Directly Modulated Tunable Single-Mode Lasers Based on a Coupled Microcavity
, You-Zeng Hao 2,3, Zhen-Ning Zhang 2,3, Bin Zhang 4, Yue-De Yang 2,3, Jin-Long Xiao 2,3,*, António L. Teixeira 1 and Yong-Zhen Huang 2,3
Round 1
Reviewer 1 Report
The article is devoted to a topical issue related to the development of simple designs of single-frequency semiconductor lasers. The advantage of the work is the presence of both theoretical studies and experimental characteristics of the developed lasers. However, the study is full of inaccuracies, and there are no comments important to understand well the advantages of the proposed approach.
For this reason, the article should be supplemented and corrected in accordance with the comments below.
The authors talk about different intermode intervals separately for the FP part and SRM when they describe the principle of tuning over a wide range of wavelengths (lines 76-81). Can the authors give the values of the intermode intervals calculated separately for the FP cavity and SRM?
“By solving the eigenmodes with complex eigenfrequency, we can evaluate the reflectivity, mode Q factors, and field distributions, respectively” (line 102-103). This is especially important for understanding the simulation results in Section 3.1.
Can the authors clarify what "the reflectivity" they mean and how it is calculated, and it should also be clarified whether the mode Q-factors were calculated for the entire HSRRL or only for a part of it, if the latter, how it is calculated?
In lines 105-106, the authors talk about the study of the mode coupling between the fundamental FP transverse modes and the WGMs. Can the authors quantify the mode coupling parameter? For example, the theoretical and experimental values of the calculated parameters of the optical coupling between waveguides are commonly used and are usually given in the manuscripts [S. O. Slipchenko et al., “Stable Lateral Far Field of Highly Dense Arrays of Uncoupled Narrow Stripe Ridge Waveguide 1060 nm Lasers,” J. Light. Technol., vol. 40, no. 9, pp. 2933–2938, May 2022. and J. T. Young, C. Wei, C. R. Menyuk, and J. Hu, “Mode coupling at avoided crossings in slab waveguides with comparison to optical fibers: tutorial,” J. Opt. Soc. Am. B, vol. 38, no. 12, p. F104, Dec. 2021.]. Can the authors disclose (describe in detail) the study of the mode coupling between the fundamental FP transverse modes and WGMs basing on the values of optical coupling parameters?
The authors claim that the HSRRLs is based on an AlGaInAs/InP (line 109), but there is no Al-containing layer in the description. In general, the description of the structure is given carelessly. Can the authors give a description with thicknesses, including n- and p-claddings, waveguides? This will allow us to speak about the typicality of the design used, and also disallow questions on the design, in particular the etching depth.
The description of the design and the sequence of operations is best accompanied by a picture showing the areas closed by SiO2, SiNx, BCB, an insulating trench between the sections, and electrodes for the sections pumping.
Since SiNx was used to limit the microresonator region (SRM), was this taken into account in the calculations?
In lines 132-134, the authors argue that the proposed design has the best quality factor for wavelengths 1520.8 1537.6 1555.2 1572.7 nm, why other equally intense reflection peaks are not taken into account, why the narrowness of the peak is associated with a high quality factor?
May the authors give calculated reflection values for selected modes with high reflection value and high Q-factors? How can the authors explain the mismatch between high Q-factors and high reflectance for some wavelengths (fig.2a)?
Since the measurements were carried out from the fiber, can the authors comment on the coupling efficiency of radiation into the optical fiber, as well as the beam divergence at the exit from the FP section.
It is customary to write “the current-voltage curve” instead of “the change of the voltage” (line 162-163).
Better, replace “contacting photolithography” with “contact photolithography”
“Semiconductor cooler (TEC)” replace with “thermoelectric cooler”
The article considers a new laser design based on coupled cavities, while it is not clear how the authors came at the claimed design: FP cavity with a width d, a length L, and an SRM with a side length a and a deformation amplitude δ. Comments should be given on how these parameters were chosen, why they are optimal?
The authors also consider the transition from square microcavity to square/rhombus microcavity (SRM) as a novelty of the work, which is determined by the deformation amplitude δ parameter. Can the authors explain how this parameter was chosen?
