By using the film mode matching method, a novel design for asymmetrical multi-section 1.55/1.31 μm wavelength splitter based on multimode interference has been proposed and simulated, which can be effectively applied to wavelength multiplexer, self-biased photodiode, and other optical devices. Compared with the conventional wavelength splitter design, the length of the novel structure has been reduced to at least 1/5, showing better performance. The presented structure is also adequate for splitting other wavelengths and more tolerable fabrications.
© 2012 OSA
Optical wavelength splitters are significant components in modern optical communication, which can increase the number of channels and the information capacity of optical fibers. Moreover, they can be applied to many devices such as demultiplexers [1, 2] and self-biased photodiodes [3, 4] to reduce the dimension greatly and improve their performance significantly.
Several approaches have been proposed to realize the wavelength splitter, such as Y branch devices , Mach-Zehnder interferometers  and multimode interference (MMI) devices . Among them, MMI devices based on the self-imaging  are good candidates due to their excellent advantages, such as ease of fabrication, low excess loss, compact size, and large fabrication tolerances .
However, when utilized as wavelength splitters, MMI based splitters still have the problem in their extensive length  because the length of devices needs to match several odd or even times of beat lengths of both wavelengths. This situation is aggravated in the InP/InGaAsP material system owing to small refractive index difference. In this work, a multi-section MMI 1.55/1.31 μm wavelength splitter concept has been demonstrated and simulated. The results show excellent characteristics and size reduction of at least 80% compared to the length of the common splitter design. In addition, the novel concept is also suitable to design splitters operating in other wavelengths and imply more tolerable fabrications.
2. Principle and device designs
The cross section of the proposed waveguide is composed of multi-layer staggered InGaAsP and InP layers with 1045 nm deep etching depth in the ridge waveguide as shown in Fig. 1(a) . The waveguide structure studied here has been successfully applied to other devices, like 90° hybrid , and high speed photodiode [3, 10, 11]. Three Q1.06 layers on the bottom of the waveguide are utilized as a mode field convertor when the waveguide is integrated with other devices. Detailed description about the function of each layer can be found in . According to the MMI theory , in order to divide two different wavelengths, for example, 1.31 μm and 1.55 μm, the length and width of the MMI should be designed to satisfy:
Figure 1(c) shows simulation results of the field distribution of conventional MMI coupler at 1.55 μm. According to the MMI theory , the optical fields in the MMI waveguide alternately and periodically moves and are redistributed as plotted in Fig. 1(c) due to phase of guided modes changing. Therefore some certain lengths must exist where most of the power distributes in the opposite side to self-imaging point as shown in the “red circle region” in Fig. 2(a) . Therefore one can design a novel wavelength splitter with appropriate width and length to maintain 1.55 μm wavelength focusing to the self-imaging point at cross output while most power of 1.31 μm light is propagating to bar port side.
The proposed asymmetrical multi-section MMI wavelength splitter is schematically shown in Fig. 2(a). From the input side, the first multimode waveguide is named after divider MMI (1) as indicated in Fig. 2(a). It is utilized to divide two lights with 1.55 μm and 1.31 μm wavelengths. The one with 1.55 μm is guided to the cross output, and most power of the one with 1.31 μm is guided to the bar output side. Then a narrower multimode waveguide is named after collection MMI (2). It serves as another MMI coupler and is located at the bar output side to collect efficiently 1.31 μm power. In order to improve the output power, another narrower multimode waveguide which is named after sub-collection MMI (3) follows to former multimode waveguide (2).
3. Results and discussion
The simulation was accomplished by the computer-aid design tool Fimmwave. Fimmwave is based on film matching method and can give an accurate prediction of the performance of passive devices and is verified by our previous works [9–11].
Firstly, 9 μm wide multimode waveguide is chosen to validate both compact dimension and suitable length for splitting two wavelengths. As shown in Fig. 1(c), the first self-imaging point is located between 690um and 720um. Then the length of divider MMI can be obtained exactly as shown in Fig. 3 . 707 μm is the optimal length to obtain maximum optical power at 1.55 μm. As illustrated in Fig. 2(c), only one 5 μm wide collection MMI is implemented to collect 1.31 μm at the first time with the motivation of shortening total device length and leaving 2 μm gap as depicted in Fig. 1(c) to avoid coupling loss. Unfortunately, no matter how long the collection MMI is selected, about 40% of 1.31 μm optical power is lost. A limited improvement can be realized using longer divider MMI (712.5 μm) at expense of increasing insertion loss of 1.55 μm by about 0.4 dB as shown in Fig. 3 and 4 . Because the input power is not concentrated on one single point as shown in Fig. 2(c) and 2(d), the power at the output of collection-MMI can’t form a perfect self-imaging point, leading relative high power loss of 1.31 μm due to mode mismatching between broader collection MMI and narrow output waveguide as shown in the inset of Fig. 4. Thus an additional narrower multimode waveguide, namely, 4 μm wide sub-collection MMI is employed to achieve a better coupling. Actually, the whole component exhibits nearly 10% higher power for 1.31 μm with a sub-collection MMI integrated as depicted in Fig. 3 and Fig. 4.
The performances of the wavelength splitter are characterized by the extinction ratio (ER) and the insertion loss (IL), which are defined asFig. 2(d).
