SEMICONDUCTOR lasers and amplifiers are important

240 JOURNAL OF LIGHTWAVE TECHNOLOGY, VOL. 28, NO. 3, FEBRUARY 1, 2010 Temperature-Dependent Saturation Characteristics of Injection Seeded Fabry Pérot Laser Diodes/Reflective Optical Amplifiers Hongyun Meng, Jung-Hyung Moon, Ki-Man Choi, and Chang-Hee Lee, Fellow, IEEE Abstract In this paper, the temperature dependence of the gain and saturation power of injection seeded Fabry Perot laser diodes/ reflective semiconductor optical amplifiers are analyzed theoretically and experimentally. For a constant gain, the saturation power increases with the ambient temperature. This dependency explains the observed variation in relative intensity noise versus injection power, as a function of temperature. Index Terms Light injection, wavelength-division-multiplexed (WDM) passive optical network, wavelength-locked Fabry Perot laser diode (F-P LD). I. INTRODUCTION SEMICONDUCTOR lasers and amplifiers are important components both for optic communication systems and for optical signal processing systems. We proposed the wavelength-locked Fabry Perot laser diode (F-P LD)-based wavelength-division-multiplexing (WDM) passive optical network, which has gained great attention due to its cost-effectiveness and colorless operation [1] [6]. We have proposed the dynamic model for the wavelength-locked F-P LD [7]. This dynamic model predicts modulation characteristics and noise evolution of the wavelength-locked F-P LDs. Unfortunately, it requires long computational time. In this paper, taking the detailed recombination mechanism and the gain as function of temperature and wavelength into account, we propose a simple model and investigate the saturation characteristics of the injection seeded F-P LDs/reflective semiconductor optical amplifiers (RSOAs). The simulation results show that the saturation power increases as we increase the temperature and explain experimentally observed features of relative intensity noise (RIN) changing with the variation of temperature. This will provide important information for designing an uncooled optical network unit for outdoor applications. Manuscript received July 29, 2009; revised October 01, 2009, November 03, 2009. First published December 15, 2009; current version published January 15, 2010. This work was supported in part by Korea Ministry of Science and Technology under the National Research Laboratory and Brain Korea 21 Project. H. Meng was with the Department of Electrical Engineering, Korea Advanced Institute of Science and Technology, Daejeon 305-701, Korea. He is now with the Laboratory of Photonic Information Technology, South China Normal University, Guangzhou 510631, China (e-mail: hymeng@scnu.edu.cn). J.-H. Moon and C.-H. Lee are with the Department of Electrical Engineering, Korea Advanced Institute of Science and Technology, Daejeon 305-701, Korea (e-mail: chl@ee.kaist.ac.kr). K.-M. Choi was with the Department of Electrical Engineering, Korea Advanced Institute of Science and Technology, Daejeon 305-701, Korea. She is now with the Network Infra Research Department, Korea Telecom Network Technology Laboratory, Daejeon 305-811, Korea (e-mail: kiman@kt.com). Digital Object Identifier 10.1109/JLT.2009.2038596 Fig. 1. ASE-injection wavelength-locked F-P LD. II. THEORY Fig. 1 shows a schematic diagram of the injection seeded F-P laser/rsoa with amplified spontaneous emission (ASE) injection, which includes an ASE optical source, an F-P LD/RSOA, an arrayed waveguide grating (AWG), and a circulator. In general, the gain depends on the bias current, facet reflectivity, the material gain, and the recombination rate. It is well known that the material gain and the recombination rate are strongly dependent on the operating temperature. For simplicity, the carrier density is assumed to be a constant along the active region of the F-P LD. Then the fiber-to-fiber spectral gain of the injected ASE can be obtained by using the nonlinear etalon theory [8]. where is the total coupling efficiency between the fiber and the F-P LD/RSOA. and are the front and the rear facet reflectivities, respectively. is the single-pass material gain, given by where is the confinement factor, is the material gain, is the internal loss, and is the active length. The single-pass phase shift is given by where is the central frequency of the input ASE, is the frequency of a resonant mode of the F-P LD, and is the group velocity. It may be noted that both the fiber-to-fiber spectral gain and the single-pass material gain are a function of the wavelength. (1) (2) (3) 0733-8724/$26.00 2010 IEEE

