Open Access
Add to BibSonomyAbstract
The modulation bandwidth enhancement of distributed reflector (DR) lasers with wirelike active regions utilizing optical injection locking is demonstrated both theoretically and experimentally. By the rate equation analysis, it is shown that DR lasers with wirelike active regions realize a low optical injection power and a large bandwidth enhancement under small operation currents. Experimentally, the small-signal bandwidth is increased to >15 GHz at a bias current of 5 mA, which is 4 times smaller than that for conventional edge-emitting lasers. A large signal modulation at 10 Gbps is also performed at the same bias current of 5 mA and voltage swing of 0.4 V pp, and error-free detection was confirmed under the low-power conditions.
©2010 Optical Society of America
1. Introduction
Owing to the explosive growth of data communications, the demand for high-bit-rate access networks has risen very rapidly. Furthermore, high-speed optical interconnects such as active optical cables for dealing with large data capacity have attracted much interest. For such short-distance applications, low power consumption is an essential feature for optical transmitters. In terms of low-power semiconductor lasers, studies of several types of lasers with a low threshold current have been reported, such as vertical-cavity surface-emitting lasers (VCSELs) [1–3], and distributed-feedback (DFB) lasers with a high index-coupling coefficient [4], [5]. A low threshold current in the sub-mA range has been successfully demonstrated in these semiconductor lasers. Lasers based on high-index-contrast waveguides, such as microdisk lasers [6], photonic crystal lasers [7], and membrane lasers [8], [9], are also candidates for the next-generation lasers required for ultralow-power-consumption optical wiring in next generation LSIs and optical interconnects; however, their light output power is not practical for optical fiber communications. A distributed reflector (DR) laser [10], consisting of DFB and distributed Bragg reflector (DBR) sections with wirelike active regions [11], can also be operated with a low threshold current and a high efficiency because of its small active volume and strong index-coupling [11], [12]. Experimentally, sub-mA threshold current operations of DR lasers have been reported [13], [14] and high external differential quantum efficiency from the front facet of approximately 50% with sub-mA threshold current has been demonstrated [15].
Another important issue is fulfilling the demands for high bandwidth without sacrificing power consumption and cost. Traditional direct modulation schemes have potential risk because a high bandwidth requires high laser operation currents. One promising solution is the use of optical injection locking (OIL) techniques [16], which has several unique advantages such as modulation bandwidth enhancement [16], [17], relative intensity noise reduction [18], and chirp-managed transmission [19,20] with a simple combination of typical optical components.
Although OIL has been actively demonstrated in typical semiconductor lasers such as Fabry-Perot (FP) lasers, DFB lasers, and vertical-cavity surface-emitting lasers (VCSELs), the best OIL schemes for access applications are still under investigation. OIL-FP lasers have a cost advantage but their bit rates for OIL-FP applications are typically below 2.5 Gbps [21]. On the other hand, OIL-DFB lasers have shown an ultrahigh resonance frequency exceeding 100 GHz [17]; however, its 3-dB bandwidth is relatively small. The high threshold and bias currents required for injection locking lead to severe power consumption problems. OIL-VCSELs are widely studied and very attractive in terms of their power consumption features, but their introduction into real-world access networks remains uncertain owing to their insufficient optical output power compared with that of edge emitting lasers.
Because DR lasers with wirelike active regions have a sub-mA range of threshold current as well as a high output power sufficient for module standards such as IEEE and ITU, they can resolve not only the issue on compatibility with the current edge-emitter-based access network but also cost and power consumption issues. In addition, the OIL of DR lasers can be realized with a lower optical injection power than that of the conventional DFB lasers owing to the structural features of wirelike active regions. Therefore, the DR laser with wirelike active regions can be a promising candidate for OIL access network systems. In this paper, we theoretically and experimentally demonstrate the OIL of DR lasers with wirelike active regions for the first time. We also show the concept of DR lasers with wirelike active regions and their theoretical superiority to the conventional DFB lasers. The experimental results are focused on bandwidth enhancement at a small operation current for the cost-effective access networks.
