -15. Why does the concentration (of the dopant?) go into the negative region in Figure 3? There are no b-h parts in Figure 5, but they are in the caption for the figure.
Minor revision
Author Response
Response to Reviewer 3 Comments
Thank you for the reviewer's comments and advice. I wrote the answers to the reviewer's comments and advice in the sentences below and attached the revised paper as a separate PDF file.
Point 1: The physical mechanism for improving erasure will be described in more detail. It is clear that due to the accumulation of holes at the IGZO-tunnel oxide interface, the voltage drop across the IGZO decreases (Figure 1d compared to Figure 1c). However, this should also lead to an increase in the potential difference between the metal gate and the tunnel oxide (this difference is not visible comparing the Figures 1d and 1c), since the total voltage drop between the metal gate and the p+ polysilicon layer is unchanged and amounts to 20 Volts. In addition, it is not clear why (comparing the Figures 1d and 1c) the band offset between p+ polysilicon and IGZO is changed? The band offset is determined only by the materials of the hetero structure and does not depend on the charge at the interface.
Response 1: Thanks for your point and advice. Reflecting the reviewer's opinion, we have corrected inaccurate figures in the band structure description in Figure 1, to be precise, the errors seen in item (d).
Point 2: All abbreviations should be explained, for example – GIDL, CMP, CTF. Authors should use consistent abbreviations, for example for Cell-On-Peri – COP, not (Cell-Peri). Line 301 10-15 A - it is necessary to use superscript to indicate the degree -15. Why does the concentration (of the dopant?) go into the negative region in Figure 3? There are no b-h parts in Figure 5, but they are in the caption for the figure.
Response 2 : Sorry, the inaccurate part of the item was a mistake in writing. We made corrections to the items pointed out by the reviewer, and the revised parts were marked in red. In addition, the doped region marked with a negative concentration in Figure 3 is because a p-type (Boron) doped region (filler) is assumed. This is because P-type doping has a relatively negative concentration compared to N-type doping.
Author Response File: Author Response.pdf
Reviewer 4 Report
Specific comments and recommendations:
1) The title of this manuscript contains abbreviations in parentheses that authors should avoid. Also, in my opinion, the title is quite long and will be difficult for potential readers to understand;
2) Section 2 has to be expanded to give the design details of the structure;
3) This manuscript lacks experimental results to serve as a reliable basis for performing a comparative analysis. Running computer simulations alone is completely insufficient as a verification, especially for an article in a journal like Electronics;
4) In the tables given at the end of the manuscript, there are too few parameters for comparative analysis. The authors have to do a detailed analysis of various literary sources.
In my opinion, minor editing of the English language is required.
Author Response
Response to Reviewer 4 Comments
Thank you for the reviewer's comments and advice. I wrote the answers to the reviewer's comments and advice in the sentences below and attached the revised paper as a separate PDF file.
Point 1: The title of this manuscript contains abbreviations in parentheses that authors should avoid. Also, in my opinion, the title is quite long and will be difficult for potential readers to understand;
Response 1: Thanks for your point and advice. Reflecting the reviewer's opinion, we modified the title of the paper to be more concise and intuitively understandable.
Point 2: Section 2 has to be expanded to give the design details of the structure;
Response 2 : Thanks for your point and advice. Reflecting the reviewer's opinion, we added in Chapter 2 the description of the design of the structure. The additions are indicated in red and further explain how our proposed structure differs from previous studies.
Point 3: This manuscript lacks experimental results to serve as a reliable basis for performing a comparative analysis. Running computer simulations alone is completely insufficient as a verification, especially for an article in a journal like Electronics;
Response 3 : I think the reviewer's opinion is perfectly valid. However, in the case of 3D NAND Flash structures, it is very difficult to make actual devices in the university laboratory at in due to the difficulty of the manufacturing process. Therefore, for structures that are not currently commercialized, such as those proposed in this paper, verification by simulation is practically the most effective. In addition, in order to reduce as much as possible the inaccuracy caused by verifying the proposed structure, which reviewer is concerned about, by simulation only, we first used the material parameters of the IGZO channel that were confirmed through actual device measurements [14]. In addition, the material parameters required for memory operation are also Since we used the parameters and models that have already been completely verified [15-17], we expect that it will not be completely different from the actual device operation. As for the last proposed structure, a three-tiered structure is currently being actually manufactured, and the actual operation result will be checked as soon as the production is completed.
Point 4: In the tables given at the end of the manuscript, there are too few parameters for comparative analysis. The authors have to do a detailed analysis of various literary sources.
Response 4 : : I think the reviewer's opinion is perfectly valid. However, since the improvement point when applying the structure proposed in this paper is that the hole erase oepration, which was impossible when using the IGZO channel, can be operated at high speed, I wanted to emphasize this part. Since the program and read operations other than the erase operation basically depend on the material characteristics of the word line and ONO film, not the channel, the proposed structure is also expected to have almost the same performance as the existing 3D NAND Flash structure.
Comments on the Quality of English Language
In my opinion, minor editing of the English language is required.
