Advancements in CMOS-Compatible Silicon Nitride Optical Modulators via Thin-Film Crystalline or Amorphous Silicon p–n Junctions

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors This paper proposed a way to modulate the passive SiN waveguide by introducing c-Si and a-Si above and beneath the nitride waveguide, and in vertical or lateral configuration, and utilizing Si p-n juction and free-carrier effect. The author did design and simulation and discussed 4 types of modulator configurations and compared their electrical and optical properties. This paper presents interesting results, but can be much stronger if the author can include measurement and characterization of fabricated devices. I suggest the author make major revisions following some of my questions below, before it can be considered to be accept for publish:   1. in the introduction: I suggest the author rephrase this sentence: "Power consumption is higher for the TM modes than for the TE, being 20 in the order of 100mW one and smaller than 23mW the other" for better readability and understanding.   2 line 57, the literature study of different SiN modulators are mostly good, but these examples are all based on electro-optic effect, including the reference 25 in which the PZT is used as a k2 material to realize the Pockels effect. To complete this literature study, there is another emerging technique of modulating SiN, which is the stress-optic and piezoelectric effect. Some good references are: 1. AlN: https://doi.org/10.1038/s41566-021-00903-x; 2: PZT: https://doi.org/10.1364/OE.467721   3. line 87, this paragraph is repetitive and similar as in introduction line 38. The author should consider rephrase or delete part of them.   4. for the optical simulation, it seems the author did not include the SiO2 thin layer between the SiN and Si as discussed in the fabrication details, which I believe should be included in the simulation for accuracy. Also, the author should mention how thin is the SiO2 layer in the fabrication section.   5. Did the author measure the propagation loss of such modulator? This is an important figure of merit when describing SiN waveguide modulator. Also I suggest the author to distinguish the losses by using "insertion loss" when talking about their "loss" results.   6. Which part of the results are from actual measurements? I did not quite understand if there is actual measurement on modulator losses, speed and tuning efficiency. If there are results from actual measurement, they should be described clearly with methods and results. If there are no actual experiements done, which is OK, the author should clearly convey that the paper is based on simulation and calculation.     7. Again, I do not see any experiment results showing this modulator can reach switching times of hundreds of MHz as declared in the abstract, therefore, such claims should be removed.    8. How are the power consumption in the table calculated? the author should include methods to present data. Comments on the Quality of English Language

NA

Author Response

The response can be found in the attached PDF.

Author Response File: Author Response.pdf

Reviewer 2 Report

Comments and Suggestions for Authors

In the manuscript, the authors reported simulation results on SiN-based optical modulators hetero-integrated with Si p-n junctions. The following technical comments should be properly addressed before recommending for publication.

 

1. Regarding novelty, SiN has already been proposed for various integrated photonics applications. Despite the benefits (and also drawbacks) from the SiN material platform, the novelty and advantages of the proposed devices in this manuscript should be further compared and highlighted with previous similar publications (comparing also their core device parameters like modulation depth or efficiency), such as other hetero-integration approaches like Refs. 25-36.

 

2. Related to the above, when mentioning heterogeneous integration approaches using 3D thin films or 2D materials, foundational references (Ref: Nature, 610, 54–60, 2022. Ref: Nature Reviews Materials 8, 498-517, 2023) should be also supplemented.

 

3. The authors need to further comment on their motivations on proposing the device structures showing in Fig. 1. 

As mentioned by the authors, SiN dose not support free carrier dispersion or EO effect. The proposed structures look a decoupling of modulation area (enabled by conventional Si p-n junctions) with the waveguiding area (SiN waveguides). The modulation is thus enabled by optical field coupling.

In this way, it can have certain benefits of reduced loss per μm, but the efficient can be much lower than pure p-n-junction-based Si modulators. A much larger device footprint is needed, and device fabrication can be also more complex.

 

4.  I assume different doping levels are used for the proposed device structure showing in Fig. 1 to mitigate the loss at the region near waveguide core. This info/ or the dope details of the color labels (e.g. dark/light blue or red) should be given.

 

5. For the top a-Si structures in Figs. 1a & 1b, chemical mechanical planarization (CMP) could be typically experimentally required after SiN patterning and SiO2 deposition in order to achieve buried waveguide structure. While in this process the top surface might be not ideal for later PVD process of Si and the actual height of SiN waveguide could be also altered by CMP. Would this impact device performance?

For the doping structures showing in Figs. 4c & 4d, the authors can provide some brief info on how to potentially realize this proposed doping structure, which looks actually not very easy to implement from my side.

 

6. Regarding the concerns on loss for modulators, how would be the proposed devices compared with 2D materials-integrated optical modulators (Ref: Nature Photonics,14, 256–262, 2020. Ref: IEEE Photonics Journal, 10, 6600217, 2018)? They can be more flexible in device implementation with also strong modulation strength despite extremely thin thickness.

 

7. The COMSOL-simulated E field distributions in Fig. 11 is not very informative, as the variation in effective mode index is more important, and the results also highly depend on the actual implementation of actual the p-n doped regions of Si.

Besides, two more quick questions:

(1) Did the authors also considering using this structures with ring resonators for footprint much reduced amplitude modulation?

(2) p-n junction based modulation typically also induce amplitude/absorption change. Will the resultant coupled phase-amplitude modulation impact the desired device functionality? 

Author Response

The response can be found in the attached PDF.

Author Response File: Author Response.pdf

Reviewer 3 Report

Comments and Suggestions for Authors

See in attached Word document

Comments for author File: Comments.docx

Author Response

The response can be found in the attached PDF.

Author Response File: Author Response.pdf

Round 2

Reviewer 1 Report

Comments and Suggestions for Authors

The author has addressed most of my questions successfully, and therefore I can recommend it to be published. 

However, there is some minor errors that needs to fixed:

1. line 66, PZT or piezo-electric materials or modulators actually has extremely low power consumption, on the order of nW. Therefore, the authors should not claim here that the piezoelectrics have power consumption problem. This claim should be moved to the thermal tuning. 

Comments on the Quality of English Language

NA

Author Response

The responses can be found in the attached PDF

Author Response File: Author Response.pdf