Open Access
Add to BibSonomyAbstract
Highly efficient, low-loss, and compact InP/InGaAsP polarization converter based on a half-ridge waveguide structure is fabricated and demonstrated experimentally. The device is fabricated by a simple self-aligned process and integrated with a ridge InP waveguide. Using a 150-μm-long device, we obtain the mode conversion of more than 96% and the on-chip loss of less than 1.0 dB over the broad wavelength range from 1510 to 1575 nm. The experimental results are explained quantitatively using the full-vector eigenmode calculation, which also reveals large fabrication tolerance of the demonstrated device.
© 2013 Optical Society of America
1. Introduction
With the continuous expansion of the channel capacity in the wavelength division multiplexing (WDM) systems and the utilization of more and more advanced coherent modulation formats, the complexity and the number of components in the optical transceivers have been increasing. InP-based photonic integrated circuits (PICs) have proved to be the promising approach toward reducing the cost and size of such highly complicated transceivers [1,2]. In order to increase the spectral efficiency further, there has recently been an increasing interest to develop monolithic PIC transceivers for polarization multiplexed (PM) signal formats [3, 4]. To realize compact and low-cost PM-PICs, simple scheme of manipulating the polarization state inside the PIC is of great importance.
To date, various types of waveguide polarization converters (PCs) have been demonstrated, including periodic loaded waveguides [5], micro-bended waveguides [6], adiabatic mode-evolution waveguides [7–9], and asymmetric waveguides [10–17]. Among them, those using asymmetric waveguides have advantages of relatively small footprint and low wavelength dependency. While several InP-based asymmetric waveguides have been proposed and experimentally demonstrated [13–17], monolithic integration of these PCs with laser diodes (LDs) and other active components has been challenging. This is largely due to two issues. First, many of these InP-based PC designs employ rib waveguide structures, which generally exhibit large modal mismatch with other typical ridge-waveguide components. Second, these schemes often require relatively complicated fabrication processes, such as strict lithographic alignment below 100 nm, hybrid (wet and dry) etching, and/or slanted dry etching, which may not be compatible with the standard LD fabrication procedures. In order to get the maximum benefit from the monolithic integration, a simple PC that can be integrated with both the active and passive components is of great interest.
We have recently proposed and numerically demonstrated a novel type of InP/InGaAsP PC, which was designed to have a particularly suitable structure for the integration with LDs and other photonic components [18]. As shown schematically in Fig. 1(a), it exhibits a half-ridge structure, having a ridge profile on one side and a deeply etched high-mesa profile on the other side. As a result, this PC should easily be integrated with both the ridge active devices such as LDs and modulators, and the high-mesa passive devices such as the arrayed waveguide gratings (AWGs), as shown in Fig. 1(b).
Fig. 1 Schematics of (a) the polarization converter integrated with ridge waveguides and (b) the polarization-multiplexed PIC with all the other components.
In this paper, we fabricate and experimentally demonstrate the half-ridge PC for the first time. The PC is integrated monolithically with standard ridge waveguides and fabricated by a self-aligned high-yield process without the need for a critical lithographic alignment. Using a 150-μm-long PC, we experimentally obtain the mode conversion of more than 96% and the on-chip loss of less than 1.0 dB over the entire C-band (1510–1575 nm). The experimental results are explained quantitatively using the full-vector eigenmode calculation, which also reveals relatively large fabrication tolerance of the demonstrated PC.
The paper is organized as follows. Section 2 describes the structure and the operation principle of the PC. In Section 3, we explain the self-aligned fabrication process. Then, in Section 4, we present the measurement results. Finally, we show the results of the eigenmode analysis and discuss the fabrication tolerance in Section 5.
2. Device structure and operation principle
Figure 1(a) shows the structure of the PC fabricated in this work. The layer stack consists of an InP substrate, a 0.5-μm-thick lattice-matched InGaAsP core layer with the photoluminescence peak at 1.3 μm, and an 1.2-μm-thick InP upper cladding. At the PC region, the waveguide has an asymmetric cross-section; one side of the waveguide is a ridge structure and the other is a high-mesa structure.
