INVITED PAPER

Vertical-cavity semiconductor optical amplifiers: recent progress and emerging applications

Staffan Björlin, Garrett Cole, Toshio Kimura, John Bowers

Electrical and Computer Engineering Department University of California Santa Barbara CA 93106, USA

Phone: +1-805-893-7003, Fax: +1-805-893-7990, Email: bjorlin@ece.ucsb.edu

Abstract: We review the development of long-wavelength vertical-cavity semiconductor optical amplifiers. General theory and device properties will be presented, and potential applications will be outlined. Results from 1540-nm VCSOAs will be reported. © 2004 Optical Society of America OCIS codes: (140.4480) Optical amplifiers; (140.3280) Laser amplifiers; (250.5980) Semiconductor optical amplifiers.

1. A brief VCSOA history Vertical-cavity semiconductor optical amplifiers (VCSOAs) are a relatively new class of devices with unique properties. Compared to more conventional amplifier technologies, VCSOAs show many advantages such as small form factor, low power consumption, high coupling efficiency to optical fiber, polarization independent gain, and potentially low manufacturing cost. Because of the current focus on inexpensive, small form factor components, VCSOAs are now attracting increasing interest. The first VCSOA was demonstrated in 1991 by Koyama, Kubota, and Iga at Tokyo Institute of Technology [1]. They used an electrically pumped GaAs/AlGaAs VCSEL structure to amplify an injected 885-nm signal. The favorable filtering properties stemming from the high-finesse VCSEL cavity was recognized; the device was not presented as an amplifier but as an active filter. No fiber-to-fiber gain was obtained but about 4 dB internal gain was reported. Two years later, in 1993, pulsed operation of an optically pumped reflection mode device, also at 850 nm, was presented by Raj et al. at France Telecom. It was presented as an amplifying photonic switch [2]. The same group introduced resonant pumping in a following generation of 850-nm devices [3] and in 1996 they presented the first long-wavelength VCSOA [4]. Also in 1996, Wiedenmann et al. at University of Ulm presented an electrically pumped reflection mode VCSOA operating at 980 nm [5]. In 1998, Wiedenmann et al. presented their second generation of devices: an electrically pumped, transmission mode VCSOA with an oxide aperture for current and mode confinement that produced up to 16 dB of gain [6]. In 1998, Lewen et al. at KTH in Sweden used a 1.55-µm VCSEL structure for what was the first electrically pumped long wavelength VCSOA [7]. They measured 18 dB of gain at 218 K (not including coupling losses). The VCSOA project at UCSB started in 1999 and lead to the demonstration of the first 1.3-µm VCSOA in 2000 [8]. These devices were fabricated using InP-GaAs wafer bonding, they were optically pumped, and operated in reflection mode. This first generation was used to fully characterize this still fairly new class of devices [9,10], to develop improved theoretical models [11,12], and to explore possible applications for VCSOAs [13,14]. A second generation of 1.3-µm devices with improved efficiency and higher gain (17 dB fiber to fiber gain) was presented in 2002 [15]. More recently, we have presented 1.55-µm VCSOAs with record-high saturation output power [16]. Tunable VCSOAs have been demonstrated at UCSB [17,18] with tuning over xx nm. Meanwhile, Calvez et al. at the University of Strathclyde, UK, have demonstrated the first GaInNAs-based devices operating in the 1.3-µm wavelength region with up to 17.7 dB of fiber-to-fiber gain [19]. The progress in continuous wave (CW) amplifier gain of VCSOAs (not including coupling losses) is summarized in Figure 1. 2. VCSOA Design The basic structure of a VCSOA consists of an active region enclosed by two mirrors. The device can be optimized for operation in either reflection mode or transmission mode, as shown schematically in Figure 2. It is easier to achieve good amplifier characteristics in devices optimized for reflection mode operation. The combined mirror loss is lower compared to transmission mode operation, so good signal gain can be achieved for lower single-pass gain. Reflection mode operation might also be a more cost effective approach since the fiber alignment, which is a very difficult and costly step in the manufacturing, is reduced from two fibers to one. However, the input and output signals need to be separated. The separation calls for an additional component (coupler or circulator), which adds complexity, cost, and signal loss. Operation in transmission mode is more attractive in some applications, e.g. integration with detectors for preamplification or array applications. It is, however, a more difficult approach as far as testing and packaging. The choice of operational mode might ultimately depend on the intended application for the VCSOA.

