We report a cost-effective transmitter optical sub-assembly using a monolithic four-wavelength vertical-cavity surface-emitting laser (VCSEL) array with 100-GHz wavelength spacing for future-proof mobile fronthaul transport using the data rate of common public radio interface option 6. The wavelength spacing is achieved using selectively etched cavity control layers and fine current adjustment. The differences in operating current and output power for maintaining the wavelength spacing of four VCSELs are <1.4 mA and <1 dB, respectively. Stable operation performance without mode hopping is observed, and error-free transmission under direct modulation is demonstrated over a 20-km single-mode fiber without any dispersion-compensation techniques.
© 2015 Optical Society of America
To accommodate the rapidly increasing mobile internet traffic, long-term evolution (LTE) and LTE-advanced (LTE-A) services started to be commercially deployed worldwide [1–3]. In 4G wireless systems, the separated baseband unit (BBU) and remote radio head (RRH) structure has been commonly utilized to reduce the size of an RRH, which is deployed outside. The compact RRH and separated BBU has several advantages such as low capital expenditure (CAPEX), low operational expenditure (OPEX), simple replacement of each component, and simultaneous operation of multiple RRHs using one BBU. Between BBU and RRH, a common public radio interface (CPRI) has been widely used for delivering baseband I/Q samples. The CPRI has many speed options as specified in the CPRI specification, and its required data bandwidth has continuously increased with the evolution of mobile communication technologies [4,5].
To save expenditure and bandwidth usage further, the CPRI should be operated in the wavelength division multiplexing (WDM) scheme with cost-effective light sources [5–11]. Many semiconductor laser sources have been examined for the purpose, including externally modulated lasers (EMLs), distributed Bragg reflectors (DBRs), and distributed feedback (DFB) lasers. However, these laser sources have some disadvantages such as high cost, high power consumption, mode hopping, and/or incompatibility of multi-wavelength lasing with the telecommunication standardization sector of the international telecommunications union (ITU-T) grid. When a vertical-cavity surface-emitting laser (VCSEL) array is used for a multi-wavelength light source, many of the limitations are overcome by its advantages, which include easy mass production, high energy efficiency, and stable single-longitudinal mode operation. However, it is still challenging to manufacture a multi-wavelength VCSEL array, and many experimental attempts were made, which include the use of a linear gradient layer, masked molecular beam epitaxy (MBE) for selective area growth, metal-organic chemical vapor deposition (MOCVD) on a non-planar substrate, and selective area MOCVD growth [12–15]. The results indicated a lasing wavelength of 980 nm, and the wavelength spacing was too wide to use in dense-WDM (DWDM) application. Recently, a current-controlled multi-wavelength VCSEL array operating in the telecommunication DWDM band was reported . However, ITU-T-grid-compatible multi-wavelength lasing was obtained using a relatively wide current-tuning range, and it resulted in a wide difference in operation bandwidth.
In this paper, we report a monolithic four-wavelength VCSEL array transmitter optical sub-assembly (TOSA) with 100-GHz wavelength spacing that uses a commercially available transistor outline (TO)-56-can packaging. The 100-GHz-spaced four-wavelength VCSEL array was realized using fine current tuning, in which the difference in operating currents was smaller than that of the previously reported one by a factor of three . To minimize the difference, we employed a selective etching technique using cavity control layers in the VCSEL growth process. The proposed TOSA is applicable to CPRI option 6.
2. Fabrication and experimental setup
To obtain four different wavelengths from a single VCSEL chip, we employed four pairs of cavity control layers and used the selective etching technique [17, 18]. One pair of cavity control layers consists of InP and InAlGaAs layers. InP is etched well in 5HCl:6H2O etchant, although InAlGaAs in the same etchant acts as an etch stop layer. On the other hand, InAlGaAs is etched well in H2PO4:H2O2:40H2O etchant, and InP acts as an etch stop layer in this etchant. By utilizing these selective etch characteristics, we selectively etched the cavity control layers to achieve VCSELs of four different wavelengths. The detailed fabrication process is as follows. After growing cavity control layers, two pairs of cavity control layers are removed using two iterations of binary-coded selective wet etching (selective etching-1 in Fig. 1). Four different cavity lengths of VCSEL could be obtained by applying binary-coded selective wet etching to one pair of cavity control layers positioned at different areas (selective etching-2 in Fig. 1). The process is completed using electrode deposition after regrowth of the top DBR mirror and cavity etching.