In the manuscript, the same cavity is called differently: whispering-gallery (WG) mode microcavity and the square/rhombus microcavity (SRM). Authors can choose one option and use it throughout the manuscript, for example, square/rhombus microcavity (SRM).
Author Response
The article is devoted to a topical issue related to the development of simple designs of single-frequency semiconductor lasers. The advantage of the work is the presence of both theoretical studies and experimental characteristics of the developed lasers. However, the study is full of inaccuracies, and there are no comments important to understand well the advantages of the proposed approach.
For this reason, the article should be supplemented and corrected in accordance with the comments below.
Q: The authors talk about different intermode intervals separately for the FP part and SRM when they describe the principle of tuning over a wide range of wavelengths (lines 76-81). Can the authors give the values of the intermode intervals calculated separately for the FP cavity and SRM?
A: The FSR is 17.2 nm for an SRM with a =15 μm, δ= 0.4 μm and d= 1.5 μm, and 1.25 nm for a 300-μm-length FP cavity, with an effective refractive index of 3.2 in simulation.
Q: “By solving the eigenmodes with complex eigenfrequency, we can evaluate the reflectivity, mode Q factors, and field distributions, respectively” (line 102-103). This is especially important for understanding the simulation results in Section 3.1. Can the authors clarify what "the reflectivity" they mean and how it is calculated, and it should also be clarified whether the mode Q-factors were calculated for the entire HSRRL or only for a part of it, if the latter, how it is calculated?
A: The reflection spectrum is calculated by the following procedure. First, set the Scattering Boundary Condition containing the plan wave on the boundary of the waveguide, set up a Domain Probe in a section of the waveguide, record “ewfd.nPoav” ("Power outflow, time average"), and calculate the ewfd.nPoav connected to the PML and the WGM, denoted as P1 and P2, respectively, then the reflectivity R=(P1-P2)/P1.
The Q factors were calculated for the entire HSRRL, which were calculated by Q=freal/fimag/2, where the freal and the fimag is the real part and imaginary part of the complex eigenfrequency, respectively.
Q: In lines 105-106, the authors talk about the study of the mode coupling between the fundamental FP transverse modes and the WGMs. Can the authors quantify the mode coupling parameter? For example, the theoretical and experimental values of the calculated parameters of the optical coupling between waveguides are commonly used and are usually given in the manuscripts [S. O. Slipchenko et al., “Stable Lateral Far Field of Highly Dense Arrays of Uncoupled Narrow Stripe Ridge Waveguide 1060 nm Lasers,” J. Light. Technol., vol. 40, no. 9, pp. 2933–2938, May 2022. and J. T. Young, C. Wei, C. R. Menyuk, and J. Hu, “Mode coupling at avoided crossings in slab waveguides with comparison to optical fibers: tutorial,” J. Opt. Soc. Am. B, vol. 38, no. 12, p. F104, Dec. 2021.]. Can the authors disclose (describe in detail) the study of the mode coupling between the fundamental FP transverse modes and WGMs basing on the values of optical coupling parameters?
A: The mode coupling between the FP transverse modes and the WGMs has been studied in ref. 21, therefore it will not be discussed in this manuscript.
Q: The authors claim that the HSRRLs is based on an AlGaInAs/InP (line 109), but there is no Al-containing layer in the description. In general, the description of the structure is given carelessly. Can the authors give a description with thicknesses, including n- and p-claddings, waveguides? This will allow us to speak about the typicality of the design used, and also disallow questions on the design, in particular the etching depth.
A: We have added the following description ''The active region of the laser wafer consists of six compressively strained quantum wells, with 6-nm-thick Al0.24GaIn0.71As quantum wells and 9-nm-thick Al0.44GaIn0.49As barrier layers, confined by the lower cladding layers of 100-nm undoped graded AlGaInAs and 140-nm N-AlInAs, and upper cladding layers of 150-nm undoped graded AlGaInAs and InAlAs. The upper confinement layers are 1.6 μm InP and P-InGaAs contacting layer. The N-doped density is about 1 × 1018 cm-3, and the P-doped densities increase from 5 × 1017 cm-3 to larger than 1 × 1019 cm-3 in the P-contacting layer." in the manuscript.
