An optical device scheme that serves simultaneously as a power combiner for upstream and wavelength demultiplexer for downstream signals is presented. The design concept is validated experimentally by an optical module based on off-the-shelf discrete optical components. An integrated device based on planar lightwave circuit (PLC) is proposed and analyzed in which a multi-mode interference (MMI) device is utilized to separate the upstream 1310 nm signal from the downstream 155x nm signals. The dense WDM function is realized through an arrayed-waveguide-grating (AWG). Design guidelines and optimization procedure for the device are discussed by way of examples.
©2009 Optical Society of America
Growing demand for high-definition television, Internet access and other bandwidth-hungry applications have fueled intense research and development for optical transmission technologies for the end users, among which the fiber-to-the-x (or FTTX, where X denotes the home, curb, cabinet, or building, etc.) has emerged as a leading access technology to overcome the bandwidth bottleneck in the last mile. In a simple power-splitting passive optical network (PON), the downstream optical signal from the optical line terminal (OLT) at the central office is delivered through a single fiber to the remote optical note where the signal is divided among all the subscribers by means of a passive splitter . The total bandwidth is shared by the optical network units (ONUs) at the subscribers’ premises. In the conventional PON architecture, the total power budget and overall bandwidth available to the end users are limited mainly by the splitting loss of the passive power splitter/combiner. To overcome this limitation, WDM PON has been proposed in which the power splitter/combiner is replaced by a wavelength multiplexer/de-multiplexer so that each ONU is assigned to a specific wavelength . The advantages of the WDM PONs are therefore obvious as the optical link power budget is no longer limited by the power splitting loss and each ONU can enjoy dedicated bandwidth which are similar to the case of point-to-point fiber optic links. A standard WDM PON architecture is depicted in Fig. 1(a) in which wavelength specific and/or tunable optical transmitters are required at both OLT and ONU sides. For the latter, this is a very costly option with the existing opteoelctronic technologies. In addition, a dual-band wavelength demultiplexer is required at the OLT, in order to separate the downstream λdn (n=1,2,3…) from the upstream λun (n=1,2,3…) bands. A number of approaches have been investigated to achieve a proper balance between performance and cost of future PON networks [3–7]. In this work, we investigate a hybrid architecture in which the WDM transmission is used only for downstream signals so as to maximize bandwidth delivered to the end users without power splitting loss. A schematic diagram for such hybrid WDM PON architecture is shown in Fig. 1(b). The gain in the downstream power budget will eliminate the need for APD at the ONUs so that less expensive PIN may be used for the receivers. For the upstream transmission, on the other hand, the traditional power combiner is used so that the same low-cost transmitters with 1310nm wavelength can be used to minimize cost. In order to compensate for the power splitting loss, high-sensitivity optical receivers such as APDs can be employed for the OLT. Such a hybrid network may be a good candidate for next-generation PONs to meet the increasing bandwidth requirements with minimum incremental increase in cost of optics.
To implement the hybrid PON architecture in Fig. 1(b), a single passive optical device that can serve as a wavelength demultiplexer for the downstream optical signals and as a power combiner for the upstream channels is required. Schemes to perform dense WDM and power splitting functions have been proposed and demonstrated [8, 9]. In particular, the device described in reference  employing Mach-Zechnder filter for CWDM and star coupler for power splitter could also be used to fulfill the requirement of hybrid WDM PON architecture discussed in this paper. We however utilize the multi-mode interference (MMI) device for the CWDM and the power splitting functions which are more compact and easy to implement. In section 2, we describe the device configuration and working principles of the optical demultiplexer/combiner module. The operation and performance of the device is demonstrated and validated by experimental results for an optical module based on off-the-shelf optical components with upstream signals at 1310nm and four DWDM channels near 1550nm. A fully monolithically integrated demultiplexer/combiner device based on planar lightwave circuits (PLCs) is proposed and discussed in detail in Section 3. The design procedure and simulation results for the individual components as well as the integrated circuit of the entire optical module are demonstrated and discussed. We close this paper by presenting our conclusion in section 4.
2. Device configuration, working principle, and system functionality
The schematic diagram for the structure of a four-channel wavelength demultiplexer power combiner is shown in Fig. 2. The upstream wavelength is 1310 nm as defined by IEEE and ITUT PON standards and the downstream wavelengths are 155x nm as defined by ITUT DWDM grids. The two wavelength bands at 1310/155x nm are separated by a coarse wavelength coupler. The upstream signals are then combined by a 4×1 power combiner whereas the downstream channels are demultiplexed by a 1×4 wavelength demultiplexer. The downstream and upstream wavelength channels are combined by 1310/155x nm couplers. It should be pointed out that the scheme in Fig. 2 is generic and flexible in the sense that each component can be implemented in accordance with its pre-defined functionality based on different device technologies. Further, the design is fully scalable to a much larger networks with N>4.