2. Theoretical analysis
Figure 1 shows the schematics of the DR laser with a double-channel high-mesa stripe surrounded by benzocyclobutene (BCB). The DR laser consists of an active DFB section with wirelike active regions of wire width W a and a passive DBR section with active regions of wire width W p. Although conventional index-coupled DFB lasers possess two lasing modes, DR lasers with wirelike active regions always oscillate on the longer-wavelength side of the stopband owing to the gain matching effect [12]. Therefore, the period in the DBR section (Λ p) is slightly longer than that in the DFB section (Λ a) in order to match the Bragg wavelength of the DBR section with the lasing wavelength. The wire width in the DBR section (W p) is smaller than that in the DFB section (W a) so that the absorption in the DBR section is suppressed owing to the lateral quantum confinement effect [11]. As a result, a high reflectivity can be obtained simply by modulating the width of the wirelike active regions without additional epitaxial growth.
Fig. 1 (a) Schematic diagram of DR laser. (b) Cross-sectional structures of wirelike active regions.
For the theoretical analysis of optical injection locked lasers, the use of the rate equation model is most effective method and most commonly used. We analyzed the OIL characteristics of DR lasers with wirelike active regions using the following three rate equations [22], [23] for the DR laser’s field amplitude A(t), the field phase difference between the DR laser and the master laser ϕ(t), and carrier numbers in the active regions of the DR laser N(t):
where g is the differential linear gain coefficient, N th the threshold carrier number, α the linewidth enhancement factor, J the number of injection carriers, γ N the carrier recombination rate, and γ P the photon decay rate, respectively. The injection terms κ, A inj(t), and Δω inj are the coupling coefficient, injected field amplitude, and detuning frequency between the master laser, and the free-running DR laser, respectively.The main differences between DR lasers with wirelike active regions and conventional DFB lasers are as follows. First, the DR lasers have a high coupling coefficient κ so that a lower optical injection power can be obtained. Since the DR lasers have a high index-coupling coefficient κ i > 300cm−1, very short cavities from 80 μm to 200 μm long are realized experimentally. Because κ is inversely proportional to the cavity roundtrip time τ rt (κ = 1/τ rt), the injected field is effectively coupled to a DR laser. It can be physically understood that the injected field amplitude is coupled to the slave laser field every time the slave field “hits” an injection facet [24]. Another feature is the small modulation response drop-off below resonance frequency. Although a high resonance frequency can be obtained with large Δω inj, the flatness of the modulation response strongly depends on slave laser bias current. As Lau et al. previously indicated [23], the modulation response drop-off is dominated by photon density inside the laser cavity and can be overcome by increasing bias current. Bias currents several times of the threshold current or several tens of mA are necessary so that the large power dissipation, which is a severe problem with OIL-DFB lasers can be eliminated. In contrast, DR lasers with wirelike active regions experience minimal drop-off owing to their high photon density. Assuming the same cavity Q and output power, a shorter cavity laser with a higher photon density is realized. Also, the longitudinal optical confinement is higher than a filling factor of wirelike active regions owing to the global effect [25]. The enhanced optical confinement enables a 1.5 fold higher photon density. Therefore, a high bandwidth under small bias currents can be achieved for DR lasers with wirelike active regions.
Figure 2 shows the comparison of the calculated modulation bandwidths of the OIL-DR lasers with wirelike active regions and the OIL-DFB laser with a λ/4 phase shift for various wavelength detuning conditions (Δλ = Δλ locked –Δλ free-running). Each laser parameter was chosen according to their typical structure; the values are listed in Table 1 . Assuming low-power OIL laser modules, the bias current of the DR laser was set to 3 mA above the threshold and a low optical injection condition of A inj = 0.5 A free-running was used for the calculations. For the DFB laser, the bias current was set to 11 mA, which provided the same relaxation oscillation frequency with the DR laser in the free-running case. Figure 2 clearly shows that the OIL-DR lasers have a higher resonance frequency, which means a higher effective optical injection power. This was already discussed in terms of the short roundtrip time and the high κ. Even with the same resonance frequency f r, for example, Δλ = 0.08 nm in Fig. 2(a) and Δλ = −0.08 nm in Fig. 2(b), the OIL-DR lasers show a relatively flat modulation response, whereas the conventional DFB lasers show the response drop-off. From this result, the small response drop-off of the OIL-DR lasers is successfully confirmed. As a result, the bandwidth enhancement of OIL-DR lasers exceeding 20 GHz at a very small net bias current of 3 mA, which cannot be realized by normal direct modulation and/or OIL using conventional DFB lasers, has been successfully demonstrated.