MENG et al.: FABRY PÉROT LASER DIODES/REFLECTIVE OPTICAL AMPLIFIERS 241 The material gain depends on the carrier density in the active region. In order to find the carrier density, the steady-state rate equation is used as follows: TABLE I LIST OF PHYSICAL PARAMETERS where is the electron charge, is the volume of the active region, is the bias current, and is the total recombination rate. and are the total intensity and the saturation intensity, respectively. is the differential gain constant at a given temperature. The saturation intensity can be shown as where E is the photon energy. As for the recombination processes in a long-wavelength semiconductor laser, the Auger recombination is important and it has strong sensitivity to temperature. The Auger recombination process can be modeled as [9] (4) (5) where is constant. Here, the temperature-dependent constants and can be given by [10] and [11] where is the chip temperature of the device, is the bandgap, and is the Boltzmann constant. At a given carrier density, the carrier lifetime in (4) can be obtained as. It may be noted that the chip temperature is higher than the ambient temperature due to the self-heating effect. For the material gain in (2) and (4), we used the following approximated model, in which both temperature and carrier density dependence are included. where is the wavelength of injection ASE, is the carrier density at transparency, is the wavelength at the gain peak, and is the bandwidth of the gain, which are shown as follows: (6) (7) (8) (9) (10) (11) (12) (13) where is a reference temperature., and are the differential gain coefficient, the peak wavelength, the bandwidth, and the carrier density at the lasing threshold of the F-P LD at the reference temperature, respectively. The material gain in (9) decreases with increasing temperature since and. The wavelength at the gain peak shifts to longer wavelength as we increase the temperature, while it shifts to shorter wavelength as we increase the carrier density. We also include the change of the gain bandwidth as function of temperature and carrier density. It increases as we increase the temperature and decreases as we increase the carrier density. With (5), we can see that as the temperature increases, the saturation intensity increases, since both the gain coefficient and the carrier lifetime decrease. The decrease of the carrier lifetime is caused by increase of the threshold carrier density and the Auger recombination. III. RESULTS AND DISCUSSIONS We selected a typical F-P LD as a reference laser for the simulation and experiment. and were assumed as 0.01 and 0.32, respectively. The channel spacing and 3 db bandwidth of the AWG was 100 GHz and 0.4 nm, respectively, in the simulation and experiment. The resolution bandwidth of the optical spectrum analyzer was 0.2 nm. The values for other parameters used in the simulation are listed in Table I. The threshold current with these simulation parameters was calculated as 24 ma at the ambient temperature of 25 C. To include heat generated by laser itself, we assumed a series resistance due to ohmic contact of the F-P LD as 5. The thermal resistance of the F-P LD package was assumed as 50 K/W. Fig. 2 shows the calculated output spectra of the F-P LD without ASE injection at different ambient temperatures and different bias currents. It may be noted that we used ambient temperature for easy comparison with experimental results, though we included self-heating effects in the model. The peak wavelength of the output envelope was shifted to the longer wavelength side as we increase the temperature. It shifts slightly to the shorter wavelength side, when we increase the injection current. It may be noted that the self-heating diminishes the blue shift induced by carrier density. This feature was also observed in the experiment. To investigate the temperature dependence of the injection current for a given fiber-to-fiber power gain (ratio of the total

242 JOURNAL OF LIGHTWAVE TECHNOLOGY, VOL. 28, NO. 3, FEBRUARY 1, 2010 Fig. 2. Output spectra at different temperature and I. Fig. 4. I versus gain at the ambient temperature of 40 C with an ASEinjection power of 011:7 dbm. Fig. 3. I versus ambient temperature with a gain of 12 db. Fig. 5. Gain versus ASE power at different ambient temperature. output power within the bandwidth of the AWG to the total injection power), we calculated the required injection current as a function of temperature with different front facet reflectivity. The injection ASE power and fiber-to-fiber gain were dbm (The corresponding injection spectral density was dbm/0.2 nm.) and 12 db, respectively. As shown in Fig. 3, the injection current increases exponentially as we increase the temperature. This can be explained as both decrease of material gain and increase of recombination rate. Internal heating also contributes to the increase of the injection current. It may be noted that thermal run away could occur, if the thermal resistance of the device is high. Another important feature is the increase of injection current with decreasing front facet