2.2 Link topology and experimental setup
Figure 2 shows one of topologies of a centralized point-to-multipoint radio access network (C-RAN) optical link based on WDM technology . In the link, a data rate of 6.144 Gb/s for each wavelength with a total capacity of 24.568 Gb/s could be supplied by one proposed TOSA for delivering baseband I/Q samples. The data rate option is defined by CPRI options 1–8 and determined by the wireless technology, which includes multiple-input multiple-output (MIMO) antennas and multi-carrier and carrier-aggregation networks. The C-RAN topology shown in Fig. 2 would be of LTE-A when carrier aggregation is utilized for increasing data rate and a higher CPRI option is needed.
Our experimental set-up, derived from a CPRI and LTE-A, is composed of a multiplexer (MUX) and demultiplexer (DeMUX) with a 20-km single-mode fiber (SMF), as shown in Fig. 3. To determine the physical reach, we considered CPRI and LTE-A specifications. A physical reach greater than 10 km was defined in the CPRI specification, considering a round trip delay between the BBU and RRH . In the case of LTE-A, a reach of 40 km was calculated from a delay of 400 μs, assuming that no equipment existed between the BBU and RRH and that the group velocity in a fiber was ~5 μs/km . As a result, we assumed a physical reach of 20 km, which is between the reach of CPRI and the reach of LTE-A, by referring to a real application of wireless networks.
We employed a 32-port arrayed waveguide grating (AWG) as a MUX and DeMUX with 100-GHz wavelength spacing for using the C-RAN optical link with WDM technology, as shown in Fig. 2. The optical link could support 8 base stations employing 8 proposed TOSAs. Because the structure of the C-RAN optical link and base station are not limited by a specific topology and wireless technology, a service provider could be chosen in accordance with present conditions such as the use of a dark fiber, frequency, and bandwidth.
3. Results and discussion
3.1 Four wavelengths with 100-GHz spacing
Fig. 4 shows the proposed four-wavelength TOSA and its evaluation board. A commercially available 8-pin TO-56 stem is employed for mounting a VCSEL array, thermistor, and thermoelectric cooler (TEC). Four pins are assigned for delivering electrical signals, and other pins are utilized for grounding, temperature monitoring, and the operation of the TEC. We measured L–I curves and wavelength variations as a function of modulated bias current to investigate the applicability to a C-RAN optical link, as shown in Fig. 2. We also directly modulated each VCSEL during the measurements to determine the lasing wavelength corresponding to the wavelength of the transmission experiment. A bias current from a power supply and a modulation signal from a pulse pattern generator (PPG) are injected into the VCSEL via a flexible printed circuit board (FPCB) after that are combined using a bias tee. The output power and wavelength are simultaneously measured using an optical splitter placed after an isolator. This measuring scheme could provide the exact value of power and wavelength at the same current. The temperature of the VCSEL array is kept constant using an externally positioned TEC controller with a thermistor and TEC inside the TOSA package.
Figure 5(a) shows L–I curves of the VCSEL array under direct modulation. The threshold current of 1.6 mA is identical for all four VCSELs. However, a kink in the L-I characteristics is observed for ch01 at an operating current of approximately 7 mA. Because the power supply in Fig. 3 shows a relatively wide resolution of 0.5 mA for current, we controlled the voltage instead of the current to inject a current. Even though the voltage resolution of 0.5 mV is sufficient for measuring the threshold current of our VCSEL, there still exists a measurement error in current when a computer reads the current value. We believe that the kink is possibly due to a measuring error that can be avoided by replacing the power supply.
