Open Access
Add to BibSonomyAbstract
A novel concavely apodized (CA) distributed feedback (DFB) semiconductor laser was theoretically analyzed and experimentally demonstrated. The CA grating profile is equivalently realized by changing the duty cycle of the sampling structure along the cavity in the middle of which an equivalent phase shift is also inserted. Because the basic grating (seed grating) is uniform, only a common holographic exposure and a µm-level photolithography are required. Therefore, the fabrication cost is highly reduced compared with the true CA grating whose index modulation continuously changes along the cavity. The experimental results show that the laser has good single longitudinal mode operation.
©2013 Optical Society of America
1. Introduction
Good single longitudinal mode (SLM) operation of distributed feedback (DFB) semiconductor lasers is very important for the optical communication systems. The SLM property has been widely studied up to now. The well-known λ/4 phase shifted (QWS) (i.e., π phase shifted) DFB laser is usually used to solve this issue [1, 2]. But it may suffer fromserious spatial hole burning (SHB) which may cause some drawbacks such as the mode instability and the decreased linearity under direct modulation [3–5]. Therefore, some particular grating structures were proposed to suppress the SHB and improve some unique features of the lasers for some special applications [6, 7]. The concavely apodized (CA) grating structure is one of the solutions [8, 9]. Though it has been proposed and theoretically analyzed earlier, it is not easy to fabricate. The Electron beam lithography is a powerful method to write complex nano-structures, but it is time-consuming and expensive. Up to now, few experimental results have been reported [10, 11]. Recently the Reconstruction-Equivalent-Chirp (REC) technique shows powerful control ability for complex grating structures. Moreover, the fabrication is low-cost [12–14]. Based on REC technique, a novel CA DFB laser, which is fabricated only using a common holographic exposure and a conventional µm-level photo-lithography, can be achieved. We have briefly reported some experiment results in Ref [15]. In this paper, the proposed laser is theoretically and experimentally studied in detail. It maintains good side mode suppression ratio (SMSR) even when chip temperature is 55°C and injection current is 100mA. It may benefit the practical applications of this kind of lasers.
2. Principle
2.1 Realization of Apodization
For a sampled Bragg grating as shown in Fig. 1(a) and based on the Fourier analysis, it can be expressed as,
Fig. 1 (a) The schematic of the uniform sampling structure, (b) the one side apodized sampling structure by varying duty cycle.
Here P is the sampling period, Δns is the index modulation of the basic grating (seed grating), Λ0 is the seed grating period, Fm is the mth order Fourier coefficient of the sampling structure.
Usually the ± 1st order coefficient F ± 1 is usually used as the working sub-grating. The F ± 1 can be expressed as [15, 16],
Here γ is the duty cycle of sampling structure. It is defined as the ratio between the length with grating in one sampling period and the sampling period P. Fig. 2 shows the curve of the Eq. (2). F ± 1 is symmetric about the duty cycle of 0.5. So if the duty cycle varies along the cavity from 0 to 0.5 or 0.5 to 1.0, the apodization can be realized equivalently.
In order to verify the equivalent apodization and the influence on the light, we linearly change the duty cycle from 0.5 to 1.0 along the sampled grating as shown in Fig. 1(b). A parameter of apodization ratio Rapodization is defined here as Rapodization = Lapodization/L. Lapodization is the grating length with varied duty cycle and L is the whole grating length. The Transfer matrix method (TMM) is used to calculate the light intensity distribution along the grating.
As shown in Fig. 3, the light penetration in the grating increases when Rapodization increases. It is easy to understand that the apodization decreases the reflectivity of the light at the front of the grating and help light to transmit deeper. Therefore, if the concavely symmetric apodized grating is used as shown in Fig. 4, the effective cavity length increases accordingly. If a DFB laser uses this kind of grating structure as resonant cavity, light will distribute in larger region along cavity. Then, light intensity will be flattened and SHB can be suppressed.
Fig. 3 The light intensity distributions along the sampled grating with one side apodization anddifferent Rapodization equal to 1.0, 0.5, 0.0 respectively.
2.2 Realization of phase shift
For ensuring the SLM, the phase shift is usually inserted in the middle of the CA grating. For the REC technique, the equivalent π phase shift can be realized only by shifting half of the sampling period [12,13]. But for the equivalent apodized grating, the insert of the equivalent shift is rather different. Based on the Fourier analysis and according to Fig. 1(a), the mth orderFourier component ΔSm of the sampling structure with the duty cycle of γ can be expressed as,
Here, S(z) denotes the sampling structure. For simplicity, the −1st order sub-grating is used, then the phase characteristic of 1-exp(j2mπγ) in Eq. (3) can be explained in Fig. 5.