Fig. 2 Calculated modulation bandwidth of (a) OIL-DR lasers and (b) conventional OIL-DFB lasers under bias current of 3 mA and injection of A inj = 0.5 A free-running.
Table 1. Laser structures and parameters
3. Fabrication process
A DR laser was fabricated by electron beam lithography (EBL), CH4/H2 reactive ion etching (RIE), and embedding growth by organometallic vapor phase epitaxy (OMVPE) [12]. The initial wafer consisted of a p-InP cladding layer, an undoped-GaInAsP optical confinement layer (OCL), 1% compressively strained (CS) Ga0.22In0.78As0.81P0.19 double-quantum-well (DQW) layers (6 nm thick) with −0.15% tensile-strained (TS) Ga0.25In0.75As0.50P0.50 barrier layers (10 nm thick), and an upper undoped GaInAsP OCL grown on a p-InP substrate. Then, EBL was carried out to form the desired wirelike patterns, which were transferred to a SiO2 mask to etch away the DQW and OCLs by CH4/H2 RIE. After that, undoped InP was regrown in the groove regions at 600°C with a lower growth speed, and an upper OCL (180 nm thick), an n-InP cladding layer, and a 50-nm-thick n+-GaInAs contact layer were grown at 650°C. After the OMVPE growth, a high-mesa stripe structure was fabricated by a combination of wet chemical etching and CH4/H2 RIE. After spin-coating BCB to make the surface flat, it was etched back by CF4/O2 RIE to open a contact window. Finally, Ti/Au was evaporated onto the p- and n-side contact layers, and a lift-off process was carried out to form a contact pad on the n-side.
4. Experimental results
4.1 Laser characteristics and experimental setup
Figure 3 shows the current-light output power (I-L) characteristics of the free-running DR laser used in the OIL experiments. Under a room-temperature continuous-wave condition, a moderate low threshold current I th of 1.7 mA and a differential quantum efficiency from the front facet (η df) of 35% were obtained. The stable single mode operation with a side-mode suppression ratio (SMSR) of 43 dB was obtained at a bias current of 2I th, as shown in Fig. 4 . The index coupling coefficient κ i was estimated to be 470 cm−1.
Figure 5 shows the schematic of the experimental setup for the OIL. The laser output was coupled to a lensed fiber with a 2 dB coupling loss. As the master laser, the tunable laser SANTEC TSL-210H was used. The optical injection was carried out through an optical circulator. Note that the same side of the laser facet was used for light injection and OIL-DR laser light output extraction. A polarization controller (PC) was used to adjust the lasing polarization to maximize the optical injection effect under a given master laser output power to evaluate OIL properly. Although the injection ratio that dominates the dynamics of the OIL lasers is the internal injection ratio R inj,int ( = A inj/A free-running), R inj,int cannot be directly measured. Hereafter, we use the external injection ratio R inj,ext defined as the ratio of injecting light that reached the laser facet to the facet output power of a free-running DR laser. The conversion between the internal and external injection ratios can be easily derived from the roundtrip travelling model as
where α m is the mirror loss for the light injection facet, L the cavity length, and R the facet reflectivity for the light injection facet, respectively. The calculated α m from the structure was 5 cm−1. The experimental setup used for large-signal measurement is not shown here; however, the same injection method and master laser were used.4.2 Results of optical injection locked DR laser
Figure 6 shows the stable locking range at various R inj,ext values with a bias current of 5 mA. The locking range at an injection ratio of 13 dB was about 0.3 nm. The bandwidth enhancement of the OIL-DR lasers was confirmed from the S 21 responses of a network analyzer, as shown in Fig. 7 . The injection ratios in Figs. 7 (a) and (b) are 5 dB and 13 dB, respectively. The data were calibrated for cable loss, bias-T loss, and photodetector loss. However, it still includes laser parasitic loss. As described above, the polarization of the master laser light output was carefully controlled to maximize the bandwidth. The 3 dB bandwidth in the case of a free-running laser was 1.7 GHz, reflecting a small bias current of 5 mA. In contrast, the 3 dB bandwidth was successfully enhanced to 7.7 GHz and 15.5 GHz at maximum at optical injection ratios of 5 dB and 13 dB, respectively. A 3 dB bandwidth of more than 15 GHz at a bias current of 5 mA was unobtainable in conventional or OIL edge-emitting lasers [20], [26]. Bias current was successfully reduced by a factor of more than 4 in DR lasers with wirelike active regions; previously reported works required bias currents in the 20-50 mA range. The bandwidth enhancement is modest but quite promising, given that the bias current is relatively low and the cleaved facet reflects a significant amount of injected power (30% reflectivity). Further bandwidth enhancement can be obtained if injection ratio or bias current is increased.