In order to validate the functionality and performance of the scheme in Fig. 2, we implemented a demultiplexer/combiner module based on off-the-shelf optical components. The individual components and experimental set-sups are shown in Fig. 3. Hewlett Packard 83437A and Fiber-Lite PL-900 DC Regulated Illuminator are chosen as the light source for 155x nm and 1330nm optical transmitters, respectively. The optical spectrum analyzer (OSA) is Hewlett Packard 86142A. The PLC 4×1 power splitter (13P0104211), 1310/1550 nm couplers (C01M310211) and 4 channel 200GHz demultiplexer (204MITU211) are commercial products from ADF Fibercom Ltd. (China). These components are connected via short lengths of SMF-28 single mode fibers with negligible loss.
The measured transmission spectra are shown in Fig. 4(a) for the upstream channels and Fig. 4(b) for the downstream channels. The insertion losses for 1310nm and 155x nm are around −7dB and −2dB, respectively. The operation principle and functionality of the proposed optical module are therefore verified experimentally.
3. Design and simulation of monolithically integrated device based on planar lightwave circuits (PLC)
The optical module based on the discrete optical components is bulky and not readily scalable for a network with large number of optical network units. It is therefore highly desirable to implement and realize the same ideas on a fully integrated device based on standard planar lightwave circuits (PLC). To demonstrate the feasibility of this idea, we adopted a buried channel waveguide structure with high index contrast material (Hydex™ ) developed at Infinera (formerly Little Optics) to demonstrate the design concept. The refractive index of the core has 17% index contrasts with respect to the SiO2 cladding. The single mode condition and the bending loss of used channel waveguide are refer to the reference . In the following design, 1μm ×1μm cross section size of the channel waveguide is assumed.
A 1×2 MMI (Component ①) is employed to separate the upstream and downstream wavelength channels. A horizontal flipped 2×1MMI is used to combine the upstream and downstream optical signals at the output (component ④) .The operating mechanism of the MMI demultiplexer is based on the self-imaging effect. The input field in a multimode waveguide is reproduced as a single or multiple images periodically along the propagation direction due to the well-known multi-mode interference effect . By properly choosing the beating length, the MMI works like a directional coupler filter to separate the upstream / downstream signals. The flat channel characteristics are obtained by tapering the input/output ports. In this design, the coupler width and the length are chosen to be 4.7 μm and 197μm, respectively. The input/output ports are configured with 140μm long taper to flatten the spectral responses. The insertion spectral response of the 1×2 MMI is shown in Fig. 6. The simulated minimum insertion loss is 0.39 dB for the upstream channel and 0.8 dB around the downstream channels. It is noted that there is a pronounced discrepancy for X and Y polarizations which could be mitigated by optimizing the width and length of the coupler .
Component ② based on the MMI principle works as a power combiner for the upstream signals. The coupler width and length are chosen to be 100 μm and 2813 μm, respectively. Taper structures of 12 μm-wide and 500-μm long are used to connect the input/output ports. The port pitch is 25μm in our design. The simulated spectral response is shown in Fig. 7. Consistent performances for X and Y polarizations are observed around the wavelength 1300 nm. The insertion loss is around 6.8 dB.
The waveguide crosses (component ④) are utilized for the intersecting waveguides in the planar configurations. The width and the length of the resonator are chosen to be twice as the port width. The spectral response of the waveguide cross is shown in Fig. 8. The insertion loss is around 0.48~0.62 dB for the wavelength range of 1200nm –1600nm.
An arrayed-waveguide-grating (AWG) is employed as the downstream demultiplexer (component ③). The AWG consists of input/output waveguides, two free propagation regions (FPRs), and multiple channel waveguides whose lengths progressively decrease with a constant difference ΔL . The path difference ΔL is evaluated by ΔL = λ 0 2/(Neff FSR). The optimized path difference in the design is chosen to 90μm The input/output ports and the channel waveguide have been tapered to reduce insertion loss . The spectral response of the 1×4 AWG is shown in Fig. 9. and the minimum insertion loss 1.0 dB is observed. Further optimization of reducing the insertion loss and eliminating the polarization sensitivity could be carried out by optimizing the width and height of the arrayed waveguides and the different diffraction orders [16, 17].
The overall spectral responses of the circuit are shown in Fig. 10. The insertion loss for upstream signal is around -8 dB and -4dB for the downstream channels.