It shows that both of the amplitude and the phase of the −1st order sub-grating changes when the duty cycle γ changes. Particularly, when γ is equal to 0.5, Eq. (3) turns out to be –j/π and has the largest amplitude. It is consistent with the normally equivalent π phase shift. But if the equivalent apodization is applied, the duty cycle is continuously changed along the cavity. Because the phase of the grating changes with the duty cycle, it is hard to determine the phase of the whole equivalent apodized sampled grating. In order to insert a phase shift, the reflective phase is then analyzed by simulation. Assuming the grating is a mirror, we found that the reflective phase in the stop-band is from 0.0 radian to about 4.0 radian when the Rapodization is 0.5 as shown in Fig. 6(a). So if the symmetric structure is used as in Fig. 4, the resonant modes at 0.0 radian and π radian can be built. But because the reflectivity is much larger around the point of 0.0 radian than that of the π radian, a prominent resonant mode at 0.0 radian can be obtained. To further verify this point, we also simulated the transmission characteristic of the structure in Fig. 4. As shown in Fig. 6(b), a transmission peak corresponding to the reflective phase of around 0.0 radian in Fig. 6(a) is formed. Fortunately, we also found that the equivalent shift can always be obtained under different Rapodization. Therefore, a phase shift can also be equivalently inserted in such a symmetric structure.
Fig. 6 (a) The reflective spectrum and the phase response of the one side apodized grating as shown in Fig. 1(b), (b) the transmission spectrum of the symmetric concavely apodized sampled grating. Here, Rapodization is 0.5.
3. Simulation analysis
Some properties of the DFB laser with proposed apodized grating structure are simulated by TMM algorithm. All of the parameters are the same except the sampling pattern. We simulated the light intensity along laser cavity as shown in Fig. 7(a) with the same output power about 14mW but different Rapodization. The light intensity is flattened when Rapodization increases. The lasing spectra are also calculated. It shows that the side mode intensity for Rapodization equal to 0.5 is smaller than that of Rapodization equal to 0.0 when bias current is around 70mA as shown in Fig. 7(b). This is because the SHB of Rapodization equal to 0.5 is much less than that of Rapodization equal to 0.0, as shown in Fig. 7(a).
Fig. 7 (a) The simulated light intensity distributions along cavity around bias current of 70mA with the same output power of 14mW and Rapodization from 0.0 to 0.75, (b) the simulated lasing spectra with Rapodization equal to 0.0 and 0.5 respectively.
4. Fabrication and experimental results
The CA DFB laser is fabricated by a conventional two-stage lower-pressure metal-organic vapor phase epitaxy (MOVPE). An InP buffer layer, a lower optical confinement layer, a multiple quantum-well (MQW) active structure and an upper optical confinement layer are successively grown on a n-InP substrate in the first epitaxial growth. The MQW structure contains five 6nm thick 1.2% compressive-strain AlGaInAs wells separated by six 9nm thick −0.45% tensile-strain InGaAsP barriers. The sampled grating is then formed on the upper separate-confinement-heterostructure (SCH) layer by a conventional holographic exposure combined with a conventional photolithography. The cavity length is 375µm. The seed grating period is 234.7nm and the refractive index is around 3.2. The 0th order wavelength in the sampled grating (m = 0 in Eq. (1)) is 1502nm where the material gain is very small to support lasing. The −1st sub-grating is used as resonator whose Bragg wavelength is 1552nm. The sampling period is 6.65µm. Due to the limited precision of the photo-mask whose minimum linewidth is 1.0µm, the designed duty cycle to form the CA profile is from 0.85 to 0.5. The anti-reflection and high-reflection (AR/HR) facet coatings with their reflections of 2.0% and 90% respectively are used here for improving the power-current (P-I) slope efficiency. This sometimes will lead to an additional side mode lasing because of the random facet phase, but is doesn't affect any results of the suppression of SHB. Except the sampled grating, all of the other fabrication processes are the same as the normal DFB laser.
Figure 8 is the measured output P-I curve and the forward voltage-current (V-I) curve. The threshold current is 13.25mA and the slope efficiency is 0.24W/A at the operation temperature of 25°C. The insert is the microscope image of the fabricated DFB laser.
Fig. 8 The measured P-I curve and V-I curve at temperature of 25°C. The insert is the microscope image of the fabricated DFB laser.
The measured lasing spectrum at bias current of 100mA is shown in Fig. 9(a). The chip is uncooled. We also measured the lasing wavelength shift when bias current is from 20mA to 100mA. The lasing wavelength changes from 1552.02nm to 1555.17nm. Usually the wavelength shift ratio is 0.1nm/°C. So it shows that the temperature is nearly up to 55°C at 100mA. Though the bias current is large and the chip temperature is high, good SLM is obtained with side mode suppression ratio (SMSR) of 48.6dB. Due to the insert of the equivalent phase shift, the main oscillating mode is located within the stop-band as shown in the enlarged figure of Fig. 9(a). Good SLM operation is maintained when bias current changes from 20mA to100mA as shown in Fig. 9(b).
Fig. 9 (a) The measured spectrum of the equivalent CA DFB laser at bias current of 100mA, (b) the measured SMSRs when bias current changes from 20mA to 100mA.
Because the apodization ratio and the sampling period can be easily changed by sampling pattern, the different light intensity distribution and the lasing wavelength can be easily controlled and fabricated on the same wafer. Figure 10 is the lasing spectrum of the equivalent CA DFB laser with different lasing wavelength fabricated on the same laser bar. So lasers with different features can be easily achieved and even monolithically integrated for future high speed optical communications.
