Effects of Current Filaments on IGBT Avalanche Robustness: A Simulation Study

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors Dear Editor
  Electrical breakdown study in BJTs may improve various circuit performances.
However, there are a few concerns that need to be addressed.
Hence, the manuscript in its current form cannot be accepted and requires MAJOR REVISIONS.    Comments are attached below.
MAJOR REVISIONS   1. Many terms in the manuscript were not defined before e.g. FS, J1, J2 etc.  If defined then still using the full form e.g. NDR. Authors have to consider subscripts or superscripts in the figure captions.   2. 10.1109/DRC55272.2022.9855818 discuss the current Filaments and its Impact on Avalanche Robustness in GaN diodes using electroluminescence images. The in situ measurements illustrate the most realistic results. Authors need to consider the already reported results and compare if there is any discrepancy.   3. The device scale bar in Fig.1 is missing. Also, Vertical distribution of the electric field in Fig. 3 is pointing in which direction upwards or downwards?   4. Line 190: the increase in temperature decreases the collision current rate of electrons and holes?    5. What is the temperature range utilized in Fig. 12?   6. What is the shape of the current filaments i.e. Hour glass or conical shaped? What is the weakest Link?     7. Conclusion needs to be improved. Comments on the Quality of English Language

The manuscript has severe grammatical mistakes hence, it requires an extensive and thorough English formatting from a Native English Speaker.

Author Response

Comment 1

Many terms in the manuscript were not defined before e.g. FS, J1, J2 etc.  If defined then still using the full form e.g. NDR. Authors have to consider subscripts or superscripts in the figure captions.

 

Author reply:

Thank you for your valuable comments. We checked the article and defined all the terms in the article. For example, the MOSFET is metal oxide semiconductor field effect transistor. FS means field stop and is defined at the end of the first sentence of the last paragraph of Part 1. J₁, J₂ and J₃ are the three p/n junctions inside the IGBT. And we added a note after the first sentence of Part 2 as follows: “The three p/n junctions of the FS-IGBT are J1, J2 and J3, which have been represented in Fig. 1.”. We also illustrate the symbols inside Eq. 1 as follows: “Where q is elementary charge and ε_Si is dielectric constant of silicon.”. Moreover, we have corrected the superscripts and subscripts of the figure captions in the article and corrected the labeling in Figures 1 and 8. For example, we changed “VCE” in the figure captions of Figures 9 and 11 to “V_CE”. Thank you again for your valuable comments.

 

 

Comment 2

10.1109/DRC55272.2022.9855818 discuss the current Filaments and its Impact on Avalanche Robustness in GaN diodes using electroluminescence images. The in situ measurements illustrate the most realistic results. Authors need to consider the already reported results and compare if there is any discrepancy.

Author reply:

Thank you very much for the literature, which helps us to visualize the current filament in the device. The movement of current filaments in this literature is consistent with our study, where the aggregation of current filaments inside the device leads to a localized temperature increase, which drives the current filaments towards the low-temperature region. However, this literature does not provide a theoretical study of the causes and intrinsic mechanisms of the formation of current filaments. Due to the differences in materials and devices, the generation and control mechanisms of avalanche current filaments that lead to GaN PN diode in this literature and Si IGBT in our study may be different. We cited this literature in the article and added a sentence at the end of the second paragraph of section 3.3 as follows: “The phenomenon of current filament movement has been experimentally observed in the literature [28], aligning with our simulation and theoretical analysis.”. Thank you again for your valuable comments and the literature provided.

 

 

Comment 3

The device scale bar in Fig.1 is missing. Also, Vertical distribution of the electric field in Fig. 3 is pointing in which direction upwards or downwards?

 

Author reply:

Thank you for your question. We have relabeled the structure of the FS-IGBT in Fig. 1, and refer to Table 1 for the specific dimensions. The vertical distribution of the electric field in Fig. 3 is pointing downwards, that is from the emitter to the collector. Thanks again for your valuable questions.

Figure 1. Cross-sectional view of the 650 V trench FS-IGBT.

 

 

 

Comment 4

Line 190: the increase in temperature decreases the collision current rate of electrons and holes?

