Memristive Properties of PANI-Polysterene/PVDF-TrFE Interface
Round 1
Reviewer 1 Report
The authors did a good work. Minor corrections have to be done before acceptance of the paper.
What the reason for evaporating the acetone at – 10°C.
Author Response
The authors did a good work. Minor corrections have to be done before acceptance of the paper.
What the reason for evaporating the acetone at – 10°C.
We thank the Reviewer for the encouraging response. As for the only question, we added the explanation (highlighted in turquoise color) that the evaporation of acetone at such low temperature was performed for obtaining PVDF-TrFE film with a high content of amorphous phase, which should incorporate individual ferroelectric crystalline inclusions.
Reviewer 2 Report
The manuscript "Memristive properties of PANI-polysterene/PVDF-TrFE interface" studied the fabrication and I-V characterization of the PANI-polysterene/PVDF-TrFE layered structures. The idea of the manuscript makes some sense. However, the technical quality is limited and the conclusion is not supported by the results. Therefore, I can not recommend publication. Below is detailed comments and questions:
What is the use of PVDF layer in the layered structure? If the authors proposed to use the polarization state to achieve distinctive memristive states, the different polarization state of the PVDF layer need to be clearly defined. Also, with PVDF in micrometer scale thickness, the voltage applied is to low to be able to change the polarization.
The I-V curves in Fig. 5 is unclear and confusing. To show distinctive memristive states, one full loop is enough. Such as these typical ones shown in Nature Materials 16(12):nmat5009. These curves in Fig. 5 could be mostly due to other factors, such as defects or film degeneration etc.
According to the SEM/EDX images, the PANI-polysterene/PVDF interface is not clear. Interlayer diffusion was clearly observed. Therefore, what the multilayer really is need to be justified, whether it is a true interlayered structure or a blend?
Author Response
The manuscript "Memristive properties of PANI-polysterene/PVDF-TrFE interface" studied the fabrication and I-V characterization of the PANI-polysterene/PVDF-TrFE layered structures. The idea of the manuscript makes some sense. However, the technical quality is limited and the conclusion is not supported by the results. Therefore, I can not recommend publication. Below is detailed comments and questions:
We thank for the comments aimed to clarify the details of our experiments and their results. Answers to each of questions is given below, and the respective additions are included to the revised version of the manuscript, where they are highlighted in green.
What is the use of PVDF layer in the layered structure? If the authors proposed to use the polarization state to achieve distinctive memristive states, the different polarization state of the PVDF layer need to be clearly defined. Also, with PVDF in micrometer scale thickness, the voltage applied is to low to be able to change the polarization.
PVDF-TrFE is one of the most widely known ferroelectric copolymers, we have added a reference motivating its usage. At the same time, new structural studies, results of which are given in our work show that the layers PVDF-TrFE obtained in the work form an amorphous matrix, which carries nanoscale ferroelectric inclusions and the performed estimates already show that the applied external voltages are sufficient to switch the polarization in individual nanocrystals.
The I-V curves in Fig. 5 is unclear and confusing. To show distinctive memristive states, one full loop is enough. Such as these typical ones shown in Nature Materials 16(12):nmat5009. These curves in Fig. 5 could be mostly due to other factors, such as defects or film degeneration etc.
We thank for this suggestion, it was taken into account and Figure 5 was reworked. In the revised version it is separated into two sub-panels, where new sub-figure (a) clearly shows one full loop indicating memristive properties, while (b) remains to discuss possible degradation issues.
According to the SEM/EDX images, the PANI-polysterene/PVDF interface is not clear. Interlayer diffusion was clearly observed. Therefore, what the multilayer really is need to be justified, whether it is a true interlayered structure or a blend?
Indeed, in the course of measuring the current-voltage characteristics, irreversible diffusion processes occur and a blend is formed (as confirmed by EDS data) – the respective explanations are added to the text.
Reviewer 3 Report
I appreciate the authors effort for this experiment. In this paper, they tried to study memrisitive properties of PANI-PS/PVDF-TrFE heterojunction. Their approach has several issues which we can summarize as following:
1) Their methodology to fabricate devices could cause some unexpected results. Instead of the dropwise they could use a spin coating method which is more consistent and gives the cleaner interface. Also, coating with ultrasonic could cause some sort of diffusion which they observed in SEM data (fig 1&2). Generally, for this type of measurement we need to have the cleaner interface between PANI-PS/PVDF-TrFE. Could you perform the experiment with spin coating method?