reflectivity. This effect is more pronounced at higher gains, as shown in Fig. 4. This feature implies that the F-P LD with higher front facet reflectivity will have better high-temperature characteristics. The F-P LD with the front facet reflectivity of 0.0001 can be considered as RSOA. Thus, the RSOA will have worse high temperature characteristics provided by the same gain medium and the same thermal resistance of the package. To investigate saturation characteristics, we calculated the fiber-to-fiber power gain as a function of the injection ASE power at different ambient temperature, as shown in Fig. 5. We maintained constant small signal gain by adjusting the injection current. It may be noted that the fiber-to-fiber power gain was calculated when the injection wavelength matches that of an F-P mode. The gain decreases due to gain saturation as we increase the injection ASE power. The saturation characteristic shows different saturation powers at different temperatures. The saturation power increases with increasing temperature, as shown in Fig. 6. This can be caused primarily by two effects: the decrease in the material gain and the increase in Auger recombination as the temperature increases. It may be noted that the saturation power increases almost exponentially as the temperature increases, regardless of the front facet reflectivity. The saturation power also depends on the front facet reflectivity. As we decrease the front facet reflectivity, the saturation power increases. It implies that the RSOA has a higher saturation power compared with the F-P LD. To verify the theoretical prediction experimentally, we measured the fiber-to-fiber power gain as a function of the ASE injection power at different temperatures, as shown in Fig. 7. and of the F-P LD used in the experiment were 0.01 and 0.32, respectively. The bias currents were 20, 25, and 75 ma

MENG et al.: FABRY PÉROT LASER DIODES/REFLECTIVE OPTICAL AMPLIFIERS 243 Fig. 6. Saturation power versus ambient temperature. Fig. 8. Measured RIN versus ASE power at different ambient temperature. Fig. 7. Gain versus ASE injection power at different temperatures. at the ambient temperatures of 25 C, 40 C, and 60 C, respectively, while the threshold current was 19 ma at the ambient temperature of 25 C. We matched the gain at the injection power of dbm (The spectral density at the peak was dbm/0.2 nm.) by changing the injection current. For different ASE injection power, the biggest gain and the smallest gain (has not been shown) have been achieved via adjusting the detuning between the ASE injection wavelength and the F-P LD lasing mode. Fig. 7 shows that the simulation results are in good agreement with the experiment results. With the simulation results, the saturation powers were, and 1.85 dbm at the temperatures of 25 C, 40 C, and 60 C, respectively. We also measured the RIN as a function of the ASE injection power at different temperatures, as shown in Fig. 8. We can get the best RIN (shown in Fig. 8) and the worst RIN (has not been shown) via changing the detuning. It can be shown that the RIN increases as temperature increases, while it decreases as ASE injection power increases. As we increase the ambient temperature from 25 Cto60 C, the RIN increases 8.47 and 0.86 db at the injection powers of and dbm, respectively. The corresponding injection spectral densities are dbm/0.2 nm and dbm/0.2 nm, respectively. For low saturation levels, the RIN shows a strong temperature dependence. However, this effect is reduced as the saturation level increases. At 60 C, the Fig. 9. Output power versus ASE power at different ambient temperature. strong dependence of the RIN on injection power suggests other mechanisms may be involved, as will be discussed shortly. The observed RIN reduction and its dependence on the temperature can be explained based on our model. To do that, we calculated the slope in output power versus input power curve. The slope is a measure of the RIN reduction, as shown in Fig. 9, in which the simulation parameters are the same as those in Fig. 7. In the linear region, there is no RIN reduction. The RIN can be reduced in the nonlinear region due to the gain saturation [12]. The RIN reduction can be calculated as (14) where the and are the RIN of the input ASE and the output laser, respectively. is the slope in output power versus input power curve (see Fig. 9). is the fiber-to-fiber gain of the injection ASE. It may be noted that (14) is valid when the side mode suppression ratio is sufficiently high. Fig. 10 shows the calculated and experimental RIN reduction as a function of the injection ASE power at different temperatures. The RIN reduction increases as the ASE injection power increases. The RIN reduction decreases as the temperature increases at the same ASE injection power due to the high saturation power.