Figure 8 shows the results of the bit-error rate test using directly modulated 10 Gbps non-return-to-zero (NRZ) 27-1 pseudorandom bit sequence (PRBS) signals. The bias current was 5 mA and the voltage swing was 0.4 V pp. The corresponding extinction ratio was 3 dB. An injection ratio of 5 dB was used since this gives sufficient bandwidth for 10 Gbps and a higher injection ratio reduces the extinction ratio owing to the interference between master laser light reflected from slave laser facet and the slave laser output. This is the case if the same port is used for injection locking and slave laser output [27]. Figure 8 shows that error-free detection was achieved. The average power received for error-free detection was −3 dBm, which was limited by the excessive noise of the electrical amplifier inserted between the photodetector and the error detector. The inset of Fig. 8 shows the corresponding eye diagrams. With help of the OIL, successful eye opening at a bias current of 5 mA was observed. The noise at the 0 or 1 level partly came from the poor noise characteristics of the master laser. The use of wavelength filters or other master lasers will improve eye diagrams. For higher extinction ratios, devices with higher efficiencies will be more effective. Also, the optimization of the injection ratio and detuning wavelength at a given modulation speed may lead to optical modulation signals of better quality.
5. Conclusion
We demonstrated the optical injection locking of a distributed reflector laser with wirelike active regions both theoretically and experimentally. Owing to the low threshold current and wirelike structural features of DR lasers, a 3 dB bandwidth of more than 15 GHz was achieved with a small bias current of 5 mA at an injection ratio of 13 dB. Also, error-free 10 Gbps detection was achieved with a bias current of 5 mA and a small voltage swing of 0.4 V pp. In conclusion, a DR laser with wirelike active regions is a promising candidate for future optical-injection-locking-based low-cost high-bit-rate access network systems.
Acknowledgments
We thank Professors Emeritus Y. Suematsu, K. Iga, and K. Furuya for their continuous encouragement, and Professors M. Asada, F. Koyama, K. Kobayashi, T. Mizumoto, Y. Miyamoto, M. Watanabe, T. Miyamoto, and H. Uenohara, Tokyo Institute of Technology, for fruitful discussions. This research was supported by a Grant-in-Aid for Scientific Research from the Ministry of Education, Culture, Sports, Science and Technology (MEXT) Japan, (# 19002009, #19686023, and #21008757), and the US DoD National Security Science and Engineering Faculty Fellowship. SeungHun Lee acknowledges the Japan Society for the Promotion of Science (JSPS) for the Research Fellowship for Young Scientists.