INVITED PAPER

Fig 1. Progress in VCSOA gain. Fig 2. Treansmission mode (left) and reflection mode operation (right). The development of VCSOAs has benefited greatly from VCSEL research. Materials and processing technologies developed for VCSELs can be directly applied to VCSOAs, and the design of the two is in many ways similar. The main difference is that strong feedback is desired for VCSELs in order to minimize the required threshold current. In VCSOAs, on the other hand, reduced feedback is advantageous in order to enable high gain without the onset of lasing. VCSOAs therefore require higher single-pass gain (more QWs) and lower mirror reflectivity than VCSELs. The large number of QWs typically used in VCSOAs makes it difficult to achieve uniform carrier distribution throughout the QWs using electrical injection. Optical pumping is an attractive way to pump VCSOAs for a number of reasons. Optical pumping generates carriers in the QWs, without the need of transporting the carrier through the structure. This results in very uniform carrier distribution throughout a large number of QWs. It also allows the entire structure to be undoped, which simplifies growth and processing, and minimizes optical losses. Furthermore, optical pumping can generate a uniform carrier distribution across a laterally large active region. Several high-performance long wavelength VCSELs have been presented that use optical pumping [20,21]. To maintain a small footprint, device and pump laser can be packaged in the same package, or even integrated into the same structure [20]. The first theoretical predictions of VCSOA performance was presented by Tombling et al. in 1994 [22]. The model that was used was based on carrier and photon rate equations and the FP equations for a cavity with gain. Karlsson et al. used a similar approach, but also analyzed the detector characteristics of VCSOAs in a 1996 publication [23]. In 1999, Kibar presented a detailed VCSOA analysis based on small signal equivalent circuits and rate equations [24]. In 2000, Piprek et al. presented a more detailed rate equation model [11]. In these early theoretical VCSOA papers—as well as in even older in-plane SOA publications—the results obtained using rate equation analysis and the FP approach did not agree. This disagreement was caused by the omission of interference between the fields that traverse the input mirror in both directions, which lead to an incorrect expression for the mirror loss in the photon rate equation. The problem was solved in 2002 by Royo et al. who showed that the mirror loss actually depends on the gain in the amplifier [25]. The balance between the reflectivity of the mirrors and the gain provided by the active region is the most important issue in VCSOA design. The reflectivity of the two mirrors has a large impact on all properties of the amplifier, and must be chosen carefully. Strong feedback, i.e. high mirror reflectivity, leads to high gain for a given value of single pass gain, but the amplifier gain is limited by lasing threshold. It also leads to poor noise figure and early saturation. Mirror reflectivity that is too low, on the other hand, simply does not provide sufficient feedback to reach high amplifier gain. For optimum performance it is desirable that the mirror reflectivity be as high as possible without enabling lasing threshold to be reached. This condition allows for operation at full population inversion, which gives the highest possible amplifier gain and gain-bandwidth product, the highest saturation output power and the lowest noise figure. 3. Applications There are a wide range of potential applications for VCSOAs. Proposed applications include optical interconnects [24,26], switching/modulation [2-4,13,27], and optical preamplification of high-bit rate receivers [14]. Optical bistability in FPSOAs is a well known phenomena that may enable the realization of all-optical logic and memory elements. Optical bistability was recently observed in a reflection mode VCSOA [28]. Optical preamplification of high-speed receivers is perhaps the most interesting application for VCSOAs. Desired properties for this application are good noise performance and polarization independent gain, which are areas of difficulty for in-plane devices. Also desired are low power consumption, compactness, and low cost,