244 JOURNAL OF LIGHTWAVE TECHNOLOGY, VOL. 28, NO. 3, FEBRUARY 1, 2010 respectively. For comparison, we show a spectrum at the temperature of 40 C with dbm ASE-injection power in the inset of Fig. 11. So the RIN reduction calculated by using (14) is not valid, since saturation characteristics in Fig. 9 are calculated only for the output power at the injection wavelength. Simulated RIN values at 60 C for low injection powers below dbm are not shown, due to the difficulty in obtaining a converged solution as the denominator in (1) becomes very small. Fig. 10. 1RIN versus ASE power at different ambient temperature. IV. DISCUSSION AND CONCLUSION We investigated performance of F-P LD/RSOA for fiber-tofiber gain of less than 20 db. It gives us about 10 db saturated gain. Then, we can achieve output power of dbm, which may be good enough for 20 km transmission. It may be noted again that the operation of F-P LD/RSOA in saturation region is highly recommended to reduce the RIN. In this case, the spectral ripple of the F-P LD diminishes considerably and differences between F-P LD and RSOA are not significant as in linear operation cases. We have investigated saturation and noise suppression characteristics of the injection seeded F-P LDs/RSOAs based on nonlinear etalon theory. With the detailed recombination mechanisms, we have analyzed the effect of the temperature on saturation power. The saturation power increases as we increase the operating temperature or decrease the front facet reflectivity. The higher saturation power at a given injection power implies less saturation effect and less noise suppression. It may be noticed that the decrease of front facet reflectivity will increase the injection current and the saturation power. Thus, better thermal design may be needed. The situation becomes more serious when the internal loss increases as the structure is modified (e.g., stripe tilt) to reduce the front facet reflectivity. Fig. 11. Output spectrum at the temperature of 60 C and 40 C (inset) with 014:7 dbm ASE injection. The dependence of the RIN on the injection power is more pronounced at higher ambient temperature. At temperatures of 25 C, 40 C, and 60 C, the experimental RIN reduction values are 5.8, 6.9, and 13.4 db, respectively, over the injection power range of to dbm. As seen in Fig. 10, there is general agreement between the simulation results and the experimental results, except for a discrepancy at the temperature of 60 C. This can be explained as a large offset between the injection wavelength and the gain peak of the lasing spectrum. When the injection wavelength is near the gain peak, the side-mode suppression ratio is high enough. Then, the most output of the F-P LD is at the injection wavelength. However, the situation becomes different, when the injection wavelength is far from the gain peak. In this case, the output of F-P LD has two dominant peaks, as shown in Fig. 11. One peak is at the injection wavelength and the other is on the gain peak. For example, the power at the injection wavelength and at the gain peak were and dbm at ASE injection power of dbm at the temperature of 60 C, while they were and, and dbm at the temperatures of 25 C and 40 C, REFERENCES [1] H. D. Kim, S. G. Kang, and C. H. Lee, A low-cost WDM source with an ASE injected Fabry-Perot semiconductor laser, IEEE Photon. Technol. Lett., vol. 12, no. 8, pp. 1067 1069, Aug. 2000. [2] S. M. Lee, K. M. Choi, S. G. Mun, J. H. Moon, and C. H. Lee, Dense WDM-PON based on wavelength-locked Fabry-Pérot laser diodes, IEEE Photon. Technol. Lett., vol. 17, no. 7, pp. 1579 1581, Jul. 2005. [3] K. M. Choi, J. S. Baik, and C. H. Lee, Color-free operation of dense WDM-PON based on the wavelength-locked Fabry-Pérot laser diodes injecting a low-noise BLS, IEEE Photon. Technol. Lett., vol. 18, no. 10, pp. 1167 1169, Mar. 2006. [4] X. F. Cheng, Y. J. Wen, Y. Dong, Z. W. Xu, X. Shao, Y. X. Wang, and C. Lu, Optimization of spectrum-sliced ASE source for injectionlocking a Fabry-Pérot laser diode, IEEE Photon. Technol. Lett., vol. 18, no. 18, pp. 1961 1963, Sep. 2006. [5] S. M. Lee, M. H. Kim, and C. H. Lee, Demonstration of a bidirectional 80-km-reach DWDM-PON with 8-Gb/s capacity, IEEE Photon. Technol. Lett., vol. 19, no. 6, pp. 405 407, Mar. 2007. [6] K. Y. Park and C. H. Lee, Intensity noise in a wavelength-locked Fabry-Perot laser diode to a spectrum sliced ASE, IEEE J. Quantum Electron., vol. 44, no. 3, pp. 209 215, Mar. 2008. [7] K. Y. Park, S. G. Mun, K. M. Choi, and C. H. Lee, A theoretical model of a wavelength-locked Fabry-Pérot laser diode to the externally injected narrow-band ASE, IEEE Photon. Technol. Lett., vol. 17, no. 9, pp. 1797 1799, Sep. 2005. [8] M. J. O Mahony, Semiconductor laser optical amplifiers for use in future fiber systems, J. Lightw. Technol., vol. 6, no. 4, pp. 531 544, Apr. 1988. [9] J. W. Wang, H. N. Olesen, and K. E. Stubkjaer, Recombination, gain and bandwidth characteristics of 1.3-m semiconductor laser amplifiers, J. Lightw. Technol., vol. LT-5, no. 1, pp. 184 189, Jan. 1987.