 

Author reply:

Thank you for your question. As you questioned, the increase in temperature complicates the collision current rates for electrons and holes. On the one hand, an increase in temperature increases the number of electrons and holes, which may increase their collision current rate under certain conditions. On the other hand, an increase in temperature leads to an enhancement of the lattice vibrations and thus to an increase in the scattering of carriers, which may counteract the effect due to the increase in the number of carriers and even reduce the collision current rate in some cases. Therefore, in this paper, we emphasize the increase in temperature, which leads to the enhancement of lattice vibrations, thereby decreasing the collision current rate of electrons and holes, and thus realizing the increase in breakdown voltage. We hope that our answers solve your questions. Thanks again for your questions.

 

 

Comment 5

What is the temperature range utilized in Fig. 12?

 

Author reply:

Thank you for your question. As shown in Figure 11, only the avalanche mode of the FS-IGBT from 0 to 10 μs is simulated in this article. And as can be seen in Fig. 11, the maximum lattice temperature inside the device is distributed between 300K and 450K. Therefore, this temperature range is also the temperature range in Figure 12. Thanks again for your valuable questions.

 

 

Comment 6

What is the shape of the current filaments i.e. Hour glass or conical shaped? What is the weakest Link?

 

Author reply:

Thank you for your question. Unfortunately, due to the limitations of 2D simulation, we cannot know the exact shape of the current filament. However, we can guess from the current density distribution graph in Figure 12. The current filament resembles a rectangle in cross-section and has a tendency to be narrower at the top and wider at the bottom, so we speculate that this current filament may be conical in three dimensions. We apologize that this article does not investigate where the current filament is the weakest link for devices. The weakest link may need to be analyzed in conjunction with experiments. Thank you for your perspective, which is of great significance for in-depth research. Thanks again for your valuable questions.

 

 

Comment 7

Conclusion needs to be improved.

 

Author reply:

Thank you for your advice. We optimized part of the description of the conclusion section. The changed conclusions are as follows: “During an avalanche breakdown in an IGBT, the avalanche electrons generated at junction J₂ move toward the collector side, leading to hole injection in the backside collector region, resulting in the formation of an additional negative differential resistance branch on the avalanche breakdown curve. This unique NDR1 branch is the main feature that distinguishes IGBTs from diodes. The avalanche current J_C at the valley point V on the IGBT avalanche breakdown curve, during the transition from the NDR1 branch to the PDR branch, is the primary factor influencing the strength of the current filament. The offset ΔV_CE of the avalanche breakdown curve at high and low temperatures plays a crucial role in regulating the movement speed of the current filament. Both of these factors are predominantly determined by α_pnp. When the α_pnp of the IGBT is larger, the negative differential resistance effect of the NDR1 branch is stronger, and the avalanche current J_C at the valley point V when the avalanche breakdown curve is converted from the NDR1 branch to the PDR branch is larger, which leads to a stronger avalanche current filament. Moreover, as the α_pnp increases, the offset ΔV_CE of the avalanche breakdown curve at high and low temperatures decreases, which results in a slower movement of the current filament and a higher temperature rise of the device. As a result, the avalanche robustness of the IGBT decreases with the increase of α_pnp. Therefore, it is crucial to find a balance in controlling α_pnp during device design.”. We hope that the optimized conclusions will give you a clearer understanding of our research. Thanks again for your valuable comments. 

 

Author Response File: Author Response.pdf

Reviewer 2 Report

Comments and Suggestions for Authors

In this paper, the effects of current filaments on IGBT avalanche robustness is investigated by simulations.

The analysis is interesting, anyway some clarifications and improvements are required. Following are some comments:

 

1. The literature overview should be extended including more recent papers on the topics.

2. It is not well clear the novelty of the proposed analysis with respect to the several papers available in the literature. This aspect must be better highlighted in the introduction.

3. What is currently missing is an experimental validation. It turns out to be difficult whether the results obtained are correct or not. How do they think they can claim that the results obtained reflect the real behaviour of the device?

Comments on the Quality of English Language

Minor editing of English language required

Author Response

Response to Reviewer 2

Comment 1

The literature overview should be extended including more recent papers on the topics.

 

Author reply:

Thank you for your valuable criticism. We extended our search for relevant literature and were successful in adding literature [7], [12], and [28]. Literature [7] of 2021 investigates the movement of current filaments inside IGBTs. The failure of IGBTs in dynamic avalanches was investigated in Literature [12] of 2020. The moving behavior of the current filament was experimentally observed in literature [28] of 2022. All these literatures support our research work in this article. Due to the recent popularity of wide bandwidth semiconductors, studies on traditional Si-based power semiconductors are slowly decreasing, so the references in this article are not very recent. We are deeply apologize for this and hope you will understand and support us. Thank you again for your criticism.

