Photonic Integrated Frequency Shifter Based on Double Side Band Modulation: Performance Analysis

Round 1

Reviewer 1 Report

The authors have designed and simulated using standard building blocks the performance of a frequency shifter in an InP photonic integrated circuit. Their approach is based on the generation of a double-sideband modulation suppressed carrier with an MZM. The authors analyze and compare two configurations for the MZM: Dual-Drive MZM and Single Drive-MZM obtaining a better performance with the former.

Unfortunately, I can not recommend the paper for publication in Photonics for the following reasons:

The results presented for the SD-MZM case have already been published by the authors in the Sensors journal (which they cite in ref. 12), namely Figures 4 upper right and bottom right and Figures 5 upper right and bottom right already appear in the paper of ref. 12. In other words, a significant part of the results presented here are not original. If the authors want to make a comparison between the two different modes of operation of an MZM, they can refer to their own paper (ref. 12) without repeating the same figures. In my opinion there is not enough novel content in this paper and more work is needed or perhaps the authors could reconsider a new submission of the paper as a letter.

I also think that the writing in section 3 of the paper should be improved to increase the readability of the paper.

Other minor comments:

I would advise the authors to also point out in Figure 2 the components that make up the frequency shifter, and those that are used to generate a comb.

There is no description of figure 5 in the article. The authors should comment on the spectra shown.

Although the results of this article are not intended to generate combs, I find figure 5 intriguing. On the top left is the optical comb generated with the DD-MZM and on the top right is the optical comb generated with the SD-MZM. In terms of span the two combs look similar, however I think the optical dual-comb generated with the SD-MZM seems better in terms of flatness. Could the authors explain why using the non-optimal configuration for the frequency shifter (FS), (i.e. operating the MZM in SD-MZM approach), the optical comb obtained is better in terms of flatness than the one obtained by operating the MZM as a DD-MZM?

The authors state that the simulated values have been obtained using official PDKs of an InP-based foundry, but do not specify which one. Given that the parameters and manufacturing tolerances of the different components vary from one foundry to another, I think it would be interesting for readers to know which PDKs have been tested and which foundry should be chosen to obtain the performances described in the article.

The authors claim that the FS can produce a frequency shift of kHz or GHz. Is the performance obtained for the FS the same in both cases or are the results frequency-dependent? They should comment on this and perhaps add some extra simulations as they only show one case.

Are the equations 1 and 2 correct? If k is the interferometric splitting ratio and 0.5 is the ideal case, shouldn't it be 1-k and k instead of 1+k and 1-k?

The operating points (voltages, currents) of the different PMs, SOAs etc. to obtain the performances shown in the article are missing. They should be included.

 

Author Response

Please see the attachment. We made major changes on the paper. We hope our answers meet your expectations.

Author Response File: Author Response.pdf

Reviewer 2 Report

In the manuscript titled ‘Photonic integrated frequency shifter based on double side band modulation: Performance Analysis’, Pérez et al present a performance analysis based on simulation of dual-comb generation using SD-MZM and DD-MZM. The simulation results are convincing and may pave the way for the establishment of dual comb laser source in experiment. However the quality of the presentation of the manuscript (including writing, figure, etc) required to be improved to make their work better understanding.

Comments:

1. I want to make sure that, are all the results (Fig. 3. Fig. 4 and Fig. 5) from simulation? If so, the authors should better add marks to point this out. For example, “XXX result is obtained by XXX simulation using XXX software.” In the current version, it is so unclear.
2. I suggest adding several notations of the RF driven frequencies (MZMs, AMZI and PMs’ driven frequencies) in Fig. 2 for better illustration.
3. I suggest labelling the subfigures by (a), (b), (c) instead of ‘upper left’, ‘upper right’, etc.
4. The authors should double check the terms of abbreviations. Make them consistent. For example, sometimes it is single driven MZM, sometime it is single-driven MZM.
5. I suggest the authors to label ‘DD-MZM’ and ‘SD-MZM’ in Fig. 3, Fig. 4 and Fig. 5 to the corresponding results.
6. What is the spectral response of the AMZI filter?
7. The title of one subfigure in Fig. 4 is ‘spectral content’, please revise.
8. In Line 194, ‘RIN, Shot and spontaneous emission noises were enabled.’ What does this mean?
9. One of the subfigures (upper right) in Fig. 4 seems to have different resolution compared with other figures. The unit of x-axis is THz, please make it consistent with other subfigures.
10. The marks in Fig. 4 are confusing. In the manuscript, it is said that ‘For the DD-MZM approach, the SMSR reached was 23 dB 184 and the CS was 41 dB.’ But 41-dB mark is in the bottom right subfigure. Again, please make it clear which result is from DD-MZM and which is from SD-MZM. Also, please make the definition of SMSR and CS clear in the figure.
11. In Fig. 5, why the optical spectra look different from DD-MZM and SD-MZM? Could the authors give some comments here?

Author Response

Please see the attachment. We made major changes. We hope the answers meet your expectations.

Author Response File: Author Response.pdf

Reviewer 3 Report

The authors present a design and performance analysis of an integrated frequency shifter based on Double Side Band modulation with Carrier Suppression (DSB-SC) intended to serve in a dual-comb system. They use official Process Design Kits (PDKs) of an Indium Phosphide based foundry, therefore the system is intended to be implemented in InP. Integrated frequency shifters are very useful in many applications in which some kind of heterodyne conversion is required, such as Dual-Comb architectures (Dual-Comb Spectroscopy or Dual-Comb Ranging) or coherent optical communications. Previous articles report integrated frequency shifters using Mach-Zehnder Interferometers or Modulators (MZMs) operated in such a way that perform Single Side Band modulation with Carrier Suppression (SSB-SC). As the authors report in this article, this approach is difficult to implement and demands a very precise control on the characteristics of each Phase Modulator (PM) and optical amplifier. To overcome this drawback, the authors propose an architecture based in DSB-SC for implementing a frequency shifter in which no such precise control is required. This architecture was proposed by the authors in [ref] using a Single-Drive MZM (SD-MZM). The article analysed here is intended to compare the performance of the system proposed in [1], with the same system when a Dual-Drive MZM (DD-MZM) is used. Nevertheless, some information should be added, and some points should be clarified before publication.  I consider that the manuscript requires major revisions before being published and in my opinion, the results are inconclusive. Below, the authors can find a summary of the main concerns that I have. As a general comment, from my point of view, I find the wording of the article a bit confusing and not very clear. At some points, it is difficult to follow the author’s dissertation. In addition, in my opinion, the manuscript could be improved in terms of writing.

