Benefits from Using Very Thin Channel Layer for TFTs

Round 1

Reviewer 1 Report

I have attched PDF for it as attached.

Comments for author File: Comments.pdf

Author Response

Please see the attachment

Author Response File: Author Response.pdf

Reviewer 2 Report

Samb et al., in the paper draft "Benefits from using very thin channel layer for TFTs" presented some important results.

Suggest accepting with minor revision with bellow comments!

Line: 84-86: Author should explain how crystalline fraction increased with increased thickness. Similar explanation is needed for surface roughness in line 101-103.

Table 1: Mobility is intrinsic property of materials. Authos should explain how mobility is increases with the thickness. Mobility thing should be discussed more with references.

Line 165-166: "At higher thickness, it is screened by the defects." - provide a reference or explain bit more here.

 

Author Response

Please see the attachment

Author Response File: Author Response.pdf

Reviewer 3 Report

Opinion:

 

Samb and his coworkers present a work in which the thickness of thin channel layer of metal oxide based transistors is proposed as a very important factor on the performance of such transistors and their stability. The work is correct, the idea is good and the proposition is interesting and deserves to be discussed. However, the manuscripts is very speculative, the arguments are in some cases only qualitative and even thought the conclusion is positive it requires to have a more quantitative support. This work could be accepted if it is complemented with a lot of references to support the arguments and a comprehensive comparison of this proposition with those mentioned in the discussion. Some small details should be also account.

 

-        Many people can think TFT stands by thin film transistors, however the authors should clarify this because the term is not public domain.

-        The introduction is poor, the theme is enough important to have a more concise presentation and all the other models and omitted references should be localized having a more extensive declaration.

-        The mention of the phenomenon in amorphous silicon based transistors should also have a more extensive presentation.

Therefore my recommendation is accepted with major revision.

 

Author Response

Please see the attachment

Author Response File: Author Response.pdf

Round 2

Reviewer 1 Report

Commet might be sent from Referee.

Comments for author File: Comments.pdf

Author Response

Dear Colleague,

 

I would to thank you again for your renewed important contribution to our paper and for the time you consumed to consider again the referenced submission.  

I really appreciate the seriousness of your reviewing.

Following your previous review report, we added a lot of references on TFTs that use different materials. We would to thank you for this important contribution to our paper.

I send you here the answers to your remarks and questions given in your second report.

 

1. Paper on silicon TFTs or general?

The paper focuses mainly on the known improvement of the parameters of Thin Film Transistors when the thickness of the channel is lower. It tries to explain the improvements that occur independently on the material used as channel layer, single crystalline silicon, amorphous or polycrystalline silicon and metal oxides.

The common origin of this improvement independently on the materials has to be found in the electrostatic interaction between the different thin layers.

 

Of course, we used the results obtained on silicon TFTs because they are our results. These results are given only to illustrate the behaviours­­ and not to support any one of our arguments. We cannot present the results obtained by other labs. HOWEVER all the related results are given in the references.

 

In the paper, after showing the improvements of the parameters of our silicon TFTs, we give the references showing that these improvements are similar to those got for other TFTs. The improvements are not unique to silicon TFTs.

 

As usual, we try to explain the improvements by specific arguments linked to the structure of our silicon TFTs. The paper can stop at this level. But in this case it will be a paper on silicon TFTs.

Even if we consider silicon TFTs only, we show that these arguments cannot explain, or at least explain partially only, the improvements.

 

As the behavior is common to all TFTs whatever the material, we tried to find a general explanation that can be used whatever the material used as channel. Effectively, the trend is universal.

For this purpose we used a model to simulate the characteristics of TFTs. Figure 5 shows the simulated behavior of the threshold voltage and subthreshold swing as a function of the channel thickness for single crystalline silicon TFT. We found the same trend as the experimental one, meaning a decrease of the threshold voltage and an increase of the subthreshold swing when the thickness increases.

The same trend is confirmed from the simulation model when introducing electrically active defects in the silicon channel, mimicking as-deposited silicon material (Figure 6).

The trend is explained from the behavior of the lateral electric field (horizontal field between the source or drain and the channel) (Figures 7 and 8). This lateral field makes more or less easy the injection of electrons from the source to the channel. The lateral field increases when the channel thickness decreases. When the material is disordered (with electrically active defects), the lateral field is nearly null at high channel layer thickness. The channel forms with a lot of difficulty in this case. The lateral field appears only when the thickness is very low leading to an easy formation of the channel.