 

 

Comment 2

It is not well clear the novelty of the proposed analysis with respect to the several papers available in the literature. This aspect must be better highlighted in the introduction.

 

 

Author reply:

Thank you for your valuable comments. We emphasize at the end of the second paragraph of the introduction that “previous studies have not established a more explicit link between the NDR effect and the nature of the current filaments, and have not revealed the control mechanism of the NDR effect on the nature of the current filaments”. Therefore, investigating the nature of the NDR effect and revealing the control mechanism of the NDR effect on the current filament and its impact on the robustness of the IGBT is the novelty of this article. We hope that our answer will make you satisfied and clear about the novelty of our research. Thanks again for your valuable criticism.

 

 

Comment 3

What is currently missing is an experimental validation. It turns out to be difficult whether the results obtained are correct or not. How do they think they can claim that the results obtained reflect the real behaviour of the device?

 

Author reply:

Thank you for your criticism. The lack of experimental validation is a major shortcoming of this article, which we greatly regret. We observed the movement of the current filament in the simulation study of this article, which was also verified in the references. The movement of the current filament during device avalanches was captured using a high-speed CCD camera in reference [28], which greatly supports the study of current filament movement in this article. And we added a sentence at the end of the second paragraph of section 3.3 as follows: “The phenomenon of current filament movement has been experimentally observed in the literature [28], aligning with our simulation and theoretical analysis.”. However, the study of current filament strength and moving speed in this article lacks experimental validation for the time being. We strongly hope that based on the simulation research and analysis in this article, we can follow up with experimental validation of this, so that we can optimize our research work. Hope you can understand and support us. Thanks again for your criticism.

 

Reviewer 3 Report

Comments and Suggestions for Authors

The manuscript "Effects of Current Filaments on IGBT Avalanche Robustness: A Simulation Study" is well written.

1. Current Filament Stability: The analysis observes that in a static avalanche breakdown scenario, the current filament generated by the IGBT remains stationary. This lack of movement might indicate a potential instability in the IGBT design, as it does not account for dynamic changes or thermal effects that could affect the device's behaviour over time. It will be more relevant as a simulation for lifetime filament.

2. Temperature Influence: The text discusses how temperature variations can impact the movement and strength of the current filament within the IGBT. An explication of a point highlighted is that temperature rise can lead to changes in the device's behaviour, affecting the current distribution and potentially reducing the IGBT's robustness. Please add more explication for current distribution.

3. Impact of Backside Hole Injection: The formation of the NDR1 branch in the avalanche breakdown curve due to backside hole injection is a distinguishing feature of IGBTs. However, this mechanism could also be seen as a potential defect as it introduces additional complexities in the device's behaviour, potentially affecting its reliability, and it is necessary to add the errors of your simulation.

4. Critical Current Density (JC) Effects: The manuscript mentions how the avalanche current density corresponding to the transition between the NDR1 branch and the PDR branch is a crucial factor in determining the strength of the current filament within the IGBT. This sensitivity to specific current densities could be problematic if not carefully controlled in the device design. Could you be more specific about the sensitivity?

5. Alpha (α) Parameter Impact: The manuscript emphasizes that the parameter αpnp plays a significant role in controlling the negative differential resistance effect and the overall avalanche breakdown characteristics of the IGBT. Managing this parameter effectively is crucial, as variations in αpnp can impact the device's performance and robustness. Could you prevent the device's performance from controlling the film dimension?

6. In my opinion, areas where the design and operation of IGBTs during avalanche breakdown may need careful consideration to ensure optimal performance and reliability.

7. The Conclusions and References are relevant to this manuscript.
8. I recommand minor revision.

on.

1.    
   
   

 

Author Response

Response to Reviewer 3

 

Comment 1

Current Filament Stability: The analysis observes that in a static avalanche breakdown scenario, the current filament generated by the IGBT remains stationary. This lack of movement might indicate a potential instability in the IGBT design, as it does not account for dynamic changes or thermal effects that could affect the device's behaviour over time. It will be more relevant as a simulation for lifetime filament.