For the conventional wavelength splitter structure, the optimum length is verified to exceed 5000 μm according to the simulation which is not shown here. The insertion loss is larger than 2 dB for both wavelengths when the device length is 5000 μm. So it makes no sense to optimize and calculate the exact length of conventional design. Therefore, the doubtless superiority of the novel design presented is not only its compact dimension whose length has been reduced to at least 1/5, but also lower insertion loss.
In this work, even higher extinction and lower insertion loss can be expected by optimizing the epitaxy layers with the same MMI wavelength splitter concept presented, we still follow the epitaxy layers of our previous products [3, 9–11], because those layers can be compatible easily with other components like photodiodes and the performance have been proved to be excellent . When the components are applied to self-biased photodetectors, optical power at 1.31 μm usually functions as power source to solar-like photodiodes. Therefore the responsivity of the self-biased photodiode can be improved since all of 1.55 μm signal power feeds to signal photodiode without consuming further power in solar-like photodiodes in this case. In addition, a compact dimension of MMI demultiplexer structures can be expected compared to the conventional structure .
The fabrication tolerance of collection MMI width has been calculated as shown in Fig. 5 . With a requirement of 1dB, the fabrication tolerance of the collection MMI width is about ± 0.5 μm. The tolerances of the other sections widths are similar, which satisfy the common fabrication requirement . Therefore the novel structure can be fabricated easily by just one step photolithography following the normal process. Besides, the wavelength splitters for other wavelengths can be developed conveniently by the same method, which shows the suitability and compatibility.
This paper proposed the novel ultra compact asymmetrical multi-section MMI wavelength splitter. In order to overcome the huge size and high insertion losses of the conventional MMI splitter, film mode matching method was utilized. Simulation results showed the insertion losses of −0.62dB and −1.46dB at 1.55 μm and 1.31μm, respectively. The size of the proposed splitter was 963.5μm, which was regarded as 80% reduction in size compared to conventional MMI splitter. It should be noted that this design methodology also applies to other wavelength MMI structures. We believe that the proposed structure is a remarkably attractive one with respect to low cost and ultra compact optical devices, such as wavelength demultiplexers, self biased photodiode, and so on.
The authors gratefully thank Prof. Jian Wu from Beijing University of Posts and Telecommunications for his continuous encouragement and help. This work was financed partly by the 100GET program of the German Federal Ministry of Education and Research, the EC’s 100GET/CELTIC project, and EC’s project MIRTHE within the FP7 program.
References and links
1. Y. Shi, S. Anand, and S. He, “A polarization-insensitive 1310/1550-nm demultiplexer based on sandwiched multimode interference waveguides,” IEEE Photon. Technol. Lett. 19(22), 1789–1791 (2007). [CrossRef]
2. W.-Y. Chiu, J.-W. Shi, Y.-S. Wu, F.-H. Huang, W. Lin, and Y. –J. Chan, “The Monolithic Integration of a wavelength-demultiplexer with evanescently coupled uni-traveling-carrier photodiodes,” IEEE Photon. Technol. Lett. 18(11), 1246–1248 (2006).
3. H.-G. Bach, R. Kunkel, G. G. Mekonnen, D. Pech, T. Rosin, D. Schmidt, T. Gaertner, and R. Zhang, “Integration potential of waveguide-integrated photodiodes: self-powered photodetectors and sub-THz pin-Antennas,” in Optical Fiber Communication Conference (OFC, 2008), OMK1.
4. H.-G. Bach, “Monolithically integrated optoelectronic subassembly,” US Patent 2009/0202197 A1.
5. N. Goto and G. L. Yip, “Y-branch wavelength multi-demultiplexer for λ=1.30 μm and 1.55 μm,” Electron. Lett. 26(2), 102–103 (1990). [CrossRef]
6. A. Tervonen, P. Poyhonen, S. Honkanen, and M. Tahkokorpi, “A guided-wave Mach-Zehnder interferometer structure for wavelength multiplexing,” IEEE Photon. Technol. Lett. 3(6), 516–518 (1991). [CrossRef]
8. L. B. Soldano and E. C. M. Pennings, “Optical multi-mode interference devices based on self-imaging: Principles and applications,” J. Lightwave Technol. 13(4), 615–627 (1995). [CrossRef]
9. R.-Y. Zhang, K. Janiak, H.-G. Bach, R. Kunkel, A. Seeger, S. Schubert, and M. Schell, “Performance of InP-based 90 °-hybrids QPSK receivers within C-Band,” in International Conference on Indium Phosphide and Related Materials (IPRM, 2011).
10. A. Beling, J. C. Campbell, H.-G. Bach, G. G. Mekonnen, and D. Schmidt, “Parallel-fed traveling wave photodetector for >100-GHz Applications,” J. Lightwave Technol. 26(1), 16–20 (2008). [CrossRef]
11. A. Beling, H.-G. Bach, G. G. Mekonnen, R. Kunkel, and D. Schmidt, “High-speed miniaturized photodiode and parallel-fed traveling-wave photodetectors based on InP,” IEEE J. Sel. Top. Quantum Electron. 13(1), 15–21 (2007). [CrossRef]
12. R. Kunkel, H.-G. Bach, D. Hoffmann, G. G. Mekonnen, R. Zhang, D. Schmidt, and M. Schell, “Athermal InP-based 90°-hybrid Rx OEICs with pin-PDs >60 GHz for coherent DP-QPSK photoreceivers,” in International Conference on Indium Phosphide and Related Materials (IPRM, 2010).