The operation of this PC is similar to that of a bulk birefringent medium with the principal axes rotated by ±45°. Due to the asymmetric cross section, the two lowest-order eigenmodes are hybridized and their electric and magnetic fields are tilted from the vertical axis. With the proper optimization of the waveguide parameters, these two eigenmodes can be adjusted to have the electric/magnetic field tilted approximately by ±45°. In such case, incident light with TE (or TM) polarization state excites the two eigenmodes with an equal magnitude. After propagating the half-beat length [Lπ = π/(β1 − β2), where βm is the propagation constant of mode m], these eigenmodes recombine into the orthogonal, i.e. TM (TE) polarization state [18]. Owing to the thick cladding and the half-ridge structure, these modes have large overlap with the TE and TM modes of a symmetric ridge waveguide, which allows low-loss integration with standard ridge components. By only controlling the etching depth, this PC should be integrated with other waveguide components as shown in Fig. 1(b).
3. Fabrication
From our numerical calculation [18], we selected the residual core thickness d shown in Fig. 1(a) to be 0.20 μm and varied the width w and the length LPC of the PC section. Two 20-μm-long tapered waveguides were inserted in between the PC section and the 2.5-μm-wide input/output ridge waveguides in order to reduce the coupling loss.
The PC was fabricated by using a simple self-aligned process as shown in Fig. 2. Following the ridge waveguide formation by Ar/Cl2-based inductively-coupled plasma reactive ion etching (ICP-RIE) (a), we deposited SiO2 from an angle by electron-beam (EB) evaporation (b). Since the ridge structure shades either side of the waveguide, we could deposit SiO2 on only one side of the waveguide. Next, a thick photoresist layer was patterned on the PC region by photolithography (c). The second layer of SiO2 was then deposited by EB evaporation from the opposite angle (d). After a lift-off process, we were left with a SiO2 hard mask layer that covered the both sides of the input/output ridge waveguides but only one side of the PC section (e). The second Ar/Cl2-based ICP-RIE was then used to form the deeply etched groove on the high-mesa side of the PC section (f). After removing the SiO2 hard mask by hydrofluoric acid, the entire structure was covered with SiO2 for passivation. Compared with a typical dry-etching fabrication process, only one additional photolithography step without critical alignment and two EB evaporation steps were required to integrate the PC. Since there is no critical lithographic alignment, the procedure is relatively simple and should be fully compatible with the fabrication process of other components.
Fig. 2 Self-aligned fabrication procedure of the PC integrated with a symmetric ridge waveguide. Cross-section structures of both the PC (left) and the symmetric ridge waveguide (right) are shown.
Scanning electron microscopic (SEM) images of the fabricated PC are shown in Fig. 3. The asymmetric structure is formed successfully and the PC is connected smoothly to the input/output ridge waveguide. On the other hand, we see that the ridge etching profile is not perfectly rectangular as shown in Fig. 1(a), but has a residual slope with a well-defined angle of 54.7°, corresponding to the (111) plane of InGaAsP. The effect of this residual etching slope will be discussed in Section 5.
Fig. 3 SEM photographs of the fabricated PC. Upper: cross-section view of the PC region. Lower: Boundary between the PC and the ridge waveguide.
4. Measurement
4.1. Setup and method
The optical measurement setup is shown in Fig. 4. The fabricated device was cleaved at the input ridge waveguide region and at the end of the PC region (as shown in Fig. 4) in order to measure the polarization state inside the PC section directly. No anti-reflection coating was applied at either facets. The total length of the measured device was 1 mm and the free-spectrum range (FSR) of Fabry-Pérot (FP) resonance between the cleaved facets was 0.36 nm in wavelength. In order to avoid the effect of this FP interference, we used an incoherent light source with the 3-dB bandwidth of 2 nm, which was generated by spectrally slicing the amplified spontaneous emission from an erbium-doped fiber amplifier (EDFA) by using an AWG. We should also note that we could not observe FP resonance with the FSR corresponding to the reflection at the ridge-PC interface. This implies that the back-reflection at the tapered interface should be negligibly small compared with those at the cleaved facets in our devices.
Fig. 4 Measurement setup for the PC. Note that the PC was cleaved at the center of the PC region in order to measure the polarization state inside the PC section directly.