 

Author reply:

Thank you for your comments. Your analysis is correct, since the current filament does not move under static avalanche, we performed thermoelectric simulations and also considered the effect of temperature changes on the current filament, thus obtaining the simulation results in Figure 11 and Figure 12. Thank you again for your comments.

 

 

Comment 2

Temperature Influence: The text discusses how temperature variations can impact the movement and strength of the current filament within the IGBT. An explication of a point highlighted is that temperature rise can lead to changes in the device's behaviour, affecting the current distribution and potentially reducing the IGBT's robustness. Please add more explication for current distribution.

 

 

Author reply:

Thank you for your valuable suggestions. Thank you again for your valuable suggestions. We added a description of the current density in Figure 12 earlier in the third paragraph of Section 3.3 as follows :“In Fig. 12(a), when N_P⁺ = 3 × 10¹⁷ cm^-3, the current filament has moved from the left-most to the rightmost end at the 5 μs. By the time 9 μs is reached, the current filament has returned to the leftmost end. While in Fig. 12(b), when N_P⁺ = 1 × 10¹⁸ cm^-3, the current filament only moves about half of the device width at 5 μs, and does not complete one round trip even at 9 μs. ” The addition of this sentence allows the reader to better understand the movement of the current filament inside the device. Thanks again for the advice.

 

 

Comment 3

Impact of Backside Hole Injection: The formation of the NDR1 branch in the avalanche breakdown curve due to backside hole injection is a distinguishing feature of IGBTs. However, this mechanism could also be seen as a potential defect as it introduces additional complexities in the device's behaviour, potentially affecting its reliability, and it is necessary to add the errors of your simulation.

 

 

Author reply:

Thank you for your valuable advice. As you have analyzed, the NDR1 branch on the avalanche breakdown curve does affect the reliability of the device. We will follow up with experiments to further calibrate our simulation model and add this error to our simulation. Thank you again for your valuable advice.

 

 

Comment 4

Critical Current Density (JC) Effects: The manuscript mentions how the avalanche current density corresponding to the transition between the NDR1 branch and the PDR branch is a crucial factor in determining the strength of the current filament within the IGBT. This sensitivity to specific current densities could be problematic if not carefully controlled in the device design. Could you be more specific about the sensitivity?

 

 

Author reply:

Thank you for your question. As you mentioned, our research in this article found that the avalanche current J_C corresponding to the valley point V of the IGBT avalanche breakdown curve at the transition from the NDR1 branch to the PDR branch is the determining factors for controlling the current filament strength. However, since the avalanche of IGBTs occurs so fast and the current density is quite high at this point, which is difficult to control. The simplest and most effective way to control the avalanche current density J_C on NDR1 during device design is to control the α_pnp to change the backside hole injection. Thanks again for your valuable comments.

 

 

Comment 5

Alpha (α) Parameter Impact: The manuscript emphasizes that the parameter αpnp plays a significant role in controlling the negative differential resistance effect and the overall avalanche breakdown characteristics of the IGBT. Managing this parameter effectively is crucial, as variations in αpnp can impact the device's performance and robustness. Could you prevent the device's performance from controlling the film dimension?

 

 

Author reply:

Thank you for your valuable question. I'm not sure that by controlling the film dimension you mean controlling the thickness of the collector T_P⁺? Of course, controlling the thickness of the collector T_P⁺ can equally change the injection of backside hole which will change the value of α_pnp. And the reason for changing the α_pnp by controlling the concentration of the collector N_P⁺ in this article is that IGBTs are all moving towards thinner thicknesses, thus limiting the thickness of the collector. Carefully controlling the thickness of the collector T_P⁺ and controlling the concentration of the collector N_P⁺ can have the same effect on changing the α_pnp. Thanks again for your valuable questions.

 

 

Comment 6

In my opinion, areas where the design and operation of IGBTs during avalanche breakdown may need careful consideration to ensure optimal performance and reliability.

 

.

Author reply:

Thank you for your valuable comments. We strongly agree with you that careful and thorough consideration is needed to improve the reliability of IGBTs during avalanches. Thank you very much for your praise and encouragement of our research work in this article. Thanks again for your valuable comments.

Round 2

Reviewer 1 Report

Comments and Suggestions for Authors

Authors responded most of the comments satisfactorily. However some comments needs to be further explained.

The temperature decrease the collision rate between the carriers when there is an increase in Lattice Vibrations? Authors need to provide references for this very strong and contrasting conclusion. 