Q: The description of the design and the sequence of operations is best accompanied by a picture showing the areas closed by SiO2, SiNx, BCB, an insulating trench between the sections, and electrodes for the sections pumping.
A: Thank you for your suggestion, we have added Figure 1(b) and the associated description to the manuscript.
Q: Since SiNx was used to limit the microresonator region (SRM), was this taken into account in the calculations?
A: Yes, the SiNx film is considered in the simulation, and We have redrawn Figure 1(a), adding SiNx to the diagram.
Q: In lines 132-134, the authors argue that the proposed design has the best quality factor for wavelengths 1520.8 1537.6 1555.2 1572.7 nm, why other equally intense reflection peaks are not taken into account, why the narrowness of the peak is associated with a high quality factor?
A: Due to the overlapping and shared carriers spatially distributed in the cavity, there is a certain competitive relationship among the modes, and the high Q modes have the competitive advantage. So, the final lasing mode of the laser is the mode with high Q factor and high reflectivity. The ratio of the center frequency of the reflection peak to the full width at half maximum represents the Q factor of this mode, so a narrow reflection has a high Q factor.
Q: May the authors give calculated reflection values for selected modes with high reflection value and high Q-factors? How can the authors explain the mismatch between high Q-factors and high reflectance for some wavelengths (fig.2a)?
A: The formation of the coupling cavity first requires that the SRM has a sufficiently high reflectivity, and secondly, among these modes with close reflectivity, the mode with high Q factor is easier to lasing, so there are mismatches between high Q-factors and high reflectance for some wavelengths.
Q: Since the measurements were carried out from the fiber, can the authors comment on the coupling efficiency of radiation into the optical fiber, as well as the beam divergence at the exit from the FP section.
A: The coupling efficiency and beam divergence of the laser have been studied in ref. 21 and ref. 24 of the manuscript. An average laser to single mode fiber coupling loss is about 3.2 dB. The far-field profiles are measured with the full width at half-maximum of 44° and 43° in the in-plane and out-of-plane directions for an HSRRL. The near-symmetric field patterns result in a high-coupling efficiency between the laser and the fiber. We add relevant descriptions and the ref. 21 in the manuscript.
Q: It is customary to write “the current-voltage curve” instead of “the change of the voltage” (line 162-163).
Better, replace “contacting photolithography” with “contact photolithography”
“Semiconductor cooler (TEC)” replace with “thermoelectric cooler”
A: Thank you for your advice, and we have corrected them in the manuscript.
Q: The article considers a new laser design based on coupled cavities, while it is not clear how the authors came at the claimed design: FP cavity with a width d, a length L, and an SRM with a side length a and a deformation amplitude δ. Comments should be given on how these parameters were chosen, why they are optimal?
The authors also consider the transition from square microcavity to square/rhombus microcavity (SRM) as a novelty of the work, which is determined by the deformation amplitude δ parameter. Can the authors explain how this parameter was chosen?
A: The optimal design of cavity size parameters has been studied in ref. 24 of the manuscript, therefore it will not be discussed in this manuscript. "which are designed considering the single-mode lasing characteristics and Q factor [24]." is added in the manuscript.
Q: In the manuscript, the same cavity is called differently: whispering-gallery (WG) mode microcavity and the square/rhombus microcavity (SRM). Authors can choose one option and use it throughout the manuscript, for example, square/rhombus microcavity (SRM).
A: Thank you for your advice, and we unify these names in the manuscript.
Reviewer 2 Report
The work of this paper provides a simple fabrication method for a multi-wavelength modulation system with a hybrid microcavity and FP cavity laser. It is possible to directly modulate at 25 Gb/s with promising performance. In the process of reading this article, I have several questions, and I hope the authors can answer them.
Q1: Why make a half square microcavity and a half rhombus structure? Does this structure have any improvement over the square or rhombus microcavity structure?
Q2: How does temperature affect device performance?