Figure 5(b) shows the lasing wavelength as a function of injection current. Four different lasing wavelengths are successfully obtained from a single VCSEL array chip. The wavelength spacing between ch01 and ch04 is 195 GHz at the same current, and the spacing corresponds to a 65-GHz spacing between adjacent VCSELs on average. A slightly larger wavelength change is observed at a higher injection current in Fig. 5(b). This result coincided with the relatively lower slope efficiency at higher injection current, as shown in Fig. 5(a). We believe this is because the non-radiative recombination process creates a temperature difference between the active layer and the temperature monitoring point.
To adapt a laser to DWDM application, a stable operation without mode hopping is essential to avoid unwanted power reduction and performance degradation caused by the filtering of MUX and DeMUX. Unlike edge-emitting lasers such as Fabry–Perot (FP) lasers and DFB lasers, a VCSEL usually operates with a single longitudinal mode because of its extremely short cavity length . A side-mode peak beside a strong main spectral peak, which is related to the single longitudinal mode, is observed at a wavelength separation of approximately 0.25 nm. A side-mode-related abrupt wavelength change was not observed throughout the operation range, as shown in Fig. 5(b).
To obtain a 100-GHz wavelength spacing corresponding to the standard ITU-T grid, we adjusted the lasing wavelength using fine current control of each VCSEL. The wavelength variation as a function of current is approximately 0.65 nm/mA at the operating current range from 4 mA to 8 mA. A 100-GHz wavelength spacing was successfully achieved within a 1.4 mA current difference between VCSELs, as shown in Fig. 6. Detailed information on wavelength spacing and the lasing wavelengths of the four VCSELs are given in Fig. 5(b).
In addition to operating wavelengths, the performance degradation of a VCSEL as a function of current should be considered. Because the output power is closely proportional to the injection current, a 3-dB down-cutoff frequency, ω3dB, would be enhanced by increasing the photon density or output power [20, 21]. Thus, the difference in optical power and operation current for each channel would be minimized to ensure uniform operation performance and convenient design of the DWDM optical link. The difference in operating current is only 1.4 mA, as mentioned above, and the output power difference is below 1 dB. We believe that these small differences between the operation currents and the output powers do not affect the design of an optical link system. The side-mode suppression ratio (SMSR) was greater than 30 dB for all VCSELs.
3.2 Transmission performances
We measured bit-error-rate (BER) curves to examine the transmission performance of the TOSA. A data rate of 6.144 Gb/s and a pseudorandom binary sequence (PRBS) pattern length of 27-1, similar to the 8B/10B line coding, are employed for CPRI option 6 application. An isolator is inserted between the TOSA and wavelength-division multiplexer, to prevent unwanted performance degradation. Even though the VCSEL DBR mirror has high reflectivity (>99%), it shows reflection-induced power penalty because the extremely short cavity length cancels the advantage of the high reflectivity . In our TOSA, the minimum reflection power without power penalty was below −30 dBm, and this minimum power could not guarantee stable operation in a real application. Therefore, we used the isolator. An erbium-doped fiber amplifier (EDFA) was utilized only for the optical eye measurement after 20-km transmission. We did not use any dispersion-compensation equipment and/or techniques during the experiment.
Figure 7 shows the BER curves of Back-to-Back (BtB) transmission without AWGs and after 20-km transmission with AWGs. We observed a negative power penalty of −2 dB and better optical eye after 20-km transmission. The negative penalty could be explained by the negative dispersion and chirp characteristics [23–25]. A pulse compression caused by factors such as increased overshooting and lowered crossing point of the optical eye were reported after transmission with negative dispersion . The shape of the optical eye measured before transmission in Fig. 7 is almost the same as that of the optical eye with negative dispersion. A negative power penalty was also reported under the presence of negative chirp in an electroabsorption modulator laser and in a chirp-managed directly modulated DFB laser [24, 25]. On the basis of these previously reported results and the measured optical eye, we inferred that the negative power penalty is caused by an effective negative dispersion due to a negative chirp characteristic of our VCSEL. However, further study is needed to confirm the reason. The extinction ratio is greater than 6.9 dB and the output power is greater than −3 dBm for all VCSELs. The sensitivity after transmission is −23 dBm at a BER of 10−12. A power budget of 20 dB including path penalty is sufficient for the C-RAN topology based on CPRI and LTE-A shown in Fig. 2.