Response on the Quality of English Language : Thanks for your point and advice. We have corrected the points you pointed out and marked them in red.
Author Response File: Author Response.pdf
Reviewer 5 Report
The manuscript entitled “A improved structure that enables hole erase operation in 3D NAND flash using indium-gallium-zinc-oxide (IGZO) channels and Cell-On-Peri (COP) structures” submitted by Yun-heub Song and co-workers for consideration for publication in the MDPI journal Electronics proposes an improved structure called 'IGZO-Filler (IF)' that can be used in the Cell-On-Peri (Cell-Peri) structure. This improvement aims to enhance the erase performance of the 'IGZO-Pillar (IP)' structure, which was previously used in the IGZO channel 3D NAND Flash structure as described in earlier studies. Overall, the paper introduces an improved structure for the IGZO channel in 3D NAND Flash, highlighting its fast erase rate and ability to address leakage current and temperature stability problems. The findings suggest potential advancements in the field of memory technology. The manuscript's subject seems to be interesting, but its presentation is more like a technical note. Language presentation is poor and not encouraging. The manuscript should be rewritten to present a higher level. I recommend publishing the manuscript in the MDPI journal Electronics only after major revision. I suggest the fundamental revision of the presented manuscript. The following question and suggestions should be considered during the manuscript's rewrite.
1. What are the reasons that this manuscript will be of interest and importance to other researchers? How can the other scientists use them?
2. The authors should follow the author guidelines of MDPI journals.
3. Language presentation must be carefully checked and revised, even the title is written with a mistake. The language of the presentation seems to be more written in colloquial language.
4. A description of all symbols and the units should be presented in the text. Partially, the current presentation can lead to misunderstanding part of the symbols.
5. All Figures should follow the same trend and size.
Language presentation is poor and not encouraging. 1. Language presentation must be carefully checked and revised, even the title is written with a mistake. The language of the presentation seems to be more written in colloquial language.
Author Response
Response to Reviewer 5 Comments
Thank you for the reviewer's comments and advice. I wrote the answers to the reviewer's comments and advice in the sentences below and attached the revised paper as a separate PDF file.
Point 1: What are the reasons that this manuscript will be of interest and importance to other researchers? How can the other scientists use them?
Response 1: The structure proposed in this paper can be said to be a new structure that enables IGZO materials, which have been impossible to perform hole erase operations due to very poor hole carrier properties despite having excellent channel characteristics, in COP structures. Therefore, in my opinion, if other researchers refer to the structure proposed in this paper, it is possible to solve the hole erase operation problem in using the IGZO channel as a means to solve the problems occurring in the channel of the current 3D NAND Flash structure. In particular, one of the problems that 3D NAND Flash manufacturers (Samsung, SK hynix, etc.) are currently struggling with is the various problems of polysilicon channels (high leakage current, low effective mobility, temperature instability, etc.), so the proposed structure can be an idea to overcome channel problems by applying IGZO channels to current products.
Point 2: The authors should follow the author guidelines of MDPI journals.
Response 2 : Thanks for your point and advice. Reflecting the reviewer's opinion, I reviewed the entire paper and revised the parts that were different from the guidelines of MDPI, and marked the modified parts in red.
Point 3: Language presentation must be carefully checked and revised, even the title is written with a mistake. The language of the presentation seems to be more written in colloquial language.
Response 3 : Thanks for your point and advice. Reflecting the reviewer's opinion, I carefully reviewed the entire paper and modified the sentences to make it easier to understand.
Point 4: A description of all symbols and the units should be presented in the text. Partially, the current presentation can lead to misunderstanding part of the symbols.
Response 4 : Thanks for your point and advice. Reflecting the reviewer's opinion, I have written explanations of all symbols and units used in the text, and marked the modified parts in red letters.
Point 5 : All Figures should follow the same trend and size.
Response 5 : Thanks for your point and advice. Again, I carefully reviewed all the figures and made sure they were consistent except for the parts necessary for explanation.
Comments on the Quality of English Language
Language presentation is poor and not encouraging. 1. Language presentation must be carefully checked and revised, even the title is written with a mistake. The language of the presentation seems to be more written in colloquial language.
Response on the Quality of English Language : Thanks for your point and advice. Again, I carefully reviewed all the sentences in the paper and revised them according to the reviewer's opinion.
Author Response File: Author Response.pdf
Round 2
Reviewer 1 Report
I appreciate the revisions by author.
Author Response
Comments and Suggestions for AuthorsI appreciate the revisions by author.
Response for reviewer 1's comment
Thank you for your consent. Your points have been of great help to the improvement and development of this paper.
Reviewer 4 Report
In this version of the manuscript, reviewer recommendations 3 and 4 were not performed. These recommendations are particularly important because, on the one hand, they are related to the validation process of the proposed structures. On the other hand, these recommendations are related to performing a comparative analysis and defining the advantages and disadvantages of the proposed structures. I recommend that after observing notes 3 and 4, the manuscript be considered again for publication in the journal Electronics.
Some sentences, especially those added in this version of the manuscript, are too long and would confuse potential readers. In my opinion, it is good to write short and concise sentences. Otherwise, the meaning of a sentence is lost to a large extent.