2) In figure 3, The authors discussed about the XRD data. Based on ref [14], it explains origin of α-PVDF-TrFE peaks which come from irradiation of x-ray and also they claim the β-PVDF-TrFE peak will disappear after x-ray irradiation (certain amount of time for example 1 hour). The x-ray exposure time is very important to have better insight for this type of analysis. Would you provide this type of information? Also Could you perform XRD after each cycle of electrical measurement?
3) In figure 4, they present the data for FTIR. Based on ref [14], first of all, observation of the peaks (FTIR for PVDF-TrFE) for very thick sample (more than 100nm) is very hard and the absorption is really high because of multiple bands. So how could you get the signal from thick layer of PVDF-TrFE (30um)?
4) The major result of this paper is summarized in fig4. It is not very clear the bias direction in experiment. they can clarify by adding inset figure into the fig 4 and makes it easier for readers to understand the experiment.
5) In Fig 4, They observed decreasing the current amplitude after each cycle. Why does the current amplitude decrease after each cycle? How could the time interval without any bias affect? Or could it be just annealing during the time which cause some unexpected results? does the samples maintain under vacuum or expose to air in each interval? It is not very clear how the rule of interface effects changing current amplitudes. they explain the reason on discussion part but it is not very clear how the charge trap will cause it .
6) Based on ref [10], they observed rectifying behavior in positive side which also have been observed in this experiment too. Why do you see hysteresis in forward bias ? You mentioned about charge traps. Could you please explain it more clearly? If you have cleaner interface, would you except to see about the are of hysteresis will enhance or suppress?
7) In the manuscript, they mention that the area of hysteresis at negative bias are increased after each cycle then suddenly dropped after the 4th cycle. based on the figure 4, the area of hysteresis decreases from cycle two to three, so why? Also why the hysteresis become so symmetrical after 4th cycle? Also if you continue cycling what would you except to see? How would you restore to the original state?
8) Could you explain or calculate the electric field you need to change the polarity of PVDF-TrFE based on the dimension of your hetero-junction ?
9) the authors claim similarity at negatives bias compare to the ref [4], can be explained as the memristive behavior. The figure 4 represents data for this experiment. It looks after 4 cycles; we see the kind of linear behavior with small hysteresis on both side. How would we make the memristive system based on this very nonlinear behavior through cycling process? Also it is very unclear how the polarization switching strength can vary after each cycling. Does it come from bulk polarization or just happen at interface? Also Could you distinguish it from charge trapping mechanism? It we you make a very flat interface with thinner layers would you be able to get better data and analyze the hysteresis area decrease or increase?
In conclusion, the paper requires certainly careful revision.
Author Response
appreciate the authors effort for this experiment. In this paper, they tried to study memrisitive properties of PANI-PS/PVDF-TrFE heterojunction. Their approach has several issues which we can summarize as following:
We thank the Reviewer for the comments and valuable suggestions. The revised version contains additional explanations, which, as we hope, clarify some misleading issues of the previous version and results of additional experiments evaluated to answer the questions related to XRD; all these additions are highlighted in yellow color in the revised manuscript.
At the same time, some comments we have to consider as interesting and valuable suggestions for further studies since they are pointed to various other approaches and generalizations apart of the discussion of the particular case considered in the present work.
1) Their methodology to fabricate devices could cause some unexpected results. Instead of the dropwise they could use a spin coating method which is more consistent and gives the cleaner interface. Also, coating with ultrasonic could cause some sort of diffusion which they observed in SEM data (fig 1&2). Generally, for this type of measurement we need to have the cleaner interface between PANI-PS/PVDF-TrFE. Could you perform the experiment with spin coating method?
Certainly, in the course of further research, we will attempt to obtain interfaces using the proposed method.
2) In figure 3, The authors discussed about the XRD data. Based on ref [14], it explains origin of α-PVDF-TrFE peaks which come from irradiation of x-ray and also they claim the β-PVDF-TrFE peak will disappear after x-ray irradiation (certain amount of time for example 1 hour). The x-ray exposure time is very important to have better insight for this type of analysis. Would you provide this type of information? Also Could you perform XRD after each cycle of electrical measurement?
Additional XRD structural studies were performed. Their results are presented in the newly added Figure 3 b and its discussion in the revised text.
3) In figure 4, they present the data for FTIR. Based on ref [14], first of all, observation of the peaks (FTIR for PVDF-TrFE) for very thick sample (more than 100nm) is very hard and the absorption is really high because of multiple bands. So how could you get the signal from thick layer of PVDF-TrFE (30um)?