MENG et al.: FABRY PÉROT LASER DIODES/REFLECTIVE OPTICAL AMPLIFIERS 245 [10] L. I. Burov, V. P. Gribkovskii, P. S. Grigelevich, M. I. Kramar, G. I. Ryabtsev, K. A. Shore, S. V. Voitikov, and R. Kragler, Theoretical analysis of the effect of amplified luminescence on the modulation response of laser diodes, Int. J. Numer. Model., vol. 14, pp. 331 343, 2001. [11] T. H. Gfroerer, L. P. Priestley, M. F. Fairley, and M. W. Wanlass, Temperature dependence of nonradiative recombination in low-band gap In Ga As/InAs P double heterostructures grown on InP substrates, J. Appl. Phys., vol. 94, no. 3, pp. 1738 1743, Aug. 2003. [12] T. Yamatoya and F. Koyama, Noise suppression of spectrum-sliced light using semiconductor optical amplifiers, Electron. Commun. Jpn. Part 2, vol. 86, no. 2, pp. 417 423, Jun. 2002. Hongyun Meng was born in 1973. He received the M.S. and Ph.D. degrees from the Institute of Modern Optics, Nankai University, Tianjin, China, in 2000 and 2003, respectively. From 2003 to 2005, he was with South China Normal University, Guangzhou, China. From 2007 to 2008, he was with Korea Advanced Institute of Science and Technology, Daejeon, Korea, as a Postdoctoral Fellow. Since 2005, he has been an Associate Professor at South China Normal University, where he is currently with the Laboratory of Photonic Information Technology. His research interests include optical amplifiers, optical lasers, and fiber sensors. Jung-Hyung Moon was born in 1980. He received the M.S. degree in electrical engineering from Korea Advanced Institute of Science and Technology (KAIST), Daejeon, Korea, in 2005. He is currently working toward the Ph.D. degree in optical communication systems at KAIST. His current research interests include lightwave systems and optical access network based on wavelength-division-multiplexing passive optical network. Ki-Man Choi was born in 1978. She received the M.S. and Ph.D. degrees from the Department of Electrical Engineering, Korea Advanced Institute of Science and Technology (KAIST), Daejeon, Korea, in 2004 and 2008, respectively. She was a Postdoctoral Fellow at KAIST, for six months. She is currently a Researcher with the Network Infra Research Department, Korea Telecom Network Technology Laboratory, Daejeon, Korea. Her research interests include optical communication and network, wavelength-division-multiplexing passive optical network, and network management. Chang-Hee Lee (S 82 M 90 SM 07) was born in 1961. He received the M.S. and Ph.D. degrees from the Department of Electrical Engineering, Korea Advanced Institute of Science and Technology (KAIST), Daejon, Korea, in 1983 and 1989, respectively. He was with Bellcore (Bell Communications Research), as a Postdoctoral Fellow, for a year. From 1989 to 1997, he was with the Electronics and Telecommunications Research Institute, as a Senior Researcher. Since 1997, he has been a Professor with KAIST, where he is currently with the Department of Electrical Engineering. He is the author or coauthor of more than 186 journal and conference papers. He is the holder of 24 U.S. patents, and more than 40 additional U.S. patents are pending. He has spent more than 20 years as an Engineer in the area of optical communications, including semiconductor lasers. He was a Technical Leader for 2.5- and 10-Gb/s optical-transmission-system development, including optical amplifiers at the Electronics and Telecommunications Research Institute. His research interest include optical communications and networks. Prof. Lee is a member of the Optical Society of America.