Author Response
The work of this paper provides a simple fabrication method for a multi-wavelength modulation system with a hybrid microcavity and FP cavity laser. It is possible to directly modulate at 25 Gb/s with promising performance. In the process of reading this article, I have several questions, and I hope the authors can answer them.
Q1: Why make a half square microcavity and a half rhombus structure? Does this structure have any improvement over the square or rhombus microcavity structure?
A: We have rewritten the sentence "Furthermore, a deformed HSRL with near-fundamental transverse mode distribution and improved far-field profiles in the FP cavity also shows good direct modulation characteristics and wavelength tunability [24] " in the manuscript.
Q2: How does temperature affect device performance?
A: Temperature is very important for the proper operation of the laser. The increase of temperature will lead to a series of performance deteriorations such as the increase of laser threshold and the decrease of output power. In the actual test, we use TEC to control the temperature of the test bench to always be 290 K. Regarding the performance comparison of the laser at different temperatures, it is not covered in this manuscript.
Reviewer 3 Report
In the manuscript, the authors continue the research of a hybrid square/rhombus-rectangular laser and demonstrate its intrinsic wavelength tuning range, a side-mode suppression ratio and a linewidth. But in order to appreciate the work highly, one need to know the answers to the following questions and comments.
Please, describe the relationship between the length of the Fabry-Perot section and the size of the rhombus. How the Q-factor spectrum will be affected?
Is it possible to specify the WG size of the mode (line 77)?
Is it possible to justify the statement that there is a single WG mode within a gain spectrum?
Please, specify the thickness of the waveguide.
Line 119: please, specify the resistance between a whispering-gallery mode microcavity and a FP microcavity.
Line 120: The “P-InGaAsP ohmic contact layer” is mentioned, but the structure has only p-type cladding InP layer (line 112) and a P+-InGaAs contact layer (line 113).
What are the manufacturing tolerances when obtaining the characteristics of the device shown in Figure 2?
Line 136: How it is clear about the better coupling between the higher-order modes in the SRM with the modes in the FP cavity from the mode intensity profiles of Hz at 1537.6 nm (Figure 2)?
Line 142: Maybe instead of "is consistent with" it is meant "similar to"?
Figure 2(a): What is the frequency or wavelength step in the calculation?
Line 163: What is the purpose of the matching resistor of 35 Ω in the circuit?
Figure 3(b): Please, comment on the longitudinal modes, intermodal interval. What effective resonator length does it correspond to?
Line 216: How the 55.5 GHz bandwidth was determined?
Line 242: What is SMF?
Line 246: Please, explain what “turn-off delay” means.
Lines 250-254: Sorry, the unreadable phrase. What “generally use electro-absorption modulated laser (EML) or add modulation”? What does this refer to?
Line 264: Replace “over” with “is over”.
Line 240: Rephrase “range, which corresponding to some specific currents has a large jitter amplitude,”
Questions on Comsol simulations:
It is not clear from Figures 1 and 2 and the text:
- was the reflection at the boundary with air taken into account?
- was the air taken into account in principle?
- were PMLs specified near the FP and SRM waveguides?
What do the authors mean by “eigenmodes with complex eigenfrequency”? The problem authors describe definitely boils down to the calculation of a single port resonator in the frequency domain. Is the physics that was used is “electromagnetic waves, frequency domain”?
What Port properties were used? Numeric // wave excitation on // Activated slit condition on interior port // Domain-backed?
Perfect Magnetic Conductor is a symmetry condition?
What Study steps were used? Parametric sweep of a wavelength/frequency parameter in the range shown in Fig.2(a) // Boundary mode analysis at the Port // Frequency Domain/Wavelength Domain?
Was the mesh physics-controlled?
What properties of the physics interface were used? Electric field components were solved for out-of-plane vector, in-plane vector or three component vector? Formulation, full field or scattered field?
Please describe how the reflection and the Q-factor are calculated?
Author Response
Please see the attachment.
Author Response File: 
Author Response.pdf
Round 2
Reviewer 3 Report
After the answers received (“the Scattering Boundary Condition containing the plane wave on the boundary of the waveguide”), there is doubt about the correctness of the simulation model and, accordingly, the correctness of the results shown in Figure 2.
