Investigation of Multi-Mesa-Channel-Structured AlGaN/GaN MOSHEMTs with SiO₂ Gate Oxide Layer

Round 1

Reviewer 1 Report

The manuscript reports on a study of a multi-mesa-channel (MMC) AlGaN/GaN MOS-HEMT utilizing SiO₂ gate dielectric. In my opinion, the manuscript cannot be published in the present form by the following reasons.

1. The MMC HEMTs have been suggested more than a decade ago. Since that time there was an intensive activity in this field aimed at understanding specific operational features of such devices. In particular, many conclusions made by authors in the manuscript, like Vth shift and better heat removal, are not new, being already published. This activity and the results obtained by other researchers, i.e. the existing background, is not properly reviewed in the introductory part of the manuscript. The only exception is Ref.[13] but its citation does not contribute to solid and sound background review. As a result, the novelty of the results reported by authors is not evident. A major part of the text in Sec.1 describes various approaches to the mesa etching and SiO₂ deposition. However, no new results relevant to these issues are reported in the manuscript. The clear goal of the study performed is not formulated as well. So, this section should be rewritten cardinally in order to indicate still existing knowledge gap, which, in turn, determines the goal of the research reported in the manuscript.
2. I believe that comparison of the performances of planar-reference and MMC HEMTs made by authors is not quite correct. Problems originate from the use of effective channel width of 3.76 μm for calculating the specific drain-source current in [A-mm]. From the technical point of view, much more important is the total current divided by the total width of the gate electrode, which is about 50 μm. If so, the MMC-HEMT would possess a worse performance compared to planar HEMT. It should be noted that up to now, a way for comparison performances of the planar and MMC HEMTs is not yet conventionally established. Therefore, many researchers (see, e.g., Ref.[13] and other relevant publications) compare the total current-voltage characteristics of the devices, pointing out also the width of individual mesa and inter-mesa spacing, as well as the total width of the gate region. Such a way enables readers to make their own assessment of the MMC-HEMT performance based on the practically important criteria. As there is no commonly accepted criteria, the conclusion that MMC-HEMT has a better performance (in terms of characteristics compared in the manuscript) cannot be regarded as correct one.
3. The incorrectness of the performance comparison, i.e. comparison of the devices operating at very different total currents, is extended, in particular, to interpretation of the obtained result. Authors invoke a better lateral heat spreading in order to explain a lower MMC-HEMT self-heating. However, much more important is here the fact that the total heat-release power is much higher in the planar devices than in MMC-HEMT operating at much lower current. This factor is ignored by authors.
4. Fig.1a with the HEMT design looks like there are some gaps between the metallic gate electrodes deposited on the top of every mesa, though Fig.1b demonstrates that the gate electrode is continuous.
5. Authors do not give in the manuscript the value of the gate length in the MMC-HEMT. Assuming it to be the same as in the planar HEMT (10 μm), one can estimate the series resistance of the 2DEG channel from the mobility and sheet electron concentration, which is ~3.3 Ω-mm. This value is about 1.5 times higher than the measured Ron of the MMC-HEMT. At that, the contact resistances were not taken into account. How such a discrepancy may be explained?
6. In my opinion, there is no reason to introduce a new name ‘Fin-Gated Nanochannel Array Structured …’ instead of ‘Multi-mesa-channel …’ well known to many readers involved into the nitride HEMT activity.
7. English should be improved substantially. It would be desirable to apply for professional proofreading service for this purpose. Now the manuscript contains many incorrectly used words, poorly understandable phrases, and some typos.

Author Response

Dear reviewer:

At first, we like to express our sincere appreciation for your useful suggestions and comments. These suggestions and comments are certainly very useful for us to revise the manuscript and valuable for our further researches.

According to your suggestions and comments, we revised the manuscript. The changes in the revised version as well as the reply to reviewers’ comments are listed on a point by point basis. Thanks!

Sincerely Yours

                                                  Ching-Ting Lee

Author Response File: Author Response.pdf

Reviewer 2 Report

A. Brief Summary:

In this article the authors demonstrated enhanced performance of a metal-oxide-semiconductor high electron mobility transistor MOSHEMT based on the AlGaN/GaN heterostructure (AlGaN/GaN MOSHEMT). The enhanced performance in DC and frequency-dependent parameters was achieved by introducing a fin-gated nano channel array structure.  