INVITED PAPER

properties that are not associated with fiber amplifiers. VCSOAs can meet all these criteria. We have demonstrated optical preamplification of a PIN receiver at 10 Gb/s. 7 dB sensitivity improvement was achieved for 11 dB of VCSOA gain, resulting in a receiver sensitivity of –26.2 dBm [14]. The inherent filtering effect of the narrow VCSOA bandwidth is advantageous for preamplifier applications, as it eliminates out-of-band noise and provides channel selection in multi-wavelength systems. However, in order to make VCSOAs more flexible for such applications, tunable devices must be developed. Tunable devices employing microelectromechanical (MEMS) tuning has demonstrated a tuning range of 11 nm [18]. Much wider tuning ranges, similar to what has been achieved for tunable VCSELs [21], can be expected from future generations of tunable VCSOAs. References [1] F. Koyama, S. Kubota, K. Iga, “GaAlAs/GaAs active filter based on vertical cavity surface emitting laser,” Electron. Lett., 27, 1093-1095, (1991). [2] R. Raj, J. A. Levenson, J. L. Oudar, M. Bensoussan, “Vertical microcavity optical amplifying switch,” Electron Lett., 29, 167-169, (1993). [3] R. Raj, J. L. Oudar, M. Bensoussan, “Vertical cavity amplifying photonic switch,” Appl. Phys. Lett., 65, 2359-2361, (1994). [4] N. Bouché, B. Corbett, R. Kuszelewicz, R. Ray, “Vertical-cavity Amplifying Photonic Switch at 1.5 µm”, IEEE Photon. Technol. Lett., 8, 1035-1037, (1996). [5] D. Wiedenmann, B. Moeller, R. Michalzik, K. J. Ebeling, “Performance characteristics of vertical-cavity semiconductor optical amplifiers,” Electron. Lett., 32, 342-343, (1996). [6] D. Wiedenmann, C. Jung, M. Grabherr, R. Jäger, U. Martin, R. Michalzik, K. J. Ebeling, “Oxide-confined vertical-cavity semiconductor optical amplifier for 980 nm wavelength”, in CLEO 98 Technical Digest, 378, (1998). [7] R. Lewén, K. Streubel, A. Karlsson, S. Rapp, “Experimental Demonstration of a Multifunctional Long-Wavelength Vertical-Cavity Laser Amplifier-Detector”, IEEE Photon. Technol. Lett., 10, 1067-1069, (1998). [8] E. S. Björlin, B. Riou, A. Keating, P. Abraham, Y-J Chiu, J. Piprek, J. E. Bowers, “1.3-µm Vertical-Cavity Amplifier,” IEEE Photon. Technol. Lett., 12, 951-953, (2000). [9] E. S. Björlin, B. Riou, P. Abraham, J. Piprek, Y-J. Chiu, K. A. Black, A. Keating, J. E. Bowers, “Long Wavelength Vertical-Cavity Semiconductor Optical Amplifiers”, IEEE J. Quantum Electron., 37, 274-281, (2001). [10] E. S. Björlin, J. E. Bowers, “Noise Figure of Vertical-Cavity Semiconductor Optical Amplifiers,” IEEE J. Quantum Electron., 38, 61-66, (2002). [11] J. Piprek, S. Björlin, J. E. Bowers, “Design and Analysis of Vertical-Cavity Semiconductor Optical Amplifiers,” IEEE J. Quantum Electron., 37, 127-134, (2002). [12] J. Piprek, E. S. Björlin, J. E. Bowers, “Optical gain-bandwidth product of vertical-cavity laser amplifiers,” Electron. Lett., 37, 298-299, (2001). [13] E. S. Björlin, J. Piprek, S. Gee, Y.