References and links
1. Y. Hayashi, T. Mukaihara, N. Hatori, N. Ohnoki, A. Matsutani, F. Koyama, and K. Iga, “Record low-threshold index-guided InGaAs/GaAlAs vertical-cavity surface-emitting laser with a native oxide confinement structure,” Electron. Lett. 31(7), 560–562 (1995). [CrossRef]
2. N. Nishiyama, C. Caneau, G. Guryanov, X. S. Liu, M. Hu, and C. E. Zah, “High efficiency long wavelength VCSEL on InP grown by MOCVD,” Electron. Lett. 39(5), 437–439 (2003). [CrossRef]
3. N. Nishiyama, C. Caneau, B. Hall, G. Guryanov, M. Hu, X. Liu, M. J. Li, R. Bhat, and C.-E. Zah, “Long-wavelength vertical-cavity surface-emitting lasers on InP with lattice matched AlGaInAs–InP DBR grown by MOCVD,” IEEE J. Sel. Top. Quantum Electron. 11(5), 990–998 (2005). [CrossRef]
4. N. Nunoya, M. Nakamura, H. Yasumoto, M. Morshed, K. Fukuda, S. Tamura, and S. Arai, “Sub-milliampere operation of 1.55 μm wavelength high index-coupled buried heterostructure distributed feedback lasers,” Electron. Lett. 36(14), 1213–1214 (2000). [CrossRef]
5. N. Nunoya, M. Nakamura, M. Morshed, S. Tamura, and S. Arai, “High-performance 1.55-μm wavelength GaInAsP-InP distributed-feedback lasers with wirelike active regions,” IEEE J. Sel. Top. Quantum Electron. 7(2), 249–258 (2001). [CrossRef]
6. M. Fujita, R. Ushigome, and T. Baba, “Continuous wave lasing in GaInAsP microdisk injection laser with threshold current of 40 μA,” Electron. Lett. 36(9), 790–791 (2000). [CrossRef]
7. H.-G. Park, S.-H. Kim, S.-H. Kwon, Y.-G. Ju, J.-K. Yang, J.-H. Baek, S.-B. Kim, and Y.-H. Lee, “Electrically driven single-cell photonic crystal laser,” Science 305(5689), 1444–1447 (2004). [CrossRef] [PubMed]
8. T. Okamoto, N. Nunoya, Y. Onodera, T. Yamazaki, S. Tamura, and S. Arai, “Optically pumped membrane BH-DFB lasers for low-threshold and single-mode operation,” IEEE J. Sel. Top. Quantum Electron. 9(5), 1361–1366 (2003). [CrossRef]
9. S. Sakamoto, H. Naitoh, M. Ohtake, Y. Nishimoto, S. Tamura, T. Maruyama, N. Nishiyama, and S. Arai, “Strongly index-coupled membrane BH-DFB lasers with surface corrugation grating,” IEEE J. Sel. Top. Quantum Electron. 13(5), 1135–1141 (2007). [CrossRef]
10. J. I. Shim, K. Komori, S. Arai, I. Arima, Y. Suematsu, and R. Somchai, “Lasing characteristics of 1.5 μm GaInAsP-InP SCH-BIG-DR lasers,” IEEE J. Quantum Electron. 27(6), 1736–1745 (1991). [CrossRef]
11. K. Ohira, T. Murayama, H. Yagi, S. Tamura, and S. Arai, “Distributed reflector laser integrated with active and passive grating sections using lateral quantum confinement effect,” Jpn. J. Appl. Phys. 42(Part 2, No. 8A8A), L921–L923 (2003). [CrossRef]
12. K. Ohira, T. Murayama, S. Tamura, and S. Arai, “Low-threshold and high-efficiency operation of distributed reflector lasers with width-modulated wirelike active regions,” IEEE J. Sel. Top. Quantum Electron. 11(5), 1162–1168 (2005). [CrossRef]
13. S. M. Ullah, R. Suemitsu, S. Lee, M. Otake, N. Nishiyama, and S. Arai, “Low-threshold-current operation of high-mesa stripe distributed reflector laser emitting at 1540 nm,” Jpn. J. Appl. Phys. 46(44), L1068–L1070 (2007). [CrossRef]