 

B. Broad Comments:

 

The novelty of the presented work rests in the fin-gated nanochannel array combined with a thin layer of silicon dioxide gate dielectric introduced into the AlGaN/GaN MOSHEMT.  The nanometer scale device structure was achieved through electron beam lithography. This configuration allows for better electric field control at the gate.  Normally HEMT devices deliver better current characteristics owing to the high electron mobility at the 2D electron gas channel however, it has a limitation of being a normally-on transistor because of difficulty in depleting the channel due to very high carrier densities.  However, due to the authors’ reported new structure that enables gate control, there is an enhancement demonstrated in positively shifting the threshold voltage of the transistor.  Moreover, they also demonstrated over 3 orders of magnitude improvement to the gate leakage current, over 3 times better subthreshold swing, and significant increase in transconductance. 

 

The overall structure of the paper is sound and mostly well-written although there are minor problems with English grammar and style. The data presentation is clear, concise and well-captioned.  However, the data and figures in the paper focuses mostly on the device structure and performance but not enough of material characterization data that would make it more useful for the readers of Coatings.

 

C. Specific Comments:

 

1. In Fig. 1 it would help the readers to also show a HR-TEM closeup at the heterostructure site to show the interfacial layer between the GaN substrate and the AlGaN layer.
2. In Section 2 within the lines 58-61, the authors should mention what metalorganic precursors they used in the MOCVD growth of GaN and AlGaN.  This is a standard way or reporting materials and methods.
3. In Fig. 3 a and b the Ids vs Vgs curves (black plots) should be in log scale in the vertical axis.  This will clearly show the location of the threshold voltage, and clearly delineate the linear and saturation region of operation of the transistor.
4. The paper needs to be improved in terms of English grammar and style. Several errors for example are:  In line 32  “high-frequency devices have attracted significantly attention.” should be “significant attention.” Also the word “Besides” is being used incorrectly in the entire paper. The authors should use the word “moreover” instead of “besides.” These are just a few of the errors in grammar and usage.  The authors are urged to carefully scan and correct the manuscript of grammatical and stylistic errors.

 

D. Recommendation:

After final review of the manuscript this reviewer recommends this paper to be reconsidered after major revision.

Author Response

Dear reviewer:

At first, we like to express our sincere appreciation for your useful suggestions and comments. These suggestions and comments are certainly very useful for us to revise the manuscript and valuable for our further researches.

According to your suggestions and comments, we revised the manuscript. The changes in the revised version as well as the reply to reviewers’ comments are listed on a point by point basis. Thanks!

Sincerely Yours

                                                  Ching-Ting Lee

Author Response File: Author Response.pdf

Reviewer 3 Report

The reviewed paper is devoted to study of AlGaN/GaN MMC-MOSHEMTs with gate-recessed structure and SiO2 gate oxide. They found that the the investigated structure enhance gate control capability and reduce self-heating effect in the channel, and that the performances of the MOSHEMTs is improved. The reviewed paper presents interest for wide scientific community.

 

I recommend to accept paper as is.

Round 2

Reviewer 1 Report

The manuscript has come for the secondary reviewing. Authors have responded convincingly the comments 4 and 5 of the previous report. However, other comments were not properly addressed. In more detail, this point of view is explained below.

1. First of all, authors are claiming throughout the manuscript that the MMC-HEMT exhibits a better performance than the fabricated planar reference HEMT. As it has been already explained in the previous report, this claim is incorrect and misguiding. The incorrectness originates from referring the MMC-HEMT characteristics to the total width of the 2DEG channel (3.76 μm) instead of the total width of the gate electrode (50 μm). If authors would make properly a background review, they could recognize that nobody tried to compare the linear saturation current densities and/or transconductances of planar and MMC HEMTs, as it has no real sense (referring of the device characteristics to the total width of the gate electrode instead of the total 2DEG channel width changes the conclusion on the device “performance” to the opposite one). That is why, the “performance comparison” in the author’s context should be excluded from the manuscript.