The polarization state of the input light was adjusted carefully by using a polarizer followed by a half-wave plate (HWP) and a quarter-wave plate (QWP), so that the TE polarization mode was excited at the input ridge waveguide. The output light was characterized by a polarization analyzer (General Photonics PSY-101) to measure its Stokes parameters S0, S1, S2, S3. From the obtained parameters, the relative power in the TE and TM modes, PTE and PTM were calculated by
Any unitary transformation of the polarization state at the output patch fibers and couplers was calibrated carefully before the measurement.4.2. Measurement results
Figure 5 shows the polarization state measured at the output of the PC section with increasing value of LPC. We see that the polarization state evolves from the input TE polarization (S1 = 1, S2 = 0, S3 = 0) to a circular polarization (S1 = 0, S2 = 0, S3 = −1) and then to a TM polarization (S1 = −1, S2 = 0, S3 = 0), as expected from the operation principle of the PC described in the previous section.
Fig. 5 Measured polarization states at the output with increasing PC length LPC (w = 1.1 μm, d = 0.20 μm, wavelength = 1.55 μm).
Figure 6(a) shows the TE/TM conversion ratio measured for several different values of the width w and length LPC. Here, we define the TE/TM conversion ratio as PTM/(PTE + PTM). The maximum conversion ratio of as large as 98% (i.e., the polarization extinction ratio of 17 dB) is obtained at w = 1.1 μm and LPC = 150 μm. The dashed lines represent the sinusoidal fitting of the measured data, expressed as
where C, Cmax, and Lπ represent the conversion ratio, the maximum conversion ratio, and the half-beat length, respectively. Cmax and Lπ are used as the fitting parameters to match the measured data. Figure 6(b) shows the fitted values of Cmax and Lπ as a function of w.Fig. 6 (a) Measured TE/TM conversion ratio as a function of the PC length LPC (dots) for w = 0.7, 0.9, and 1.1 μm (d = 0.20 μm, wavelength = 1.55 μm). The dashed curves are the sinusoidal fit represented by Eq. (2). (b) The fitted values of the maximum conversion ratio Cmax and the half-beat length Lπ as a function of w.
The propagation loss at the 1.1-μm-wide PC section is measured to be 0.57 dB/mm by a cut-back method using a long PC. Since the half-beat length of this PC is 150 μm, the total propagation loss at the PC section should be less than 0.1 dB. The cut-back method is also applied to a reference symmetric ridge waveguide fabricated on the same chip, from which we derive the propagation loss at the ridge waveguide section as well as the coupling loss from/to the input/output fibers. By using these values, we estimate the total on-chip loss of the PC to be less than 1.0 dB, including the loss at the tapered section between the PC and the ridge waveguide.
Figure 7 shows the conversion ratio as a function of center wavelength of the input light for the case of w = 1.1 μm and LPC = 150 μm. The conversion ratio of more than 96% was obtained over the wavelength range from 1510 to 1575 nm, covering the entire C-band.
Fig. 7 Measured wavelength dependency of the TE/TM conversion ratio (w = 1.1 μm, d = 0.20 μm, LPC = 150 μm).
5. Numerical simulation
We perform numerical simulation using the full-vector finite difference method [19] and compared it with the measurement results. The cross-sectional structure assumed in the calculation is shown in Fig 8. To reflect the exact structure of the fabricated device shown in Fig. 3, we introduced the residual slope on the InGaAsP core at the ridge side with an angle of 54.7° with respect to the horizontal axis. The refractive indices of the InP, InGaAsP, and background are set to 3.17, 3.40, and 1.45, respectively.
Fig. 8 The cross-sectional structure assumed in the calculation (right), which reflects the actual structure of the fabricated device (left).
Figure 9 shows the calculated maximum conversion ratio Cmax and the half-beat length Lπ as a function of w in the case of d = 0.20 μm, together with the measured data. We see that the calculation agrees well with the measurement result. Although we find from the calculation that the influence of the etching slope is actually non-negligible on the PC performance, we should note that this shape is perfectly reproducible in our ICP-RIE process. We, therefore, believe that the effect of this non-rectangular etching profile would not be an issue in practice, once we derive a proper design that takes into account the actual etching profile.
Fig. 9 Calculated maximum conversion ratio Cmax (a) and half-beat length Lπ (b) as a function of w in the case of d = 0.20 μm together with the measured data (dots).