Author Response
Response to Reviewer 4 Comments
Thank you for the reviewer's comments and advice. I wrote the answers to the reviewer's comments and advice in the sentences below and attached the revised paper as a separate PDF file.
Comments and Suggestions for Authors
In this version of the manuscript, reviewer recommendations 3 and 4 were not performed. These recommendations are particularly important because, on the one hand, they are related to the validation process of the proposed structures. On the other hand, these recommendations are related to performing a comparative analysis and defining the advantages and disadvantages of the proposed structures. I recommend that after observing notes 3 and 4, the manuscript be considered again for publication in the journal Electronics.
Author's Response to Reviewer's Comments and Suggestions
Thank you for your suggestions and comments. I revised the paper reflecting your point and suggestion.
- To improve the accuracy of this paper, I added TFT behavior using real IGZO channel[18] and simulation results reproducing them to Figure 4. As shown in the figure, the simulation results behave very similarly to the actual measurement results, and for this reason, it is believed that the reliability of the physical and electrical characteristics of the IGZO channel used in this paper can be verified. The added figures and contents were shown in red letters.
- The comparison items presented in the last table were classified in more detail and detailed comparisons were added. In particular, not only performance improvement but also process cost aspects were added so that those who read the paper could refer to this structure when applying it. Also, the added contents were marked in red letters.
Comments on the Quality of English Language
Some sentences, especially those added in this version of the manuscript, are too long and would confuse potential readers. In my opinion, it is good to write short and concise sentences. Otherwise, the meaning of a sentence is lost to a large extent.
Author's Response to Reviewer's Comments on the Quality of English Language
Thank you for your suggestions and comments. I revised the paper reflecting your point and suggestion. For the too long sentences you pointed out, I tried to increase readability by properly dividing them. The modified parts were also marked in red letters.
Author Response File: Author Response.pdf
Reviewer 5 Report
The revised manuscript entitled “A improved structure enabling hole erase operation when using IGZO channel in 3D NAND Flash structure to which COP (Cell-On-Peri) structure is applied” resubmitted by Yun-heub Song and co-workers for reconsideration for publication in the MDPI journal Electronics presents definitely a higher level than the first submission. The Authors have performed all the required corrections and added additional explanations. Also, additional explanations are properly presented in the answers to the Reviewer's comments. Therefore, I consider the revised manuscript is suitable for publication in the MDPI journal Electronics. My congratulations to the Authors. I wish You all the best and the next good papers.
I would only suggest changing the title “A improved..” to “An improved…”.
I would only suggest changing the title “A improved..” to “An improved…”.
Author Response
Comments and Suggestions for Authors
The revised manuscript entitled “A improved structure enabling hole erase operation when using IGZO channel in 3D NAND Flash structure to which COP (Cell-On-Peri) structure is applied” resubmitted by Yun-heub Song and co-workers for reconsideration for publication in the MDPI journal Electronics presents definitely a higher level than the first submission. The Authors have performed all the required corrections and added additional explanations. Also, additional explanations are properly presented in the answers to the Reviewer's comments. Therefore, I consider the revised manuscript is suitable for publication in the MDPI journal Electronics. My congratulations to the Authors. I wish You all the best and the next good papers.
I would only suggest changing the title “A improved..” to “An improved…”.
Author's Response to Reviewer's Comments and Suggestions
Thank you very much for agreeing to publish this paper. Also, the point you pointed out was my mistake and now it has been corrected and marked in red.
Author Response File: Author Response.pdf
Round 3
Reviewer 4 Report
According to the overall recommendation definitions for this manuscript, further experiments are necessary. Also, further analyzes are needed to highlight the advantages, as well as define some disadvantages of the proposed structures. Therefore, I do not recommend that this manuscript be accepted for publication in the journal Electronics.
I have no complaints about the quality of the English language.
Author Response
Response to Reviewer 4 Comments
Thank you for the reviewer's comments and advice.
Comments and Suggestions for Authors
According to the overall recommendation definitions for this manuscript, further experiments are necessary. Also, further analyzes are needed to highlight the advantages, as well as define some disadvantages of the proposed structures. Therefore, I do not recommend that this manuscript be accepted for publication in the journal Electronics.
Author's Response to Reviewer's Comments and Suggestions
Thank you for your suggestions and comments. I totally agree with the reviewer. And accordingly, our research team is working to build and measure the actual device of the proposed structure. However, it is currently difficult to secure the operation result due to leakage current problem of insulating film and trap problem. Now, we're doing our best to solve these problems, but unfortunately, it's hard to solve them by the revision period of the paper. In addition, the purpose of this paper is to explain that when an oxide semiconductor such as IGZO is used as a channel in a 3D NAND Flash structure, erase operation that cannot be solved by the current structure is possible with the proposed structure. Therefore, I believe that the material parameters of IGZO channel, which reflect actual measurement results, and the material parameters and simulation models used in many previous simulation papers, are not perfect but can show some reliable simulation results. Therefore, I sincerely ask the reviewer to agree to accept of this paper.