We have proposed and demonstrated the design of a novel optical device that can serve as a wavelength demultiplexer for downstream signals and power combiner for upstream signals in a high- performance and low-cost hybrid WDM-TDM PON architecture. The device is first implemented and verified experimentally as an optical module based on off-the-shelf commercial optical components. Further, a fully integrated scheme for implementation of the demultiplexer/combiner is designed and simulated based on planar lightwave circuits (PLC). The design procedure and key optimization considerations are discussed in detail. The simulation results show that the proposed design can satisfy the spectral requirements defined for the PON systems.
References and links
1. P. W. Shumate, “Fiber-to-the-home: 1977–2007,” J. Lightwave Technol. 26, 1093–1103 (2008). [CrossRef]
2. R. W. Heron, T. Pfeiffer, D. T. van Veen, J. Smith, and S. S. Patel, “Technology innovations and architecture solutions for the next-generation optical access network,” Bell Labs Tech. J. 13, 163–181 (2008). [CrossRef]
3. S. G. Mun, S. M. Lee, K. Okamoto, and C. H. Lee, “A multiple star WDM-PON using a band splitting WDM filter,” Opt. Express 16, 6260–6266 (2008), http://www.opticsinfobase.org/oe/abstract.cfm?URI=OPEX-16-9-6260. [CrossRef] [PubMed]
4. C. H. Lee, W. V. Sorin, and B. Y. Kim, “Fiber to the home using a PON infrastructure,” J. Lightwave Technol. 24, 4568–4583 (2006). [CrossRef]
5. I. Sankawa, F. Yamamoto, Y. Okumura, and Y. Ogura, “Cost and quantity analysis of passive double-star optical-access-network facilities for broadband service multiplexing,” J. Lightwave Technol. 24, 3625–3634 (2006). [CrossRef]
6. I. Tsalamanis, E. Rochat, S. D. Walker, M. C. Parker, and D. M. Holburn, “Experimental demonstration of cascaded AWG access network featuring bi-directional transmission and polarization multiplexing,” Opt. Express. 12, 764–769 (2004), http://www.opticsinfobase.org/oe/abstract.cfm?URI=OPEX-12-5-764. [CrossRef] [PubMed]
7. A. Banerjee, Y. Park, F. Clarke, H. Song, S. H. Yang, G. Kramer, K. Kim, and B. Mukherjee, “Wavelength-division-multiplexed passive optical network (WDM-PON) technologies for broadband access: a review,” J. Opt. Networking 4, 737–758 (2005). [CrossRef]
8. Y. Inoue, A. Himeno, K. Moriwaki, and M. Kawachi, “Silica-based arrayed-waveguide grating circuit as opticalsplitter/router,” Electron. Lett. 31, 726–727, (1995). [CrossRef]
9. Y. Li, L. Cohen, C. Henry, E. Laskowski, and M. Cappuzzo, “Demonstration and application of a monolithic two-PONs-in-one device,” in Proceedings of the Twenty-second European Conference on Optical Communication ECOC ’96, (NEXUS Media Ltd., Oslo, Norway, 1996), pp. 123–126.
10. B. Little, “A VLSI photonics platform,” in Optical Fiber Communication Conference, pp. 444–445 (2003).
11. C. L. Xu, X. B. Hong, and W. P. Huang, “Design optimization of integrated BiDi triplexer optical filter based on planar lightwave circuit,” Opt. Express 14, 4675–4686 (2006), http://www.opticsinfobase.org/oe/abstract.cfm?URI=OPEX-14-11-4675. [CrossRef] [PubMed]
12. APSS, “Apollo photonics solution suite,” Apollo Inc., Hamilton, Ontario Canada.
13. L. B. Soldano and E. C. M. Pennings, “Optical multimode interference devices based on self-Imaging -principles and applications,” J. Lightwave Technol. 13, 615–627 (1995). [CrossRef]
14. J. Lin, “Theoretical investigation of polarization-insensitive multimode interference splitters on silicon-on-insulator,” IEEE Photon Technol. Lett. 20, 1234–1236 (2008). [CrossRef]
15. M. K. Smit and C. vanDam, “PHASAR-based WDM-devices: Principles, design and applications,” IEEE J. Sel. Top. Quantum Electron. 2, 236–250 (1996). [CrossRef]
17. L. H. Spiekman, M. R. Amersfoort, A. H. deVreede, F. P. G. M. vanHam, A. Kuntze, J. W. Pedersen, P. Demeester, and M. K. Smit, “Design and realization of polarization independent phased array wavelength demultiplexers using different array orders for TE and TM,” J. Lightwave Technol. 14, 991–995 (1996). [CrossRef]