To estimate fabrication tolerance of our device, we calculate the sensitivity of Cmax and Lπ on both w and d. Figure 10 shows the calculated Cmax and Lπ as a function of w and d. We can calculate the net conversion ratio for a given LPC by using Fig. 10 and Eq. (2). When we set LPC = 100 μm for instance, the allowed regimes for w and d to achieve the conversion over 90% are calculated to be 0.84±0.06 μm and 0.31±0.02 μm, respectively. We believe that these values of fabrication tolerance are readily achievable with the current state-of-the-art InP PIC fabrication technology [20].
Fig. 10 Calculated maximum conversion ratio Cmax (a) and half-beat length Lπ (b) as a function of w and d (wavelength = 1.55 μm).
6. Conclusion
We have fabricated and experimentally demonstrated a simple and compact InP/InGaAsP polarization converter. With a 150-μm-long device, the TE/TM conversion of 96% was obtained over a broad wavelength range of 1510–1575 nm. The eigenmode calculation agreed well with the measurement result, which also revealed relatively large tolerances to fabrication errors. Owing to the LD-compatible half-ridge cross-sectional structure and simple self-aligned fabrication process, the demonstrated PC should be useful in realizing various types of polarization-multiplexed PICs with integrated LDs.
Acknowledgments
This work was supported by Grant-in-Aid for Scientific Research (S) #20226008 and Research Fellowships for Young Scientists, Japan Society for the Promotion of Science.
References and links
1. R. Nagarajan, C. H. Joyner, J. Richard, P. Schneider, J. S. Bostak, T. Butrie, A. G. Dentai, V. G. Dominic, P. W. Evans, M. Kato, M. Kauffman, D. J. H. Lambert, S. K. Mathis, A. Mathur, R. H. Miles, M. L. Mitchell, M. J. Missey, S. Murthy, A. C. Nilsson, F. H. Peters, S. C. Pennypacker, J. L. Pleumeekers, R. A. Salvatore, R. K. Schlenker, R. B. Taylor, H.-S. Tsai, M. F. V. Leeuwen, J. Webjorn, M. Ziari, D. Perkins, J. Singh, S. G. Grubb, M. S. Reffle, D. G. Mehuys, F. A. Kish, and D. F. Welch, “Large-scale photonic integrated circuits,” J. Sel. Top. Quantum Electron. 11, 50–65 (2005) [CrossRef] .
2. T. Fujisawa, S. Kanazawa, H. Ishii, N. Nunoya, Y. Kawaguchi, A. Ohki, N. Fujiwara, K. Takahata, R. Iga, F. Kano, and H. Oohashi, “1.3-μm 4×25-Gb/s monolithically integrated light source for metro area 100-Gb/s Ethernet,” IEEE Photon. Technol. Lett. 23, 356–358 (2011) [CrossRef] .
3. M. Kato, P. Evans, S. Corzine, J. Gheorma, M. Fisher, M. Raburn, A. Dentai, R. Salvatore, I. Lyubomirsky, A. Nilsson, J. Rahn, R. Nagarajan, C. Tsai, B. Behnia, J. Stewart, D. Christini, M. Missey, A. Spannagel, D. Lambert, S. Agashe, P. Liu, D. Pavinski, M. Reffle, R. Schneider, M. Ziari, C. Joyner, F. Kish, and D. Welch, “Transmitter PIC for 10-channel x 40Gb/s per channel polarization-multiplexed RZ-DQPSK modulation,” in proceedings of Optical Fiber Communication Conference (OFC2009), OThN2 (2009).
4. R. Nagarajan, J. Rahn, M. Kato, J. Pleumeekers, D. Lambert, V. Lal, H.-S. Tsai, A. Nilsson, A. Dentai, M. Kuntz, R. Malendevich, J. Tang, J. Zhang, T. Butrie, M. Raburn, B. Little, W. Chen, G. Goldfarb, V. Dominic, B. Taylor, M. Reffle, F. Kish, and D. Welch, “10 channel, 45.6 Gb/s per channel, polarization-multiplexed DQPSK, InP receiver photonic integrated circuit,” J. Lightwave Technol. 29, 386–395 (2011) [CrossRef] .