2. Sec.1 of the manuscript does not present a satisfactory review of the existing background on the topic of the reported study. A proper review should include, at least, the info about:

- who has suggested the MMC-HEMT design and in what publications (with citing pioneering papers and available reviews)

- what are advantages of the MMC-HEMT design compared to conventional ones revealed in previous studies (with appropriate citing of key publications)

- who has suggested to use the gate dielectrics in III-nitride HEMTs in order to suppress the gate leakage and what dielectrics are commonly used for this purpose (with citing pioneering papers and available reviews)

- what is the knowledge gap still remained after the previous studies

- what is the goal and novelty of the study reported in the manuscript an how do they address the identified knowledge gap

I would like to emphasize that the use of SiO₂ film in the MMC-HEMT cannot be regarded by itself as the novelty of the reported study since (i) this dielectric is among those (SiO₂ , Al₂O₃ , and SiN_x) conventionally utilized in MIS-HEMTs, (ii) its capability of gate leakage suppression is well known from the studies of MOS- and MIS-HEMTs, and (iii) the use of the oxide gate dielectrics in MMC-HEMTs has been already demonstrated (see, in particular, the paper by Hsin-Ying Lee et al., Materials 14, 5474 (2021)).

Identification of novelty of the reported study is an important point of the paper, as a simple demonstration of the author’s capability of fabricating devices like MMC-HEMTs and measuring their characteristics is still insufficient for publication. Since it was not done in the revised manuscript, authors just reproduced in Sec.3 the facts and conclusions (V_th shift, self-heating, gate leakage, etc.) already known from other publications. This cannot be a subject of new publication.

3. English may be polished additionally. In particular, the title may be made less cumbersome and more clear fro readers, like ‘Investigation of Multi-Mesa-Channel AlGaN/GaN MOSHEMTs with SiO₂ Gate Dielectric’; the word ‘array’ may be omitted everywhere from the combination ‘multi-mesa-channel array’, etc.

Author Response

Dear reviewer:

At first, we like to express our sincere appreciation for your useful suggestions and comments. These suggestions and comments are certainly very useful for us to revise the manuscript and valuable for our further researches.

According to your suggestions and comments, we revised the manuscript. The changes in the revised version as well as the reply to reviewers’ comments are listed on a point by point basis. Thanks!

Sincerely Yours

                                                  Ching-Ting Lee

Author Response File: Author Response.pdf

Reviewer 2 Report

After careful consideration of the revised manuscript and response letter of the authors, this reviewer recommends publication of this article in Coatings pending the editor’s requirements.

Author Response

Dear reviewer:

Thank you for your comments.

Sincerely Yours

                                                  Ching-Ting Lee

Author Response File: Author Response.pdf

Round 3

Reviewer 1 Report

It looks like my discussion with authors of the reviewed manuscript has reached a deadlock. There are two principal points of disagreement.

The first one is different assessment of the MMC-HEMT performance. Authors insist that MMC-HEMT exhibits a higher transconductance and saturation current ( measured in A/mm) compared to conventional planar HEMT. In my opinion, the above conclusion is based on a trick, i. e. on the use of an effective 2DEG channel width instead of the conventionally invoked width of the gate electrode. If the gate-electrode width is used for estimation of the transconductance and saturation current, than they should be reduced by a factor of ~13, resulting in an opposite relationship between the parameters of the MMC- and planar HEMTs. As the conclusion made by authors is dependent on the way of calculation, it should be regarded as unreliable and, therefore it should be avoided in the manuscript.

The second point concerns identification of novelty of the reported study. Authors believe that combination of SiO₂ as gate dielectric with the PEC mesa etching and (NH4)₂S_x surface treatment in MMC-HEMT fabrication just provides the novelty of the study. In my opinion, all these technological approaches are well known from the studies of conventional devices and no specific features are visible in their application to MMC-HEMTs. So, the novelty of the reported study is not evident. Moreover, many effects discussed in Sec.3, e. g. shift of the threshold voltage in MMC-HEMT, different self-heating in MMC- and planar HEMTs, gate leakage reduction due to use of a gate dielectric, have been already reported in the other studies, which was not mentioned in the background review made in Sec.1.

These principal points do not allow me regarding the manuscript as ready for publication. As the authors disagree with my comments, there is no sense to continue our discussion. So, I would like to ask Editors either to transfer the manuscript to one more referee for independent expertize or to make a decision on the manuscript publication by themselves.