Design of a Power Splitter Based on a 3D MMI Coupler at the Fibre-Tip
, Martin Feiler 1, Tomas Mizera 2, Anton Kuzma 1,*, Dusan Pudis 2 and Frantisek Uherek 1
Round 1
Reviewer 1 Report
This paper presents a design of two 3D IP-dip polymer-based MMI power splitters operating in the near-infrared part of the spectrum at a wave-length of 1550 nm. There are some issues that should be clarified before it is being considered to be accepted for publication in electronics.
1. What is the theory of the design of the MMI power splitters? It is able to design
1×4 splitter and 1×9 splitter, if it is possible to design any other 1×n splitters using this method?
2. In Figure 1, one with direct fibre attachment and one using a tapered input waveguide, there is no discussion on the difference of power distribution structures between the two input waveguides. And what are the the dimensions of the tapered fiber of the input waveguide?
3. On page 8, it is said :”In a 1×4 splitter, the lowest out-put crosses the half-power mark (-3 dB) at an offset value of 1.7 μm. In the case of a 1×9 splitter, this occurs at 2.4 μm.” What is the meaning of it in Figure 11?
Author Response
- What is the theory of the design of the MMI power splitters? It is able to design
1×4 splitter and 1×9 splitter, if it is possible to design any other 1×n splitters using this method?
The design approach for any MMI power splitter is to determine the width (WMMI) and the length (LMMI) of the coupler based on the dimensions and character of the input mode. WMMI has to be high enough to encompass the chosen number of outputs (N) without crosstalk between the output waveguide. LMMI can then be determined analytically by using Equation 1 in the article or by numerical simulations such as Beam Propagation Method (BPM). In the case of our 3D power splitters, this approach is much the same, the width is determined in regards to a single axis of propagation and due to the symmetricity of the 3D structure, the splitting is then present in the other axis (2 in case of the 1x4 splitter, 3 in the case of 1x9 splitter). In other words, the same WMMI and LMMI could be used for a planar MMI power splitter resulting in a symmetrical 1x2 or 1x3 split.
The same method is applicable to any other 1xN splitter, however, with increasing N there is an increase in insertion losses. Also one has to take the dimension limitations of the fabrication technique into account with a higher number of outputs.
- In Figure 1, one with direct fibre attachment and one using a tapered input waveguide, there is no discussion on the difference of power distribution structures between the two input waveguides. And what are the the dimensions of the tapered fiber of the input waveguide?
The difference in the tapered input waveguide profile does not have a significant impact on the MMI coupler function, thus the discussion has been omitted to present the simulations of the coupler itself more clearly. The different profiles of the input waveguides change the input mode field which is then replicated in the split outputs. The function of the MMI coupler is unchanged regardless of the input used, provided the 9 µm fibre is used as the input.
The dimensions of the tapered input waveguide are 10 µm in width on the fibre side, and 9 µm in width on the MMI side, with a length of 60 µm in both fabricated structures.
The dimensions of the tapered input waveguide have been added to the article.
- On page 8, it is said :”In a 1×4 splitter, the lowest out-put crosses the half-power mark (-3 dB) at an offset value of 1.7 μm. In the case of a 1×9 splitter, this occurs at 2.4 μm.” What is the meaning of it in Figure 11?
Figure 11 has been missing the lines that signify the offset values mentioned. This has been fixed. The half power mark is the threshold where at least one of the output losses crosses -3dB value, compared to the output losses without any offset of the input fibre.
Reviewer 2 Report
Dear Authors,
I found this article interesting, however I advise some revision. These are my suggestions:
1. You have two constructions. One without the input waveguide and second with tapered input waveguide (TIW) . In Fig. 7 is shown the 1x4 MMI with this TIW. In your previous article "3D Polymer-Based 1 × 4 MMI Splitter" the square profile in the middle is recognized as optical splitter part. So where is this TIW and how looks the construction without the TIW?
2. You did the measurements with NSOM technique, however the results are unreadable in this form. It's impossible to compare the results of simulation and measurements and there is no table where such results are combined.
3. There is something wrong with the "sentence" in lines 147-149.
4. Maybe it will be better to label the chapter 3 as "Simulation" and add chapter 4 where you will compare the results of simulation with the measurements and discuss it.
5. Please add the XYZ axis orientation in Fig. 2.
Author Response
- You have two constructions. One without the input waveguide and second with tapered input waveguide (TIW). In Fig. 7 is shown the 1x4 MMI with this TIW. In your previous article "3D Polymer-Based 1 × 4 MMI Splitter" the square profile in the middle is recognized as optical splitter part. So where is this TIW and how looks the construction without the TIW?
The MMI power splitters without TIW have only been used in simulations as the tapered input did not introduce significant changes to the output modal field distributions. However, the TIW is necessary for the MMI splitters as the direct fibre to MMI attachment introduces distortion with unpredictable interference patterns in a real-world scenario. In the simulation, the interface defects are not accounted for.
- You did the measurements with NSOM technique, however the results are unreadable in this form. It's impossible to compare the results of simulation and measurements and there is no table where such results are combined.
Only the number of output modal fields and their shape are directly comparable between the simulation and NSOM measurements. It is not possible to directly measure power in the monitored areas of the MMI couplers by NSOM. Furthermore, the power figures acquired by simulation do not include losses caused by surface imperfections. The NSOM measurements clearly show the same number of maxima and sufficient modal field separation, thus comparison table was not added.
- There is something wrong with the "sentence" in lines 147-149.
The sentence has been revised to be clearer.
- Maybe it will be better to label the chapter 3 as "Simulation" and add chapter 4 where you will compare the results of simulation with the measurements and discuss it.
The only directly comparable metrics between NSOM measurements and simulation are the number of maxima and modal field diameters. Therefore, the proposed chapter 4 would be rather short. By keeping the realization and measurements after the MMI simulations the results we are trying to convey are represented better.
- Please add the XYZ axis orientation in Fig. 2.
The XYZ axis indicators have been added to Figure 2.