14. S. M. Ullah, S. Lee, R. Suemitsu, N. Nishiyama, and S. Arai, “GaInAsP/InP distributed reflector lasers and integration of front power monitor by using lateral quantum confinement effect,” Jpn. J. Appl. Phys. 47(6), 4558–4565 (2008). [CrossRef]
15. T. Shindo, S. Lee, D. Takahashi, N. Tajima, N. Nishiyama, and S. Arai, “Low-threshold and high-efficiency operation of distributed reflector laser with wirelike active regions,” IEEE Photon. Technol. Lett. 21(19), 1414–1416 (2009). [CrossRef]
16. L. Chrostowski, X. Zhao, and C. J. Chang-Hasnain, “Microwave performance of optically injection-locked VCSELs,” IEEE Trans. Microw. Theory Tech. 54(2), 788–796 (2006). [CrossRef]
17. E. K. Lau, X. Zhao, H.-K. Sung, D. Parekh, C. Chang-Hasnain, and M. C. Wu, “Strong optical injection-locked semiconductor lasers demonstrating > 100-GHz resonance frequencies and 80-GHz intrinsic bandwidths,” Opt. Express 16(9), 6609–6618 (2008). [CrossRef] [PubMed]
18. L. Chrostowski, C.-H. Chang, and C. Chang-Hasnain, “Reduction of relative intensity noise and improvement of spur-free dynamic range of an injection locked VCSEL,” Proc. IEEE LEOS Annu. Meeting Conf. 2, 706–707 (2003).
19. X. Zhao, B. Zhang, L. Christen, D. Parekh, W. Hofmann, M. C. Amann, F. Koyama, A. E. Willner, and C. J. Chang-Hasnain, “Greatly increased fiber transmission distance with an optically injection-locked vertical-cavity surface-emitting laser,” Opt. Express 17(16), 13785–13791 (2009). [CrossRef] [PubMed]
20. D. Parekh, B. Zhang, X. Zhao, Y. Yu, W. Hofmann, M. C. Amann, A. E. Willner, and C. J. Chang-Hasnain, “90-km Single-Mode Fiber Transmission of 10-Gb/s Multimode VCSELs under Optical Injection Locking, ” in Proc. OFC/NFOEC, 2009, paper OTuK7.
21. Q. T. Nguyen, L. Bramerie, G. Girault, O. Vaudel, P. Besnard, J.-C. Simon, A. Shen, G.-H. Duan, and C. Kazmierski, “16x2.5 Gbit/s Downstream Transmission in Colorless WDM-PON based on Injection-Locked Fabry-Perot Laser Diode using a single Quantum Dash mode-locked Fabry-Perot laser as multi-wavelength seeding source,” in Proc. OFC/NFOEC, 2009, paper OThA3.
22. A. Murakami, K. Kawashima, and K. Atsuki, “Cavity resonance shift and bandwidth enhancement in semiconductor lasers with strong light injection,” IEEE J. Quantum Electron. 39(10), 1196–1204 (2003). [CrossRef]
23. E. K. Lau, H.-K. Sung, and M. C. Wu, “Frequency response enhancement of optical injection-locked lasers,” IEEE J. Quantum Electron. 44(1), 90–99 (2008). [CrossRef]
24. R. Lang and K. Kobayashi, “External optical feedback effects on semiconductor injection laser properties,” IEEE J. Quantum Electron. 16(3), 347–355 (1980). [CrossRef]
25. A. Champagne, R. Maciejko, D. M. Adams, G. Pakulski, B. Takasaki, and T. Makino, “Global and local effects in gain-coupled multiple-quantum-well DFB lasers,” IEEE J. Quantum Electron. 35(10), 1390–1401 (1999). [CrossRef]
26. N. B. Terry, N. A. Naderi, M. Pochet, A. J. Moscho, L. F. Lester, and V. Kovanis, “Bandwidth enhancement of injection-locked 1.3 μm quantum-dot DFB laser,” Electron. Lett. 44(15), 904–905 (2008). [CrossRef]
27. W. Yang, P. Guo, D. Parekh, W. Hofmann, M. C. Amann, and C. J. Chang-Hasnain, “Physical Origin of Data Pattern Inversion in Optical Injection-Locked VCSELs,” in Frontiers in Optics, OSA Technical Digest Series, (Optical Society of American, 2009), paper FTuW2.