Authors response to comment 2 is not convincing. The experimental evidence and theoretical study represents more realistic values. Authors need to provide a literature survey for the experimental evidence and theoretical values for avalanche breakdown in Si-IGBT structures.

Still there are many Grammatical errors in the manuscript.

Comments on the Quality of English Language

Still there are many Grammatical errors in the manuscript.

Author Response

A point-by-point response to the reviewers’ comments

 

We thank the reviewers for their positive and constructive comments and suggestions. We have addressed all the comments point-by-point and revised the manuscript accordingly. In this response letter, comments from the reviewers are summarized in black italic typeface. Our responses are in a regular blue typeface. Our changes to the manuscript are in red. Some of the modified figures have been clearly indicated. See the revised version of the manuscript for detailed modification information.

=============================================================

 

Response to Reviewer 1

Comment 1

The temperature decrease the collision rate between the carriers when there is an increase in Lattice Vibrations? Authors need to provide references for this very strong and contrasting conclusion. 

 

Author reply:

Thank you for your valuable comments. We searched the literature and found strong evidence in Prof. B. Jayant Baliga's book “Fundamentals of Power Semiconductor Devices”. It is clearly stated in the last paragraph on page 766 of the book that an increase in temperature leads to a decrease in impact ionization, as shown in the figure below, and we cite that literature in the article.

Thanks again for your valuable questions.

 

 

 

 

 

Comment 2

Authors response to comment 2 is not convincing. The experimental evidence and theoretical study represents more realistic values. Authors need to provide a literature survey for the experimental evidence and theoretical values for avalanche breakdown in Si-IGBT structures.

Author reply:

Thank you for your valuable comments. We reviewed a large amount of literature and eventually found the following relevant papers with experimental validation. The Breglio, G team observed current images of IGBTs during UIS by means of an ultrafast infrared camera system in the literature published in 2011 and 2013. Images of current filaments of IGBTs under UIS avalanche conditions acquired by time resolved emission (TRE) microscopy observations in the research results of the Endo, K team in 2016 and 2019. Both research teams experimentally observed the movement of the current filaments of IGBTs during avalanches, confirming our simulation analysis in this article. I hope you'll be satisfied with our response this time. Thanks again for your valuable comments.

 

29. Riccio, M.; Irace, A.; Breglio, G.; Spirito, P.; Napoli, E.; Mizuno, Y. Electro-Thermal Instability in Multi-Cellular Trench-IGBTs in Avalanche Condition: Experiments and Simulations. In Proceedings of the 2011 IEEE 23rd International Symposium on Power Semiconductor Devices and ICs; IEEE: San Diego, CA, USA, May 2011; pp. 124–127.
30. Breglio, G.; Irace, A.; Napoli, E.; Riccio, M.; Spirito, P. Experimental Detection and Numerical Validation of Different Failure Mechanisms in IGBTs During Unclamped Inductive Switching. IEEE Trans. Electron Devices 2013, 60, 563–570, doi:10.1109/TED.2012.2226177.
31. Endo, K.; Nagamine, S.; Saito, W.; Matsudai, T.; Ogura, T.; Setoya, T.; Nakamae, K. Direct Photo Emission Motion Observation of Current Filaments in the IGBT under Avalanche Breakdown Condition. In Proceedings of the 2016 28th International Symposium on Power Semiconductor Devices and ICs (ISPSD); IEEE: Prague, Czech Republic, June 2016; pp. 367–370.
32. Endo, K.; Nakamae, K. Temporally- and Spatially-Resolved Observations of Current Filament Dynamics in Insulated Gate Bipolar Transistor Chip During Avalanche Breakdown. IEEE Trans. Device Mater. Relib. 2019, 19, 723–727, doi:10.1109/TDMR.2019.2953197.

 

Comment 3

Still there are many Grammatical errors in the manuscript.

Thank you for your feedback regarding the grammatical errors in the manuscript. We have carefully reviewed and revised the entire document to address these issues.

Author Response File: Author Response.pdf

Reviewer 2 Report

Comments and Suggestions for Authors

The paper is now suitable for publication.

Author Response

Thank you for your positive evaluation of the revised manuscript. I am pleased to hear that it is now suitable for publication. I appreciate your guidance and support throughout the review process.

Round 3

Reviewer 1 Report

Comments and Suggestions for Authors

My previous comments have been addressed well.

No further comment. I recommend the acceptance of this paper.