1) When calculating eigenfrequencies (Eigenfrequency study step in Comsol), the Comsol fundamentally does not take into account radiation sources. 2) The Scattering Boundary Condition (SBC) is fundamentally not set on the inner boundary of the model, because in fact, from a mathematical point of view, this condition is analogous to impedance. If with the help of SBC the authors tried to set the source of radiation, then it was not taken into account anyway.
Nevertheless, it is possible to calculate the quality factor of such a system through the study of eigenfrequencies by finding eigenfrequencies in the wavelength range one interested in using Region method with a frequency range and manually analyzing and sorting the resulting modes.
In summary, I propose either to recalculate the results, or to omit the section on calculations in Comsol.
My question “Is it possible to specify the WG size of the mode (line 77)?” does not refer to FSR, but to “short optical size”, which is mentioned in line 77.
In the manuscript, the authors still do not provide the thickness and material of the n-waveguide indicated in the answer (2 μm).
Can it be claimed that the difference between the device size and the design value must be less than 1% to ensure the device performance presented in the manuscript.
Why not to present the following in the manuscript: “The FSR of FP cavity is 0.89 nm in figure 3(b), which correspond to an effective FP cavity length of 375 nm, with group refractive index of 3.6. Due to the extra optical path provided by the SRM, the effective FP cavity length is longer than the actual FP cavity.”?
Author Response
Q: After the answers received (“the Scattering Boundary Condition containing the plane wave on the boundary of the waveguide”), there is doubt about the correctness of the simulation model and, accordingly, the correctness of the results shown in Figure 2.
1) When calculating eigenfrequencies (Eigenfrequency study step in Comsol), the Comsol fundamentally does not take into account radiation sources. 2) The Scattering Boundary Condition (SBC) is fundamentally not set on the inner boundary of the model, because in fact, from a mathematical point of view, this condition is analogous to impedance. If with the help of SBC the authors tried to set the source of radiation, then it was not taken into account anyway.
Nevertheless, it is possible to calculate the quality factor of such a system through the study of eigenfrequencies by finding eigenfrequencies in the wavelength range one interested in using Region method with a frequency range and manually analyzing and sorting the resulting modes.
In summary, I propose either to recalculate the results, or to omit the section on calculations in Comsol.
A：We agree with your description of Comsol simulation, In the previous description we did not describe clearly the simulation models, we rewrite the parts about simulation models in more detail in the manuscript. The model used for the simulation of the Q factor is shown in Figure 1(a), and as you described, we obtained the complex frequency and mode Q factor by direct calculation by using the Eigenfrequency study step in Comsol. The model used for the simulation of the reflection spectrum is shown in Figure 1(c) and (d), the SBC is set on the outer boundary, and the SBC set on the entire outer boundary is divided into two parts, which is set in the SBC at the end face of the waveguide with incident field: Wave given by H Field, and SBC set to No incident field for other outer boundary settings. The reflection spectrum calculated using this model is similar to the reflection spectra calculated by the FDTD method using Rsoft in Ref. (21), (22) and (24).
"The simulation of the resonator. A two-dimensional (2D) finite element method (FEM) (COMSOL Multiphysics 5.0) is utilized to calculate the mode reflectivity spectra of the microcavity. An HSRRL consists of an FP cavity with a width d, a length L, and an SRM with a side length a and a deformation amplitude δ. Here, the structure parameters are taken to be a=15 μm, δ= 0.4 μm, d= 1.5 μm, and L=300 μm, respectively, which are designed considering the single-mode lasing characteristics and Q factor [24].