1. In the introduction, the authors make a good rationale for their design based on DSB-SC arguing that this would be easier to implement. They do a summary of the State-of-the-Art, focusing on the drawbacks of the other integrated approaches. Nevertheless, the article lacks a more detailed account of the integration of the proposed system in InP. Moreover, it would be positive to mention some characteristics about the foundry and the used PDK, which would be helpful to understand the integration of this system in InP. In addition, the integration of Phase Modulators (PMs) in InP comes with some limitations that are not studied. Only reverse voltage can be applied to the Electro-Optic Modulators Phase Modulators (EOPMs) and the maximum ratings should be considered.
2. Regarding the introduction, I would have explained what an Optical Frequency Comb (OFC) is and the dual-comb technique. As long as this article is intended to analyse the performance of the proposed architecture in a dual-comb system, it would be helpful a short explanation about this technique. Likewise, a short explanation about OFC generation using PMs would be good. Even taking into account that this article is aimed at specialized readers, this would contribute to a better understanding of the manuscript.
3. In my opinion, the presentation of the analytical model in Section 2 could be improved. I would suggest starting with a description of an EOPM, its equations, and its integration on InP, also defining important parameters such as the voltage at which a phase shift of 180º is achieved, i.e. V_π. Regarding V_π, I would suggest to make the analysis of its dependence with the length of the EOPM in this point. Next, I would suggest presenting the different architectures of the MZMs and the development of the equations based on the equations of a single EOPM.
4. In Section 2, the definition of Single Drive and Dual Drive MZMs is not correct. According to the literature and the manufacturers, a SD-MZM consist of two parallel PMs which share a common ground and are biased with opposite voltages. A DD-MZM consist of two parallel PMs, which are biased with independent voltages. Taking into account that integrated EOPMs could only be biased with reverse voltages, implementation of push-pull operation is not straightforward. Using these definitions, the presented discussion about the mitigation of the chirp effect would be inaccurate since this is associated with push-pull operation and not with the MZMs being Single-Drive or Dual-Drive (A SD-MZM also can be operated in push-push, where there is no mitigation of the chirp effect).
5. In my opinion, the manuscript lacks a detailed development and justification of the equations in Section 2. These equations are presented without previous analysis and without a reference containing their development. The detailed model development would be helpful to better understand the proposed analytical model and the discussion about the needed value of V_RF in each architecture (Single-Drive and Dual-Drive), which is not clear and straightforward.
6. In Section 2, the authors sketch the difference between a symmetric MZI and Asymmetric MZI (AMZI). In my opinion, a more detailed explanation about how an AMZI works and its functionality is needed, since AMZIs are used in the dual-comb system to filter one of the side bands resulting from the DSB-SC and play an important role in the frequency shifter system.
7. In Section 3, the authors present the obtained results. In my opinion, the presentation of the results could be improved. The point in where the results are measured should be presented more clearly. In addition, in the graphs where there are more than one result, introducing a legend will be helpful. Furthermore, the axes of some graphs could be improved: In Figure 3, upper left is not specified which one of the voltages of the two EOPMs is swept; in Figure 4, the figures at the centre would be clearer if the frequency is expressed in MHz; in Figure 5 the figures do not have the same scale, and the frequency could be expressed in MHz in the bottom figures. Moreover, the caption of the figures could be improved in writing.
8. In Section 3, in my opinion, the explanation of how the Side Mode Supression Ratio (SMSR) and the Carrier Supression (CS) are measured should be improved. Furthermore, in line 167, the authors conclude that a low V_π is used, but it would be better to quantify this value.
9. In Figure 4, the electrical spectrum of the SD-MZM presents four harmonics at 7, 8, 9 and 10 MHZ. Could the authors explain the appearance of these harmonics? Which is the effect caused by them in the system? Why in the case of DD-MZM the spectrum does not have these harmonics? How is the SMSR and the CS measured in this case?
10. In my opinion, when presenting the results of the dual-comb system, it would be interesting to characterize the generated OFCs (spectral width or comb tones within a certain bandwidth, Carrier-to-Noise Ratio…) and present the two combs alone. This would clarify a lot the obtained results in the electrical domain. Moreover, the results would be clearer if the authors were to relate the spectral widths in optical and electrical domains through the compression factor, since the electrical comb seems to be narrower (has less tones) than the optical one. That is why I find some issues that should be clarified: How are the OFCs combined with each other at the output of the architecture depicted in Figure 2? Why the optical dual-comb for a DD-MZM architecture is different from the SD-MZM? Why the combs in the electrical domain have fewer tones than in the optical domain?

 

[1] Betancur-Pérez, A., Martín-Mateos, P., de Dios, C., & Acedo, P. (2020). Design of a multipurpose photonic chip architecture for thz dual-comb spectrometers. Sensors (Switzerland), 20(21), 1–14. https://doi.org/10.3390/s20216089

Author Response

Please see the attachment. We made major changes. We hope the answers meet your expectations.

Author Response File: Author Response.pdf

Round 2

Reviewer 1 Report

The authors have replied to the comments and implemented changes that substantially improve the paper. Although I believe that the amount of new content compared to their previous work is not remarkable, I believe that with the improvements made it can be accepted for publication in Photonics.

There are two small typos, one in the caption of figure 3, "optcial filter", another on line 406, I believe it is Fig. 5 e instead of Fig. 4 e.

Reviewer 3 Report

I consider the manuscript for its publication