We apologize for the annoying typo. In fact, the absorption spectra for the studied structures were studied and presented. This issue is corrected in the text
4) The major result of this paper is summarized in fig4. It is not very clear the bias direction in experiment. they can clarify by adding inset figure into the fig 4 and makes it easier for readers to understand the experiment.
We thank for this suggestion, it was taken into account and Figure 5 was reworked. In the revised version it is separated into two sub-panels, where new sub-figure (a) clearly shows one full loop indicating memristive properties, while (b) remains to discuss possible degradation issues.
5) In Fig 4, They observed decreasing the current amplitude after each cycle. Why does the current amplitude decrease after each cycle? How could the time interval without any bias affect? Or could it be just annealing during the time which cause some unexpected results? does the samples maintain under vacuum or expose to air in each interval? It is not very clear how the rule of interface effects changing current amplitudes. they explain the reason on discussion part but it is not very clear how the charge trap will cause it .
Decrease in the amplitude of the current-voltage characteristics with an increase in the number of measurement cycles may indicate the presence of long-term conduction relaxation in the sample under study. Long-term traps capture charge carriers and can hold them longer than the gap between cycles (as shown in Fig.6). Thus, carriers that could participate in the conductivity are excluded from it.
6) Based on ref [10], they observed rectifying behavior in positive side which also have been observed in this experiment too. Why do you see hysteresis in forward bias ? You mentioned about charge traps. Could you please explain it more clearly? If you have cleaner interface, would you except to see about the are of hysteresis will enhance or suppress?
The hysteresis of the current-voltage characteristic observed at the positive bias voltage is due to an injection of carriers from the electrode to the trap site located in the interface between semiconductor-dielectric layers as well as in the case considered in [Biswas, B., Chowdhury, A., Mallik, B., Tuning of electrical conductivity and hysteresis effect in poly(methyl methacrylate)–carbon nanotube composite films. RSC Advances 2012, Vol. 3, pp. 3325-3332]. Additional explanations are included into the text.
7) In the manuscript, they mention that the area of hysteresis at negative bias are increased after each cycle then suddenly dropped after the 4th cycle. based on the figure 4, the area of hysteresis decreases from cycle two to three, so why? Also why the hysteresis become so symmetrical after 4th cycle? Also if you continue cycling what would you except to see? How would you restore to the original state?
Since this diffusion process is irreversible, it leads to irreversible degradation of the memristive properties of the structure under study. Diffusion leads to the formation of blend at the phase boundaries with increasing number of measurement cycles. Since interface (abrupt interface of electron states) degrades, it becomes impossible to apply an external voltage sufficient to switch the polarization of ferroelectric crystallites. Therefore, with an increase in the number of cycles, the hysteresis of the current –voltage characteristics of the structure becomes due only to the processes of capture and release of carriers from traps. An addition discussion is given in the text.
8) Could you explain or calculate the electric field you need to change the polarity of PVDF-TrFE based on the dimension of your hetero-junction ?
The revised section “Discussion” contains now the respective estimation.
9) the authors claim similarity at negatives bias compare to the ref [4], can be explained as the memristive behavior. The figure 4 represents data for this experiment. It looks after 4 cycles; we see the kind of linear behavior with small hysteresis on both side. How would we make the memristive system based on this very nonlinear behavior through cycling process? Also it is very unclear how the polarization switching strength can vary after each cycling. Does it come from bulk polarization or just happen at interface? Also Could you distinguish it from charge trapping mechanism? It we you make a very flat interface with thinner layers would you be able to get better data and analyze the hysteresis area decrease or increase?
Memristive properties of the structure under study emerges due to the both polarization switching processes and the capture and release of charge carriers from traps localized near the interface. At the same time, diffusion leads to the formation of a blend at the phase boundaries with increasing number of measurement cycles. Since the interface (abrupt interface of electron states) degrades, it becomes impossible to apply an external voltage sufficient to switch the polarization of ferroelectric crystallites. Therefore, with an increase in the number of cycles, the hysteresis of the current –voltage characteristics of the structure becomes to exist only due to the processes of capture and release of carriers from traps. The respective discussion is added to the text.
Surely, a flat interface with thinner layers possible would be able to provide better data and analyze. However, we are not sure that the phenomenon under study is possible in a fully crystalline thin PVDF-TrFE layer. We thank the reviewer for the proposed approach and will use it as one of directions for further research.
Round 2
Reviewer 3 Report
Thanks for addressing the comments.
Author Response
Dear Rewiever,
Thank your for a positive response.
Sincerely yours,
Nikita A. Emelianov
On behalf of all co-authors