The simulated boundary uses Perfectly Matched Layer (PML), and the dark blue and dark gray parts in Figure 1(a) represent the BCB region boundary and the air boundary, respectively. For the use of the PML layer, the influence of the simulation boundary reflection on the simulation results is avoided. In addition, the symmetry boundary condition (Perfect Magnetic Conductor, PMC) is taken along the horizontal center dashed line to investigate the mode coupling between the fundamental FP transverse modes and WGMs. The coupled cavity has an effective refractive index (neff) of 3.2 covered with 200nm silicon nitride with a refractive index of 2, and the surrounding materials are bis-benzocyclobutene (BCB) and air with refractive indices of 1.54 and 1, respectively. The eigenfrequency calculation uses the physics interfaces electromagnetic waves, frequency domain, and add the study “Eigenfrequency”. The calculation yields the complex frequency of the eigenfrequency, where the real part of freal represents the resonant frequency and the imaginary part of fimag represents the loss. The Q factors were calculated by Q=freal/fimag/2 for the entire HSRRL.
The model of reflection spectrum simulation is show in the Fig. 1(c) and (d). Set the Scattering Boundary Condition (SBC) containing the plan wave at the end of the waveguide as the source, and set up a Domain Probe as the monitor in a section of the waveguide, record “ewfd.nPoav” ("Power outflow, time average"), and calculate the ewfd.nPoav connected to the PML (Figure 1(d)) and the WGM (Figure 1(c)), denoted as P1 and P2, respectively, then the reflectivity R=(P1-P2)/P1." is added in the manuscript.
Q: My question “Is it possible to specify the WG size of the mode (line 77)?” does not refer to FSR, but to “short optical size”, which is mentioned in line 77.
A: For a square microcavity with side length a, the effective optical pathlength is 2√2a×neff, where the neff is the effective refractive index. For the SRM, due to the small deformation delta (only 0.4 μm), the effective optical pathlength is close to that of a square microcavity, about 43 μm with neff of 3.2 in simulation.
Q: In the manuscript, the authors still do not provide the thickness and material of the n-waveguide indicated in the answer (2 μm).
A: We have added the material structure of the wafer in the manuscript and added references to aid in the description.
The wafer is to first grow 500nm N-InP Buffer layer with N-doped density of 1 × 1018 cm-3 on N-InP substrate, and then grow separate-confinement heterostructure (SCH) and QWs, and finally grow the P-doped confinement layer and the ohmic contact layer. The FP cavity is a deeply etched ridge waveguide structure, and the height of the whole ridge waveguide is more than 4 μm, wherein the thickness of QWs and SCH is 0.2 μm, and the upper confinement layers are 1.6 μm P-InP and 0.2 μm P-InGaAs contacting layer，the lower confinement layer is 2 μm N-InP, including 500 nm N-InP Buffer layer and 1.5 μm N-InP substrate.
Q: Can it be claimed that the difference between the device size and the design value must be less than 1% to ensure the device performance presented in the manuscript.
A: The size of the device we actually fabricated differs from the design value by less than 1%, but it does not mean that must be less than 1% to ensure the device performance. So, we do not add relevant descriptions in the manuscript.
Q: Why not to present the following in the manuscript: “The FSR of FP cavity is 0.89 nm in figure 3(b), which correspond to an effective FP cavity length of 375 nm, with group refractive index of 3.6. Due to the extra optical path provided by the SRM, the effective FP cavity length is longer than the actual FP cavity.”?
A: Thank you for your suggestion, we have added the description to the manuscript.
Round 3
Reviewer 3 Report
Line 77: please, replace “size” with “path”
Lines 133-134: please, add “the lower confinement layer is 2 μm N-InP, including 500 nm N-InP Buffer layer and 1.5 μm N-InP substrate” and compare your answer and the text in the manuscript.
Line 112: delete “add”
Line 116: replace “show” with “shown”
Line 117: replace “plan” with “plane”
Please check grammar and English style.
Author Response
Q: Line 77: please, replace “size” with “path”
Lines 133-134: please, add “the lower confinement layer is 2 μm N-InP, including 500 nm N-InP Buffer layer and 1.5 μm N-InP substrate” and compare your answer and the text in the manuscript.
Line 112: delete “add”
Line 116: replace “show” with “shown”
Line 117: replace “plan” with “plane”
A: Thanks very much for your comments. We have revised these typos in the manuscript.
Q: Please check grammar and English style.
A: We have polished this manuscript carefully, please see if the revised version met the English presentation standard.