Fig. 10 The measured spectrum of the equivalent CA DFB laser at bias current of 100mA and the lasing wavelength is 1554.6nm.
5. Conclusion
A novel CA DFB semiconductor laser is theoretically analyzed and experimentally demonstrated. Good SLM operation is obtained even at large bias current and high chip temperature. Because both the phase shift and the apodization are equivalently realized by changing the sampling structure while the seed grating is uniform, the cost of the fabrication is low.
Acknowledgment
This research was supported by the National Nature Science Foundation of China under Grant 61090392 and 61090391, National “863” project under Grant 2011AA010300, the Fundamental Research Funds for the Central Universities and PAPD, Jiangsu Province, China.
References and links
1. K. Utaka, S. Akiba, K. Sakai, and Y. Matsushima, “Longitudinal-mode behaviour of λ/4-shifted InGaAsP/InP DFB lasers,” Electron. Lett. 21(9), 367–369 (1985). [CrossRef]
2. M. Usami, S. Akiba, and K. Utaka, “Asymmetric λ/4-shifted InGaAsP/InP DFB lasers,” IEEE J. Quantum Electron. 23(6), 815–821 (1987). [CrossRef]
3. J. E. A. Whiteaway, G. H. B. Thompson, A. J. Collar, and C. J. Armistead, “The design and assessment of λ/4 phase-shifted DFB laser structures,” IEEE J. Quantum Electron. 25(6), 1261–1279 (1989). [CrossRef]
4. A. J. Lowery and H. Olesen, “Dynamics of mode-instabilities in quarter-wave-shifted DFB semiconductor lasers,” Electron. Lett. 30(12), 965–967 (1994). [CrossRef]
5. J. Chen, R. J. Ram, and R. G. Helkey, “Linearity and third-order intermodulation distortion in DFB semiconductor lasers,” IEEE J. Quantum Electron. 35(8), 1231–1237 (1999). [CrossRef]
6. M. Okai, N. Chinone, H. Taira, and T. Harada, “Corrugation-pitch-modulated phase-shifted DFB laser,” IEEE Photon. Technol. Lett. 1(8), 200–201 (1989). [CrossRef]
7. G. P. Agrawal, J. E. Geusic, and P. J. Anthony, “Distributed feedback lasers with multiple phase-shift regions,” Appl. Phys. Lett. 53(3), 178–179 (1988). [CrossRef]
8. G. G. Morthier, K. David, P. Vankwikelberge, and R. G. E. Baets, “A new DFB-laser diode with reduced spatial hole burning,” IEEE Photon. Technol. Lett. 2(6), 388–390 (1990). [CrossRef]
9. G. G. Morthier and R. G. E. Baets, “Design of index-coupled DFB lasers with reduced longitudinal spatial hole burning,” J. Lightwave Technol. 9(10), 1305–1313 (1991). [CrossRef]
10. D. W. Wiesmann, C. David, R. Germann, D. Emi, and G.-L. Bona, “Apodized surface-corrugated gratings with varying duty cycle,” IEEE Photon. Technol. Lett. 12(6), 639–641 (2000). [CrossRef]
11. F. Girardin, G.-H. Duan, and T. Anna, “Modeling and measurement of spatial-hole-burning applied to amplitude modulated coupling distributed feedback lasers,” IEEE J. Quantum Electron. 31(5), 834–841 (1995). [CrossRef]
12. J. Li, H. Wang, X. Chen, Z. Yin, Y. Shi, Y. Lu, Y. Dai, and H. Zhu, “Experimental demonstration of distributed feedback semiconductor lasers based on reconstruction-equivalent-chirp technology,” Opt. Express 17(7), 5240–5245 (2009). [CrossRef] [PubMed]
13. Y. Shi, X. Chen, Y. Zhou, S. Li, L. Lu, R. Liu, and Y. Feng, “Experimental demonstration of eight-wavelength distributed feedback semiconductor laser array using equivalent phase shift,” Opt. Lett. 37(16), 3315–3317 (2012). [CrossRef] [PubMed]
14. Y. Shi, X. Chen, Y. Zhou, S. Li, L. Li, and Y. Feng, “Experimental demonstration of the three phase shifted DFB semiconductor laser based on Reconstruction-Equivalent-Chirp technique,” Opt. Express 20(16), 17374–17379 (2012). [CrossRef] [PubMed]
15. Y. Shi, R. Gu, and X. Chen, “A concave tapered DFB semiconductor laser based on reconstruction-equivalent-chirp technology”, Photonics Global Conference (PGC), 9882 (2010). [CrossRef]
16. Y. Shi, J. Li, L. Jia, S. Liu, and X. Chen, “An apodized DFB semiconductor laser realized by varying duty cycle of sampling Bragg grating and reconstruction-equivalent-chirp technology,” Opt. Commun. 283(9), 1840–1844 (2010). [CrossRef]