We successfully demonstrated a cost-effective TOSA utilizing a monolithic four-wavelength VCSEL array with 100-GHz spacing and compact packaging based on TO-56-can. The 100-GHz wavelength spacing was achieved through selective etching of the cavity control layers and fine wavelength adjustment through current control. The operating current difference between four VCSELs is within 1.4 mA when the spacing is maintained. The VCSEL array shows stable operation performance without side-mode-related mode hopping. Error-free transmission over a 20-km SMF with two AWGs is achieved at the data rate of 6.144 Gb/s, which corresponds to CPRI option 6. No dispersion-compensation technique is employed during the transmission experiment. We believe that our proposed multi-wavelength TOSA based on a VCSEL array would be a strong candidate for a cost-effective light source in future-proof centralized BBU architecture.
This research was funded by the MSIP (Ministry of Science, ICT & Future Planning), Korea in the ICT R&D Program 2014. [14-000-05-002]
References and links
1. A. Ghosh, R. Ratasuk, B. Mondal, B. Managalvedhe, and T. Thomas, “LTE-advanced: next-generation wireless broadband technology,” IEEE Wireless Commun. 17(3), 10–22 (2010). [CrossRef]
2. S. M. Kim, “A study on the output power optimization of mobile WiMAX base station,” J. Kor. Inf. Commun. Soc. 35(4), 341–349 (2010).
3. G. Yuan, X. Zhang, W. Wang, and Y. Yang, “Carrier aggregation for LTE-Advanced mobile communication systems,” IEEE Commun. Mag. 48(2), 88–93 (2010). [CrossRef]
4. A. B. Ericsson, Huawei Technologies Co. Ltd, NEC Corporation, Alcatel Lucent, and Nokia Siemens Networks GmbH & Co. KG., “CPRI specification v6.0,” http://www.cpri.info/downloads/CPRI_v_6_0_2013-08-30.pdf.
5. F. Ponzini, L. Giorgi, A. Bianchi, and R. Sabella, “Centralized radio access networks over wavelength-division multiplexing: a plug-and-play implementation,” IEEE Commun. Mag. 51(9), 94–99 (2013). [CrossRef]
6. J. H. Lee, S.-H. Cho, K. H. Doo, S.-I. Myong, J. H. Lee, and S. S. Lee, “CPRI transceiver for mobile front-haul based on wavelength division multiplexing,” in Proceedings of the International Conference on ICT Convergence (ICTC), (2012), Paper VI-4.5. [CrossRef]
7. P. Chanclou, A. Pizzinat, F. Leclech, T. L. Reedeker, Y. Lagadec, F. Saliou, B. Leguyader, L. Guillo, Q. Deniel, S. Gosselin, S. D. Le, T. Diallo, R. Brenot, F. Lelarge, L. Marazzi, P. Parolari, M. Martinelli, S. O’dull, A. Gebrewold, D. Hillerkuss, J. Leuthold, G. Gavioli, and P. Galli, “Optical fiber solution for mobile fronthaul to achieve cloud radio access network,” in Proceedings of the Future Network and Mobile Summit (FutureNetworkSummit), (2013), pp. 1–11.