5. Y. Shani, R. Alferness, T. Koch, U. Koren, M. Oron, B. I. Miller, and M. G. Young, “Polarization rotation in asymmetric periodic loaded rib waveguides,” Appl. Phys. Lett. 59, 1278–1280 (1991) [CrossRef] .
6. C. van Dam, L. H. Spiekman, F. P. G. M. van Ham, F. H. Groen, J. J. G. M. van der Tol, I. Moerman, W. W. Pascher, M. Hamacher, H. Heidrich, C. M. Weinert, and M. K. Smit, “Novel compact polarization convertors based on ultra short bends,” IEEE Photon. Technol. Lett. 8, 1346–1348 (1996) [CrossRef] .
7. M. R. Watts and H. A. Haus, “Integrated mode-evolution-based polarization rotators,” Opt. Lett. 303, 138–140 (2005) [CrossRef] .
8. T. Barwicz, M. R. Watts, M. A. Popović, P. T. Rakich, L. Socci, F. X. Kärtner, E. P. Ippen, and H. I. Smith, “Polarization-transparent microphotonic devices in the strong confinement limit,” Nat. Photon. 1, 57–60 (2006) [CrossRef] .
9. J. Zhang, T.-Y. Liow, M. Yu, G.-Q. Lo, and D.-L. Kwong, “Silicon waveguide based TE mode converter,” Opt. Express 18, 25264–25270 (2010) [CrossRef] [PubMed] .
10. J. Z. Huang, R. Scarmozzino, G. Nagy, M. J. Steel, and R. M. Osgood Jr., “Realization of a compact and single-mode optical passive polarization converter,” IEEE Photon. Technol. Lett. 12, 317–319 (2000) [CrossRef] .
11. B. M. Holmes and D. C. Hutchings, “Realization of novel low-loss monolithically integrated passive waveguide mode converters,” IEEE Photon. Technol. Lett. 18, 43–45 (2006) [CrossRef] .
12. H. Fukuda, K. Yamada, T. Tsuchizawa, T. Watanabe, H. Shinojima, and S. ichi Itabashi, “Polarization rotator based on silicon wire waveguides,” Opt. Express 16, 2628–2635 (2008) [CrossRef] [PubMed] .
13. H. El-Refaei, D. Yevick, and T. Jones, “Slanted-rib waveguide InGaAsP-InP polarization converters,” J. Light-wave Technol. 22, 1352–1357 (2004) [CrossRef] .
14. L. M. Augustin, J. J. G. M. van der Tol, E. J. Geluk, and M. K. Smit, “Short polarization converter optimized for active-passive integration in InGaAsP-InP,” IEEE Photon. Technol. Lett. 19, 1673–1675 (2007) [CrossRef] .
15. D. M. Beggs, M. Midrio, and T. F. Krauss, “Compact polarization rotators for integrated polarization diversity in InP-based waveguides,” Opt. Lett. 32, 2176–2178 (2007) [CrossRef] [PubMed] .
16. S.-H. Kim, R. Takei, Y. Shoji, and T. Mizumoto, “Single-trench waveguide TE-TM mode converter,” Opt. Express 17, 11267–11273 (2009) [CrossRef] [PubMed] .
17. C. Alonso-Ramos, S. Romero-García, A. Ortega-Moñux, I. Molina-Fernández, R. Zhang, H. G. Bach, and M. Schell, “Polarization rotator for InP rib waveguide,” Opt. Lett. 37, 335–337 (2012) [CrossRef] [PubMed] .
18. T. Tanemura, T. Amemiya, K. Takeda, A. Higo, and Y. Nakano, “Simple and compact InP polarization converter for polarization-multiplexed photonic integrated circuits,” in proceedings of IEEE Conference on Lasers and Electro-Optics Society Annual Meeting (LEOS’09) , 436–437 (2009) [CrossRef] .
19. P. Lüsse, P. Stuwe, J. Schule, and H.-G. Unger, “Analysis of vectorial mode fields in optical waveguides by a new finite difference method,” J. Lightwave Technol. 12, 487–494 (1994) [CrossRef] .
20. M. Smit, J. van der Tol, and M. Hill, “Moore’s law in photonics,” Laser & Photon. Rev. 6, 1–13 (2012) [CrossRef] .