-J. Chiu, J. E. Bowers, A. Dahl, P. Abraham, “1.3µm vertical-cavity amplifying switch,” OSA TOPS Optical Amplifier and Their Applications, 60, 154-160, (2001). [14] E. S. Björlin, J. Geske, J. E. Bowers, “Optically preamplified receiver at 10 Gbit/s using vertical-cavity SOA,” Electron. Lett., 37, 1474- 1475, (2001). [15] E. S. Björlin, P. Abraham, D. Pasquariello, J. Piprek, Y.-J. Chiu, J. E. Bowers, “High Gain, High Efficiency Vertical-Cavity Semiconductor Optical Amplifiers,” in Proc. 14th Indium Phosphide and Related Materials Conference, 307-310, (2002). [16] E. S. Björlin, T. Kimura, Q. Chen, C. Wang, J. E. Bowers, “High output power 1540 nm vertical cavity semiconductor optical amplifiers,” Electron. Lett., 40, (2004). [17] T. Kimura, S. Björlin, J. Piprek, J. E. Bowers, “High-Temperature Characteristics and Tunability of Long-Wavelength Vertical-Cavity Semiconductor Optical Amplifiers,” IEEE Photon. Technol. Lett., 15, 1501-1503, (2003). [18] Q. Chen, G. D. Cole, E. S. Björlin, T. Kimura, S. Wu, C. S. Wang, N. C. MacDonald, J. E. Bowers, “First Demonstration of a MEMS Tunable Vertical-Cavity SOA,” IEEE Photon. Technol. Lett., 16, (2004). [19] S. Calvez, A. H. Clark, J.-M. Hopkins, R. Macaluso, P. Merlin, H. D. Sun, M. D. Dawson, T. Jouhti, M. Pessa, “1.3 µm GaInNAs optically- pumped vertical cavity semiconductor optical amplifier,” Electron. Lett., 39, 100-102, (2003). [20] V. Jayaraman, T. J. Goodnough, T. L. Beam, F. M. Ahedo, R. A. Maurice, “Continuous-wave operation of single-transverse-mode 1310-nm VCSELs up to 115 degrees C,” IEEE Photon. Technol. Lett., 12, 1595-1597, (2000). [21] D. Vakhshoori, J.-H. Zhou, M. Jiang, M. Azimi, K. McCallion, C.-C. Lu, K. J. Knopp, J. Cai, P. D. Wang, P. Tayebati, H. Zhu, P. Chen, “C-band tunable 6 mW vertical-cavity surface-emitting lasers,” in OFC 2000 Tech. Dig. Postconf. Edition, PD13-1-PD13-3, (2000). [22] C. Tombling, T. Saitoh, T. Mukai, “Performance Predictions for Vertical-Cavity Semiconductor Laser Amplifiers,” IEEE J. Quantum Electron., 30, 2491-2499, (1994). [23] A. Karlsson, M. Höijer, “Analysis of a VCLAD: Vertical-Cavity Laser Amplifier Detector,” IEEE Photon. Technol. Lett., 7, 1336-1338, (1995). [24] O. Kibar, “VCSEL-based Digital Free-Space Optoelectronic interconnections,” Ph.D. dissertation (UCSD, San Diego, CA 1999). [25] P. Royo, R. Koda, L. A. Coldren, “Vertical Cavity Semiconductor Optical Amplifiers: Comparison of Fabry-Perot and Rate Equation Approaches,” IEEE J. Quantum Electron., 38, 279-284, (2002). [26] N. Suzuki, M. Ohashi, M. Nakamura, “A Proposed Vertical-Cavity Optical Repeater for Optical Inter-Board Connections”, Photon. Technol. Lett., 9, 1149-1151, (1997). [27] P. Wen, M. Sanchez, O. Kibar, S. C. Esener, “Low-voltage, high contrast-ratio, low-noise VCSEL modulator,” in Proc. OSA Topical Mtg. on Optical Amplifiers and Their Applications, 265-266, (2000). [28] P. Wen, M. Sanchez, Matthias Gross, S. Esener, “Observation of bistability in a Vertical-Cavity Semiconductor Optical Amplifier (VCSOA),” Optics Express, 10, 1273-1278, (2002).