We show in this paper that only this explanation based on electrostatic arguments can be acceptable to understand the general behavior observed whatever the material of the channel. So why, the paper is general to the TFTs and not be applied to silicon TFTs only.

 

1. “In practical applications, very thin channel is adopted for the TFTs”

You are right. Of course following their name TFTs use thin channel layer. However, the paper shows and explain why it is better to use the thinnest channel layer.

 

1. “For the discussion on silicon TFTs, gate insulator is important in terms of reliability. Authors described in the case of SiN gate insulator. Discuss for the case of SiO2 gate insulator.”

For the stability of TFTs, you are right, both channel layer and gate insulator are important in terms of reliability. We had to use both usual SiN and SiO2 insulators. However, the purpose of the present paper is only to discuss what benefit we can expect when using thinnest channel layer. So why we use the same insulator and varying only the thickness of the channel layer in the purpose to highlight the benefit of using thinner layer.

 

The paper is not a paper on the stability of TFTs. Such a paper can involve only the specific properties of the channel material and of the insulator. For your information, TFTs that use SiO₂ are generally less stable than TFTs that use SiN. We have a lot of experimental results on this comparison. If you have other results on other TFTs, we will be happy to collaborate in future in the preparation of a paper on the stability of TFTs.

 

1. “For the merit of thin channel for sharp S value, authors explained only by the flatness (smoothness) which relates to the trap states (energy states) due to the SiN/Si insulator. The trap states density per area in poly-silicon channel (under gate) should

be taken into account?”

The higher subthreshold swing when the thickness of the channel layer increases can be physically due to the higher surface roughness or the higher deep state density in the channel as shown the following equation of the subthreshold swing/

where C_ins is the capacitance per area unit of the gate insulator, and  are the deep state density in the channel and the interface state density respectively. An increase of the interface state density N_SS due to an increase of the surface roughness or an increase of the deep state density N_ep lead to an increase of S.

In our case, the crystalline fraction inside the channel layer improves when the thickness of the layer increases. This means that normally Nep decreases when the thickness increases. Following the previous equation S has to decrease consequently! The experimental result is opposite. The deep defect density cannot be involved to explain the experimental behavior of S.

 

 

1. “Include the grain size in Table 1”

The grain size was added in Table 1.

 

1. Explain the crystalline volume. (in line 63. Analysis? )

The crystalline volume is deduced from Micro Raman analyses at a wavelength of 632.8 nm. The Raman spectrum can be fitted by three Gaussian peaks that correspond to the crystalline band at 520 cm^-1 I_c, the amorphous band at 480 cm^-1 I_a and an intermediate component peaked at 494--510 cm^-1 I_i which is associated with bond dilation at grain boundaries. The crystalline volume fraction F_c is determined as

where l_c, I_i and l_a are  the integrated intensities of the 3 peaks.

Reference 14 gives more detail on the analysis.

 

1. “Benefits are not described in Conclusion”

The benefits are described in lines 7 to 9 of the conclusion. The main benefits are a decrease of the subthreshold slope and a better stability under gate bias stress when the thickness of the channel layer decreases

 

1. “For the technical words, Deposited crystalline, ”deposited poly-crystalline “is recommended to use in place of “deposited crystalline,,,,”.

Thank you for this remark. We thank that we changed all our “deposited crystalline” by “deposited micro-poly crystalline” following your first review report. We found two other lines with “deposited crystalline”. We corrected these 2 lines.

 

1. “The order the reference number is irregular.”

Thank you. The order was corrected.

 

1.  C →℃  SF6→SF6

°C and SF₆ are corrected.

 

 

I hope you will be satisfied by these explanations and I would to thank you again for your interest in this work.

As I said you just previously, I would appreciate if we can collaborate in future around the problematic of TFTs, after this reviewing. Your high expertise in the domain, as demonstrated by your review, is really appreciated by us.

 

Sincerely yours.

 

Dr. Samb

Reviewer 3 Report

I am glad to see we have a new version of the manuscript and more glad because I have a version which deserves to be published. Thank you for your seriois and profesional work.

Author Response

Thank you very much for agreeing to publish the paper

Best regards

Samb