8. S.-H. Cho, J. H. Lee, K. H. Doo, S. I. Myong, J. H. Lee, and S. S. Lee, “Experimental investigations of RIN of spectrally sliced ASE injected RSOA in WDM-PON for mobile fronthauling applications based on CPRI line rate option 3,” in Proceedings of the International Conference on ICT Convergence (ICTC), (2012), Paper VI-4.3. [CrossRef]
11. W. Lee, S.-H. Cho, J. Park, B. K. Kim, and B. Kim, “Noise suppression of spectrum-sliced WDM-PON light sources using FP-LD,” ETRI J. 27(3), 334–336 (2005). [CrossRef]
12. C. J. Chang-Hasnain, M. W. Maeda, N. G. Stoffel, J. P. Harbison, L. T. Florez, and J. Jewell, “Surface emitting laser arrays with uniformly separated wavelengths,” Electron. Lett. 26(13), 940–942 (1990). [CrossRef]
13. H. Saito, I. Ogura, and Y. Sugimoto, “Uniform CW operation of multiple-wavelength vertical-cavity surface-emitting lasers fabricated by mask molecular beam epitaxy,” IEEE Photon. Technol. Lett. 8(9), 1118–1120 (1996). [CrossRef]
14. F. Koyama, T. Mukaihara, Y. Hayashi, N. Ohnoki, N. Hatori, and K. Iga, “Two-dimensional mutiwavelength surface emitting laser arrays fabricated by nonplanar MOCVD,” Electron. Lett. 30(23), 1947–1948 (1994). [CrossRef]
15. G. G. Ortiz, S. Q. Luong, S. Z. Sun, J. Cheng, H. Q. Hou, G. A. Vawter, and B. E. Hammons, “Monolithic, multiple-wavelength vertical-cavity surface-emitting laser arrays by surface-controlled MOCVD growth rate enhancement and reduction,” IEEE Photon. Technol. Lett. 9(8), 1069–1071 (1997).
16. W. Hofmann, E. Wong, G. Böhm, M. Ortsiefer, N. H. Zhu, and M. C. Amann, “1.55-μm VCSEL arrays for high-bandwidth WDM-PONs,” IEEE Photon. Technol. Lett. 20(4), 291–293 (2008). [CrossRef]
17. A. Karim, J. Piprek, P. Abraham, D. Lofgreen, Y.-J. Chiu, and J. E. Bowers, “1.55-μm vertical-cavity laser arrays for wavelength-division multiplexing,” IEEE J. Sel. Top. Quantum Electron. 7(2), 178–183 (2001). [CrossRef]
18. J.-H. Kim, B.-S. Yoo, J.-H. Shin, W.-S. Han, O.-K. Kwon, Y.-G. Ju, and H.-W. Song, “Fabrication method of densely spaced 1.55-μm multiple-wavelength vertical-cavity surface-emitting laser array structure for the application of dense wavelength division multiplexing,” Jap. J. Appl. Phys. 43(1), 137–139 (2004). [CrossRef]
19. H. J. Unold, S. W. Z. Mahmoud, R. Jager, M. Grabherr, R. Michalzik, and K. J. Ebeling, “Large-area single-mode VCSELs and the self-aligned surface relief,” IEEE J. Sel. Top. Quantum Electron. 7(2), 386–392 (2001). [CrossRef]
20. L. A. Coldren and S. W. Corzine, Diode Lasers and Photonic Integrated Circuits (John Willey & Sons, 1995), Chap. 2.
21. L. A. Coldren and S. W. Corzine, Diode Lasers and Photonic Integrated Circuits (John Willey & Sons, 1995), Chap. 5.
22. N. Nishiyama, C. Caneau, S. Tsuda, G. Guryanov, M. Hu, R. Bhat, and C.-E. Zah, “10-Gb/s error-free transmission under optical reflection using isolator-free 1.3-μm InP-based vertical-cavity surface-emitting lasers,” IEEE Photon. Technol. Lett. 17(8), 1605–1607 (2005). [CrossRef]
23. K. Sato, S. Kuwahara, and Y. Miyamoto, “Chirp characteristics of 40-Gb/s directly modulated distributed-feedback laser diodes,” J. Lightwave Technol. 23(11), 3790–3797 (2005). [CrossRef]
24. K. Yamada, K. Nakamura, Y. Matsui, T. Kunii, and Y. Ogawa, “Negative-chirp electroabsorption modulator using low-wavelength detuning,” IEEE Photon. Technol. Lett. 7(10), 1157–1158 (1995). [CrossRef]
25. J. Yu, Z. Jia, M.-F. Huang, M. Haris, P. N. Ji, T. Wang, and G.-K. Chang, “Applications of 40-Gb/s chirp-managed laser in access and metro networks,” J. Lightwave Technol. 27(3), 253–265 (2009).